article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
simplicial volumes are homotopy invariants for oriented closed connected manifolds , defined as the infimal weighted number of singular simplices needed to represent the fundamental class with respect to the given coefficients .
originally , gromov introduced the simplicial volume with efficients @xcite in his study of mostow rigidity . in the context of the connection of simplicial volume with betti numbers or @xmath1-betti numbers , integral versions of
simplicial volume play a central role @xcite .
we will consider the most basic integral version : the _ integral simplicial volume _ of an oriented closed connected @xmath2-manifold @xmath3 is defined as @xmath4 where @xmath5 for any chain @xmath6 in @xmath7 ( in reduced form ) .
integral simplicial volume lacks most of the convenient inheritance properties of ordinary simplicial volume ( such as multiplicativity under finite coverings and gluing formulae ) , and hence is rather difficult to control . in the following
, we will characterise manifolds with integral simplicial volume equal to @xmath0 , and derive an elementary non - computability statement for integral simplicial volume : [ thm : odd1 ] let @xmath8 and let @xmath3 be an oriented closed connected @xmath2-manifold .
then @xmath9 if and only if @xmath2 is odd and @xmath3 is homotopy equivalent to @xmath10 . in view of the positive solution of the poincar conjecture , the conditions in theorem [ thm : odd1 ] are also equivalent to @xmath11 .
[ cor : noncomp ] let @xmath12 be odd . 1 .
it is not decidable whether a given triangulated oriented closed connected @xmath2-manifold @xmath3 satisfies @xmath9 .
2 . in general , the integral simplicial volume of triangulated oriented closed connected @xmath2-manifolds is not computable .
the basic strategy of the proof of theorem [ thm : odd1 ] is as follows : * we show that any manifold of integral simplicial volume @xmath0 of dimension @xmath13 is simply connected .
this step is based on quotients of the standard simplex that model singular _ cycles _ consisting of a single simplex ( section [ sec : model ] ) . *
similar to the betti number estimates for integral simplicial volume ( * ? ?
* example 14.28 ) , we will use poincar duality and an explicit formula for the cap - product to prove that any manifold of integral simplicial volume equal to @xmath0 is an integral homology sphere ( lemma [ lem : oddhomology ] ) . *
finally , we apply the hurewicz and whitehead theorems to conclude .
more generally , this technique also gives a characterisation for multiples of the fundamental class ( theorem [ thm : odd1mult ] ) , which is interesting in the context of integral approximations of simplicial volume ( section [ subsec : intapprox ] ) .
we then derive the non - computability statement of corollary [ cor : noncomp ] from theorem [ thm : odd1 ] by comparison with the sphere - recognition problem .
alternatively , this result can also be obtained from weinberger s non - computability result for ordinary simplicial volume of homology spheres , which is based on far less elementary techniques ( * ? ? ?
* theorem 2 , p. 88 ) . more concretely , already
the case of integral simplicial volume equal to @xmath14 is more involved than theorem [ thm : odd1 ] : for example , in even dimensions , spheres are _ not _ the only oriented closed connected manifolds of integral simplicial volume equal to @xmath14 ( proposition [ prop : isvs1 ] ) .
moreover , the oriented closed connected @xmath15-manifolds with integral simplicial volume equal to @xmath14 are precisely ( up to homotopy / homeomorphism ) @xmath16 , @xmath17 , and @xmath18 ( proposition [ prop : isv32 ] , remark [ rem:32 ] ) . in section [ sec :
model ] , we study the combinatorics of singular cycles that consist of a single singular simplex . section [ sec:1 ] contains a proof of theorem [ thm : odd1 ] and a generalisation to multiples of the fundamental class ; moreover , we discuss the relation with integral approximations of simplicial volume . in section [ sec:2 ] , we discuss some aspects of the case that the integral simplicial volume equals @xmath14
. finally , in section [ sec : noncomp ] we briefly introduce the relevant setup from computability theory and derive corollary [ cor : noncomp ] .
singular cycles that consist of a single singular simplex have a very restricted combinatorial structure .
the corresponding model spaces for such cycles are simply connected ( proposition [ prop : modelpi1 ] ) , but fail to be closed manifolds in general ( proposition [ prop : modelmfd ] ) .
a model space associated with a cycle consisting of a single simplex is a standard simplex where the faces are glued according to their cancellation in the singular boundary .
let @xmath19 be odd .
* an _ @xmath2-matching _ is a bijection @xmath20 . * for @xmath21
, we denote the inclusion of the @xmath22-th face of @xmath23 by @xmath24 .
* let @xmath25 be an @xmath2-matching .
then the _ model space associated with @xmath25 _ is @xmath26 where @xmath27 is the equivalence relation generated by @xmath28 we denote the canonical projection @xmath29 by @xmath30 .
let @xmath19 be odd and let @xmath25 be an @xmath2-matching .
then @xmath31 where @xmath32 is the singular boundary operator .
hence , @xmath33 is a singular cycle .
we write @xmath34 \in h_n(m_\pi;{\mathbb{z}}).\ ] ] every singular class that is represented by a single singular simplex is a push - forward of one of these model classes : [ lem : cyclematching ] let @xmath35 be a topological space , let @xmath36 , and let @xmath37 be a singular simplex . 1 .
if @xmath38 is a ring with unit and @xmath39 in @xmath40 , then @xmath2 is odd .
if @xmath41 in @xmath42 , then @xmath2 is odd and there exists an @xmath25 and a continuous map @xmath43 such that @xmath44 \in h_n(x;{\mathbb{z}}).\ ] ] let @xmath38 be a ring with unit and let @xmath45 . because @xmath46 is an @xmath38-basis of @xmath47 , this sum has an even number of summands ; in particular , @xmath2 is odd .
we now consider the case @xmath48 . in order for these terms to cancel
there has to be a bijection @xmath49 such that @xmath50 holds for all @xmath51 .
hence , @xmath52 factors over the projection map @xmath53 , i.e. , there is a continuous map @xmath54 with @xmath55 . by construction , @xmath56 $ ] .
[ rem : cw ] let @xmath19 be odd and let @xmath25 be an @xmath2-matching .
the face lattice of @xmath23 induces a cw - structure on the quotient @xmath57 .
[ rem : susp ] let @xmath19 be odd , let @xmath25 be an @xmath2-matching , and let @xmath58 be the @xmath59-matching with @xmath60 then @xmath61 , where @xmath62 denotes the unreduced suspension . indeed , this can be seen as follows : let @xmath63 & \longrightarrow \delta^{n+2}\\ ( x , t ) & \longmapsto \begin{cases } ( 1-t ) \cdot x + t \cdot ( \frac14 \cdot e_{n+1 } + \frac34 \cdot e_{n+2 } ) & \text{if $ t \geq 0$}\\ ( 1-|t| ) \cdot x + |t| \cdot ( \frac34 \cdot e_{n+1 } + \frac14 \cdot e_{n+2 } ) & \text{if $ t < 0$}. \end{cases } \end{aligned}\ ] ] then it is not hard to show that the map @xmath64 ^ 2 & \longrightarrow m_{\pi ' } = \delta^{n+2 } / \sim_{\pi ' } \\ ( x , t , s ) & \longmapsto \begin{cases } ( 1-s ) \cdot \varphi ( x , t ) + s \cdot e_{n+2 } & \text{if $ t \geq 0 $ , $ s\geq 0$}\\ ( 1-|s| ) \cdot \varphi ( x , t ) + |s| \cdot e_{n+1 } & \text{if $ t \geq 0 $ , $ s < 0$}\\ ( 1-s ) \cdot \varphi ( x , t ) + s \cdot e_{n+1 } & \text{if $ t
< 0 $ , $ s\geq 0$}\\ ( 1-|s| ) \cdot \varphi ( x , t ) + |s| \cdot e_{n+2 } & \text{if $ t < 0 $ , $ s < 0 $ } \end{cases } \end{aligned}\ ] ] induces a well - defined homeomorphism @xmath65 ( figure [ fig : suspension ] ) .
there is only one @xmath0-matching , and clearly the corresponding model space is homeomorphic to @xmath66 . hence , by induction , for all odd @xmath19 the matching @xmath25 given by @xmath67 leads to the model space @xmath68
. however , not all model spaces are homotopy spheres ( proposition [ prop : modelmfd ] ) . # 1 # 1 circle ( 0.066 )
; we now show that these model spaces are simply connected : [ prop : modelpi1 ] let @xmath69 be odd , and let @xmath25 be an @xmath2-matching .
then @xmath70 is simply connected .
clearly , @xmath70 is connected .
we now use the cellular structure from remark [ rem : cw ] to compute the fundamental group of @xmath70 ( with respect to the basepoint given by the @xmath71-vertex of @xmath23 ) : as first step , we describe the _
@xmath71-skeleton _ of @xmath70 : if @xmath72 satisfying @xmath73 are matched through @xmath25 , then the @xmath74-faces @xmath75 \quad \text{and } \\
& [ 0,\dots , j-1 , j , \dots , k-2 , k -1 , k+1 \dots , n ] \end{aligned}\ ] ] of @xmath23 are glued and so give the same @xmath74-cell in @xmath76 . in particular , the vertices of @xmath23 numbered by @xmath77 all project to the same @xmath71-cell of @xmath70
let @xmath78 denote the equivalence relation on @xmath79 generated by @xmath80 then one can check easily for @xmath81 that the @xmath22-th and the @xmath82-vertex of @xmath23 lead to the same @xmath71-cell in @xmath70 if and only if @xmath83 . as next step , we describe a spanning tree for the _
@xmath0-skeleton _ of @xmath70 : using the above description of the @xmath71-skeleton , it is not hard to see that the @xmath0-cells of @xmath70 corresponding to the edges @xmath84 \bigm| \text{$j \in \{0,\dots , n\}$ is maximal in its $ \sim$-equivalence class } \bigr\}\ ] ] of @xmath23 is the set of edges of a spanning tree for the @xmath0-skeleton of @xmath70 .
@xmath14-skeleton _ of @xmath70 now gives us the presentation @xmath85 of the fundamental group , where @xmath86 } \bigm| j , k \in \{0,\dots , n\},\ j
< k\bigr\}\\ r : = & \;\{s_e \mid e \in t\}\\ & \;\cup \ { s_{[k,\ell ] } \cdot s_{[j , k ] } = s_{[j,\ell ] } \mid j , k,\ell \in \{0,\dots , n\},\ j < k < \ell\}\\ & \;\cup \ { s_{[j , k ] } = s_{[j',k ' ] } \mid j , k , j',k ' \in \{0,\dots , n\},\ j < k , j'<k ' ,
\ \sigma_\pi \circ [ j , k ] = \sigma_\pi \circ [ j',k ' ]
\}. \end{aligned}\ ] ] using the first and second type of relations , we immediately see that @xmath87 is generated by @xmath88 } \mid j \in \{0,\dots , n-1\},\ [ j , j+1 ] \not\in t \}$ ] .
so it suffices to show that these elements are trivial in @xmath87 .
let @xmath89 with @xmath90 \not\in t$ ] .
by construction of @xmath91 , this means that there exist @xmath92 with @xmath93 and the property that @xmath94 and @xmath95 are matched through @xmath25 .
we now consider the following cases : * let @xmath96 .
because @xmath94 and @xmath95 are matched through @xmath25 , we have @xmath97 = \sigma_\pi \circ [ j+1 , n],\ ] ] and so @xmath98 } = s_{[j+1,n]}$ ] in @xmath87 . then , the @xmath14-face @xmath99 $ ] of @xmath23 shows that @xmath100 } = s^{-1}_{[j+1,n ] } \cdot s_{[j , n ] } = 1\ ] ] in @xmath87 . *
the case of @xmath101 can be treated in the same way as the previous case .
* let @xmath102 and @xmath103 . because @xmath104 there is at least one other matching between some @xmath105 with @xmath106 through @xmath25 .
the same argument as in the first case yields @xmath107 } = 1 $ ] in @xmath87 . on the other hand
, @xmath108 gives @xmath109 } = s_{[1,2 ] } = \dots = s_{[n-1,n]}\ ] ] in @xmath87 .
in particular , @xmath110 } = s_{[j',j'+1 ] } = 1 $ ] in @xmath87 . hence , @xmath87 is the trivial group .
these model spaces are ( after sufficiently many barycentric subdivisions ) @xmath2-dimensional pseudo - manifolds .
however , in general , they do not have the homotopy type of oriented closed connected manifolds : [ prop : modelmfd ] 1 .
let @xmath111 , and let @xmath25 be an @xmath2-matching . then @xmath70 is homotopy equivalent to @xmath10 .
2 . for every odd @xmath12
there exists an @xmath2-matching @xmath25 such that the model space @xmath70 does _ not _ have the homotopy type of an oriented closed connected @xmath2-manifold .
in particular , @xmath112 .
_ ad 1 . _ the matchings @xmath113 and @xmath114 lead to @xmath66 and @xmath115 , respectively ( remark [ rem : susp ] ) .
it remains to consider the matching @xmath25 given by @xmath116 recall that any simply connected integral homology sphere is a homotopy sphere by the hurewicz and whitehead theorems . by proposition [ prop : modelpi1 ]
it is thus sufficient to compute @xmath117 using the cellular structure from remark [ rem : cw ] : a straightforward calculation shows that : * by proposition [ prop : modelpi1 ] , we have @xmath118 and @xmath119 . * the cellular chain complex of @xmath70 in degree @xmath14 has the @xmath14-cells corresponding to the @xmath14-faces @xmath120 $ ] and @xmath121 $ ] as a basis ; in degree @xmath0 , a basis is given by the @xmath0-faces @xmath122 , [ 0,3]$ ] , and with respect to these bases , the boundary map is represented by the matrix @xmath123 thus , there are no cellular cycles in degree @xmath14 and so @xmath124 . * because @xmath25 is a matching , the unique @xmath15-cell of @xmath70 is a cycle ; as @xmath70 has no cells in dimension at least @xmath125 , it follows that @xmath126 and @xmath127 for all @xmath128
. therefore , @xmath129 .
_ we first prove the statement in dimension @xmath130 : to this end , we consider the matching @xmath25 given by @xmath131 using the cellular structure from remark [ rem : cw ] one can calculate that @xmath132 - [ 0,1,3,4,5]$''}\\ { \mathbb{z } } & \text{if $ j=5 $ , generated by~``$[0,\dots,5]$''}\\ 0 & \text{if $ j \geq 6 $ } \end{cases } \end{aligned}\ ] ] for all @xmath133 .
hence , @xmath117 is not compatible with poincar duality , and so @xmath70 can not have the homotopy type of an oriented closed connected @xmath130-manifold .
let @xmath134 . taking successive trivial extensions of this matching @xmath25 as in remark [ rem : susp ] leads to model spaces homeomorphic to @xmath135 of dimension @xmath136 .
because @xmath137 is not compatible with poincar duality , these model spaces do not have the homotopy type of an oriented closed connected @xmath138-manifold .
in this section , we will prove theorem [ thm : odd1 ] and its generalisation to multiples of the fundamental class ( theorem [ thm : odd1mult ] ) .
we follow the strategy outlined in the introduction . in particular
, we will use the model spaces introduced in section [ sec : model ] and the following refinement of the classical betti number bound ( * ? ? ?
* example 14.28 ) : [ lem : oddhomology ] let @xmath19 be odd and let @xmath3 be an oriented closed connected fold . 1 .
let @xmath38 be a ring with unit , let @xmath139 be invertible in @xmath38 , and suppose that @xmath140 .
then for all @xmath141 we have @xmath142 2 . in particular :
if @xmath143 , then for all @xmath144 we have @xmath145 if @xmath35 is a space and @xmath146 , then @xmath147 for example , @xmath148 holds for all oriented closed connected manifolds @xmath3 .
it suffices to prove the first part .
because of @xmath149 there is a singular simplex @xmath150 that is a singular cycle and that satisfies @xmath151 = \pm m \cdot \fcl m \in h_n(m;{\mathbb{z}})$ ] ; without loss of generality , we may assume @xmath151 = m \cdot \fcl m$ ] . because @xmath152 is a unit in @xmath38 , the poincar duality homomorphism @xmath153 & \longmapsto ( -1)^{(n - k ) \cdot k } \cdot \bigl [ f(\bface { n - k } \sigma ) \cdot \fface k \sigma \bigr ] \end{aligned}\ ] ] with @xmath38-coefficients is an isomorphism for all @xmath154 .
hence , it is enough to show that @xmath155 is the zero map for all @xmath156 : if @xmath82 is even , then the @xmath82-simplex @xmath157 can not be cycle ( lemma [ lem : cyclematching ] ) , and so the fact that @xmath155 as described above is well - defined shows that @xmath158 . is a boundary , whence every cocycle vanishes on @xmath159 . ]
now let @xmath160 be odd and let @xmath25 be an @xmath2-matching for @xmath52 as in lemma [ lem : cyclematching ] .
we then define @xmath161 and @xmath162 .
moreover , for a sequence @xmath163 in @xmath164 we write @xmath165_\sigma \colon \delta^r & \longrightarrow m \\ ( t_0 , \dots , t_r ) & \longmapsto \sigma \biggl(\sum_{s=0}^r ( -1)^s \cdot t_s \cdot e_{j_s}\biggr ) \end{aligned}\ ] ] for the corresponding subsimplex of @xmath52 , where , @xmath166 denote the standard unit vectors of @xmath167 . for example , @xmath168_\sigma$ ] . with this notation and the relation @xmath169 for all @xmath170
, we obtain @xmath171_\sigma \hspace{-2 cm } & \\ & = \sum_{j=0}^{k+1 } ( -1)^{j } \cdot [ 0,\dots,\widehat j , \dots , k+1]_\sigma\\ & = \sum_{j \in i } [ 0,\dots,\widehat{\pi(j ) } , \dots , k+1]_\sigma - \sum_{j \in \pi(i ) } [ 0,\dots,\widehat j , \dots k+1]_\sigma + \sum_{j \in \overline i } ( -1)^j \cdot [ 0,\dots , k]_\sigma \\ & = 0 + \sum_{j \in \overline i } ( -1)^j \cdot \fface k \sigma .
\end{aligned}\ ] ] because @xmath82 is odd , also @xmath172 is odd ; moreover , @xmath25 matches even with odd indices , so the number of even elements and the number of odd elements in @xmath173 differ by @xmath0
. therefore , @xmath174_\sigma = \pm \fface k \sigma.\ ] ] thus , @xmath175 = 0 $ ] in @xmath176 , which implies @xmath177 .
let @xmath19 be odd , and let @xmath3 be an oriented closed connected @xmath2-manifold with @xmath178 . then @xmath179 .
moreover , @xmath180 for odd @xmath2 , as is witnessed by the cycle @xmath181 given by the canonical projection .
conversely , let @xmath8 and @xmath3 be an oriented closed connected @xmath2-manifold that satisfies @xmath143
. then @xmath2 is odd by lemma [ lem : cyclematching ] .
because @xmath66 is the only oriented closed connected fold we may assume @xmath104 .
we follow the strategy outlined in the introduction : the manifold @xmath3 is simply connected : in view of lemma [ lem : cyclematching ] there is an @xmath2-matching @xmath25 and a continuous map @xmath182 such that @xmath183 because @xmath70 is simply connected , @xmath184 lifts to a map @xmath185 to the universal covering @xmath186 of @xmath3 .
thus , @xmath187 and so @xmath188 .
in particular , @xmath189 is compact and @xmath190 is a finite covering .
transfer shows that @xmath191 , where @xmath192 is the number of sheets of @xmath25 .
therefore , @xmath193 , and so @xmath194 is simply connected .
we have @xmath195 by lemma [ lem : oddhomology ] . hence , @xmath178 : any simply connected integral homology @xmath2-sphere is homotopy equivalent to @xmath10 by the hurewicz and whitehead theorems .
more generally , this technique also gives the following characterisation for multiples of the fundamental class : [ thm : odd1mult ] let @xmath19 be odd and let @xmath3 be an oriented closed connected @xmath2-manifold
. then the following are equivalent : 1 .
there exists @xmath196 with @xmath197 .
the manifold @xmath3 is dominated by @xmath10 , i.e. , there is a map @xmath198 of non - zero degree .
there is a subsequence of @xmath199 that is bounded by @xmath0 .
the implication `` 3 @xmath200 1 '' is clear . for the implication ``
2 @xmath200 3 '' we use self - maps @xmath201 of arbitrarily high degree , the fact that @xmath180 and that @xmath202 holds for all continuous maps @xmath203 and all @xmath146
. we will now prove `` 1 @xmath200 2 '' : without loss of generality we may assume that @xmath204 . let @xmath205 with @xmath206 .
we will now proceed as in the proof of theorem [ thm : odd1 ] : * reduction to the simply connected case : because @xmath207 , lemma [ lem : cyclematching ] implies that there is an @xmath2-matching @xmath25 and a continuous map @xmath208 satisfying @xmath209 the model space @xmath70 is simply connected ( proposition [ prop : modelpi1 ] ) ; hence , @xmath184 lifts to a map @xmath210 to the universal covering @xmath189 of @xmath3 . because @xmath211 it follows that @xmath212 , and so @xmath189 is compact and @xmath213 is a non - zero multiple of @xmath214 with integral @xmath215-norm equal to @xmath0 .
* we have @xmath216 by lemma [ lem : oddhomology ] and the first step .
* existence of a non - zero degree map @xmath217 : combining the previous steps with the rational hurewicz theorem ( * ? ? ?
* theorem 8.6 ) implies that the rational hurewicz homomorphism @xmath218 _ * & \longmapsto h_n(g;{\mathbb{z}})(\fcl { s^n } ) \otimes 1 \end{aligned}\ ] ] is surjective , and any ( pointed ) continuous map @xmath219 that is mapped via @xmath220 to an element of @xmath221 has non - zero degree .
composing @xmath222 with the universal covering map @xmath223 hence shows that @xmath3 is dominated by @xmath10 . determining the exact relation between ordinary simplicial volume and @xmath1-betti numbers of oriented closed connected aspherical manifolds is a long - standing open problem @xcite .
estimates between betti numbers and simplicial volumes can be obtained in the presence of some integrality of coefficients @xcite , based on poincar duality and a counting argument as in lemma [ lem : oddhomology ] .
hence , we are led to the question in which sense ordinary simplicial volume can be approximated by integral versions of simplicial volume and how these approximations relate to betti numbers and @xmath1-betti numbers . on the one hand , one can use the action of the fundamental group to introduce more flexibility in the coefficients .
this leads to stable integral simplicial volume ( based on the integral simplicial volumes of all finite coverings ) and integral foliated simplicial volume ( based on probability spaces with an action of the fundamental group ) @xcite .
in contrast to @xmath1-betti numbers that can be computed by the corresponding integral approximations @xcite , stable integral simplicial volume and integral foliated simplicial volume of oriented closed connected hyperbolic manifolds of dimension at least @xmath125 does _ not _ coincide with ordinary simplicial volume @xcite . on the other hand , ordinary simplicial volume of an oriented closed connected manifold @xmath3 coincides with ( * ? ?
* remark 5.4 ) @xmath224 therefore , one can use the sequence @xmath225 to introduce refined versions of vanishing of simplicial volume .
the strongest such vanishing occurs in case the sequence @xmath226 contains a subsequence bounded by @xmath0 .
this is exactly the situation of theorem [ thm : odd1mult ] , which gives a complete geometric characterisation of such manifolds . more generally , the property that @xmath226 contains a bounded subsequence is related to weak flexibility of homology classes ( * ?
* theorem 5.1 , remark 5.4 ) .
however , it is not clear how @xmath226 is related to @xmath1-betti numbers in the aspherical case .
in the previous section , we gave a complete classification of manifolds with integral simplicial volume equal to @xmath0 .
the natural next step is to look at manifolds with integral simplicial volume equal to @xmath14 . which oriented closed connected manifolds @xmath3 satisfy @xmath227 ?
however , this problem seems to be much harder than the case of integral simplicial volume equal to @xmath0 . in the following
, we will give some partial answers .
we begin with a small , but helpful , observation , generalising the parity aspect of lemma [ lem : cyclematching ] : [ lem : parity ] let @xmath3 be an oriented closed connected manifold of even dimension .
then @xmath228 is even .
let @xmath229 and let @xmath230 be an integral fundamental cycle . because @xmath23 has an odd number of @xmath74-faces and because the singular @xmath2-simplices of @xmath3 form a @xmath231-basis of @xmath232 , the definition of the boundary operator yields that @xmath233 is even
. then also @xmath234 is even and thus @xmath228 is even .
we start with the simple and well - known case of dimension @xmath14 : ( # 1,#2)(#1,#2 ) circle ( 0.1 ) ; # 1#1 circle ( 0.1 ) ; ( 0,2 ) ( 0,0 ) ( 0,2 ) ( 2,2 ) ; ( 0,0 ) ( 2,0 ) ( 2,2 ) cycle ; ( 0,0 )
( 2,2 ) cycle ; ( 0,0 ) ( 2,0 ) ( 2,2 ) ( 0,0 ) node[anchor = east ] @xmath71 ; ( 2,0 ) node[anchor = west ] @xmath0 ; ( 2,2 ) node[anchor = west ] @xmath14 ; ( 1.33,0.66 ) node @xmath52 ; ( 2,0 ) ( 0,0 ) ( 2,0 ) ( 2,2 ) ; ( 0,0 ) ( 0,2 ) ( 2,2 ) cycle ; ( 0,0 ) ( 0,2 ) ( 2,2 ) cycle ; ( 0,0 ) ( 0,2 ) ( 2,2 ) ( 0,0 ) node[anchor = east ] @xmath71 ; ( 0,2 ) node[anchor = east ] @xmath0 ; ( 2,2 ) node[anchor = west ] @xmath14 ; ( 0.66,1.33 ) node @xmath235 ; [ prop : surfaces ] 1 .
we have @xmath236 .
2 . if @xmath237 is an oriented closed connected surface of genus @xmath238 , then @xmath239 . in even dimension , there is no cycle consisting of a single simplex ( lemma [ lem : cyclematching ] ) .
thus , oriented closed connected surfaces @xmath237 satisfy @xmath240 .
clearly , @xmath241 and @xmath242 ; the latter can be seen from the fundamental cycle depicted in figure [ fig : toruscycle ] .
let @xmath237 be an oriented closed connected surface of genus @xmath243 .
then there is an explicit triangulation of the regular @xmath244-gon that shows that @xmath245 @xcite . on the other hand
, we have @xmath246 no fundamental cycle on @xmath237 realises the optimal value @xmath247 because no straight singular simplex on @xmath248 has maximal volume @xcite . therefore , @xmath249 .
because @xmath250 is even ( lemma [ lem : parity ] ) , we obtain the desired estimate @xmath251 . as a next step
, we consider @xmath15-manifolds .
[ prop : isv32 ] the following oriented closed connected @xmath15-manifolds @xmath3 satisfy @xmath252 : @xmath253 in view of theorem [ thm : odd1 ] , all of these manifolds have integral simplicial volume at least @xmath14 .
_ ad @xmath16 : _ by proposition [ prop : isvs1 ] below , @xmath254 .
_ ad @xmath17 : _ we describe an integral fundamental cycle for @xmath17 that consists of two singular simplices : we view @xmath17 as quotient of the @xmath15-ball @xmath255 modulo the antipodal action on @xmath256 .
we consider the singular @xmath15-simplices @xmath257 on @xmath17 as depicted in figure [ fig : rp3 ] .
then @xmath258 and looking at the local degree shows that @xmath259 represents @xmath260 .
( 0,0 ) arc ( 180:360:1.5 ) ; ( 0,0 ) ( 1,-0.5 ) ( 3,0 ) ( 3,0 ) arc ( 0:180:1.5 ) cycle ; ( 0,0 ) node[anchor = east ] @xmath14 ; ( 1,-0.5 ) node[anchor = north ] @xmath0 ; ( 2,0.5 ) node[anchor = south ] @xmath71 ; ( 3,0 ) node[anchor = west ] @xmath15 ; ( 3,0 ) arc ( 0:180:1.5 ) ; ( 0,0 )
( 1,-0.5 )
( 2,0.5 ) cycle ; ( 1,-0.5 ) ( 2,0.5 ) ; ( 0,0 ) ; ( 3,0 ) ; ( 1,-0.5 ) ; ( 2,0.5 ) ; ( 1.5,-2.2 ) node[anchor = south ] @xmath52 ; ( 3,0 ) arc ( 0:180:1.5 ) ; ( 3,0 ) ( 2,0.5 ) ( 0,0 ) ( 0,0 ) arc ( 180:360:1.5 ) cycle ; ( 0,0 ) node[anchor = east ] @xmath15 ; ( 1,-0.5 ) node[anchor = north ] @xmath0 ; ( 2,0.5 ) node[anchor = south ] @xmath71 ; ( 3,0 ) node[anchor = west ] @xmath14 ; ( 0,0 ) arc ( 180:360:1.5 ) ; ( 0,0 ) ( 1,-0.5 )
( 2,0.5 ) cycle ; ( 1,-0.5 ) ( 2,0.5 ) ; ( 0,0 ) ; ( 3,0 ) ; ( 1,-0.5 ) ; ( 2,0.5 ) ; ( 1.5,-2.2 ) node[anchor = south ] @xmath235 ; _ ad @xmath18 : _ the construction of a fundamental cycle of @xmath18 consisting of two singular simplices is a little bit more delicate : we view the lense space @xmath18 as quotient of the @xmath15-ball @xmath255 , where points on the southern hemisphere of @xmath256 are identified with points on the northern hemisphere after reflection at the equator and rotating around the north - south axis by @xmath261 .
we consider the two singular @xmath15-simplices @xmath262 on @xmath18 as depicted in figure [ fig : l31 ] . by construction , @xmath263 and
looking at the local degree ( inside the ball ) shows that @xmath264 represents @xmath265 .
we now manipulate the local degree of @xmath266 in order to obtain a fundamental cycle of @xmath18 : we may construct @xmath266 in such a way that @xmath266 on the middle slice @xmath267 of @xmath268 is the obvious parametrisation of the equatorial disk in @xmath255 .
let @xmath269 be the @xmath270-shaped pillow obtained from @xmath271 $ ] by collapsing the vertical intervals in @xmath272 $ ] . passing to the universal covering of
@xmath18 one can easily construct a continuous map @xmath273 that coincides with @xmath274 on the top and the bottom of @xmath269 and that has local degree @xmath275 .
we then glue @xmath276 into @xmath266 at @xmath237 ; because gluing @xmath269 into @xmath268 at @xmath237 yields a space that is homeomorphic to @xmath268 ( while preserving the boundary @xmath277 ) , we thus obtain a singular simplex @xmath278 of local degree @xmath279 with @xmath280 . then @xmath281 and looking at the local degree shows that the cycle @xmath282 represents @xmath283 .
the lense space @xmath18 admits an obvious deomposition into two tetrahedra ; however , it is not hard to see that there is _ no _ affine parametrisation of these tetrahedra that yields an integral fundamental cycle of @xmath18 .
( 1.5,0 ) circle ( 1.5 ) ; ( 1.5,0 ) circle ( 1.5 ) ; ( 1.5,0 ) circle ( 1.5 and 0.5 ) ; ; ; ; ; ( 240:1.5 and 0.5 )
( 240:1.5 and 0.9 ) ; ( 120:1.5 and 0.5 ) arc ( 120:240:1.45 and 0.7 ) ; ( 240:1.5 and 0.9 ) arc ( 240:360:1.5 and 0.8 ) ; ( 240:1.5 and 0.5 ) node[anchor = south ] @xmath15 ; ( 240:1.5 and 0.9 ) node[anchor = north ] @xmath71 ; ( 120:1.5 and 0.5 ) node[anchor = south ] @xmath0 ; ( 0:1.5 and 0.5 ) node[anchor = west ] @xmath14 ; ( 1.5,-2.2 ) node[anchor = south ] @xmath52 ; ( 1.5,0 ) circle ( 1.5 ) ; ( 1.5,0 ) circle ( 1.5 ) ; ( 1.5,0 ) circle ( 1.5 and 0.5 ) ; ; ; ; ; ( 240:1.5 and 0.5 ) ( 240:1.5 and 0.9 ) ; ( 120:1.5 and 0.5 ) arc ( 120:240:1.45 and 0.7 ) ; ( 240:1.5 and 0.9 ) arc ( 240:360:1.5 and 0.8 ) ; ( 240:1.5 and 0.5 ) node[anchor = south ] @xmath71 ; ( 240:1.5 and 0.9 ) node[anchor = north ] @xmath15 ; ( 120:1.5 and 0.5 ) node[anchor = south ] @xmath14 ; ( 0:1.5 and 0.5 ) node[anchor = west ] @xmath0 ; ( 1.5,-2.2 ) node[anchor = south ] @xmath266 ; [ rem:32 ] if @xmath3 is an oriented closed connected @xmath15-manifold that is _ not _ homotopy equivalent / homeomorphic to @xmath16 , @xmath17 or @xmath18 , then we have @xmath284 : the model space construction of section [ sec : model ] can be extended to the case of singular cycles that consist of more than one simplex .
if @xmath3 has a fundamental cycle @xmath285 consisting of two singular simplices @xmath286 , then ( depending on the signs of the coefficients ) there are two types of matchings to be considered : * if @xmath287 ( or @xmath288 , then one obtains a bijection @xmath289 where the sets @xmath290 and @xmath291 correspond to the faces of @xmath52 and @xmath235 respectively . *
if @xmath292 , then one obtains a bijection @xmath293 similarly to the proof of proposition [ prop : modelpi1 ] , a lengthy calculation shows that the ( connected components of the ) corresponding model spaces constructed out of two copies of @xmath268 whose facets are glued according to these matchings have cyclic fundamental group .
moreover , we obtain ( e.g. , by computer calculations ) the integral homology groups of these model spaces in table [ table : computation32plus ] and table [ table : computation32minus ] . llllllll & & & & + & & & & 0 & 1 & 2 & 3 + + llllllll & & & & + & & & & 0 & 1 & 2 & 3 + + hence , all these model spaces have first integral homology group isomorphic to @xmath294 because these model spaces have cyclic ( in particular , abelian ) fundamental group , these are also the only possible isomorphism types of fundamental groups of such model spaces .
let @xmath25 be the matching corresponding to the considered fundamental cycle @xmath285 of @xmath3 , let @xmath295 be the associated model space , let @xmath296 be the model class and let @xmath297 be the continuous map modelling @xmath285 .
then @xmath298 , and @xmath299 ^ -{h_k(f;{\mathbb{z } } ) } & h_k(m;{\mathbb{z } } ) \\ h^{3-k}(n_\pi;{\mathbb{z } } ) \ar[u]^{\args\cap \beta_\pi } & h^{3-k}(m;{\mathbb{z } } ) \ar[l]^-{h^{3-k}(f;{\mathbb{z } } ) } \ar[u]^{\cong}_{\args \cap \fcl m } } \end{aligned}\ ] ] is commutative for all @xmath134 in view of the naturality of the cap - product .
the right vertical arrow is an isomorphism by poincar duality .
hence , @xmath300 is split surjective .
moreover , covering theory shows that @xmath301 is surjective . because @xmath302 is cyclic
, also @xmath303 is cyclic and thus isomorphic to @xmath304 .
as we have seen above , @xmath305 is split surjective , and so @xmath306 is _ split _ surjective .
therefore , the fundamental group of @xmath3 is isomorphic to @xmath294 now the classification of oriented closed connected @xmath15-manifolds @xcite shows that @xmath3 is homotopy equivalent ( even homeomorphic ) to @xmath115 , @xmath17 , @xmath18 , or @xmath16 . as we know that @xmath307 the only remaining candidates for @xmath3 are @xmath17 , @xmath18 , or @xmath16 , as claimed .
the example of the torus generalises as follows : [ prop : isvs1 ] for all @xmath308 we have @xmath309 theorem [ thm : odd1 ] shows that @xmath310 because @xmath311 .
therefore , it suffices to construct a fundamental cycle @xmath312 of @xmath313 consisting of two singular simplices @xmath314 .
we now describe the construction of such singular simplices ; to this end we construct the two factors of these maps separately .
the @xmath66-factors are defined on @xmath315 by @xmath316 in order to understand the definition of the @xmath317-factors we first give the corresponding matchings : * if @xmath2 is even , then we consider the matching @xmath25 given by @xmath318 moreover , we define @xmath319 * if @xmath2 is odd , then we consider the matching @xmath25 given by @xmath320 moreover , we define @xmath321 we choose @xmath322 and a homeomorphism @xmath323 that maps the point associated with @xmath324 to @xmath325 .
we denote the constant map with value @xmath325 by @xmath326 and the projection corresponding to the above homeomorphism by @xmath327 . for @xmath328
we then define @xmath329 because @xmath330 , these maps can be extended to a continuous map @xmath331 with @xmath332 for all @xmath328 . in the same way , we construct @xmath333 using @xmath334 . finally , we set @xmath335 because the @xmath336 and @xmath334 are compatible with the matching @xmath25 , the faces of @xmath337 and @xmath338 cancel according to @xmath25 . also @xmath339 and @xmath340
are compatible with the matching @xmath25 because for all @xmath341 we have @xmath342 , and so @xmath343 etc .
thus , @xmath312 is a singular cycle on @xmath344 .
it remains to show that @xmath312 represents @xmath345 . to this end
we choose singular cocycles @xmath346 and @xmath347 dual to @xmath348 and @xmath349 respectively ; without loss of generality , we may assume that @xmath350 .
then the cohomological cross - product @xmath351 is a cocycle on @xmath344 that is dual to @xmath352 .
therefore , it suffices to show that @xmath353 . using the explicit description of the cohomological cross - product via the alexander - whitney map , we obtain @xmath354 therefore , @xmath312 is a fundamental cycle of @xmath344 .
in particular , in even dimensions spheres are not the only oriented closed connected manifolds with integral simplicial volume equal to @xmath14 .
we now turn our attention to products of higher - dimensional spheres . in certain cases
, we can show that these have integral simplicial volume larger than @xmath14 .
however , the general case is still wide open . [
lem : betti2 ]
let @xmath3 be an oriented closed connected @xmath2-manifold with @xmath252 .
then for all @xmath134 we have : 1 . if @xmath82 is even , then @xmath355 is cyclic .
if @xmath82 is odd , then @xmath355 can be generated by two elements .
because of @xmath227 , there exist singular simplices @xmath356 such that @xmath357 or @xmath358 is a cycle in @xmath232 representing @xmath359 . without loss of generality
, we may assume that @xmath360 = \fcl m \in h_n(m;{\mathbb{z}})$ ] where @xmath361 .
it suffices to treat the case @xmath141 . as in the proof of lemma [ lem : oddhomology ]
, we consider the poincar duality isomorphism @xmath362 & \longmapsto ( -1)^{(n - k ) \cdot k } \cdot \bigl [ f(\bface { n - k } \sigma ) \cdot \fface k \sigma\\ & \phantom{\longmapsto ( -1)^{(n - k ) \cdot k}\quad } + \varepsilon \cdot f(\bface { n - k } \tau ) \cdot \fface k \tau\bigr ] \end{aligned}\ ] ] clearly , the image of @xmath155 can be generated by two elements or less .
if @xmath82 is even , then neither @xmath157 nor @xmath363 is a cycle ( lemma [ lem : cyclematching ] ) ; hence , there is ( up to integral rescaling ) at most one @xmath231-linear combination of @xmath157 and @xmath364 that is a cycle .
thus , the image of @xmath155 is cyclic . because @xmath155 is an isomorphism and hence surjective
, the result follows .
in general , these bounds can not be improved as the example of the torus in dimension @xmath14 shows ( proposition [ prop : surfaces ] ) .
let @xmath19 be even .
then @xmath365 on the one hand , @xmath366 is even ( lemma [ lem : parity ] ) . on the other hand , by lemma [ lem : betti2 ]
, we have @xmath367 because @xmath368 and @xmath2 is assumed to be even .
[ rem : s2s4 ] we have @xmath369 indeed , @xmath370 is even by lemma [ lem : parity ] .
because this integral simplicial volume is non - zero it suffices to prove that @xmath371 . as in remark [ rem:32 ]
we consider model spaces for cycles consisting of two singular @xmath372-simplices .
all such cycles are modelled by matchings that are bijections @xmath373 calculating the homology of the model spaces of all possible matchings ( via a computer ) shows that they all have second integral homology that does not surject onto @xmath231 . on the other hand ,
if @xmath370 were equal to @xmath14 , then the cap - product argument of remark [ rem:32 ] shows that the model map on second integral homology would need to be surjective .
this shows that @xmath371 .
if @xmath3 and @xmath374 are oriented closed connected manifolds , then the ordinary simplicial volume ( with efficients ) satisfies @xcite @xmath375 the upper estimate is based on the homological cross - product and works in the same way also for the integral simplicial volume .
but the lower estimate depends on the hahn - banach theorem , and the argument does not carry over to the integral setting .
hence , one might ask whether also @xmath376 holds .
we will now briefly introduce the relevant setup for computability and derive the non - computability statement of corollary [ cor : noncomp ] from theorem [ thm : odd1 ] : for @xmath19 we let @xmath377 be the countable set of all finite @xmath2-dimensional simplicial complexes whose vertices form a subset of @xmath378 and we let @xmath379 be the subset of all such complexes whose geometric realisation is an oriented closed connected topological @xmath2-manifold .
we now consider an injection @xmath380 with the property that for every @xmath381 the number @xmath382 can be calculated by integer arithmetic from the set of simplices , and conversely such that for every @xmath383 the complex @xmath95 with @xmath384 can be reconstructed by integer arithmetic .
for example , such a map @xmath385 can be obtained by considering suitable products of prime numbers . using @xmath385
, we can interpret notions from computability on @xmath379 : a function @xmath386 is _ computable _ [ resp .
_ semi - computable _ ] , if the associated function @xmath387 is a partial recursive [ resp .
_ @xmath388-recursive _ ] function @xmath389 with domain @xmath390 .
a subset @xmath391 is called _ decidable _ [ resp .
_ semi - decidable _ ] if there is a computable [ resp .
semi - computable ] function @xmath386 satisfying @xmath392 and @xmath393 . for the exact definition of ( @xmath388-)recursive functions and the relation with turing machines we refer to the literature @xcite . for all @xmath19 the function
@xmath399 is semi - computable in the following sense : if @xmath400 , then the set @xmath401 is semi - decidable : let @xmath402 .
if @xmath403 is a cycle , then the simplicial approximation theorem and an inductive construction of ( relative ) homotopies shows that there is a @xmath404 and a singular chain @xmath405 with the following properties : * we have @xmath406 = [ c']$ ] in @xmath407 and @xmath408 . *
the chain @xmath409 is a _ generalised simplicial chain _ , i.e. , it consists of singular simplices @xmath410 that are geometric realisations of simplicial maps from the @xmath82-fold barycentric subdivision of @xmath23 ( viewed as simplicial complex in the standard way ) to @xmath95 .
moreover , all the following operations can be performed by turing machines : enumerate all barycentric subdivisions of @xmath23 , enumerate all simplicial maps between two finite simplicial complexes , check whether a generalised simplicial @xmath231-chain on a finite simplicial complex in @xmath379 is a fundamental cycle , and compute the @xmath215-norm of a generalised simplicial singular @xmath231-chain .
hence , there is a turing machine acting on inputs @xmath95 from @xmath379 such that : if @xmath411 , then the turing machine stops and accepts @xmath95 .
if @xmath412 , then the turing machine does not stop or , if it stops , does not accept @xmath95 .
in other words , the set @xmath413 is semi - decidable .
the proof uses models of singular cycles , similar to the ones discussed in the previous sections .
a more detailed description of such combinatorial models of cycles and their associated homology classes can be found in the discussion of weak flexibility ( * ? ? ?
* section 5.1 ) .
let @xmath19 . then there exist only finitely many different combinatorial types of singular @xmath2-cycles with @xmath231-coefficients and @xmath215-norm at most @xmath152 .
these finitely many combinatorial types can be combined into a finite connected simplicial complex @xmath35 and a class @xmath415 with the following property : if @xmath3 is an oriented closed connected @xmath2-manifold with @xmath414 , then there is a continuous map @xmath416 satisfying @xmath417 now let @xmath418 ; in particular , @xmath419 .
then by thom s work @xcite on representing homology classes through manifolds , there is an oriented closed connected @xmath2-manifold @xmath374 and a continuous map @xmath420 with @xmath421 composing these maps , we see that @xmath374 dominates every oriented closed connected @xmath15-manifold @xmath3 with @xmath414 through a map of degree @xmath0
. however , every oriented closed connected @xmath15-manifold can dominate only finitely many other homeomorphism types of oriented closed connected @xmath15-manifolds @xcite .
m. gromov .
_ metric structures for riemannian and non - riemannian spaces _ with appendices by m. katz , p. pansu , and s. semmes , translated by s. m. bates . _ progress in mathematics
_ , 152 , birkhuser , 1999 .
h. j. munkholm .
simplices of maximal volume in hyperbolic space , gromov s norm , and gromov s proof of mostow s rigidity theorem ( following thurston ) . in _ topology symposium , siegen 1979 _ , _ lecture notes in mathematics _ , 788 , pp . 109124 , springer , 1980 . | integral simplicial volume is a homotopy invariant of oriented closed connected manifolds , defined as the minimal weighted number of singular simplices needed to represent the fundamental class with integral coefficients .
we show that odd - dimensional spheres are the only manifolds with integral simplicial volume equal to @xmath0 .
consequently , we obtain an elementary proof that , in general , the integral simplicial volume of ( triangulated ) manifolds is _ not _ computable . |
a heffter array @xmath0 is an @xmath1 matrix with nonzero entries from @xmath2 such that 1 .
each row contains @xmath3 filled cells and each column contains @xmath4 filled cells , 2 . the elements in every row and column sum to 0 in @xmath2 , and 3 .
for every @xmath12 , either @xmath13 or @xmath14 appears in the array .
the notion of a heffter array @xmath0 was first defined by archdeacon in @xcite .
it is shown there that a heffter array with a pair of special orderings can be used to construct an embedding of the complete graph @xmath15 on a surface .
the connection is given in the following theorem .
[ heffter1]@xcite given a heffter array @xmath0 with compatible orderings @xmath16 of the symbols in the rows of the array and @xmath17 on the symbols in the columns of the array , then there exists an embedding of @xmath15 such that every edge is on a face of size @xmath3 and a face of size @xmath4 .
moreover , if @xmath16 and @xmath17 are both simple , then all faces are simple cycles .
we will not concern ourselves with the ordering problem in this paper and will just concentrate on the construction of the heffter arrays .
( in a subsequent paper ( see @xcite ) we will address the ordering problem in more detail . ) we refer the reader to @xcite for the definition of a simple ordering and the definition of compatible orderings .
theorem [ heffter1 ] can be stated in design theoretic terms .
a @xmath8cycle system on @xmath18 points is a decomposition of the edges of @xmath19 into @xmath8cycles .
a @xmath8cycle system @xmath20 on @xmath19 is _ cyclic _ if there is a labeling of the vertex set of @xmath19 with the elements of @xmath21 such that the permutation @xmath22 preserves the cycles of @xmath20 .
a biembedding of an @xmath7cycle system and a @xmath8cycle system is a face 2-colorable topological embedding of the complete graphs @xmath15 in which one color class is comprised of the cycles in the @xmath7cycle system and the other class contains the cycles in the @xmath8cycle system .
we refer the reader to @xcite for general information on @xmath8cycle systems , to @xcite for cyclic @xmath8cycle systems and to @xcite for information about biembedings of cycle systems .
while not using heffter arrays , @xcite and @xcite gives a related idea of embedding a single 2-fold triple system .
the following proposition about cycle systems follows from theorem [ heffter1 ] .
[ heffter - cycle ] assume there exists a heffter array @xmath0 with compatible orderings @xmath16 of the symbols in the rows of the array and @xmath17 on the symbols in the columns of the array such that @xmath16 and @xmath17 are both simple .
then there exists a biembedding of a cyclic @xmath7cycle system @xmath23 and a cyclic @xmath8cycle system @xmath24 both on @xmath25 points .
a heffter array is called an _ integer _
heffter array if condition 2 in the definition of heffter array above is strengthened so that the elements in every row and every column sum to zero in @xmath11 . in this paper
we will concentrate on constructing _ square integer _ heffter arrays . if the heffter array is square , then @xmath26 and necessarily @xmath27 .
so for the remainder of this paper define a heffter array @xmath28 to be an @xmath29 array of integers satisfying the following : 1 .
each row and each column contains @xmath30 filled cells , 2 . the symbols in every row and every column sum to 0 in @xmath11 , and 3 . for every element @xmath31 either @xmath13 or @xmath14 appears in the array . in @xcite the authors study the case where the heffter array has no empty cells . in that paper
it is shown that there is an _ integer _
@xmath32 if and only if @xmath33 and @xmath34 ( mod 4 ) and in general that there is an @xmath32 for all @xmath33 . a _ shiftable _ heffter array @xmath35 is defined to be a heffter array @xmath0 where every row and every column contain the same number of positive and negative numbers . from @xcite we have that there is an @xmath36 if and only if @xmath37 and @xmath18 are even .
the notation @xmath38 denotes a _
@xmath39 . in this paper
we extend the idea of shiftable to any array .
an @xmath40 array @xmath41 of integers ( possibly with empty cells ) is _ shiftable _ if each row and each column contains the same number of positive and negative numbers .
let @xmath41 be a shiftable array and @xmath13 a nonnegative integer .
if @xmath13 is added to each positive element and @xmath14 is added to each negative element , then all of the row and column sums remain unchanged .
let @xmath42 denote the array where @xmath13 is added to all the _ positive _ entries in @xmath41 and @xmath14 is added to all the _ negative _ entries .
if @xmath41 is an integer array , define the _ support _ of @xmath41 as the set containing the absolute value of the elements contained in @xmath41 .
so if @xmath41 is shiftable with support @xmath23 and @xmath13 a nonnegative integer , then @xmath43 has the same row and column sums as @xmath41 and has support @xmath44 . in the case of a shiftable heffter array @xmath45 , the array @xmath46 will have row and column sums equal to zero and support @xmath47 .
the following lemma gives necessary conditions for the existence of @xmath28 and @xmath45 .
[ necessary ] if there exists an @xmath28 , then necessarily @xmath48 ( mod 4 ) .
furthermore , if there exists an @xmath45 , then necessarily @xmath30 is even and @xmath49 ( mod 4 ) . given an @xmath28 , in order for each row to sum to zero , each row must contain an even number of odd numbers .
hence the entire array contains an even number of odd numbers .
now , the support of the @xmath28 is the set @xmath50 .
there will be an even number of odd numbers in @xmath23 exactly when @xmath48 ( mod 4 ) .
if there is an @xmath45 , then clearly @xmath30 must be even in order to have the same number of positive and negative entries in each row .
it follows that @xmath51 ( mod 4 ) , hence @xmath49 ( mod 4 ) . we give some examples of heffter arrays @xmath28 and @xmath45 .
[ ex1 ] the following are @xmath52 , @xmath53 and @xmath54 respectively .
-9 & 3 & 6 & + & -11 & 1 & 10 + 5 & & -7 & 2 + + 1 & & 18 & -9 & -10 + -6 & 2 & & 19 & -15 + -11&-12&3&&20 + 16&-7&-13&4 & + + & & -14&16&10&-12 + -1&3&&&17&-19 + 2&-4&&&-18&20 + 21&-23&-5&7 & & + -22&24&6&-8 & & + + -8&2&&&&27&-21 + -15&-16&3&&&&28 + 22&-9&-17&4 & & & + & 23&-10&-18&5 & & + & & 24&-11&-19&6 & + & & & 25&-12&-20&7 + + 3&-4&-6&8&11&-12 & & + & & -13&17&14&-19&-21&22 + & & 15&-16&-18&20&23&-24 + -33&34&&&-25&29&26&-31 + 35&-36&&&27&-28&-30&32 + 38&-43&-45&46&&&-37&41 + -42&44&47&-48&&&39&-40 + + in this paper we will prove the existence of heffter arrays @xmath28 for many of the possible values of @xmath18 and @xmath30 .
the paper is organized as follows . in section [ section2 ]
we cover the case when @xmath30 is even and in section [ section3 ] we deal with the case when @xmath55 ( mod 4 ) . in both of these cases
we prove that the necessary conditions from lemma [ necessary ] are sufficient . in section [ section4 ]
we consider the case when @xmath56 ( mod 4 ) . in this case
we do not get a complete solution , however we construct @xmath28 for many values of @xmath18 and @xmath30 .
the results are summarized in section [ section5 ] .
we give two direct constructions which both yield shiftable heffter arrays with an even number @xmath30 of cells per row and column . from lemma [ necessary ] , if @xmath57 ( mod 4 ) , then @xmath18 must be even , while if @xmath58 ( mod 4 ) , then any @xmath59 is possible .
we will construct heffter arrays in both of these cases .
the first construction covers the case when @xmath18 is even and the second one covers the case when @xmath18 is odd .
[ th1 ] there exists an @xmath45 for all @xmath60 even with @xmath61 . for @xmath62 let @xmath63 , @xmath64 and @xmath65 be @xmath66 arrays defined by @xmath67 [ cols="^,^",options="header " , ] & dne&dne +
much of this research was done at the university of queensland while the second and fourth authors were visiting ( at separate times ) .
we thank the ethel raybould visiting fellowship and the turkish science foundation 2219 program for providing funds for these visits .
m.j . grannell and t.s .
designs and topology . in surveys in combinatorics 2007 ,
a. hilton and j. talbot , eds .
london mathematical society lecture note series , 346 .
cambridge , uk : cambridge university press , ( 2007 ) , 121 @xmath68 174 . | a heffter array @xmath0 is an @xmath1 matrix with nonzero entries from @xmath2 such that _ i _ ) each row contains @xmath3 filled cells and each column contains @xmath4 filled cells , _ ii _ ) every row and column sum to 0 , and _
iii _ ) no element from @xmath5 appears twice .
heffter arrays are useful in embedding the complete graph @xmath6 on an orientable surface where the embedding has the property that each edge borders exactly one @xmath7cycle and one @xmath8cycle .
archdeacon , boothby and dinitz proved that these arrays can be constructed in the case when @xmath9 , i.e every cell is filled . in this paper
we concentrate on square arrays with empty cells where every row sum and every column sum is @xmath10 in @xmath11 .
we solve most of the instances of this case .
_ dedicated to the memory of our friend and colleague scott vanstone .
_ + 05b30 , 05c10 |
the fe / mg abundance ratio , one of the most important parameters for estimating the age of a stellar system after its initial starburst , is based on the delay of iron enrichment by type ia supernovae and compared with the formation of such @xmath4-elements as magnesium by type ii supernovae ( see hamann & ferland 1999 for a review ) . the delay of iron formation was generally thought to be @xmath5 gyr ( e.g. , yoshii et al . 1996 ) , while more recent studies suggest that this delay can be @xmath6 gyr , depending on environmental conditions of the star formation process @xcite .
the ( uv)/ emission - line ratio of qsos is considered a probable indicator of the fe / mg abundance ratio , which has been measured in various redshift ranges : @xmath7 by ( * ? ? ?
* hereafter the99 ) , @xmath8 by ( * ? ? ?
* hereafter paper i ) , @xmath9 by @xcite , @xmath10 by @xcite , and @xmath11 by @xcite .
the median values of the / emission - line ratios are almost constant at all redshifts , while large scatter dominates the distribution of the ratios at @xmath2 , indicating the diversity of the formation histories of high - redshift qsos . on the other hand ,
recent photoionization calculations of fe@xmath12 atoms in the broad - line regions of qsos show that the ( uv)/ emission - line ratio is sensitive not only to fe and mg abundance but also to physical conditions , including microturbulence @xcite .
however , the median value of the / emission - line ratio is expected to converge on a significantly smaller value when the age of the universe becomes less than @xmath6 1 gyr , assuming that the substantial scatter caused by the diverse characteristics ( or microturbulence velocities , etc . ) of the broad - line regions in qsos has the same contribution at any redshift . in this paper , we report the / emission - line ratios of four qsos at @xmath0 observed by the oh - airglow suppressor ( ohs ; iwamuro et al . 2001 ) and cisco @xcite mounted on the subaru telescope .
in @xmath13 we analyze the data by the same method as in paper i , and then we compare the results with other observations in @xmath14 . in @xmath15
we discuss the expected fe / mg abundance evolution of qsos using the combined samples of these results , the new sloan digital sky survey ( sdss ) archival data , and the previous samples in paper i. throughout the paper we adopt a cosmology with @xmath16 km s@xmath17 mpc@xmath17 , @xmath18 , and @xmath19 .
the observations were carried out on 2002 february 27 and march 2 and again on 2003 march 20 and 21 , using ohs and cisco mounted on the subaru telescope . since the and are redshifted to the @xmath20 and @xmath21 bands , respectively , for @xmath0 qsos , the @xmath21-band spectra were obtained by cisco and the @xmath22-band spectra by ohs separately in each observation run .
the typical exposure sequence is 200 s @xmath23 3 @xmath23 4 positions ( cisco ) or 1000 s @xmath23 4 positions ( ohs ) , in which the object is moved about @xmath24 along the slit by nodding the telescope after every three exposures ( cisco ) or one exposure ( ohs ) .
the slit widths were @xmath25 ( cisco ) and @xmath26 ( ohs ) , providing spectral resolutions of 330 and 210 , respectively .
after this exposure sequence , a nearby sao star with a spectral type of a or f was observed as a spectroscopic standard to remove the telluric atmosphere absorption features and to correct the instrumental response .
the pixel scale was @xmath27/pixel with the infrared secondary mirror of the telescope , while the seeing size was @xmath28 . to calibrate the relative flux between the @xmath21- and @xmath22-band spectra , short imaging exposures of 20 s @xmath23 3 @xmath23 4 positions
were executed by cisco in both the @xmath29 ( 1.972.30 @xmath30 m ) and @xmath20 ( 1.501.79 @xmath30 m ) bands , except for sdss 1030 + 0524 .
the observation log is summarized in table 1 .
the obtained data were reduced using iraf with typical reduction procedures for infrared images ( see paper i ) .
the resultant one - dimensional spectra in the @xmath21 and @xmath22 bands were combined using the photometric results of the imaging observation by cisco . as a result , through a circular aperture of 22 diameter , the average fluxes between 1.972.30 and 1.501.79 @xmath30 m correspond to the @xmath29- and @xmath20-band magnitudes respectively .
the extraction procedure for and emission lines from the combined spectra is the same as in paper i : nonlinear @xmath31 fitting for the ( moffat function ) and for the continuum ( power law ) profile with six free parameters , after subtraction of the template from the composite qso spectrum of the large bright quasar survey ( lbqs ; francis et al .
1991 ) and the balmer continuum template from the uv spectrum of 3c 273 @xcite with various multiplying factors ( see eq .
( 1)(4 ) in paper i ) .
the fitting range is 21503300 , which is also the wavelength range for our definition of flux .
note that the strength of the emission is determined by shape comparison between our template and the observed spectral features ( large bumps and small depressions ) , which does not require the assumption of the power - law continuum level ( we need the assumption of the shape of the template instead ) .
although the moffat function is very useful for fitting the various profiles of emission lines with the minimum number of parameters , a serious problem is that the fitted profile sometimes has very broad wings . to prevent such unexpected broad wings , we set the lower limits at @xmath32 for the moffat power ( eq . ( 1 ) in paper i ) and reanalyzed all the data fitted with @xmath33 in paper i ( see appendix ) .
the object spectra with fitted components are shown in figure 1 , and the numerical results are listed in table 2 .
the spectra with wavelengths longer than 1.80 @xmath30 m are obtained by cisco , and the shorter parts by ohs . the strength of the emission is mainly determined by the break feature at rest 2200 , which is observed by ohs with superior signal - to - noise ratios .
the wavelength range between 1.80 and 1.95 @xmath30 m corresponds to the atmospheric absorption band , whose contribution to the fitting results is very small because of large errors .
as shown in figure 1 and table 2 , we detected strong emission from sdss j1148 + 5251 , which is consistent with @xcite .
sdss j1048 + 4637 also shows significant emission , while the / ratio is not as much as the value of @xmath348.1 reported by @xcite .
the absorption - line feature at 1.29 @xmath30 m in sdss j1048 + 4637 is the broad absorption line component of ] reported by @xcite .
the observed spectra of these objects fit well into our template spectrum of in every detail .
although @xcite reported a smaller / ratio of @xmath344.7 for sdss j1148 + 5251 , this inconsistency is mostly caused by the difference in the template under emission .
on the other hand , sdss j1030 + 0524 and sdss j1630 + 4012 show almost no emission or any break feature at rest 2200 .
these results for sdss j1030 + 0524 are consistent with freudling et al.(2003 ) , as well as the unknown absorption - line feature at 1.57 @xmath30 m . as a result ,
even at @xmath0 , the ( uv)/ emission - line ratios still show large diversity .
figure 2 shows the / ratio and the absolute ab magnitude at rest 2500 as a function of redshift , for which we reanalyzed the sdss data on the basis of the dr1 quasar catalog @xcite .
the other @xmath1 data are the same as those plotted in figure 7 of paper i , except for four objects listed in the appendix .
the median values and the standard deviation of the sample distribution are shown by the squares with the cross bars in this figure whose numerical values are listed in table 3 . although the number of samples at @xmath2 is insufficient to make a statistical argument , we can deduce the trends of the evolution of the / ratio from figure 2 .
first , the median / ratios at @xmath2 are almost constant , while the upper envelope of the distribution plotted by the dashed line in figure 2 shows a probable declination toward high redshift .
second , the median distribution settles into lower ratios at @xmath3 , with small scatter compared to the other redshift . here ,
the downward error bars from these median points are almost equal to the typical fitting errors of the samples in each bin ( column 3 in table 3 ) , while the longer upward error bars are affected by the substantial scatter of the / ratios .
assuming that the fe / mg abundance ratio affects these rough trends , these small / ratios at @xmath3 originated with the dilution of the fe abundance with the outflow gas from low - mass stars after @xmath343 gyr from the initial starburst @xcite .
the scatter of the sample at this redshift ( a factor of @xmath342 between the top and bottom of the cross bars ) corresponds to the maximum contribution of the scatter caused by the differences in the physical conditions of the broad - line regions in high - redshift qsos @xcite .
if this substantial scatter of factor of @xmath342 is the universal nature of the qsos , independent of the redshift , the larger scatter of the @xmath2 samples are affected by the difference in the fe / mg abundance ratio or the time passage after the initial starburst .
this idea is also consistent with the declination of the upper envelope of the / distribution toward high redshift , which is expected to be the initial abundance evolution of qsos .
an alternative explanation for the variation of the / ratios at @xmath35 may be their luminosity dependence such that the qsos of higher luminosities , in a low - redshift ( @xmath360.1 - 0.6 ) sample from _ hubble space telescope _ ( @xmath37 ) uv archives @xcite as well as in a high - redshift ( @xmath363 - 5 ) sample from near - infrared observations @xcite , are found to have higher / ratios .
the larger variation of the / ratios for the brighter qsos suggests that either the ionizing photon flux arriving at the broad - line - emitting clouds or the physical characteristics of the clouds would scatter more significantly for brighter qsos . to estimate their respective contributions to
the / ratio is a theoretical challenge that requires an elaborate photoionization model of fe@xmath12 in the broad - line regions of qsos .
we thank the canadian astronomy data centre , which is operated by the herzberg institute of astrophysics , national research council of canada .
we would like to express our thanks to the members of the sdss project .
funding for the sdss has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the us department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org .
the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , the university of pittsburgh , princeton university , the united states naval observatory , and the university of washington .
this work has been supported by a grant - in - aid for scientific research ( a ) , japan ( 14204016 ) .
this work is also supported by a grant - in - aid for the 21st century coe `` center for diversity and universality in physics , '' and in part by a grant - in - aid for center - of - excellence research ( 07ce2002 ) of the ministry of education , culture , sports , science , and technology ( mext ) of japan .
we re - analyzed all the data having broad wings ( @xmath33 ) in paper i with lower limits at @xmath32 for the moffat profile of @xmath38^{-b}.\ ] ] all the data were fitted with @xmath39 ( lorentzian ) , and the results are listed in table 4 .
three out of seven objects listed in table 4 are not plotted in figure 2 because we adopted the results of the ohs data rather than those of the99 and the results of the @xmath37 fos archive rather than those of @xcite ( see table 2 and table 4 in paper i ) .
barth , a. j. , martini , p. , nelson , c. h. , & ho , l. c. 2003 , , 594 , l95 dietrich , m. , hamann , f. , appenzeller , i. , & vestergaard , m. 2003 , , 596 , 817 fan , x. et al .
2001 , , 122 , 2833 fan , x. et al .
2003 , , 125 , 1649 francis , p. j. , hewett , p. c. , foltz , c. b. , chaffee , f. h. , weymann , r. j. , & morris , s. l. 1991 , , 373 , 465 freudling , w. , corbin , m. r. , & korista , k. t. 2003 , , 587 , l67 friaca , a. c. s. & terlevich , r. j. 1998 , , 298 , 399 hamann , f. & ferland , g. 1999 , , 37 , 487 iwamuro , f. , motohara , k. , maihara , t. , hata , r. , & harashima , t. 2001 , , 53 , 355 iwamuro , f. , motohara , k. , maihara , t. , kimura , m. , yoshii , y. , & doi , m. 2002 , , 565 , 63 kinney , a. l. , bohlin , r. c. , blades , j. c. , & york , d. g. 1991 , , 75 , 645 maiolino , r. , juarez , y. , mujica , r. , nagar , n. m. , & oliva , e. 2003 , , 596 , l155 maiolino , r. , oliva , e. , ghinassi , f. , pedani , m. , mannucci , f. , mujica , r. , & juarez , y. 2003 , submitted , ( astro - ph/0312402 ) matteucci , f. & recchi , s. 2001 , , 558 , 351 motohara , k. et al .
2002 , , 54 , 315 schneider , d. p. et al .
2003 , , 126 , 2579 thompson , k. l. , hill , g. j. , & elston , r. 1999 , , 515 , 487 tsuzuki , y. 2004 , ph.d .
thesis , university of tokyo verner , e. , bruhweiler , f. , verner , d. , johansson , s. , & gull , t. 2003 , , 592 , l59 wills , b. j. , netzer , h. , & wills , d. 1985 , , 288 , 94 yoshii , y. , tsujimoto , t. , & nomoto , k. 1996 , , 462 , 266 yoshii , y. , tsujimoto , t. , & kawara , k. 1998 , , 507 , l113 lcccccccc & 6.28 & 17.67 & & 2002 feb.27 & cisco & 2400 & 080 & 1 + & & & 18.57 & 2002 mar .
2 & ohs & 4000 & 088 & + & 6.23 & 17.12 & 17.76 & 2003 mar.20 & cisco & 2400 & 056 & 2 + & & & 17.83 & 2003 mar.21 & ohs & 4000 & 071 & + & 6.43 & 16.98 & 17.70 & 2003 mar.20 & cisco & 2400 & 065 & 2 + & & & 17.62 & 2003 mar.21 & ohs & 4000 & 062 & + & 6.05 & 18.40 & 19.25 & 2003 mar.20 & cisco & 2400 & 071 & 2 + & & & 19.18 & 2003 mar.21 & ohs & 8000 & 068 & + lccccccc & 6.311 & 3590 & 5.96@xmath40 & [email protected] & 0.99@xmath42 & 21.0 & 22.9 + & 6.203 & 4050 & [email protected] & [email protected] & [email protected] & 81.8 & 25.7 + & 6.403 & 3020 & 137@xmath4116 & [email protected] & [email protected] & 333 & 41.7 + & 6.065 & 2690 & [email protected] & [email protected] & [email protected] & 52.2 & 55.8 + cccc [email protected] & 4.48@xmath43 & 0.84 & 30 + [email protected] & 3.36@xmath44 & 0.82 & 1101 + [email protected] & 3.49@xmath45 & 0.90 & 1101 + [email protected] & 3.34@xmath46 & 0.90 & 1101 + [email protected] & 3.05@xmath47 & 0.92 & 1101 + [email protected] & 2.78@xmath48 & 0.93 & 1101 + [email protected] & 2.79@xmath49 & 0.81 & 1101 + [email protected] & 2.81@xmath50 & 0.84 & 1101 + [email protected] & 3.00@xmath51 & 0.95 & 1101 + [email protected] & 3.75@xmath52 & 1.24 & 1101 + [email protected] & 5.11@xmath53 & 1.67 & 1099 + [email protected] & 5.50@xmath54 & 1.45 & 6 + [email protected] & 4.40@xmath55 & 1.76 & 9 + [email protected] & 5.69@xmath56 & 2.70 & 8 + [email protected] & 2.38@xmath57 & 2.02 & 4 + lcccccc & 4.527 & 4050 & [email protected] & 98.5 & 46.6 & the99 + & 3.288 & 4610 & [email protected] & 110 & 46.1 & the99 + & 3.290 & 3430 & [email protected] & 71.3 & 29.1 & the99 + & 0.089 & 3020 & [email protected] & 155 & 79.1 & @xmath58 + & 0.065 & 2320 & [email protected] & 238 & 40.4 & @xmath37 + & 0.109 & 3010 & [email protected] & 138 & 40.9 & @xmath37 + & 0.064 & 4200 & [email protected] & 424 & 143 & @xmath58 + | near - infrared spectra of four qsos located at @xmath0 are obtained with the oh - airglow suppressor mounted on the subaru telescope
. the / emission - line ratios of these qsos are examined by the same fitting algorithm as in our previous study of @xmath1 qsos .
the fitting results show that two out of the four @xmath0 qsos have significant emission in their rest - uv spectra , while the other two have almost no features .
we also applied our fitting algorithm to more than 10,000 sdss qsos and found two trends in the distribution of / against redshift : ( 1 ) the upper envelope of the / distribution at @xmath2 shows a probable declination toward high redshift , and ( 2 ) the median distribution settles into lower ratios at @xmath3 with small scatter compared to the other redshift .
we discuss an fe / mg abundance evolution of qsos with a substantial contribution from the diverse nature of the broad - line regions in high - redshift qsos . |
we examined a series of high quality single crystals of hg-1201 ranging from deeply underdoped to optimal doping with @xmath83 s of 55 k , 65 k , 71 k , 79 k , 81 k , 91 k , and 94 k. these samples were grown by the self - flux method @xcite and annealed to achieve different oxygen concentrations @xcite , with @xmath83 determined within @xmath92 k by magnetic susceptibility measurements .
hole concentration was determined using methods described in ref .
samples were polished under nitrogen flow using 0.3 @xmath93 m grit films to prevent surface oxidation .
the measurements reported herein were performed using 100 fs pulses from a mode - locked ti : sapphire laser at 800 nm center wavelength and @xmath94j@xmath95@xmath96 fluence .
pump pulses induced a small ( @xmath97 ) fractional change in reflectivity that was monitored by time - delayed probe pulses . at the low laser fluence
used in this study , samples are weakly perturbed by the pump pulse , in the sense that the density of photogenerated quasiparticles is much less than the thermal equilibrium value . in this regime
the decay rate is a measure of the quasiparticle recombination rate in thermal equilibrium .
measurements in zero magnetic field at berkeley were performed under vacuum in an oxford continuous - flow liquid he cryostat .
magneto - optic measurements at berkeley were performed in a 6 t oxford spectramag cryostat , and those at mit in a 7 t janis cryostat .
the mean - field hamiltonian is given by @xmath98 @xcite , with @xmath99 diagonalizing the cdw and sc subsystems via successive bogoliubov transformations yields energy eigenvalues @xmath100 .
the eigenvectors give the composite qp operators @xmath101 , where @xmath102 in the above , @xmath38 and @xmath103 are assumed to be real , and the phases of @xmath39 and @xmath9 are determined by the phases of @xmath42 and @xmath44 , respectively .
we calculate the coherence factors following ref .
@xcite , starting with a quasiparticle interaction of the form @xmath104 .
we assume that the interaction is spin - independent .
the four four electronic transitions that involve the same qp operators , @xmath105 must be summed coherently before calculating the squared modulus of the matrix element that determines the quasiparticle scattering rate . expressing the four bilinear operators in terms of the quasiparticle ( @xmath106 ) operators yields , @xmath107 where , @xmath108 and @xmath109 .
summing over recombination channels and then squaring yields a coherence factor for quasiparticle recombination given by @xmath110 .
2 tranquada jm , sternlieb bj , axe jd , nakamura y , uchida s ( 1995 ) evidence for stripe correlations of spins and holes in copper oxide superconductors _ , nature _ 375 : 561 - 563 .
hoffman j _
( 2002 ) a four unit cell periodic pattern of quasi - particle states surrounding vortex cores in bi@xmath111sr@xmath111cacu@xmath111o@xmath112 _ science _
, 295 : 466 - 469 .
howald c , eisaki h , kaneko n , kapitulnik a ( 2003 ) coexistence of periodic modulation of quasiparticle states and superconductivity in bi@xmath111sr@xmath111cacu@xmath111o@xmath112 _ proceedings of the national academy of sciences _ 100 : 9705 - 9709 .
vershinin michael _
( 2004 ) local ordering in the pseudogap state of the high - tc superconductor bi@xmath111sr@xmath111cacu@xmath111o@xmath112 _ science _ 303:1995 - 1998 .
( 2008 ) charge - density - wave origin of cuprate checkerboard visualized by scanning tunnelling microscopy _ nature physics _
4:696 - 699 .
wu tao , _ et al . _
( 2011 ) magnetic - field - induced charge - stripe order in the high - temperature superconductor yba@xmath111cu@xmath1133o@xmath114 _ nature _ 477 : 191 - 194 .
ghiringhelli g _
( 2012 ) long - range incommensurate charge fluctuations in ( y , nd ) ba@xmath111cu@xmath113o@xmath6 _ science _ 337 : 821 - 825 .
chang j ( 2012 ) direct observation of competition between superconductivity and charge density wave order in yba@xmath1112cu@xmath113o@xmath115 _ nature physics _ 8 : 871 - 876 .
le tacon m _
( 2014 ) inelastic x - ray scattering in yba@xmath1112cu@xmath113o@xmath116 reveals giant phonon anomalies and elastic central peak due to charge - density - wave formation _ nature physics _ 10:52 - 58 .
( 2014 ) charge order driven by fermi - arc instability in bi@xmath111sr@xmath117la@xmath118cuo@xmath119 _ science _ 343 : 390 - 392 . da silva neto eh _ et al . _
( 2014 ) ubiquitous interplay between charge ordering and high - temperature superconductivity in cuprates _ science _ 343 : 393 - 396 .
et al . _ , ( 2014 ) charge order and its connection with fermi - liquid charge transport in a pristine high - tc cuprate _
_ 5 : 5875 .
leboeuf d _
_ , ( 2013 ) thermodynamic phase diagram of static charge order in underdoped yba@xmath111cu@xmath113o@xmath114 _ nat .
_ 9 : 79 - 83 .
efetov kb , meier h , ppin ( 2013 ) pseudogap state near a quantum critical point _
_ 9 : 442 - 446 .
sachdev s , la placa r ( 2013 ) bond order in two - dimensional metals with antiferromagnetic exchange interactions _ phys .
_ 111 : 027202 .
davis jc , and lee dh ( 2013 ) concepts relating magnetic interactions , intertwined electronic orders , and strongly correlated superconductivity _ proc .
110 : 17623 - 17630 .
hayward lauren e , hawthorn david g , melko roger g , sachdev subir ( 2014 ) angular fluctuations of a multicomponent order describe the pseudogap of yba2cu3o6 + x _ science _ 343 : 1336 - 1339 .
wang y , chubukov ( 2014 ) a charge - density - wave order with momentum ( 2 q , 0 ) and ( 0 , 2 q ) within the spin - fermion model : continuous and discrete symmetry breaking , preemptive composite order , and relation to pseudogap in hole - doped cuprates _ phys . rev .
b _ 90 : 035143 .
schrieffer jr ( 1964 ) _ theory of superconductivity _
benjamin , new york ) .
kohn w , sherrington d , ( 1970 ) two kinds of bosons and bose condensates _ rev .
_ 42:1 .
averitt r.a . and talor a.j .
( 2002 ) ultrafast optical and far - infrared quasiparticle dynamics in correlated electron materials _ j. phys . condens . matter _ 14
: r1357 demsar j _ et al . _
( 1999 ) superconducting gap @xmath120 , the pseudogap @xmath121 , and pair fluctuations above @xmath0 in overdoped y@xmath122ca@xmath118ba@xmath111cu@xmath113o@xmath123 from femtosecond time - domain spectroscopy _ phys .
_ 82 : 4918 .
han sg , vardeny zv , wong ks , symko og , koren g ( 1990 ) , femtosecond optical detection of quasiparticle dynamics in high - tc yba@xmath111cu@xmath113o@xmath124 superconducting thin films _ phys .
_ 65 : 2708 - 2711 . hcker m. _ et al . _
( 2014 ) competing charge , spin and superconducting orders in yba@xmath111cu@xmath113o@xmath114 _ phys .
b _ 90 : 054514 .
blanco - canosa s , _
( 2014 ) resonant x - ray scattering study of charge - density wave correlations in yba@xmath111cu@xmath113o@xmath6 _ phys .
b _ 90 : 054513 .
hinton jp _
( 2013 ) new collective mode in yba@xmath111cu@xmath113o@xmath6 observed by time - domain reflectometry _ phys
b _ 88 : 060508(r ) .
schuller ik , gray ke ( 2006 ) time - dependent ginzburg
landau : from single particle to collective behavior _ j. superconductivity and novel magnetism _ 19:401 - 407 .
rothwarf , allen , taylor bn , ( 1967 ) measurement of recombination lifetimes in superconductors _ phys .
lett _ 19 : 27 .
kabanov vv , demsar j , podobnik b , mihailovic d ( 1999 ) quasiparticle relaxation dynamics in superconductors with different gap structures : theory and experiments on yba@xmath111cu@xmath113o@xmath123 _ phys .
b _ 59 : 1497 - 1506 .
( 2004 ) single - quasiparticle stability and quasiparticle - pair decay in yba@xmath111cu@xmath113o@xmath125 _ phys .
b _ 70 : 014504 .
kabanov vv , demsar j , mihailovic d , ( 2005 ) kinetics of a superconductor excited with a femtosecond optical pulse _ phys .
lett _ 95:147002 .
hanaguri t _
( 2009 ) coherence factors in a high - tc cuprate probed by quasi - particle scattering off vortices _ science _ 323 : 923 - 926 .
lee j _ et al . _
( 2009 ) spectroscopic fingerprint of phase - incoherent superconductivity in the underdoped bi@xmath111sr@xmath111cacu@xmath111o@xmath126 _ science _ 325 : 1099 - 1103 .
density waves in solids _ ( addison - wesley , reading , ma , 1994 ) .
hamidian m _
( 2015 ) atomic - scale electronic structure of the cuprate d - symmetry form factor density wave state _ nature phys .
online publication _
smallwood cl _
( 2012 ) tracking cooper pairs in a cuprate superconductor by ultrafast angle - resolved photoemission _ science _ 336 : 1137 - 9 .
smallwood cl _
( 2015 ) influence of optically quenched superconductivity on quasiparticle relaxation rates in bi@xmath111sr@xmath111cacu@xmath111o@xmath112 _ phys .
b _ 92 : 161102(r ) .
norman mr , randeria m , ding h , campuzano jc ( 1998 ) phenomenology of the low - energy spectral function in high-@xmath0 superconductors _ phys .
b _ 57 : r11093 .
corson j , mallozzi r , orenstein j , eckstein jn , bozovic , i ( 1999 ) vanishing of phase coherence in underdoped bi@xmath111sr@xmath111cacu@xmath111o@xmath112 _ nature _ 398 : 221 - 223 .
halperin bi , nelson dr ( 1979 ) resistive transition in superconducting films _
j. low temp .
_ 36 : 599616 .
hashimoto m _
( 2010 ) particle - hole symmetry breaking in the pseudogap state of bi2201 _ nat .
_ 6:414 - 418 .
( 2011 ) from a single - band metal to a high - temperature superconductor via two thermal phase transitions _ science _ 331 : 1579 - 1583 .
levin k , mills dl and cunningham sl ( 1974 ) incompatibility of bcs pairing and the peierls distortion in one - dimensional systems .
i. mean - field theory _ phys .
b _ 10 : 3821 .
tinkham m _ introduction to superconductivity 2nd edition _
( dover publications , mineola , new york 1996 ) zhao x , _ et al . _
( 2006 ) crystal growth and characterization of the model high - temperature superconductor hgba@xmath111cuo@xmath127 _ advanced materials _ 18 : 3243 - 3247 .
n , _ et al . _
( 2008 ) demonstrating the model nature of the high - temperature superconductor hgba@xmath111cuo@xmath4 _ phys .
b _ 78 : 054518 .
yamamoto a , hu w - z , and tajima s ( 2000 ) thermoelectric power and resistivity of hgba@xmath111cuo@xmath4 over a wide doping range _ phys .
b _ 63 : 024504 .
time - dependence of @xmath128 for a range of temperatures that spans @xmath0 is shown for three underdoped samples with @xmath12 , @xmath13 , and @xmath14 k , in ( a ) , ( b ) , and ( c ) , respectively .
curves are offset for clarity by an ammount proportional to the temperature .
the quasiparticle recombination time is extracted for each temperature from the exponential fits shown as dashed black lines.,width=642 ] ( a)-(f ) the temperature dependence of @xmath129ps@xmath130 is plotted for a series of doping levels .
values of @xmath18 as determined from the inflection points of the curves are indicated by the vertical blue arrows , and @xmath83 is indicated by red arrows .
( g ) @xmath18 as function of hole concentration , @xmath19 , as obtained from the inflection points in figs .
( a)-(f ) are plotted as blue squares in a @xmath131 phase diagram for the hg-1201 system .
also shown is the onset temperature of charge order observed by x - ray scattering @xcite ( green circle ) and the critical temperatures for superconductivity ( open circles).,width=642 ] ( a ) temperature dependence of @xmath2 for a series of underdoped hg-1201 samples .
curves are offset for clarity , as indicated in the figure .
( b ) temperature dependence of the positive component of the transient reflectivity , @xmath132 and ( c ) the quasiparticle lifetime , @xmath2 , for ybco ortho viii with @xmath28 k ( black circles ) and hg-1201 with @xmath133 k ( blue squares ) .
the temperature axis is normalized to @xmath0.,width=642 ] ( a ) momentum space diagram of the electron eigenstates that determine the rate of quasiparticle recombination .
the region defined by the square is the first brillouin zone of a cuprate superconductor - inside are the four `` fermi arcs '' shown as blue curves . the pairing of electrons and holes at @xmath62 and @xmath134 drives cdw formation , whereas electron pairing at @xmath62 , @xmath135 states ( blue lines ) is responsible for superconductivity
. quasiparticles are excited states of paired condensates formed from a superposition of two normal phase electron eigenstates .
cdw quasiparticles , represented by the red lines , are superpositions of electrons at @xmath62 and @xmath134 , whereas sc quasiparticles , represented by blue lines , are superpositions of electrons and holes . in the mixed sc - cdw condensate that forms in underdoped cuprate superconductors ,
quasiparticles become superpositions of four , rather than just two electron eigenstates .
the rate at which such quasiparticles scatter back into the condensate depends sensitively on the mixing angle of sc and cdw components , as well as on the condensate phase - correlation time .
the temperature dependent mixing angle , @xmath71 ( shown in ( b ) ) , phase fluctuation rate , @xmath136 ( solid line in ( c ) ) , and reciprocal of the normal state lifetime , @xmath73 ( gray dashed line in ( c ) ) produce the fit to the quasiparticle lifetime in the hg-1201 ( @xmath27 k ) sample displayed in ( d).,width=642 ] ( a ) temperature dependence of the recombination time @xmath2 in zero field and 6 tesla applied normal to the surface of hg-1201 ( @xmath27 k ) .
( b ) temperature dependence of @xmath2 at 0 and 7 tesla in hg-1201 ( @xmath137 ) k. ( c ) the quasiparticle recombination rate , @xmath138 , at 35 k , at @xmath0 = 71 k , and at 80 k , plotted vs @xmath139 for the underdoped hg-1201 sample with @xmath27 k.,width=642 ]
here we describe details of the fits to @xmath11 to demonstrate that the values of @xmath7 discussed in the main text are independent of the fitting procedure . fig .
s1 ( a ) shows time - resolved reflectivity data for five temperatures below , at , and above @xmath0 for the hg-1201 sample with @xmath133 k. the three - component fits , displayed in red , are of the form described in the text , @xmath140 . fig s1 ( b ) and
( c ) display the temperature dependence of the parameters extracted from these fits .
for each temperature , the fit extends from the delay at which @xmath8 first attains 50% of its peak value to @xmath141 ps .
the parameters @xmath79 and @xmath142 account for both the finite rise time of @xmath10 and the presence of the negative pg component .
these two contributions can not be easily disentangled , but they can be captured in the fit by a single negative term in @xmath10 . as can be clearly seen ,
the structure in @xmath2 near @xmath0 is not present in the other fit parameters .
this confirms that the peak in @xmath7 is representative of the rate of quasiparticle recombination and is not related to structure in the other parameters .
the time - independent component @xmath24 reflects the long - lived photo - induced heating of the system and the non - exponential decay that occurs at low temperatures .
this component is much smaller than the exponential components and does not affect the other fit parameters . in order to further demonstrate that the structure reported in @xmath2 is not related to the parameters @xmath79 and @xmath142 , we can limit the fit to an exponential decay plus constant , for example by considering @xmath10 only after it has decreased to 75@xmath143 of its peak value
. fits to a two - component form @xmath144 are displayed in fig .
s2 ( a ) and the resulting fit parameters are compared to those from the three component fits in fig .
s2 ( b)-(c ) .
it is clear that the quantities of interest , @xmath145 and @xmath2 , are not affected by the details of the fit procedure .
( a ) select @xmath146 curves from the @xmath133 k sample of hg1201 .
the three - component fits are shown in red .
( b ) temperature dependence of the three amplitudes @xmath145 , @xmath79 , and @xmath24 .
( c ) temperature dependence of the rise and decay time constants @xmath142 and @xmath147 .
@xmath142 is multiplied by 10 for easy comparison.,width=642 ] ( a ) select @xmath146 curves from the @xmath133 k sample of hg1201 .
the two - component fits are shown in red .
( b ) comparison between @xmath148 extracted from the three - component fit ( blue ) and the two - component fit ( red ) .
( c ) comparison between @xmath149 extracted from the three - component fit ( blue ) and the two - component fit ( red).,width=642 ] | the condensation of an electron superfluid from a conventional metallic state at a critical temperature @xmath0 is described well by the bcs theory . in the underdoped copper - oxides ,
high - temperature superconductivity condenses instead from a nonconventional metallic `` pseudogap '' phase that exhibits a variety of non - fermi liquid properties .
recently , it has become clear that a charge density wave ( cdw ) phase exists within the pseudogap regime , appearing at a temperature @xmath1 just above @xmath0 .
the near coincidence of @xmath0 and @xmath1 , as well the coexistence and competition of cdw and superconducting order below @xmath0 , suggests that they are intimately related .
here we show that the condensation of the superfluid from this unconventional precursor is reflected in deviations from the predictions of bsc theory regarding the recombination rate of quasiparticles .
we report a detailed investigation of the quasiparticle ( qp ) recombination lifetime , @xmath2 , as a function of temperature and magnetic field in underdoped hgba@xmath3cuo@xmath4 ( hg-1201 ) and yba@xmath3cu@xmath5o@xmath6 ( ybco ) single crystals by ultrafast time - resolved reflectivity .
we find that @xmath7 exhibits a local maximum in a small temperature window near @xmath0 that is prominent in underdoped samples with coexisting charge order and vanishes with application of a small magnetic field .
we explain this unusual , non - bcs behavior by positing that @xmath0 marks a transition from phase - fluctuating sc / cdw composite order above to a sc / cdw condensate below .
our results suggest that the superfluid in underdoped cuprates is a condensate of coherently - mixed particle - particle and particle - hole pairs .
first observed as stripe - like order in the `` 214 '' cuprates @xcite , as checkerboard order in vortex cores @xcite , and subsequently at the surface of bi and cl based compounds @xcite , the universality of cdw order in the cuprate phase diagram has been established , through nmr @xcite and x - ray scattering @xcite probes . in yba@xmath3cu@xmath5o@xmath6 ( ybco )
near hole concentration 1/8 application of large magnetic fields stabilizes long - range cdw order at a temperature that approaches @xmath0 @xcite .
the near - degeneracy of the characteristic temperatures of cdw and sc phases suggests that these two order parameters are related , as opposed to simply coexisting and competing .
several theoretical works have suggested that the same short - range antiferromagnetic fluctuations drive the formation of cdw and sc states @xcite .
moreover , the temperature dependence of the cdw amplitude in ybco , as determined from x - ray scattering , can be reproduced by a model of fluctuating , multi - component order , of which cdw and sc states are two projections @xcite .
the basic premise of bcs theory is that the sc condensate is made up of cooper pairs , which are bound states of two electrons with opposite momenta and spin @xcite .
subsequent to bcs , kohn and sherrington @xcite showed that a cdw state is likewise a pair condensate , but of electron and holes , whose net momentum determines the wavelength of charge order .
quasiparticles ( or broken pairs ) are the fundamental excitations of paired condensates such as the sc and cdw states .
it is important for what follows to note that , although quasiparticles in sc and cdw states are fermions , they have properties distinct from the quasiparticles that constitute the normal state .
sc quasiparticles are phase - coherent linear superpositions of normal state electrons and holes , while cdw quasiparticles are superpositions of electrons ( or holes ) . our experiments probe quasiparticle states in underdoped cuprates through time - resolved measurements of their lifetime against recombination , whereby two quasiparticles of opposite spin re - pair and scatter into the condensate . as we discuss below , this scattering rate is sensitive to the phase - coherence of quasiparticle superposition states . to measure the recombination lifetime
, we first generate a nonequilibrium quasiparticle population by photoexcitation with an ultrashort optical pulse .
we use a low pump fluence so as to probe the linear regime in which the photogenerated quasiparticle population is small compared to its thermal value near @xmath0 .
the rate of return to equilibrium is measured by resolving the photoinduced change in optical reflectivity , @xmath8 , as a function of time , @xmath9 , after absorption of the pump pulse .
a wealth of experiments , as reviewed for example in @xcite , have demonstrated that the appearance of a @xmath10 signal reflects the opening of a gap ( or gaps ) at the fermi level and that its amplitude is proportional to the nonequilibrium quasiparticle population . in figure 1
we show @xmath11 at several temperatures for underdoped samples of hg-1201 with @xmath12 , @xmath13 , and @xmath14 k. at high temperature we observe a short - lived negative component of @xmath8 that is associated with the pseudogap ( pg ) @xcite . with decreasing @xmath15
a larger amplitude positive component with a much longer lifetime appears and quickly dominates the signal . in cuprates with near - optimal doping
this positive , long - lived component appears close to @xmath0 @xcite and was therefore associated with the onset of superconductivity .
however , this association breaks down in underdoped samples in which the positive component is already large at @xmath0 ( data at @xmath0 are highlighted in red ) . in figures 2a - f
we plot the maximum value of @xmath16 ( which occurs at @xmath17 1 ps after photoexcitation ) as a function of @xmath15 in underdoped hg-1201 samples with @xmath0 s ranging from 55 to 91 k. the temperature ( @xmath18 ) at which the positive component of @xmath10 appears is indicated by a blue down arrow in each panel .
note that @xmath10 continues to increase continuously with further decrease of @xmath15 , without a clear feature at the critical temperature for superconductivity ( indicated by red arrows ) .
both @xmath18 and @xmath0 are plotted as a function of hole concentration , @xmath19 , in fig .
2 g .
the onset temperatures of positive @xmath10 outline a dome that peaks on the underdoped side of the phase diagram and extends to temperatures 130 k above @xmath0 . based on a correlations with other probes
, we believe that the appearance of the slow , positive component of @xmath10 at the temperatures @xmath20 shown in fig .
2 g corresponds to the onset of local cdw order .
there is a clear correspondence between the dome of @xmath18 as determined by @xmath21 in hg-1201 and the region of the phase diagram where a cdw is detected in ybco @xcite .
although the phase space region of cdw in hg-1201 is yet to be mapped in as much detail as in ybco , a cdw has been detected in hg-1201 samples with @xmath0=71 k @xcite at a temperature ( indicated by the green circle in fig .
2 g ) coincident with @xmath18 .
another correlation linking @xmath10 to the cdw in underdoped cuprates is that the positive component of @xmath10 in ybco ortho - viii ( to be discussed further below ) has the same temperature dependence as a zero wavevector vibrational mode that arises from cdw - induced zone - folding @xcite .
we turn now to measurements of the @xmath15 dependence of the recombination lifetime of quasiparticles for @xmath22 .
the decay curves in fig .
1 were fit using a function of the form , @xmath23 , where the first term describes qp recombination , the second term accounts for finite risetime and the presence of a negative pg component , and the constant offset @xmath24 captures a long - lived contribution that we attribute to local heating by the pump pulse ( see supplement for details on the fitting procedure ) .
figure 3a displays the evolution of @xmath7 with hole concentration in the hg-1201 system . at each hole concentration
we observe structure in the @xmath15-dependence of the quasiparticle recombination time at @xmath0 . in underdoped samples
there is a peak in @xmath7 at @xmath0 that is most prominent in the @xmath2571 k sample and decreases in amplitude at lower and slightly higher hole concentration in a manner that appears to be correlated with the strength of the cdw . figs . 3b and 3c
compare @xmath26 in hg-1201 ( @xmath27 k ) and ybco ortho - viii ( @xmath28 k ) , illustrating that generality of the phenomena described above .
fig . 3b and 3c
show the temperature dependence of the amplitude @xmath29 and @xmath2 respectively for the two underdoped samples . in
both we observe the onset of positive @xmath21 well above @xmath0 and a smooth variation through the sc transition . as shown in fig .
3c , peak in @xmath7 centered on @xmath0 is strikingly similar in the two representative samples .
the @xmath15-dependence of the quasiparticle lifetime in the sample of hg-1201 at near - optimal doping ( topmost data set of fig .
3a ) is qualitatively different from what is seen in underdoped samples . in
the near - optimal sample @xmath2 grows monotonically as @xmath30 , suggesting a tendency to diverge as @xmath31 goes to zero , while in underdoped samples the peak in @xmath2 at @xmath0 is a small feature on a smooth background .
the behavior of @xmath2 in near - optimal hg-1201 , which is observed in other optimally doped cuprates as well @xcite , can be understood within the mean - field theory of superconductivity .
the theoretical description of @xmath7 in the context of bcs began in the 1960 s and its subsequent history is reviewed in ref .
@xcite .
in the mean - field picture , the relaxation of a nonequilibrium quasiparticle population is described by a pair of coupled equations : a landau - khalatnikov equation for the energy gap and a boltzmann - like equation that governs the quasiparticle distribution @xcite . according this analysis ,
@xmath32 ( where @xmath33 is the electron inelastic scattering time ) and diverges as the superconducting gap @xmath34 vanishes as @xmath15 approaches @xmath0 from below . in more recent work on cuprate superconductors , effects associated with the phonon bottleneck @xcite
are included , which leads to replacing @xmath33 by the lifetime of phonons with energy greater than @xmath35 .
the behavior of @xmath7 in underdoped samples is at odds with the mean - field picture described above .
two observations continuous variation of @xmath36 for @xmath22 and the smoothly varying background in @xmath7 that underlies the small peak suggest that a quasiparticle gap has opened well above @xmath0 .
given a pre - existing gap , it is very difficult to explain the modulation of the quasiparticle lifetime near @xmath0 . in view of the difficulties with the mean - field picture ,
we are led to consider the onset of phase coherence at @xmath0 , rather than gap - opening , as the origin of the structure in @xmath7 .
coherence effects have been observed previously elastic qp scattering , as detected by qp interference in scanning tunneling microscopy experiments @xcite .
however , the implications of coherence for a process in which two quasiparticles scatter into the condensate have not previously been considered in the context of the cuprates . in a phase - incoherent state the recombination rate is proportional to the square modulus of the interaction matrix element between electrons and holes . in a phase - coherent paired state , recombination is more complex because , as mentioned previously , qps are linear superpositions of electrons and holes . as a result ,
the matrix element for recombination reflects multiple channels that can interfere constructively or destructively , depending on the nature of the pairing state and qp interactions @xcite .
the rate of qp recombination in a paired condensate is the product of the normal state rate and a `` coherence factor '' ( @xmath37 ) that is a function of the bogoliubov coefficients @xmath38 and @xmath39 . for the case of time - reversal invariant
qp interactions the coherence factors for sc and cdw condensates are given by @xmath40 @xmath41 where @xmath42 and @xmath43 are the amplitudes of the sc gap at the two qp momenta , @xmath44 and @xmath45 are the analogous cdw gaps , and @xmath46 and @xmath47 are the qp energies @xcite . in the limit that @xmath46 and @xmath47 approach the gap energy , these factors reduce to @xmath48 and @xmath49 , while in the normal state @xmath50 .
we note that the result @xmath51 is related to the @xmath52 phase shift between occupied and unoccupied states , which has recently been observed in cdw states in bi2212 @xcite .
the simplest scenario , in which a phase - fluctuating superconductor becomes fully phase coherent at @xmath0 , is inconsistent with a peak in @xmath7 ( or local minimum in recombination rate ) .
instead , sc coherence yields a doubling of the recombination rate , corresponding to the factor of two jump in @xmath37 from 1/2 to 1 upon crossing from a normal to sc state .
however , we have found that a model that takes into account the dual presence of fluctuating cdw and sc order leads to a singular feature in @xmath7 that agrees well with experiment . to formulate this model quantitatively ,
we derived the recombination coherence factor for a state with coexisting sc and cdw order .
the composite sc / cdw condensate is made of particle - hole quadruplets , pairing electrons at @xmath53 and holes at @xmath54 separated by the cdw wavevector @xmath55 .
the structure of the quasiparticle eigenstates of the composite sc / cdw condensate is shown schematically in fig .
4a . based on these eigenstates , we determined the dependence of the mixed state coherence factors on the quasiparticle energy ( see methods ) .
as time and angle - resolved photoemission measurements observe rapid thermalization of quasiparticles to gap edge after photo - excitation @xcite , we focus on the coherence factor in the limit that @xmath56 , which simplifies the expression for @xmath37 considerably . in this limit ,
the qp recombination coherence factor becomes , @xmath57 where @xmath58 and @xmath59 are the relative phases of the qp pair undergoing recombination . in the case of @xmath60-wave pairing
, these phases will be @xmath61 or @xmath52 , depending on which @xmath62 points the qps occupy .
stm qp interference demonstrates that the strongest scattering channels are between states with the same sign of gap amplitude @xcite , so we restrict our attention to these recombination processes . in the presence of phase fluctuations ,
the @xmath63 factors are replaced by their ensemble average @xmath64 , where @xmath65 is the phase - correlation time and @xmath66 is the qp lifetime in the fully incoherent regime .
the coherence factor obtained by this substitution would apply to systems with coexisting , but independent , cdw and sc order .
however , in the light of evidence that these orders are strongly coupled in underdoped cuprates , we consider a description in terms of a multi - component order parameter @xcite whose amplitude @xmath67 is constant and whose fluctuations are described by a single phase @xmath68 . with these assumptions
the coherence factor can be written , @xmath69 where @xmath70 .
fig .
4d shows a fit to @xmath7 for the hg-1201 @xmath27 k sample using eq .
4 . figs .
4b and 4c show the temperature dependence of the three parameters from which the fits were generated . the mixing angle @xmath71 is plotted in fig
4b . the reciprocal of the coherence time , @xmath72 , and reciprocal of the incoherent recombination time , @xmath73 ,
are plotted as solid and dashed lines , respectively in fig .
4c .
the @xmath15-dependence of @xmath66 is determined by a polynomial fit to @xmath7 that ignores the peak at @xmath0 .
the quasiparticle decoherence rate @xmath72 is constrained to be constant below @xmath0 and to vary @xmath74 above the transition , as suggested by angle - resolved photoemission spectra @xcite .
the slope of @xmath72 vs. @xmath75 extracted from our fit ( @xmath76 thz / k ) is consistent with estimates of coherence times obtained from optical conductivity @xcite .
the peak in @xmath7 shown in fig .
4d arises from the interplay between the mixing angle @xmath71 and the onset of quasiparticle coherence .
starting at temperatures well above @xmath0 , the order parameter is cdw - like with a short coherence time .
as the temperature is lowered towards @xmath0 and the cdw becomes more coherent , @xmath77 dips below its normal state value of @xmath78 and the qp lifetime is enhanced .
however , as the order parameter crosses over from cdw to sc - like near @xmath0 , @xmath77 increases , giving rise to a peak @xmath2 . to further examine the relationship between phase coherence and quasiparticle recombination , we investigated the effect of magnetic field , @xmath79 , on @xmath7 .
an overview of @xmath80 in fields applied perpendicular to the cu - o planes is shown in figure 5 .
fig .
5a compares the quasiparticle lifetime in zero field and 6 tesla for a sample of hg-1201 with @xmath25 71 k. the peak in @xmath7 is entirely washed out by the field and replaced by a smooth background that is described by the model parameter @xmath81 discussed previously . in fig
5c , the change in recombination rate caused by the field , @xmath82 , is plotted vs. @xmath79 at three temperatures : well below , above , and at @xmath83 .
it is clear that strong dependence of @xmath2 on @xmath79 is observed only near @xmath0 .
the field dependence of the quasiparticle lifetime is qualitatively different in the near optimal hg-1201 sample ( shown in fig .
5b ) , where we find that the maximum @xmath2 shifts to lower @xmath15 but is not reduced , consistent with what is expected for a mean - field gap opening transition .
the @xmath79 dependence of @xmath2 can be understood to be a consequence of dephasing in the vortex liquid induced by the field . assuming a total dephasing rate , @xmath84 , where @xmath85 is the dephasing rate associated with vortex diffusion , yields @xmath86 \{1-\exp\left[-\gamma(b)\tau_0\right]\}.\ ] ] the dashed line in fig .
5c is a fit to this functional form , with @xmath87(0.08 thz / t)@xmath79 .
this result can be compared with the theory of phase fluctuations in the vortex liquid @xcite , where it is shown that @xmath88 , where @xmath89 and @xmath90 are the coherence and magnetic length , respectively . equating this estimate with the measured @xmath85 yields a reasonable value for the coherence length of 2.4 nm , strongly suggesting that @xmath79-induced dephasing accounts for wipeout of the @xmath2 peak . to summarize , we have described measurements and analysis of the photoinduced transient reflectivity , @xmath26 in representative ybco and hg-1201 samples of underdoped cuprates . the onset of @xmath10 with decreasing temperature is correlated with the appearance of the incommensurate cdw detected previously by nmr , stm , and x - ray scattering measurements .
this correlation leads us to conclude that , although the density of states depression known as the pseudogap is formed at a higher temperature , the cdw either enhances it , or leads to a new gap of different origin . a correlation between cdw and gap formation
is suggested as well by recent arpes measurements @xcite . we focused attention on two aspects of @xmath8 as @xmath15 was lowered through @xmath0 . in underdoped samples ,
@xmath10 increases continuously through @xmath0 , suggesting a smooth variation of the gap at the fermi surface hot spots ( or fermi - arc tips ) .
second , while the gap varies continuously , the quasiparticle recombination time exhibits a narrow local maximum at @xmath0 .
we proposed that this peak in @xmath2 indicates a crossover from fluctuating cdw to sc / cdw order , occurring as the condensate coherence time slides through the time window set by the background recombination time @xmath91 2 - 3 ps .
the link between condensate coherence and the peak in @xmath7 was further supported by the observation that magnetic field causes the peak to disappear into the background .
our results show that quasiparticle recombination provides a new method for probing the onset of coherence in systems characterized by a fluctuating multi - component order parameter .
* acknowledgements : * synthesis and characterization of hg-1201 samples performed at the university of minnesota was supported by the department of energy , office of basic energy sciences , under award no .
de - sc0006858 .
n.b . acknowledges the support of fwf project p2798 .
optical measurements and modeling performed at lawrence berkeley national lab was supported by the director , office of science , office of basic energy sciences , materials sciences and engineering division , of the u.s .
department of energy under contract no .
de - ac02 - 05ch11231 . |
the rich chemistry of carbon is due to the capability of its electronic configuration to adjust to different bonding situation .
carbon atoms realize three main hybridization schemes of the valence orbitals : @xmath1 , @xmath2 , and @xmath0 hybridization . @xmath1
bonding tends to form three - dimensional networks , as in diamond and the related amorphous structures .
@xmath2-hybridized orbitals lead to two - dimensional networks , at the basis of graphene , graphite , fullerenes , nanotubes , ribbon structures , and other forms of wide current interest for their unique mechanical and electronic properties .
@xmath0-hybridized carbon forms linear structures ( carbynes ) , which , compared to their chemically more stable @xmath1 ( alkanes ) and @xmath2 ( alkenes ) hydrogenated counterparts , tend to be rigid , thus potentially appealing as backbones for molecular nanotechnology
. carbynes of increasing length and varied terminations are being synthesized @xcite and characterized.@xcite at the same time , the fabrication of carbon atomic chains from stretched nanotubes or graphene was achieved by controlled electron irradiation in transmission electron microscopes.@xcite the remarkable robustness of these carbon chains under irradiation combined with the ease of electron - beam fabrication at the nm scale can provide a route to the synthesis of actual nano - devices based on carbynes .
recent investigations have elucidated the prominent role of molecular oxygen as a primary source of chemical degradation of @xmath0 carbon @xcite . in a pure - carbon environment carbynes
are far more stable @xcite .
effects of longitudinal straining @xcite and lateral bending @xcite were investigated very recently by _ ab - initio _ methods .
the outcome of those studies is that carbyne chains are at one time extremely stiff against longitudinal straining and very soft against a bending deformation , to the extent that even extreme bending affects only moderately their bonding properties.@xcite in the present paper we focus on the mechanical properties of carbynes in the perspective of exploiting them as construction materials for nano - engineering .
we use _ ab - initio _
simulations @xcite to predict the stiffness of the carbynes for both longitudinal strain and for bending , and their ultimate tensile strength .
the main results of these simulations is that a sp carbon chain ( and it matters little whether it is of cumulenic all double bonds , or polyynic with alternating single - triple bonds ) exhibits not only an exceptional mechanical stiffness against longitudinal deformations , but also a small but nonzero rigidity against bending .
accordingly , a c@xmath3 chain resembles a thin beam characterized by a finite rigidity against buckling , rather than a string .
we find that , e.g. , a single - carbyne c@xmath4 `` pillar '' can withstand a modest but non - negligible axial compression force @xmath5 nn before buckling .
in simulations , one can tune the carbyne length by fixing the position of the end carbon atoms of a selected chain section , and leave all other atoms free to relax to their equilibrium position , compatible with the imposed strain . as well know to civil engineers ,
as soon as the compressive stress exceeds a critical value , the straight configuration of a beam subject to a purely axial strain @xmath6 turns unstable and `` buckles '' spontaneously to a curved arc . in the simulations of highly strained configurations , it is straightforward to address either the metastable straight configuration , or the stable buckled configuration investigated in ref . . to characterize the mechanical properties of a carbyne segment , and in particular to evaluate the buckling transition , one needs to identify the correct linear - response parameteres in a simple model where the effects of different kinds of deformations can be decoupled and studied separately .
the simplest continuum model describes a carbyne as a strained thin beam of a homogeneous material , thereby ignoring its discrete atomic structure .
three basic deformations can be applied to the beam : longitudinal traction or compression ; lateral bending , as sketched in fig .
[ model : fig ] ; torsion around its axis .
correspondingly , the elastic deformation energy can be decomposed as : @xmath7 where @xmath8 is the tensile energy increase due to longitudinal elongation / shortening , @xmath9 is the energy due to the pure flexion of the beam , and @xmath10 is the energy increase due to the twisting of the beam .
remarkably , even this third term is important for the essentially 1d cumulenes.@xcite in the present work we will however mostly ignore torsion , and concentrate on the first two terms . in the linear - response ( small - strain ) regime , we write the tensile energy in terms of the strained length @xmath11 relative to the fully relaxed length @xmath12 as @xmath13 where @xmath14 is the elastic stiffness of the beam , and @xmath15 is the length - independent beam elastic tension : in the continuum mechanics of 3d solids @xmath16 , the product of the material s
elasticity modulus times the beam transverse area .
figure [ model : fig ] indicates the relevant geometrical quantities characterizing the pure - bending deformations of a carbyne of fully relaxed length @xmath12 . in a circular - arc geometry ensuing from the application of a small bending moment @xmath17 at the beam ends , the bending energy can be derived by the linear relation of the deflection angle @xmath18 with the applied bending moment : @xmath19 here @xmath20 is the length - independent bending stiffness of the beam : in 3d continuum mechanichs @xmath21 , the product of the elasticity modulus times the second moment of area of the beam cross section .
( the dimension of @xmath20 is force@xmath22area . ) for a circular deformation , the bending energy @xmath9 is therefore @xmath23 it is convenient to express all quantities in terms of clearly defined geometrical quantities , e.g. the distance @xmath24 between the end points of the chain and the the maximum deflection @xmath25 from straight geometry .
we have : @xmath26 with the implicit substitution of the expressions ( [ angle ] ) and ( [ length ] ) , the elastic energy of the beam stretched and/or bent in a circular arc is therefore @xmath27
we determine the linear - response parameters @xmath15 and @xmath20 by means of simulations based on the density functional theory ( dft ) in the local density approximation ( lda ) .
the time - honored lda is one in many functionals being used for current dft studies of molecular systems : other functionals often improve one or another of the systematic defects of lda ( underestimation of the energy gap , small overbinding and overestimation of the vibrational frequencies ) , but to date no functional is universally accepted to provide systematically better accuracy than lda for all properties of arbitrary systems . for a covalent system of @xmath28 and @xmath29 electrons as the one studied here ,
lda is appropriate , and we expect our results to change by a few percent at most if the calculations were repeated using some other popular functional.@xcite we compute the total adiabatic energies by means of the plane - waves dft code quantum espresso.@xcite ry@xmath30 pev , and push atomic coordinates relaxation until all force components are smaller than @xmath31 ry@xmath32 pn .
we use ultrasoft pseudopotentials,@xcite for which a moderate cutoff for the wave function / charge density of @xmath33 ry is sufficient . to address isolated molecules in the repeated - cell geometry implied by plane waves ,
we make sure that at least @xmath34 nm of vacuum separates all atoms in adjacent periodic images . ]
we consider the two different limiting structures of carbynes : polyynes ( or @xmath35carbynes ) with alternating single triple bonds ( @xmath36c@xmath37c@xmath36c@xmath37c@xmath36 ) , contrasted to cumulenes or @xmath38carbynes , with nearly - equal - length ( double ) bonds ( = c = c= ) .
a typical polyynic chain is obtained when each end carbon atom forms a single bond with a ligand ( e.g. a hydrogen atom , as in diacetylene ) , while a typical cumulene is obtained when each end carbon forms a double bond , for example to a ch@xmath39 group , as in ethylene .
we focus on a c@xmath4 chain segment , and compare its mechanical response in its hc@xmath39-c@xmath4-c@xmath39h polyynic and h@xmath39c - c@xmath4-ch@xmath39 cumulenic realizations . for the polyyne , we select hydrogenacetylide terminations rather than a single hydrogen because the hc@xmath39- group allows us to impose a bending moment at the ends of the c@xmath4 chain as in fig . [ model : fig ] , by displacing a c atom along the chain continuation rather , than a chemically different h atom , thereby probing the intrinsic properties of carbyne , rather than those of a specifically - terminated compound .
we also consider c@xmath40 polyyne of different lengths , to check for size effects , and a @xmath41 end - twisted c@xmath4 cumulene , which we must treat within the local spin - density approximation ( lsda ) since twisting induces a peculiar total spin-1 electronic state , with a twofold - degenerate level occupied by two parallel - spin electrons.@xcite we do not attempt any serious size scaling , for two main reasons : ( i ) very long carbynes are mechanically unstable ( thus unsuitable for nanomechanical applications ) anyway , and ( ii ) the long - range interaction effects demonstrated recently @xcite would make a proper _ ab - initio _ size - scaling determination of the mechanical ( in particular bending ) properties of long carbynes prohibitively expensive . in the present work we do not deal with the exotic odd-@xmath42 c@xmath3 chains , which are chemically less stable and more difficult to synthesize.@xcite to evaluate @xmath15 , starting from the fully relaxed carbyne we stretch or compress a straight c@xmath4 chain section by fixing the positions of the end carbon atoms , and letting all other atoms free to relax to their resulting equilibrium positions under stress .
we extract an estimate of @xmath15 by fitting eq . to the resulting values of the excess energy as a function of the chain elongation , as reported in fig .
[ kfit : fig ] . [ cols="<,^,^,^,^,^,^,^ " , ] to evaluate the bending stiffness @xmath20 , we force the chain to bend by imposing a small lateral displacement of the two atoms adjacent to the c@xmath4 section of the chain , while keeping the first and last atom in the chain section bounded to remain laterally undisplaced , following the scheme of figs .
[ model : fig ] and [ gfit : fig ] ( inset ) .
we then let all other degrees of freedom ( including the axial position of the laterally bounded atoms ) of the chain relax .
the c@xmath4 section relaxes to form a circular arc , which we fit by minimizing the squared distance of the relaxed atomic positions in the @xmath43 plane from the circular arc @xmath44 , whose radius @xmath45 is determined by taking into account the separation @xmath24 of the terminal atoms of the c@xmath4 chain section , namely those kept at @xmath46 . ] to extract the geometrical quantities indicated in fig .
[ model : fig ] .
note that this procedure differs substantially from the fixed - end approach of ref .
, where longitudinal and bending strains are applied at the same time , and deviations extend well outside the linear - response region . in the language of that work , the maximum arc - chord ratio in our calculations is @xmath47 .
we repeat this procedure for several values of the fixed lateral displacements . with the computed dft energies and structures
, we extract the values of @xmath20 by fitting the angular dependency of @xmath9 in eq .
, see fig . [ gfit : fig ] .
the resulting elasticity parameters @xmath15 and @xmath20 are collected in table [ values : tab ] . despite the different bonding configuration of polyyne and cumulene , the overall longitudinal elasticity @xmath15 values are remarkably similar .
the value of @xmath15 is compatible with a sound velocity @xmath48 km / s for long - wavelength longitudinal vibrations ( @xmath49 is the linear mass density ) . in view of the long - range interactions
demonstrated recently @xcite , the weak dependence of @xmath15 on the chain length is remarkable .
the bending stiffness @xmath20 is comparatively more sensitive to the chain length , see also fig .
[ gscale : fig ] .
if this velocity is used to evaluate the highest - frequency brillouin - zone boundary phonon in a simplest mass - spring chain , one obtains a frequency @xmath50 thz , i.e. a wave number @xmath51 @xmath52 , expectedly slightly above the region of the observed high - frequency raman and infrared modes emerging as the characteristic spectroscopic signatures of the carbynes.@xcite the cumulene is anisotropic : one can make it bend either within the molecular plane , as sketched in the inset of fig .
[ gfit : fig ] , or perpendicularly to it , resulting in differently oriented double bonds @xcite being perturbed .
we find that bending is significantly softer for the out - of - plane distortion than for the in - plane one , with a @xmath53% anisotropy . the proposed cosine expression for the strain energy given in eq . ( 9 ) of ref .
is compatible with our eq . for small curvature
( although a small amount of longitudinal strain energy is neglected there ) .
the relation @xmath54 obtained based on the average bond length @xmath55 pm and orientation - dependent bonding strength @xmath56 ev as obtained in ref .
yields @xmath57 nn/(nm)@xmath58 , in good agreement with our result , especially with the c@xmath59 chain value .
this agreement is even more surprising in view of the extreme deformations introduced in ref . , and indicates that the linear - response coefficient of eq
. extends well outside its validity region , provided that the cosine espression eq .
( 9 ) of that work is adopted . according to elementary continuum mechanics,@xcite the critical axial load that a beam can sustain in its straight configuration before the onset of the buckling instability
is given by euler s formula @xmath60 corresponding to a critical strain @xmath61 for a critical sinusoidal lateral deformation profile . by substituting the computed values for @xmath15 and @xmath20
, we find that the c@xmath4 carbyne sustains a critical strain of nearly @xmath62% , under the action of a critical force of approximately @xmath63 nn , before buckling , as reported in table [ values : tab ] . the corresponding critical strain and force for cumulene ( in the softer out - of - plane direction )
are essentially equal .
the computed values are relevant when both ends of c@xmath4 are pinned ( hinged , free to rotate ) , while if both ends were frozen , the critical force and strain would be four times larger .
likewise , following eqs . and , @xmath64 decays rapidly with size . for example , for c@xmath65 ,
whose length is approximately 2.15 times that of c@xmath4 , @xmath66 nn , i.e. approximately 20% of the c@xmath4 value only .
direct calculation done for the c@xmath59 polyyne in hc@xmath39-c@xmath59-c@xmath39h agrees with this scaling , with a small deviation due mostly to a smaller value of @xmath20 , relative to the c@xmath4 value ( while @xmath15 is practically coinciding , see table [ values : tab ] ) .
we find it rather surprising that , despite the presence of in principle softer single bonds , the polyynic chain is essentially as hard to compress as double - bonds based cumulene . the same essentially equivalent stiffness of polyyne and cumulene is found against bending deformations , but here the result is less unexpected .
even a maximally twisted cumulene in a high - spin electronic state exhibits very similar mechanical properties , see table [ values : tab ] .
we check the obtained buckling critical point for possible effects of chain discreteness or anharmonicity , by running actual simulations .
we compare pairs of dft simulations at fixed strain : one bound to the linear configuration as in fig .
[ kfit : fig ] , and one starting off in a slightly curved geometry . in the curved simulations , below the buckling instability
, the relaxing chain goes back to straight , and recovers the same energy and outward force that the chain produces on the pinned end atoms as in the straight calculation .
in contrast , above the buckling instability , a curved shape stabilizes , with a net decrease in total energy and in the force acting on the pinned end atoms . crosses in fig .
[ kfit : fig ] report the fixed - end excess energy of the polyynic chain when allowed to relax in a buckled geometry .
we find that the actual buckling instability occurs very close to the linear - response continuum - model value @xmath67 of eq . .
relaxations near instability are rather delicate , since equilibrium is almost indifferent , providing a manifold of almost - equivalent geometrical configurations which give a hard time to the optimization algorithm .
the present work demonstrates that it is possible to go a long way in describing an immaterialy thin object such as a monoatomic carbon chain with the mechanics of bulk construction elements , not unlike it was done for 2d graphene in refs . . eventually , carbyne chains turn out rigid enough that the 886 pm - long free - standing c@xmath4 polyyne can sustain a compressive strain of nearly 0.5% before buckling . on the tensile side ,
we extend our calculations well outside the linear - response region , to estimate the ultimate tensile strength of carbynes .
the ultimate tension @xmath68 is computed as the maximum force that the carbyne section produces in sustaining an externally imposed longitudinal strain @xmath69 .
the result in the 10 nn region agrees with previous determinations @xcite ( but disagrees significantly with the tersoff - brenner model determination of ref . ) , and indicates that a single molecular chain could be used as a rope to lift a mass as heavy as one microgram without breaking !
if one could pack carbynes with a lateral density of one per @xmath70 nm@xmath58 cross - section , one would obtain a material characterized by a remarkable ultimate tensile strength of the order 300 gpa , comparable to that of carbon nanotubes.@xcite despite the remarkable mechanical properties of carbynes , their chemical reactivity makes them unsuitable to many applications in real world mechanical situations , where carbon fiber or carbon nanotubes provide superior stability with comparable mechanical properties .
however in clean well - isolated nano - engineered devices one could envisage that the usage of c@xmath3 chains as structural or elastic elements may lead to consistent advantages over traditional solutions .
for example , the soft bending degree of freedom could be exploited for the construction of sensitive accelerometers . indeed , according to our evaluation of @xmath20 , a 0.1 @xmath49 g mass hanging at the end of a 37 nm - long c@xmath71 chain ( with the other end bonded to a fixed substrate )
would deflect laterally by as much as 1 nm under an acceleration of 0.01 ms@xmath72 , i.e. one thousandth of the earth gravitational field . | as @xmath0 carbon chains ( carbynes ) are relatively rigid molecular objects , can we exploit them as construction elements in nanomechanics ? to answer this question , we investigate their remarkable mechanical properties by _ ab - initio _ total - energy simulations .
in particular , we evaluate their linear response to small longitudinal and bending deformations and their failure limits for longitudinal compression and elongation . |
the first stage in the formation of low - mass stars is considered to be the time when molecular clouds fragment into a number of cold ( @xmath5 10 k ) and dense ( @xmath6 @xmath7 ) gravitationally bound cores ( e.g. * ? ? ?
these cores are usually referred to as starless or pre - stellar cores and were first identified in molecular line surveys carried out by myers and co - workers @xcite .
comparison of these cores with the iras point - source catalogue revealed that they were not associated with infrared sources , evidence that they represented the earliest stages of star formation .
observations of starless cores therefore provide an opportunity to study the initial conditions of protostellar collapse .
the cold and dense cores can be probed by studying their thermal ( continuum ) emission , which provides information about the dust properties ( ( e.g .
* ) ; @xcite ) , and by studying their line emission or absorption , which provides information on the chemistry in these regions ( ( e.g . * ? ? ?
* ) ; @xcite ; @xcite ) .
dense cores provide the environment necessary for the formation of molecules such as , and complex organic carbon bearing species @xcite .
the homonuclear molecule is difficult to observe in quiescent molecular clouds and so the second most abundant molecule is used as a molecular tracer .
is usually optically thick in molecular clouds and so its isotopologues ( e.g. , , , and arranged in order of decreasing abundance ) have been used to trace gas temperature and cloud mass in more opaque regions ( ( e.g .
* ) ; @xcite ) .
infrared and millimetre observations of cold cores have provided evidence of freezing out of gaseous species onto dust grain surfaces ( e.g. * ? ? ?
* ; * ? ? ?
this freeze out results in a change in grain sizes , which affects the emissivity of the grains in ways that are currently not well understood ( ( e.g . * ? ? ?
* ) ; @xcite ) .
in this paper we present far - infrared , broadband observations of three starless cores carried out using the spectral and photometric imaging receiver array ( spire ) instrument @xcite on the herschel space observatory ( hereafter referred to as _
herschel , _ * ? ? ?
spire consists of an imaging fourier transform spectrometer ( fts ) with two hexagonally packed bolometer arrays : slw ( 447 - 990 ghz ) , and ssw ( 958 - 1546 ghz ) and a photometer with three hexagonally packed bolometer arrays : psw ( 250 band ) , pmw ( 350 band ) and plw ( 500 band ) . far - infrared observations from ground based telescopes are restricted to transmission `` windows '' in the atmosphere ( e.g. * ? ? ?
as a result , spectral energy distributions ( seds ) of starless cores have been constructed using flux densities obtained from a small number ( 3 - 5 ) of continuum maps ( ( e.g . * ? ? ?
* ; * ? ? ?
* ) ; @xcite ) .
these maps are obtained using facility instruments on different telescopes having different beam sizes , each sensitive to structure on a different scale .
interpretation of seds obtained by combining data from different instruments taken under different atmospheric conditions is challenging .
fitting grey body functions to such seds using @xmath8 minimization techniques exposes the degeneracy between emissivity spectral index ( @xmath2 ) and temperature ( @xmath1 ) @xcite . _
herschel _ provided unfettered access to the far - infrared wavelength region , free from atmospheric absorption , and a stable operating platform .
since the peak of the seds for cold and dense cores falls within the range of the spire fts , analysis of the resulting spectra allows one to retrieve more robust values of @xmath1 , @xmath2 and column densities .
this paper is the first in a series that will present results from our analysis of spire fts spectra of a number of star forming regions . in this paper
we present the analysis of spectra corresponding to the brightness peaks of three starless cores , l1521e , l1521f and l1689b .
the spatial extent and variation of dust temperature and emissivity spectral index will be the subject of a subsequent paper .
l1521e is located in the taurus star - forming region at a distance of @xmath9=140 pc @xcite .
it is a triple - lobed pre - stellar core situated at the south - eastern end of the l1521 filament @xcite .
l1521e was initially identified as a very young core by @xcite who showed that the abundances of carbon - rich molecules in l1521e are higher than in other dark cloud cores and comparable to those in tmc-1 .
studies carried out by @xcite have supported this claim by showing that l1521e has low abundance and no depletion , implying that l1521e is less chemically processed and therefore recently contracted to its present density .
@xcite estimate the age of l1521e to be @xmath10 yr .
@xcite have studied l1521e and other starless cores by combining ground - based 24 , 70 and 160 @xmath3 m data from the multiband imaging photometer ( mips ) on the _ spitzer space telescope _ ( _ spitzer _ , * ? ? ?
* ) , 450 and 850 @xmath3 m data from the submillimetre common - user bolometer array ( scuba ) on the james clerk maxwell telescope ( jcmt ) and 90 , 170 and 200 @xmath3 m data from the imaging photopolarimeter ( isophot ) on the _ infrared space telescope _
( _ iso _ ) . despite the differences between these facility instruments , in particular their beam profiles
, @xcite were able to derive core properties with a limited number of assumptions .
their modified blackbody fit to the l1521e sed with a fixed dust emissivity index ( @xmath2 = 2 ) resulted in a dust temperature of k. by using observations of l1521e , @xcite derived a core radius of 0.031 pc , a mass of 2.4 m@xmath11 and a density of @xmath12 = @xmath7 for the brightest position .
l1521f ( aka mc27 ; @xcite ; @xcite ) is also located in the taurus star - forming region at the same distance as l1521e .
l1521f appears isolated with a strong central condensation as shown from the 160 @xmath3 m _ spitzer _ map presented by @xcite .
it is a dense starless core harbouring a low luminosity ( @xmath13 = 0.05 l@xmath11 ) object l1521f - irs @xcite . due to its low luminosity
, l1521f did not appear in the iras catalogue @xcite , but was observed by _
l1521f - irs together with l1544 @xcite are the best known examples of evolved starless cores .
@xcite obtained a dust temperature of 9 @xmath14 2 k , a mass of m@xmath15 and a column density of @xmath16 = @xmath17 . the typical error in the derived @xmath16
is @xmath14 20 - 30 % . in their mapping surveys of taurus
, @xcite argued that the high central density ( @xmath010@xmath18 @xmath7 ) and infall asymmetry seen in the ( 3 - 2 ) lines from l1521f indicate that the core is in its earliest stages of gravitational collapse with a free - fall timescale of 10@xmath19 to 10@xmath20 yr .
besides the high central density and infall asymmetry , l1521f also shows molecular depletion and enhanced deuterium fractionation @xcite .
@xcite explain this by suggesting that l1521f is less chemically evolved and has a molecular hole of radius @xmath21 2000 au .
@xcite have speculated that the spatially compact ( @xmath2130 ) line wings seen in ( 3 - 2 ) lines may be due to bipolar outflow emission similar to that seen from l1014-irs by @xcite . a well - defined bipolar scattered light nebula seen at short wavelengths (
@xmath215 @xmath3 m ) by @xcite also suggests the presence of molecular outflows in l1521f similar to those from low - mass protostars .
l1689b is located in the ophiuchus star - forming region at a distance of @xmath9=120 pc ( @xcite ; @xcite ) .
millimeter and mid - infrared maps of l1689b have shown that it is an elongated core , sharp - edged in the north - south direction @xcite .
( 2 - 1 ) , ( 2@xmath22 - 1@xmath23 ) and ( 3 - 2 ) observations of l1689b have revealed line asymmetry typical of infall ( @xcite ; @xcite ; @xcite ) .
this suggests a supercritical core has formed and the object is entering the phase of dynamical contraction @xcite .
@xcite compared their ( 2 - 1 and 3 - 2 ) observations and millimetre / submillimetre continuum observations of l1689b with results from a spherically symmetric radiative transfer model and concluded that freeze - out of co is occurring towards the centre of l1689b .
@xcite and @xcite fitted grey body functions to l1689b seds with a fixed dust emissivity index ( @xmath2 = 2 ) and obtained dust temperatures of k and 11.6 k respectively .
@xcite attempted to disentangle the effects of temperature variations along the line of sight by applying an inverse - abel transform based technique and obtained a temperature of k. the mass and column density of l1689b derived by @xcite are @xmath24 = m@xmath11 and @xmath16 = @xmath17 respectively .
typical errors on @xmath16 were reported to be @xmath14 2030 % .
@xcite derived a mass of 11 m@xmath11 .
spire and pacs photometer maps of the three starless cores : l1521e , l1521f and l1689b were obtained as part of the gould belt guaranteed time key programme for the study of star formation @xcite .
this programme produced large area maps of the l1688 cloud in the rho ophiuchus star forming region and the l1521 cloud in the taurus star forming region .
the l1521e and l1521f observations correspond to 1342202254 and that for l1689b to 1342205093 .
these observations were obtained using the spire pacs parallel mode and processed using the herschel interactive processing environment ( hipe ) version 13.0 @xcite .
details for these observations are provided in table [ tab : obsdetailsall ] .
the data are publicly available through the herschel science archive ( hsa )
. images of each core obtained at the three spire photometer wavebands are shown in figure [ fig : spirephoto ] . [ cols= " < , <
, > , < , < , < , > " , ] [ tab : obsdetailsall ] the calibration uncertainties associated with spire and pacs photometer maps is in the region of 5% or less @xcite .
the starless cores ( l1521e , l1521f and l1689b ) were observed using the _ herschel _ spire fts as part of the `` evolution of interstellar dust '' key program under the ism specialist astronomy group number 4 ( sag4 , * ? ? ? * ) .
a summary of the observation details is provided in table [ tab : obsdetailsall ] .
three sparsely sampled observations were obtained for each core studied : _ on ( on - source ) , _ int ( intermediate or on the wings of the source ) and _ off ( off - source ) .
the observations were taken in low spectral resolution mode , yielding a spectral resolving power ( @xmath25 ) of @xmath020 60 across the frequency range 447 1546 ghz ( 671 194 ) . figure [ fig : spirefootprints ] shows the spire fts detector array footprints for these observations over - plotted onto spire photometer pmw maps .
the spire fts data were reduced using hipe version 14.0 , which applies a major correction to all extended - source calibrated fts spectra correcting for the far - field coupling efficiency of the fts feed - horns ( see @xcite and valtchanov et al .
( in preparation ) for details ) .
prior to hipe version 14.0 , flux densities of extended sources observed with the fts were found to be significantly lower than those from the spire photometer by factors of between 1.4 and 1.7 . from the spire fts sparsely sampled observations , spectra were obtained for each of the detectors in the short wavelength ( ssw ) and long wavelength ( slw ) arrays @xcite . in this paper
we discuss results from analyzing spectra for the central detectors ( slwc3 and sswd4 ) for the spire fts sparse sampled observations pointed at the on - source positions of the three cores ( table [ tab : obsdetailsall ] ) .
the uncertainty in the derived intensity for sparsely sampled observations of reasonably bright and extended sources observed in the central detectors of the spire fts is 7 % @xcite .
figure [ fig : spirephoto ] shows the spire photometer maps for each of the three starless cores ; each map measures 6@xmath266 and is centred on the source in question ( table [ tab : obsdetailsall ] ) .
the spire photometer beam at each wavelength is represented by a circle on the lower left corner of each map . on the lower right corner of each map
is a linear scale in pc .
contours are shown at 50 , 40 , 30 , 20 and 10% of the peak flux density indicated and show that all three cores are extended with respect to the largest spire photometer beam fwhm of 36.4 at 500 @xmath3 m , which corresponds to 5096 au = 0.025 pc at an assumed distance of 140 pc .
for each source , the far - infrared images show that there is a similarity in the extent of the continuum emission , evidence that the 250 , 350 and 500 @xmath3 m maps are tracing the same dust component . in the limit of small optical depth , emission from dust particles can be described using a modified blackbody ( grey body ) function @xcite : @xmath27},\ ] ] where @xmath28 is the flux density per reference beam in jy ( 1 jy= ) and @xmath29 is the planck blackbody emission from dust at temperature @xmath1 : @xmath30},\ ] ] @xmath31 is the optical depth , which can be expressed in terms of the emissivity of dust ( specific opacity ) , @xmath32 ( in @xmath33 g@xmath34 ) , the density of the gas - dust mixture @xmath35 , and the path length @xmath13 by ( e.g. * ? ? ?
* ) @xmath36 the emissivity of the dust at frequency @xmath37 is parametrized by @xcite as : @xmath38},\ ] ] for a standard dust - to - gas ratio of 1:100 .
@xmath3 (= 2.8 ) is the mean molecular weight of interstellar material in a molecular cloud ( with 71 % molecular hydrogen gas , 27 % helium and 2 % metals ) per hydrogen atom ( e.g. * ? ? ?
* ) , @xmath39 is the mass of the hydrogen atom , @xmath16 is the hydrogen column density , @xmath40 is the beam solid angle subtended either by the reference beam ( for spectra corrected using the semiextendedcorrector tool ( sect ) which will be discussed in section [ sec : mismatch ] ) or by the adopted source extent ( for extended calibrated spectra ) and @xmath41 is the opacity in units of @xmath33 per hydrogen molecule .
@xmath42 is the planck constant , @xmath43 is the boltzmann constant , @xmath44 is the speed of light , @xmath45 = 0.1 @xmath33/g is the emissivity of dust grains at a frequency of 1000 ghz @xcite .
we use the dust opacity law from @xcite to be consistent with other papers on dense cores ( e.g. * ? ? ?
* ; * ? ? ?
however , recent results from dust models have shown that @xmath45 can vary by up to a factor of 2.0 ( e.g. * ? ? ?
@xmath2 is the dust emissivity index ( 1 @xmath46 ) .
small values of @xmath2 ( @xmath471 ) have been attributed to dust grains of mm size or larger @xcite while values of @xmath2 between 1 and 2 have been attributed to changes in material composition of dust grains which result from grain growth @xcite .
the origin of values of @xmath48 2 is not yet well understood @xcite .
the total mass ( gas and dust ) contained within a circle of radius @xmath49/2 ( where @xmath49 is the diameter in m ) from the centre of the core , @xmath50 can be obtained from equation [ eqn:1 ] @xcite as @xmath51 the number density @xmath12 and hydrogen column density @xmath16 are calculated as : @xmath52 @xmath53 respectively .
the diameter of each core is derived from the angular source sizes calculated using sect ( discussed in section [ sec : mismatch ] ) . in this way , fits of a grey body function to the far - infrared / submillimetre seds provide values for the dust mass and molecular hydrogen column density .
the derived masses for the three sources have been compared with their jeans masses to determine if they are stable against gravitational collapse .
the jeans mass can be expressed as ( e.g. * ? ? ?
* ) : @xmath54 where @xmath55 is the jeans radius which we assume to be equal to the core radius @xmath56
. cores with @xmath57 are considered to be stable against gravitational collapse while cores with @xmath58 are unstable and will collapse to form protostars .
the spire instruments use conical feed horns to couple the telescope beam onto the individual detectors .
the photometer uses single - mode feed horns whose beams are well described by gaussian profiles with fwhm values of 18.1 , 24.9 and 36.4 for the psw , pmw and plw , respectively @xcite . to accommodate the broad spectral range of the two bands , the fts feed - horns are by necessity multi - moded , which results in complex wavelength dependent beam profiles @xcite . @xcite
used hermite - gaussian functions to approximate the spire fts beams and have provided coefficients that can be used to reconstruct the beam profiles at each wavelength in the spire band .
the calibration of the spire fts is performed using solar system objects and other astronomical targets as outlined by @xcite and @xcite .
the standard data reduction pipeline for the spire fts was designed to work for two extreme cases of objects : those whose spatial extent is smaller than the beam and those whose spatial extent is much larger than the beam @xcite .
the calibration scheme for the former is referred to as point - source calibration and , for the latter , extended - source calibration .
neither of the calibration schemes work for sources whose spatial extent falls in - between the two extremes as is the case for l1521f . a large gap and a difference in slope between the ssw and slw spectra in the overlap region is an indicator that the applied calibration is not appropriate for the source being studied .
this is a result of the fact that the beam diameter for the slw detectors is almost twice as large as that for the ssw detectors in the overlap region . to illustrate this point , figure [ fig : source_simulations ] shows the impact of assuming extended- and point - source calibration for a fictitious instrument similar to the spire fts when viewing a grey body of temperature 15 k and dust emissivity index of 1.45 and where the source size is 50 ( top row ) , 25 ( centre row ) and 15 ( bottom row ) , respectively . to simplify the problem
, it is assumed that the instrument has a constant beam size of 19 in the ssw and 35 in the slw bands , respectively .
the left column shows the brightness expected from the grey body when using extended - source calibration .
the right column shows the spectra produced when using point - source calibration . in the top row ,
the source size is 50 , which is larger than both the ssw and slw beam sizes . since in this case
the source will appear extended to both beams , it is appropriate to use the extended source calibration which yields the continuous spectrum , seamless in the overlap region , as shown in the top left panel . however , if point source calibration was used it would lead to a large gap in the overlap region as shown in the top right figure . in the centre row ,
the source size is 25 , which is larger than the ssw beam size but smaller than that for slw . since in this case
the source is extended to ssw and point - like to slw , it is appropriate to use extended source calibration for ssw and point - source calibration for slw . in the left figure ,
the spectrum for ssw ( red ) is correct , while the spectrum for slw ( blue ) is not . in the right figure ,
the spectrum for slw ( blue ) is correct while that for ssw ( red ) is not . in the bottom row , a source size of 15 has been assumed .
this is smaller than both the ssw and slw beams .
the source appears as a point source to both ssw and slw and so applying extended - source calibration ( left panel ) is inappropriate for both bands .
as is expected , applying point - source calibration ( right panel ) results in a continuous spectrum , again seamless in the overlap region .
figure [ fig : source_simulations ] illustrates that important information on the source size is conveyed by studying the overlap region between the ssw and slw bands .
the gap is large when an observation of an extended source is point - source calibrated ( top right panel ) . in this case
the larger slw beam measures more flux than the smaller ssw beam .
the gap is smaller for a source size that is extended in the ssw band , but point - like in the slw band ( right panel in the centre row ) . applying extended - source calibration to an observation of a source that is extended in the ssw band , but point - like in the slw band ( left panel in the centre row ) results in a gap where the slw spectrum is lower than the ssw spectrum .
point - source and extended source calibrated spectra for the centre detectors are shown in figure [ fig : point_ext_allcores ] . as discussed above ,
the large mismatch between point - source calibrated slw and ssw spectra is an indication that these sources are not point - like .
the mismatch is minimal for extended source calibrated spectra indicating that these sources are close to being fully extended .
of the three cores , extended source calibration works well for l1521e and l1689b while a significant difference exists in the overlap region for l1521f indicating that it must be approached differently .
in fact , as will be shown in section [ sec : mismatch ] , l1521f is not as extended as the other two sources and extended - source calibration is therefore inappropriate . for comparison purposes
, the bottom panels of figure [ fig : point_ext_allcores ] also show flux densities of the cores derived from extended - source calibrated spire photometer maps , using circular apertures having the same size as the fts beam , at the corresponding wavelengths .
the photometer data shown represent the mean of the colour corrected flux densities of all map pixels in the circular apertures used .
the error bars include both the statistical uncertainty from the spectra and the overall calibration uncertainties . before the combined slw and ssw spectra
can be fitted with a grey body function it is essential that the gap and differences in slope in the overlap region that result from source - size issues be addressed .
we have explored two methods of overcoming this challenge .
the first method is background subtraction .
the background can be determined from a select number of off - axis detectors or from observations of the same region slightly offset from the target .
the effect of emission from the local background in which cores would be found was recognized in preparing the `` evolution of interstellar dust '' key program under the ism specialist astronomy group ( sag4 , * ? ? ?
as a result the proposal included three observations for each of the starless cores , two of which were pointed slightly away from the source as discussed in section [ sec : specobs ] and figure [ fig : spirefootprints ] .
figure [ fig : ext_allcores ] shows the extended - source calibrated spectra for the on - source positions for l1521e ( top left ) and l1689b ( bottom left ) and the corresponding differences between extended - source calibrated spectra for the on - source and for the off - source positions ( right column ) . in this paper only data from the central detectors ( slwc3 and sswd4 ) from each footprint shown in figure [ fig : spirefootprints ] is used .
spectra for the other detectors will be discussed in a subsequent paper focused on the spatial variation of dust parameters .
figure [ fig : ext_allcores ] shows that subtracting a background corrects for the differences in slope in the overlap region between slw and ssw spectra .
fitting a grey body function to the resultant sed for l1689b results in a lower temperature than that from fitting the uncorrected sed .
this is a clear indication that spectra measured by the spire fts beam consist of emission from both a hot extended foreground or background and a cold core .
this is consistent with the interpretation that the centres of these self - gravitating cores are cold while the outer regions , which are heated by an external radiation field , are warmer .
if the emission from the warmer dust in the outer layers is not removed from the measured data , the derived core masses will be biased towards lower values @xcite .
fitting to the l1521e sed a more extended core before and after background subtraction results in almost the same temperatures and dust emissivity indices implying that the background is a less important contributor to the measured flux .
the relative importance of background flux can be seen by examining the spectra in figure 5 : for l1521e , the background contributes @xmath010% of the total brightness ( figure 5a vs. 5b ) , whereas the background contribution is @xmath030% for l1689b ( figure 5c vs. 5d )
. for comparison purposes , figure [ fig : ext_allcores ] also shows flux densities of the cores derived from extended - source calibrated spire and the photodetector array camera and spectrometer ( pacs , * ? ? ?
* ) photometer maps using circular apertures having the same size as the fts beam at the corresponding wavelengths for the on - source and off - source positions .
the photometer data shown in figure [ fig : ext_allcores ] represent the mean of the colour corrected flux densities of all map pixels in the circular apertures used . as mentioned previously , the error bars include contributions from the statistical analysis and calibration uncertainties . an equivalent background subtraction method as that used for the fts spectra has also been applied to the photometer data and the resultant seds fitted with grey body functions .
it can be seen that the brightnesses measured by the spire and pacs photometers are consistent with those from the fts , yielding consistent temperatures and dust emissivity indices especially for background subtracted seds .
the derived parameters from fitting a grey body to the fts spectra are based upon 45 independent spectral points whereas the fit to the photometer data is based upon four points . as a result fts spectra
provide a better constraint to grey body fitting .
a similar exercise carried out on spectra from l1521f showed that background subtraction alone can not correct for the gap in the overlap region between slw and ssw spectra .
extended source calibration appears to be overcorrecting the spectra for this source and so l1521f is not as extended as the other two cores .
a second method for correcting the gap in the overlap region between slw and ssw spectra is to use the `` semiextendedcorrector '' tool ( sect ) as discussed by @xcite .
the tool corrects sparse sampled point - source calibrated spectra using the measured beam profile @xcite and an image of the source distribution .
the spectrum is corrected to a default reference beam of 40 . in the case of l1521f
, the spectra were corrected to a reference beam equal to the source size derived from sect .
sect applies a correction for the forward coupling efficiency , but does not correct for effects such as the efficiency with which the reconstructed beam shape couples to the source , the deviation of the model from the true source distribution , and the response far from the axis @xcite .
the tool is therefore limited to sources that are not extended .
for this reason , sect was only used to correct spectra for l1521f .
figure [ fig : point_sect_allcores ] shows the results from applying the sect tool to point - source calibrated and background subtracted spectra for l1521f and fitting grey body functions to extract the dust temperature and emissivity index .
figure [ fig : point_sect_allcores ] also shows colour corrected flux densities of the core derived from point - source calibrated spire and pacs photometer maps .
we used circular apertures derived by assuming that the measured map for l1521f is a convolution of a gaussian source profile and the instrumental beam profile at the corresponding wavelengths for the on - source and off - source positions .
the error bars are as described above .
the source sizes required to eliminate the gap in the overlap region using sect for the background subtracted spectra are 61 , 29.5 and 39 which corresponds to 0.041 , 0.019 and 0.023 pc for l1521e , l1521f and l1689b , respectively ( the latest sect updates do not allow determination of source sizes for sources larger than 60 since these are considered extended ) .
these source sizes are in remarkable agreement with 0.031 pc derived by @xcite for l1521e and 0.015 and 0.024 pc derived by @xcite for l1521f and l1689b , respectively .
the smallest source sizes for l1521e , l1521f and l1689b estimated from the photometer maps using the 50 % contour are @xmath080 , @xmath035 and @xmath0120 respectively .
it is clear that source sizes for l1521e , l1521f derived from the spire fts spectra and photometer maps are in agreement while those for l1689b are not . l1689b appears larger than the other two cores on the photometer maps most likely due to the more extended foreground / background .
performing background subtraction reveals a much smaller cold core .
the derived source size for l1521f is slightly less than the smallest slw beam size , but larger than that for the maximum ssw beam indicating that l1521f is semi - extended .
the derived size for l1521e is a factor of @xmath01.5 larger than the largest slw beam indicating that l1521e is extended .
the derived size for l1689b is slightly less than that for the largest slw beam and therefore lies at the boundary between semi - extended and extended sources . for the reasons discussed in previous paragraphs , analysis for l1521e and l1689b
is based on spectra corrected only for background emission ( figure [ fig : ext_allcores ] ( b ) and ( d ) ) while for l1521f we use the spectrum corrected for background emission and for source extent by sect ( figure [ fig : point_sect_allcores ] ( b ) ) . since the intermediate positions are closer to the on - source positions the possibility exists for the intermediate position to be contaminated by emission from the core itself .
for this reason we have chosen the off - source position when performing the background subtraction . however , we find that using spectra from either the intermediate or off - source positions for the background subtraction produces consistent results for @xmath1 and @xmath2 .
once corrected for effects of source spatial extent , the spire fts and spire / pacs photometer seds were separately fitted using equation [ eqn:1 ] to derive @xmath1 and @xmath2 . in fitting the grey body function to our data we used the mpfit function which employs the levenberg - marquardt technique to solve least square problems @xcite .
uncertainties in the input fts spectra include calibration uncertainties and the standard deviation from averaging spectra for 55 repetitions ( 110 scans ) .
uncertainties in the input photometer data include calibration uncertainties and the standard deviation from averaging flux densities for all pixels within an aperture .
the best grey body fits to the seds are shown in figure [ fig : ext_allcores ] ( for l1521e and l1689b ) and in figure [ fig : point_sect_allcores ] ( for l1521f ) and the results are presented in table [ tab : starlessresults ] .
this allows us to derive the core masses , densities and column densities of the cores ( section [ sec : starlesscoresanalysis ] ) .
the error values in the derived quantities do not incorporate uncertainties in the value of @xmath45 ( equation [ eqn:3 ] ) .
lccc + & l1521e & l1521f & l1689b + @xmath59 [ ] & 61 & 29.5 & 39 + @xmath49 [ pc ] & 0.041 & 0.020 & 0.023 + @xmath1 [ k ] & [email protected] & [email protected] & [email protected] + @xmath2 & [email protected] & [email protected] & [email protected] + @xmath61 [ ] & 263 & 250 & 246 + @xmath62 [ @xmath2610@xmath63 & [email protected] & [email protected] & [email protected] + @xmath24 [ m@xmath11 ] & [email protected] & [email protected] & [email protected] + @xmath65 [ @xmath26 10@xmath18 @xmath7 ] & [email protected] & [email protected] & [email protected] + @xmath67 [ @xmath26 10@xmath68 @xmath17 ] & [email protected] & [email protected] & [email protected] + @xmath70 [ m@xmath11 ] & 0.55 & 0.43 & 0.34 + @xmath71 & 1.7 & 0.2 & 1.5 + [ tab : starlessresults ] the sed for l1521e from the spire fts data peaks at a wavelength of = 263 which is in the spire band . our derived dust temperature ( @xmath72 = 9.8 @xmath140.2 ) for l1521e is slightly larger than @xmath1 = 8.1 @xmath14 0.4 derived by @xcite by assuming a dust emissivity index of 2 .
the value of @xmath2 derived in this work ( @xmath2= 2.6 @xmath60 0.9 ) is consistent with the often assumed value of @xmath2 = 2 based on studies of dielectric functions of silicate and graphite grains by @xcite .
the fwhm size of l1521e ( 0.04 pc ) obtained from sect is slightly larger than 0.031 pc reported by @xcite .
the derived optical depth for l1521e and for the other two cores is @xmath73 as is expected in these cold and dense environments .
the derived core mass is @xmath24 = m@xmath11 which is less than 2.4 m@xmath11 derived by @xcite .
the column density is @xmath16 = @xmath17 .
using our derived core diameter of 0.041 pc , the derived column density can be converted to a density of @xmath12 = @xmath7 which is 2.8@xmath74 lower than the minimum in the range @xmath12 = @xmath7 derived by @xcite .
the ratio of the core mass to the jeans mass is 1.7 implying that l1521e is unstable to gravitational collapse .
the conclusion drawn from this analysis is consistent with that presented in papers cited above and in section [ l1521eback ] , namely that l1521e is a very young core that has recently contracted to its present density .
fitting to the spire fts sed for l1521f results in = 250 which is also in the spire band and yields a temperature , @xmath72 of k and a dust emissivity index of .
our temperature is greater than 9 @xmath14 2 k reported by @xcite , which again was obtained by assuming a dust emissivity index of 2 .
the large differences in the derived temperatures is most likely due to the degeneracy between @xmath2 and @xmath1 , which is such that larger values of @xmath2 correspond to lower dust temperatures and vice versa . the higher temperature and lower core mass ( @xmath24 = m@xmath11 ) that we obtain support the existence of a protostar in the l1521f core as reported by @xcite .
the low value of @xmath2 could be explained by the change in the dust properties as the temperature of the protostar increases .
it could also be due to temperature variations along the line of sight ( e.g. * ? ? ?
* ) and therefore complete modelling of the core , which is beyond the scope of this paper , is necessary to provide a definitive conclusion . to the best of our knowledge ,
this is the first time a far - infrared observation of a dense core has resulted in such a low value for @xmath2 .
our low @xmath71 ratio of 0.3 which is less than 1.0 is also an indication that the core is stable against gravitational collapse .
the column density of the gas in the envelope is @xmath16 = @xmath17 . using our derived core diameter of 0.02 pc
, the column density of the envelope can be converted to a density of @xmath12 = @xmath7 . for l1689b ,
the spire fts sed peaks at = 246 which , again , is within the spire band .
we obtained a temperature of k and a dust emissivity index of .
our derived temperatures are consistent with k derived by @xcite which again was obtained by assuming a dust emissivity index of 2 .
we also derived a core mass of m@xmath11 and a column density of @xmath17 . using our derived core diameter of 0.023 pc
, the derived column density can be converted to a density of @xmath12 = @xmath7 .
the @xmath71 ratio of 1.5 implies that l1689b is gravitationally unstable to collapse .
in fact studies by @xcite and @xcite have shown that l1689b is undergoing dynamical contraction .
the _ herschel _ space observatory provides a unique platform to study the far - infrared universe .
low spectral resolution spire fts spectra of three starless cores : l1521e , l1521f and l1689b have been analyzed to derive their dust properties .
the calibration schema employed makes interpretation of spire spectra challenging when sources are neither point - like nor extended .
since the three studied cores fall into this regime , there is a mismatch between the overlap regions of the slw and ssw spectra .
two methods have been used to address this problem : background subtraction and the use of sect .
the resulting corrected spectra have been fitted with grey body functions .
previous studies of starless cores have been based on @xmath03 - 5 points from photometer maps covering the seds sparsely . as a result , it has been common practice to assume @xmath2 = 2 . the broadband spire fts spectra provide tighter constraints on the sed shape allowing us to have both the dust temperature and emissivity index to vary as free parameters . for comparison purposes ,
we have also fitted grey body functions to flux densities derived from spire and pacs photometer maps . using our fts spectra
, we derive temperatures , @xmath72 , of k , k and k and dust emissivity indices , @xmath2 , of , , , respectively for l1521e , l1521f and l1689b .
dust emissivity indices for l1521e and l1689b are consistent with @xmath2 = 2 usually assumed in modelling starless cores . we derive core masses of , and m@xmath11 for l1521e , l1521f and l1689b , respectively .
since the peaks for the seds for all three cores fall within the spire fts band , we believe that our derived core masses are more accurate than those previously recorded in literature .
our derived core properties are robust and will help improve radiative transfer models of starless and pre - stellar cores .
the dust in l1521f is significantly warmer that that for the other two cores suggesting that it is being heated by an internal radiation source .
l1521f could now be a low - mass protostar .
_ herschel _ is an esa space observatory with science instruments provided by european - led principal investigator consortia and with important participation from nasa .
spire has been developed by a consortium of institutes led by cardiff univ .
uk and including univ .
lethbridge ( canada ) ; naoc ( china ) ; cea , lam ( france ) ; ifsi , univ .
padua ( italy ) ; iac ( spain ) ; stolkholm observatory ( sweden ) ; imperial college london , ral , ucl - mssl , ukatc , univ .
sussex ( uk ) ; caltech , jpl , nhsc , univ.colorado ( usa ) .
this development has been supported by national funding agencies : csa ( canada ) ; naoc ( china ) ; cea , cnes , cnrs ( france ) ; asi ( italy ) ; mcinn ( spain ) ; snsb ( sweden ) ; stfc ( uk ) ; and nasa ( usa ) .
the authors would like to thank ivan valtchanov and trevor fulton for their valued input in the processing of spire fts data .
, d.a.n . and m.h.d.vdw .
acknowledge support from the csa and nserc and the latter also by the lundbeck foundation .
research at the centre for star and planet formation is funded by the danish national research foundation and the university of copenhagen s programme of excellence . | dust grains play a key role in the physics of star - forming regions , even though they constitute only @xmath01 % of the mass of the interstellar medium .
the derivation of accurate dust parameters such as temperature ( @xmath1 ) , emissivity spectral index ( @xmath2 ) and column density requires broadband continuum observations at far - infrared wavelengths .
we present _
fourier transform spectrometer ( fts ) measurements of three starless cores : l1521e , l1521f and l1689b , covering wavelengths between 194 and 671 @xmath3 m .
this paper is the first to use our recently updated spire - fts intensity calibration , yielding a direct match with spire photometer measurements of extended sources .
in addition , we carefully assess the validity of calibration schemes depending on source extent and on the strength of background emission .
the broadband far - infrared spectra for all three sources peak near 250 @xmath3 m .
our observations therefore provide much tighter constraints on the spectral energy distribution ( sed ) shape than measurements that do not probe the sed peak .
the spectra are fitted using modified blackbody functions , allowing both @xmath1 and @xmath2 to vary as free parameters .
this yields @xmath1 of k , k and k and corresponding @xmath2 of , and for l1521e , l1521f and l1689b respectively .
the derived core masses are , and @xmath4 , respectively .
the core mass / jeans mass ratios for l1521e and l1689b exceed unity indicating that they are unstable to gravitational collapse , and thus pre - stellar cores . by comparison ,
the elevated temperature and gravitational stability of l1521f support previous arguments that this source is more evolved and likely a protostar .
= 1 [ firstpage ] stars : formation dust , extinction infrared : ism submillimetre . |
physics experiments at colliders have mapped production of strange hadron species over a variety of systems and energies .
such measurements have commonly included production of the @xmath1 but , due to the complexities of its observation , rarely the @xmath0 .
these two hyperons share the same quark content , differing only in isospin and mass , separated by a mere 77 mev / c@xmath6 @xcite . reconstructing the @xmath0
requires reconstructing the @xmath1 , so comparing the ratio of yields of the two species will be exclusive of its feed - down from @xmath0 decays , but contributions from resonance decays into both species will be included . ] provides a simple test of whether conditions are different for these two similarly - composed hyperons .
are they produced similarly ?
are their final state interactions the same ? without measurements , these traits may only be assumed . above threshold energies for production of @xmath0 , the experimental data for the value of the ratio @xmath5 cluster around 1/3 , matching the ratio of isospin degeneracy factors ( @xmath7 ) of the @xmath8 and @xmath1 ground states ( @xmath9=3 , @xmath10=1 ) , as will be shown later in this paper .
however there are absences in measurements for hadronic collisions at moderate to high energies , and for collisions of sizable nuclei .
high energy nuclear collisions are of particular interest for the possibility of opening new channels of production via partonic degrees of freedom @xcite .
models often start with the production of all particles , followed by the rapid decay of resonance states .
both the @xmath1 and @xmath0 receive contributions from decays of strange resonances , but those feeding the @xmath1 are much more numerous .
knowledge of these contributions is pertinent to understanding the full picture , and efforts to measure these are underway @xcite .
= 5.0 in the predictions for @xmath5 from a variety of particle production models are as follows : the event generator hijing / b@xmath11 @xcite gives 0.37 for @xmath12 gev d+au collisions ; the thermus @xcite statistical thermal model gives 0.36 at @xmath13 mev , @xmath14 ; the sudden hadronization " @xcite thermal model gives @xmath150.33 at @xmath13 mev ; and the alcor @xcite quark coalescence model gives 0.20 .
both thermal models feature little sensitivity to input parameters , as seen in figure [ fi : ratiocont ] and elsewhere @xcite , with nearly no dependence upon chemical potentials . though there is some spread in the predicted values
, it should be noted that the @xmath16 @xmath17 ( taken from direct production before resonance contributions ) is much higher .
so the final value of @xmath5 may reflect even more upon resonance yields than the direct production of @xmath1 and @xmath0 .
for example , quark coalescence treats the @xmath1 and @xmath0 equally : the entirety of their difference then results from the copious resonances @xcite .
thermal models place a penalty on production of more massive states , so they begin with a primordial value below that of quark coalescence , but have less resonance contributions to drop the final ratio .
the star detector is capable of reconstructing the decay @xmath18 with excellent signal - to - noise @xcite .
reconstruction of photon conversions ( @xmath19 e@xmath3e@xmath4 ) in detector material can be accomplished similarly @xcite , the validity of which is underscored by the ability to visualize the structure of detector material from identified conversion points as seen in figure [ fi : svt](left ) .
focusing on low transverse momentum ( @xmath20 ) photons then provides sufficient resolution to identify @xmath0 through invariant mass of @xmath21 pairs when combinatoric backgrounds are adequately low , shown in figure [ fi : svt](right ) . to extract a yield , the shape of the background under the invariant mass peak
is obtained by rotating candidate @xmath1 and @xmath2 daughters from same events to avoid mixing event classes .
preliminary signals for @xmath0 have been observed via this technique in @xmath22 gev p+p collisions , as well as @xmath23 and 200 gev cu+cu collisions at rhic , but we concentrate here on minimum bias 200 gev d+au collisions .
= 2.4 in = 2.3 in
by using the same @xmath1 candidates in both numerator and denominator of @xmath5 , efficiencies for finding the @xmath1 cancel out and we are left with three main corrections .
the first is weak decay feed - downs from doubly or triply strange hadrons , which is essentially an issue for @xmath1 only .
based on preliminary analyses of @xmath1 , @xmath24 , and @xmath25 in minimum bias 200 gev d+au data in star , approximately 20% of the @xmath1 candidates at all @xmath20 come from such feed - down @xcite .
= 3.1 in the second and third correction factors are the conversion probability of photons in detector material and the reconstruction efficiency of those conversions .
these are obtained via embedding simulated @xmath0 decays into real data .
the conversion probability is difficult to determine accurately as it involves simulating every possible conversion volume properly , but there seems to be rather little dependence on the @xmath20 of the parent @xmath0 .
in contrast , the reconstruction efficiency drops rapidly with decreasing parent @xmath20 , due mostly to reduced acceptance for low @xmath20 electrons ( which are found only for conversions in the gas volume of the star tpc ) . to improve statistics , we exploit the expectation that @xmath26 and sum baryon+antibaryon counts under the assumption @xmath27 . as the number of reconstructed counts are still small and bounded by zero yield , the errors are large and asymmetric .
determination of the measured ratio thus involves a monte carlo allowing values to vary appropriately within their combined statistical and systematic errors , and we report the most probable value with errors determined by the rms of outcomes above and below it .
we reconstruct sufficient numbers of @xmath0 to determine @xmath5 with significance in two @xmath20 bins from minimum bias 200 gev d+au ( @xmath28 ) .
we find preliminary @xmath5 = @xmath29 for @xmath0 @xmath20 = 2.5 - 3.5 gev / c , and @xmath30 for @xmath0 @xmath20 = 3.5 - 4.5 gev / c .
combining the two bins leads to @xmath5 = @xmath31 and this is shown along with past experiment results in figure [ fi : pastres ] .
we also find @xmath32 , which does not discount the summation of baryons+antibaryons for our ratio .
the large error bars on @xmath5 prevent definitive model differentiation , but the low value suggests that strong resonance contributions are prevalent even in d+au collisions .
11 albrecht h ( argus collaboration ) 1987 _ phys .
_ b * 183 * 419 + bogolyubsky m y @xmath33 @xmath34 1986 _ yad .
fiz . _ * 43 * 1199 + bogolyubsky m y @xmath33 @xmath34 1989 _ yad .
fiz . _ * 50 * 683 + eisenstein r a ( ps185 collaboration ) 1994 _ phys . atom .
nucl . _ * 57 * 1680 + sullivan m w @xmath33 @xmath34 1987 _ phys .
_ d * 36 * 674 + baldini a @xmath33 @xmath34 1988 _ total cross - sections for reactions of high - energy particles _ vol 12 ed madelung o ( berlin : springer , landolt - brnstein ) + acciarri m @xmath33 @xmath34 ( l3 collaboration ) 1994 _ phys .
_ b * 328 * 223 + acciarri m @xmath33 @xmath34 ( l3 collaboration ) 2000 _ phys .
b * 479 * 79 + sewerin s @xmath33 @xmath34 ( cosy-11 collaboration ) 1999 _ phys .
lett . _ * 83 * 682 + kowina p @xmath33 @xmath34 ( cosy-11 collaboration ) 2004 _ proc .
tenth int . conf . on hadron spectroscopy ( aschaffenburg ) _ vol 1 ( secaucus , heidelberg : springer ) 852 + ( kowina p @xmath33 @xmath34 ( cosy-11 collaboration ) 2004 _ preprint _ nucl - ex/0401020 ) + yuldashev b s @xmath33 @xmath34 ( fnal-343 collaboration ) 1991 _ phys . rev . _
d * 43 * 2792 bir t s and zimnyi j 1983 _ nucl .
phys . _ a * 395 * 525 eidelman s @xmath33 @xmath34 2004 _ phys .
_ b * 592 * 1 koch p , rafelski j , and greiner w 1983 _ phys .
_ b * 123 * 151 koch p and rafelski j 1985 _ nucl .
a * 444 * 678 markert c 2005 _ j. phys .
g : nucl . part .
phys . _ * 31 * s897 vance s e , gyulassy m , wang x n 1998 _ phys .
b * 443 * 45 wheaton s and cleymans j 2005 _ j. phys .
g : nucl . part .
phys . _ * 31 * s1069 letessier j and rafelski j 2002 _ hadrons and quark gluon plasma _
cambridge monographs on particle physics , nuclear physics , and cosmology ( cambridge ; new york : cambridge university press ) bir t s , lvai p , and zimnyi j 1995 _ phys .
_ b * 347 * 6 lvai p 2004 _ private communications _ adler c @xmath33 @xmath34 ( star collaboration ) 2002 _ phys .
* 89 * 092301 adams j @xmath33 @xmath34 ( star collaboration ) 2004 _ phys .
_ c * 70 * 044902 cai x ( star collaboration ) 2005 _ j. phys .
g : nucl . part .
* 31 * s1015 | while yields of @xmath0 have been measured in many different colliding systems , no measurements exist in high energy nuclear collisions @xcite .
it also remains unexplored whether the relative yields of @xmath0 and @xmath1 change in collisions of heavy ions where a dense medium may permit alternate production mechanisms , such as quark coalescence @xcite .
the star detector at rhic is able to reconstruct the electromagnetic @xmath0 decay into @xmath1 plus @xmath2 via the weak decay of the @xmath1 and @xmath2 conversions into e@xmath3e@xmath4 pairs in the detector material .
we report here on our measurement of the ratio @xmath5 in minimum bias 200 gev d+au collisions .
we also compare the measured ratio to expectations from various models . |
the effective temperature @xmath10 is one of the fundamental properties of a star . in principle
, it may be determined from the stellar angular diameter @xmath11 and the bolometric flux @xmath12 according to : @xmath13 where @xmath14 is the stefan - boltzmann constant . in practice , however , the difficulties and limitations that arise when measuring angular diameters restrict the _ direct _ determination of temperatures to a relatively low number of stars . the problem is worse for main sequence stars although interferometric ( see e.g. kervella et al .
2004 ) and transit ( brown et al .
2001 ) observations have been recently performed . even though there are alternative ways to determine effective temperatures , all of them require the introduction of models , which makes the results somewhat dependent on the models adopted . in the range of spectral types
f0-k5 , the effective temperature may be determined from the spectroscopic conditions of excitation and ionization equilibrium of fe lines ( e.g. takeda et al . 2002a , hereafter t02a ) , synthetic photometry ( e.g. bessell et al .
1998 , hereafter b98 ; houdashelt et al .
2000 ) , fitting of the balmer line - profiles ( e.g. fuhrmann 1998 , hereafter f98 ; barklem et al . 2002 ) , and line - depth ratios ( e.g. gray 1994 ; kovtyukh et al .
2003 , hereafter k03 ) .
our approach is the use of the infrared flux method ( irfm ) , which will be described in
[ sect : irfm ] and compared to the other methods in
[ sect : irfmcomparison ] . in purely spectroscopic methods ,
the temperature is determined by constraining its value in such a way that the abundance of does not depend on the excitation potential of the lower level of the lines used and that the average abundance of fe derived from and lines are equal ( see e.g. t02a ) .
this actually sets a constraint on both @xmath3 and @xmath15 ( surface gravity ) so degenerate solutions may be found .
the technique has been applied to a wide variety of samples : stars in the thin and thick disks of the galaxy ( bensby et al .
2003 , hereafter b03 ) , planet - hosting stars ( santos et al . 2004 , hereafter sim04 ) , globular cluster stars ( sneden et al . 2004 ) , and so on .
evidence is being collected , however , towards the fact that the spectroscopic method has systematic errors caused by model uncertainties , such as non - lte effects ( note also that most spectroscopic studies use 1d - lte classical model atmospheres ) . for instance , the spectroscopic temperatures are @xmath16 k higher than those determined with the irfm for planet - hosting stars ( ramrez & melndez 2004b , hereafter rm04b ; sim04 ) , surface gravities are @xmath17 dex higher than those derived from hipparcos parallaxes for @xmath0}}>-0.7 $ ] ( allende prieto et al . 1999 , sim04 ) , and ionization equilibrium of fe is not reproduced by classical model atmospheres neither in the hyades cool dwarfs ( yong et al .
2004 ) nor in the field metal - rich k dwarfs ( allende prieto et al .
2004 , hereafter ap04 ) .
synthetic spectra and filter transmission functions have been used to generate grids of colors in the @xmath3 , @xmath0}}$ ] , @xmath15 space using both kurucz ( 1979 ) atlas models ( b98 ) and gustafsson et al .
( 1975 ) marcs models ( houdashelt et al .
synthetic colors are often put into the observational systems with samples of field stars ( usually with metallicities near solar ) with reliable temperatures or calibrated color - temperature relations .
zero point corrections are finally applied until the @xmath3 scale resembles the absolute one .
scale obtained with direct temperatures ( i.e. stellar angular diameter and bolometric flux measurements ) as the _ absolute _ @xmath3 scale . ] since this is made with field stars of solar metallicity , the zero point corrections are strictly valid only for synthetic colors with @xmath0}}\sim0 $ ] , and its validity for other metallicities relies on the capability of models to reproduce real stellar spectra . although observations and temperatures derived with other methods are involved in the procedure , the synthetic @xmath3 scales retain the basic properties of the model atmospheres adopted and are thus useful to test them .
the strength of the hydrogen balmer lines is very sensitive to the effective temperature below 8000 k and it is only weakly affected by the surface gravity .
profile fitting is the best approach to an estimate of @xmath3 with balmer lines , normally within 110 k of random error .
barklem et al .
( 2002 ) found @xmath18 k for the sun with up - to - date line opacities and a reasonable agreement between their temperatures and those obtained with the irfm for @xmath19 k. castelli et al .
( 1997 ) , however , show that the results from balmer line - profile fitting are strongly dependent on the models and the details of the convection treatment . to satisfactorily reproduce the profiles in the solar spectrum using atlas9 models with overshoot , for example
, the temperature of the sun has to be increased up to 5900 k. with the line - depth ratio technique ( gray & johansson 1991 , gray 1994 ) very small internal errors ( @xmath20 k ) for individual stars may be achieved if several calibrated line - depth ratios are used . even though the temperature scale derived with this technique may be less accurate due to systematic errors , the line - depth ratios may be used to look for very small time dependent @xmath3 variations , like those produced by photospheric inhomogeneities and starspots .
the solar temperature with this method results to be 44 k below the standard value of 5777 k and an empirical correction of 44 k has to be applied ( k03 ) . since nothing guarantees that this correction is applicable to the whole range of atmospheric parameters , the systematics ,
if present , are not removed .
the aim of this work is to extend and improve the internal accuracy of the best available empirical temperature - color calibrations based on the irfm , namely , those given by alonso et al .
( 1996b , hereafter aam96b ; 1999b , hereafter aam99b ) for the @xmath21 and strmgren @xmath22-@xmath23 systems .
melndez & ramrez ( 2003 ) and ramrez and melndez ( 2004a , hereafter rm04a ) extensions to the vilnius , geneva , @xmath24 and ddo systems will also be revised . in this paper
we describe our irfm implementation and test our results .
the calibration of the metallicity - dependent @xmath3 vs. color relations is the subject of a companion paper ( hereafter referred to as part ii ) .
this paper is distributed as follows : in [ sect : irfm ] the irfm and its present implementation are explained , the sample and input data adopted as well as their effects on the temperatures to be derived are described in [ sect : sample ] .
the calculation of the final temperatures and angular diameters is given in
[ sect : irfmtemps ] while the fundamental tests ( e.g. comparison with measured angular diameters ) and comparison with other methods are presented in
[ sect : irfmcomparison ] .
conclusions are summarized in
[ sect : conclusions ] .
the irfm was first introduced by blackwell & shallis ( 1977 ) as a mean to simultaneously determine the effective temperature and angular diameter of a star .
the basic idea of the method is to use the emergent monochromatic flux predicted by models ( @xmath25 ) and , similarly to eq .
( [ eq : def1 ] ) , given by @xmath26}},{\log g})\ , \label{eq : def2}\ ] ] where @xmath27 is the corresponding flux as measured on earth , to solve the system of equations ( [ eq : def1 ] ) and ( [ eq : def2 ] ) for the unknowns @xmath3 and @xmath11 .
the irfm thus requires @xmath12 , @xmath0}}$ ] , @xmath15 and @xmath27 to be known _ a priori_. the use of
the infrared is due to the relatively low dependence of @xmath25 on @xmath3 in this region of the spectrum ( see e.g. fig . 1 in blackwell et al .
1979 ) .
a more practical approach to the irfm was given by blackwell et al .
( 1980 ) , who defined the @xmath28-factors as the quotients between the total and monochromatic fluxes .
the comparison of the observational @xmath28-factor ( @xmath29 ) with a grid of theoretical @xmath28-factors ( @xmath30 , note that @xmath31 is a grid in the @xmath3 , @xmath0}}$ ] , @xmath15 space at a given wavelength ) allows to easily automatize the method .
the irfm is often applied using more than one infrared wavelength ( @xmath32 ) , and a different @xmath3 is obtained for each @xmath32 . a critical ingredient of the irfm is the absolute flux calibration of the standard star from which the @xmath27 fluxes of the problem stars are calculated . if @xmath33 are the absolute monochromatic fluxes for the standard star , then @xmath34 where @xmath35 is the observed magnitude in a given band for which @xmath32 is , for example , the mean wavelength of the filter . the @xmath36 ( @xmath37 )
factors are required since the filters have finite widths and are not usually narrow in the infrared .
it would be ideal to calculate the @xmath36 factors from spectrophotometric observations but in practice models are adopted ( see the appendix in aam94a for a discussion on this particular point ) . for the irfm
, the absolute flux calibration for the standard star must be done in such a way that the temperatures obtained in different wavelengths are close to each other and , when properly averaged , consistent with the direct temperatures .
a calibration of this kind is that of aam94a and is the one we have used in our work .
even though it was derived using direct temperatures of giant stars , this calibration provides good temperatures for dwarf stars as well , as it was shown in rm04b , where the comparison of irfm angular diameters with measured ones for 7 main sequence stars showed good agreement ( see also [ sect : irfmvsdiam ] for an update ) .
the models we have adopted for the @xmath25 fluxes are kurucz seds ( spectral energy distributions ) as given in the library of theoretical spectra by lejeune et al .
( 1997 ) . from them and
the response functions of the @xmath38 , @xmath39 , @xmath40 and @xmath41 filters of the tcs system ( aam94b , aam98 ) we calculated grids of @xmath36 and @xmath31 factors in the ranges 3500 k@xmath42 k , @xmath43}}<+0.5 $ ] , and @xmath44 , where kurucz models are available . figure [ fig : qr ] illustrates the dependence of the temperature indicator in the irfm ( the @xmath45 factors , see [ sect : irfmtemps ] ) as a function of atmospheric parameters and filter . in the ranges of most interest to us
( @xmath46 k for giants , @xmath47 k for dwarfs ) , the use of more sophisticated models will not have a substantial effect on the irfm given that the model dependence of the method is relatively small . also , unlike the uv and optical regions of the spectrum , the density of lines in the infrared is relatively low .
the seds we took from lejeune et al .
( 1997 ) are originally from kurucz ( 1992 ) .
they were synthesized with kurucz multi - million line list and atmosphere models with scaled solar abundances and an approximate treatment of convective overshooting .
the models were computed with a microturbulent velocity of 2 km s@xmath48 .
the use of classical model atmospheres may not be a big issue for the irfm .
it is remarkable that the three - dimensional hydrodynamical simulations suggest that the effective temperature scale derived with the irfm using one - dimensional hydrostatic models is the same as the one obtained with hydrodynamical models for turn - off stars with @xmath49}}<0 $ ] , and that for stars in the same metallicity range and approximately solar temperature the difference is only about 20 k ( asplund & garca prez 2001 ) .
our sample consists of stars in the ranges of atmospheric parameters for which few data points were available in aam work , like the metal - poor and metal - rich extremes of the regions around @xmath50 k and below 5000 k. the planet - hosting star sample is a metal - rich one given the well stated metallicity - enhancement of stars with planets ( see e.g. gonzalez et al .
2001 , sim04 ) .
other metal - rich stars were taken from castro et al .
( 1997 ) and pompeia et al .
the very metal - poor stars added are mainly from ryan et al .
( 1991 , 1996 ) , norris et al .
( 2001 , mostly hot fg unevolved stars ) , and mcwilliam et al .
( 1995 ) ; while the metal - poor cool dwarfs are from yong & lambert ( 2003a , 2003b ) .
the distribution of our sample stars in the @xmath3 vs. @xmath0}}$ ] plane is given in fig .
[ fig : newsample ] . the distribution of the combined sample ( i.e. including aam stars ) in @xmath3 and @xmath0}}$ ] is shown in fig .
[ fig : sample ] .
the total sample is shown with the solid histogram , the dotted histogram is the aam sample .
some of the peaks in the giant @xmath0}}$ ] distribution correspond to giants in clusters included in aam work while those at @xmath0}}=0.0,-1.0,$ ] and @xmath51 in the dwarf @xmath0}}$ ] distribution are the stars with kinematic metallicity assignment ( see [ sect : feh ] ) .
note how the dwarf metal - rich extreme is now much more populated , mainly with the planet - hosting stars .
the very metal - poor end of both dwarfs and giants is also better represented , although still in small numbers .
some hot stars ( @xmath52 k ) are also included but they are almost in the limit of applicability of the irfm .
the infrared photometry for the aam sample was taken from aam94b and aam98 .
this photometry is given in the tcs system and has a mean accuracy of 0.02 mag .
for our sample , johnson and 2mass photometry was adopted and transformed into the tcs system by means of the transformation equations of aam94b and aam98 ( johnson to tcs ) and the following transformation equations ( 2mass to tcs ) we derived for both dwarfs : @xmath53 with standard deviations of @xmath54 ( @xmath55 ) , @xmath56 ( @xmath57 ) and @xmath58 ( @xmath59 ) , respectively ( @xmath60 is the number of stars in every fit ) ; and giants : @xmath61 with standard deviations of @xmath62 ( @xmath63 ) , @xmath54 ( @xmath64 ) and @xmath65 ( @xmath63 ) , respectively .
the @xmath66 magnitudes were mostly obtained from the different catalogs included in the general catalogue of photometric data ( mermilliod et al .
1997 , hereafter gcpd ) . for the globular and open clusters we adopted the reddening corrections given by kraft & ivans ( 2003 ) and aam99a , respectively . for the field stars in aam sample ,
their @xmath67 values were adopted , while for our sample stars the reddening was estimated as explained below .
we estimated @xmath67 values from the maps of schlegel et al .
( 1998 ) and burstein & heiles ( 1982 ) , several extinction surveys included in hakkila et al .
( 1997 ) code , and empirical laws by bond ( 1980 ) and chen et al .
( 1998 ) . for stars closer than 75 pc ,
@xmath68 was adopted .
the arenou et al .
( 1992 ) extinction model included in hakkila et al .
code seems to systematically overestimate the reddening of stars with @xmath69 kpc , and so we have given a lower weight to the @xmath67 values obtained from arenou et al .
map for stars closer than 500 pc .
a latitude - dependent weight was adopted for the reddening obtained from the schlegel et al . and burstein & heiles maps , in such a way that the high latitude objects ( @xmath70 ) had a larger weight and the low latitude objects ( @xmath71 ) a weight that was close to zero .
there is evidence that the @xmath67 values obtained from the schlegel et al .
map , which is based on dust maps by dirbe / cobe observations , overestimates the reddening , so we have multiplied the schlegel et al .
@xmath67 values by 0.8 ( arce & goodman 1999 ; chen et al .
1999 ; beers et al . 2002 ; dutra et al .
2002 , 2003 ) .
the extinction ratios @xmath72 , or @xmath73 given in table [ t : red ] were obtained from schlegel et al .
( 1998 ) , using the appropriate effective wavelengths of the filters .
rr @xmath74 & 2.74 + @xmath75 & 2.72 + @xmath76 & 0.84 + @xmath77 & 0.56 + @xmath78 & 0.36 + @xmath79 & 0.38 + @xmath80 & 0.15 [ t : red ] our basic reference for the @xmath0}}$ ] values is the cayrel de strobel et al .
( 2001 ) catalog .
we have updated the catalog with recent spectroscopic determinations , the most important of them being mishenina & kovtyukh ( 2001 ) , santos et al .
( 2001 ) , heiter & luck ( 2003 ) , stephens & boesgaard ( 2002 ) , takada - hiday et al .
( 2002 ) , and yong & lambert ( 2003a ) . from now on
, we will refer to this updated catalog as c03 . for the stars with as many or more than four entries in the catalog , the standard deviation of the different measurements was adopted as the error bar for @xmath0}}$ ] , which was assumed to be the average of the values reported in c03 . for those with less than four entries a mean error of 0.15 dex was adopted .
based on the c03 catalog and strmgren photometry , we were able to calculate the metallicities of @xmath8120% of the sample dwarf stars from the following photometric calibration : for @xmath82 , with @xmath83 dex : @xmath84 } } & = & -4.29 - 66.0m_1 + 444.2m_1(b - y ) \nonumber \\ & & -782.4m_1(b - y)^2 \\ & & + ( 0.966 - 37.8m_1 - 1.707c_1)\log\eta\ , \nonumber\end{aligned}\ ] ] where @xmath85 $ ] . for @xmath86 , with @xmath87 dex : @xmath84 } } & = & -3.864 + 48.6m_1 - 108.5m_1 ^ 2 \nonumber\\ & & -85.2m_1(b - y)+190.6m_1 ^ 2(b - y ) \\ & & + [ 15.7m_1 - 11.1c_1 + 17.7(b - y)]c_1\ .\nonumber\end{aligned}\ ] ] for @xmath88 , with @xmath89 dex : @xmath84 } } & = & -2.63 + 26.0m_1 - 41.3m_1 ^ 2 - 45.4m_1(b - y ) \nonumber \\ & & + 74.0m_1 ^ 2(b
- y)+17.0m_1c_1\ .\end{aligned}\ ] ] a mean error of 0.2 dex was adopted for these photometric metallicities . these formulae follow schuster & nissen ( 1989 ) parameterization , but our relations are based on a much larger number of stars ( @xmath90 ) and they extend to late k stars .
the formulae are valid for @xmath91 and @xmath92}}<0.4 $ ] . for f dwarfs they extend to @xmath0}}\sim-3.5 $ ] .
the major part of the dispersion comes from the metal - poor stars , for @xmath0}}>-1.5 $ ] the standard deviations reduce to 0.1 dex .
the metallicities derived by sim04 were found consistent with c03 and were thus adopted for the sample of planet - hosting stars .
although the temperatures from the irfm and those derived from a purely spectroscopic approach differ by 100 k the metallicities are not severely affected if another @xmath3 scale is adopted ( j. melndez & i. ramrez 2005 , in preparation ) .
the photometric metallicities derived by aam96a and aam99a were also adopted , when necessary , and a mean error of 0.25 dex was assumed .
kinematic metallicities were assigned as a last resource .
for these stars ( @xmath20% of the dwarf and field giant samples ) a mean error of 0.5 dex in @xmath0}}$ ] was adopted .
the criteria was @xmath0}}=0.0 $ ] for thin disk stars , @xmath0}}=-1.0 $ ] for thick disk stars , and @xmath0}}=-1.5 $ ] for halo stars .
the temperatures for these stars are very unreliable and should be taken only as rough estimates .
most of the metallicites for the giant stars in clusters have been taken from kraft & ivans ( 2003 ) .
for the clusters not included in kraft & ivans work , aam99a values were adopted ( see aam99a for the references ) .
as we did with the metallicities , we adopted the mean values from the c03 catalog when as many or more than four entries were available .
these include both spectroscopic and hipparcos gravities and even though it has been suggested that a systematic difference of the order of 0.1 dex exists between them , it has almost no effect on the @xmath3 derived with the irfm ( see e.g. fig . 4 in aam96a ) .
for the stars with less than four entries a mean error of 0.30 dex was adopted .
the exact value of @xmath15 is not required for the irfm , a distinction between giant and dwarf is , in principle , reasonable .
thus , when no spectroscopic or hipparcos gravities were available , the @xmath15 value was estimated from the spectral type and luminosity class as given in simbad and the representative @xmath15 values given by straiys ( 1995 ) .
an error bar of 0.6 dex was assumed in this case .
the @xmath15 values of aam96a and aam99a were used for the remaining stars ( 0.7 dex of mean error ) .
the surface gravities of the cluster giants were calculated from their distance moduli ( kraft & ivans 2003 ) and the bolometric corrections of aam99a , taking @xmath93 as an estimate of the stellar masses .
bolometric fluxes have been calculated from the @xmath40 magnitudes and @xmath94 colors , using the calibrations of aam95 and aam99a , which are based on integrated @xmath95 photometry .
these calibrations have a mean accuracy of 1% and are in reasonable agreement with other flux determinations ( aam99a , rm04b ) .
systematically inaccurate fluxes lead to offsets and/or trends in the temperature scale . as an illustration
, we have calculated self - consistent bolometric fluxes and temperatures from the theoretical bolometric correction ( bc ) scale of b98 and compared them to ours for 13 dwarf stars to compare measured diameters and direct temperatures with the irfm ( see table [ t : treced ] ) . ] in fig .
[ fig : a95b98 ] .
we derived @xmath12 values from their tables of bc by calculating an initial flux @xmath96 from our irfm @xmath3 and the adopted parameters .
then , using the @xmath12 from the calibration of aam95 and @xmath96 we got @xmath97 and iterated .
this procedure is safe , because it converges simultaneously to @xmath12 and @xmath3 independently of the input values , which is not true if we use the measured angular diameters to iterate in @xmath3 .
figure [ fig : a95b98 ] shows how different bolometric flux scales lead to different temperature scales .
lower temperatures are clearly correlated to lower fluxes , although the correlation is not linear . while the difference in the fluxes increases by 7% ( from @xmath98% at @xmath99 k to @xmath100% at @xmath101 k ) , the difference in the temperature scales increases only by about 2% ( from @xmath102 k or @xmath103% at @xmath99 k to 20 k or @xmath104% at @xmath101 k ) .
also shown in fig .
[ fig : a95b98 ] as dotted lines are the flux differences that would produce a @xmath3 scale offset of @xmath10550 k in the irfm.^{2.7}$ ] . ]
note how this explains almost completely the differences found with b98 .
the comparison of aam95 fluxes with b98 may be an extreme case ( see rm04b ) so even if the adopted bolometric fluxes are inaccurate , a conservative estimate is that the irfm temperatures are reliable within 50 k of systematic error . in rm04b
we showed that the measurements of bolometric fluxes are within @xmath106% of the fluxes from aam95 calibration , so the typical errors due to the fluxes are probably around 20 k. the situation with the calibration for giants ( aam99a ) is similar , as it is shown in fig .
[ fig : fluxg ] , where the fluxes from the calibration for about 30 stars to compare measured diameters and direct temperatures with the irfm ( see table [ t : theteffg ] ) . ]
are compared to the mean of the measurements we found in the literature ( see [ sect : irfmvsdiam ] ) and the fluxes by mozurkewich et al .
( 2003 , hereafter m03 ) . as before ,
we conclude that the effect of the flux calibration on the @xmath3 scale may not be larger than 50 k of systematic error , and that typically it should be around 25 k.
from the definitions of the @xmath36 and @xmath31 factors it is shown that @xmath107 where the left hand side is calculated from the models and filter transmission functions while the right hand side is obtained entirely from observations ( adopting a calibration for the bolometric fluxes ) .
the @xmath108 factors are thus considered as the temperature indicators of the irfm . in this work ,
the bolometric flux calibration adopted has the functional form @xmath109 .
then , from eq .
( [ eq : qr ] ) we deduce the relative error in the observed @xmath108 factor : @xmath110 ^ 2+\left(\frac{\delta\phi}{\phi}\right)^2\right\}^{1/2}\ .
\label{eq : errph}\ ] ] note that @xmath111 is a function of @xmath94 and its error @xmath112 . then , adopting mean errors of @xmath113 and @xmath114 we obtained the approximate uncertainties @xmath115% plotted in fig .
[ fig : errph ] . since , roughly , @xmath116 independently of the band ( fig .
[ fig : qr ] ) , the @xmath117 given in fig .
[ fig : errph ] is a reasonable estimate of the uncertainty in @xmath3 introduced by photometric errors .
thus , in our implementation of the irfm we have adopted a mean error of 3% in @xmath108 as the error due to the photometry .
the errors introduced by the atmospheric parameters @xmath0}}$ ] and @xmath15 were explicitly considered since their updated values are a major improvement to aam work . in the best cases
four temperatures were derived , one for each of the bands : @xmath38 , @xmath39 , @xmath40 , and @xmath41 . only the first three of them were available in the dwarf sample .
the sensitivity of @xmath118 ( the @xmath108 factors in the @xmath38 band ) to @xmath3 decreases for @xmath119 k for both dwarf and giants ( fig .
[ fig : qr ] ) and thus no @xmath120 ( temperature derived from the @xmath38 band ) below 5000 k was used to get the final @xmath3 . the same reason restricted the use of @xmath121 to @xmath122 k for the dwarf sample .
also , not all the giant stars have @xmath41 magnitudes measured and so not all of them have a @xmath123 .
figure [ fig : tjhkl ] shows that , for our sample stars , the temperatures derived in the different bands are consistent .
the dispersion is larger for the dwarf sample but no trends are observed . the same result is obtained when the updated temperatures for the aam stars are included .
the adopted irfm effective temperature and its error was obtained from : @xmath124}{\sum_i(\delta t_i)^{-1}}\ \ \ \ \ , \ \ \ \
t_i)^{-1}}\ \ , \label{eq : teff}\ ] ] where @xmath125 ; and @xmath60 is the number of temperatures available .
the angular diameters were calculated from eq .
( [ eq : def1 ] ) using the adopted bolometric fluxes ( from aam95 or aam99a calibrations ) and the irfm temperatures .
their errors were estimated by propagating a mean error of 1.5 % in the bolometric fluxes and the errors in @xmath3 .
table [ table : oursample ] contains our sample stars , their adopted atmospheric parameters @xmath15 , @xmath0}}$ ] , fluxes from aam95 or aam99a calibrations , reddening corrections , @xmath94 colors in the tcs system ( transformed from either johnson or 2mass photometry ) , irfm temperatures and their errors , and angular diameters and their errors .
the table has been ordered alpha - numerically by the star standard names : hd numbers have been preferred , then bd and then g numbers .
stars with other nomenclatures have also been ordered alpha - numerically .
a more detailed version of this table as well as tables for the combined sample , including the updated @xmath3 and atmospheric parameters of the aam stars , are available upon request .
lrrcccccccccccc bd -09 1261 & 4.00 & -0.93 & 2.034e-09 & 0.000 & 2.923 & 3967 & 148 & 0.157 & 0.012 + bd -10 3166 & 4.41 & 0.45 & 2.764e-09 & 0.010 & 1.953 & 5228 & 67 & 0.105 & 0.003 + bd -13 3442 & 3.92 & -2.95 & 2.381e-09 & 0.000 & 1.226 & 6442 & 80 & 0.064 & 0.002 + bps cs 16085 - 0050 & 1.00 &
-3.10 & 4.759e-10 & 0.010 & 2.225 & 4876 & 55 & 0.050 & 0.001 + bps cs 22166 - 0030 & 4.00 & -3.36 & 1.239e-10 & 0.033 & 1.461 & 6209 & 78 & 0.016 & 0.000 + & & & & & & & & & + [ table : oursample ] by comparing the effective temperatures we derived for the aam sample with the aam96a and aam99a temperatures ( fig .
[ fig : control ] ) we find good agreement ( within error bars ) and thus show that the present @xmath3 scale is consistent with that by aam within the ranges in common . a closer inspection of fig .
[ fig : control ] reveals very small systematic differences of about 40 k ( 0.7% at solar temperature ) for dwarf stars and 25 k ( 0.6% at 4500 k ) for giants , which are , however , well within the uncertainties of our results ( 1.3% ) .
the scatter is due to the random nature of the differences between our input data and that adopted by aam , and the extreme cases can be easily identified by comparing the two sets of input data . less than 9 % ( 20 % ) of the giant ( dwarf ) stars in fig .
[ fig : control ] exhibit differences larger than our mean uncertainty of 60 k ( 75 k ) .
if we take into account the uncertainties in aam temperatures these numbers are even smaller .
one important improvement to the irfm @xmath3 scale is the use of more accurate input parameters .
even though individually the temperature differences between the temperatures we obtained and those by aam are relatively small , overall they greately help to reduce the dispersion in the @xmath126}}$ ] relations ( part ii ) , thus allowing to separate real effects from random errors .
the revised irfm temperature scale may then have a better _ internal _ accuracy , which is a very useful improvement , particularly in the light of the need to reduce ( or even better , recognize ) the sources of uncertainty in chemical abundance studies .
with the internal uncertainty reduced , the systematics can be better quantified .
also important is the effort to include stars representative of regions of particular interest in the atmospheric parameters space , like the metal - rich ( e.g. for planetary - host star studies ) and metal - poor ( e.g. for globular cluster and halo star studies ) extremes . as an example of the importance of these selected regions we show in fig .
[ fig : vk ] the @xmath3 vs. @xmath94 relation defined by our sample stars as well as aam96b relations for @xmath0}}=0 $ ] and @xmath0}}=-3 $ ] ( photometry sources and reddening corrections employed are discussed in detail in part ii ) .
the aam96b relation for metal - poor cool dwarfs seems to be in agreement with ours ( even if the aam96b relation is slightly extrapolated ) but the corresponding aam96b relation for hot metal - poor stars differs from ours .
the ` redder ' points near 6500 k on the dwarf @xmath3 vs. @xmath94 plane are all very metal - poor stars ( @xmath0}}\sim-3 $ ] ) .
this is the place where our results show the most significant deviations with respect to aam temperatures .
in part ii we show that the synthetic @xmath3 scales based on both marcs and kurucz models for metal - poor stars are consistent with our temperatures but not with aam96b in this region .
the inclusion of these stars allows to better define the trends of the irfm temperature scale in that region , which has proven to be of fundamental importance , for example , in the controversy of the primordial lithium abundance ( melndez & ramrez 2004 ) .
since we have shown that our stars effective temperatures are essentially consistent with the aam @xmath3 scale , in the comparisons below , the stars with irfm temperatures are those for the whole sample ( including the aam stars ) , for which we have homogeneously calculated irfm temperatures .
the irfm was explicitly developed to solve for both @xmath3 and @xmath11 simultaneously from the observed and theoretical data .
one of the fundamental tests of validity of the irfm is thus the comparison with measured angular diameters .
richichi & percheron ( 2002 , hereafter charm ) have collected most of the published stellar angular diameters to date , measured mainly with lunar occultation ( lo ) and long - baseline interferometry ( lbi ) techniques .
a large number of giant stars constitute the major part of the catalog and only a small fraction of main sequence stars is included .
we have compared the angular diameters that result from our irfm implementation with those given in charm in fig .
[ fig : theg ] ( see also table [ t : theteffg ] ) . the stars in charm
have been grouped according to the technique employed to derive their angular diameters .
first , a comparison is made with the indirect ( ind ) diameters of cohen et al .
( 1999 ) , then with lbi , and finally with lo diameters .
rrrrrrccrrlr 3546 & -0.72 & 1.77 @xmath105 0.08 & lbi & & 4935 @xmath105 52 & 5.81e-07 & & & 1.71 @xmath105 0.05 & 4857 @xmath105 111 & + 3627 & 0.17 & 4.21 @xmath105 0.05 & lbi & 4.14 @xmath105 0.04 & 4343 @xmath105 45 & 2.07e-06 & 2.12e-06 & 2.12e-06 & 4.18 @xmath105 0.11 & 4355 @xmath105 30 & 4392 @xmath105 27 + 3712 & -0.10 & 5.58 @xmath105 0.13 & lbi & 5.61 @xmath105 0.06 & 4553 @xmath105 48 & 4.68e-06 & 4.70e-06 & 4.70e-06 & 5.72 @xmath105 0.15 & 4614 @xmath105 55 & 4602 @xmath105 29 + 6860 & -0.07 & 14.08 @xmath105 0.76 & lbi & 13.75 @xmath105 0.14 & 3824 @xmath105 41 & 1.20e-05 & 1.26e-05 & 1.26e-05 & 12.98 @xmath105 0.34 & 3719 @xmath105 102 & 3763 @xmath105 23 + 9927 & -0.01 & 3.76 @xmath105 0.07 & lbi & & 4380 @xmath105 48 & 1.56e-06 & 1.60e-06 & & 3.57 @xmath105 0.09 & 4292 @xmath105 43 & + 10380 & -0.27 & 2.81 @xmath105 0.03 & lbi & & 4132 @xmath105 46 & 8.56e-07 & & & 2.97 @xmath105 0.08 & 4247 @xmath105 28 & + 12533 & -0.07 & 7.73 @xmath105 0.20 & lbi & 7.81 @xmath105 0.08 & 4259 @xmath105 45 & 6.12e-06 & 6.66e-06 & 6.66e-06 & 7.47 @xmath105 0.19 & 4277 @xmath105 57 & 4254 @xmath105 27 + 12929 & -0.25 & 6.87 @xmath105 0.04 & lbi & 6.83 @xmath105 0.07 & 4501 @xmath105 50 & 6.19e-06 & 6.33e-06 & 6.33e-06 & 6.73 @xmath105 0.18 & 4478 @xmath105 21 & 4493 @xmath105 28 + 18884 & 0.00 & & & 13.24 @xmath105 0.26 & 3718 @xmath105 46 & 9.42e-06 & & 9.57e-06 & 12.17 @xmath105 0.35 & & 3578 @xmath105 37 + 29139 & -0.18 & 20.87 @xmath105 0.57 & lbi & 21.10 @xmath105 0.21 & 3883 @xmath105 44 & 3.06e-05 & 3.34e-05 & 3.33e-05 & 20.10 @xmath105 0.55 & 3895 @xmath105 56 & 3871 @xmath105 24 + 29139 & -0.18 & 20.26 @xmath105 1.63 & lo & & 3883 @xmath105 44 & 3.06e-05 & 3.35e-05 & & 20.10 @xmath105 0.55 & 3956 @xmath105 160 & + 44478 & -0.05 & 13.97 @xmath105 0.16 & lbi & 15.12 @xmath105 0.15 & 3667 @xmath105 44 & 1.11e-05 & 1.12e-05 & 1.12e-05 & 13.55 @xmath105 0.38 & 3623 @xmath105 25 & 3483 @xmath105 22 + 44478 & -0.05 & 13.45 @xmath105 0.21 & lo & & 3667 @xmath105 44 & 1.11e-05 & & & 13.55 @xmath105 0.38 & 3684 @xmath105 32 & + 61338 & 0.00 & 3.07 @xmath105 0.78 & lo & & 3869 @xmath105 48 & 7.44e-07 & & & 3.16 @xmath105 0.09 & 3925 @xmath105 498 & + 62509 & -0.02 & 7.98 @xmath105 0.06 & lbi & 7.98 @xmath105 0.08 & 4833 @xmath105 50 & 1.16e-05 & 1.15e-05 & 1.18e-05 & 7.99 @xmath105 0.20 & 4823 @xmath105 25 & 4858 @xmath105 30 + 62721 & -0.27 & 3.01 @xmath105 0.00 & lbi & & 3988 @xmath105 48 & 6.94e-07 & & & 2.87 @xmath105 0.08 & 3894 @xmath105 15 & + 76294 & -0.01 & 3.24 @xmath105 0.08 & lbi & & 4817 @xmath105 50 & 1.86e-06 & & & 3.22 @xmath105 0.08 & 4806 @xmath105 61 & + 80493 & -0.26 & 7.69 @xmath105 0.26 & lbi & 7.54 @xmath105 0.08 & 3851 @xmath105 42 & 4.03e-06 & 4.10e-06 & 4.10e-06 & 7.42 @xmath105 0.20 & 3798 @xmath105 65 & 3836 @xmath105 24 + 94264 & -0.20 & 2.54 @xmath105 0.03 & lbi & & 4670 @xmath105 51 & 1.05e-06 & & & 2.58 @xmath105 0.07 & 4701 @xmath105 33 & + 99998 & -0.39 & 3.21 @xmath105 0.03 & lbi & & 3919 @xmath105 45 & 7.98e-07 & & & 3.19 @xmath105 0.09 & 3905 @xmath105 23 & + 99998 & -0.39 & 3.20 @xmath105 0.51 & lo & & 3919 @xmath105 45 & 7.98e-07 & & & 3.19 @xmath105 0.09 & 3911 @xmath105 312 & + 102224 & -0.44 & 3.23 @xmath105 0.02 & lbi & & 4378 @xmath105 46 & 1.36e-06 & & & 3.33 @xmath105 0.09 & 4447 @xmath105 22 & + 112142 & 0.00 & 5.90 @xmath105 0.70 & lo & & 3647 @xmath105 47 & 1.75e-06 & & & 5.45 @xmath105 0.16 & 3504 @xmath105 208 & + 113226 & 0.11 & 3.23 @xmath105 0.06 & lbi & 3.28 @xmath105 0.03 & 5049 @xmath105 59 & 2.25e-06 & 2.23e-06 & 2.21e-06 & 3.23 @xmath105 0.09 & 5035 @xmath105 47 & 4981 @xmath105 31 + 124897 & -0.55 & 20.87 @xmath105 0.52 & lbi & 21.37 @xmath105 0.25 & 4231 @xmath105 49 & 4.69e-05 & 4.89e-05 & 4.86e-05 & 20.95 @xmath105 0.58 & 4284 @xmath105 56 & 4226 @xmath105 29 + 135722 & -0.40 & 2.75 @xmath105 0.01 & lbi & 2.76 @xmath105 0.03 & 4834 @xmath105 50 & 1.35e-06 & 1.37e-06 & 1.41e-06 & 2.72 @xmath105 0.07 & 4831 @xmath105 20 & 4851 @xmath105 32 + 150997 & -0.28 & 2.52 @xmath105 0.11 & lbi & 2.62 @xmath105 0.03 & 4948 @xmath105 54 & 1.26e-06 & 1.27e-06 & 1.26e-06 & 2.52 @xmath105 0.07 & 4947 @xmath105 111 & 4841 @xmath105 36 + 164058 & -0.15 & 10.04 @xmath105 0.21 & lbi & 9.86 @xmath105 0.13 & 3927 @xmath105 42 & 8.29e-06 & 8.38e-06 & 8.40e-06 & 10.23 @xmath105 0.27 & 3974 @xmath105 45 & 4013 @xmath105 30 + 189319 & 0.00 & 6.09 @xmath105 1.35 & lbi & 6.23 @xmath105 0.06 & 3877 @xmath105 41 & 2.75e-06 & 2.86e-06 & 2.86e-06 & 6.04 @xmath105 0.16 & 3899 @xmath105 433 & 3858 @xmath105 24 + 197989 & -0.12 & 5.23 @xmath105 0.60 & lbi & 4.61 @xmath105 0.05 & 4710 @xmath105 52 & 3.57e-06 & 3.61e-06 & 3.63e-06 & 4.67 @xmath105 0.12 & & 4757 @xmath105 30 + 198149 & -0.18 & 2.65 @xmath105 0.04 & lbi & & 4907 @xmath105 54 & 1.38e-06 & 1.42e-06 & & 2.67 @xmath105 0.07 & 4960 @xmath105 42 & + 210745 & 0.25 & 5.31 @xmath105 0.11 & lbi & 5.23 @xmath105 0.05 & 4482 @xmath105 51 & 3.21e-06 & 3.27e-06 & 3.27e-06 & 4.89 @xmath105 0.13 & & 4351 @xmath105 27 + 214868 & -0.25 & 2.63 @xmath105 0.05 & lbi & & 4303 @xmath105 47 & 7.69e-07 & 7.59e-07 & & 2.60 @xmath105 0.07 & 4260 @xmath105 44 & + 217906 & -0.11 & 17.02 @xmath105 1.42 & lbi & 17.98 @xmath105 0.02 & 3648 @xmath105 43 & 1.50e-05 & 1.55e-05 & 1.52e-05 & 15.93 @xmath105 0.45 & 3557 @xmath105 149 & 3448 @xmath105 13 + 221115 & 0.04 & 1.61 @xmath105 0.17 & lbi & & 4980 @xmath105 63 & 4.79e-07 & 4.89e-07 & & 1.53 @xmath105 0.05 & 4877 @xmath105 258 & + 222107 & -0.50 & 2.66 @xmath105 0.08 & lbi & & 4605 @xmath105 49 & 1.09e-06 & & & 2.70 @xmath105 0.07 & 4637 @xmath105 72 & [ t : theteffg ]
the ind diameters are in excellent agreement with the irfm for @xmath127 mas , where stars in common are found . these are diameters derived from absolutely calibrated stellar spectra in the infrared , which is the main reason for the good agreement .
these are not direct measurements and will not be used to derive direct temperatures in the comparisons below .
the lbi technique uses the interference of the wavefronts from the starlight entering the telescope to resolve smaller angular sizes . for the stars with large diameters ,
more than one lbi entry is often found in charm .
the mean value of those entries and its standard deviation for each star were adopted for the comparison ( hence the large error bars of some points in fig .
[ fig : theg ] ) , otherwise the single value and its internal error were used .
comparison with the irfm for 29 stars shows agreement within 1.5 mas for @xmath128 mas and better than 0.5 mas below that value .
less measurements with lo are available due to the limited sky coverage of the moon ( @xmath129 ) and other observational difficulties .
nevertheless , the lo diameters are reliable and have been shown to be well scaled to the lbi ones . when compared to the irfm , a very good agreement is also found . the differences with the five lo measurements ( average in some cases ) are less than 0.5 mas in the range @xmath130 mas .
from the charm angular diameters and the bolometric flux measurements of blackwell et al .
( 1979 , 1990 ) , blackwell & lynas - gray ( 1994 , 1998 ) , leggett et al .
( 1986 ) , and m03 ; we obtained the direct temperatures of the stars in common with our field giant sample and compared them with those from the present irfm implementation in fig .
[ fig : teffg ] ( see also table [ t : theteffg ] ) . when no flux measurements were available ,
the fluxes from aam99a calibration were adopted .
although the error bars are larger than 200 k for some stars , we may safely state that a reasonable agreement is found in the ranges covered , i.e. 3500 k@xmath6 k and @xmath7}}<0.2 $ ] .
the mean difference @xmath131 , with the direct temperatures from lbi measurements only , is @xmath132 k. the irfm temperatures start to become larger than the direct ones at @xmath133 k. discarding the points below that limit , the difference reduces to only @xmath134 k. the zero point of the irfm @xmath3 scale for giants may thus be the absolute one , but the reader should be cautioned that this result is currently only proved for stars in the ranges : 3800 k@xmath6 k and @xmath7}}<0.2 $ ] , where the comparison is made and no trends are found .
a more recent interferometric work is that of m03 , who determined the angular diameters of 85 stars , 19 of which are included in our field giant sample .
these diameters allow for a better comparison than the previous one with charm given the homogeneity of m03 results .
the comparison between our irfm diameters and the measurements is shown in fig .
[ fig : thegm ] ( see also table [ t : theteffg ] ) . below @xmath135
mas the agreement is excellent but above that limit the irfm diameters appear to be lower than those by m03 .
the reason for this is the very low temperatures of the latter group of stars . as we stated above
, the irfm produces high temperatures compared to the direct ones below 3800 k. thus , we have plotted the stars with @xmath133 k with open symbols in fig .
[ fig : thegm ] to show that they are the ones that depart most from the 1:1 line . note also that m03 work uses optical interferometry and so their corrections from uniform - disk to limb - darkened diameters may be inaccurate due to uncertainties in the modeling of the tio lines ( dyck & nordgren 2002 ) .
m03 also measured bolometric fluxes and provide direct temperatures .
they are compared to the irfm temperatures of the 19 stars in common in fig .
[ fig : teffgm ] . as we did before ,
the stars with @xmath133 k have been plotted with open symbols .
it is more clear now that the irfm temperatures exceed the direct ones for the very late - type giants .
excluding them , however , the agreement is excellent , as the mean difference @xmath131 is only @xmath136 k. the excellent agreement found with the measured angular diameters and direct temperatures of the giant stars is not completely surprising , given that the monochromatic flux calibration in the infrared was derived in such a way that the direct temperatures of a given number of giant stars were well reproduced by the irfm ( aam94a ) .
the situation for the main sequence stars is more complicated since the angular diameters of dwarf stars are very difficult to measure .
kervella et al .
( 2004 ) have collected uniform - disk angular diameters of several dwarfs and subgiants , all measured by interferometry , and have uniformly corrected them for limb darkening . in fig .
[ fig : thed1 ] we compare the irfm diameters with those given by kervella et al .
( 2004 ) for 10 dwarf stars and 2 subgiants ( see also table [ t : treced ] ) .
the two subgiants are hd 121370 and hd 198149 .
the original sources for the measured angular diameters include lane et al .
( 2001 ) , kervella et al .
( 2003 ) , nordgren et al .
( 1999 , 2001 ) , and m03 ; as well as some other papers in preparation ( see kervella et al .
2004 ) . for procyon ( hd 61421 ) and hd 121370 more than one value is given by kervella et al .
( 2004 ) ; only the more accurate vlti measurement was considered .
also shown in fig .
[ fig : thed1 ] and table [ t : treced ] is the comparison of the irfm results with the transit observations of brown et al .
( 2001 ) for the planet - hosting star hd 209458 .
rrrlccccl 10700 & -0.54 & 2.08 @xmath105 0.03 & 5372 @xmath105 65 & 1.163e-06 & 1.154e-06 & 1,2,3 & 2.047 @xmath105 0.059 & 5319 @xmath105 43 + 16160 & -0.03 & 0.94 @xmath105 0.07 & 4714 @xmath105 67 & 1.687e-07 & 1.725e-07 & 2,3 & 1.013 @xmath105 0.032 & 4930 @xmath105 185 + 22049 & -0.12 & 2.15 @xmath105 0.03 & 5015 @xmath105 56 & 1.024e-06 & & & 2.204 @xmath105 0.059 & 5078 @xmath105 40 + 26965 & -0.28 & 1.65 @xmath105 0.06 & 5068 @xmath105 63 & 5.313e-07 & 5.313e-07 & 1,2,3 & 1.555 @xmath105 0.045 & 4920 @xmath105 91 + 61421 & -0.01 & 5.48 @xmath105 0.05 & 6591 @xmath105 73 & 1.837e-05 & 1.822e-05 & 3,4,5,6,7,8 & 5.405 @xmath105 0.145 & 6532 @xmath105 39 + 88230 & -0.12 & 1.29 @xmath105 0.04 & 3950 @xmath105 161 & 1.373e-07 & & & 1.301 @xmath105 0.108 & 3967 @xmath105 63 + 121370 & 0.25 & 2.20 @xmath105 0.03 & 6038 @xmath105 75 & 2.210e-06 & 2.206e-06 & 1,8 & 2.234 @xmath105 0.065 & 6081 @xmath105 47 + 128620 & 0.20 & 8.51 @xmath105 0.02 & 5759 @xmath105 70 & 2.677e-05 & & & 8.548 @xmath105 0.244 & 5771 @xmath105 23 + 128621 & 0.20 & 6.01 @xmath105 0.03 & 5201 @xmath105 65 & 8.653e-06 & & & 5.957 @xmath105 0.173 & 5178 @xmath105 23 + 131977 & 0.09 & 1.23 @xmath105 0.03 & 4571 @xmath105 52 & 1.921e-07 & 2.010e-07 & 2 & 1.149 @xmath105 0.031 & 4469 @xmath105 57 + 198149 & -0.18 & 2.67 @xmath105 0.04 & 4907 @xmath105 54 & 1.377e-06 & 1.414e-06 & 1 & 2.670 @xmath105 0.071 & 4939 @xmath105 41 + 209100 & -0.02 & 1.89 @xmath105 0.02 & 4642 @xmath105 54 & 5.069e-07 & 5.000e-07 & 2 & 1.810 @xmath105 0.050 & 4527 @xmath105 29 + 209458 & -0.01 & 0.23 @xmath105 0.02 & 5993 @xmath105 71 & 2.279e-08 & & & 0.230 @xmath105 0.006 & 6049 @xmath105 202
[ t : treced ] the comparison shows that the irfm diameters are in excellent agreement with the measurements , to better than 0.2 mas in @xmath11 .
also , this result seems to be valid for the wide temperature range from 4000 k to 6500 k , although it is restricted to almost solar metallicity ( fig .
[ fig : thed2 ] ) . the stars in table [ t : treced ] span the metallicity range from @xmath137 to 0.25 dex .
table [ t : treced ] also contains the bolometric fluxes from aam95 calibration and those measured by several authors . when a measured @xmath12 was available , the direct temperature was derived from it and the measured angular diameter using eq .
( [ eq : def1 ] ) , otherwise the flux from aam95 calibration was adopted .
the mean of the measurements of @xmath12 is given in table [ t : treced ] .
the direct temperatures and their errors are also given in table [ t : treced ] , a mean error of 1.5% in @xmath12 was assumed .
the comparison between direct and irfm temperatures is shown in fig .
[ fig : teffd ] .
the three points that depart most from the 1:1 line are hd 16160 , hd 26965 and hd 209458 .
the uncertainty in the angular diameters measured for these stars is larger than 4% and , thus , they are not useful to constraint any @xmath3 scale within 200 k. excluding these three points , the mean difference @xmath138 is @xmath139 k. consequently , the zero point of the irfm @xmath3 scale for dwarfs is also in good agreement with the direct @xmath3 scale , given that the @xmath139 k difference is well within the average uncertainty of the irfm temperatures ( @xmath140 k ) , at least for stars with temperatures between 4000 k and 6500 k and metallicities in the range @xmath9}}<0.25 $ ] .
the temperatures of these three stars are very well documented , although no general agreement concerning their exact values ( the sun being the exception ) has been achieved , and should be used to test any @xmath3 scale . using the photometry given by aam94a ,
the present irfm temperature of the sun results in @xmath141 k which is 21 k lower than the standard value .
aam96a obtained @xmath142 k and thus the difference with our solar @xmath3 may only be attributed to the details of the calculations .
the fact that the irfm does not exactly reproduce the solar temperature has been suggested to imply that the zero point of the irfm @xmath3 scale is off by some amount .
this is by no means conclusive , as the photometry of the sun is very uncertain and the errors in the observed magnitudes and colors are easily propagated to the temperature derived with this method .
procyon ( hd 61421 ) , whose spectral type , as given in simbad , is f5iv - v , has an irfm temperature of @xmath143 k. aam96a obtained @xmath144 k , the difference is mainly due to a slightly different set of @xmath0}}$ ] and @xmath15 values adopted .
its direct temperature is @xmath145 k , as given in table [ t : treced ] .
these different results may be consistent , considering the size of the error bars .
being a bright star , the photometry has a considerable uncertainty and may be the reason for the discrepancy , if any . using the spectroscopic method , takeda et al .
( 2002b , hereafter t02b ) obtained 6576 k ; edvardsson et al .
( 1993 , hereafter e93 ) , on the other hand , obtained 6704 k from a theoretical calibration of the @xmath146 color .
a fit to spectrophotometric data by castelli et al .
( 1997 ) resulted in @xmath147 k. fitting of the balmer line - profiles in the spectrum of procyon leads to @xmath148 k , @xmath149 k , and @xmath150 k according to castelli et al .
( 1997 ) , f98 and ap04 , respectively . for the giant star arcturus ( hd 124897 )
, the present irfm implementation gives @xmath151 k. that of aam99a was @xmath152 k. its direct temperature , as given by griffin & lynas - gray ( 1999 ) is @xmath153 k ; the value we derived for this star in table [ t : theteffg ] from several angular diameter and bolometric flux measurements is @xmath154 k , in perfect agreement with griffin & lynas - gray result .
m03 also derived a direct temperature for arcturus , which is slightly lower ( 4226 k ) , but in better agreement with the irfm .
although the irfm temperature of arcturus is lower by about 60 k , the temperature derived from several temperature - color calibrations based on the irfm is closer to the direct one ( 4296 k , rm04a ) .
thus , the use of several of these calibrations allows to get rid of errors in single irfm temperatures , which are mainly based on infrared magnitudes alone .
we will come back to this point in part ii of this work .
in [ sect : intro ] we discussed some of the other methods developed to derive @xmath3 . as it is the case with the irfm
, they all introduce models to complement the available observations and so the differences that we will describe below are partly due to the different models adopted , although the observational input parameters may also be the reason of the discrepancies , if any . the comparison with the spectroscopic temperatures derived by three different groups is given in fig . [ fig : compsp ] . in this and the next two figures the filled circles are dwarf stars , the open circles are field giants and the open squares are giants in clusters .
t02b applied spectroscopic equilibrium to obtain the temperatures of 32 f - k dwarfs and subgiants , 23 of which are included in our dwarf sample .
the comparison shows that the irfm temperatures are higher by about 35 k but the dispersion is large ( @xmath155 k ) , mainly due to a metallicity effect . below solar metallicity ,
t02b temperatures are systematically lower ; excluding the stars with @xmath0}}<0 $ ] , the mean difference reduces to 10 k , but then a clear trend with temperature is observed .
comparison with b03 spectroscopic temperatures of a sample of thin and thick disk stars suggests a mean difference of @xmath156 k , in which the dispersion is , again , mainly due to a trend of the differences with metallicity .
it seems that the differences reduce to zero at @xmath0}}\sim0.25 $ ] and the b03 temperatures become systematically higher by 110 k as lower metallicities are reached .
a similar difference was found in rm04b between the irfm temperatures and the spectroscopic ones of sim04 for the planet - hosting stars but that one did not show any trend with metallicity , and only a slight trend with temperature , probably due to non - lte effects and/or the bolometric fluxes adopted for the irfm . in fig .
[ fig : compsp ] is also shown a comparison of the irfm temperatures of 21 metal - poor field giants ( @xmath157}}<-0.55 $ ] ) and 14 giants in the globular clusters m3 ( @xmath0}}\sim-1.5 $ ] ) , m71 ( @xmath0}}\sim-0.8 $ ] ) and m92 ( @xmath0}}\sim-2.4 $ ] ) with those derived spectroscopically by sneden et al .
( 1991 ) , kraft et al .
( 1992 ) , and shetrone ( 1996 ) , which we refer jointly as lt ( lick - texas group ) .
the differences have a mean value of @xmath158 k , they show no large trends with @xmath3 nor @xmath0}}$ ] , and it is also independent of the population , i.e. it is basically the same for the giants in the field and in the three clusters .
thus , besides an offset of 65 k , the spectroscopic @xmath3 scale of the metal - poor @xmath159}}<-0.5)$ ] giants is differentially consistent with the irfm . using spectra synthesized from marcs models , the filter transmission functions of the @xmath22-@xmath23 system , and a set of stars to fix the zero point ;
e93 calculated a grid of synthetic colors from which the temperatures of their sample stars were determined .
we found 50 stars ( 1 giant ) in common with our sample .
the comparison with the irfm is shown in fig .
[ fig : compeb ] where a mean difference of @xmath160 k is observed .
no trend is apparent with metallicity , and probably a slight trend with temperature is present , in the sense that the irfm temperatures and those by e93 may be in agreement for @xmath161 k but with a decreasing difference for lower temperatures . two sets of temperatures based on the fitting of the balmer line - profiles are shown in fig .
[ fig : compeb ] . note how even though f98 and ap04 temperatures were derived using the same principle , the theoretical prescriptions adopted have an important effect on the final results .
f98 temperatures are higher by about 100 k below 5800 k but slightly lower than the irfm ones above 6000 k. there is apparent agreement at solar metallicity but disagreement both above and below @xmath0}}\sim0 $ ] .
the solar temperature is thus well reproduced by both the irfm and f98 . on the other hand ,
ap04 temperatures show very good agreement with ours in the wide temperature range from 4800 k to 6600 k and it is likely that they are close to the irfm ones even at low metallicites ( @xmath0}}\sim-0.8 $ ] ) .
note that the zero point of the e93 theoretical calibration was set using a sample of stars with temperatures derived from balmer line - profile fitting ( cayrel de strobel 1990 ) and so the similarity between the differences found with e93 and f98 may be due to the same basic technique employed .
we recall , however , that this technique is very model dependent , as shown by the most recent and improved results of ap04 , and that it is very likely that the results from the irfm are in good agreement with those from the modeling of hydrogen lines .
a method that was not discussed in
[ sect : intro ] is the one that uses surface brightness vs. color relations based on measurements of stellar angular diameters ( di benedetto 1998 , hereafter db98 ) .
since the irfm is also based on measured angular diameters , the zero points of both the irfm and the db98 @xmath3 scales should be the same .
this is actually shown in fig .
[ fig : compdik ] , where the temperatures of 25 stars ( 12 giants ) in common with our sample are compared to those given by db98 .
the mean difference is @xmath162 k , as expected , and the dispersion is fully consistent with the mean errors of the irfm and db98 temperatures .
the good agreement of the zero point is also true if dwarfs and giants are considered separately .
note , however , that the comparison with db98 results is limited to @xmath0}}>-0.6 $ ] , although the only point below @xmath163 dex , a giant star with @xmath0}}=-1.64 $ ] , shows a db98 temperature consistent with the irfm .
finally , the line - depth ratio technique has been applied to 181 dwarfs by k03 , 44 of them are included in our sample .
it is claimed that very small errors , of the order of 10 k or less , may be achieved with this technique if several calibrated ratios are used . the comparison with our irfm temperatures , however , shows a dispersion that is larger than the one expected by the combination of the errors in the irfm temperatures ( @xmath140 k ) with those suggested by k03 , which will be essentially equal to the uncertainty from the irfm .
it is true that very small _ individual _ uncertainties ( that is , for a single star ) may be achieved with the line - depth ratio technique , but some systematic effects may enlarge the dispersion in the @xmath3 scale derived with this technique , which is , however , potentially useful for studying intrinsic temperature fluctuations .
k03 adjusted their @xmath3 scale to reproduce the solar temperature , hence the good agreement with the irfm around 5800 k and solar metallicity . from 5500 k to 6200 k , however , k03 temperatures become lower than the irfm ones . from 4500 k to 5500 k , a mean difference of about @xmath164 k is observed , and it seems to go to zero again at lower temperatures .
effective temperatures with a mean accuracy of about 1.3% have been determined with the infrared flux method for a selected sample of stars as well as for the alonso et al .
( aam ) sample with the aim of calibrating the metallicity - dependent temperature vs. color relations in a companion publication .
the status of this particular irfm implementation is revised and it is shown that the irfm temperatures are well scaled to the direct ones for main sequence stars . the case for giant stars
was proved in the past and the present work confirms the good agreement .
the present work improves the irfm @xmath3 scale by the use of up - to - date spectroscopic atmospheric parameters and the inclusion of stars in the ranges where few or no data were previously studied .
our irfm approach , however , is largely based on the aam study , and a comparison of our updated to aam temperatures shows agreement within error bars .
we point out that the influence of the bolometric flux calibrations adopted on our irfm temperatures is important because they may introduce systematic trends in the resulting @xmath3 scale .
however , we have also shown that the systematic trends may not be larger than 50 k , and so , if differences larger than 50 k are found with other @xmath3 scales , they should not be attributed solely to the fluxes adopted in the irfm .
this number ( 50 k ) is a conservative estimate , and thus even smaller differences may not be totally due to errors in the bolometric fluxes .
comparison of the stellar angular diameters from the irfm with those measured by interferometry , lunar occultation or transit observations shows good agreement .
this work thus provides accurate stellar angular diameters , which are essential to calibrate extremely high - resolution long - baseline interferometric observations ( e.g. vlt and chara interferometers ) .
when combined with bolometric flux measurements , the interferometric angular diameters produce a temperature scale with a zero point that is only 10 k below that derived with the irfm for the giant stars with @xmath0}}>-0.7 $ ] and temperatures above 3800 k and below 5000 k. the corresponding zero points of the dwarf @xmath3 scales differ by only about 20 k ( the irfm being larger ) , and it is valid for @xmath0}}>-0.55 $ ] and temperatures between 4000 k and 6500 k. no stars outside those ranges were available for comparison , and we can not guarantee that the zero point of the irfm @xmath3 scale is essentially the absolute one everywhere in the atmospheric parameters space , as it is the case in the ranges mentioned above .
the irfm temperature of the sun is only about 20 k below the standard value , and those of the well studied stars procyon and arcturus are consistent with the direct ones and those derived with other methods within the error bars .
no perfect agreement is expected for these bright stars because the photometric uncertainties associated are large and easily propagated in the irfm .
consequently , no empirical shifts to reproduce the temperature of the sun or other well studied stars were applied .
the comparison with the temperatures obtained with other methods shows that the temperature scale of the lower main sequence is still unreliable at the 100 k level because larger discrepancies are found .
although consistent with the temperatures from balmer line - profile fitting , the irfm produces temperatures that are about 100 k lower that those derived from the excitation equilibrium of , but exceptions are found .
the temperature scale of giant stars is better defined , probably to a 50 k level or even better .
the spectroscopic temperatures of the metal - poor giants , both in the field and in clusters , are about 65 k higher that those from the irfm , but otherwise differentially consistent with them . with the irfm temperatures described in this paper and the atmospheric parameters adopted ,
we will calibrate a set of homogeneous temperature vs. color relations ( part ii ) .
the fact that the input parameters are somewhat inhomogeneous ( several sources have been considered ) will not have a significant effect on these relations but they may have some effect on the irfm temperatures .
heiter , u. , & luck , r. e. 2003 , abundance analysis of planetary host stars . in proceedings of iau symp .
210 , modelling of stellar atmospheres " , ed .
n. piskunov , w.w .
weiss , & d.f .
gray , asp conference series | the infrared flux method ( irfm ) has been applied to a sample of 135 dwarf and 36 giant stars covering the following regions of the atmospheric parameters space : 1 ) the metal - rich ( @xmath0}}\gtrsim0 $ ] ) end ( consisting mostly of planet - hosting stars ) , 2 ) the cool ( @xmath1 k ) metal - poor ( @xmath2}}\lesssim-3 $ ] ) dwarf region , and 3 ) the very metal - poor ( @xmath0}}\lesssim-2.5 $ ] ) end .
these stars were especially selected to cover gaps in previous works on @xmath3 vs. color relations , particularly the irfm @xmath3 scale of a. alonso and collaborators .
our irfm implementation was largely based on the alonso et al .
study ( absolute infrared flux calibration , bolometric flux calibration , etc . ) with the aim of extending the ranges of applicability of their @xmath3 vs. color calibrations .
in addition , in order to improve the internal accuracy of the irfm @xmath3 scale , we recomputed the temperatures of almost all stars from the alonso et al .
work using updated input data .
the updated temperatures do not significantly differ from the original ones , with few exceptions , leaving the @xmath3 scale of alonso et al .
mostly unchanged . including the stars with updated temperatures , a large sample of 580 dwarf and 470 giant stars ( in the field and in clusters ) , which cover the ranges : 3600 k@xmath4 k , @xmath5}}\lesssim+0.5 $ ] , have @xmath3 homogeneously determined with the irfm .
the mean uncertainty of the temperatures derived is 75 k for dwarfs and 60 k for giants , which is about 1.3% at solar temperature and 4500 k , respectively .
it is shown that the irfm temperatures are reliable in an absolute scale given the consistency of the angular diameters resulting from the irfm with those measured by long - baseline interferometry , lunar occultation and transit observations . using the measured angular diameters and bolometric fluxes
, a comparison is made between irfm and direct temperatures , which shows excellent agreement , with the mean difference being less than 10 k for giants and about 20 k for dwarf stars ( the irfm temperatures being larger in both cases ) .
this result was obtained for giants in the ranges : 3800 k@xmath6 k , @xmath7}}<0.2 $ ] ; and dwarfs in the ranges : 4000 k@xmath8 k , @xmath9}}<0.25 $ ] ; and thus the zero point of the irfm @xmath3 scale is essentially the absolute one ( that derived from angular diameters and bolometric fluxes ) within these limits .
the influence of the bolometric flux calibration adopted is explored and it is shown that its effect on the @xmath3 scale , although systematic , is conservatively no larger than 50 k. finally , a comparison with temperatures derived with other techniques is made .
agreement is found with the temperatures from balmer line - profile fitting and the surface - brightness technique .
the temperatures derived from the spectroscopic equilibrium of lines are differentially consistent with the irfm but a systematic difference of about 100 k and 65 k ( the irfm temperatures being lower ) is observed in the metal - rich dwarf and metal - poor giant @xmath3 scales , respectively . |
the data were collected with the belle detector at kekb @xcite , an asymmetric energy double storage ring collider with @xmath218.2 gev electrons and @xmath213.6 gev positrons . in this analysis ,
a data sample of @xmath22 taken at the @xmath2(5s ) cm energy of @xmath2110869 mev is used .
a data sample of @xmath23 taken in the continuum , 60 mev in center - of - mass energy below the @xmath2(4s ) , and a data sample of @xmath24 taken at the @xmath2(4s ) resonance energy , were also used for this analysis .
the continuum and @xmath2(4s ) data , which were collected with the same vertex detector , provide comparable detector systematics .
all experimental conditions of data - taking at the @xmath2(5s ) were nearly the same as at the @xmath2(4s ) or continuum runs except for beam energies and luminosities .
the belle detector is a general - purpose large - solid - angle magnetic spectrometer that consists of a silicon vertex detector ( svd ) , a central drift chamber ( cdc ) , an array of aerogel threshold cherenkov counters ( acc ) , a barrel - like arrangement of time - of - flight scintillation counters ( tof ) , and an electromagnetic calorimeter comprised of csi(tl ) crystals ( ecl ) located inside a superconducting solenoidal coil with a 1.5 t magnetic field . an iron flux - return located outside the coil is instrumented to detect @xmath25 mesons and to identify muons ( klm ) .
the detector is described in detail elsewhere @xcite .
a geant - based detailed simulation of the belle detector is used to produce monte carlo ( mc ) event samples and determine efficiencies .
charged tracks are reconstructed using hit information from the cdc and svd .
kaon and pion mass hypotheses are assigned to the charged tracks using a likelihood ratio @xmath26 ; the likelihood @xmath27 is obtained by combining information from the cdc ( specific ionization ( @xmath28 ) ) , tof , and acc . we require @xmath29 ( @xmath30 ) for kaon ( pion ) candidates @xcite .
the resulting identification efficiency varies from 86@xmath10 to 91@xmath10 ( 94@xmath10 to 98@xmath10 ) for kaons ( pions ) , depending on the momentum .
standard procedures are used to identify leptons .
electrons are identified using a combination of the specific ionization measurement from the cdc , the acc response , and the electromagnetic shower position , shape and energy measurements from the ecl @xcite .
the electron identification probability is required to exceed 0.5 .
muons are identified with klm hit positions and penetration depth @xcite ; the standard muon prerejection cuts are applied and the muon likelihood ratio is required to be larger than 0.8 . with these requirements , the combined reconstruction and identification efficiency is flat in momentum and around @xmath31 for both electrons and muons with momentum larger than 1gev / c .
the efficiency decreases with momentum in the lower momentum region .
lepton tracks are required to satisfy @xmath32 cm and @xmath33 cm , where @xmath34 and @xmath35 are the distances of closest approach to the nominal interaction point in the plane perpendicular to the beam axis ( @xmath36 plane ) and along the beam direction , respectively . to remove leptons produced in @xmath37 decays and photon conversions , we require @xmath38mev / c@xmath39 , @xmath40mev / c@xmath39 and @xmath41mev / c@xmath39 , where @xmath42 is any charged particle and @xmath43(@xmath44 ) is any identified electron ( muon ) reconstructed in the event .
charge - conjugate combinations are also implied .
finally , momentum requirements @xmath45gev / c and @xmath46gev / c in the center - of - mass ( cm ) system are applied , and the polar angle requirements @xmath47 and @xmath48 in the laboratory system are applied to select regions of high reconstruction efficiency .
only the cleanest decay mode @xmath49 is used to measure the @xmath4 meson yield .
the invariant mass of @xmath50 candidates is required to be within of the nominal @xmath51 mass .
the @xmath51 helicity angle distribution is expected to be proportional to ; therefore , the requirement is applied .
the helicity angle @xmath52 is defined as the angle between the directions of the @xmath53 and @xmath4 momenta in the @xmath51 rest frame .
although the @xmath4 mass resolution in the kinematic region studied is only @xmath54mev / c@xmath39 , we select @xmath55 combinations within @xmath56mev / c@xmath39 of the nominal @xmath4 mass , because sufficient sidebands are needed to fit the @xmath4 signal .
the large background under the @xmath4 signal from @xmath57 continuum events ( @xmath58 or @xmath19 ) is significantly suppressed by the @xmath4 selection requirements and , additionally , by the same - sign lepton requirement .
therefore , to avoid uncertainties due to a variation of selection efficiency between the @xmath2(5s ) , @xmath2(4s ) and continuum data samples , no topological cuts are applied in this analysis to suppress continuum .
the @xmath5 mesons are produced at the @xmath2(5s ) through the intermediate @xmath59 , @xmath60 , @xmath61 or @xmath62 channels , with subsequent @xmath63 decay .
these channels are not separated in this analysis ; the signal mc simulation generates 93@xmath10 of events in the @xmath62 channel , and 7% of events in the @xmath60 and @xmath61 channels .
the @xmath64 production cross - section at the @xmath2(5s ) has been measured in ref .
@xcite to be @xmath65 nb , and the fraction of @xmath15 events over all @xmath64 events is taken from the pdg @xcite : @xmath66 .
to obtain the number of @xmath4 mesons produced from @xmath5 decays , we compare the normalized @xmath4 momentum distribution in the @xmath2(5s ) , @xmath2(4s ) and continuum data samples using a procedure similar to that described in ref.@xcite .
the normalized momentum is defined as @xmath67 , where @xmath68 is the @xmath4 momentum in the cm , and @xmath69 is the expected @xmath4 momentum if the @xmath4 were produced in the process @xmath70 at the same cm energy . the comparison of the normalized momentum @xmath71 distributions of the @xmath2(5s ) and continuum data samples ( fig .
1a ) and of the @xmath2(4s ) and continuum data samples ( fig .
1b ) shows excellent agreement in the region @xmath72 , where @xmath64 events can not contribute .
the continuum distributions are normalized using the energy - corrected luminosity ratio . to obtain these distributions ,
the @xmath4 mass spectra are fitted by a gaussian to describe the signal and a linear function to describe the background in each bin of @xmath71 .
the gaussian width is fixed to the value obtained from mc simulation ; the gaussian mean is fixed to that obtained from fitting the @xmath4 signal in the whole @xmath73 range .
the signal normalization and the background parameters are allowed to float .
normalized momenta @xmath71 for the @xmath2(5s ) and continuum data samples ( a ) , and for the @xmath2(4s ) and continuum data samples ( b ) .
the points with error bars are the @xmath2(5s ) and @xmath2(4s ) data , while the histograms show the normalized continuum.,title="fig : " ] normalized momenta @xmath71 for the @xmath2(5s ) and continuum data samples ( a ) , and for the @xmath2(4s ) and continuum data samples ( b ) .
the points with error bars are the @xmath2(5s ) and @xmath2(4s ) data , while the histograms show the normalized continuum.,title="fig : " ] continuum contributions were subtracted from the @xmath74 distributions for the @xmath2(5s ) and @xmath2(4s ) data samples ( fig .
1 ) , and the obtained distributions are used to calculate the number of @xmath4 mesons produced from @xmath5 decays . the @xmath2(4s ) distribution is normalized to the number of the @xmath64 pairs produced at the @xmath2(5s ) that do not hadronize to @xmath5 mesons . to obtain this normalization
, we use the number of @xmath75 pairs in our @xmath2(4s ) sample , @xmath76 .
the number of @xmath64 pairs in the @xmath2(5s ) data sample that do not hadronize to @xmath5 mesons is obtained from the formula : @xmath77 finally , this normalization is calculated , @xmath78 .
we find that about 2/3 of @xmath4 mesons produced at the @xmath2(5s ) originate from @xmath5 decays .
after the @xmath75 contribution is subtracted and all bins are summed over the interval @xmath79 , the number of @xmath4 mesons produced from @xmath5 decays is obtained : @xmath80 .
the number of same - sign @xmath4 and lepton events is obtained at the @xmath2(5s ) , @xmath2(4s ) and continuum data samples as a function of lepton momentum for the whole @xmath4 momentum range @xmath81gev / c . to obtain these distributions , the @xmath4 mass spectra are fitted in each bin of @xmath82 in a manner similar to that described in the previous section .
the following backgrounds must be subtracted from the @xmath2(5s ) data to obtain the same - sign @xmath4 and lepton events from @xmath5 decays : * background due to @xmath2(5s)@xmath83 decays .
this is the dominant background , which is evaluated using the momentum - corrected @xmath2(4s)@xmath84 data .
* background due to continuum under the @xmath2(5s ) resonance .
this background is significantly suppressed by the lepton requirement .
it is evaluated using the continuum data sample . *
tracks misidentified as leptons in @xmath2(5s)@xmath85 events .
this background is important for low lepton momentum . it is evaluated using mc simulation of @xmath2(5s)@xmath85 events .
* residual electrons from photon conversions , muons from charged kaon decays , and leptons from @xmath37 decays in @xmath2(5s)@xmath85 events .
this background is expected to be small and is also evaluated using @xmath2(5s)@xmath85 mc simulation .
the lepton momentum distribution obtained at the @xmath2(4s ) was corrected to take into account the @xmath14 meson momentum difference between the @xmath2(5s ) and @xmath2(4s ) samples ; this resulted in a smearing ( up to @xmath21300 mev / c at high lepton momenta ) with only a small global effect . to subtract the continuum contribution ,
lepton momenta are scaled from continuum to the @xmath2(5s ) energy using the same scale factor as for the @xmath71 distributions .
the raw lepton momentum distributions for the @xmath2(5s ) sample are shown in fig . 2 , together with the backgrounds discussed above .
the background contributions are not large in the momentum region greater than 1.2 gev / c and are subtracted from the raw @xmath2(5s ) lepton momentum distribution .
a small correction is also applied to the electron distribution to take into account the radiative energy losses .
uncertainties in this correction are small and are included in the systematic error . after the background subtraction and the radiative losses correction ,
the final lepton momentum distributions are corrected for efficiencies .
these efficiencies are obtained from mc simulation .
( 5s ) sample ( data points with error bars ) , together with estimates of the contributions from background ( 1 ) ( solid histograms ) , background ( 2 ) ( dotted histograms ) and backgrounds ( 3 ) and ( 4 ) ( dashed histograms).,title="fig:"](5s ) sample ( data points with error bars ) , together with estimates of the contributions from background ( 1 ) ( solid histograms ) , background ( 2 ) ( dotted histograms ) and backgrounds ( 3 ) and ( 4 ) ( dashed histograms).,title="fig : " ]
the final lepton momentum distributions , after the background subtraction and efficiency correction , are fitted with the sum of two contributions : one from primary leptons and one from secondary leptons .
primary leptons , produced directly in semileptonic @xmath5 decays , have a momentum spectrum significantly harder than that of secondary leptons , which are produced in the subsequent decays of @xmath5 daughters such as @xmath4 , @xmath86 and @xmath87 mesons , and @xmath88 leptons .
we expect a @xmath89% contribution from each of the latter three daughters relative to the @xmath4 contribution .
we take the @xmath4 , @xmath86 , @xmath87 and @xmath88 production ratios from our signal mc simulation to obtain the shape of the secondary lepton momentum distribution , and we vary these ratios within reasonable limits . the resulting systematic uncertainty is found to be small , because the lepton spectra resulting from @xmath4 , @xmath86 , @xmath87 and @xmath88 decays are similar . in the fit ,
we use fixed polynomial parametrizations for both the primary and the secondary lepton momentum spectra as determined from our mc simulation ( fig .
3 ) , which contains the most up - to - date knowledge of specific @xmath5 decays based on experimental measurements and theoretical models .
thus these shapes result from a semi - empirical description of @xmath5 decays .
decays in the mc simulation , superimposed with the polynomial parametrizations used in the final fit ( curves).,width=283 ] decays .
the solid curves show the results of the fits , and the dotted curves show the fitted contributions from primary and secondary leptons.,title="fig : " ] decays .
the solid curves show the results of the fits , and the dotted curves show the fitted contributions from primary and secondary leptons.,title="fig : " ] the final leptonic momentum distributions produced in @xmath5 decays ( fig .
4 ) obtained after background subtractions and an efficiency correction ( all corrections described above are also applied ) are used to extract the numbers of primary and secondary leptons .
we fit the data with a function which includes the sum of two terms with fixed shapes and floating normalizations . based on the mc simulation
, we expect that the momentum ranges selected in our analysis will contain 96.7@xmath10 of the primary electrons ( @xmath45gev / c ) and 89.4@xmath10 of the primary muons ( @xmath46gev / c ) .
the only free parameters in the fit are the numbers of primary and secondary leptons .
we fit separately the electron spectrum and the muon spectrum , as shown in .
we also fit the two distributions simultaneously , assuming equal electron and muon production rates in @xmath5 decays , for both primary and secondary leptons .
the results of the three binned @xmath90-fits are given in table 1 .
they represent our measured numbers of a same - sign leptons accompanying @xmath91 meson in events with @xmath5 pairs from the @xmath2(5s ) sample . [
tab : bfr2 ] @l@ @c@ @c@ @c@ subsample & number of & number of & fit , + & primary leptons & secondary leptons & @xmath92 + electron & @xmath93 & @xmath94 & 0.99 + muon & @xmath95 & @xmath96 & 1.42 + combined & @xmath97 & @xmath98 & 1.16 + finally , dividing these numbers by the full number of @xmath4 mesons from @xmath5 decays in our @xmath2(5s ) data sample ( see section 3 ) and applying an additional factor of 2 due to the @xmath5 mixing effect , we obtain the following semileptonic branching fractions : @xmath99 @xmath100 @xmath101 where the last one represents an average over electrons and muons . table 2 lists the systematic uncertainties , which are combined in quadrature to obtain the total systematic uncertainty . uncertainties due to continuum and @xmath75 background subtraction are estimated by varying the measured normalization factors by @xmath102 ; the shapes of the distributions are not varied .
uncertainties due to the shapes of the primary and secondary lepton momentum spectra are evaluated by varying them within the expected uncertainties in the @xmath5 decay branching fractions .
the obtained branching fractions can be compared with the pdg value @xmath103 , which is theoretically expected to be approximately the same , neglecting a small possible lifetime difference and small corrections due to electromagnetic and light quark mass difference effects .
[ tab : bfr2 ] @l@ @c@ source & relative uncertainty + 1 . lepton misidentification & @xmath104 + 2 . signal and
background shapes in @xmath4 mass fit & @xmath105 + 3 .
determination of full @xmath4 meson number & @xmath106 + 4 .
continuum background modeling & @xmath107 + 5 .
@xmath75 background modeling & @xmath107 + 6 .
mc lepton efficiency determination & @xmath106 + 7 .
variation of @xmath86 , @xmath87 and @xmath88 contributions & @xmath104 + 8 .
shapes of primary and secondary lepton momentum spectra & @xmath108 + sum in quadrature & @xmath109 +
the total inclusive semileptonic @xmath6 decay branching fraction is measured for the first time using a 23.6fb@xmath1 data sample collected at the @xmath2(5s ) resonance with the belle detector at the kekb asymmetric energy @xmath3 collider .
the obtained value , @xmath110 , is in a good agreement with the total inclusive semileptonic @xmath12 decay branching fraction .
we thank the kekb group for the excellent operation of the accelerator , the kek cryogenics group for the efficient operation of the solenoid , and the kek computer group and the national institute of informatics for valuable computing and super - sinet network support .
we acknowledge support from the ministry of education , culture , sports , science , and technology of japan and the japan society for the promotion of science ; the australian research council and the australian department of education , science and training ; the national science foundation of china and the knowledge innovation program of the chinese academy of sciences under contract no .
10575109 and ihep - u-503 ; the department of science and technology of india ; the bk21 program of the ministry of education of korea , the chep src program and basic research program ( grant no .
r01 - 2005 - 000 - 10089 - 0 ) of the korea science and engineering foundation , and the pure basic research group program of the korea research foundation ; the polish state committee for scientific research ; the ministry of education and science of the russian federation and the russian federal agency for atomic energy ; the slovenian research agency ; the swiss national science foundation ; the national science council and the ministry of education of taiwan ; and the u.s.department of energy . 99 w .-
( particle data group ) , j. phys .
g * 33 * , 1 ( 2006 ) .
m. artuso _
( cleo collaboration ) , phys .
lett . * 95 * , 261801 ( 2005 ) .
g. bonvicini _
( cleo collaboration ) , phys .
lett . * 96 * , 022002 ( 2006 ) .
a. drutskoy _
et al . _ ( belle collaboration ) , phys .
98 * , 052001 ( 2007 ) .
a. drutskoy _
et al . _ ( belle collaboration ) , phys .
rev . d * 76 * , 012002 ( 2007 ) .
i. bigi , b. blok , m. a. shifman and a. vainshtein , phys .
lett . b * 323 * , 408 ( 1994 ) . a. f. falk , m. b. wise and i. dunietz , phys . rev .
d * 51 * , 1183 ( 1995 ) .
e. bagan , p. ball , v. m. braun and p. gosdzinsky , phys .
lett . b * 342 * , 362 ( 1995 ) .
m. b. voloshin , phys .
rev . d * 51 * , 3948 ( 1995 ) .
m. neubert and c. t. sachrajda , nucl .
phys . b * 483 * , 339 ( 1997 ) .
s. kurokawa and e. kikutani , nucl .
instr . and meth . a * 499 * , 1 ( 2003 ) and other papers included in this volume .
a. abashian _
et al . _ ( belle collaboration ) , nucl .
instr . and meth . a * 479 * , 117 ( 2002 ) .
k. hanagaki _
et al . _ ,
instr . and meth . a * 485 * , 490 ( 2002 ) .
a. abashian _
et al . _ , nucl .
instr . and meth . a * 491 * , 69 ( 2002 ) . | inclusive semileptonic @xmath0 decays are studied for the first time using a 23.6fb@xmath1 data sample collected on the @xmath2(5s ) resonance with the belle detector at the kekb asymmetric energy @xmath3 collider .
these decays are identified by the means of a lepton accompanied by a same - sign @xmath4 meson originating from the other @xmath5 in the event .
the semileptonic branching fractions are measured in the electron and muon channels to be and , respectively .
assuming an equal electron and muon production rate in @xmath5 decays , a combined fit yields an average leptonic branching fraction .
although @xmath5 mesons have been studied by lep and tevatron experiments for a long time , only a few @xmath5 decay branching fractions have been measured @xcite .
in addition to these very high energy experiments , an alternative source of @xmath5 mesons has been explored recently , where @xmath5 mesons are produced by @xmath3 colliders running at the @xmath2(5s ) energy .
first evidence for inclusive and exclusive @xmath5 decays at the @xmath2(5s ) was found by the cleo collaboration @xcite , which collected a data sample of 0.42 fb@xmath1 , and by the belle collaboration @xcite , which used a data sample of 1.86fb@xmath1 . however , the statistical significance of these first measurements was very limited . in 2006
the belle collaboration collected an additional 21.7 fb@xmath1 of data at the @xmath2(5s ) , allowing one to study several @xmath5 decay modes with improved statistical accuracy .
one of the most interesting @xmath5 decay modes , for which the branching fraction can be measured for the first time using the new @xmath2(5s ) data , is the total inclusive semileptonic decay @xmath6 .
charge - conjugate modes are implicitly implied everywhere in this paper .
the @xmath7 collisions at the @xmath2(5s ) resonance are well - suited for the @xmath8 measurement , while it is probably unfeasible at very high energy colliders .
historically , the total inclusive semileptonic @xmath9 decay branching fraction was measured many years ago using data obtained at the @xmath2(4s ) ( see pdg for the full list of the measurements @xcite ) .
the measured values covered the range of 9.7 to 11.0@xmath10 ( the most recent pdg average value is ( 10.78 @xmath11 0.18)% ) and were in a poor agreement with the theoretically predicted branching fraction of about 12% @xcite .
later , explanations for this difference were proposed in several theoretical papers @xcite ; however this discrepancy is not yet completely understood .
measurements of the total semileptonic @xmath12 , @xmath13 and @xmath5 branching fractions , together with corresponding well - measured lifetimes , determine the semileptonic widths .
these widths for the @xmath12 , @xmath13 and @xmath5 mesons are expected to be equal , neglecting small corrections due to electromagnetic and light quark mass difference effects .
if any significant difference between the @xmath12 , @xmath13 and @xmath5 semileptonic widths were observed , it would indicate an unknown source of lepton production in @xmath14 decays .
the correlated production of a @xmath4 meson and a same - sign lepton at the @xmath2(5s ) resonance is used in this analysis to measure @xmath8 .
the method exploits the fact that , at the @xmath2(5s ) , the dominant production of a @xmath4 meson and a same - sign fast lepton comes from the @xmath15 state . neither the @xmath16 continuum nor @xmath17 states ( except for a small contribution due to @xmath18 @xmath12 mixing effect ) can result in a same - sign @xmath19-quark ( i.e. , @xmath4 meson ) and primary lepton final state .
simultaneous production of a same - sign @xmath4 meson and @xmath20 from one @xmath5 decay is totally negligible . |
semiconductor coupled microcavity systems , also called photonic molecules ( pms ) , are attracting considerable attention as efficient test - beds for quantum and nonlinear optics @xcite .
the high optical nonlinearities combined with tight light confinement open up new routes for the study of light - matter interaction in non - equilibrium driven dissipative systems . even at the semiclassical level , already two coupled cavities or
photonic dimers may display a wide range of rich nonlinear dynamical phenomena such as instabilities and bifurcations @xcite . in this context ,
micro and nanolasers prove useful to investigate nonlinear dynamics in the low photon number regime .
either cavity quantum electrodynamics ( cqed ) microlasers or photonic band - gap materials can provide large enough spontaneous emission factors ( @xmath0 ) such that strong nonlinearities may take place in an optical cavity with , ultimately , few photons .
large purcell factors in the former , and the suppression of leaky modes in the latter can be tailored in order to obtain high @xmath0-factors . as a consequence , the saturation photon number , @xmath1 ,
may be significantly lowered @xcite .
micro and nanolasers may display rich nonlinear dynamical behaviors even at a single laser level , such as injection locking phenomena @xcite , self - pulsing @xcite and superradiant giant photon bunching @xcite .
recently , spontaneous symmetry breaking in coupled photonic crystal nanolasers has been demonstrated with @xmath2 intracavity photons ( @xmath3 ) @xcite , paving the way to explore a wealth of nonlinear dynamical phenomena in photonic dimers operating in a laser regime .
these systems can also be useful for applications such as flip - flop optical memories and logic gates in active photonic circuitry @xcite . in this article
we show , both theoretically and experimentally , the asymmetric interaction between strongly coupled modes in a photonic crystal nanolaser dimer
. we will analyze this phenomenon in the framework of `` two discrete sites '' bogatov effect , in the sense that the dynamic population grating takes the form of oscillations of the population imbalance between two optical defects at the rabi frequency ( fig .
[ fig1]a ) .
mode competition is at the heart of multimode laser dynamics , and may occur as long as different modes share the same gain medium .
for instance , mode switching and bistability have been observed in standard microcavity systems , such as vertical cavity semiconductor lasers ( vcsels ) @xcite and micropillar lasers @xcite , and they have also been predicted in photonic crystal coupled cavities @xcite .
these mechanisms usually rely on cross - gain saturation effects .
however , mode interaction in semiconductor cavities may also result from stimulated scattering due to a dynamic carrier population grating in the gain medium , which oscillates at the beat note frequency between adjacent cavity modes ( @xmath4 ghz ) .
such interaction , known as the `` bogatov effect '' , is asymmetrical in the sense that the laser mode at the blue side of the spectrum transfers its energy to the mode at the red side @xcite .
this phenomenon is a consequence of nonlinear dispersion , and it is also known to occur in free - running edge - emitters @xcite , in vcsels supporting two orthogonal polarizations @xcite , and in vertical external cavity lasers ( vecsels ) leading to a slow light regime @xcite .
yet , the observation of this effect in photonic dimers remains elusive to date . for moderate @xmath0-factors ( @xmath5 ) as in our case , an accurate description of the system can still be obtained in terms of mean - field equations accounting for the dynamics of two coupled complex field amplitudes in the left ( l ) and right ( r ) nanocavities filled with a quantum well gain medium @xcite(see fig .
[ fig1]b ) .
the evolution of the field and carrier densities are governed by @xmath6 where @xmath7 and @xmath8 are normalized as the photon and carrier numbers in the cavities , respectively , @xmath9 is the cavity loss rate , @xmath10 is the two - level radiative recombination rate , @xmath11 the henry factor , @xmath12 the carrier number at transparency , @xmath13 the pump rate and @xmath14 is the total carrier recombination rate . the complex inter - cavity coupling constant quantifies frequency ( @xmath15 ) and loss ( @xmath16 ) splitting as a result of the evanescent coupling , and @xmath17 are langevin noise terms accounting for spontaneous emission with rate @xmath18 where @xmath19 is the bimolecular radiative recombination rate and @xmath20 the volume of the active medium .
the expression of @xmath21 is standard ; it is chosen as to preserve the proper correlations between the photon and carrier numbers , see @xcite supplementary material for more details . in order to address mode interaction , we first project eqs .
[ eq : alr ] and [ eq : nlr ] on the basis of eigenmodes , @xmath22 and @xmath23 , corresponding to bonding ( b ) and anti - bonding ( ab ) modes of the photonic dimer respectively ( fig .
[ fig1]a ) .
dynamic grating bogatov effects can be described by considering a superposition of monochromatic fields , @xmath24 , which induces oscillation of the carrier number at the rabi frequency @xmath25 : @xmath26 where @xmath27 is the mean carrier number , @xmath28 is the population imbalance between the cavities , @xmath29 and @xmath30 are the steady state values , and @xmath31 and @xmath32 are small amplitude complex oscillations . replacing these expressions into eqs .
[ eq : alr ] and [ eq : nlr ] , and neglecting spontaneous emission for the moment , we obtain : @xmath33 equation [ grating ] stands for the discrete grating complex amplitude , and it is at the origin of asymmetric interaction between the modes ( fig . [
the stimulated gain originating from eq .
[ grating ] for the bonding ( @xmath34 ) and the anti - bonding ( @xmath35 ) modes can be approximated , in the limit of large rabi splitting , by @xmath36 equations [ stimgain ] contains three important features : i ) the mode interaction is anti - symmetric , meaning that the sign is reversed when changing the direction of the interaction ( scattering from @xmath37 to @xmath38 or viceversa ) ; ii ) for @xmath39 , i.e. blue - detuned bonding mode , the positive gain contribution is experienced by the red - detuned ab mode ; and iii ) this effect is proportional to the @xmath11-factor , thus revealing the nonlinear index effect as the main mechanism for the stimulated scattering .
note that saturation term in eq .
[ grating ] is negligible provided that @xmath40 , where @xmath41 is total photon number , and @xmath42 . in our case
this approximation is justified since @xmath43 and @xmath41 is few times @xmath44 in a normal laser operation regime .
in order to experimentally realize the discrete bogatov effect , we have designed and fabricated two evanescently - coupled active photonic crystal ( phc ) l3 cavities ( three holes missing in the @xmath45k direction of a triangular lattice ) in a semiconductor free standing membrane ( fig . [ fig1]b ) .
the size of surrounding holes has been modified to both increase beaming efficiency ( band - folding technique ) and control the inter - cavity coupling strength ( barrier engineering ) .
details on the phc structure can be found in refs .
@xcite and @xcite .
both single and coupled cavities have been etched in inp membranes containing ingaas / ingaasp quantum wells .
the measured q - factor at transparency is @xmath46 ( @xmath47 , where @xmath48 is the photon lifetime in the cavity ) and the spontaneous emission factor ( @xmath0 ) is @xmath49 0.02 for coupled cavities . the linewidth enhancement factor is @xmath50 .
two modes are observed in the coupled - cavity system : the b mode for a symmetric superposition of the single cavity modes , and the ab mode for an anti - symmetric superposition , which can be clearly identified in the far field images ( fig .
[ fig1]c ) . 8@xmath9 ( a ) and k @xmath49 12@xmath9 ( b ) .
c , d ) superposition of 10 numerical ll - curves for k = 8@xmath9 ( c ) and k = 12@xmath9 ( d ) .
e ) black line : effective loss split parameter @xmath51 .
markers : pump power at the switching point normalized by pump power at threshold plotted versus normalized mode splitting ( @xmath52/@xmath9 , bottom axis ) and rabi frequency ( k/@xmath9 , top axis ) .
the corresponding wavelength splittings from the leftmost to the rightmost points are : 1.92 , 2.66 , 3.25 , 4.06 and 4.8 nm .
blue line : fit of the experimental data using eq .
[ ps ] and @xmath51 ( see text ) .
the star indicates the operation point of the spontaneous symmetry breaking ( ssb ) transition reported in @xcite .
the red numbers represent the modification of the holes in the central row of the phc structure in order to tune the amount of evanescent coupling between the nanocavities .
the inset shows a structure where central holes diameters were decreased by 15% . ]
the effective refractive index of the symmetric mode of the underlying w1 waveguide ( one row of holes omitted in the @xmath16k direction of the triangular lattice ) is smaller than the index of the anti - symmetric mode .
hence the b mode is blue - detuned with respect to the ab mode @xcite .
our strategy to observe bogatov mode interaction is to induce a mode switching by stabilizing the system close to the laser threshold at the blue most mode ( b ) . for this ,
the phc lattice parameter is tuned such that the optical resonances lie on the red slope of the gain curve ( fig .
[ fig1]c ) ( maximum at @xmath53 @xmath49 1510 nm and fwhm @xmath54 nm ) . as a result
, the b mode has larger enough gain and , even though the ab mode has lower optical losses @xcite , the b mode experiences larger net gain .
therefore , the blue - detuned mode ( b ) is the lasing one close to laser threshold .
the coupled nanolasers are optically pumped at room temperature .
we use optical periodic ( 50 khz - repetition rate ) and incoherent ( @xmath55 nm ) pumping consisting either of a ramp for the acquisition of the output vs. input power ( ll - curve ) or short pulses ( @xmath56 ps - pulse duration ) to perform statistics on the output pulse energies . in both cases , the pump spot is located at the center of the coupled cavity system .
the emission is collected with a n.a.=0.95 microscope objective , and its back focal plane is imaged through a lens to obtain the far - field pattern ( fig .
[ fig1]c ) .
two single mode fibers coupled to microscope objectives are used to spatially select two regions of the far field : the center corresponds to b - mode intensity , and one of the lateral lobes to the ab - mode .
these optical signals are sent to two identical low noise ( 200 fw@xmath57 ) , 660 mhz - bandwidth avalanche photodiode ( apd ) detectors .
this is similar to the experimental technique of ref .
@xcite , but now the two detectors monitor intensity fluctuations from modes rather than individual cavities .
the optical rabi frequency is related to the laser mode splitting @xmath52 as @xmath58 @xcite . in fig .
[ switch - quasicw]a , b we show the simultaneous output intensities of both modes as a function of the instantaneous pump power for two different rabi mode splittings ( @xmath59 and @xmath60 ) .
this plot consists of a superposition of 40 sequences of quasi - cw pumping .
it can be clearly observed that , at threshold , lasing takes place on the b - mode , demonstrating that it has higher net gain as expected . in fig .
[ switch - quasicw]a the b mode loses stability and turns off at p @xmath61 2.5p@xmath62 , while the ab mode turns on right after .
this reveals energy transfer from the blue - detuned to the red - detuned mode .
in addition we observe that , from sequence to sequence , the instability threshold fluctuates in a range 2.1 @xmath63 p / p@xmath62 @xmath63 2.9 .
once the bifurcation point is crossed , the system remains stable on the ab - mode up to large values of the pumping .
a similar behavior is observed for larger coupling strength ( fig .
[ switch - quasicw]b ) , and in this case the switching point occurs at a larger pumping level , p @xmath49 3.8p@xmath62 .
note that this is consistent with the dependence of the stimulated gain on the cavity coupling of eq .
[ stimgain ] , i.e. @xmath64 .
next , we explore the full range of experimentally available coupling levels by extensively exploiting the barrier engineering technique . to this aim
we have fabricated a number of different coupled cavity systems with varying evanescent coupling .
laser emission experiments on these cavities have been carried out subsequently , and the results are depicted in fig .
[ switch - quasicw]e . here
, the pump power at the switching point is plotted as a function of the normalized frequency splitting of the laser modes . ,
dashed line : k = 12@xmath9 .
b ) second order cross - correlation function as a function of normalized pump power for k = 8@xmath9 ( solid line ) and k = 12@xmath9 ( dashed line ) . ] our last experimental study deals with the statistics of the laser emission .
since a few hundreds of photons fill the cavities ( in our case @xmath65 ) , stochastic fluctuations play a fundamental role , which is typical in nanolasers .
we have investigated the consequences of this in the context of mode switching by performing a statistical study of the mode fluctuations . with this aim
we have implemented a short - pulse pumping scheme developed in @xcite which allows us to obtain second order auto and cross - correlation functions of the emitted photon number @xmath66 : @xmath67 where @xmath68 is the standard deviation .
note that eq .
[ g2 ] holds in the large photon number limit @xcite .
figure [ g2_exp ] depicts self and cross second order correlations functions obtained from time traces containing @xmath69 pulses , with @xmath70 ( solid line ) and @xmath71 ( dashed line ) . before the switch
, lasing occurs in the b - mode and thus @xmath72 1 while the ab mode develops super - poissonian fluctuations .
we point out that in this close - to - threshold pump region highly influenced by noise , the actual @xmath73 might be larger compared to the observed one , which is limited by the resolution of our experimental technique .
a decrease of @xmath74 is visible , which is the signature of a growing anti - correlation of the mode fluctuations .
the mode switching occurs as @xmath75 which takes place at @xmath76 with a value of @xmath77 for @xmath70 , and @xmath78 with a value of @xmath79 for @xmath71 .
note that at the switching points the modes become indistinguishable in terms of the second moments of their energy fluctuations .
furthermore , for @xmath71 , a `` soft '' region for 3 @xmath63 p / p@xmath62 @xmath63 4.3 is characterized by a plateau in @xmath80 , which corresponds to both modes present in the optical spectrum .
this region also exists when the coupling is smaller , but the plateau has a larger cross - correlation , @xmath81
. this could be due to a stronger effect of spontaneous emission noise for small k as a result of a closer proximity of the switching point to the laser threshold .
the experimental results can be compared with the full solutions of the coupled field equations in presence of spontaneous emission , that we modeled as langevin forces .
figures [ switch - quasicw]c , d ) show simulated time traces for two coupling parameters @xmath70 and @xmath71 .
note that the energy transfer in the form of a switch mechanism is well reproduced by the model .
the jitter of the mode switching , in turn , is larger in the experiment , which could be explained as a consequence of other sources of noise ( e.g. mechanical vibrations and thermal effects ) .
we now theoretically investigate the nonlinear dynamical origin of the mode switching and explain the dependence of the switch - power upon the coupling parameter observed in fig .
[ switch - quasicw]e .
to this aim , we have performed an analytical reduction of the deterministic part of eqs .
[ eq : alr ] and [ eq : nlr ] ( @xmath82 ) to a one - dimensional ( 1d ) dynamical system able to capture the essence of the mode switching dynamics . by exploiting the specific scaling of the parameters and of the governing time scales ,
it is possible to reduce the time evolution of the system to the dynamics of the fractional photon population imbalance between the two modes , @xmath83 . in this framework ,
the b and ab solutions are represented by the two extreme values @xmath84 . under the assumption of large rabi splitting and small population imbalance it is possible to show that the dynamics of @xmath85 is approximately ruled by the following 1d differential equation : @xmath86 where @xmath87 is an effective 1d potential .
the factor @xmath88 is a 1d nonlinear optical loss splitting rate which reads : @xmath89 corresponding to the three values of the normalized pump power @xmath90 are @xmath91 and @xmath92 .
( c ) bifurcation diagrams for the photon number in a single cavity in the b and ab modes as a function of pump power ( @xmath60 ) .
a switching occurs from the b to the ab mode through a hopf bifurcation at @xmath93 .
( d ) blow - up around the bifurcation points and the narrow periodic bridge solution .
full and dotted lines correspond to stable and unstable solutions , respectively . ]
equations [ eq : zztop ] and [ eq : gamz ] contain the essence of the mode switching phenomenon induced by the bogatov effect in the coupled cavity system .
one of the fixed points @xmath84 is stable ( node ) and the other one unstable ( saddle ) , and depending on the sign of @xmath94 the stable one will be bonding ( @xmath95 ) or anti - bonding ( @xmath96 ) .
this is illustrated in fig .
[ theo]a , where the potential @xmath97 is represented for three different pumping parameters : before ( @xmath98 ) and after ( @xmath99 ) the switching point , and for an intermediate pump level ( @xmath100 ) . the mode switching depicted in fig .
[ theo]b is thus represented by the change of sign of @xmath94 .
setting @xmath101 in eq .
[ eq : gamz ] allows us to find the locus of the mode switching pump power @xmath102 as a function of the imaginary ( @xmath15 ) and real ( @xmath16 ) parts of the coupling parameter : @xmath103 when using eq .
[ ps ] to fit the experimental results , one has to take into account the gain difference between b and ab modes , since they are spectrally located on the slope of the gain function ( see fig . [ fig1]c ) . a natural way to include this effect in the model
is to replace @xmath16 by an effective @xmath51 parameter which accounts for net optical losses ( loss split minus gain difference ) .
hence , the effective loss split parameter depends on mode spacing , @xmath104 .
we consider the following phenomenological saturable function @xmath105 $ ] , where @xmath106 is the `` zero - detuning '' loss split , @xmath107 is the large @xmath52 limit , and @xmath108 is a saturation parameter .
we fix @xmath109 for @xmath110 ( or @xmath111 ) , for consistency with the operation point of the symmetry breaking ( ssb ) transition , whose parameters have been determined in @xcite .
two effects are captured by @xmath51(@xmath52 ) : a first - order linear increase for @xmath52 @xmath61 @xmath112 which comes from the non - zero slope of the gain curve ( fig .
[ fig1]c ) , and higher - order saturation terms .
the latter can be attributed to saturation effects not taken into account in the original model , such as spectral hole burning @xcite .
our expression of @xmath104 is inserted into eq .
[ ps ] to fit the experimental data , and the result is plotted in fig .
[ switch - quasicw]e .
it is interesting to note that two regimes can be identified : the rabi regime , where the nonlinear laser frequency shift is smaller than the rabi splitting , i.e. @xmath113 , and the josephson regime for weaker rabi splitting , i.e. @xmath114 . the ssb transition lies within the josephson region ( see fig . [
switch - quasicw]e ) , where ultra fast ( @xmath2 ghz ) oscillations have been predicted to emerge through secondary ( hopf ) bifurcations @xcite . in this regime
, @xmath51 is rapidly increasing and bogatov mode interaction can be observed as well ( see point at @xmath115 ) .
indeed , for @xmath115 , also ssb can be observed after the mode switching instability .
for increasing @xmath52 values we enter into the rabi regime , where @xmath51 strongly saturates to @xmath116 . in this regime of nearly constant @xmath51 ,
the linear dependence of the switch power upon the coupling parameter k predicted in eq .
[ ps ] is established . within this simple description ,
the switching point takes place at a single point @xmath102 since the switching mechanism is represented by an exchange of stability between the two solutions .
first , the presence of spontaneous emission modifies this picture and induces a `` transition region '' in the pump parameter with a nonzero measure .
second , the single bifurcation point is a consequence of the large @xmath15 approximation used to derive eq .
[ eq : zztop ] . relaxing this hypothesis ,
i.e. for finite k - values , the switching mechanism becomes more complex : a non - trivial connection of b and ab solutions in phase space comes up as the pump parameter is varied
. this question can be elucidated by computing a bifurcation diagram for the full deterministic dynamical system of eqs .
[ eq : alr ] and [ eq : nlr ] , shown in fig .
it can be observed that the stable b - mode undergoes a supercritical hopf bifurcation at p @xmath49 3.765 p@xmath62 .
in other words , there is a switching zone ( see the inset of fig . [
theo]c ) where the only stable solution is a limit cycle , which manifests itself as a cavity beating at the rabi frequency .
this limit cycle , in presence of noise , is the nonlinear dynamical representation of the soft transition region from the blue - detuned to the red - detuned mode where the coupled nanolasers operate in a dual - frequency regime . such a fast nonlinear beating dynamically delocalizes photons in the coupled cavity system . as a result
strong anti - correlated intensity fluctuations for b and ab modes are expected , which is consistent with the observed minimum of the cross - correlation function at the switching point ( fig .
[ g2_exp ] ) .
hence , we conclude that the observation of a dip in the mode cross - correlation is the statistical consequence of a dual frequency operation region given by an ultrafast ( @xmath117 thz ) noisy limit cycle rather than , for instance , bi - stable mode switching scenarios previously reported for bi - modal microlasers @xcite .
in conclusion , we have revealed for the first time asymmetric energy transfer between eigenmodes in a semiconductor photonic dimer .
we have experimentally shown that a mode switching occurs from the blue - detuned ( bonding ) to the red - detuned ( anti - bonding ) modes as the pump power is increased .
we have identified the basic mechanism underlying such mode interaction as a `` discrete bogatov effect '' , i.e. the asymmetric stimulated light scattering induced by two - site population oscillations at the optical rabi frequency .
the predicted scaling of the onset of switching with respect to the coupling parameter in a mean - field model is shown to be in good agreement with the experimental results for different coupling strengths .
we claim that this phenomenon is generic in semiconductor coupled cavities , and dominates against cross - gain saturation effects , therefore it could be used to model nonlinear mode coupling in probabilistic theories such as bimodal birth - death models in micro / nano - laser photonic dimers .
furthermore , it should also be scalable to n - coupled cavities , such that energy can be expected to flow from the blue - most to the red - most hybrid modes of the photonic molecule with the increase of pump power .
the authors acknowledge i. sagnes and g. beaudoin for the fabrication of the samples .
this work is supported by a public grant overseen by the french national research agency ( anr ) as part of the `` investissements davenir '' program ( labex nanosaclay , reference : anr-10-labx-0035 ) and funding from the anr project optiroc . j.j .
acknowledges financial support from the ramn y cajal fellowship and project combina ( tec2015 - 65212-c3 - 3-p ) .
m. abbarchi , a. amo , v. g. sala , d. d. solnyshkov , h. flayac , l. ferrier , i. sagnes , e. galopin , a. lemaitre , g. malpuech , and j. bloch , `` macroscopic quantum self - trapping and josephson oscillations of exciton polaritons , '' nat phys * 9 * , 275279 ( 2013 ) .
p. hamel , s. haddadi , f. raineri , p. monnier , g. beaudoin , i. sagnes , a. levenson , and a. yacomotti , `` spontaneous mirror - symmetry breaking in coupled photonic - crystal nanolasers , '' nature photonics * 9 * , 311315 ( 2015 ) . s. r. k. rodriguez , a. amo , i. sagnes , l. le gratiet , e. galopin , a. lemaitre , and j. bloch , `` interaction - induced hopping phase in driven - dissipative coupled photonic microcavities , '' nat commun * 7 * ( 2016 ) . e. schlottmann , s. holzinger , b. lingnau , k. ldge , c. schneider , m. kamp , s. hfling , j. wolters , and s. reitzenstein , `` injection locking of quantum dot microlasers operating in the few photon regime , '' arxiv preprint arxiv:1604.02817 ( 2016 ) .
f. jahnke , c. gies , m. aszmann , m. bayer , h. a. m. leymann , a. foerster , j. wiersig , c. schneider , m. kamp , and s. hofling , `` giant photon bunching , superradiant pulse emission and excitation trapping in quantum - dot nanolasers , '' nat commun * 7 * ( 2016 ) .
l. liu , r. kumar , k. huybrechts , t. spuesens , g. roelkens , e .-
geluk , t. de vries , p. regreny , d. v. thourhout , r. baets , and g. morthier , `` an ultra - small , low - power , all - optical flip - flop memory on a silicon chip , '' nature photonics * 4 * , 182187 ( 2010 ) .
h. a. m. leymann , c. hopfmann , f. albert , a. foerster , m. khanbekyan , c. schneider , s. hofling , a. forchel , m. kamp , j. wiersig , and s. reitzenstein , `` intensity fluctuations in bimodal micropillar lasers enhanced by quantum - dot gain competition , '' physical review a * 87 * , 053819 ( 2013 ) . c. redlich , b. lingnau , s. holzinger , e. schlottmann , s. kreinberg , c. schneider , m. kamp , s. hfling , j. wolters , s. reitzenstein _ et al . _ ,
`` mode - switching induced super - thermal bunching in quantum - dot microlasers , '' new journal of physics * 18 * , 063011 ( 2016 ) .
a. m. yacomotti , l. furfaro , x. hachair , f. pedaci , m. giudici , j. tredicce , j. javaloyes , s. balle , e. a. viktorov , and p. mandel , `` dynamics of multimode semiconductor lasers , '' phys .
a * 69 * , 053816 ( 2004 )
. k. d. choquette , r. p. schneider , jr .
, k. l. lear , and r. e. leigenbuth , `` gain - dependent polarization properties of vertical - cavity lasers , '' ieee journal of selected topics in quantum electronics * 1 * , 661 ( 1995 ) .
a. el amili , b .- x .
miranda , f. goldfarb , g. baili , g. beaudoin , i. sagnes , f. bretenaker , and m. alouini , `` observation of slow light in the noise spectrum of a vertical external cavity surface - emitting laser , '' phys .
lett . * 105 * , 223902 ( 2010 ) .
s. haddadi , p. hamel , g. beaudoin , i. sagnes , c. sauvan , p. lalanne , j. a. levenson , and a. m. yacomotti , `` photonic molecules : tailoring the coupling strength and sign , '' optics express * 22 * , 12359 ( 2014 ) . | two coupled - cavity systems , or `` photonic dimers '' , are efficient test - beds for both fundamental optics the realization of quantum correlated states , josephson physics , and so forth , and applications such as optical flip - flop memories . in this work we report on the first observation of nonlinear mode interaction in a photonic dimer formed by two semiconductor photonic crystal coupled nanolasers . for this , we investigate energy transfer between hybrid modes , which manifests as a switching from the blue - detuned ( bonding ) to the red - detuned ( anti - bonding ) modes .
an mean - field model allows us to explain this phenomenon as stimulated scattering due to carrier population oscillations in the cavities at the rabi frequency .
such asymmetrical mode interaction is universal in semiconductor laser photonic molecules , and unveils the origin of cross - correlation dips in the statistics of mode fluctuations . |
multiparticle hadroproduction has been extremely useful along decades to understand the strong interaction dynamics @xcite . in ref.@xcite
we explored the possibility of applying well - known techniques based on inclusive correlations and moments of multiplicity distributions to the quest for a hidden sector ( hs ) at tevatron / lhc experiments . in all fairness ,
multiparticle production is still not fully understood , while our proposal is based on possible departures of @xmath2anomalous from @xmath2standard events .
nonetheless , one might expect that , once stringent selection criteria are applied to events , distinctive features associated to a hs would become useful in the search strategy and subsequent interpretation of new phenomena in multiparticle dynamics . in this work we focus on unparticle physics @xcite in particular , though the main ideas can be generally applied to hidden valley models @xcite . in these models ,
the standard model ( sm ) is accompanied by a hs of new particles not been yet observed due to typically an energetic barrier or a weak coupling to sm particles .
we will assume further that particles coming from a hs can decay back to sm particles @xcite , thereby modifying the conventional pattern of the parton cascade in multiparticle production .
so far most signatures of new physics ( like jets , missing energy , high-@xmath1 leptons or photons , displaced vertices ) have been considered on the transverse plane with respect to the beam direction . in a complementary way
, we advocate that a new stage of matter might also show up in _ soft _ physics of underlying events _ tagged by hard products _ , e.g. through particle ( pseudo)-rapidity correlations ( either integrated or not ) .
collisions , subsequently decaying into final - state sm particles through cluster formation .
the number of unparticles sources at the onset of the cascade may fluctuate , and a large ( and @xmath2continuously variable )
mass of the unparticle stuff would induce additional long - range correlations among the final - state sm particles . ]
experimental results in multibody production along decades have steadily supported the tendency of produced particles to merge into correlated groups @xcite .
this experimental evidence has traditionally led to the view of a two - step process for high - energy hadron collisions . the resulting multiplicity distribution is thus given by the convolution for particle emission sources ( strings , clusters / clans , fireballs ... ) with the decay / fragmenting distribution of sources .
however , observable consequences could derive from a new stage of matter ( stemming from a hs ) generated at the onset of the parton shower ( see fig.1 ) leading to a 3-step scenario in hadronic collisions at very high energy .
the inclusive 2-particle ( rapidity ) correlation function is defined as @xmath3 where @xmath4 denotes the inelastic @xmath0 cross section and subscripts 1 and 2 refer to the two considered particles event by event .
@xmath5 is usually split in two terms : @xmath6 where the short - range ( sr ) part is generally assumed to be more sensitive to dynamical correlations , while @xmath7 stands for long - range ( lr ) correlations usually due to the mixing of different topologies . in this work
we focus on the central rapidity region . according to our study @xcite , the @xmath7 piece in eq.(2 )
should become enhanced in a 3-step scenario wrt a standard 2-step cascade as a consequence of larger fluctuations of the number ( and mass ) of the primary sources of partons .
in addition , the correlation length in the @xmath8 term should become larger because of the ( presumably ) higher mass of the intermediate hidden ( un)particle stuff . therefore longer and
stronger correlations between particles can be expected from both sr and lr terms in eq.(2 ) whenever a hs appears in the cascade .
nevertheless , other conventional sources of possible lr effects in multiparticle production @xcite should be properly taken into account .
-scaling for @xmath9 ( upwards ) , in 2-step ( dotted lines ) and 3-step ( solid lines ) cascades ; middle : @xmath10 versus @xmath11 for 2-step and 3-step cascades ; right : @xmath12 versus @xmath11 setting @xmath13 , @xmath14 ( @xmath15 , @xmath16 ) in a 2-step ( 3-step ) cascade , respectively .
see @xcite for more details.,title="fig : " ] -scaling for @xmath9 ( upwards ) , in 2-step ( dotted lines ) and 3-step ( solid lines ) cascades ; middle : @xmath10 versus @xmath11 for 2-step and 3-step cascades ; right : @xmath12 versus @xmath11 setting @xmath13 , @xmath14 ( @xmath15 , @xmath16 ) in a 2-step ( 3-step ) cascade , respectively .
see @xcite for more details.,title="fig : " ] -scaling for @xmath9 ( upwards ) , in 2-step ( dotted lines ) and 3-step ( solid lines ) cascades ; middle : @xmath10 versus @xmath11 for 2-step and 3-step cascades ; right : @xmath12 versus @xmath11 setting @xmath13 , @xmath14 ( @xmath15 , @xmath16 ) in a 2-step ( 3-step ) cascade , respectively .
see @xcite for more details.,title="fig : " ] fluctuations in small phase space regions ( intermittency ) have been commonly described by the scaled moments @xmath17 of a multiplicity distribution @xmath18 ( corresponding to the normalized phase - space integral over the @xmath11-particle density function ) as the rapidity interval under study @xmath19 is split into @xmath20 bins of equal size @xmath21 @xcite . on the other hand , the fractality nature of multiparticle hadroproduction
is deeply connected with intermittent behaviour exhibiting a power - law dependence of the multiplicity moments with the cell size @xcite .
if self - similar dynamical fluctuations exist , @xmath17 should obey a power - law increase ( @xmath20-scaling ) at small @xmath22 , i.e. @xmath23 where the @xmath2intermittency exponents
@xmath24 ( @xmath25 ) are related to the anomalous fractal dimensions @xmath10 @xcite as @xmath26 moreover , the power scaling between @xmath27 and @xmath28 ( @xmath29-scaling ) , @xmath30^{\beta_q}\ \
\ln{f_q}(m)=\beta_q\ln{f_2}(m)\ ] ] represents another interesting relation , where the @xmath12 coefficient can be expressed as @xmath31 monofractality ( @xmath32 , @xmath33 ) implies that @xmath34 , which can be tested from ( 5 ) .
brax and peschanski @xcite proposed a better approximation than eq.([eq : beta ] ) using a lvy stable law description of multiparticle production : @xmath35 where @xmath36 is the lvy index ( also known as the degree of multifractality ) which permits an estimation of the cascading rate ; @xmath36 should be in principle restricted to the interval @xmath37 ( region of stability ) with @xmath38 characterizing monofractal behaviour .
instead of the law expressed in eq.([eq : levy ] ) , @xmath12 has been parametrized in the ginzburg - landau model of phase transitions as @xmath39 where @xmath40 now implies monofractality .
the motivation for all these approaches in heavy - ion collisions comes from the need of a signal for quark - gluon plasma formation through , e.g. , a ( second - order ) phase - transition , since the correlation length would then diverge , and the system behave as a simple fractal @xcite .
we are certainly not considering here such kind of phase - transition , but extra lr correlations from a hs leading to a similar effect on both @xmath36 and @xmath41 estimators . indeed our results shown in fig.2 are consistent with the interpretation that multiparticle production including a hs approaches monofractality ; the intermittency exponents , all @xmath10 values and @xmath36 and @xmath41 estimators turn out to be smaller in a 3-step scenario than in a 2-step cascade .
let us finally note that the ( shannon ) entropy , defined for a multiplicity distribution @xmath18 as @xmath42 @xcite , might become another indicator of a hs in multiparticle production . for example , it is relevant to verify to what extent entropy is additive @xcite by checking whether @xmath43 using particles belonging to regions @xmath44 and @xmath45 well separated in rapidity space ( e.g. forward and backward hemispheres ) .
intermittency and ( multi)fractality can be sensitive to a hs altering the pattern of the parton cascade in high - energy inelastic @xmath0 collisions . from our study in @xcite
we conclude that longer and stronger correlations between emitted particles , and larger scaled moments in multiplicity distributions , should be expected than in a conventional qcd cascade , while the pattern of the multiparticle system approaches monofractality - or merely becomes less fractal .
the challenge remains however in tagging those _ anomalous _ underlying events ( where such new phenomena would show up through soft physics ) by appropriate selection cuts , to be compared with a control sample of standard events .
more work requiring prior tuning and monte carlo generation of events at lhc energies is required before drawing any conclusion about the feasibility of our proposal . | the study of multiparticle dynamics in hadron - hadron collisions at tevatron and lhc could provide useful information on new physics in addition to the expected signatures on the transverse plane .
we suggest that an analysis of inclusive correlations between emitted particles in @xmath0 inelastic collisions , and factorial moments of multiplicity distributions , may be helpful in uncovering ( un)particles from hidden sectors , using underlying events tagged by hard products like high-@xmath1 leptons and photons , and applying stringent selection criteria like event shape variables , etc .
address = departamento de fsica terica and ific , university of valencia - csic + dr .
moliner , 50 46100 burjassot , spain |
noncommutative geometry @xcite is a vast generalization of differential geometry to the quantum world . however , many of its successful applications are in ordinary , commutative differential geometry . for instance
, it turned out that kasparov s bivariant @xmath2-theory provided the right context for index theory which naturally extends to foliations @xcite . here
, a central role is played by the shriek or wrong - way map @xmath3 associated to any ( @xmath2-oriented ) smooth map @xmath4 between smooth manifolds .
it is an element in the bivariant @xmath2-theory @xmath5 of kasparov @xcite and is defined using the principal symbol of a suitably defined pseudodifferential operator of order @xmath6 .
we refer to @xcite ( and @xcite for the general case of maps between foliations ) for full details . for the special case that @xmath7 is a point
the shriek map is the fundamental class @xmath8 $ ] of the manifold in the @xmath2-homology group @xmath9 .
the wrong - way functoriality of the shriek map was stated as a problem in @xcite and proven in @xcite .
it says that for two maps @xmath10 and @xmath11 we have @xmath12 where @xmath13 denotes the internal kasparov product in @xmath1-theory .
now , for the special case that @xmath14 consist of a single point , this implies that the fundamental class of @xmath15 factorizes in @xmath1-theory as follows : @xmath16 = f !
\hot_{c(y ) } [ y].\ ] ] moreover , if @xmath4 is a submersion , the shriek map @xmath17 can equivalently be described by a family of pseudodifferential operators of order zero acting on the fibers of @xmath18 ( * ? ? ?
* proposition 2.9 ) .
a natural first question that arises is whether ( and under which conditions ) one can explicitly find unbounded @xmath1-cycles ( as introduced in the early 1980s by connes and baaj julg in @xcite ) that represent the respective classes @xmath8 , [ y]$ ] and @xmath3 .
if @xmath15 and @xmath7 are spin@xmath0 manifolds , it is clear @xcite that the dirac operators @xmath19 and @xmath20 represent the corresponding fundamental classes .
moreover , if @xmath21 is a riemannian submersion , one expects that there is a family of dirac operators @xmath22 on the fibers @xmath23 ( for all @xmath24 ) that gives an unbounded representative of the class @xmath3 in @xmath5 .
the immediate next question is then whether we can write the factorization formula as a tensor sum of these unbounded @xmath1-cycles , that is to say , whether in some sense @xmath25 with @xmath26 a suitable connection on the bundle of vertical spinors .
even though the above are natural questions , and their affirmative answers appear to be folklore to practitioners of ( unbounded ) @xmath1-theory , we have not found a detailed written account on it in the literature .
of course , the bounded @xmath1-cycles that enter @xcite are very much differential in nature and essentially already dictate the form of the corresponding unbounded @xmath1-cycles .
the work of bismut @xcite ( see also ( * ? ? ?
* chapter 10 ) and ( * ? ? ? * chapter 4 ) ) comes very close , at least in spirit , but does not connect to @xmath1-theory ( even though that would provide the right context for the atiyah singer index theorem for families ) .
also , the mere existence of the unbounded @xmath1-cycles and the validity of the tensor sum as an unbounded representative of the internal kasparov product would follow from @xcite . there , the authors work with the unbounded form of the internal @xmath1-product that was the topic of @xcite and use a theorem by kucerovsky @xcite to check if an unbounded @xmath1-cycle is a representative of the internal product of ( the bounded classes of ) two other unbounded @xmath1-cycles . without questioning the power of their approach , the geometric context sketched above makes it most natural to explicitly construct the unbounded @xmath1-cycles in terms of ( families of ) dirac operators and establish for them an explicit tensor sum decomposition as in equation [ eq : tensor - sum ] .
moreover , as we will see below ( theorem [ thm : fact ] ) there is an obstruction to having an exact factorization of unbounded @xmath1-cycles given by a ( bounded ) curvature form .
in fact , this is the key point of working with unbounded @xmath1-cycles instead of merely their homotopy classes in bounded @xmath1-theory : one captures the metric aspect of the geometry which is encoded by an unbounded operator . from this point of view
it is thus not surprising that curvature enters in the ( unbounded product ) formula for submersions . in this paper
, we will present in full detail the construction of an unbounded representative for @xmath17 in the case of a riemannian submersion @xmath27 of spin@xmath0 manifolds @xmath28 and @xmath29 .
it is given by a family of dirac operators on the fibers of @xmath30 and is defined in terms of a vertical clifford connection acting on a vertical clifford module .
after we have shown that the tensor product of the vertical clifford module with the hilbert space of @xmath31-spinors on @xmath29 is unitarily equivalent to the hilbert space of @xmath31-spinors on @xmath28 , we state our main result in theorem [ thm : fact ] .
we establish that the tensor sum @xmath32 is unitarily equivalent to @xmath33 , up to an explicit ( bounded ) curvature term .
as already mentioned , such a curvature term can not be obtained by analyzing the bounded @xmath1-product , but only arises as an obstruction to having an exact factorization of unbounded @xmath1-cycles .
we exemplify our results by homogeneous spaces for simply - connected compact semisimple lie groups .
note that this class and the corresponding factorization in @xmath1-theory is also being studied in @xcite . as a special case in this class
we obtain the factorization in @xmath1-theory of the dirac operator on the total space of the hopf fibration which was previously obtained in @xcite .
we also explain the term @xmath34 that appeared in the tensor sum in _ loc .
_ as coming from the curvature of the hopf fibration .
finally , our construction generalizes the projectability results on circle and torus principal bundles of @xcite and the @xmath1-factorization results of torus principal bundles of @xcite .
we would like to thank alain connes and nigel higson for useful suggestions .
we are grateful to simon brain , bram mesland and adam rennie for fruitful discussions .
the first author also wants to thank matthias lesch for a sequence of inspiring conversations back in 2012 concerning the possibility of applying unbounded @xmath1-theory techniques in the context of fiber bundles .
we start by giving a brief overview of riemannian submersions , referring to ( * ? ? ?
* chapter iii.d ) and ( * ? ? ?
* chapter 10.1 ) for more details .
let @xmath28 and @xmath29 be compact riemannian manifolds without boundaries and let @xmath27 be a smooth and surjective map .
we let @xmath35 and @xmath36 denote the vector fields on @xmath28 and @xmath29 and the hermitian forms on the vector fields coming from the riemannian metrics are denoted by @xmath37 and @xmath38 .
the forms on @xmath28 and @xmath29 are denoted by @xmath39 and @xmath40 , respectively .
we obtain a @xmath41-module homomorphism : @xmath42 and we equip the @xmath41-module @xmath43 with the hermitian form @xmath44 we say that @xmath45 is a _ riemannian submersion _ when @xmath46 is surjective and @xmath47 is an isometric isomorphism , where @xmath48 is the orthogonal complement of @xmath49 .
a vector field @xmath15 on @xmath28 is called _ vertical _ if @xmath50 , and _ horizontal _ if @xmath51 .
we also write @xmath52 and @xmath53 so that we have a direct sum decomposition @xmath54 moreover , @xmath55 , isometrically .
we denote by @xmath56 the orthogonal projection onto @xmath57 , where @xmath58 is the adjoint of @xmath46 with respect to the hermitian forms on @xmath35 and @xmath43 . if @xmath7 is a vector field on @xmath29 , the unique vector field @xmath59 on @xmath28 such that @xmath60 is called the _ horizontal lift _ of @xmath7 , thus @xmath61 . in what follows
, we will assume that @xmath27 is a riemannian submersion .
we remark that the lie - bracket @xmath62 $ ] of vector fields on @xmath28 restricts to a lie - bracket on the vertical vector fields @xmath57 but that the same need not be true for the horizontal vector fields .
we record the following : [ lma : hor - vert ] if @xmath15 is a vertical vector field on @xmath28 and @xmath59 a horizontal lift of a vector field @xmath7 on @xmath29 , then @xmath63 $ ] is vertical .
this follows since @xmath64)(f ) & = [ x , y_h](f \ci \pi ) = x\big ( y_h(f \ci \pi ) \big ) \\ & = x\big ( ( y \ot 1)(f ) \big ) = x\big ( y(f ) \ci \pi ) = 0 \end{split}\ ] ] for all @xmath65 .
recall that on @xmath28 and @xmath29 there are levi - civita connections @xmath66 and @xmath67 , respectively .
they are metric and torsionfree , and as a result they satisfy koszul s formula : @xmath68,z}_m - \inn{[y , z],x}_m + \inn{[z , x],y}_m \\ & \qquad + x(\inn{y , z}_m ) + y ( \inn{z , x}_m ) - z ( \inn{x , y}_m),\nonumber\end{aligned}\ ] ] for all _ real _ vector fields @xmath69 and a similar formula holds for @xmath67 .
[ lma : levi - civitas ] let @xmath66 and @xmath67 be the levi - civita connections on @xmath28 and @xmath29 , respectively .
if @xmath70 and @xmath59 are horizontal lifts of @xmath15 and @xmath7 in @xmath36 , respectively , then @xmath71.\ ] ] without loss of generality we may assume that @xmath15 and @xmath7 are real vector fields on @xmath29 .
using koszul s formula for the levi - civita connections , we determine both the vertical and horizontal part of the left - hand - side .
first , for a real vertical vector field @xmath14 we have @xmath72,z } _ m - \inn{[y_h , z],x_h}_m + \inn{[z , x_h],y_h}_m \\ & \qq + x_h ( \inn{y_h , z}_m ) + y_h ( \inn{z , x_h}_m ) - z ( \inn{x_h , y_h}_m)\\ & \q = \inn{[x_h , y_h],z}_m.\end{aligned}\ ] ] we have used lemma [ lma : hor - vert ] , the fact that horizontal and vertical parts are orthogonal , and that @xmath73 so that a vertical @xmath14 vanishes on it
. next , suppose @xmath74 is a horizontal lift of a real vector field @xmath14 on @xmath29 .
then we may use that for any horizontal lifts @xmath75 , @xmath76 together with the fact that @xmath77 - [ x , y]_h$ ] is vertical , this yields @xmath78 we will now define metric connections on the vertical and horizontal vector fields . using the levi - civita connection on @xmath28 we can define a metric connection @xmath79 on @xmath57 by @xmath80 in terms of the orthogonal projection @xmath81 and the inclusion @xmath82 . on the horizontal vector fields we define a metric connection using the pullback connection on @xmath83 .
the pullback connection @xmath84 is defined by @xmath85 for all @xmath86 , @xmath87 and @xmath88 , where @xmath89 for all @xmath90 , @xmath91 .
the pullback connection @xmath92 is a metric connection on @xmath43 .
we then define the metric connection @xmath93 on @xmath94 by @xmath95 note that this implies that the following holds @xmath96 for all @xmath87 and all @xmath97 . on the direct sum @xmath98
we may combine the above two connections @xmath79 and @xmath84 to define the _ direct sum connection _ on @xmath35 as @xmath99 this connection is metric , but might have torsion in general . in @xcite bismut ( _ cf . _
* section 10.1 ) compared the direct sum connection to the levi - civita connection on @xmath28 , finding that the difference can be expressed in terms of a three - tensor @xmath100 defined as follows : [ defn : tensors ] we introduce : 1 . the _ second fundamental form _
@xmath101 defined by @xmath102,py}_m\ ] ] for real vector fields @xmath69 .
the _ curvature of the fiber bundle @xmath45 _ is the element @xmath103 given by @xmath104 , pz}_m\ ] ] for real vector fields @xmath69 .
3 . the tensor @xmath100 defined by @xmath105 for @xmath69 .
we provide a more tangible formula for the second fundamental form : [ prop : expression - tensor ] we have that @xmath106,y}_m - \inn{[z , y],x}_m\ ] ] for all real vector fields @xmath107 and @xmath108 . in particular , we have the symmetry @xmath109 for all @xmath69 .
this follows from koszul s formula for the levi - civita connection @xmath66 , after restriction to the vertical vector fields using the orthogonal projection @xmath110 .
indeed , for real vertical vector fields @xmath111 and a real horizontal vector field @xmath14 we have that @xmath112,y}_m + \inn{[y , z],x}_m + z ( \inn{x , y}_m)\end{aligned}\ ] ] from which the result follows at once .
[ p : omedir ] the levi - civita connection @xmath66 is related to the direct sum connection @xmath113 by the following formula @xmath114 we follow the proof in ( * ? ? ?
* proposition 10.6 ) . without loss of generality
we may assume that @xmath69 are all real .
we repeatedly use lemma [ lma : hor - vert ] and the fact that @xmath115 is a tensor in @xmath116 which is anti - symmetric in @xmath7 and @xmath14 because both @xmath66 and @xmath113 are metric .
we proceed by a case - by - case study of whether @xmath111 and @xmath14 are horizontal or vertical .
moreover , tensoriality makes it sufficient to consider in the horizontal case only horizontal lifts of vector fields on @xmath29 . 1 . if @xmath116 are all horizontal lifts then @xmath113 reduces to @xmath117 and the result readily follows from lemma [ lma : levi - civitas ] and since then also @xmath118 .
2 . if both @xmath7 and @xmath14 are vertical then @xmath119 and @xmath120 both coincide with @xmath121 and @xmath122 .
if @xmath15 and @xmath7 are vertical and @xmath14 is horizontal , then @xmath120 vanishes whereas @xmath123,x}_m + \half \inn{[z , x],y}_m - \half z ( \inn{x , y}_m ) \\ & = -s(x , y , z ) = \omega(x)(y , z).\end{aligned}\ ] ] 4 .
if @xmath15 and @xmath7 are horizontal lifts and @xmath14 is vertical , then @xmath124 and lemma [ lma : levi - civitas ] implies that @xmath125,z}_m\\ & = -\half \omega(x , y , z ) = \omega(x)(y , z).\end{aligned}\ ] ] 5 .
finally , if @xmath15 is vertical and if @xmath7 and @xmath14 are horizontal lifts then @xmath126 by since @xmath127 , and lemma [ lma : hor - vert ] gives that @xmath128,z}_m - \half \inn{[y , z],x}_m + \half \inn{[z , x],y}_m \\ & = \half \omega(y , z , x ) = \omega(x)(y , z ) .\qedhere\end{aligned}\ ] ]
we now assume that the riemannian manifolds @xmath28 and @xmath29 that enter the riemannian submersion @xmath27 are even - dimensional @xmath129 manifolds . hence letting @xmath130 denote the complex clifford algebra generated by the vector fields on @xmath28 ,
we have a @xmath131-graded finitely generated projective module @xmath132 over @xmath41 together with a hermitian form @xmath133 and an isomorphism of @xmath131-graded @xmath134-algebras @xmath135 the grading operator on @xmath132 is denoted by @xmath136 .
we fix similar data and notation for the @xmath129-manifold @xmath29 .
we choose even metric connections @xmath137 and @xmath138 on @xmath132 and @xmath139 which are clifford connections for the levi - civita connections on @xmath28 and @xmath29 , respectively .
thus , they satisfy the relation @xmath140 for @xmath141 .
these even metric clifford connections are unique up to a purely imaginary one - form
. we will now pull back the clifford action and the clifford connection on @xmath29 to a horizontal clifford action and clifford connection on @xmath28 .
the _ horizontal spinor module _ is defined as the pullback of the spinor module on @xmath29 : @xmath142 the horizontal spinor module is @xmath131-graded with grading operator @xmath143 and it has a hermitian form defined by @xmath144 we let @xmath145 denote the @xmath131-graded @xmath134-subalgebra generated by the horizontal vector fields and refer to it as the _ horizontal clifford algebra_. the _ horizontal clifford action _ is then defined by @xmath146 for all @xmath147 , where @xmath148 for all @xmath149 , @xmath91 .
we equip @xmath150 with the pullback connection @xmath151 defined as in .
the connection @xmath152 is then even and metric and it is compatible with the horizontal clifford action in the sense that @xmath153 = c_h ( \na^h_x(y ) ) \q x \in \sx(m ) \
, , \ , \ , y \in \sx_h(m).\ ] ] [ l : horiclif ] the horizontal clifford action @xmath154 is an isomorphism of @xmath131-graded @xmath134-algebras .
since the @xmath155-module @xmath139 is finitely generated projective , it follows that we have an isomorphism @xmath156 of @xmath131-graded @xmath134-algebras . indeed , any simple tensor @xmath157 with @xmath158 and @xmath159 acts on @xmath160 via the assignment @xmath161 .
furthermore , using that @xmath139 is a spinor module on @xmath29 we obtain the isomorphism @xmath162 of @xmath131-graded @xmath134-algebras .
finally , since @xmath163 , isometrically , it follows that @xmath164 as @xmath131-graded @xmath134-algebras .
we will now focus on constructing a vertical clifford action and clifford connection following ( * ? ? ?
* section 10.2 ) ( to some extent ) .
we let @xmath165 denote the @xmath131-graded @xmath134-subalgebra generated by the vertical vector fields and we refer to it as the _ vertical clifford algebra_. the _ vertical spinor module _ is defined to be the @xmath131-graded @xmath41-module @xmath166 with @xmath131-grading operator @xmath167 , where @xmath168 is the @xmath131-grading operator on the dual module @xmath169 induced by the @xmath131-grading on @xmath150 .
the _ vertical clifford action _ is defined by @xmath170 for all @xmath171 .
furthermore , we have the hermitian form on @xmath172 defined by @xmath173 for @xmath174 and @xmath175 , @xmath176 .
the element @xmath177 acts on @xmath132 according to the identifications @xmath178 the reason for calling this module on @xmath28 the `` vertical spinor module '' lies in the following two results : [ l : vertclif ] the vertical clifford action @xmath179 is an isomorphism of @xmath131-graded @xmath134-algebras .
in particular , we obtain a @xmath2-orientation of the riemannian submersion @xmath45 , see ( * ? ? ?
* section @xmath180 , @xmath181 ) and ( * ? ? ?
* appendix b ) . since
@xmath169 and @xmath132 are finitely generated projective modules over @xmath182 and @xmath41 , respectively , we have the isomorphisms : @xmath183 of @xmath131-graded @xmath134-algebras . moreover , since @xmath132 defines a @xmath129-structure on @xmath28 we have that @xmath184 where the tensor product on the right - hand side is graded . now , from lemma
[ l : horiclif ] we know that @xmath185 and hence that @xmath186 where the last isomorphism of @xmath131-graded @xmath134-algebras is given explicitly by @xmath187 for @xmath188 , @xmath189 and homogeneous @xmath190 .
[ prop : spinors - vert - hor ] we have an isometric isomorphism of @xmath131-graded @xmath41-modules : @xmath191 which is compatible with the respective clifford actions under the isomorphism @xmath192 ( graded tensor product ) .
remark here that @xmath193 is @xmath131-graded with grading operator @xmath194 and that it carries the hermitian form @xmath195 notice also that @xmath196 acts on @xmath193 according to the rule @xmath197 for @xmath189 and homogeneous @xmath190 .
by lemma [ l : horiclif ] we have the following isomorphism of @xmath41-modules : @xmath198 we leave it to the reader to verify that this isomorphism is compatible with the hermitian forms , the gradings and the clifford actions .
our next task is to find an explicit even metric clifford connection on @xmath172 and compare the tensor sum of it with @xmath152 with the clifford connection @xmath137 .
we let @xmath199 denote the dual connection , on @xmath169 , described by the formula @xmath200 for real vector fields @xmath97 and @xmath201 .
we remark that the naive choice @xmath202 for the connection acting on @xmath203 does not work as this connection is _ only _ well - defined on @xmath204 . instead , we need to introduce correction terms , that make this connection well - defined on @xmath172 , expressed in terms of the tensor @xmath205 of definition [ defn : tensors ] .
we apply the notation @xmath206 and @xmath207 for the musical isomorphisms .
we then define the homomorphisms @xmath208 and @xmath209 by the formulae @xmath210
\\ \wit c(\om_1 \we \om_2 ) & : = \om_1 \ot c(\om_2^\sharp ) -
\om_2 \ot c(\om_1^\sharp ) \end{split}\ ] ] for all @xmath211 .
[ l : commcliff ] we have the relation @xmath212 = \wit c(\om)(x)\ ] ] for all @xmath97 and @xmath213 . without loss of generality
we may assume that @xmath214 for real vector fields @xmath215 and @xmath216 on @xmath28 .
furthermore , we may assume that @xmath97 is a real vector field .
we thus have that @xmath217 $ ] and an application of the clifford relation @xmath218 ( for _ real _ vector fields @xmath219 ) entails that @xmath220 = \frac{1}{4 } \big[c(x ) , [ c(x_1),c(x_2)]\big ] \\ &
\q = \inn{x , x_1}_m c(x_2 ) - \inn{x , x_2}_m c(x_1 ) = \wit c(\om_1 \we \om_2)(x ) \qedhere \end{split}\ ] ] [ prop : spin - connections ] the following formula defines an even metric clifford connection @xmath221 on @xmath222 : @xmath223 where @xmath97 , @xmath224 and @xmath225 .
let @xmath97 be real and let @xmath201 and @xmath225 be given .
we start by showing that @xmath226 is well - defined on @xmath203 .
thus , let @xmath227 be real .
we need to establish that @xmath228 we compute as follows : @xmath229 \big)(s ) \end{split}\ ] ] where the second identity follows since both @xmath230 and @xmath152 are clifford connections , see and .
the identity in is now a consequence of proposition [ p : omedir ] and lemma [ l : commcliff ] .
it is clear that @xmath231 commutes with the grading operator @xmath232 on @xmath172 .
we continue by showing that @xmath221 is a clifford connection for the action of @xmath233 .
this amounts to proving the identity @xmath234 ( \bra{\xi } \ot s)= c_v(\nabla^v_x(z ) ) ( \bra{\xi } \ot s)\ ] ] for any vertical vector field @xmath14 .
we thus compute as follows : @xmath235 \big ) s \end{split}\ ] ] it therefore suffices to verify that @xmath236 = c ( \na^v_x(z)).\ ] ] but this is a consequence of proposition [ p : omedir ] and lemma [ l : commcliff ] .
we finally show that @xmath237 is metric .
thus , let @xmath238 and @xmath239 be two extra elements .
by lemma [ l : horiclif ] we may find a unique @xmath240 with @xmath241 .
using that both @xmath230 and @xmath152 are metric clifford connections we then compute that @xmath242(t)}_{\se_m}. \end{split}\ ] ] the fact that @xmath237 is metric can now be derived from proposition [ p : omedir ] and lemma [ l : commcliff ] .
the following result is a restatement of ( * ? ? ?
* lemma 10.13 ) .
mean curvature _ of the riemannian submersion @xmath45 is defined as the @xmath243-form @xmath244 where @xmath245 is the _ trace _ defined by @xmath246 for real vector fields @xmath247 .
[ lma : mean - curv ] we have the identity @xmath248 to ease the notation we introduce the elements @xmath249 and @xmath250 in @xmath251 defined by @xmath252 and we thus have that @xmath253 we claim that @xmath254 to prove the first of these identities we write @xmath255 for real vector fields @xmath256 and @xmath257 , and we notice that the symmetry relation in proposition [ prop : expression - tensor ] implies that @xmath258 using the clifford relation we then obtain that @xmath259 + c(x_j^2 ) [ c(x_i^1 ) , c(x_k^3 ) ] \big ) \\ & \q = 2 \sum_{i , j , k } \inn{x_i^1,x_j^2}_m \cd c(x_k^3 ) = 2 \big ( \tex{tr } \ot ( c \ci \sharp ) \big)(s ) \ , , \end{split}\ ] ] proving the first identity in .
to verify the second identity in we write @xmath260 for real vector fields @xmath261 and @xmath262 .
we then have that @xmath263 using the clifford relation we obtain that @xmath264 + c(y_k^3 ) [ c(y_i^1 ) , c(y_j^2 ) ] \\
& \qqq - c(y_i^1 ) [ c(y_j^2 ) , c(y_k^3 ) ] \big ) \\ & = - \sum_{i , j , k } [ c(y_i^1 ) , c(y_j^2 ) ] c(y_k^3 ) = - ( c \ot c)(1 \ot
\sharp)(\om ) \ , , \end{split}\ ] ] and this ends the proof of the lemma .
in this section we come to the main result of this paper , which is the factorization in unbounded @xmath1-theory of the dirac operator @xmath33 on @xmath28 in terms of a vertical operator @xmath265 and the dirac operator @xmath266 on @xmath29 for a riemannian submersion @xmath267 of compact spin@xmath268 manifolds .
we let @xmath269 and @xmath270 denote the hilbert space completions of the spinor modules @xmath132 and @xmath139 with respect to the hermitian forms and the measures @xmath271 and @xmath272 coming from the riemannian metrics .
our task is then to find a @xmath273-correspondence from @xmath274 to @xmath275 , together with a self - adjoint and regular unbounded operator @xmath265 on @xmath15 , such that @xmath276 and in such a way that the operator @xmath33 corresponds to the tensor sum @xmath277 for some metric connection @xmath278 on @xmath15 .
first , let us translate the results of the previous section on connections on the vertical spinor bundle on @xmath28 to a hilbert @xmath273-module over @xmath275 .
in fact , since @xmath172 is a @xmath41-module , it becomes a @xmath155-module ( denoted @xmath279 ) using the pullback map @xmath280 .
moreover , @xmath279 can be equipped with a @xmath155-valued inner product @xmath281 defined by integration along the fibers : @xmath282 using the measures coming from the riemannian metrics on the riemannian submanifolds of @xmath28 : @xmath283 , @xmath284 .
we define the hilbert @xmath273-module @xmath15 to be the completion of @xmath279 in the norm coming from this inner product and the @xmath273-norm on @xmath275 .
there is a left action of @xmath274 on @xmath15 ( coming from the left action of @xmath41 on @xmath172 ) and this gives @xmath15 the structure of a @xmath273-correspondence from @xmath274 to @xmath275 .
moreover , @xmath15 is @xmath131-graded with grading operator @xmath285 induced by the grading operator @xmath286 on @xmath172 . in order to relate the interior tensor product @xmath287 to @xmath269 we fix
coordinate charts @xmath288 on @xmath29 and @xmath289 on a model fiber @xmath290 together with a local trivialization @xmath291 putting @xmath292 we have a coordinate chart @xmath293 on @xmath28 with @xmath294 we refer to such a chart as a _
fibration chart_. combining the fibration chart @xmath293 with the riemannian metric we obtain the positive invertible matrix of smooth functions @xmath295 furthermore , letting @xmath296 denote the projection @xmath297 onto the last @xmath298 copies of @xmath299 , we have the following invertible positive matrix of smooth functions : @xmath300 we may also combine the coordinate chart @xmath288 with the riemannian metric on the base obtaining the invertible positive matrix of smooth functions : @xmath301 by taking determinants , we thus obtain the three positive functions @xmath302 we define the upper triangular matrix of smooth functions @xmath304 by putting @xmath305 for @xmath306 we compute that @xmath307 and we thus have that @xmath308 using that @xmath45 is a riemannian submersion we conclude that @xmath309 .
therefore , since @xmath310 is upper - triangular and @xmath243 on the diagonal , we obtain that @xmath311 proving the lemma . using that @xmath315
( see lemma [ l : horiclif ] ) and the fact that @xmath139 ( and hence also @xmath150 ) are finitely generated projective modules , it follows that @xmath316 is an isomorphism of @xmath131-graded left modules over @xmath41 . by a density argument
we obtain the result of the proposition if we can establish that @xmath317 for all @xmath318 , @xmath319 and @xmath320 . to ease the notation we put @xmath321 where we suppress the identifications @xmath322
. moreover , without loss of generality , we assume that @xmath323 .
we expand on the right - hand side of : @xmath324 now , for each @xmath325 we compute that @xmath326 where @xmath327 is lebesgue measure on @xmath328 . moreover , @xmath329 for all @xmath330 . using lemma [ l : indepchart ]
we thus have that @xmath331 and the result of the proposition can be obtained by noting that @xmath332 we show that @xmath336 is symmetric , thus that @xmath337 without loss of generality , assume that @xmath338 where @xmath339 is a fibration chart .
we then notice that @xmath340 and similarly for @xmath341 .
to ease the notation , we define the vertical vector fields @xmath342 let @xmath325 . using that @xmath237 is a metric clifford connection ( see proposition [ prop : spin - connections ] )
we obtain that @xmath343 using the det / log relationship @xmath344 we then reduce the proof of to a verification of the identity @xmath345 of vertical vector fields .
however , using koszul s formula for the levi - civita connection on @xmath28 , we see that @xmath346 , \pa/\pa \si_l}_m \big ) \cd
\pa/\pa \si_j \\ &
\q = \frac{1}{2}\sum_{l > \dim(b ) } \tex{tr}\big ( \pa/\pa \si_l ( qgq ) \cd ( qgq)^{-1 } \big ) \cd x_l \ , , \end{split}\ ] ] thus proving and thereby also that @xmath336 is symmetric .
let @xmath350 denote the symbol of the first order differential operator @xmath351 .
by lemma [ l : symmetry ] and @xcite , it suffices to verify that @xmath352 is invertible whenever @xmath353 , @xmath354 and @xmath355 is non - trivial ( where @xmath356 denotes the vertical tangent bundle ) .
but this follows from the local formula @xmath357 = i \sum_{j = 1}^{\dim(f ) } c_v(e_j ) e_j(f)$ ] which holds for all @xmath88 .
indeed , this formula implies that @xmath358 for all @xmath354 .
next , consider the dirac operator @xmath359 .
it is defined as the closure of the unbounded operator @xmath360 defined locally for any local orthonormal frame @xmath361 for @xmath36 consisting of real vector fields . in order to form the unbounded kasparov product of the vertical and the horizontal components we need to lift the dirac operator @xmath266 to an unbounded selfadjoint operator on the hilbert space @xmath287 . to carry this out ,
we need a metric connection on the hilbert @xmath273-module @xmath15 .
we let @xmath364 denote the @xmath273-correspondence from @xmath275 to @xmath275 defined as the completion of the smooth form @xmath365 with respect to ( the norm coming from ) the @xmath275-valued inner product @xmath366 .
[ l : locmean ] let @xmath288 and @xmath289 be coordinate charts on @xmath29 and @xmath290 , resp . and suppose that @xmath372 is a local trivialization .
for any vector field @xmath368 , we have the local formula : @xmath373 where @xmath293 is the coordinate chart given by @xmath374 , @xmath375 . without loss of generality ,
we assume that @xmath368 is real . recall that @xmath376 where @xmath377 is given by @xmath378,y}_m - \inn { x , [ z_h , y]}_m \big ) \ , .\ ] ] we have the local formula @xmath379 using the explicit formula for @xmath380 a direct computation then shows that @xmath381 the result of the lemma thus follows from the det / log relationship : @xmath382 the first two identities can be verified by a straightforward computation , so we focus on the third identity . without loss of generality
we assume that @xmath386 where @xmath339 is a fibration chart of the form @xmath387 and @xmath388 .
we may also assume that @xmath389 for some @xmath390 .
using that @xmath237 is metric , by proposition [ prop : spin - connections ] , we then have that @xmath391 to ease the notation , we put @xmath392 . using lemma [ l : locmean ] we compute the integrand in the above expression : @xmath393 combining these computations
we thus obtain that @xmath394 this ends the proof of the proposition .
the result follows since @xmath397 is a metric connection , by proposition [ p : metconx ] , and since @xmath398 = i \sum_{\al = 1}^{\dim(b ) } c(f_\al ) f_\al(h)$ ] , whenever @xmath399 has support inside the domain of a local orthonormal frame @xmath400 of vector fields on @xmath29 .
the tensor sum we are after is given by the odd symmetric unbounded operator @xmath401 where the domain is the image of @xmath395 in @xmath287 .
the closure of the symmetric unbounded operator @xmath402 will be denoted by @xmath403 . before we start comparing @xmath404 with the dirac operator @xmath33 we present a useful lemma on the various connections involved .
we recall that @xmath405 was introduced in definition [ defn : tensors ] and that the maps @xmath208 and @xmath209 are defined in .
[ l : commconn ] let @xmath406 be given . for all @xmath97
we then have the identities @xmath407 = [ \na_x^{\se_h},t ] + \frac{1}{4 } [ t , c ( \omega(x))]\end{split}\ ] ] of endomorphisms of @xmath132 , where we are suppressing the identifications @xmath408 without loss of generality , we may assume that @xmath409 for some vector field @xmath87 . the fact that @xmath137 and @xmath152 are clifford connections implies that we can use proposition [ p : omedir ] for the corresponding connections on vector fields . indeed : @xmath410 = c(\nabla^h_x(y_h ) ) + c ( \omega(x)(y_h,\cdot)^\sharp ) \\ & \qquad = [ \na_x^{\se_h } , c_h(y_h ) ] + \tilde c ( \omega(x))(y_h)\\ & \qquad = [ \na_x^{\se_h } , c_h(y_h ) ] + \frac{1}{4 } [ c(y_h ) , c(\omega(x ) ) ] \ , , \end{aligned}\ ] ] where we also used lemma [ l : commcliff ] in passing to the last line .
let us fix an element @xmath413 in the core of @xmath414 , with @xmath415 of the form @xmath416 .
thus , @xmath224 and @xmath225 .
we will show that @xmath417 since @xmath418 descends to an isomorphism of the core @xmath395 for @xmath414 with the core @xmath132 for @xmath33 the above identity will prove the result of the theorem . without loss of generality
, we may assume that @xmath419 and @xmath420 for some open set
@xmath421 admitting real local orthonormal frames @xmath422 and @xmath423 defined on @xmath424 and @xmath425 , resp . using this observation
we compute the vertical part : @xmath430 using the observation in one more time , we compute the horizontal part : @xmath431 applying lemma [ lma : mean - curv ] and the above computations , we see that the sum of the vertical and the horizontal part is given by @xmath432 [ thm : fact ] suppose that @xmath267 is a riemannian submersion of even - dimensional spin@xmath0-manifolds . then the even spectral triple @xmath433 is the unbounded kasparov product of the even unbounded kasparov module @xmath434 with the even spectral triple @xmath435 up to the curvature term @xmath436 .
we will show that the bounded transforms of the above unbounded kasparov modules coincide with the shriek maps defined in @xcite .
the result then follows from the wrong - way functoriality of the shriek map .
specifically , we know that the factorization @xmath437 \hot_{c(b ) } \pi ! = [ b]$ ] holds in @xmath1-theory ( _ cf .
_ equation ) . recall ( * ? ? ?
* proposition 2.9 ) that for any @xmath2-oriented submersion @xmath4 the class @xmath3 can be represented by a continuous family of pseudodifferential operators of order @xmath6 ( parametrized by @xmath24 ) on the vertical spinor bundle . for each @xmath24 , the principal symbol is given by @xmath438 for any element @xmath439 , where @xmath440 .
in particular , if @xmath15 is a spin@xmath0 manifold and @xmath7 consists of a single point , the bounded transform @xmath441 of the dirac operator is a representative of the shriek map , which in this case is the fundamental class @xmath8 $ ] in @xmath2-homology .
moreover , from the construction of the unbounded kasparov module @xmath434 we have that @xmath442 is a continuous family of pseudodifferential operators of order zero with the correct principal symbols , see the proof of proposition [ p : unboukas ] .
we thus conclude that the spectral triples @xmath433 and @xmath443 represent the respective fundamental classes @xmath437 \in kk_0(c(m),\c)$ ] and @xmath444 \in kk_0(c(b),\c)$ ] , and that the unbounded kasparov module @xmath445 represents the class @xmath17 in @xmath446 . as a special case of theorem [ thm : fact ] we can consider the following situation .
let @xmath447 be a simply - connected compact semisimple lie group and @xmath448 a closed subgroup and consider @xmath449 .
the killing form on the corresponding lie algebra @xmath450 then induces a riemannian metric on @xmath447 and on @xmath451 .
in fact , we can identify @xmath452 in terms of the lie algebra @xmath453 of @xmath448 .
it is clear that then @xmath30 is a riemannian submersion .
moreover , under the above assumptions on @xmath447 both @xmath447 and @xmath451 are spin manifolds ( for a construction of dirac operators @xmath454 and @xmath455 on @xmath447 and @xmath451 we refer to @xcite ) .
hence , theorem [ thm : fact ] applies and yields a factorization of @xmath454 as a tensor sum @xmath456 up to an explicit curvature - term . in particular
, this applies to the hopf fibration @xmath457 for which the explicit factorization was already obtained in @xcite .
in fact , one can compute that @xmath458 in terms of the vertical @xmath459 and horizontal vector fields @xmath460 .
the curvature term appearing in theorem [ thm : fact ] above is then @xmath461 \gamma_1= -\frac 12,\ ] ] where we have expressed @xmath462 and @xmath463 in terms of the pauli matrices @xmath464 , with the convention used in @xcite .
this is an independent check for the @xmath465 appearing in ( * ? ? ?
* theorem 6.31 ) , at the same time giving meaning to it as a curvature term .
more generally , such homogeneous spaces are subject of study in @xcite . | we establish the factorization of dirac operators on riemannian submersions of compact spin@xmath0 manifolds in unbounded @xmath1-theory .
more precisely , we show that the dirac operator on the total space of such a submersion is unitarily equivalent to the tensor sum of a family of dirac operators with the dirac operator on the base space , up to an explicit bounded curvature term .
thus , the latter is an obstruction to having a factorization in unbounded @xmath1-theory . we show that our tensor sum represents the bounded @xmath1-product of the corresponding @xmath1-cycles and connect to the early work of connes and skandalis . |
the characteristic polynomials of random matrices have attracted considerable attention in the last few years , one reason being that their correlations seem to reflect the correlations of the eigenvalues @xcite . in this paper
we continue with our investigation @xcite of the second - order correlation function of the characteristic polynomial of a sample covariance matrix .
this function is given by @xmath0 where @xmath1 , @xmath2 , and @xmath3 is a complex or real sample covariance matrix defined as follows : * complex sample covariance matrices .
* let @xmath4 be a distribution on the real line with expectation @xmath5 , variance @xmath6 and finite fourth moment @xmath7 , and for given @xmath1 with @xmath8 , let @xmath9 denote the @xmath10 matrix whose entries @xmath11 are complex random variables whose real and imaginary parts are independent , each with distribution @xmath4 .
let @xmath12 denote the conjugate transpose of @xmath13 .
then the hermitian @xmath14 matrix @xmath15 is called the ( unrescaled ) _ sample covariance matrix _ associated with the distribution @xmath4 .
* real sample covariance matrices .
* let @xmath4 be a distribution on the real line with expectation @xmath5 , variance @xmath16 and finite fourth moment @xmath7 , and for given @xmath1 with @xmath8 , let @xmath9 denote the @xmath10 matrix whose entries @xmath11 are real random variables with distribution @xmath4 .
let @xmath17 denote the transpose of @xmath18 .
then the symmetric @xmath14 matrix @xmath19 is called the ( unrescaled ) _ sample covariance matrix _ associated with the distribution @xmath4 .
we are interested in the asymptotic behavior of the values @xmath20 as @xmath21 , for certain choices of the sequences @xmath22 , @xmath23 , @xmath24 , @xmath25 . in a recent paper @xcite we considered the `` square '' case where the difference @xmath26 is fixed .
we showed that in this situation , the second - order correlation function of the characteristic polynomial of a complex sample covariance matrix is asymptotically given ( after the appropriate rescaling ) by the sine kernel in the bulk of the spectrum , by the airy kernel at the soft edge of the spectrum , by the bessel kernel at the hard edge of the spectrum .
moreover , similar results were obtained for real sample covariance matrices .
the purpose of this note is to derive similar results for the `` non - square '' case where the difference @xmath26 tends to infinity in such a way that the ratio @xmath27 tends to some constant @xmath28 sufficiently quickly .
we will show that in this situation , the second - order correlation function of the characteristic polynomial of a complex sample covariance matrix is asymptotically given ( after the appropriate rescaling ) by the sine kernel in the bulk of the spectrum , by the airy kernel at the edge of the spectrum .
note that both edges of the spectrum are `` soft '' when @xmath28 .
these results are in consistence with results by ben arous and pch @xcite , tao and vu @xcite , soshnikov @xcite , pch @xcite , and feldheim and sodin @xcite , who obtained similar results for the ( more relevant ) correlation function of the eigenvalues , yet under stronger assumptions on the underlying distributions .
ben arous and pch @xcite proved the occurrence of the sine kernel in the bulk of the spectrum for a certain class of complex sample covariance matrices with @xmath29 .
very recently , tao and vu @xcite extended this result to a quite general class of sample covariance matrices , still with @xmath30 .
soshnikov @xcite and pch @xcite established the occurrence of the airy kernel at the upper edge of the spectrum for real and complex sample covariance matrices whose underlying distributions are symmetric with exponential tails ( or at least finite 36th moments ) .
recently , feldheim and sodin @xcite proposed another approach to obtain these results , which also works for the lower edge of the spectrum .
thus , our results add some support to the wide - spread expectation that correlation functions in random matrix theory are universal , and subject to very weak moment conditions only .
it is well - known that the asymptotic distribution of the global spectrum of the rescaled sample covariance matrix @xmath31 is given ( both in the complex case and in the real case ) by the marenko - pastur distribution of parameter @xmath32 , which has the density @xmath33 where @xmath34 and @xmath35 . note that @xmath36 as well as @xmath37 implicitly depend on @xmath32 , but this dependence will be kept implicit throughout this paper
also , note that @xmath36 are the solutions to the equation @xmath38 unless otherwise mentioned , we will always assume that @xmath39 for all @xmath40 , @xmath41 as @xmath21 , and @xmath42 as @xmath21 , where @xmath43}0;1{]}$ ] .
our main results are as follows : [ bulk-1 ] suppose that @xmath44 and @xmath45 as @xmath21 .
let @xmath46 denote the second - order correlation function of a complex sample covariance matrix satisfying our standing moment conditions .
for any @xmath47 , @xmath2 , setting @xmath48 @xmath49 , @xmath50 , we have @xmath51 where @xmath52 and @xmath53 is defined by @xmath54 [ bulk-2 ] suppose that @xmath44 and @xmath45 as @xmath21 .
let @xmath46 denote the second - order correlation function of a real sample covariance matrix satisfying our standing moment conditions .
for any @xmath47 , @xmath2 , setting @xmath48 @xmath49 , @xmath50 , we have @xmath55 where @xmath56 and @xmath57 is the sine kernel defined by @xmath58 [ edge-1 ] suppose that @xmath44 and @xmath59 as @xmath21 .
let @xmath46 denote the second - order correlation function of a complex sample covariance matrix satisfying our standing moment conditions .
for any @xmath2 , setting @xmath60 @xmath61 , @xmath62 , we have @xmath63 for @xmath64 and , when @xmath65 , @xmath66 for @xmath67 , where @xmath52 and @xmath68 is the airy kernel defined by @xmath69 [ edge-2 ] suppose that @xmath44 and @xmath59 as @xmath21 .
let @xmath46 denote the second - order correlation function of a real sample covariance matrix satisfying our standing moment conditions .
for any @xmath2 , setting @xmath60 @xmath61 , @xmath62 , we have @xmath70 for @xmath64 and , when @xmath65 , @xmath71 for @xmath67 , where @xmath56 and @xmath72 is defined by @xmath73 note that in all the cases , the limit is almost universal in that it depends on the underlying distribution only via a multiplicative factor containing the fourth cumulant .
also , note that we obtain the sine kernel in the complex setting not only for @xmath30 but also for @xmath74 .
this paper is organized as follows .
section 2 contains the proofs of theorems [ bulk-1 ] and [ edge-1 ] .
( theorems [ bulk-2 ] and [ edge-2 ] may be proved similarly . ) in section 3 we state some auxiliary results concerning bessel functions which are needed for the proofs .
the starting point for the proofs of theorems [ bulk-1 ] and [ edge-1 ] will be the following exponential - type generating function for the second - order correlation function of the characteristic polynomial of the ( unrescaled ) sample covariance matrix satisfying the moment conditions from section 1 : [ gf ] for any @xmath75 , @xmath76 , @xmath77 where @xmath52 and @xmath78 in the complex setting , @xmath79 and @xmath80 in the real setting , and @xmath81 denotes the modified bessel function of order @xmath82 .
note that @xmath83 is an even entire function of @xmath84 , as follows directly from the power series representation of the modified bessel function .
thus , the right - hand side in extends to a meromorphic function on the complex plane , with a singularity at @xmath85 . in the subsequent calculations
, we will always consider the principal branch of @xmath86 .
we only give the proofs for the complex setting , the proofs for the real setting being very similar .
we will always assume that @xmath41 and @xmath42 for some constant @xmath43}0,1{]}$ ] .
( the case where @xmath87 remains bounded is already considered in @xcite . )
furthermore , we will usually omit the index @xmath88 of the parameters , and also of several auxiliary functions defined below . by proposition [ gf ]
, we have the integral representation @xmath89 where @xmath90 denotes a contour around the origin .
similarly as in @xcite , we will show that the main contribution to the contour integral in comes from a small neighborhood of the point @xmath91 . to determine the asymptotic behavior of the values @xmath92 in theorems [ bulk-1 ] and [ edge-1 ]
, we proceed similarly as in @xcite , but we have to address some additional complications arising from the circumstance that @xmath93 . to deal with the this problem , we will use various well - known _ uniform asymptotics _ for the modified bessel function of large positive order ( see chapters 10 and 11 in olver @xcite ) . for convenience , we state these results ( and several bounds deduced therefrom ) in section 3 of this paper .
we will choose the contour @xmath94 consisting of the following parts @xmath95 , @xmath96 ( see figure [ newcontour ] ) : * @xmath97 , @xmath98 , * @xmath99 , @xmath100 , * @xmath101 , @xmath102 , * @xmath103 , @xmath104 , * @xmath105 , @xmath106 , here @xmath107 is a positive real number which will finally be chosen sufficiently large , and @xmath108 for the bulk of the spectrum and @xmath109 for the edge of the spectrum .
.,width=249 ] the rough idea behind the choice of this contour is the following : after the appropriate rescaling , the main contribution to the integral comes from the part @xmath110 , for which we can use the uniform asymptotics for the modified bessel function in the positive half - plane .
the other parts are essentially error terms , which can be made arbitrary small by picking @xmath111 and @xmath88 sufficiently large . however ,
to prove this , we need rather precise approximations and bounds for the modified bessel function close to the imaginary axis , and this is our main motivation for the choice of the above contour . for the contours @xmath112 and @xmath113
, we can use the uniform bound for the modified bessel function _ near _ the imaginary axis , but away from the turning points @xmath114 .
for the contours @xmath115 and @xmath116 , we can use the uniform bounds and for the modified bessel function _ on _ the imaginary axis , including the vicinity of the turning points @xmath114 . in the following proofs
, we adopt the convention that all asymptotic bounds may depend on the sequence @xmath117 as well as on the `` shift parameters '' @xmath118 , unless otherwise indicated .
a similar remark applies to large positive constants , typically denoted by @xmath119 . after these comments , we turn to the proofs of the main theorems .
let @xmath118 be as in theorem [ bulk-1 ] , and let @xmath120 .
we assume throughout this proof that @xmath88 is sufficiently large . by ( [ intrep ] ) and the definition of @xmath121
, we have @xmath122 where @xmath123 @xmath124 and @xmath94 is the contour specified above , with @xmath108 .
note that @xmath90 implicitly depends on @xmath111 . for any @xmath125 $ ]
, let @xmath126 where @xmath127 .
note that by we have @xmath128 for @xmath129 $ ] .
we will show that for any @xmath130 , there exists some @xmath120 such that the following holds : @xmath131 @xmath132 @xmath133 then , by symmetry , similar bounds hold for the integrals along @xmath112 and @xmath115 . using these results ,
it is easy to see that @xmath134 now , by laplace inversion , we have , for @xmath135 , @xmath136 , @xmath137 ( see p.245 in @xcite ) .
we therefore get @xmath138 now replace the local shift parameters @xmath139 with @xmath140 and observe that @xmath141 ( which is a simple consequence of ) to obtain . to prove , we write @xmath142 where @xmath143 and @xmath144 since @xmath145 a simple calculation yields @xmath146 ( recall our convention that implicit constants in @xmath147-terms may depend on @xmath148 , @xmath118 , which are regarded as fixed ) .
we first prove . in doing so
, we use the notation @xmath149 to denote a bound involving an implicit constant depending also on @xmath111 ( in addition to @xmath148 , @xmath118 ) .
let @xmath150 .
substituting @xmath151 and @xmath152 in the integral on the left - hand side in , we obtain @xmath153 now , using taylor expansion , we have the following approximations , for @xmath154 : @xmath155 for the second approximation , we have also used the uniform asymptotics for the modified bessel function .
indeed , since @xmath156 @xmath157 and @xmath158 for @xmath159 , @xmath160 , it follows from that @xmath161 whence the approximation for @xmath162 . inserting the above approximations into and recalling the definition of @xmath163 yields .
we now prove and . for @xmath164 , write @xmath165 with @xmath166 $ ] , @xmath167 $ ] .
since @xmath120 , we then have the estimates @xmath168 @xmath169 @xmath170 @xmath171 in particular , and show that @xmath172 is uniformly bounded .
moreover , @xmath173 the proof of and is therefore reduced to showing that for any @xmath130 , there exists some @xmath120 such that @xmath174 for the first integral in , note that for @xmath175 $ ] and fixed @xmath120 , we have @xmath156 and @xmath176 writing @xmath177 with @xmath175 $ ] and using , as well as the uniform bound for the modified bessel function , it therefore follows that for all sufficiently large @xmath40 ( depending on @xmath111 ) , @xmath178 where @xmath119 is a positive constant which does not depend on @xmath111 and which may change from step to step in the sequel . combining this with the estimates
@xmath179 , @xmath180 , it follows that @xmath181 since for @xmath120 sufficiently large , the right - hand side is clearly bounded above by @xmath182 , this proves . for the second integral in , note that for @xmath183 , we have @xmath184 and @xmath185 from and .
thus , we may use the bounds and for the modified bessel function on the imaginary axis .
writing @xmath186 with @xmath187 $ ] , we have , for all sufficiently large @xmath40 , @xmath188 and @xmath189 hence , by and , for all sufficiently large @xmath40 , @xmath190 where @xmath119 is a positive constant which depends only on @xmath191 and which may change from occurrence to occurrence in the sequel .
it follows that @xmath192 since for @xmath120 sufficiently large , this is clearly @xmath193 for all sufficiently large @xmath40 , this proves .
the proof of theorem [ bulk-1 ] is complete now .
let @xmath194 , @xmath2 , and @xmath120 .
when we use the symbols @xmath195 and @xmath196 , we mean the upper sign for @xmath64 and the lower sign for @xmath67 .
we assume throughout this proof that @xmath88 is sufficiently large .
again , by ( [ intrep ] ) and the definition of @xmath121 , we have @xmath197 where @xmath198 @xmath199 and @xmath94 is the contour specified above , with @xmath109 .
note that @xmath90 implicitly depends on @xmath111 . for any @xmath125
$ ] , let @xmath200 we will show that for any @xmath130 , there exists some @xmath120 such that the following holds : @xmath201 @xmath202 @xmath203 then , by symmetry , similar bounds hold for the integrals along @xmath112 and @xmath115 . using these results , it is easy to see that @xmath204 substituting @xmath205 , @xmath206 , and shifting the path of integration back to the line @xmath207 ( which is easily justified by cauchy s theorem ) , we obtain @xmath208 making the replacements @xmath209 , @xmath210 and multiplying by @xmath211 , we further obtain @xmath212 by the integral representation for the airy kernel ( see proposition 2.2 in @xcite ) , the latter expression is equal to @xmath213 , whence and . to prove , we proceed similarly as in the preceding proof . to begin with , we write @xmath214 where @xmath215 and @xmath144 similarly as above , we then have @xmath216 starting from this representation , the proof is by and large similar to the preceding proof
. however , a notable difference is given by the fact the leading - order terms cancel out now , which is why we have to keep track of several additional terms in the asymptotic approximations .
we begin with the contour integral for @xmath110 .
let @xmath217 .
substituting @xmath151 and @xmath218 in the integral on the left - hand side in , we obtain @xmath219 similarly as in the preceding proof , we use the notation @xmath149 to denote a bound involving an implicit constant depending also on @xmath111 ( in addition to @xmath148 , @xmath118 ) .
then , putting @xmath220 for abbreviation and using taylor expansion , we have the approximations , for @xmath154 , @xmath221 for the second approximation , we have used the uniform asymptotics for the modified bessel function .
indeed , since @xmath222 @xmath223 and @xmath224 for @xmath159 , @xmath160 , it follows from that @xmath225 now , by straightforward expansion , @xmath226 plugging this into the preceding expression and using once again the assumption @xmath227 yields the approximation for @xmath228 . inserting the above approximations into and ordering with respect to fractional powers of @xmath88
, we obtain @xmath229 with @xmath230 by the characterizing equation for the edge of the spectrum , we have @xmath231 thus , the terms in the large round brackets cancel out , and the remaining sum can be simplified to @xmath232 recall that @xmath233 at the upper edge and the lower edge of the spectrum , respectively .
plugging this into the expression for @xmath234 and simplifying , we obtain @xmath235 and the proof of is complete .
the proof of is similar to that of . setting @xmath236 with @xmath175 $ ] and assuming that @xmath40 is large enough ,
we have the following bounds : @xmath237 let us comment on the second bound ( the others bounds being straightforward ) .
the uniform asymptotic approximation is not applicable anymore , since @xmath238 approaches the imaginary axis as @xmath239 : @xmath240 however , we may use the uniform asymptotic bound instead , since @xmath241 @xmath242 , as follows from a similar estimate as in .
we thus obtain , for all @xmath40 sufficiently large ( depending on @xmath111 ) , @xmath243 where @xmath244 is a constant not depending on @xmath111 .
since @xmath245 this entails the asserted bound by substituting @xmath246 and simplifying .
putting it all together and doing similar simplifications as in the proof of , we find that the integral on the left - hand side in is bounded above by @xmath247 where @xmath119 is a positive constant which does not depend on @xmath111 ( and which may change from step to step as usual ) and @xmath248 the real part of this expression is obviously of order @xmath249 .
it therefore follows that @xmath250 since for @xmath120 sufficiently large , the right - hand side is clearly bounded above by @xmath182 , this proves .
it remains to prove .
using similar estimates as in and , we see that it suffices to show that @xmath251 for @xmath120 sufficiently large .
moreover , by similar estimates as in and , we have @xmath252 and @xmath253 for all @xmath254 $ ] , so we can use and to bound @xmath255 .
let @xmath256 be a small constant which will be chosen later , let @xmath257 denote the subset of those @xmath254 $ ] such that @xmath258 , and let @xmath259 denote the complement of this subset . using the stronger bound on @xmath259 and the weaker bound on @xmath257 , we then have , for sufficiently large @xmath40 , @xmath260 here @xmath261 denotes the length of the interval @xmath257 , @xmath119 is a constant which depends only on @xmath191 , and @xmath262 is a constant which may additionally depend on @xmath263 .
( of course , both constants may change from occurrence to occurrence as usual . ) clearly , once @xmath264 is fixed , we may make the first term arbitrarily small by choosing @xmath107 sufficiently large .
hence , to complete the proof of , it remains to establish an appropriate bound on the second term , on @xmath265 . to this purpose , first note that @xmath266 and therefore @xmath267 for the upper edge of the spectrum , we have @xmath268 , which implies that the interval @xmath257 is empty for @xmath263 sufficiently small and @xmath88 sufficiently large , and is proven .
for the lower edge of the spectrum , we have @xmath269 as well as @xmath270 ( as is readily verified ) and therefore , for @xmath263 sufficiently small and @xmath88 sufficiently large , @xmath271 \cap 2 \arcsin \left ( \frac{\xi}{(1-\gamma_\infty)(1 + 2\varepsilon)};\frac{\xi}{(1-\gamma_\infty)(1 - 2\varepsilon ) } \right ) \,.\end{aligned}\ ] ] in particular , @xmath272 for some positive constant @xmath273 depending only on @xmath191 and @xmath32 .
moreover , @xmath261 can be made arbitrarily small by picking @xmath263 sufficiently close to zero .
thus , is also proven .
the proof of theorem 1.3 is complete now .
in this section we state some uniform asymptotic approximations for bessel functions from the literature ( see chapters 10 and 11 in olver @xcite ) . moreover
, we deduce several asymptotic bounds for bessel functions which are sufficient for controlling the error bounds in the preceding proofs .
these asymptotic bounds should be well - known , but we have not been able to find explicit statements suiting our purposes in the literature . throughout this section ,
we assume that @xmath75 and @xmath274 .
by proposition 3.1 we have the following bound on @xmath282 , uniformly in the set @xmath283 : @xmath284 using proposition 3.2 it can be shown that this bound in fact remains valid up to the imaginary axis : fix @xmath264 , and consider @xmath288 with @xmath75 , @xmath289 @xmath290 @xmath291 .
we then have the uniform asymptotic approximation .
we now use the following asymptotic approximations for the airy function and its derivative , which hold for @xmath159 , @xmath292 ( see chapters 11.1 and 11.8 in olver @xcite ) : @xmath293 @xmath294 using the inequalities @xmath295 , @xmath296 , and the fact that @xmath297 , @xmath298 are holomorphic functions , it follows that @xmath299 @xmath300 uniformly in @xmath301 such that @xmath302 and @xmath292 .
it can be checked that for @xmath75 , @xmath303 , and for some @xmath304 depending only on @xmath264 .
hence , @xmath305 @xmath306 uniformly in @xmath303 , where @xmath307 is defined as in and @xmath308 . inserting this into , it follows that @xmath309 the @xmath277-bound holding uniformly in @xmath303 . thus , for @xmath310 , we have @xmath311 when @xmath312 and @xmath313 when @xmath314 .
a comparison of the definitions and ( together with a careful discussion of the choice of branches ) shows that @xmath315 which completes the proof of the lemma .
[ jboundlemma ] there exists a constant @xmath244 such that for all sufficiently large @xmath75 , @xmath316 more precisely , for any @xmath264 , there exists a constant @xmath317 such that for all sufficiently large @xmath75 , @xmath318 these bounds are not the best possible , but they are sufficient for our purposes . in terms of the modified bessel function , they read @xmath319 and @xmath320 respectively .
this is the form in which they have been used in the last section .
we start from the uniform asymptotic approximation , which we now consider for @xmath321 only .
note that @xmath322 is a decreasing function of @xmath323 with @xmath324 for @xmath325 and @xmath326 for @xmath327 .
there exists a constant @xmath328 such that latexmath:[\[\begin{aligned } \label{jbound - tmp } for all @xmath321 for all sufficiently large @xmath75 .
we now use the bounds @xmath330 which follow from well - known asymptotic approximations for the airy function and its derivative ( see section 11.1 in olver @xcite ) . in particular , these bounds imply that @xmath331 and @xmath332 throughout the real line .
it therefore follows from that there exist constants @xmath333 such that @xmath334 for all @xmath321 for all sufficiently large @xmath75 . to deduce , use the first term inside the minimum and observe that since @xmath335 for @xmath336 , @xmath337 for @xmath338 , and @xmath339 is an analytic function of @xmath340 with @xmath341 ( see section 11.10 in olver @xcite ) , @xmath342 | we consider the sample covariance matrices of large data matrices which have complex matrix entries and which are non - square in the sense that the difference between the number of rows and the number of columns tends to infinity .
we show that the second - order correlation function of the characteristic polynomial of the sample covariance matrix is asymptotically given by the sine kernel in the bulk of the spectrum and by the airy kernel at the edge of the spectrum .
similar results are given for real sample covariance matrices . |
with the observation of zero resistivity in sr@xmath0ruo@xmath1 below @xmath5 , maeno _ et al .
_ @xcite discovered the first layered perovskite compound without cuo@xmath0 planes showing superconductivity .
recent experimental and theoretical research provides considerable evidence that the superconducting state of sr@xmath0ruo@xmath1 is of an odd - parity @xmath2-wave symmetry . classifying the superconducting states of a generalized bcs theory@xcite , the mean fields @xmath6 are split into singlet and triplet pairing states making use of the antisymmetry condition @xmath7 .
while singlet states are described through an even scalar function @xmath8 in the form @xmath9 , triplet states take the form @xmath10 where @xmath11 is an odd vector function@xcite . taking into account the crystal symmetry , the @xmath12-dependent functions @xmath8 and @xmath13 have to belong to an irreducible representation of the lattice point group , in our case the tetragonal point group @xmath14 .
initial proposals of odd parity superconductivity in sr@xmath0ruo@xmath1 were founded on the itinerant ferromagnetism of related ruthenate compounds and on the similarities between sr@xmath0ruo@xmath1 and @xmath15he in the fermi liquid corrections produced by electronic correlations@xcite .
recently , muon spin rotation ( @xmath4sr ) measurements@xcite have observed that a spontaneous magnetization begins to develop below @xmath16 .
this finding is most naturally explained in terms of a time reversal ( @xmath17 ) symmetry breaking state implying a multiple component representation of the @xmath14 point group .
furthermore , since the electronic structure of sr@xmath0ruo@xmath1 is quasi - two dimensional , a gap function with a strong @xmath18-dependence is unlikely . of the remaining two odd or even two - component representations of @xmath14 , the odd @xmath19 representation with basis functions
@xmath20 is the only without nodes at @xmath21 and therefore is most likely to be realized in sr@xmath0ruo@xmath1 . in ginzburg - landau ( gl )
theory , we deal with order parameters @xmath22 depending only on the center - of - mass coordinate @xmath23 , defined through @xmath24 as usual we switch the transformation behavior from the basis functions @xmath25 to the order parameter components @xmath26 . the components @xmath27 then share the rotation - inversion symmetry properties of ( @xmath28 ) , and the broken @xmath17-state is @xmath29 . based on this order parameter , one of us@xcite examined the vortex lattice structures of sr@xmath0ruo@xmath1 near @xmath30 : for an applied finite field @xmath31 along a high symmetry direction in the basal @xmath32-plane , two vortex lattice phases have been found with a second superconducting transition between them , while a square vortex lattice has minimal energy when the field is along the @xmath33-axis .
band structure calculations@xcite reveal that the density of states near the fermi surface is mainly due to the four ru 4@xmath34 electrons , which hybridize strongly with the o 2@xmath2 orbitals and give rise to three bands crossing the fermi surface , labeled @xmath35 , @xmath36 and @xmath37 ( see ref . ) . within the model of orbital dependent superconductivity @xcite
, the orientation of the square lattice mentioned above depends on which of the fermi surface sheets is responsible for the superconductivity .
recent experiments using sans@xcite show a square vortex lattice aligned with the crystal lattice in a wide range of the phase diagram , indicating that the @xmath38 sheet ( due to the @xmath39 - wannier functions ) of the fermi surface shows the superconducting transition@xcite . assuming that ferromagnetic spin fluctuations are responsible for the superconducting state , this finding is in agreement with @xmath40o nmr experiments@xcite showing strong ferromagnetic spin enhancements only in the ru 4@xmath41 orbitals . in this paper
we investigate the single vortex core structure and the geometry of the vortex lattice near the lower critical field @xmath3 .
we start from the ginzburg - landau theory for the two - component order parameter @xmath27 and solve numerically for the vortex core structure of an isolated vortex .
next , we derive an extended london model valid for fields near @xmath3 and large gl parameters @xmath42 and determine the geometric structure of the vortex lattice .
we find hexagonal lattices near @xmath3 , deforming into square lattices upon increasing the applied field .
taking the @xmath38 sheet as relevant for superconductivity , the orientation of the square lattices is in agreement with experiments@xcite .
the individual vortex cores show square deformations .
this vortex core structure is observable through scanning tunneling microscopy ( stm ) measurements since the in - plane coherence length of 900 is well within the spatial resolution required by stm . alternatively
, bitter decoration studies near @xmath3 can be used to detect the predicted field dependent transition between hexagonal and square lattice , which would complete the picture found by sans - measurements for higher applied fields@xcite .
a further characteristic of the hex - to - square transition is the peak splitting in the field distribution calculated in sec .
[ sec4 ] and measurable in @xmath4sr experiments .
note that a similar transition has been predicted in borocarbide @xcite and high-@xmath16 superconductors @xcite and has been observed recently in the borocarbide materials @xcite .
we start from the ginzburg - landau free - energy density ( @xmath43 ) for the two - dimensional representation @xmath44 of the tetragonal point group , with the ginzburg - landau coefficients determined within a weak - coupling approximation in the clean limit .
these two approximations seem reasonable for sr@xmath0ruo@xmath1 given the ratios@xcite @xmath45 and @xmath46 ( here @xmath16 denotes the critical temperature , @xmath47 the fermi energy , @xmath48 the mean free path and @xmath49 the coherence length ) . in usual dimensionless units ,
we have@xcite @xmath50\nonumber\\ & & -\frac{1+\nu}{2}\,[(d_y \eta_+)(d_y \eta_-)^*+(d_y \eta_-)(d_y \eta_+)^*]\nonumber\\ & & + \frac{1-\nu}{2\ , i}\,[(d_x \eta_+)(d_y \eta_-)^*+(d_y \eta_+)(d_x \eta_-)^*]\nonumber\\ & & -\frac{1-\nu}{2\ , i}\,[(d_x \eta_+)^*(d_y \eta_-)+(d_y \eta_+)^*(d_x \eta_-)],\end{aligned}\ ] ] where @xmath51 , @xmath52 , @xmath53 , @xmath43 is in units @xmath54 , lengths are in units of the penetration depth @xmath55 , @xmath56 is in units @xmath57 , @xmath58 , and @xmath59 .
the anisotropy @xmath60 is a parameter that measures the tetragonal distortion of the fermi surface ( @xmath61 denotes averaging over the fermi surface ; note @xmath62 and @xmath63 for a cylindrical geometry ) .
the free energy @xmath43 is invariant under a simultaneous rotation of @xmath64 around the @xmath33-axis and changing sign of @xmath65 .
this implies that all results derived below for @xmath66 can be transformed into the corresponding results for positive @xmath67 by a simple rotation , and we limit ourselves to the case @xmath66 .
further we deal only with external fields applied along the @xmath33-axis allowing us to omit all terms containing @xmath68-derivatives . at zero
applied field the two degenerate solutions @xmath69 and @xmath70 minimize the free energy @xmath43 . for nonzero
applied field , the degeneracy is lifted , and depending on the direction of the applied field only one component @xmath71 is stable . throughout this paper we will fix the dominant order parameter to be @xmath72 .
the asymptotic vorticity of the dominant @xmath72-component then determines the field direction along the positive or negative @xmath68-axis . in the spatially inhomogeneous situations considered below , the second component @xmath73 is driven through the mixed gradient terms [ gradient terms such as @xmath74 and decays far away from the defect .
note , that to each solution @xmath75 a symmetric counterpart with reversed field direction exists @xmath76 below , we first examine the asymptotic solutions of the gl - equations for @xmath63 .
second , we use the @xmath63-solution and carry out perturbation theory in @xmath65 to obtain the asymptotic solutions for @xmath77 ( we ignore the coupling to the vector potential in the latter case ) . in the cylindrical case ( @xmath63 )
the problem exhibits rotational symmetry . restricting to fields @xmath78 , the solution has the structure @xmath79 where @xmath80 are polar coordinates and @xmath81 denotes the phase winding of the topological defect driving the vortex in the main superconducting component .
the relative phase @xmath82 between @xmath72 and @xmath73 corresponds to an angular momentum difference @xmath83 and comes in through the specific form @xmath84 of the mixed gradient terms for @xmath63 .
the solutions for @xmath85 are the most stable ones .
the small @xmath86 expansion to cubic order takes the form @xmath87 where @xmath88 , @xmath89 , and @xmath90 are the first expansion coefficients of @xmath72 , @xmath73 , and @xmath91 , to be determined by matching this small @xmath86 expansion to the asymptotics at @xmath92 .
the latter reads @xmath93 where the @xmath94 stands for @xmath85 .
the asymptotic form of the gl - equations requires the three functions @xmath95 , @xmath96 , and @xmath97 to be of the form @xmath98 if we ignore the vector potential @xmath99 . for the general case ( @xmath100 ) the main asymptotics is an exponential decay @xmath101 which is a consequence of transverse screening . from fig .
[ expofig ] we see , that for reasonably large @xmath42 this exponent @xmath102 approaches 1 , which equals the usual penetration depth in dimensionless units .
we proceed now with the anisotropic case @xmath103 ( here we neglect the vector potential ) .
an expansion to first order in @xmath65 produces the following result for small @xmath86 @xmath104 and for large @xmath86 @xmath105 the additional coefficients @xmath106 and @xmath107 are the first expansion coefficients of the higher angular momentum states ( @xmath107 from @xmath108 and @xmath107 from @xmath109 ) , which arise for @xmath77 . comparing the condensation energy for the two cases @xmath110 we find that the @xmath111-case is the more stable : the admixed @xmath73 order parameter rises linearly in the center rather than cubic as is the case for @xmath112 , and the asymptotic phase turns of @xmath73 are the same in the center and for large @xmath86 ( which is also an argument against off - centered zeroes as seen for example in @xmath113-superconductors ) .
both points lead to a higher gain in condensation energy , which should stabilize the @xmath111-solution .
our numerical ginzburg - landau results ( described in the next paragraph ) for the critical fields @xmath114 confirm this finding : for @xmath115 and @xmath116 , we obtain @xmath117 demonstrating the stability of the @xmath111-solution compared to @xmath118 .
the corresponding @xmath119-wave result is @xmath120 . to obtain complete numerical solutions ,
we minimize the free energy density on a two - dimensional grid .
the minimization is carried out by relaxing an initial configuration @xmath121 , @xmath122 and @xmath123 iteratively .
indeed we find that the energetically favorable @xmath111-solution shows no off - centered zeroes in @xmath73 .
a contour plot of the absolute value of the dominant @xmath72 ( a ) and @xmath73-component ( b ) is displayed in fig .
[ eta - glfig ] .
the admixed @xmath73 solution shows a fourfold structure with a maximal amplitude along the axes .
in addition the vortex core center of the dominant order parameter is distorted with marked elongations along the diagonals of the crystal lattice for @xmath66 ( and analogously along the axes for @xmath67 ) .
this behavior is understood recalling the properties of the fermi surface . for @xmath66 ,
the density of states ( dos ) at the fermi surface is larger along the axes than along the diagonals .
a larger dos however implies also a larger effective mass , which reduces the coherence length and thus the size of the vortex .
finally , the distribution of the absolute value of the @xmath124-field ( see fig . [ a - glfig ] ) also shows clear square deformations aligned with the underlying crystal lattice . on the other hand , in the unstable @xmath118-case the admixed component @xmath73 exhibits a negative phase turn around the center , and consequently four off - centered zeroes with positive phase turn are needed to match up this phase @xmath125 at small @xmath86 to the asymptotic phase @xmath126 at large @xmath86
this pattern is qualitatively well known from analog gl - calculations for @xmath119-admixture in @xmath113 high-@xmath16 superconductors@xcite .
the dominant component @xmath72 and the @xmath124-field show similar features as in the stable case discussed above .
we wish to determine the form of the vortex lattice close to @xmath3 . instead of solving the ginzburg - landau free energy directly , here
we follow an idea recently developed in the context of @xmath113-wave superconductors with a weak admixing of an @xmath119-component@xcite .
we first integrate out the admixed component , leading to a one - component free energy density .
we assume @xmath127 and @xmath128 and use london approximation to obtain an effective free energy density which depends only on the field @xmath129 and which can be minimized with respect to the geometry of the vortex lattice .
though our results will be quantitatively inaccurate for the low-@xmath42 sr@xmath0ruo@xmath1 material , we still expect to obtain a qualitatively correct picture . the ginzburg - landau equation for @xmath73 to first order in @xmath130 reads @xmath131 within london theory the dominant component @xmath72 has modulus unity ( see eq .
( [ asf ] ) ) and the first three terms lead to the expression @xmath132 . solving formally for @xmath73 and substituting this expression into the free energy density , we obtain the effective one - component ginzburg - landau free energy @xmath133 with @xmath134 here we have consistently neglected all terms of order @xmath135 , @xmath136 , and higher .
following the usual scheme for london approximation we replace the order parameter @xmath72 and its derivative by the superfluid velocity @xmath137 @xmath138 keeping only terms up to second order in @xmath137 and up to first order in the anisotropy parameter @xmath65 , we obtain the london free energy in fourier representation @xmath139 \big).\end{aligned}\ ] ] here , all quadratic expressions in @xmath137 and @xmath129 have to be understood as @xmath140 ( the results for vorticity @xmath85 are the same on this level of calculation ; they would differ if one included higher orders in @xmath137 ) .
next , we eliminate @xmath137 .
we obtain an equation connecting @xmath129 and @xmath137 by variation of the london free energy eq .
( [ londfed ] ) with respect to the vector potential @xmath141 .
the corresponding euler - lagrange equation reads @xmath142 \big).\ ] ] we solve this equation for @xmath137 up to first order in @xmath65 and substitute the solution back into the london free energy density ( eq . ( [ londfed ] ) ) , to obtain @xmath143 b^2 \nonumber \\ &
\equiv & -\frac{1}{2 } + j(\bbox{q},\nu,\kappa)\ , b^2,\end{aligned}\ ] ] where we introduced the short notation @xmath144 the variation with respect to @xmath145 leads to the extended london equation @xmath146 ( in the limit @xmath147 we recover the well - known result @xmath148 ( no sources ) ) .
we generate the vortex lattice by introducing the corresponding source terms at the vortex core positions into the london equations . a convenient form for these source terms is@xcite @xmath150 where the sum leads over all positions @xmath151 of the vortex lattice . the magnetic field and the gibbs free energy for the vortex lattice read @xmath152 where the specific form of the vortex lattice is given by the reciprocal lattice vectors @xmath153
we parameterize the unit cell in real space through the angles @xmath154 and @xmath155 , the ratio of the lengths of the unit vectors @xmath156 and the average field @xmath157 ( see fig . [ unitcellfig ] ) , where the first three parameters describe the geometry and @xmath157 the size of the cell .
the lattice geometry depends on the applied field @xmath31 and is determined by minimizing the gibbs free energy @xmath158 given in eq .
( [ gibbs ] ) with respect to the unit cell parameters . writing the operator @xmath159 in the form
@xmath160 it is obvious that a change of sign in @xmath65 still corresponds to a rotation in real and reciprocal space by @xmath161 .
we thus restrict the discussion again to negative values of @xmath65 .
the resulting vortex lattices exhibit a centered rectangular unit cell with the main axes aligned along the diagonals of the underlying crystal . without loss of generality
we thus assume @xmath162 and @xmath163 in the remainder of the paper . in fig .
[ bmin - hfig ] we plot the angle @xmath155 minimizing the gibbs free energy for different values of @xmath65 versus the field @xmath31 ( fig .
[ bmin - hfig]a : @xmath164 , fig . [ bmin - hfig]b : @xmath165 ) . for large enough values of the anisotropy parameter @xmath65
, the vortex lattice evolves from hexagonal at low applied fields to square at high fields .
the lattices found in our calculations ( @xmath166 ) are always aligned with the underlying lattice ( see fig .
[ contfig ] ) . for decreasing @xmath42 , the crossover to a square lattice approaches to the lower critical field @xmath3 and occurs at lower anisotropies @xmath65 , an effect which is believed to be even stronger if one would include all nonlinearities into the theory .
these results are in good agreement with recent experimental data of riseman et al .
@xcite , who find an aligned square lattice in sr@xmath0ruo@xmath1 down to very small fields in the phase diagram .
apart form the vortex lattice symmetry , the shape of the individual vortices as well as the corresponding field distributions is of interest .
the former , because stm - measurements are able to determine the form of the individual vortices directly , and the latter since the field distribution is the typical result of @xmath4sr - measurements .
while the shape of the @xmath129-field around the vortex core is perfectly circular for @xmath167 over the whole range of the applied field @xmath31 , we find square deformations of the @xmath129-field for negative values of @xmath65 ( see fig.[contfig ] ) .
this is in agreement with the single vortex results obtained in the previous section , which also exhibits the corresponding deformation of the fields around the vortex core .
experiments using @xmath4sr can observe the field distribution @xmath168 in fig .
[ ph - hfig ] , characteristic results are shown for two cases @xmath169 ( a ) and @xmath170 ( b ) . while at @xmath167 the field distribution shows only one peak with a small shoulder towards the minimal field value , the situation for @xmath103 is quite different .
with increasing @xmath171 the @xmath63-peak splits into two due to the appearance of two nonequivalent classes of saddle points in the @xmath129-field . upon further increase of @xmath171 , the low field peak gradually vanishes and finally develops into a broad shoulder .
we have investigated the single vortex and vortex lattice structure of sr@xmath0ruo@xmath1 near @xmath3 , assuming a weak - coupling , clean limit description of the 2d odd - parity @xmath44 symmetry of the superconducting state in sr@xmath0ruo@xmath1 .
the stable single vortex solution ( denoted @xmath172 ) takes the asymptotic form @xmath173e^{-i \theta})$ ] .
a second low energy vortex solution has the asymptotic form @xmath174e^{i \theta})$ ] ( denoted @xmath175 ) .
solutions of the latter type appear to be the minimal ones for @xmath34-wave superconductors with an admixed @xmath119-wave component .
here , the symmetry allows for the competing @xmath172-solution which has a lower energy as it contains fewer nodes and decays slower near the vortex core .
the stable @xmath172-solution shows clear four - fold deformations due to the anisotropy of the density of states at the fermi surface .
current experimental results on the vortex lattice structure point towards an elongation of the vortex along the crystal lattice diagonals , in agreement with our theoretical analysis .
a direct proof of such a structure should be observable through stm measurements . the structure of the vortex lattice has been determined using an extended london approach where the admixed @xmath73 component has been integrated out .
the resulting london theory predicts that with increasing field the vortex lattice will continuously deform from a hexagonal to a square lattice .
the orientation of the square vortex lattice depends on the anisotropy parameter @xmath65 : for @xmath66 the square vortex lattice is aligned with the crystal lattice whereas for @xmath67 the orientation of the vortex lattice is rotated by @xmath64 .
comparing to recent experimental results of riseman _
et al._@xcite , who found the vortex and crystal lattices aligned in almost the entire field range , we conclude that @xmath66 .
the continuous hex - to - square transition should be observable through bitter decoration experiments .
an alternative signature of this transition is given through the splitting in the peak of the field distribution @xmath176 .
these field distributions should be observable in @xmath4sr experiments .
we thank g. blatter , e. m. forgan , g. luke , a. mackenzie , y. maeno , t. m. rice , and especially m. sigrist for helpful discussions .
we gratefully acknowledge support by the swiss nationalfonds and d. f. a. further acknowledges support from the natural science and engineering research council of canada .
t. m. riseman , p. g. kealy , e. m. forgan , a. p. mackenzie , l. m. galvin , a. w. tyler , s. l. lee , c. ager , d. mck .
paul , c. m. aegerter , r. cubitt , z. q. mao , s. akima , and y. maeno , nature ( london ) , in press . | based on a two dimensional odd - parity superconducting order parameter for sr@xmath0ruo@xmath1 with @xmath2-wave symmetry , we investigate the single vortex and vortex lattice structure of the mixed phase near @xmath3 .
ginzburg - landau calculations for a single vortex show a fourfold structure with an orientation depending on the microscopic fermi surface properties .
the corresponding extended london theory is developed to determine the vortex lattice structure and we find near @xmath3 a centered rectangular vortex lattice . as the field is increased from @xmath3 this lattice continuously deforms until a square vortex lattice is achieved . in the centered rectangular phase the field distribution , as measurable through @xmath4sr experiments , exhibits a characteristic two peak structure ( similar to that predicted in high temperature and borocarbide superconductors ) . |
quasars are luminous , accreting supermassive black holes ( smbhs ) that are thought to reside at the center of almost every massive galaxy ( e.g. , salpeter 1964 ; zeldovich & novikov 1964 ; lynden - bell 1969 ; kormendy & richstone 1995 ; richstone et al.1998 ) . despite their high luminosities and perhaps different bh fuelling mechanism ,
quasars are no different from their low - luminosity counterparts , active galactic nuclei ( agns ) : they both represent stages of building up smbhs , and they interact with the host galaxies in a self - regulated way ( e.g. , silk & rees 1998 ; magorrian et al .
1998 ; ferrarese & merritt 2000 ; gebhardt et al .
2000 ; kauffmann & haehnelt 2000 ; wyithe & loeb 2002 , 2003 ; king 2003 ; di matteo et al . 2005
; hopkins et al . 2006 ; croton et al .
2006 ) . therefore understanding the properties of quasars / agns , such as their abundance ( luminosity function ) , spatial distributions ( clustering ) , physical properties ( spectral energy distributions ; black hole masses , etc . ) and the dependence of these properties on redshift , is crucial to understand galaxy formation and smbh growth within the standard hierarchical structure formation framework . in this work we focus on the spatial clustering of quasars .
similar to galaxies , quasars can be used to trace the large - scale distribution of the underlying dark matter .
because of their high luminosity , quasars can be seen at redshift @xmath8 in a single survey , allowing them to be used to map the dark matter ( dm ) distribution over this entire redshift range . however , owing to the enormously large spatial volume probed and the fact that the quasar number density is much smaller than that of galaxies , the analysis of quasar clustering has only become practical in recent years , namely after large - scale surveys such as the 2df quasi - stellar object ( qso ) redshift survey ( 2qz , croom et al.2004 ) and the sloan digital sky survey ( sdss , york et al .
results on quasar clustering have been reported using the 2qz samples ( e.g. , porciani , magliocchetti & norberg 2004 ; croom et al .
2005 ; da ngela et al . 2005 ; porciani & norberg 2006 ) , the sdss samples ( e.g. , myers et al . 2006 , 2007a , 2007b ; shen et al .
2007a , 2008a ; ross et al .
2008 ) , and the 2df - sdss lrg and qso survey ( 2slaq , croom et al .
2008 ) sample ( e.g. , da ngela et al . 2008 ) .
complementary to these results , quasar clustering has also been studied through cross - correlations with large galaxy samples ( e.g. , adelberger & steidel 2005 ; coil et al .
2007 ; padmanabhan et al . 2008 ) , and surveys for close quasar pairs ( e.g. , hennawi et al . 2006 ; myers et al . 2008 ) . while the statistics of quasar clustering are not yet comparable to those of local galaxy clustering , recent measurements of the quasar two - point correlation function ( cf ) already offer valuable information about the physical properties of quasars , such as their host dark matter halo masses and duty cycles .
the amplitude of the quasar correlation function suggests that quasars live in massive dark matter halos , and therefore are biased tracers of the underlying dark matter ( e.g. , kaiser 1984 ; bardeen et al .
1986 ) ; the bias increases with increasing redshift ( e.g. , croom et al . 2005 ; myers et al . 2007a ; shen et al .
2007a ; da ngela et al .
2008 ; ross et al .
2008 ) . by comparing the relative abundance of quasars and host halos
, one can infer the average duty cycles of quasar activity ) , but the fraction of such companion quasars gives a negligible contribution to the quasar luminosity function as revealed in dedicated binary quasar surveys ( e.g. , hennawi et al .
2006 ; myers et al .
therefore no detailed halo occupation distribution modeling is needed to connect the halo abundance to the quasar luminosity function through the duty cycle . on the other hand
, this simple halo model assumes that halo mass is the only parameter that determines the clustering properties of dark matter halos .
recently it has been suggested that halo clustering also depends on the assembly history and recent merger activity ( e.g. , gao et al .
2005 ; wechsler et al .
2006 ; furlanetto & kamionkowski 2006 ; wetzel et al . 2007 ) ; such effects will complicate the interpretation of quasar clustering measurements ( e.g. , wyithe & loeb 2008 ) .
nevertheless we neglect these complications in the current work , since the magnitude of such effects is still quite uncertain at this moment . ]
( e.g. , cole & kaiser 1989 ; martini & weinberg 2001 ; haiman & hui 2001 ) . with substantial uncertainties ,
the estimated host dm halo mass is @xmath9 ( e.g. , porciani et al .
2004 ; croom et al . 2005 ; myers et al . 2006 , 2007a ; shen et al .
2007a ; da ngela et al .
2008 ; padmanabhan et al .
2008 ) , and the estimated quasar lifetime is between a few myr and @xmath10 yr .
quasar clustering measurements have also facilitated several theoretical investigations on the cosmic evolution of smbhs within the hierarchical structure formation paradigm ( e.g. , hopkins et al .
2007 ; shankar et al .
2008a ; white et al .
2008 ; croton 2008 ; shankar et al . 2008b ; wyithe & loeb 2008 ) .
previous investigations also examined the potential luminosity dependence of quasar clustering .
if quasar luminosity is a good indicator of the host halo mass , then more luminous quasars live in more massive halos , and hence should have stronger clustering .
however , the current observations only show weak luminosity dependence of quasar clustering at @xmath11 ( e.g. , adelberger & steidel 2005 ; croom et al .
2005 ; porciani & norberg 2006 ; myers et al .
2007a ; da ngela et al .
2008 ) , albeit with large error bars and , in most cases , limited dynamical range ( @xmath12 dex in luminosity ) .
if indeed quasar clustering depends only weakly on luminosity , then low luminosity quasars can also live in high mass halos and vice versa , which means there is substantial scatter in the host halo mass quasar luminosity relation .
this is naturally expected in some evolutionary quasar models , where quasar luminosity also depends on the evolving accretion rate onto the smbh ( e.g. , hopkins et al .
2005 ; hopkins et al . 2006 ; shen et al .
2007b ) . based on numerical simulations of quasar light curves and prescriptions for the relation between quasar peak luminosity and host halo mass , lidz et al .
( 2006 ) were able to show a weak luminosity dependence of quasar clustering at @xmath13 for intermediate luminosity ranges .
they argue that the peak luminosity is a better tracer of the host halo mass than the instantaneous luminosity , although in practice one can only observe the instantaneous luminosity . an alternative approach is to study quasar clustering as a function of black hole mass , which might better correlate with host halo mass than luminosity ( e.g. , ferrarese 2002 ; baes et al .
unfortunately for most quasars , the only estimates of their bh mass are virial bh masses ( e.g. , mclure & jarvis 2002 ; vestergaard & peterson 2006 ; shen et al .
2008c ) , which are based on broad line widths and continuum luminosities measured from single - epoch spectra .
substantial uncertainties and biases exist for these estimates ( e.g. , shen et al .
2008b ) , and any potential difference in clustering may be washed out by these complications ( see the discussion in [ subsec : bh_vir_vs_mhalo ] ) .
nevertheless , we will use these virial bh mass estimates to test any dependence of quasar clustering on bh mass , which in turn can offer some insight on the accuracy of these virial mass estimates .
there are also sub - classes of quasars that show characteristic properties different from the `` regular '' quasar population .
quasars have a distribution of colors at a given redshift . using a sample of 4576 uniformly - selected quasars from sdss , richards et al .
( 2003 ) found that at each redshift the @xmath14 color follows an approximate gaussian distribution with a red tail , and they proposed to use the deviation from the median @xmath14 color as a redshift - independent color indicator .
thus quasars can be divided into blue , red and reddened populations ( e.g. , richards et al .
2003 ) , where the reddened population is most likely dust - reddened in the quasar host galaxy ( e.g. , hopkins et al .
there are also significant differences ( at least in the ensemble average ) between the emission line properties for blue and red quasars ( e.g. , richards et al .
moreover , a small fraction ( @xmath15 ) of quasars are radio - loud , although they have similar optical properties to radio - quiet quasars . it will be interesting to study their clustering properties for these quasar sub - classes to see if their large - scale environment is responsible for their different properties .
the purpose of this paper is to investigate various dependences of quasar clustering using a homogeneous spectroscopic sample drawn from the sdss fifth data release ( dr5 , adelman - mccarthy et al .
in particular we focus on the dependences on luminosity , virial bh mass , quasar color and radio loudness .
the redshift evolution of quasar clustering / bias factor and redshift distortions at @xmath16 are reported in a companion paper ( ross et al .
2008 ) . in
[ sec : data ] we describe our sample .
we present clustering measurements in [ sec : cf ] and discuss our results in [ sec : diss ] .
we adopt a flat @xmath17cdm cosmology with cosmological parameters consistent with the three year wmap observations ( spergel et al .
2007 ) : @xmath18 , @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 .
comoving distance will be used throughout unless otherwise specified .
the sdss uses a dedicated 2.5-m wide - field telescope ( gunn et al.2006 ) with a drift - scan camera with 30 @xmath24 ccds ( gunn et al .
1998 ) to image the sky in five broad bands ( @xmath25 ; fukugita et al . 1996 ) .
the imaging data are taken on dark photometric nights of good seeing ( hogg et al.2001 ) , are calibrated photometrically ( smith et al .
2002 ; ivezi et al .
2004 ; tucker et al .
2006 ) and astrometrically ( pier et al.2003 ) , and object parameters are measured ( lupton et al .
2001 ; stoughton et al .
quasar candidates for spectroscopy are selected from the imaging data using their colors ( richards et al.2002 ) , and are arranged in spectroscopic plates ( blanton et al.2003 ) to be observed and confirmed with a pair of double spectrographs .
the sdss spectroscopy prohibits two targets closer than @xmath26 to be assigned spectroscopic fibers on one plate due to fiber collision . although serendipitous close pairs can be observed on overlap plates ( hennawi et al .
2006 ) , their completeness is poor in the sdss main quasar sample . fig . 1 in ross et al .
( 2008 ) shows the minimal projected comoving separations corresponding to the @xmath26 fiber collision scale as function of redshift .
these are the smallest scales that can be systematically probed by the sdss spectroscopic quasar sample at each redshift .
investigations on quasar clustering below these scales can only be achieved with other dedicated quasar samples ( e.g. , hennawi et al .
2006 ; myers et al .
2008 ) .
we take the dr5 spectroscopic quasar catalog ( schneider et al . 2007 ) as our parent sample from which we draw our clustering samples .
this catalog contains 77,429 bona fide quasars that have luminosities larger than @xmath27 ( using a slightly different cosmology in that paper ) and have at least one broad emission line ( @xmath28 ) or have interesting / complex absorption features .
about half of the quasars in this catalog were selected from a uniform algorithm ( as described in richards et al .
2002 ) implemented after dr1 ( abazajian et al .
2003 ) , which is flux limited to @xmath29 at @xmath30 and @xmath31 at @xmath32 .
there are a few ( @xmath33 ) @xmath34 quasars at @xmath30 which were selected by the high-@xmath35 ( @xmath36 ) branch of the targeting algorithm ( richards et al .
2002 ) ; we reject these objects when constructing the homogenous clustering sample .
our final clustering sample thus includes all uniformly - selected quasars with limiting magnitude @xmath29 at @xmath37 and @xmath31 at @xmath38 .
this homogeneous clustering sample includes 38,208 quasars covering the redshift range @xmath39 ; the angular coverage of this sample is identical to the high-@xmath35 ( @xmath38 ) quasar sample studied in shen et al .
( 2007a , the _
all _ sample ) , and covers a solid angle of @xmath40 deg@xmath41 .
we do not consider further refinements of the sample , such as using only _ good _ imaging fields as in richards et al .
( 2006 ) and shen et al .
( 2007a ) , because at low @xmath35 , quasars are found by uv excess , and are not as subject to subtle problems in the photometry as are high-@xmath35 quasars , and we generally found consistent results with and without including bad fields ( e.g. , see appendix c in ross et al .
although we may gain better control over systematics , such refinements will reduce the number of quasars in each subsample and therefore decrease the s / n for clustering measurements .
the entire clustering sample is shown in the left panel of fig .
[ fig : z_l_bh_div ] in the space of redshift and @xmath42band absolute magnitude ( @xmath43corrected to @xmath44 , see below and richards et al .
this homogeneous quasar sample is also used in the companion paper ( ross et al .
2008 ) , where it is referred to as the uniform sample .
we divide our sample into six redshift bins : @xmath45 $ ] , @xmath46 $ ] , @xmath47 $ ] , @xmath48 $ ] , @xmath49 $ ] , and @xmath50 $ ] , where we have deliberately avoided straddling two redshift deserts of quasars at @xmath51 and @xmath52 , caused by the inefficiency of quasar color selection at these redshifts due to the contamination from f and gk stars ( fan 1999 ; richards et al .
2002 ; see also fig .
1 in richards et al .
note that these redshift bins are different from those used in ross et al .
while the redshift evolution of quasar clustering is not the focus of this paper , we measure quasar clustering for these redshift bins independently from ross et al .
( 2008 ) as a consistency check . within each redshift bin
, we further divide the sample according to luminosity or virial black hole mass , where we take the virial mass estimates from shen et al .
( 2008b ) .
these redshift grids are shown in fig .
[ fig : z_l_bh_div ] for @xmath53 ( left ) and @xmath54 ( right ) .
we use the @xmath43corrected absolute @xmath42band magnitude @xmath55 normalized at @xmath44 ( e.g. , richards et al .
2006 ) as a luminosity indicator .
the conversion between this non - standard absolute magnitude and @xmath56 is given by equation ( 1 ) in richards et al .
for a universal power - law spectral energy distribution ( sed ) with @xmath57 we simply have @xmath58 .
there is also a tight correlation between @xmath55 and the bolometric luminosity @xmath59 estimated from the quasar spectrum , as shown in fig .
2a of shen et al . ( 2008b ) .
this correlation is described by is used in fitting @xmath60 .
if we do not fix @xmath61 then the best fitted parameters are @xmath62 , @xmath63 , which is fully consistent with a nominal slope @xmath64 between luminosity and magnitude . ] : @xmath65 where the gaussian scatter of the correlation is @xmath66 . while the @xmath53 and @xmath54 grids in fig .
[ fig : z_l_bh_div ] can be used to study the luminosity dependence and virial mass dependence at fixed redshift , the small number of objects in each bin will limit the quality of clustering measurements . we thus construct four other divisions using a larger redshift range @xmath67 to increase the signal - to - noise ratio , which are the following : \(a ) _ median luminosity ( bright and faint samples ) .
_ the sample is divided by the median luminosity at each redshift , which is used to test luminosity dependence .
in addition we also divide the sample into the top @xmath68 luminous quasars and the remaining less luminous quasars , since if these brightest quasars live in the rarest halos then their clustering will be much stronger than the less luminous ones because of the rapid increase of linear bias factor with halo mass ( e.g. , kaiser 1984 ; mo & white 1996 ) .
\(b ) _ median virial mass ( high and low mass samples ) . _
the sample is divided by the median virial bh mass at each redshift , which is used to test virial bh mass dependence . as in the luminosity case
, we also divide the sample into the top @xmath68 massive quasars and the remainder .
\(c ) _ color division ( reddened , red and blue samples ) .
_ the sample is divided according to _ relative _ quasar colors .
we use @xmath69 , i.e. , the deviation from the median @xmath14 color at each redshift for our sample values are re - calculated for our uniformly - selected sample , not those tabulated in the schneider et al .
( 2007 ) catalog . ] , as a color indicator ( described in detail in richards et al .
the reddened quasars are classified as those with @xmath70 from the median value @xmath71 , where @xmath72 is the standard deviation of the distribution of @xmath69 ; the red quasars are classified as those with @xmath73 ; the blue quasars are those with @xmath74 .
these subsamples are used to test color dependence .
\(d ) _ radio division ( radio and non - radio samples ) . _
the sample is divided into first - detected and undetected sources using the target selection information included in the dr5 quasar catalog .
the first - detected and undetected quasars within this redshift range @xmath67 are complete to @xmath29 in our uniform sample .
the first - detected sample contains quasars that are generally radio - loud with typical radio - loudness @xmath75 ( jiang et al .
these subsamples are used to test radio dependence . to simplify the discussion we use _ first - detected / undetected _ and _ radio - loud / quiet _ interchangeably in the following sections , but we acknowledge that they are two different concepts .
note that the flux limit of the first survey is bright enough that moderately radio - loud objects ( @xmath76 ) may be undetected .
since all the four divisions include quasars from a wide redshift range , it is crucial to ensure that the redshift distributions are similar for whichever subsamples are compared .
the bottom panels in figs .
[ fig : median_l_bh_div][fig : radio_div ] show that our choice of subdividing these samples yields almost identical redshift distributions for comparing subsamples .
the disadvantage of these divisions is that they intermix quasars of different luminosity , mass etc .
, at various redshifts .
however , if the dependences of clustering on these properties are monotonic and do not evolve significantly over the redshift range @xmath77 , we still expect to see differences in the averaged sense .
the simplest way to measure clustering is the redshift space correlation function @xmath78 where @xmath79 is the redshift space separation of quasar pairs . to minimize the effects of redshift distortions and redshift errors ,
the projected correlation function @xmath80 is often used ( e.g. , davis & peebles 1983 ) , where @xmath81 is the projected separation . in our clustering studies , both @xmath78 and @xmath80 will be presented .
to compute @xmath78 and @xmath80 , we use the landy - szalay estimator ( landy & szalay 1993 ) . for error estimation
we use jackknife resampling as described in shen et al .
( 2007a ) .
only the diagonal elements in the covariance matrix are used in the @xmath82 fitting because we found the covariance matrix is generally too noisy ; on the other hand , we find consistent results with or without including off - diagonal elements in the fitting ( shen et al . 2007a ; appendix b of ross et al . 2008 ) .
we refer to our previous paper ( shen et al .
2007a ) and ross et al .
( 2008 ) for more detailed descriptions on the estimators used , error estimations and generating random catalogs .
in some cases , the sample is too sparse to make useful auto - correlation measurements , so we cross - correlate it with a larger quasar sample matched in redshift to boost the clustering signal ( e.g. , coil et al . 2007 ; shen et al .
2008a ; padmanabhan et al .
2008 ) .
although our sample represents the largest spectroscopic quasar sample to date , we are limited by shot noise , because our sample is much more sparse than low redshift galaxy samples ( e.g. , zehavi et al . 2005 ; padmanabhan et al .
2007 ) , type 2 agn samples ( e.g. , constantin & vogeley 2006 ; li et al .
2006 ) , or photometric quasar samples ( e.g. , myers et al .
2006 , 2007a , 2007b ) .
the volume number density of the sdss main quasar sample is even lower than the number densities of the 2qz ( croom et al .
2004 ) and the 2slaq ( croom et al . 2008 ) spectroscopic samples studied by several authors ( e.g. , porciani et al . 2004 ; croom et al . 2005 ; da ngela et al . 2005 ; da ngela et al .
2008 ) , because sdss does not reach as far down the quasar luminosity function as the other surveys .
the sparseness of the sample directly impacts our strategy of measuring the correlation function and estimating its error .
in particular , only a handful of quasar pairs will be present for the smallest projected @xmath81 bins when we compute @xmath83 , which not only makes the results noisy but also puts the usage of @xmath83 into question .
binning effects may also become important for these sparse - pair bins .
for this reason , we develop a maximum - likelihood method in [ subsec : ml_cf ] , which we use to fit the smallest scales .
in addition to statistical issues with our sample , there might also be some systematics that are difficult to quantify . as already mentioned earlier , near redshifts @xmath51 and @xmath84 quasars are close to the stellar locus in color - color space , and the selection efficiency drops rapidly .
the number of observed quasars near these redshifts is reduced and pairs with similar colors are preferentially lost , hence the correlation function measurements suffer . both statistical fluctuations and systematic issues
can lead to negative correlations at certain scales . including or excluding these negative bins in the power - law model fitting
sometimes yield quite different results .
in addition , the jackknife errors we are using may underestimate the real errors for small number statistics .
for all these reasons , we caution that the actual uncertainties in our clustering measurements may be larger than our formal estimates .
in this section we use the traditional binned correlation function and power - law models to measure the clustering properties .
we use an alternative un - binned , maximum - likelihood approach in [ subsec : ml_cf ] for small to intermediate scales probed in our sample . before we continue , we want to clarify several issues with fitting a single power - law model of @xmath85 or @xmath86 to the raw cf data points .
first of all , the underlying cf is not a perfect power - law at all scales ; even if it were , because of the statistical fluctuations of cf data for sparse samples , the best - fit parameters of power - law models would depend on the fitting range .
hence the best - fit parameters should be quoted with fitting ranges . secondly ,
for most of our subsamples the data are of insufficient quality for a flexible power - law - index fit , hence we have fixed the power - law indices to be @xmath87 and @xmath88 , close to the observed values ( e.g. , porciani et al .
2004 ; croom et al . 2005
; shen et al .
thirdly , occasionally some cf data points are negative due to statistical fluctuations or unknown systematics ; including those negative points are necessary , but , because of these outliers and the nature of power - law models , the @xmath82 fit will have high model rejection probability .
hence we perform fits both with and without negative cf data points within the fitting range . for most of the cases the resulting correlation lengths
are consistent within @xmath89 errors , but there are exceptions . given the possibility that some of the negative points might be caused by systematic effects ( i.e. , close to the redshift deserts or at characteristic scales of preferentially losing pairs ) , we believe the actual clustering strength should lie somewhere in between these two limits .
quasars are biased tracers of the underlying dark matter .
the linear _ scale - independent _ bias factor is normally defined as @xmath90 where @xmath91 is the matter correlation function , extrapolated to high redshift according to the linear growth model .
we compute @xmath92 using a linear cdm power spectrum with transfer function from eisenstein & hu ( 1999 ) , and estimate @xmath61 using the integrated correlation function to marginalize nonlinear effects and scale - dependent biasmpc by @xmath93 , below the uncertainty from the clustering measurement itself ( @xmath94 ; see table 1 ) . ]
@xmath95 where we choose @xmath96 and @xmath97 ( note these choices are slightly different from ross et al . 2008 ) .
we use the best - fit power - law models to compute the integrated @xmath98 for quasars , and estimate the matter correlation function at the median redshift of the quasar sample . for cross - correlation functions with power - law models and fixed power - law index @xmath88
, we simply have @xmath99 where subscript `` cross '' refers to the cross - correlation and @xmath100 , @xmath101 refer to auto - correlations of sample 1 and 2 .
note that with equation ( [ eqn : cross_cf ] ) we have implicitly assumed that both samples 1 and 2 are perfectly correlated with each other , i.e. , they both trace the underlying dm distribution exactly .
lcccccccccc * all luminosities * & & & & & & & & & & + @xmath102 & 7902 & 0.50 & 45.57 & @xmath103 $ ] & @xmath104 & @xmath105 $ ] & @xmath106 & @xmath107 & @xmath108 $ ] & @xmath109 + & & & & & & & & & @xmath110 $ ] & @xmath111 + @xmath112 & 9975 & 1.13 & 46.27 & @xmath113 $ ] & @xmath114 & @xmath115 $ ] & @xmath116 & @xmath117 & @xmath108 $ ] & @xmath118 + & & & & & @xmath119 & & @xmath120 & @xmath121 & @xmath110 $ ] & @xmath122 + @xmath123 & 11304 & 1.68 & 46.60 & @xmath113 $ ] & @xmath124 & @xmath115 $ ] & @xmath125 & @xmath126 & @xmath108 $ ] & @xmath127 + & & & & & @xmath128 & & @xmath129 & @xmath130 & @xmath110 $ ] & @xmath131 + @xmath132 & 3828 & 2.18 & 46.86 & @xmath113 $ ] & @xmath133 & @xmath115 $ ] & @xmath134 & @xmath135 & @xmath108 $ ] & @xmath136 + & & & & & @xmath137 & & @xmath138 & @xmath139 & @xmath110 $ ] & @xmath140 + @xmath141 & 2693 & 3.17 & 46.73 & @xmath113 $ ] & @xmath142 & @xmath143 $ ] & @xmath144 & @xmath145 & @xmath108 $ ] & @xmath146 + & & & & & @xmath147 & & @xmath148 & @xmath149 & @xmath110 $ ] & @xmath150 + & & & & & & & & & @xmath151 $ ] & @xmath152 + @xmath153 & 1788 & 3.84 & 47.07 & @xmath143 $ ] & @xmath154 & @xmath143 $ ] & @xmath155 & @xmath156 & @xmath108 $ ] & @xmath157 + & & & & & @xmath158 & & @xmath159 & @xmath160 & @xmath110 $ ] & @xmath161 + + @xmath141 & & & & & & & & & & + @xmath162 & 1262 & 3.17 & 46.67 & & & & & & @xmath108 $ ] & @xmath163 + & & & & & & & & & @xmath110 $ ] & @xmath164 + & & & & & & & & & @xmath151 $ ] & @xmath165 + @xmath166 & 1416 & 3.17 & 46.91 & & & & & & @xmath108 $ ] & @xmath167 + & & & & & & & & & @xmath110 $ ] & @xmath168 + & & & & & & & & & @xmath151 $ ] & @xmath169 + + @xmath170 & & & & & & & & & & + bright & 15291 & 1.40 & 46.56 & @xmath171 $ ] & @xmath172 & @xmath173 $ ] & @xmath174 & @xmath175 & & + & & & & & @xmath176 & & @xmath177 & @xmath178 & & + faint & 15228 & 1.40 & 46.31 & @xmath171 $ ] & @xmath179 & @xmath173 $ ] & @xmath180 & @xmath181 & & + & & & & & @xmath182 & & & & & + @xmath68 most luminous - cross & 3050 & 1.40 & 46.84 & @xmath171 $ ] & @xmath183 & @xmath173 $ ] & @xmath184 & @xmath185 & & + & & & & & & & @xmath186 & @xmath187 & & + @xmath188 least luminous & 27469 & 1.40 & 46.39 & @xmath171 $ ] & @xmath189 & @xmath173 $ ] & @xmath190 & @xmath191 & & + @xmath68 most luminous - auto & 3050 & 1.40 & 46.84 & & & & @xmath192 & @xmath193 & & + & & & & & & & @xmath194 & @xmath195 & & + high - mass & 13414 & 1.35 & 46.51 & @xmath171 $ ] & @xmath196 & @xmath173 $ ] & @xmath197 & @xmath198 & & + & & & & & @xmath199 & & @xmath200 & @xmath201 & & + low - mass & 13443 & 1.35 & 46.37 & @xmath171 $ ] & @xmath202 & @xmath173 $ ] & @xmath203 & @xmath204 & & + & & & & & @xmath205 & & @xmath206 & @xmath207 & & + @xmath68 most massive - cross & 2698 & 1.35 & 46.64 & @xmath171 $ ] & @xmath208 & @xmath173 $ ] & @xmath209 & @xmath210 & & + & & & & & @xmath211 & & & & & + @xmath188 least massive & 24159 & 1.35 & 46.42 & @xmath171 $ ] & @xmath212 & @xmath173 $ ] & @xmath213 & @xmath214 & & + @xmath68 most massive - auto & 2698 & 1.35 & 46.64 & & & & @xmath215 & @xmath216 & & + reddened - cross & 2505 & 1.40 & 46.31 & @xmath171 $ ] & @xmath217 & @xmath173 $ ] & @xmath218 & @xmath219 & & + & & & & & @xmath220 & & @xmath221 & @xmath222 & & + red & 12697 & 1.40 & 46.41 & @xmath171 $ ] & @xmath223 & @xmath173 $ ] & @xmath224 & @xmath225 & & + & & & & & & & @xmath226 & @xmath227 & & + blue & 15317 & 1.40 & 46.43 & @xmath171 $ ] & @xmath228 & @xmath173 $ ] & @xmath229 & @xmath230 & & + first - cross & 2173 & 1.30 & 46.38 & @xmath171 $ ] & @xmath231 & @xmath173 $ ] & @xmath232 & @xmath233 & & + & & & & & @xmath234 & & @xmath235 & @xmath236 & & + non - first & 28346 & 1.40 & 46.42 & @xmath171 $ ] & @xmath237 & @xmath173 $ ] & @xmath238 & @xmath239 & & + & & & & & @xmath240 & & & & & + first - auto & 2173 & 1.30 & 46.38 & & & & @xmath241 & @xmath242 & & + & & & & & & & @xmath243 & @xmath244 & & + in the companion paper ( ross et al .
2008 ) we study the redshift evolution of quasar clustering in detail . here
we only provide brief discussion on this topic .
there have been several studies of the redshift evolution of quasar clustering , based on large survey samples such as 2qz , sdss , and 2slaq ( e.g. , croom et al .
2005 ; myers et al .
2006 , 2007a ; porciani & norberg 2006 ; shen et al . 2007a ; da ngela et al .
2008 ; ross et al .
2008 ) . while the redshift evolution at @xmath11 is still controversial to some extent ( cf .
porciani & norberg 2006 ; da ngela et al .
2008 ; ross et al .
2008 ) owing to the uncertainties in the clustering measurements for individual redshift bins , high redshift ( @xmath32 ) sdss quasars have much stronger clustering than their low redshift ( @xmath11 ) counterparts ( shen et al .
2007a ) , reflecting the large bias of their host halos .
we have performed independent clustering measurements using our redshift bins indicated in the left panel of fig . [
fig : z_l_bh_div ] , but with all the quasars in each redshift bin , i.e. , no luminosity cut .
the best - fit values of correlation length and linear bias are tabulated in table [ table : cf_result ] .
our results are in good agreement with those of ross et al .
we see a slight trend of increasing clustering with redshift at @xmath2 , but the uncertainties are large in each redshift bin . at each redshift
our sample spans a range in luminosity ( typically @xmath245 magnitudes ) .
hence the next question is whether or not quasar clustering depends on luminosity . on theoretical grounds ,
if quasar luminosity perfectly correlates with host halo mass , we expect that higher - luminosity quasars live in more massive halos and therefore are more strongly clustered
. weak dependence of quasar clustering on luminosity would imply that a quasar s luminosity is not a good tracer of its halo mass there is substantial scatter in @xmath246 and/or @xmath247 correlations ( e.g. , lidz et al .
observationally an important caveat is that the errors on the clustering measurements must be small enough that any difference caused by varying host halo mass is discernible .
no strong luminosity dependence has yet been reported ( e.g. , adelberger & steidel 2005 ; porciani & norberg 2006 ; myers et al .
2007a ; da ngela et al .
2008 ) ; but most of the studies are for @xmath11 and the error bars are large .
we attack this problem by first examining the @xmath67 samples divided by the median luminosity at each redshift ( [ subsec : sample ] and the top - left panel of fig .
[ fig : median_l_bh_div ] ) , which offers the best signal - to - noise ratio , although it mixes objects at different redshifts and different luminosities .
the correlation function for the _ bright _ and _ faint _ quasar samples are shown in the upper two panels in fig . [ fig : l_dep ] .
the projected correlation functions are consistent with one another given the errors .
we then compare the results for the @xmath68 most luminous quasars and the remainder , as shown in the bottom panels in fig .
[ fig : l_dep ] , where we use cross - correlation for the @xmath68 most luminous quasars with the remaining @xmath188 less luminous quasars .
we detect stronger clustering strength for the most luminous quasars at the @xmath248 level ( e.g. , see table [ table : cf_result ] ) .
this is somewhat expected since the halo bias factor increases rapidly with mass .
if these brightest quasars live in the rarest and most massive halos , their clustering will be appreciably stronger than the fainter quasars , an effect that is detectable even with our current sample size .
finally we compare the clustering in each luminosity - redshift bin as indicated in the left panel of fig .
[ fig : z_l_bh_div ] , since the luminosity dependence of clustering may evolve with redshift .
the results are shown in fig .
[ fig : lz_fine ] . given the large uncertainties , no appreciable luminosity dependence is observed in any of the redshift bins . the comparison of different luminosity samples becomes increasingly difficult at high redshift , and hence we defer the discussion of the luminosity dependence of clustering at @xmath38 to [ subsec : ml_cf ] . for the overall quasar population , one can also compare the clustering amplitudes between our sample and the 2qz/2slaq samples ( e.g , porciani et al .
2004 ; croom et al . 2005 ; porciani & norberg 2006 ; da ngela et al . 2008 ) or the sdss photometric quasar samples ( e.g. , myers et al . 2006 , 2007a ) .
for the same redshift ranges at @xmath11 , all samples have comparable clustering amplitude within the error bars .
the average luminosity is the highest for our sdss spectroscopic quasar sample and the lowest for the 2slaq sample ( differing by a factor of * @xmath249 * ) , which also indicates weak luminosity dependence of quasar clustering at @xmath11 for all but the highest luminosities .
we discuss the implications of this weak luminosity dependence at @xmath250 in [ sec : diss ] .
we have divided the sample in virial bh masses using the measurements from shen et al .
( 2008b ) .
quasars have a range of eddington ratios at fixed bh mass ( e.g. , kollmeier et al .
2006 ; shen et al .
2008b ) , thus the instantaneous quasar luminosity is not a good tracer of the bh mass .
if bh mass is a superior tracer of its host halo mass , we might expect a bh mass dependence of quasar clustering .
we test this possibility using virial bh masses measured from single - epoch quasar spectra .
however , we note that virial bh masses are _ not _ true bh masses ( see the extensive discussion in shen et al . 2008b ) . the scatter between virial and true bh masses may also dilute any potential clustering dependence , which we will return to in [ subsec : bh_vir_vs_mhalo ] . as in the luminosity dependence case , we show the comparison of our @xmath67 _ high mass _ and _ low mass _ samples in fig .
[ fig : bh_dep ] . while there is some indication that the _ high mass _ sample has a stronger clustering amplitude in redshift space , it is completely consistent with no difference in the projected correlation function ( see table [ table : cf_result ] for the best - fit @xmath251 values ) .
however , when the sample is divided into the top @xmath68 massive quasars and the remainder , the clustering is stronger for the quasars with highest virial bh masses , as shown in the bottom panels of fig .
[ fig : bh_dep ] .
the difference in the clustering strength for the cross - correlation of the @xmath68 most massive quasars and the auto - correlation of the remaining @xmath188 is significant at the @xmath248 level ( see table [ table : cf_result ] ) .
similarly we show the comparison between virial mass bins at each redshift bin ( as indicated in the right panel of fig .
[ fig : z_l_bh_div ] ) in fig .
[ fig : bhz_fine ] .
again no appreciable mass dependence is observed , although the s / n is poor in most bins .
quasars have different colors , either intrinsic or dust reddened . as discussed in
[ subsec : sample ] , we have followed the approach of richards et al .
( 2003 ) to divide our quasar sample by median @xmath14 color excess @xmath69 for @xmath67 , which removes the mean color at each redshift [ e.g. , division ( _ c _ ) in [ subsec : sample ] ; see fig . [
fig : color_div ] ] . the red and blue quasar samples are sufficiently large to perform auto - correlations , and we show the results in the upper two panels of fig . [
fig : color_dep ] .
the red and blue quasars are consistent with having similar clustering properties .
the reddened quasar sample is too sparse to do auto - correlation , so we cross - correlate the blue@xmath252red quasar sample with the reddened quasar sample , and show the results in the bottom two panels of fig . [ fig : color_dep ] .
the reddened quasars have similar clustering amplitude .
thus our data indicate there is weak or no dependence of quasar clustering on colors .
finally we investigate if there is any dependence of quasar clustering on radio activity .
using our @xmath67 first - detected and undetected samples as indicated in fig .
[ fig : radio_div ] , we show their clustering comparison in fig .
[ fig : radio_dep ] , where we use auto - correlation for the first - undetected quasars and cross - correlate them with the first - detected quasars .
the first - detected quasars are appreciably more strongly clustered ( cross correlation length @xmath253 ) than the first - undetected quasars ( @xmath254 ) , which indicates that radio quasars live in more massive dark matter halos .
these first - detected quasars also tend to have systematically larger virial bh masses than those first - undetected quasars by @xmath255 dex ( e.g. , shen et al .
however , the difference in their clustering remains when we consider a radio - undetected quasar sample with virial masses matched to those of the radio sample .
thus more massive host halos , and denser environments may be related to the triggering of radio activity . at the smallest separations ,
there is only a handful of pairs distributed among the bins , hence binning effects may become important , and the usage of the projected correlation function @xmath80 becomes questionable . here
we design a maximum - likelihood ( ml ) method to use in this situation .
poisson statistics are assumed to apply , hence we only focus on the smallest bins where the pairs are statistically independent , i.e. , the total number of pairs within each bin is much less than the total number of quasars in the sample .
we take a power - law model for the underlying real space correlation function @xmath256 .
then we compute the expected numbers of quasar - quasar and random - quasar pairs for a specified quasar sample , within a comoving cylindrical volume with projected radius @xmath257 to @xmath258 and half - height @xmath259mpc ; the usage of a cylindrical volume is to minimize the effects of redshift distortions and errors , and we found our results are insensitive to the value of @xmath260 . in mathematical form , the expected number of quasar pairs within the cylindrical annulus volume is : @xmath261dh\times n_{q}\equiv 2\pi rh(r)dr\ , \ ] ] where @xmath262 , @xmath263 is the number of quasars included in the sample under consideration , and @xmath264 is the number of random - quasar pairs within interval @xmath265 divided by the height @xmath266 , averaged over @xmath263 quasars . to compute
we have divided by a factor of 2 because we are calculating the number of pairs , not the number of companions . ]
@xmath267 we have used the number density of quasars of the sample under study as a function of redshift and the polygons which define the angular geometry of our sample ( see appendix b of shen et al .
2007a ; and hamilton & tegmark 2004 for the description of spherical polygons ) .
following croft et al .
( 1997 , also see stephens et al .
1997 ) we write the likelihood function as @xmath268 where @xmath269 , the expected number of pairs in the interval @xmath265 , and the index @xmath270 runs over all the elements @xmath265 in which there are no pairs . defining the usual quantity @xmath271
we have @xmath272\ , \ ] ] with all the model independent terms removed . here
the summation is over all @xmath273 quasar pairs we found in our sample , and @xmath274 $ ] is the range of scales over which we search for quasar pairs .
the values of @xmath275 and @xmath276 are chosen to avoid fiber collision effects and to ensure that the pairs found are statistically independent , i.e. , @xmath277mpc and generally @xmath278mpc .
but we find that different ranges of @xmath274 $ ] do yield different results , which do not always overlap within 1@xmath72 .
hence we examine different ranges of @xmath279 $ ] for each sample .
we then minimize @xmath280 with respect to @xmath281 and obtain its @xmath89 uncertainty .
again , because of the noisy data we fix @xmath282 in our fitting procedure . compared with binned cfs , the ml procedure has several drawbacks .
the information from pairs with large separations ( @xmath283 ) is not used , and the merits of the ml method suffer if there is a paucity of pairs due to systematic effects ( see below ) . for these reasons , we only apply this ml method and compare the ml estimates with our binned cf results as a consistency check ( see table [ table : cf_result ] ) .
we start with the two luminosity bins at @xmath141 shown in the left panel of fig .
[ fig : z_l_bh_div ] . in fig .
[ fig : highz_ldep ] we show the binned cfs for these two luminosity bins . on small scales ( @xmath284mpc ) , an apparent deficit of correlation is seen for the fainter bin with respect to the brighter bin . on larger scales ,
the signal - to - noise ratios of the cfs are too small to distinguish any difference in clustering strength . for the ml approach we choose a comoving volume of a cylinder annulus with half - height @xmath259mpc and @xmath274=[2,15]\ h^{-1}$]mpc , @xmath110\ h^{-1}$]mpc , and @xmath151\ h^{-1}$]mpc .
the best fit ml correlation lengths are listed in table [ table : cf_result ] , which show no detectable correlations at small scales ( @xmath284mpc ) for the fainter bin , consistent with the binned cf results .
the relative paucity of small - separation pairs in the fainter @xmath141 luminosity bin is much stronger than seen in the @xmath285 luminosity bins .
however , we suspect this is at least partly caused by some not well - understood systematic effects in our sample .
the lower redshift cut @xmath286 is close to the redshift desert at @xmath51 where the quasar selection efficiency drops .
the completeness becomes worse close to the magnitude cut @xmath31 at @xmath287 , as seen in our sample ( also see fig .
17 of richards et al .
as we argued above , image quality becomes important in selecting quasars close to the faint magnitude cut , and preferential loss of quasar pairs with similar colors is possible .
therefore we do not claim that the apparent weaker clustering of fainter quasars at @xmath141 and at scales @xmath288mpc is significant . on the other hand ,
our sample is unable to give clear evidence of a dependence of clustering on luminosity at large scales at @xmath289 because of the sample sparseness .
we hope that ongoing and future surveys of high - redshift quasars at the faint luminosity end can help resolve these issues .
we now apply the ml method to several subsamples in different redshift bins , and compare the results with those using binned method in table [ table : cf_result ] .
the ml method yields results consistent with the binned method within error bars .
in fact , the difference in the best - fit values of @xmath290 for various fitting ranges in the ml approach is fully reflected in the binned cf data points , i.e. , we get small values of @xmath290 for fitting ranges where the binned cf data points fall below the power - law fit .
we now discuss the dependences ( or lack thereof ) of quasar clustering on various quantities that we have found .
our clustering measurements of various subsamples are summarized in fig .
[ fig : bias_compile ] in terms of the bias factor .
the goal is to compare the observed clustering measurements with theoretical predictions ; or alternatively , to use these measurements to constrain theoretical models , i.e. , the form and scatter of various relations , such as relations of true bh mass versus host halo mass , luminosity versus true bh mass / host halo mass , and virial bh mass versus true bh mass , etc .
we assume that halo mass is the only relevant quantity that correlates with clustering strength .
hence the question in any theoretical model is how to populate halos with quasars of various physical properties ( luminosity , bh mass etc ) , given constraints from the halo abundance , the quasar luminosity function and quasar clustering . in this subsection
we outline the basic logic in modeling the luminosity dependence of quasar clustering .
suppose that at fixed host halo mass , the quasar luminosity follows a correlation given by : @xmath291 where @xmath292 and @xmath293 are constants , and @xmath294 is a gaussian random deviate . for simplicity
we have assumed here a power - law form with log - normal scatter , but more complex correlations are possible .
this correlation ( [ eqn : l_m ] ) gives the probability distribution of @xmath295 as , @xmath296 ^ 2}{2\sigma_l^2}\bigg\}\ .\ ] ] using bayes s theorem we can derive the probability distribution of host halo mass at fixed luminosity @xmath297 , given our knowledge of the halo mass function @xmath298 and the halo duty cycle @xmath299 , the probability of a halo hosting an active quasar : @xmath300 where the halo mass function @xmath298 is usually extracted from simulations .
generally there will be a malmquist - type bias because of the scatter @xmath294 and the bottom - heavy halo mass function .
but the exact magnitude of this bias as function of luminosity also depends on the halo duty cycle @xmath299 , which itself may well be a function of halo mass and redshift , and is constrained via the quasar luminosity function : @xmath301 therefore given the assumed correlation eqn .
( [ eqn : l_m ] ) , halo mass function , and @xmath299 constrained from the quasar luminosity function , we obtain the mass distribution of halos hosting quasars shining at any fixed luminosity @xmath302 via eqn .
( [ eqn : prob_m ] ) .
we also must know the linear halo bias factor @xmath303 as function of halo mass in order to make clustering predictions .
this is somewhat problematic , as noted in white et al .
various fitting functions for the halo bias exist in the literature , which differ in both magnitude and shape ( e.g. , mo & white 1996 ; jing 1998 ; sheth & tormen 1999 ; sheth et al .
2001 ; tinker et al .
2005 ; basilakos et al .
2008 ) . in practice
one needs a halo bias formula that is tested against large numerical simulation sets at the ranges of halo mass and redshift of interest ( e.g. , see the discussion in shankar et al .
finally the effective bias factor of quasars at fixed luminosity is the average over all the halos that host them : @xmath304 which is to be compared with observations .
the correlation in eqn .
( [ eqn : l_m ] ) is analogous to a light bulb model in which a quasar shines at a more or less fixed luminosity ( i.e. , a fraction of the eddington luminosity of its smbh ) when it is on , with a log - normal scatter @xmath294 in luminosity .
theoretical models predict a scaling @xmath305 or @xmath306 depending on energy conservation ( e.g. , silk & rees 1998 ; wyithe & loeb 2002 ) or momentum conservation ( e.g. , king 2003 ) during self - regulated black hole growth .
recent measurements of the @xmath307 relation ( e.g. , tremaine et al .
2002 ) favor the latter scaling , @xmath306 , as revealed in some hydrodynamic simulations ( e.g. , di matteo et al .
a more sophisticated , and perhaps more physically motivated , estimate for @xmath308 has been put forward by p. hopkins and collaborators , based upon simulated quasar light curves and correlations between quasar peak luminosity and halo mass ( e.g. , hopkins et al .
2005 ; lidz et al . 2006 ) .
simple phenomenological models , however , such as the one that uses eqn .
( [ eqn : l_m ] ) , may fit the observations equally well ( e.g. , shankar et al .
2008a ; croton 2008 ; shankar et al .
a full exploration of these models is beyond the scope of this paper . to gain some crude sense of how our measurements compare with theoretical models
, we use the diagnostic fig . 3 in lidz et al .
( 2006 ) , which shows predictions for the quasar bias as a function of luminosity at @xmath44 .
their models use the halo bias formula from sheth et al .
the dynamical range in the median luminosity of each bin probed by our sample is quite narrow , @xmath309 dex , compared to @xmath310 dex for the 2qz+2slaq sample used in da ngela et al .
( 2008 ) . for our bright and faint quasar samples as defined in
[ subsec : sample ] and in the upper - left panel of fig .
[ fig : median_l_bh_div ] , the median redshift is @xmath311 , and the median luminosities are @xmath312 and @xmath313 .
note that @xmath314 corresponds to _ rest - frame _ @xmath315band luminosity @xmath316 in lidz et al .
( 2006 ) . below and around this luminosity , even the light bulb model in their fig .
3 ( without any scatter @xmath72 ) gives weak luminosity dependence of quasar clustering .
this is consistent with our measurements , as well as da ngela et al .
( 2008 ) , who probe only a @xmath12 dex dynamical range in luminosity at redshift below @xmath317 .
it is also marginally consistent with the results at @xmath318 in adelberger & steidel ( 2005 ) , whose dynamical range in agn luminosity is @xmath317 dex . when the scatter @xmath294 in eqn .
( [ eqn : l_m ] ) is incorporated , the weighted bias factor at fixed luminosity will drop due to the scatter from more abundant less massive , and less biased halos , but it can be fine - tuned to yield an even flatter luminosity dependence if , for instance , the halo duty cycle @xmath319 decreases at an appropriate rate when @xmath320 decreases . one must also correct the overall linear bias of quasars by adjusting the normalization of the correlation between luminosity and halo mass ( eqn .
[ eqn : l_m ] ) , using constraints from the observed quasar luminosity function and the theoretical halo mass function , following the logical flow we just formulated above . at this stage , the majority of the measurements of the luminosity dependence of quasar clustering for moderate luminosities are consistent with theoretical predictions of weak dependence , but are unable to distinguish different models .
uncertainties in the quasar luminosity function , halo bias factor , and halo mass function , which are all needed to make a self - consistent model , further obscure the situation .
since the halo bias factor increases very rapidly with mass , various models are sensitive at the high - luminosity end of quasar clustering .
unfortunately , previous studies were based on samples containing too few bright quasars to perform correlation analyses ; cross - correlating those bright quasars with much larger galaxy samples is a promising approach ( e.g. , adelberger & steidel 2005 ; padmanabhan et al .
2008 ) , but statistically significant results are yet to come . we have shown for the first time that the most luminous quasars indeed are more strongly clustered than their fainter counterparts .
the median luminosity for our top @xmath68 most luminous quasar sample is @xmath321 , a factor of @xmath322 dex larger than the median luminosity for the remaining quasars . with this increment in luminosity
, we are probing the rarest and most biased halos . using the sheth et al .
( 2001 ) halo bias formula , we estimate a halo mass @xmath323 for the @xmath68 most luminous quasars and @xmath6 for the remainder .
we may also expect a correlation between host halo mass and black hole mass .
there have been claims for such a correlation at low redshift , where the circular velocity is used as a surrogate for halo mass and the central velocity dispersion is used as a surrogate for bh mass ( e.g. , ferrarese 2002 ; baes et al .
2003 ) , although significant uncertainty remains due to limited sample size and systematic effects ; for instance , it is argued that this correlation is dependent on hubble type ( e.g. , courteau et al . 2007 ; ho 2007 ) .
but for now let us neglect the systematic effects , and , analogous to eqn .
( [ eqn : l_m ] ) , write the correlation between central bh mass and halo mass as : @xmath324 where the constants @xmath292 and @xmath325 define the normalization and scaling of this correlation , and the random gaussian deviate @xmath326 denotes the scatter of this correlation . neglecting the scatter @xmath326 and taking @xmath327
as measured in ferrarese ( 2002 ) , one dex in @xmath328 corresponds to @xmath329 dex in @xmath320 .
the virial masses we measure for our quasar sample are scattered from the true masses by @xmath330 dex ( see discussion in shen et al .
2008b ) . to use eqn .
( [ eqn : bh_mhalo ] ) and its inverse relation we need a model to disentangle the intrinsic bh masses from the estimated virial bh masses .
shen et al . ( 2008b ) proposed a simple model of the intrinsic bh mass and eddington ratio distributions which simultaneously reproduces the quasar luminosity and virial bh mass distributions in different luminosity bins for sdss quasars . in this model ,
the intrinsic bh mass has a power law distribution , and the eddington ratio has a log - normal distribution at fixed intrinsic bh mass with a dispersion @xmath331 dex ( see 4.3 and 4.4 in shen et al .
2008b for details ) .
we follow the methodology outlined in 4.3 in that paper and use the model parameters for the mgii virial mass relation in their table 2 to simulate both true and virial bh masses and bolometric luminosities .
these parameters are found to be redshift - independent ( shen et al . 2008b ) .
we then use equation ( [ eqn : mi_lbol ] ) to simulate the @xmath42band absolute magnitudes and impose the sdss magnitude cut ( @xmath29 at @xmath67 ) until we achieve the same redshift distribution of observed virial bh masses as shown in fig .
[ fig : median_l_bh_div ] .
our simulated virial and true bh masses are shown in fig .
[ fig : sim_bh ] , where they are divided using the same definition of our high / low mass subsamples in [ subsec : sample ] .
the simulated virial bh masses reproduce the distribution of observed virial masses well .
the differences in the median values between these two subsamples are @xmath331 dex and @xmath332 dex for the virial and true bh masses respectively , i.e. , there is not much distinction between the true masses for the apparently high and low mass populations .
the dynamical range in the median @xmath333 is then only @xmath334 dex given our assumed simple model without the scatter @xmath326 .
take the typical host halo mass @xmath335 , the expected difference in bias for the _ low _ and _ high _ mass samples would be only @xmath93 , well below our @xmath89 uncertainties @xmath3 .
thus it is not surprising that we do not observe an appreciable difference in the clustering strength of the _ high _ and _ low _ mass quasar samples , given our current precision of the clustering measurements . as in the luminosity dependence case ,
when the scatter @xmath326 is included , the expected dependence could be even weaker .
our clustering measurements of reddened , red and blue quasar populations show consistent results , with differences that are undiscernible given our uncertainty levels .
a weak dependence on color is expected if the color of a quasar is determined by the physical conditions in the host galaxy of the smbh more than by its large - scale environment . also , there is no strong correlation of color with either luminosity or bh mass , indicating that this diversity in quasar color occurs at all mass scales that our sample probes .
these reddened quasars are most likely dust - reddened in the immediate nuclear environs and/or host galaxy of the quasar , as argued by richards et al .
( 2003 ) and hopkins et al .
( 2004 ) , and there is a strong connection between dust - reddening and broad absorption line quasars ( balqsos ; e.g. , reichard et al .
reassuringly , there seems to be no appreciable difference between the clustering strength of balqsos and regular quasars either ( shen et al .
2008a ) .
the most significant difference in clustering strength we have found is for the first - detected quasar sample .
the relative bias factor of radio - detected to radio - undetected quasars is @xmath336 , using the correlation length values in table 1 .
this implies that , on average , radio - detected quasars live in more massive halos than radio - undetected ones . using the formula in sheth et al .
( 2001 ) for the halo bias factor , the estimated host halo mass is @xmath337 for first - detected quasars , and @xmath338 for first - undetected quasars , at median redshift @xmath339 .
if one uses the jing ( 1998 ) fitting formula for the halo bias factor instead ( which gives the largest bias factor for fixed halo mass compared with others ) , the resulting halo masses are @xmath340 and @xmath341 for first - detected and undetected quasars respectively .
accurate determinations of the host halo masses for the radio - loud and radio - quiet populations again require better clustering measurements and robust halo bias factor from simulations .
the trend observed in this study is consistent with low redshift ( @xmath342 ) observations by mandelbaum et al .
( 2008 ) based on clustering measurements ( also see lin et al . 2008 and wake et al . 2008 ) and lensing maps of radio - loud type 2 agn .
both studies find that for comparable virial bh masses ( or galaxy properties ) , radio quasars / agns reside in more massive halos than their regular counterparts , indicating that radio quasars / agn may follow a different bh - halo relation from the normal population .
we have performed clustering measurements of quasars from a uniformly selected sample in the fifth data release of sdss , and focussed on the dependences of quasar clustering on luminosity , virial bh mass , quasar color and radio loudness . 1
. strong luminosity dependence of quasar clustering at @xmath2 is _ not _ detected for typical quasar luminosities .
while this is consistent with several theoretical predictions , we caution that the dynamical range in luminosity probed in our sample is narrow and our current sample size ( especially when split in luminosity / redshift ) is not yet large enough to yield accurate clustering measurements . however , the @xmath68 most luminous quasars in our sample are more strongly clustered , which is likely caused by the fact that halo bias increases rapidly with mass at the high mass end .
the final data release of sdss - ii ( dr7 , abazajian et al . 2008 ) will double the spectroscopic quasar sample size and boost the clustering signal by a factor of @xmath343 , thus providing superior results over the current study .
our measurements show weak or no dependence of quasar clustering on virial bh mass at @xmath2 .
this is at least partly due to the fact that virial masses are scattered around true bh masses , where only the latter are presumably correlated with host halo masses .
the limited dynamical range in bh mass and our limited sample size are other barriers to detecting this mass dependence .
however , even if we have accurate bh mass estimates , the dependence of quasar clustering on bh mass may still be weak if there is substantial scatter around the relation between bh mass and halo mass .
thus the combination of precise bh mass estimates ( i.e. , improved virial or new bh mass estimators ) and high - quality clustering measurements in future surveys can be used to test how smbhs grow in dark matter halos in the hierarchial structure formation framework .
3 . blue , red and reddened quasars have similar clustering strengths , at least indistinguishable at the current levels of measurement uncertainty and dynamical range probed by our sample .
radio - loud quasars are more strongly clustered than are radio - quiet quasars .
this implies that radio - loud quasars live in more massive dark matter halos and denser environments than radio - quiet quasars , consistent with local @xmath342 observations for radio - loud type 2 agn ( mandelbaum et al .
2008 ) and radio galaxies ( lin et al .
2008 ; wake et al .
radio quasars cluster more strongly than a regular quasar sample matched in virial bh mass , indicating that the bh - halo connection might be different for the radio population .
massive dark matter halos may provide the necessary hot medium needed for radio activity .
future quasar clustering analyses should aim at both broader dynamical ranges and larger sample sizes .
faint quasar detections are the focus of many ongoing or planned wide - angle survey projects such as boss ( schlegel et al .
2007 ) and pan - starrs ( kaiser et al . 2002 ) , which will both broaden the dynamical range and enlarge the sample size . because the halo bias factor increases rapidly at the high - mass end
, the clustering of the most luminous and most massive quasars will provide the best test of theoretical models ( e.g. , lidz et al .
2006 ; hopkins et al . 2007 ) . to obtain accurate measurements of quasar clustering at the high luminosity end
one needs a much larger control sample to cross correlate with them ; this control sample could be either a low luminosity quasar sample or a high redshift galaxy sample .
we thank the anonymous referee for helpful comments .
this work was partially supported by nsf grants ast-0707266 ( ys and mas ) and ast-0607634 ( npr , dvb , and dps ) .
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/. the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
luminosity bins .
_ left : _ redshift space correlation function .
_ right : _ projected correlation function . despite the absence of detectable correlation for the fainter sample ( triangles ) at @xmath288mpc , which may be caused by systematic effects ( see the text ) , there is no appreciable difference at large scales ( @xmath344mpc ) for the two samples , although the error bars are large.,title="fig:",scaledwidth=45.0% ] luminosity bins .
_ left : _ redshift space correlation function .
_ right : _ projected correlation function . despite the absence of detectable correlation for the fainter sample ( triangles ) at @xmath288mpc , which may be caused by systematic effects ( see the text ) , there is no appreciable difference at large scales ( @xmath344mpc ) for the two samples , although the error bars are large.,title="fig:",scaledwidth=45.0% ] | using a homogenous sample of @xmath0 quasars with a sky coverage of @xmath1 drawn from the sdss data release five quasar catalog , we study the dependence of quasar clustering on luminosity , virial black hole mass , quasar color , and radio loudness . at @xmath2 ,
quasar clustering depends weakly on luminosity and virial black hole mass , with typical uncertainty levels @xmath3 for the measured correlation lengths .
these weak dependences are consistent with models in which substantial scatter between quasar luminosity , virial black hole mass and the host dark matter halo mass has diluted any clustering difference , where halo mass is assumed to be the relevant quantity that best correlates with clustering strength .
however , the most luminous and most massive quasars are more strongly clustered ( at the @xmath4 level ) than the remainder of the sample , which we attribute to the rapid increase of the bias factor at the high - mass end of host halos .
we do not observe a strong dependence of clustering strength on quasar colors within our sample . on the other hand ,
radio - loud quasars are more strongly clustered than are radio - quiet quasars matched in redshift and optical luminosity ( or virial black hole mass ) , consistent with local observations of radio galaxies and radio - loud type 2 agn .
thus radio - loud quasars reside in more massive and denser environments in the biased halo clustering picture . using the sheth et al .
( 2001 ) formula for the linear halo bias , the estimated host halo mass for radio - loud quasars is @xmath5 , compared to @xmath6 for radio - quiet quasar hosts at @xmath7 . |
the challenge in facility location problems is to place @xmath0 service centers or facilities so that @xmath3 demand points are optimally served ( see for example ref . for an overview ) .
facilities can be hospitals , supermarkets , fire stations , libraries , warehouses , or any other supply centers providing vital resources to the population living at the demand points ( e.g. , households or cities ) .
here we consider the case where the demand points are at regular intervals along a one - dimensional geographic object , such as a road or a river , and where every demand point is a possible location for a facility .
the number of people @xmath4 who require the facilities services is assumed to be known at each demand point .
this number is typically very heterogeneous across geographic space .
depending on the context , there are different strategies for the placement of the facilities . in this article , we concentrate on the @xmath0-median problem , an important special case , where the objective is to minimize the average distance between a person s demand point and the nearest facility .
( a recent summary of the vast literature on the @xmath0-median problem can be found in ref . ) .
-median problem along a line .
the population @xmath4 is known at the demand points @xmath5 . in this article ,
the distance between neighboring demand points is assumed to be constant .
facilities will be placed on @xmath0 of these @xmath3 demand points .
( in the figure , @xmath6 . )
their locations @xmath7 are to be determined so that the average distance between a demand point and the nearest facility , weighted by @xmath4 , is minimized .
after the facilities have been located , the line can be divided into @xmath0 segments @xmath8 so that the @xmath9-th segment corresponds to the service region of the @xmath9-th facility .
, width=325 ] let us call the facility locations from left to right @xmath7 .
these positions are chosen among the demand points @xmath5 , which are equidistant ( i.e. , @xmath10 for @xmath11 ) along a line ( see fig . [ pmed_intro ] ) . if the population at @xmath12 is denoted by @xmath13 , the @xmath0-median problem consists of minimizing the cost function for a given location problem
is constant , we could in principle directly minimize the numerator in eq .
[ pmedian ] and ignore the denominator .
we have decided to keep the denominator so that @xmath14 equals the average distance .
@xmath14 can then be more easily compared across different location problems . ]
@xmath15 because only trips to the nearest facility play a role in eq .
[ pmedian ] , the line along which the demand points are located can be partitioned into @xmath0 segments or service regions .
demand points belong to the same segment if and only if they share the same closest facility , see fig .
[ pmed_intro ] .
the length of facility @xmath9 s service region is given by @xmath16 we will now take a closer look at the relation between @xmath17 and the population density around facility @xmath9 .
at first sight , it is plausible that the spatial density of facilities should follow the same trend as the population density : where there are more people there should be proportionately more facilities .
however , as we will see shortly , the @xmath0-median solution does not follow this rule that would give every facility an equal number of customers .
instead facilities are less abundant per capita in the high - demand regions than in the low - demand regions .
( gray histogram ) varies little between neighboring demand points , @xmath4 can be approximated by a continuous function @xmath18 ( black curve ) .
( b ) the function @xmath19 is defined as the length @xmath20 of the segment covering position @xmath21 .
strictly speaking , @xmath22 is a piecewise constant function .
however , if the spatial variations in @xmath4 are sufficiently small , @xmath22 can be approximated by a continuous function ( indicated by the dotted curve ) .
, width=325 ] for a spatially heterogeneous population distribution @xmath13 , it is difficult to deduce this general trend directly from eq . [ pmedian ] . with certain approximations , however , the problem becomes analytically tractable ; essentially , we translate the line of reasoning developed in ref . and for the two - dimensional @xmath0-median problem to the one - dimensional case .
first we define the population density @xmath23 which is the number of people per unit length in the vicinity of @xmath21 .
equation [ pmedian ] can be rewritten as @xmath24 where we have used the new notation to replace sums by integrals .
if we allow @xmath18 to be piecewise constant , this expression is still exact , but later it will be more convenient to approximate @xmath18 with a continuous function ( fig . [
cont_approx]a ) .
next we define @xmath19 to be the length of the segment serviced by the facility closest to @xmath21 ( see fig .
[ cont_approx]b ) .
the average distance from facility @xmath25 to a point @xmath21 inside its service region is equal to @xmath26 , where @xmath27 depends on the exact location of the facility .
for example , if @xmath28 is close to the center of the segment , @xmath29 . in the spirit of a mean - field approximation
, we will now assume that @xmath18 varies little over the size of a segment .
then we can replace the exact distance , @xmath30 , in the numerator of eq .
[ pmed_cont_approx ] with its average @xmath31 , @xmath32 the index @xmath25 was dropped in eq .
[ pmed_mf_approx ] assuming that most facilities will be close to the center of their service region so that @xmath27 is approximately constant .
unlike in eq .
[ pmedian ] , the locations @xmath28 no longer appear explicitly in eq .
[ pmed_mf_approx ] .
instead we have to find the function @xmath19 that minimizes @xmath14 subject to the constraint that there are @xmath0 facilities .
this constraint can be expressed as @xmath33 introducing a lagrange multiplier @xmath34 , the problem is equivalent to finding the zero of the functional derivative @xmath35 = 0,\ ] ] solved by @xmath36 the lagrange multiplier can be eliminated by inserting this expression into eq . [ constraint ] .
after some algebra , @xmath37^{-1/2}. \label{sqrt_scaling}\ ] ] the lengths of the service regions are thus inversely proportional to the square root of the population density .
the spatial density of facilities @xmath38 increases @xmath39 , but the per - capita density @xmath40 decreases @xmath41 with growing population .
the square - root scaling is a compromise providing most services where they are most needed , namely in the densely populated regions , but still leaving sufficient resources in sparsely populated regions where travel distances are longer .
this result implies an economy of scales : in crowded cities fewer facilities per capita can supply a larger population than in rural areas .
if facilities and demand points are not restricted to be along a line , but can be placed in two - dimensional space , the scaling exponent is @xmath1 instead of @xmath42 @xcite ( see section [ conclusion ] ) .
however , economies of scale are also predicted in two dimensions .
empirical studies have indeed reported this effect for certain classes of real facilities @xcite .
the calculation in the previous section assumes that the population density @xmath23 varies little within a service region . as we can see from eq .
[ sqrt_scaling ] , this implies that the segment length @xmath19 is also a smooth function ( fig .
[ cont_approx]b ) .
real census data , however , typically reveal strongly varying populations even on small spatial scales . in fig .
[ test_sets]a d , we show population numbers near three us interstate highways and the navigable mississippi river .
the data were generated from the us census of the year 2000 .
first , interstates 5 , 10 , 90 and the mississippi river were parameterized by arc length and markers were placed at regular 1-km intervals
. then census blocks within 10 km of the highways or the mississippi were identified and their population assigned to the nearest kilometer marker . as is clear from fig .
[ test_sets]a d , neither of the four populations is a smooth function . whether the assumptions behind eq .
[ sqrt_scaling ] are valid , is questionable , but it turns out that the scaling law for the service regions still holds with surprising accuracy . as a function of position @xmath21 along ( a ) interstate 5 , ( b ) 10 , ( c ) 90 , ( d ) the navigable part of the mississippi river .
the small squares below the @xmath21-axes indicate the optimal @xmath0-median positions of @xmath43 facilities .
( e ) map of the roads , the river , and the facility locations .
, width=325 ] versus the mean population @xmath44 in this service region .
lines indicate least - squares fits to eq .
[ log - log - scaling ] .
scaling is in good agreement with the analytic prediction @xmath45.,width=325 ] to compute the scaling exponent , @xmath46 facilities are placed on each of the four test data sets .
the optimal locations are calculated with the efficient algorithm of ref . .
their positions along the roads and the river in geographic space are shown in fig .
[ test_sets]e .
the segment lengths @xmath17 are calculated for each facility @xmath47 . in fig .
[ scaling_exact_opt ] , @xmath17 is plotted versus the mean value of @xmath4 inside the segment , denoted by @xmath48 . and @xmath49 , of the segment are half - integers , @xmath48 is defined as @xmath50 if @xmath51 or @xmath49 is an integer , @xmath52 or @xmath53 is added to the sum ( i.e. , half of the population is assigned to the facility on the right , half to the equally distant facility on the left ) . ]
ordinary least - squares fits of @xmath54 to the data yield slopes @xmath55 ( i-5 ) , @xmath56 ( i-10 ) , @xmath57 ( i-90 ) , and @xmath58 ( mississippi river ) , close to the prediction @xmath59 of eq .
[ sqrt_scaling ] .
the correlations are strong ; @xmath60 is consistently bigger than @xmath61 . assuming that the residuals are log - normally distributed , the predicted value @xmath62 is in all cases within the 95% confidence intervals .
thus , the equivalent of eq .
[ sqrt_scaling ] , @xmath63 , obtained by replacing the continuous variables @xmath18 and @xmath22 by their discrete counterparts @xmath48 and @xmath17 , is a good approximation .
this observation demonstrates that scaling at the exact @xmath0-median configuration is robust even in the presence of strong spatial fluctuations .
that the square - root scaling of the service regions is discernible even for realistically heterogeneous input , establishes a potential link to previous empirical work .
data collected in ref . suggest , at least for certain classes of facilities , a sublinear dependence of service facilities on population numbers .
it has been conjectured that the @xmath0-median model @xcite or a generalization thereof @xcite might explain this trend .
admittedly , we are looking in this article at a simplified linear geometry . yet
that sublinear scaling is robust even for substantially noisy input , might be viewed as supporting evidence for this conjecture . however , there is more to the problem than first meets the eye . although it is mathematically convenient to assume that facilities are placed to minimize an objective function such as eq .
[ pmedian ] , it is far from clear that the exact minimum will be achieved in reality .
decisions about facility locations are probably more haphazard in real life .
for example , site selections may be swayed by political interests , short - term fluctuations in property prices , or based on an incomplete knowledge of the actual demand .
even if the best effort is made to reach the global optimum , `` accidents of history '' may keep the facility locations trapped in a costlier local optimum .
it seems overly optimistic to draw conclusions about the scaling of real service regions only from the best of all solutions . the available literature for real facility distributions @xcite rather than the numerically optimal ones discussed in sec .
[ exact_numerical ] also justifies cautious skepticism , as some significant differences to the @xmath0-median result have been observed in reality , albeit in two dimensions .
how many facility configurations with costs near , but not necessarily equal to , the global minimum exist ?
there is no simple way to answer this question .
although the algorithm of ref . can find the global optimum very efficiently , it does not provide information about non - optimal solutions .
scanning all possible configurations is out of the question because their number is too vast . even for our smallest test data
set ( i-5 ) there are @xmath64 different ways to locate the facilities .
the situation is reminiscent of many - particle systems in physics where one wishes to calculate the large number of micro - states at a certain energy level out of an even larger number of all conceivable micro - states . in that context
, statistical mechanics has developed many powerful numerical tools .
we will build on this analogy in order to estimate the number of non - optimal facility locations .
let us call @xmath65 the number of facility locations with costs between @xmath14 and @xmath66 .
the function @xmath67 plays the role of the `` density of states '' in statistical mechanics .
as we will see , @xmath68 increases very rapidly as @xmath14 exceeds the minimum @xmath69 , so that it will be more convenient to work with its logarithm , the entropy @xmath70 .
our aim is to calculate @xmath71 with monte carlo simulations .
several methods exist @xcite ; here we apply the wang - landau algorithm @xcite .
first , the range of possible costs is divided into small discrete intervals of length @xmath72 .
then a random walk through the set of facility locations is performed and we count , in the form of a histogram , how often each interval is visited . the main idea behind the wang - landau algorithm is to bias the random walk in such a manner that all intervals are visited equally often .
for such a `` flat histogram '' we obtain equally good statistics for all intervals , an advantage when @xmath73 is the basis of further calculations .
we describe details of our implementation in app .
[ wang_landau ] .
( i.e. , the logarithm of the density of states ) versus the cost @xmath14 .
the inset shows the same four curves as the main panel , but with rescaled abscissa @xmath74 where @xmath3 is the number of demand points.,width=325 ] from calculations for four different empirical population distributions ( fig .
[ entropy ] ) it is clear that @xmath71 is singular at @xmath69 , the smallest possible cost .
thus , @xmath71 increases enormously in the vicinity of @xmath69 and the density of states @xmath75 grows even more rapidly .
the results for four different empirical population distributions suggest that @xmath71 follows approximately the same curve ( inset of fig .
[ entropy ] ) if regarded as a function of @xmath74 , where @xmath3 is the total number of demand points @xmath76 .
therefore , it appears to be a universal feature that for all realistic populations a large number of different possible configurations must be considered if the assumption of optimality is relaxed .
this observation raises the question : can the scaling relation of eq .
[ sqrt_scaling ] still be observed if facility locations are not exactly optimal , but are among the numerous configurations achieving almost but not exactly @xmath69 ?
if we randomly select a facility configuration with a cost in the interval @xmath77 $ ] , we can formally obtain the scaling exponent @xmath78 from eq . [ sqrt_scaling ] as follows .
first , we log - transform the segment lengths @xmath17 and the population density @xmath48 .
then a least - squares linear fit to eq .
[ log - log - scaling ] will be performed to calculate @xmath78 .
this procedure can be coupled with the wang - landau algorithm so that , at every step in the random walk through configuration space , we compute @xmath78 , the cost @xmath14 , and at the end of the algorithm the mean value @xmath79 as a function of @xmath14 . ,
( b ) the coefficient of determination @xmath60 as a function of the cost @xmath14.,width=325 ] the results , shown in fig .
[ exp_rsq]a , indicate that @xmath79 is approximately @xmath62 at the minimum cost @xmath69 for all four numerical test sets , as anticipated by our earlier calculations . as the cost increases
, @xmath78 also increases , indicating a decreasing dependence of the segment lengths on the population .
this behavior makes sense because the facility locations become more random as we move away from the optimum .
interestingly , the overall trend how @xmath79 increases with @xmath14 is similar in all four cases . in particular ,
the behavior near the minimum is noteworthy because @xmath79 increases most rapidly near @xmath69 . in other words , the analytic prediction at @xmath69 which provides us with the only easily calculable reference point for @xmath78 is unfortunately at the point where small deviations can also cause the greatest changes in @xmath78 . together with the least - squares exponent @xmath78
, we can also obtain other statistical measures from linear regression , such as the coefficient of determination @xmath60 ( fig .
[ exp_rsq]b ) .
it can take values between @xmath80 and @xmath81 ; the higher its value , the stronger the correlation between @xmath20 and @xmath44 . in our numerical test sets , @xmath60 takes its highest value ( @xmath82 ) at @xmath69 and decreases as we move toward higher costs following a slightly sigmoidal curve toward values around zero . at very large costs , @xmath60 increases again because the solution is effectively an `` obnoxious facility '' location where facilities are in sparsely populated regions so that @xmath20 and @xmath44 are positively ( instead of negatively ) correlated . for costs near @xmath69 , however , an increase in @xmath79 is coupled with a reduction in @xmath60 .
in this article , we have studied the one - dimensional @xmath0-median problem . in one dimension
, the exact optimum can be calculated numerically in polynomial time ; in two dimensions @xcite and on arbitrary graphs @xcite , the @xmath0-median problem is np - complete so that no polynomial - time algorithm is currently known .
previous empirical studies of scaling in real facility locations have usually dealt with two - dimensional densities .
the approximate analytic result in one dimension , @xmath83 ( eq . [ sqrt_scaling ] ) , can be easily generalized for arbitrary dimension @xmath84 , where the size of a @xmath84-dimensional voronoi cell @xmath22 is predicted to scale as @xmath85 .
the scaling of the facility density @xmath38 with the population density @xmath18 thus remains sublinear in all dimensions .
numerical optimization in two dimensions , based on us census data , yields indeed an exponent in excellent agreement with the predicted exponent @xmath86 @xcite . in 1977
, stephan implicitly proposed that the @xmath0-median model might explain empirical scaling relations between the area and population density of subnational administrative units ( e.g. , states , provinces , counties ) @xcite .
although he later generalized the objective function as more data became available @xcite , the notion that facilities may self - organize towards sublinear scaling has remained attractive , as proved by the recent rediscovery of stephan s model by um et al .
@xcite .
however , as the work shown here underlines , one has to be careful when interpreting empirical data .
increased spatial noise in the facility distribution can lead to different exponents and reduced correlations .
the situation investigated here portrays only one special scenario how randomness might be present , namely as a uniform probability distribution over all costs in an interval @xmath77 $ ] .
it is also conceivable that not all configurations within this range are equally likely , so that the best - fit exponents may behave differently .
we may also replace the @xmath0-median model by a different optimization principle ( e.g. , competitive facility location such as the hotelling model @xcite ) which can change the exponent at the optimum .
however , we believe that a steep increase in the number of possible configurations is a generic tendency of most models that relax the constraint of strict optimization even to a small degree .
the author thanks m. e. moses , s. banerjee , b. blasius , and h. youn for stimulating discussions .
the author acknowledges support from imperial college .
the wang - landau algorithm @xcite is designed to calculate the density of states @xmath67 for @xmath14 in some interval @xmath87 $ ] .
first the interval is divided into small sub - intervals of length @xmath72 .
the key element of the wang - landau algorithm is to visit every sub - interval @xmath88 $ ] with a probability @xmath89^{-1}$ ] .
initially , the density of states is of course unknown this is why we need the algorithm in the first place but we will recursively obtain better estimates for @xmath68 as the calculation proceeds . at the beginning we set @xmath90 for all intervals
@xmath88 $ ] .
simultaneously we maintain a histogram @xmath91 , which counts how often a cost between @xmath14 and @xmath92 is encountered during the course of a random walk . at the beginning @xmath93 for all @xmath14 .
the random walk through the set of facility locations proceeds as follows .
starting from an arbitrary initial configuration @xmath94 , a new set of facility positions @xmath95 is generated with probability @xmath96 .
in addition , a uniform random number @xmath97 $ ] is generated .
if @xmath98 , the current value of @xmath99 is multiplied by a constant factor @xmath100 , @xmath91 is incremented by 1 , and @xmath101 becomes the next step in the random walk .
otherwise , the move is rejected , and we increment @xmath67 and @xmath91 instead of @xmath99 and @xmath102 . following wang s and landau s original paper , ref . , we initially set @xmath100 equal to the euler number @xmath103 .
when the histogram @xmath91 is sufficiently `` flat '' , @xmath100 is replaced by its square root ( i.e. , @xmath104 ) . for practical purposes ,
the histogram is treated as flat if the maximum number of visits recorded by @xmath91 is less than @xmath105 more than the minimum .
if this condition is satisfied , all @xmath91 are reset to 0 , and the procedure is iterated until @xmath106 from an intermediate set of facility locations @xmath107 , we generate the new set @xmath101 by shifting one random facility one step to the left or to the right with equal probability .
exceptions are made if the facility is already at one of the edges of the line or adjacent to another facility .
let us define @xmath108 to be the number of facilities on the edges ( @xmath109 and @xmath110 ) plus twice the number of facility pairs occupying neighboring demand points .
then the non - zero step probabilities are given by @xmath111 = 1/(2p-\nu)\;\text{if $ r_1\neq q_1$},\\ & p[(r_1 , \ldots , r_{i-1 } , r_i , r_{i+1 } , \ldots , r_p ) \rightarrow ( r_1 , \ldots , r_{i-1 } , r_i-1 , r_{i+1 } , \ldots , r_p ) ] = 1/(2p-\nu)\;\text{if $ r_i-1\neq r_{i-1}$ , $ i=2,\ldots , p$},\\ & p[(r_1 , \ldots , r_{i-1 } , r_i , r_{i+1 } , \ldots , r_p ) \rightarrow ( r_1 , \ldots , r_{i-1 } , r_i+1 , r_{i+1 } , \ldots , r_p ) ] = 1/(2p-\nu)\;\text{if $ r_i+1\neq r_{i+1}$ , $ i=1,\ldots , p-1$},\\ & p[(r_1 , \ldots , r_{p-1 } , r_p ) \rightarrow ( r_1 , \ldots , r_{p-1 } , r_p+1 ) ] = 1/(2p-\nu)\;\text{if $ r_p\neq q_n$}. \end{aligned}\ ] ] in principle , we are able to explore all costs between the globally minimal @xmath69 and maximal @xmath112 . in practice , we have to reduce the search interval . on one hand , @xmath112 is orders of magnitude larger than @xmath69 and we are interested only in @xmath68 near @xmath69 . on the other hand , @xmath68 increases so quickly that , close to @xmath69 , the random walk is extremely unlikely to propose a step decreasing the cost .
therefore , we confine the random walk to intervals @xmath87 $ ] which become smaller as @xmath113 approaches @xmath69 .
we interpolate between all estimates of @xmath68 , which all differ from the real @xmath68 by a multiplicative constant , with a straightforward least - squares algorithm to obtain a single curve for the entropy @xmath114 over all measured values of @xmath14 .
there is exactly one constant left to be fixed because the wang - landau algorithm can calculate the entropy only up to an additive constant .
we adopt the normalization that @xmath115 at the extrapolated maximum . | the @xmath0-median problem is a common model for optimal facility location .
the task is to place @xmath0 facilities ( e.g. , warehouses or schools ) in a heterogeneously populated space such that the average distance from a person s home to the nearest facility is minimized .
here we study the special case where the population lives along a line ( e.g. , a road or a river ) .
if facilities are optimally placed , the length of the line segment served by a facility is inversely proportional to the square root of the population density .
this scaling law is derived analytically and confirmed for concrete numerical examples of three us interstate highways and the mississippi river . if facility locations are permitted to deviate from the optimum , the number of possible solutions increases dramatically .
using monte carlo simulations , we compute how scaling is affected by an increase in the average distance to the nearest facility .
we find that the scaling exponents change and are most sensitive near the optimum facility distribution .
quantitative studies in many branches of science frequently reveal scaling laws where two sets of observables are related by a power law over several orders of magnitude .
examples range from astronomy ( e.g. , kepler s third law ) to biology where , for example , kleiber s law states that the metabolic rates of mammals scale approximately as the three - quarter power of their body mass @xcite . here we look at a problem from economic geography , the relationship between the spatial distribution of a population and the distribution of service establishments ( e.g. , post offices or gas stations ) .
physicists typically enjoy the luxury of measuring scaling exponents in carefully designed and repeatable experiments . in biology and the social sciences , by contrast , the exact circumstances of an experiment are generally more difficult to control and to repeat . as a consequence
, power - law exponents are frequently obfuscated by noise in the measurement and in the process generating the scaling law itself .
the remaining uncertainty can lead to heated debates if , for example , the scaling exponent in kleiber s law is not truly @xmath1 instead of @xmath2 @xcite .
the available geographic data for the distribution of service establishments leave similar room for interpretation so that various scaling laws have been proposed @xcite .
facing such controversies , theorists often try to calculate the `` correct '' exponent from deterministic models .
one recurring idea is that scaling should emerge naturally from some appropriate model if an objective function ( energy dissipation @xcite , earnings @xcite , travel distance @xcite , etc . )
is optimized .
this approach has led to elegant theories , but it leaves one key problem unaddressed . knowing that evolutionary biology , human decisions , or other processes shaping the available empirical data are intrinsically stochastic , there is in principle a huge variety of outcomes .
how many different solutions are conceivable ?
how close to optimal does the observed solution need to be in order to exhibit the theoretically predicted scaling exponent ? here we study a model which serves as an example of computational techniques suited to address these questions .
the model is the @xmath0-median problem of optimal facility location along a strongly heterogeneously populated line ( e.g. , a transcontinental highway ) .
the task is to place @xmath0 facilities along the line and find the configuration that minimizes an objective function , in this case the average distance to the nearest facility @xcite . ignoring small - scale heterogeneity in the population ,
an analytic calculation predicts a simple scaling law for the length of the line segments served by different facilities .
the exact optimum locations can be computed numerically for realistic input data and are in good agreement with the analytic prediction . using techniques from statistical physics
, we calculate the number of possible facility locations for non - minimal costs .
with monte carlo simulations we will then quantify how deviations from the optimum make it less likely to find the theoretical exponent . |
the discovery of a standard model ( sm)-like higgs boson at the cern large hadron collider ( lhc ) @xcite provides proof of the existence of weakly coupled scalar particles in nature .
extensions of the scalar sector of the sm are thus of great interest , particularly when they could be connected to the solution of other mysteries of the sm , such as the nature of dark matter ( dm ) in the universe @xcite .
extensions of the sm higgs sector involving an additional `` inert '' scalar multiplet , the lightest state of which is stable and hence a possible dark matter candidate , have been well - studied in the singlet @xcite , doublet @xcite , and triplet @xcite cases ( for recent summaries of the experimental status of these models , see , e.g. , refs .
@xcite , @xcite and @xcite ) .
more recently , multiplets from larger representations of @xmath10 have been investigated in the context of dark matter @xcite . in this paper , we expand upon the study in ref .
@xcite , which focuses on models in which the sm is extended by a scalar electroweak multiplet with half - odd - integer weak isospin , @xmath11 , where @xmath12 is an even integer that counts the number of complex fields in the multiplet .
this scalar multiplet has weak hypercharge such that @xmath13 . ]
@xmath14 ( the same as that of the sm higgs doublet ) , and transforms as odd under an imposed global @xmath3 symmetry , allowing for the lightest member to be a stable dark matter candidate . in particular , we focus on the models with weak isospin @xmath15 ( sextet , @xmath16 ) and @xmath17 ( octet , @xmath18 ) .
larger complex scalar multiplets are disallowed by perturbative unitarity considerations involving the scattering of two scalars into two gauge bosons @xcite .
smaller multiplets , such as the @xmath19 case , have previously been studied in the literature @xcite . in these models
, the large number of new scalar particles leads to a plethora of electroweak pair - production processes accessible at the lhc via an @xmath20-channel photon , @xmath6 , or @xmath7 .
because of the @xmath3 symmetry , these scalars can decay only to lighter scalars within the large multiplet via the emission of @xmath6 or @xmath7 bosons , terminating at the lightest state @xmath5 of the large multiplet , which is stable and escapes the detector .
these production and decay modes produce similar phenomenology to supersymmetric electroweak gauginos , and the most promising searches to constrain these models are similarly those with a combination of missing transverse energy ( @xmath21 ) and leptons .
we avoid signatures with too many high - energy jets in order to avoid large qcd backgrounds , such as @xmath22 .
in particular , we have recast a number of existing searches for supersymmetric electroweak gauginos by both the atlas @xcite and cms @xcite collaborations which use data taken at a centre - of - mass energy of 8 tev ( lhc8 ) .
this paper is organized as follows . in sec .
[ sec : models ] we introduce the two large - multiplet models first studied in ref .
we re - express the parameter freedom in the models in terms of physical masses and mass splittings , which are most physically relevant for the kinematics of pair production and decays .
we also translate the theoretical and indirect experimental constraints on the models previously studied in ref .
@xcite into this new parameter space . in sec .
[ sec : decays ] we discuss the pattern of decay branching fractions of each of the scalars . in sec .
[ sec : collider ] we recast a collection of lhc8 searches for supersymmetric particles to constrain our models .
we conclude in sec .
[ sec : conclusions ] .
some mathematical details and simulation issues are relegated to the appendices .
we consider two models that extend the sm through the addition of a single , large electroweak multiplet of complex scalars , @xmath0 , which has quantum numbers under @xmath23 of @xmath24 , where @xmath25 or @xmath26 is the size of the multiplet . in these models , the most general gauge - invariant scalar potential that preserves a @xmath3 symmetry under which @xmath4 is given by @xmath27 + \mathcal{o}({\sigma}^4)\ ; , \label{eq : conjugatepotential}\end{aligned}\ ] ] where @xmath28 is the sm @xmath10 doublet . here
@xmath29 and @xmath30 are the higgs doublet and the large scalar multiplet in the conjugate representation , respectively .
the conjugation matrix , @xmath31 , is an antisymmetric @xmath32 matrix equal to @xmath33 for the su(2)@xmath34 doublet , and is given in appendix [ app : massesmixing ] for the @xmath25 and @xmath35 representations .
the @xmath36 and @xmath37 matrices are the generators of @xmath10 in the doublet and @xmath12-plet representations , respectively .
the term @xmath38 can only be non - zero when @xmath12 is an even number , i.e. , for half - odd - integer values for the total isospin , @xmath39 , of the scalar multiplet . together with the constraint @xmath40 ( @xmath41 ) for complex scalar multiplets that arises from perturbative unitarity of scattering amplitudes involving the large multiplet s weak charge @xcite , this limits the models of interest to the cases @xmath15 ( @xmath25 ) and @xmath17 ( @xmath35 ) .
has been studied in ref .
@xcite . ] for these two cases , the large multiplet is given in the electroweak basis by @xmath42 note that the conjugate of the charged state @xmath43 is written as @xmath44 , which is not the same as @xmath45 . when the @xmath46 term in eq .
( [ eq : conjugatepotential ] ) vanishes , the lagrangian preserves a @xmath47 symmetry ( models with such a @xmath47-symmetric potential have been studied in ref .
the @xmath46 term breaks the @xmath47 down to a @xmath3 symmetry @xcite , and splits the complex neutral component of @xmath0 into its real and imaginary parts , @xmath48 and @xmath49 , with different masses .
furthermore , the @xmath46 term induces and controls the amount of mixing between the charged states with the same electric charge , @xmath50 and @xmath51 .
the mass eigenstates are defined for @xmath52 in terms of a mixing angle @xmath53 such that @xmath54 with @xmath55 .
since there is only one state with @xmath56 , it remains unmixed .
expressions for the mixing angles are given in eqs .
( 7 ) and ( 14 ) of ref .
@xcite , and the details are summarized in appendix [ app : massesmixing ] for completeness .. the expressions in eqs .
( 7 ) and ( 14 ) of ref .
@xcite are correct . ]
for these models to contain a dark matter candidate , we require that the lightest ( stable ) member of the large multiplet be electrically neutral .
this occurs only when @xmath57 .
in addition , and without loss of generality , we choose the real part of @xmath58 to be the lightest member of the large multiplet ; this constrains the sign of @xmath46 such that @xmath59 for the sextet model and @xmath60 for the octet .
the masses of the physical states in terms of the mass of the neutral real particle , @xmath61 , and the lagrangian parameters @xmath62 and @xmath46 are given by @xcite @xmath63 , \nonumber \\
m_{\zeta^{+\frac{n}{2}}}^2 & = & { \ensuremath{m_{\zeta^{0,r}}}}^2 - \frac{n}{8}v^2\left[\lambda_3 + 2 ( -1)^{\frac{n}{2}+1}\lambda_4\right ] , \label{eq : masses } \end{aligned}\ ] ] where the notation is such that the sign in @xmath64 forces the relation @xmath55 .
the physical scalars arising from the large multiplet always occur in the same hierarchy , given from lightest to heaviest by : @xmath65 sample mass spectra are shown in fig .
[ fig : massspectra ] , where the two left plots are for @xmath25 and the two right plots are for @xmath35 . for each of the two plots for each multiplet
, we have fixed one of @xmath62 or @xmath46 and varied the other to illustrate the effect of varying these parameters . or @xmath46 .
the left two plots are for the @xmath16 model and show , from top to bottom , @xmath66 ( dashed black ) , @xmath67 ( dashed red ) , @xmath68 ( dashed blue ) , @xmath69 ( solid green ) , @xmath70 ( solid blue ) , @xmath71 ( solid red ) , and @xmath72 ( solid black ) . the right two plots are for the @xmath18 model and show , from top to bottom , @xmath66 ( dashed black ) , @xmath67 ( dashed red ) , @xmath68 ( dashed blue ) , @xmath73 ( dashed green ) , @xmath74 ( solid brown ) , @xmath75 ( solid green ) , @xmath70 ( solid blue ) , @xmath71 ( solid red ) , and @xmath72 ( solid black).,title="fig:",scaledwidth=24.0% ] or @xmath46 .
the left two plots are for the @xmath16 model and show , from top to bottom , @xmath66 ( dashed black ) , @xmath67 ( dashed red ) , @xmath68 ( dashed blue ) , @xmath69 ( solid green ) , @xmath70 ( solid blue ) , @xmath71 ( solid red ) , and @xmath72 ( solid black ) . the right two plots are for the @xmath18 model and show , from top to bottom , @xmath66 ( dashed black ) , @xmath67 ( dashed red ) , @xmath68 ( dashed blue ) , @xmath73 ( dashed green ) , @xmath74 ( solid brown ) , @xmath75 ( solid green ) , @xmath70 ( solid blue ) , @xmath71 ( solid red ) , and @xmath72 ( solid black).,title="fig:",scaledwidth=24.0% ] or @xmath46 .
the left two plots are for the @xmath16 model and show , from top to bottom , @xmath66 ( dashed black ) , @xmath67 ( dashed red ) , @xmath68 ( dashed blue ) , @xmath69 ( solid green ) , @xmath70 ( solid blue ) , @xmath71 ( solid red ) , and @xmath72 ( solid black ) .
the right two plots are for the @xmath18 model and show , from top to bottom , @xmath66 ( dashed black ) , @xmath67 ( dashed red ) , @xmath68 ( dashed blue ) , @xmath73 ( dashed green ) , @xmath74 ( solid brown ) , @xmath75 ( solid green ) , @xmath70 ( solid blue ) , @xmath71 ( solid red ) , and @xmath72 ( solid black).,title="fig:",scaledwidth=24.0% ] or @xmath46 .
the left two plots are for the @xmath16 model and show , from top to bottom , @xmath66 ( dashed black ) , @xmath67 ( dashed red ) , @xmath68 ( dashed blue ) , @xmath69 ( solid green ) , @xmath70 ( solid blue ) , @xmath71 ( solid red ) , and @xmath72 ( solid black ) .
the right two plots are for the @xmath18 model and show , from top to bottom , @xmath66 ( dashed black ) , @xmath67 ( dashed red ) , @xmath68 ( dashed blue ) , @xmath73 ( dashed green ) , @xmath74 ( solid brown ) , @xmath75 ( solid green ) , @xmath70 ( solid blue ) , @xmath71 ( solid red ) , and @xmath72 ( solid black).,title="fig:",scaledwidth=24.0% ] as seen in eq .
( [ eq : conjugatepotential ] ) , the states of @xmath0 interact with the sm through a coupling to the higgs doublet , as well as through their gauge - kinetic terms .
the couplings to the photon and the higgs boson are diagonal in the mass basis , and thus do not induce decays .
however , off - diagonal vertices involving two different scalar mass eigenstates and a @xmath7 or @xmath6 boson do exist .
thus , the decays of the mass eigenstates in @xmath0 to lighter members of @xmath0 occur only through emission of a @xmath6 or @xmath7 , which may be off - shell depending on the mass splitting involved .
this leads to a distinctive decay pattern , which is shown for the @xmath16 and @xmath18 models in fig .
[ fig : chargespectrum ] for a typical parameter point .
( left ) and @xmath18 ( right ) models for a typical parameter point .
the upper branch of states includes the @xmath76 , @xmath77 and @xmath78 , while the lower branch includes @xmath5 and @xmath79 , each sorted by electric charge .
decays are shown as diagonal ( @xmath6 emission ) or vertical ( @xmath7 emission ) arrows . on -
shell decays are indicated by a solid line and off - shell decays by a dashed line.,title="fig:",scaledwidth=40.0% ] ( left ) and @xmath18 ( right ) models for a typical parameter point .
the upper branch of states includes the @xmath76 , @xmath77 and @xmath78 , while the lower branch includes @xmath5 and @xmath79 , each sorted by electric charge .
decays are shown as diagonal ( @xmath6 emission ) or vertical ( @xmath7 emission ) arrows . on -
shell decays are indicated by a solid line and off - shell decays by a dashed line.,title="fig:",scaledwidth=40.0% ] reference @xcite studied theoretical and indirect experimental constraints on the model parameters @xmath80 , which we apply here : * the zeroth partial wave @xmath81 scattering amplitude involving pairs of scalars must satisfy perturbative unitarity , @xmath82 .
this results in constraints for the sextet model of @xmath83 6.59 , @xmath84 8.48 , and @xmath85 4.25 , and for the octet model of @xmath83 3.10 , @xmath84 5.46 , and @xmath85 2.74 .
* the scalar potential should not have an alternative minimum in which @xmath0 gets a vacuum expectation value . a sufficient condition to ensure
this is to set @xmath86 in eq .
( [ eq : conjugatepotential ] ) . *
the electroweak oblique parameters @xmath87 , @xmath39 , and @xmath88 @xcite , receive contributions from electroweak gauge boson self - energy diagrams involving the states in @xmath0 .
the measured oblique parameters constrain @xmath89 and favour @xmath90 .
this constraint causes the @xmath91 state to tend to cluster with the heavier scalars , as can be seen in the first and third plots in fig .
[ fig : massspectra ] . this will be the case unless the overall spectrum is highly compressed .
* the sm higgs boson decay width to two photons receives contributions from the charged scalars in @xmath0 running in the loop .
the lhc measurement of the @xmath92 rate constrains @xmath93 more strongly than the unitarity constraint when @xmath94 gev . since @xmath93 does not appear explicitly in the mass formulae in eq .
( [ eq : masses ] ) or affect any of the electroweak production and decay rates , we satisfy this constraint by setting @xmath95 for this work .
we checked numerically that this choice does not further restrict the range of the remaining parameters beyond the previous three constraints . in this study , it is most convenient to reparameterize the models in terms of physical masses and mass splittings .
we take as inputs @xmath61 as in eq .
( [ eq : masses ] ) , and replace @xmath62 and @xmath46 with the mass splitting parameters @xmath96 and @xmath97 , defined as @xmath98 here @xmath96 parameterizes the mass of the highest - charged state , @xmath99 , in terms of the mass splitting @xmath100 as a fraction of the mass splitting between the lightest and heaviest state of @xmath0 , @xmath101 ( see fig .
[ fig : massspectra ] ) , while @xmath97 parameterizes @xmath102 as a fraction of the maximum such splitting @xmath103 allowed after the theoretical and indirect experimental constraints are imposed on the model .
the numerical values of @xmath103 allowed by the theoretical and indirect experimental constraints are given in table [ tbl : dm0max ] for the four values of @xmath61 used in our simulation .
we study four mass slices , with @xmath104 , 120 , 180 , and 300 gev .
( lhc8 does not provide any exclusions for @xmath61 larger than 300 gev in these models . )
c6emc4.5emc4.5em & + & @xmath105 & @xmath106 + @xmath107 & @xmath108 & @xmath109 + @xmath110 & @xmath111 & @xmath112 + @xmath113 & @xmath114 & @xmath115 + @xmath116 & @xmath117 & @xmath118 + the mass splitting parameters are normalized to fall in the ranges @xmath119 $ ] and @xmath120 $ ] .
from eq .
( [ eq : masses ] ) , it is clear that as @xmath121 , @xmath122 , and as @xmath123 , @xmath124 .
furthermore , as illustrated in fig .
[ fig : massspectra ] , the mass splittings among the states are maximized when @xmath125 , while the spectrum collapses into two tightly - clustered groups of states when @xmath122 or 1 .
also , because @xmath97 parameterizes the overall mass splitting between the neutral real and imaginary scalars , the entire spectrum becomes compressed as @xmath126 , whereas the splitting between the heavier and lighter states is maximized when @xmath127 .
the parameter regions allowed by the theoretical and indirect experimental constraints studied in ref .
@xcite are plotted in figs .
[ fig : r1r2envconstl3l4_n6 ] and [ fig : r1r2envconstl3l4_n8 ] in the @xmath96@xmath97 plane , for the four @xmath61 values that we study .
the yellow shaded regions are allowed . on each plot
we also show the contours of constant @xmath62 and @xmath46 .
@xmath97 plane , for four values of @xmath61 in the @xmath16 model .
the allowed region is shaded in yellow .
contours of constant @xmath62 are shown as solid curves with pink labels and contours of constant @xmath46 are shown as dashed horizontal lines with blue labels .
curves are labelled with their respective @xmath62 or @xmath46 value.,title="fig:",scaledwidth=24.0% ] @xmath97 plane , for four values of @xmath61 in the @xmath16 model .
the allowed region is shaded in yellow .
contours of constant @xmath62 are shown as solid curves with pink labels and contours of constant @xmath46 are shown as dashed horizontal lines with blue labels .
curves are labelled with their respective @xmath62 or @xmath46 value.,title="fig:",scaledwidth=24.0% ] @xmath97 plane , for four values of @xmath61 in the @xmath16 model .
the allowed region is shaded in yellow .
contours of constant @xmath62 are shown as solid curves with pink labels and contours of constant @xmath46 are shown as dashed horizontal lines with blue labels .
curves are labelled with their respective @xmath62 or @xmath46 value.,title="fig:",scaledwidth=24.0% ] @xmath97 plane , for four values of @xmath61 in the @xmath16 model .
the allowed region is shaded in yellow .
contours of constant @xmath62 are shown as solid curves with pink labels and contours of constant @xmath46 are shown as dashed horizontal lines with blue labels .
curves are labelled with their respective @xmath62 or @xmath46 value.,title="fig:",scaledwidth=24.0% ] but for the @xmath18 model.,title="fig:",scaledwidth=24.0% ] but for the @xmath18 model.,title="fig:",scaledwidth=24.0% ] but for the @xmath18 model.,title="fig:",scaledwidth=24.0% ] but for the @xmath18 model.,title="fig:",scaledwidth=24.0% ]
the mass splittings and preferred decay modes between the physical scalars in our models vary substantially over the allowed parameter space . understanding the collider constraints requires an understanding of the branching fractions of the unstable states over the @xmath96@xmath97 plane . in this section , we provide insight into the experimental results through examining the primary decay modes of the heavier branch of states over the full @xmath96@xmath97 plane , for @xmath104 , 120 , 180 , and 300 gev
. decays in the @xmath16 model are shown in figs .
[ fig : decays_z0i_n6][fig : decays_h22_n6 ] and decays in the @xmath18 model are shown in figs .
[ fig : decays_z0i_n8][fig : decays_h23_n8 ] . we do not plot the decay branching fractions of the lighter states , as these states always cascade down to @xmath72 through @xmath6 emission as illustrated in fig .
[ fig : chargespectrum ] . except for small regions in the transition from one dominant decay mode to another ,
the dominant mode has a branching fraction of at least 50% .
we note that decays of a heavy state to another state in the heavier branch , such as @xmath128 , are never dominant .
this is because they are highly suppressed by the lack of available phase space , making them effectively negligible over the entire parameter space .
our numerical results were obtained as follows .
we implemented the two models in ` feynrules ` version 2.0.26 @xcite and generated the corresponding universal feynrules output ( ufo ) files .
we used these with ` madgraph5_amc@nlo ` version 2.1.2 @xcite to compute the decay branching ratios of all the new scalars in a grid over the @xmath96@xmath97 plane for each of the four @xmath61 slices .
in the @xmath25 model . shown are @xmath129 ( orange / red ) and @xmath130 ( purple / black ) .
the decay @xmath131 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath129 ( orange / red ) and @xmath130 ( purple / black ) .
the decay @xmath131 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath129 ( orange / red ) and @xmath130 ( purple / black ) .
the decay @xmath131 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath129 ( orange / red ) and @xmath130 ( purple / black ) .
the decay @xmath131 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath129 ( orange / red ) and @xmath130 ( purple / black ) .
the decay @xmath131 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints . ] in the @xmath25 model . shown
are @xmath132 ( yellow ) , @xmath133 ( orange / red ) , and @xmath134 ( purple / black ) .
the decay @xmath135 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath132 ( yellow ) , @xmath133 ( orange / red ) , and @xmath134 ( purple / black ) .
the decay @xmath135 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath132 ( yellow ) , @xmath133 ( orange / red ) , and @xmath134 ( purple / black ) .
the decay @xmath135 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath132 ( yellow ) , @xmath133 ( orange / red ) , and @xmath134 ( purple / black ) .
the decay @xmath135 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath132 ( yellow ) , @xmath133 ( orange / red ) , and @xmath134 ( purple / black ) .
the decay @xmath135 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints . ] in the @xmath25 model .
shown are @xmath136 ( yellow ) , @xmath137 ( orange / red ) , and @xmath138 ( purple / black ) .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath136 ( yellow ) , @xmath137 ( orange / red ) , and @xmath138 ( purple / black ) .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath136 ( yellow ) , @xmath137 ( orange / red ) , and @xmath138 ( purple / black ) .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath136 ( yellow ) , @xmath137 ( orange / red ) , and @xmath138 ( purple / black ) .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath25 model . shown
are @xmath136 ( yellow ) , @xmath137 ( orange / red ) , and @xmath138 ( purple / black ) .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints . ] in the @xmath35 model . shown
are @xmath129 ( orange / red ) and @xmath130 ( purple / black ) .
the decay @xmath131 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model .
shown are @xmath129 ( orange / red ) and @xmath130 ( purple / black ) .
the decay @xmath131 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath129 ( orange / red ) and @xmath130 ( purple / black ) .
the decay @xmath131 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath129 ( orange / red ) and @xmath130 ( purple / black ) .
the decay @xmath131 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model .
shown are @xmath129 ( orange / red ) and @xmath130 ( purple / black ) .
the decay @xmath131 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints . ] in the @xmath35 model .
shown are @xmath132 ( yellow ) , @xmath133 ( orange / red ) , and @xmath134 ( purple / black ) .
the decay @xmath135 is severely kinematically suppressed over the entire parameter space and never dominates
. the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath132 ( yellow ) , @xmath133 ( orange / red ) , and @xmath134 ( purple / black ) .
the decay @xmath135 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath132 ( yellow ) , @xmath133 ( orange / red ) , and @xmath134 ( purple / black ) .
the decay @xmath135 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath132 ( yellow ) , @xmath133 ( orange / red ) , and @xmath134 ( purple / black ) .
the decay @xmath135 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath132 ( yellow ) , @xmath133 ( orange / red ) , and @xmath134 ( purple / black ) .
the decay @xmath135 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints . ] in the @xmath35 model . shown
are @xmath139 ( yellow ) , @xmath137 ( orange / red ) , and @xmath138 ( purple / black ) .
the decay @xmath140 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath139 ( yellow ) , @xmath137 ( orange / red ) , and @xmath138 ( purple / black ) .
the decay @xmath140 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath139 ( yellow ) , @xmath137 ( orange / red ) , and @xmath138 ( purple / black ) .
the decay @xmath140 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath139 ( yellow ) , @xmath137 ( orange / red ) , and @xmath138 ( purple / black ) .
the decay @xmath140 is severely kinematically suppressed over the entire parameter space and never dominates
. the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath139 ( yellow ) , @xmath137 ( orange / red ) , and @xmath138 ( purple / black ) .
the decay @xmath140 is severely kinematically suppressed over the entire parameter space and never dominates .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints . ] in the @xmath35 model . shown
are @xmath141 ( yellow ) , @xmath142 ( orange / red ) , and @xmath143 ( purple / black ) .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath141 ( yellow ) , @xmath142 ( orange / red ) , and @xmath143 ( purple / black ) .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath141 ( yellow ) , @xmath142 ( orange / red ) , and @xmath143 ( purple / black ) .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath141 ( yellow ) , @xmath142 ( orange / red ) , and @xmath143 ( purple / black ) .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints.,title="fig:",scaledwidth=24.0% ] in the @xmath35 model . shown
are @xmath141 ( yellow ) , @xmath142 ( orange / red ) , and @xmath143 ( purple / black ) .
the darker shade of each colour indicates that the gauge boson is emitted on - shell , while the lighter shade indicates it is off - shell .
the region below the yellow curve is allowed by indirect and precision electroweak constraints . ]
the branching ratios vary over the parameter space due to variation of the scalar mass splittings and the mixing angles @xmath53 , which in turn control the gauge couplings involved in the decays .
the variation in the dominant decay mode with @xmath96 is primarily due to variation in the decay couplings , which leads to the essentially vertical stripes in figs .
[ fig : decays_z0i_n6][fig : decays_h23_n8 ] .
the dependence on the mass splittings appears in regions of parameter space where one or more decays are transitioning between off - shell and on - shell .
for example , in fig .
[ fig : decays_z0i_n6 ] , the @xmath129 decay goes on - shell at a lower @xmath97 value than the other decay @xmath130 , leading the former to encroach into the parameter region that would have otherwise been dominated by the latter .
this effect also shows up in fig .
[ fig : decays_h21_n6 ] : because @xmath144 , the decay of @xmath145 to @xmath146 goes on - shell while the decay to @xmath147 is still off - shell , leading to a small region of parameter space in which the former branching ratio surpasses the latter ( the small island of black in the yellow region near the left side of the plot for @xmath148 gev ) , despite the relative size of the couplings favoring decays to the latter . to illustrate the interplay of coupling strengths and mass splittings in the pattern of branching ratios
, we examine in more detail two slices through the parameter space of figs .
[ fig : decays_h22_n6 ] and [ fig : decays_z0i_n8 ] . in fig .
[ fig : z0i_br_dm_c_n8_m180 ] we show a slice through the @xmath149 gev panel of fig . [ fig : decays_z0i_n8 ] ( decays of @xmath66 in the @xmath18 model ) at @xmath150 . the relevant mass splittings ( left plot in fig .
[ fig : z0i_br_dm_c_n8_m180 ] , normalized to the mass of the daughter gauge boson ) are nearly constant with @xmath96 , as are the @xmath151 and @xmath152 couplings ( middle plot in fig .
[ fig : z0i_br_dm_c_n8_m180 ] ) . instead , the variation in the branching ratios is due to the variation in the @xmath153 coupling , which passes through zero around @xmath154 .
this behavior can be traced back to the variation of the mixing angle @xmath155 .
this results in a big dip in the @xmath129 branching ratio in favor of the decay @xmath130 ( right plot in fig .
[ fig : z0i_br_dm_c_n8_m180 ] ) . for small @xmath96 where the decay to @xmath156 does dominate over @xmath157 ,
the @xmath153 coupling is still smaller than the @xmath152 coupling ; the larger branching fraction is instead due to the larger available phase space for the former mode .
note also that , although the @xmath158 coupling is significantly stronger than either of the others , the resulting decay is severely kinematically suppressed and never dominates .
this is a generic feature of our models : due to the pattern of mixings among the isospin eigenstates , the gauge couplings between pairs of scalars tend to be largest when their mass splittings are smallest .
the exception is the @xmath152 coupling , which does not depend on any mixing angles . in fig .
[ fig : h22_br_dm_c_n6_m120 ] we show a slice through the @xmath148 gev panel of fig . [ fig : decays_h22_n6 ] ( decays of @xmath68 in the @xmath25 model ) at @xmath159 .
the three - band structure in fig .
[ fig : decays_h22_n6 ] emerges in a natural way from the interplay between couplings and mass splittings . at small @xmath96
the @xmath160 coupling is sizable and the available phase space is large , so this decay dominates . as @xmath96 increases , the mass of @xmath69 increases
( see also fig .
[ fig : massspectra ] ) , and the @xmath161 branching fraction drops due to the squeezing of the phase space , in spite of the growth of the @xmath160 coupling .
meanwhile , the @xmath162 coupling is passing through zero around @xmath163 .
this allows the decay to @xmath164 to dominate for @xmath96 between about 0.3 and 0.7 . for high values of @xmath96 ,
the rising @xmath162 coupling and its larger phase space allow the decay to @xmath165 to become dominant . as a function of @xmath96 in the @xmath35 model for @xmath149 gev and @xmath150 ( compare fig .
[ fig : decays_z0i_n8 ] ) . shown
are the decays to @xmath156 ( solid black curves ) , @xmath157 ( dashed red curves ) , and @xmath166 ( dotted blue curves ) .
note that the @xmath153 squared coupling ( solid black curve ) touches zero near @xmath154.,title="fig:",scaledwidth=30.0% ] as a function of @xmath96 in the @xmath35 model for @xmath149 gev and @xmath150 ( compare fig .
[ fig : decays_z0i_n8 ] ) . shown
are the decays to @xmath156 ( solid black curves ) , @xmath157 ( dashed red curves ) , and @xmath166 ( dotted blue curves ) .
note that the @xmath153 squared coupling ( solid black curve ) touches zero near @xmath154.,title="fig:",scaledwidth=30.0% ] as a function of @xmath96 in the @xmath35 model for @xmath149 gev and @xmath150 ( compare fig .
[ fig : decays_z0i_n8 ] ) . shown
are the decays to @xmath156 ( solid black curves ) , @xmath157 ( dashed red curves ) , and @xmath166 ( dotted blue curves ) .
note that the @xmath153 squared coupling ( solid black curve ) touches zero near @xmath154.,title="fig:",scaledwidth=30.0% ] as a function of @xmath96 in the @xmath25 model for @xmath148 gev and @xmath159 ( compare fig .
[ fig : decays_h22_n6 ] ) . shown
are the decays to @xmath167 ( dashed red curves ) , @xmath168 ( solid black curves ) , and @xmath169 ( dotted blue curves ) .
note that the @xmath170 squared coupling ( dotted blue curve ) touches zero near @xmath163.,title="fig:",scaledwidth=30.0% ] as a function of @xmath96 in the @xmath25 model for @xmath148 gev and @xmath159 ( compare fig .
[ fig : decays_h22_n6 ] ) . shown
are the decays to @xmath167 ( dashed red curves ) , @xmath168 ( solid black curves ) , and @xmath169 ( dotted blue curves ) .
note that the @xmath170 squared coupling ( dotted blue curve ) touches zero near @xmath163.,title="fig:",scaledwidth=30.0% ] as a function of @xmath96 in the @xmath25 model for @xmath148 gev and @xmath159 ( compare fig .
[ fig : decays_h22_n6 ] ) .
shown are the decays to @xmath167 ( dashed red curves ) , @xmath168 ( solid black curves ) , and @xmath169 ( dotted blue curves ) .
note that the @xmath170 squared coupling ( dotted blue curve ) touches zero near @xmath163.,title="fig:",scaledwidth=30.0% ] finally , we note that the maximum allowed mass splitting @xmath171 is considerably smaller in the @xmath18 model than in the @xmath16 model ( see table [ tbl : dm0max ] ) . while on - shell decays first appear in the @xmath148 gev plots for the @xmath16 model , they only appear in the @xmath172 gev plots for the @xmath18 model .
lhc constraints on models with stable neutral particles ( i.e. , wimp dark matter candidates ) typically arise from searches for excess events with large missing transverse energy ( @xmath21 ) recoiling against some combination of visible matter .
these types of signatures have been studied both in the context of supersymmetry and in generic mono - jet searches by both atlas and cms using lhc8 data .
searches for @xmath173 are typically only constraining for signals produced via qcd - strength interactions because of the large @xmath174 backgrounds ; nevertheless , we include them here for completeness .
signatures involving @xmath175 , for which the sm backgrounds are smaller , are typically more constraining for models in which the new physics is produced with electroweak - strength cross sections , as in our models . these @xmath175 signatures will be responsible for the direct - search constraints on our models .
we take advantage of the large number of possible pairs of scalars that can be produced via @xmath20-channel @xmath176 , @xmath7 , or @xmath6 exchange in our models by simulating all such processes and combining the resulting signal events .
the following searches were reproduced : * atlas opposite - sign dileptons with @xmath21 and no jets @xcite ; * atlas trilepton plus @xmath21 @xcite ( the two studies employ distinct methodologies , and are reproduced separately ) ; * atlas four or more leptons @xcite ; * atlas dileptons with razor variables @xcite ; * atlas hadronic di-@xmath177 plus @xmath21 @xcite ; * atlas same - sign dileptons plus jets @xcite ; * atlas monojet @xcite ; * atlas multi - jets plus @xmath21 @xcite ; * cms multi - leptons ( dilepton , trilepton , multi - lepton ) with @xmath21 @xcite ; * cms monojet @xcite .
we performed the simulation following the methodology of ref .
@xcite . for each of the four mass slices , @xmath104 , 120 , 180 , and 300 gev in each model , we generated parameter points on a grid in @xmath96 and @xmath97 , focusing on the region that is allowed by the theoretical and indirect experimental constraints .
we also simulated a few additional points for added sensitivity in regions where the experimental sensitivity was varying rapidly .
the parameter points that we simulated are shown as green dots on our exclusion plots .
we do not simulate parameter points for which @xmath178 gev .
as we will see , this compressed - spectrum region is particularly difficult to detect because the leptons resulting from the decays tend to be too soft to pass the selection cuts .
indeed , we will find that parameter regions with @xmath179 gev can not be excluded using the lhc8 searches that we recast .
this parameter region could be picked up by @xmath173 searches , but we find that the scalar pair production cross sections are too small for lhc8 to have any sensitivity in this channel .
very small mass splittings could result in macroscopic decay lengths for charged scalars ; while dedicated searches for such signatures exist , we will not consider them here . using ` madgraph5_amc@nlo ` and our ` feynrules ` model files , we generated 50,000 events for each of the possible pairs of scalars that can be produced via @xmath20-channel @xmath180 and @xmath181 exchange , in association with zero or one hard jet(s ) .
this amounted to 32 possible scalar pairings in the @xmath16 model and 44 in the @xmath18 model .
each simulated event sample was then matched and merged , decayed , showered , and hadronized using ` pythia 6 ` @xcite .
the events were then passed through the ` delphes 3 ` @xcite detector simulation multiple times : for each of the experimental searches that we recast , we adjusted the identification efficiencies and jet algorithm settings in ` delphes 3 ` to match the working point used in the experimental analysis .
we then used the ` seer ` @xcite analysis program to apply the trigger and kinematic cuts of each experimental analysis and determine the total signal cross section from all production processes that fall into each of the signal regions .
additional gaussian smearing of the @xmath21 beyond that already present in ` delphes 3 ` was introduced via ` seer ` to account for the effect of pileup @xcite .
this was necessary in order to reproduce the cut - flow tables for each of the experimental searches .
at each simulated parameter point , we select as most sensitive the experimental search that yields the largest value of @xmath182 , where @xmath183 is the number of signal events computed from the cross section in the signal region for the appropriate integrated luminosity and @xmath184 is the corresponding 95% confidence level exclusion threshold from the experimental analysis . if @xmath185 for the most sensitive search , the parameter point is excluded .
we do not combine significances from different search channels to do so would require knowledge of the statistical correlations among the various searches which we do not have . in between parameter points , we interpolate linearly in the value of @xmath182 found for the most sensitive search at each point .
we find that the most sensitive analyses are the atlas dilepton , trilepton and four - lepton searches .
several other searches also exclude parameter space for low values of @xmath61 , but this is primarily due to the large pair production rates when the scalar masses are low , which overcome the inherently smaller acceptances and/or larger backgrounds in these searches .
our results are shown in fig .
[ fig : results_n6 ] for the @xmath25 model and fig .
[ fig : results_n8 ] for the @xmath18 model . for each of the mass slices @xmath104 , 120 , 180 , and 300 gev
, we show the excluded region of the @xmath96@xmath97 plane ( darker - colored regions above and to the left of the solid black line ) .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
as before , the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints . in each plot the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev , below which the spectrum is highly compressed .
we now describe the results for each model . in the @xmath16 model ,
the parameter space is strongly excluded at low @xmath61 so long as @xmath188 gev . as @xmath61 increases , the excluded region shrinks until only a small sliver of parameter space around @xmath189 is excluded for @xmath149 gev .
lhc8 makes no exclusion at @xmath172 gev or above .
0.44 model from all examined lhc8 analyses .
the darker - colored regions above and to the left of the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] 0.44 model from all examined lhc8 analyses .
the darker - colored regions above and to the left of the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] 0.1 model from all examined lhc8 analyses .
the darker - colored regions above and to the left of the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] 0.44 model from all examined lhc8 analyses .
the darker - colored regions above and to the left of the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] 0.44 model from all examined lhc8 analyses .
the darker - colored regions above and to the left of the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] 0.1 model from all examined lhc8 analyses .
the darker - colored regions above and to the left of the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] for @xmath190 gev , shown in the upper left panel of fig .
[ fig : results_n6 ] , the total scalar pair production cross section before cuts ranges from @xmath191 pb for small @xmath97 to @xmath192 pb for large @xmath97 . the atlas @xmath193 searches
@xcite provide the most sensitive exclusion for all parameter points scanned .
due to the large signal cross sections , even small acceptance rates result in large numbers of events in the experimental signal regions , resulting in the strong exclusion as indicated by dark colors .
the excluded region ends abruptly at @xmath194 gev . for @xmath97 values below this boundary ,
the highest lepton @xmath195 is rarely above 10 gev , unless the scalars are produced boosted or in association with an energetic jet , which is rare .
the atlas and cms offline leptonic triggers are typically set at @xmath196 gev , resulting in a drastic plunge in acceptance .
similarly , the @xmath21 tends to be low when the spectrum is compressed ; most ( but not all ) of the searches require @xmath197 gev , further reducing sensitivity to the compressed - spectrum region .
this kinematic acceptance boundary is apparent in many of the exclusion plots for both the @xmath16 and @xmath18 models . for @xmath198 gev , shown in the upper right panel of fig .
[ fig : results_n6 ] , the total scalar pair production cross section before cuts ranges from @xmath199 pb for small @xmath97 to @xmath200 pb for large @xmath97 .
the atlas @xmath193 searches @xcite provide the most sensitive exclusion for @xmath201 , in particular the search region with same - flavor opposite - sign dilepton mass smaller than the @xmath7 mass , which picks up the off - shell @xmath7 boson decays . for @xmath202 and @xmath203 ,
the most sensitive search is the atlas four - lepton search @xcite , due to the smaller backgrounds and the predominance of @xmath130 and @xmath133 decays in this region ( see figs .
[ fig : decays_z0i_n6 ] and [ fig : decays_h21_n6 ] ) .
the unexcluded region in the upper right corner of this plot results from the loss of the four - lepton signatures as the decays involving @xmath6 emission go on shell ( the resulting two - lepton final states suffer from larger backgrounds ) . for @xmath204 gev , shown in the lower left panel of fig .
[ fig : results_n6 ] , the total scalar pair production cross section before cuts ranges from @xmath205 pb for small @xmath97 to @xmath206 pb for large @xmath97 .
the atlas @xmath193 searches @xcite are responsible for the exclusion between @xmath97 values of about 0.2 and 0.3 , while the atlas four - lepton search @xcite is responsible for the thin strip of excluded parameter space near @xmath207 . within the parameter region allowed by theoretical and indirect experimental constraints ,
only a small sliver of parameter space around @xmath189 is excluded by the atlas @xmath193 searches . the largest contribution to the @xmath193 signal
is from the @xmath208 pair production mode , which decays in this region of parameter space mainly to @xmath209 , followed by very soft decays of the @xmath210 states . as @xmath97 increases from 0.3 to 0.4 , the @xmath211 decay transitions to @xmath212 due to the nearby kinematic threshold ( see fig .
[ fig : decays_h21_n6 ] ) so that the third lepton is lost .
the production cross sections for the heavier scalars also decrease with increasing @xmath97 . for @xmath172 gev , shown in the lower right panel of fig .
[ fig : results_n6 ] , the total scalar pair production cross section before cuts ranges from @xmath213 pb for small @xmath97 to @xmath214 pb for large @xmath97 . after accounting for branching ratios and acceptances ,
it is clear that there is simply insufficient luminosity collected at lhc8 to result in any sensitivity to the model for this and higher masses . in the @xmath18 model , the parameter space at low @xmath61
is again excluded so long as @xmath215 gev .
the excluded region shrinks rapidly with increasing @xmath61 until lhc8 can make no exclusion at @xmath149 gev or above . 0.44 model from all examined lhc8 analyses .
the darker - colored regions above the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] 0.44 model from all examined lhc8 analyses .
the darker - colored regions above the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] 0.1 model from all examined lhc8 analyses .
the darker - colored regions above the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] 0.44 model from all examined lhc8 analyses .
the darker - colored regions above the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] 0.44 model from all examined lhc8 analyses .
the darker - colored regions above the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] 0.1 model from all examined lhc8 analyses .
the darker - colored regions above the solid black line are excluded at 95% confidence level by at least one analysis .
the dashed black line indicates how the excluded region would expand if the signal cross sections were all increased by a @xmath186-factor of 1.2 .
the parameter region below the thick teal curve is allowed by the theoretical and indirect experimental constraints .
the thick horizontal light pink line indicates the @xmath97 value at which @xmath187 gev.,title="fig : " ] the most important difference between the two models is that the maximum mass splitting between the lightest and heaviest scalars , @xmath171 , is considerably smaller in the @xmath18 model than in the @xmath16 model , by about a factor of two ( see table [ tbl : dm0max ] ) .
therefore , a much larger fraction of the allowed @xmath18 model parameter space lies in the compressed region with @xmath216 gev ( below the thick light pink horizontal line in fig .
[ fig : results_n8 ] ) .
the smaller mass splittings over the entire @xmath18 parameter space result in softer leptons , jets , and @xmath21 and hence overall lower acceptance rates in the lhc8 analyses .
this results in weaker exclusions in the @xmath18 model in spite of the larger number of scalars , their higher weak and electric charges ( leading to more like - sign leptons from longer decay chains ) , and the correspondingly higher total scalar pair production cross sections before cuts .
the emitted leptons are just too soft . for @xmath104 gev , shown in the upper left panel of fig .
[ fig : results_n8 ] , the total scalar pair production cross section before cuts ranges from @xmath217 pb for small @xmath97 to @xmath218 pb for large @xmath97 . the entire parameter space with @xmath202 is excluded .
similarly to the @xmath16 model , the excluded region ends around @xmath194 gev , below which the highest lepton @xmath195 is rarely above 10 gev , so that most events fail the offline leptonic triggers .
the most sensitive exclusion comes from the published atlas @xmath193 study @xcite .
this study captures a broader range of signal characteristics than the corresponding conference note @xcite , and includes signal regions that are more sensitive to low dilepton invariant masses ( called the @xmath219 bins 14 in ref .
in particular , the best constraints on the @xmath18 model come from @xmath219 bin 2 , in which the invariant mass of the same - flavor opposite - sign lepton pair lies between 12 and 40 gev , the transverse mass of the remaining lepton and the @xmath21 lies below 80 gev , and @xmath220 gev . the most constraining bin in ref .
@xcite is similar , but with a softer cut on @xmath21 , which results in larger background .
the atlas four - lepton search @xcite also redundantly excludes part of the parameter space , but loses sensitivity for larger values of @xmath96 , where the dominant decay of @xmath221 changes from @xmath164 to @xmath165 . for @xmath148 gev , shown in the upper right panel of fig .
[ fig : results_n8 ] , the total scalar pair production cross section before cuts ranges from @xmath222 pb for small @xmath97 to @xmath223 pb for large @xmath97 .
the excluded region is understood similarly to the previous mass slice , with the same mechanisms at work . when the smaller cross sections are combined with branching ratios and acceptance rates , the signal events in the published atlas @xmath193 selection @xcite wind up falling into multiple signal bins , rather than being as concentrated in a single bin as in the @xmath104 gev slice . since we do not have information on the correlations of the systematic uncertainties among different bins
, we can not employ the full usefulness of the @xmath224 method to combine the bins to extend the exclusion boundaries .
in addition , decays to pairs of off - shell @xmath7 bosons are common , which reduces the trilepton signal from @xmath225 production .
a four - lepton search is intrinsically less sensitive than a trilepton search due to the smaller leptonic branching ratio of the @xmath7 compared to the @xmath6 . for @xmath149 gev , shown in the lower left panel of fig .
[ fig : results_n8 ] , the total scalar pair production cross section before cuts ranges from @xmath226 pb for small @xmath97 to @xmath227 pb for large @xmath97 .
none of the scanned parameter space is excluded by lhc8 searches .
a small region of parameter space at @xmath228 and @xmath229 would be excluded if a signal @xmath186-factor of 1.2 were included .
the sensitivity in this region comes from the atlas opposite - sign @xmath230 search with no jets @xcite , which is sensitive to the off - shell @xmath7 decays from the @xmath221 and @xmath66 that are produced in association with @xmath231 states .
jets from the decays of the @xmath231 states are soft enough that they evade the jet veto in the search . for @xmath172 gev , shown in the lower right panel of fig .
[ fig : results_n8 ] , the signal rate after cuts in any channel is less than 10% of that required for an exclusion over the entire scanned parameter space .
we studied the direct - search constraints from lhc8 data on models in which the higgs sector is extended by a @xmath3-odd complex scalar electroweak multiplet @xmath0 with isospin @xmath1 ( @xmath16 ) or @xmath232 ( @xmath18 ) and the same hypercharge as the sm higgs doublet .
these models can be probed by recasting the dedicated searches for supersymmetric particles performed by the cms and atlas collaborations .
we showed that , even after imposing theoretical and indirect experimental constraints ( mainly from the electroweak oblique parameters ) , the lhc8 searches provide nontrivial further constraints on the parameter space of the two models .
the data from lhc8 excludes the majority of the remaining parameter space for @xmath104120 gev in both the @xmath16 and @xmath18 models , except for parameter regions in which the entire spectrum is compressed within a mass splitting @xmath216 gev .
the loss of sensitivity in the compressed spectrum region is caused by the extreme softness of the cascade decay products , leading most of the signal events to fail the trigger and minimum energy requirements for lepton tagging .
the lhc8 constraints rapidly disappear for @xmath233 gev , until there is little sensitivity above @xmath149 gev .
these results are in general agreement with the lhc8 constraints on the electroweak gauginos in supersymmetric models studied in ref .
@xcite , in which the same collection of lhc8 searches showed sensitivity to masses of the lightest supersymmetric particle primarily smaller than 100 gev .
the lhc8 constraints on our models are much weaker than those on the scalar septet extension of the sm higgs sector @xcite studied in ref .
@xcite . using similar searches from atlas and cms , ref .
@xcite found that septet masses below @xmath234 gev were robustly excluded , and that multi - lepton searches provided the strongest constraints .
the dramatically enhanced lhc8 sensitivity to the septet model compared to our sextet and octet models is due to the nonzero vacuum expectation value carried by the septet and the mixing between the neutral scalar in the septet model and the sm higgs boson in that model .
these features allow the lightest septet state to decay into sm particles , particularly into energetic gauge boson pairs , leading to final states with two to four high - energy leptons .
the mass reach of the lhc8 exclusion is then controlled by the decrease in the production cross section with increasing scalar masses . with
run 2 of the lhc underway at 13 tev center - of - mass energy , the next round of searches will likely provide stronger constraints on our models , extending to higher scalar masses .
the pair production cross section for fixed scalar masses grows with increasing proton - proton center - of - mass energy due to the rise in the antiquark parton density .
higher integrated luminosity will also aid the searches , but with diminishing returns .
the analyses that we use compare an observed event number to theoretical predictions for the background cross section ; unless these backgrounds can be normalized from data , such searches suffer from systematic uncertainties that ultimately limit the benefit of greater luminosity .
in addition , higher instantaneous luminosity will necessitate an increase in trigger thresholds , which will further reduce the sensitivity to the compressed spectrum regions .
discovery prospects at a higher energy proton - proton collider would face similar issues . finally , we would be remiss not to mention the fact that the scalars in @xmath0 could be pair produced via @xmath20-channel photon and @xmath7 exchange at a lepton collider such as the international linear collider ( ilc ) @xcite
. such a machine would be able to probe scalar masses up to the pair production threshold @xmath235 with small backgrounds .
ilc search prospects in the compressed spectrum region would require a dedicated study .
we would like to thank mikhail batygov , wade hong , hugues beauchesne , and kevin earl for their help in setting up and troubleshooting computing resources , and katy hartling for discussions about the models .
this work was supported by the natural sciences and engineering research council of canada .
in this section we give some of the mathematical details used in the derivation of the mass spectrum and mixing angles in sec .
[ sec : models ] . for a complex scalar multiplet @xmath0 with hypercharge @xmath14 ( normalized so that @xmath13 ) , the most general gauge - invariant and @xmath3-invariant
renormalizable scalar potential was given in eq .
( [ eq : conjugatepotential ] ) , in which @xmath236 and @xmath237 are the conjugate multiplets . here
@xmath238 is the second pauli matrix and the conjugation matrix @xmath31 for the large multiplet is an anti - diagonal @xmath32 matrix . for @xmath16 and 8 ,
the matrix @xmath31 is given by @xmath239 taking @xmath46 real and working in unitarity gauge , the term involving @xmath46 in the scalar potential of eq .
( [ eq : conjugatepotential ] ) reduces to @xmath240,\ ] ] where @xmath241 . for @xmath16
the generators @xmath37 are given by @xmath242 @xmath243 while for @xmath18 they are @xmath244 @xmath245 the terms @xmath246 and @xmath247 split the masses of @xmath72 and @xmath66 and cause mixing between states with the same electric charge but different isospin . for @xmath16 or 8 ,
these two terms can be written as @xmath248 writing the neutral state @xmath58 in terms of its real and imaginary components , @xmath249 , we find a mass splitting between the components , @xmath250 \equiv m^2 + \frac{1}{2}v^2\lambda_n , \nonumber \\
m_{\zeta^{0,i}}^2 & = & m^2 + \frac{1}{2}v^2\left[\lambda_2 + \frac{1}{4}\lambda_3 + \frac{n}{2}(-1)^{n/2}\lambda_4\right ] = m_{\zeta^{0,r}}^2 + \frac{n}{2}(-1)^{n/2}v^2\lambda_4.\end{aligned}\ ] ] the mass matrices for the pairs of scalars with positive electric charge @xmath251 are given in the basis @xmath252 by @xmath253 which we diagonalize to find the mass eigenvalues , @xmath254 the mass eigenstates @xmath255 and @xmath256 are defined in terms of the weak eigenstates by eq .
( [ eq : alphadef ] ) such that @xmath255 is the lighter state and @xmath256 is the heavier state .
the mixing angle @xmath257 $ ] is given by @xmath258 there is only one state with @xmath259 .
its mass is given by @xmath260
for this study , it is convenient to describe the parameter space in terms of the mass differences @xmath261 and @xmath100 , since these two mass splittings are monotonic in @xmath262 and @xmath263 . in terms of the original lagrangian parameters ,
the mass splittings are given by @xmath264 } - { \ensuremath{m_{\zeta^{0,r}}}}\;.\label{appeq : dmn20}\end{aligned}\ ] ] we defined normalized versions of these splittings , each lying in the range @xmath265 $ ] , as @xmath266 the maximum value of @xmath102 allowed by the theoretical and indirect experimental constraints ( mainly from the electroweak oblique parameters ) , @xmath103 , can be written parametrically for the @xmath16 and @xmath18 models as @xmath267 the numerical values of @xmath268 for our chosen mass slices were given in table [ tbl : dm0max ] .
the relations in eqs .
( [ appeq : r1r2def ] ) , ( [ appeq : dm0 ] ) , and ( [ appeq : dmn20 ] ) can be inverted to obtain @xmath62 and @xmath46 as follows : @xmath269 , \\
\label{appeq : l4reparam}\lambda_4 & = ( -1)^{\frac{n}{2 } } \frac{2 r_2\ , { \ensuremath{\delta m_0^{\mathrm{max}}}}}{n\,v^2}\left[2 { \ensuremath{m_{\zeta^{0,r}}}}+ r_2\ , { \ensuremath{\delta m_0^{\mathrm{max}}}}\right]\;.\end{aligned}\ ] ] the physical masses can also be expressed as @xmath270 \delta m_{0}^{\mathrm{max } } } , \\
\label{appeq : newmasses } m_{\zeta^{+\frac{n}{2 } } } & = { \ensuremath{m_{\zeta^{0,r}}}}+ r_1 r_2 \delta m_{0}^{\mathrm{max}}\;,\end{aligned}\ ] ] where @xmath271 ^ 2}\;.\ ] ]
our models include parameter regions in which the spectrum is highly compressed . in such regions ,
simulation problems can arise because the automatic parton - level computation of partial widths and branching ratios using ` madwidth ` @xcite does not take into account hadronization effects arising from decays involving a highly off - shell @xmath6 or @xmath7 boson .
for example , consider the case of @xmath272 decay for a parameter point at which the mass splitting @xmath273 is larger than the mass of a @xmath274 but smaller than the mass difference of the @xmath275 system , as in table [ tab : spectra ] .
physically , a hadronically - decaying off - shell @xmath6 will emerge in this case as a single @xmath276 .
however , ` madwidth ` treats all decay products as bare partons and will decay the off - shell @xmath6 to a quark - antiquark pair . when ` pythia ` @xcite hadronizes the resulting quark - antiquark pair , it tries to create two pions
, finds that this is forbidden by the available phase space , and discards the event . c6emc7emc8em state & mass [ gev ] & @xmath277 [ gev ] + @xmath72 & 80.0000 &
+ @xmath278 & 80.1960 & 0.196 + @xmath279 & 80.8320 & 0.636 + @xmath280 & 82.1524 & 1.3204 + @xmath221 & 89.2535 & 7.1011 + @xmath281 & 89.8254 & 0.5719 + @xmath66 & 90.0000 & 0.1746 + this is a problem because it removes generated events that could otherwise potentially be detected via , e.g. , a hard initial - state jet radiation .
we solve this problem by manually removing any hadronic @xmath6 branching ratio from the ` madgraph ` ` paramcard ` files in which the @xmath6 invariant mass is less than the mass of two correspondingly - flavored mesons and rescaling the remaining @xmath6 decay branching ratios to sum to one .
should we worry about changing the potential detector signature of the very soft @xmath6 boson ?
a simple kinematical argument indicates no . in the rest frame of the decaying scalar ,
the momentum of the outgoing decay products must satisfy , e.g. , @xmath282 for the above example . in this frame
the decay product(s ) of the @xmath6 will have momentum less than the scalar mass splitting . to make the @xmath6 decay product(s ) visible in the detector
, they must be boosted to an ultrarelativistic velocity ; for example , a 10 gev charged pion requires a relativistic boost factor @xmath283 .
this requires that the parent scalar itself be boosted by the same relativistic boost factor : for a scalar mass around 80 gev , this implies a transverse momentum of more than 5 tev , which is not even kinematically possible at the 8 tev lhc .
99 g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 716 * , 1 ( 2012 ) [ arxiv:1207.7214 [ hep - ex ] ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 716 * , 30 ( 2012 ) [ arxiv:1207.7235 [ hep - ex ] ] . for a recent pedagogical review , see , e.g. , g. b. gelmini , arxiv:1502.01320 [ hep - ph ] . v. silveira and a. zee , phys .
b * 161 * , 136 ( 1985 ) ; m. j. g. veltman and f. j. yndurain , nucl .
b * 325 * , 1 ( 1989 ) ; j. mcdonald , phys .
d * 50 * , 3637 ( 1994 ) [ hep - ph/0702143 ] ; c. p. burgess , m. pospelov and t. ter veldhuis , nucl .
b * 619 * , 709 ( 2001 ) [ hep - ph/0011335 ] .
n. g. deshpande and e. ma , phys .
d * 18 * , 2574 ( 1978 ) .
t. araki , c. q. geng and k. i. nagao , phys .
d * 83 * , 075014 ( 2011 ) [ arxiv:1102.4906 [ hep - ph ] ] .
j. m. cline , k. kainulainen , p. scott and c. weniger , phys .
d * 88 * , 055025 ( 2013 ) [ arxiv:1306.4710 [ hep - ph ] ] ; g. belanger , b. dumont , a. goudelis , b. herrmann , s. kraml and d. sengupta , phys .
d * 91 * , 115011 ( 2015 ) [ arxiv:1503.07367 [ hep - ph ] ] .
s. y. ayazi and s. m. firouzabadi , jcap * 1411 * , 005 ( 2014 ) [ arxiv:1408.0654 [ hep - ph ] ]
. m. cirelli , n. fornengo and a. strumia , nucl .
b * 753 * , 178 ( 2006 ) [ hep - ph/0512090 ] ; m. cirelli , a. strumia and m. tamburini , nucl . phys .
b * 787 * , 152 ( 2007 ) [ arxiv:0706.4071 [ hep - ph ] ] ; m. cirelli and a. strumia , new j. phys . * 11 * , 105005 ( 2009 ) [ arxiv:0903.3381 [ hep - ph ] ] ; t. hambye , f .- s .
ling , l. lopez honorez and j. rocher , jhep * 0907 * , 090 ( 2009 ) [ erratum jhep * 1005 * , 066 ( 2010 ) ] [ arxiv:0903.4010 [ hep - ph ] ] . k. earl , k. hartling , h. e. logan and t. pilkington , phys .
d * 88 * , 015002 ( 2013 ) [ arxiv:1303.1244 [ hep - ph ] ] .
k. earl , k. hartling , h. e. logan and t. pilkington , phys .
d * 90 * , 055029 ( 2014 ) [ arxiv:1311.3656 [ hep - ph ] ] .
y. hamada , k. kawana and k. tsumura , phys .
b * 747 * , 238 ( 2015 ) [ arxiv:1505.01721 [ hep - ph ] ] .
k. hally , h. e. logan and t. pilkington , phys .
d * 85 * , 095017 ( 2012 ) [ arxiv:1202.5073 [ hep - ph ] ] .
s. s. abdussalam and t. a. chowdhury , jcap * 1405 * , 026 ( 2014 ) [ arxiv:1310.8152 [ hep - ph ] ] .
the atlas collaboration , atlas - conf-2013 - 049 .
the atlas collaboration , atlas - conf-2013 - 035 .
g. aad _ et al . _ [ atlas collaboration ] , jhep * 1404 * , 169 ( 2014 ) [ arxiv:1402.7029 [ hep - ex ] ] .
the atlas collaboration , atlas - conf-2013 - 036 . the atlas collaboration , atlas - conf-2013 - 089 .
the atlas collaboration , atlas - conf-2013 - 007 .
the atlas collaboration , atlas - conf-2013 - 028 .
the atlas collaboration , atlas - conf-2013 - 068 .
the atlas collaboration , atlas - conf-2012 - 147 .
the atlas collaboration , atlas - conf-2013 - 047 .
cms collaboration , cms - pas - sus-13 - 006 .
cms collaboration , cms - pas - sus-13 - 017 .
cms collaboration , cms - pas - exo-12 - 048 .
m. e. peskin and t. takeuchi , phys .
lett . * 65 * , 964 ( 1990 ) ; phys .
d * 46 * , 381 ( 1992 ) .
a. alloul , n. d. christensen , c. degrande , c. duhr and b. fuks , arxiv:1310.1921 [ hep - ph ] .
j. alwall , r. frederix , s. frixione , v. hirschi , f. maltoni , o. mattelaer , h .- s .
shao and t. stelzer _ et al .
_ , jhep * 1407 * , 079 ( 2014 ) [ arxiv:1405.0301 [ hep - ph ] ] . j. alwall , c. duhr , b. fuks , o. mattelaer , d. g. ozturk and c. h. shen , arxiv:1402.1178 [ hep - ph ] . t. a. w. martin and d. morrissey , jhep * 1412 * , 168 ( 2014 ) [ arxiv:1409.6322 [ hep - ph ] ] .
t. sjostrand , s. mrenna and p. z. skands , jhep * 0605 * , 026 ( 2006 ) [ hep - ph/0603175 ] .
j. de favereau _ et al . _
[ delphes 3 collaboration ] , jhep * 1402 * , 057 ( 2014 ) [ arxiv:1307.6346 [ hep - ex ] ] .
t. a. w. martin , arxiv:1503.03073 [ hep - ph ] .
the atlas collaboration , atl - phys - pub-2013 - 004 j. hisano and k. tsumura , phys .
d * 87 * , 053004 ( 2013 ) [ arxiv:1301.6455 [ hep - ph ] ] ; s. kanemura , m. kikuchi and k. yagyu , phys .
d * 88 * , 015020 ( 2013 ) [ arxiv:1301.7303 [ hep - ph ] ] . c. alvarado , l. lehman and b. ostdiek , jhep * 1405 * , 150 ( 2014 ) [ arxiv:1404.3208 [ hep - ph ] ] .
h. baer , t. barklow , k. fujii , y. gao , a. hoang , s. kanemura , j. list and h. e. logan _ et al .
_ , arxiv:1306.6352 [ hep - ph ] . | we study the lhc search constraints on models that extend the standard model ( sm ) with an inert , complex scalar electroweak multiplet , @xmath0 , with isospin @xmath1 ( sextet ) or @xmath2 ( octet ) and identical hypercharge to the sm higgs doublet . imposing a global @xmath3 symmetry under which @xmath4 , the lightest member of @xmath0 is stable and we require that it be neutral ( @xmath5 ) to avoid cosmological constraints from charged relics .
pair production of scalars by electroweak interactions followed by cascade decays to @xmath5 through @xmath6 and @xmath7 emission produces signatures similar to those of supersymmetric electroweak gauginos , and we constrain the models by recasting a collection of such searches made with data from the 8 tev run of the lhc . we find that there is no sensitivity from these searches to the compressed spectrum regime , in which the mass splittings between the lightest and heaviest states in @xmath0 are less than about 20 gev . in the remaining parameter space , we find significant exclusions for @xmath8 gev in the sextet model , and @xmath9 gev in the octet model . |
in the past decades , searching for the mysterious particle , majorana fermion , proposed by ettore majorana in 1937 @xcite has continued to be an important issue in high - energy physics .
recently , the field of topological superconductors and superfluids@xcite in condensed matter physics has been drawing much attention since these topological matters are capable of hosting majorana fermions . instead of seeking for it in high - energy experiments , topological matters in condensed matter systems may shed light on finding this elusive majorana fermion .
majorana fermion , half state of a fermion , obeys non - abelian statistics and possesses the non - local property .
both of the significant ingredients render such topological matters a great potential to realize quantum computers @xmath12 the dream which scientists have been pursuing for over 30 years .
topological matters are classified by distinct topological numbers based on dimension and symmetries of the system . in analogous to the framework of topological insulators characterized by a @xmath13 invariance , non - centrosymmetric superconductors which break inversion symmetry but preserve time reversal symmetry are generally termed as time - reversal - invariant ( tri ) superconductors .
inversion symmetry breaking leads to the mixing of spin - singlet and spin - triplet cooper pairings@xcite . as a result
, intensive studies have been dedicated to investigate fully gapped tri superconductors for their potential to be a topological nontrivial matter , termed as tri topological superconductors@xcite . in order to have fully gapped superconductors , many proposals for creating majorana fermions rely on the proximity effect of @xmath14-wave superconductors@xcite .
the disadvantage of these approaches lies in the fact that most @xmath14-wave superconductors have low critical temperatures and majorana fermions can be generated only in very low temperatures .
it is therefore desirable to seek alternative approaches based on high-@xmath0 cuprate superconductors . in real materials , the heavy - fermion superconductor , @xmath15 @xcite
, discovered in the early days and the recent finding of superconductivity at the interface between @xmath16 and @xmath17@xcite are typical examples of tri superconductors in which spin - orbit interaction comes into play due to lack of inversion center .
recently , it is argued that the contact of heavy metals with superconductors can induce the rashba spin - orbit interaction in the interface@xcite . in particular , for au / ybco interface , it is estimated that the rashba spin splitting energy at fermi surface is of the order 200 mev@xcite .
there is also experimental evidence showing that there may exist intrinsic spin - orbit interaction in cuprate superconductors@xcite . from theoretical point of view , however , the strength of available spin orbit interaction is too small to turn a @xmath3-wave cuprate superconductor into a fully gapped tri superconductor
. therefore , even though the ideal scheme is to generate majarona fermions via tri topological superconductors , its feasibility is low in high @xmath0 cuprates and it is necessary to seek for alternative approaches . in this paper , we investigate effects of the rashba spin - orbit interaction on high-@xmath0 cuprate superconductors .
based on the mean - field theory of _ t - j model@xcite , we show that the gap function @xmath3-vector for the induced spin - triplet pairing is generally not aligned with the internal magnetic field of the spin - orbit interaction .
this fact respects the @xmath18 point group symmetry and for typical strength of the rashba interaction , the superconductor is still gapless with each nodal point of the original @xmath3-wave superconducting state being split .
the resulting nodal points are protected by symmetries in the symmetry class diii@xcite .
furthermore , due to the splitting nodal structures , zero - energy majorana edge modes that are determined by projection of bulk nodes to edges@xcite always exist for any orientations of interfaces that are not exactly in @xmath19 $ ] or @xmath20 $ ] directions .
in particular , near @xmath21 $ ] direction in which @xmath3-wave is subdominant to @xmath2-wave , there exist sizable dispersive majorana edge states .
our results indicate that the presence of these majorana modes gives rises to a small plateau in tunneling spectrum near zero bias peak .
furthermore , these majorana modes always result in 4@xmath7-josephson effect in typical junctions and hence flux trapped in tricrystal experiments jump in unit of two flux quanta .
these phenomena may have been already observed in experiments and their connections to other experiments will be discussed .
_ this paper is organized as follows . in sec .
[ model ] , by including the rashba spin - orbit interaction , we derive superconducting phase diagram based on the mean - field theory of the effective low - energy hamiltonian for the strong coupling limit of the hubbard model . in sec .
[ gapless ] , the gapless regime of @xmath22 wave resulted from typical strength of rashba interaction is investigated .
edge states and their relations to nodal point structures are examined . in particular , how the edge states of @xmath6 edge change in the presence of the rashba spin - orbit interaction changes is also examined . in sec .
[ tunnel ] , we examine the tunneling spectrum for the nis junction and periodicity of josephson current in a sis@xmath8 junction .
the contribution due to the dispersive and flat - band majorana modes are examined .
furthermore , the relation of 4@xmath7 periodicity and flat - band majorana modes are derived for josephson junctions .
the conditions for the @xmath7-ring in the tricrystal experiment to hold majorana fermions are also examined in details . finally , in sec .
[ conclusion ] , we conclude and discuss possible connections of our results to experimental observations .
cuprate superconductors in contact with other materials ( electrode or substrate ) . through the interface between the cuprate and the underlying material , spin - orbit interactions that break inversion symmetry
are induced in the cuprate superconductor either through proximity effect or in the layer near the interface . here
the cuprate superconductor can make further tunnel junction with other superconductors via the edge contact.,title="fig:",width=170 ] +
the parent compound of high-@xmath0 cuprate superconductor is well - known to be an antiferromagnetic mott insulator . in spite of the fact that a complete theory for high-@xmath0 superconductors is under debate , a common consensus for grasping the essential physics behind the emergence of superconductivity lies in the 2d hubbard model within large-_u limit , or the equivalent _ t - j model@xcite .
however , in typical experiments , high-@xmath0 cuprate superconductors are often in the form of thin films with contact to other materials along @xmath23-axis as illustrated in fig .
[ fig1 ] . through the interface between the cuprate and the underlying material , spin - orbit interactions that break inversion symmetry
may be induced either in the cuprate superconductor by proximity effect or induced in the layer near the interface . in particular , for au / ybco interface , it is estimated that the energy splitting due to the spin - orbit interaction is of the order 0.2 ev@xcite .
the largest splitting reported is also around 0.3ev@xcite .
it is therefore necessary to include effects due to the spin - orbit interaction up to the spitting energy being around 0.3ev . _
_ in the presence of the rashba spin - orbit interaction , however , there is no full understanding of corresponding superconducting phases . from the symmetry point of view , the rashba spin - orbit interaction breaks the inversion symmetry . as a result
, parity is not conserved and hence it is generally true that singlet pairing will be mixed with the triplet pairing in any superconducting phases . to be explicit , we shall start from the 2d hubbard model with the rashba spin - orbit interaction .
the effect of dispersion along @xmath23-axis will be examined at the end of sec .
iv . in the strong coupling limit ,
exchange interactions give arise an extended @xmath24 model which can be generally expressed as @xmath25 with @xmath26_{\alpha \beta } \cdot\hat{z } ( \tilde{c}^\dagger_{i,\alpha}\tilde{c}_{j,\beta } ) + h.c .
, \nonumber \\ \\
\nonumber h_j&=&h_{j_1}+h_{j_2}+h_{j_3},\\ \nonumber&= & \sum_{<ij>}j_1\left(\mathbf{s}_i\cdot \mathbf{s}_j-\frac{1}{4}n_in_j\right)+j_2\left[\mathbf{d } \cdot\left(\mathbf{s}_i\times \mathbf{s}_j\right)\right],\\ & + & j_3\sum_{<ij>}\left[\mathbf{s}_{i}\cdot\mathbf{s}_{j}+\frac{1}{4}n_in_j-2(\mathbf{d } \cdot \mathbf{s}_i)(\mathbf{d } \cdot \mathbf{s}_j)\right ] .
\nonumber \\ \end{aligned}\ ] ] here @xmath27 is the lattice constant .
@xmath28 , @xmath29 , and @xmath30 for sites @xmath31 and @xmath32 are nearest , next nearest , and the third nearest neighbors , respectively , and @xmath33 for longer distance . @xmath1
characterizes the energy strength of the rashba spin - orbit interaction .
@xmath34 satisfies the no - double - occupancy constraint . for electrons in terms of the on - site coulomb interaction ( hubbard @xmath35 ) ,
three spin interactions are given by @xmath36 , @xmath37 and @xmath38 , representing the heisenberg interaction , the dzyaloshinskii - moriya ( dm ) interaction and the spin dipole - dipole interaction respectively .
the orbital vector @xmath39 that couples to spins is related to the lattice constant @xmath27 connecting sites in nearest neighbors along @xmath40(@xmath41 ) direction by @xmath42 . to satisfy the no - double - occupancy constraint
, one can resort to the slave - boson method@xcite or adopt gutzwiller approximations by using renormalized parameters@xcite .
in low doping regime , both approaches yield the same mean - field hamiltonian .
we shall adopt the gutzwiller approximation . in the gutzwiller approximations ,
strength of spin interactions remain the same in the low doping regime .
furthermore , @xmath43 is replaced by @xmath44 .
the spin interactions can be generally factorized as summations of particle - particle channels and particle - hole channels@xcite . to investigate superconducting phases , one focuses on particle - particle channels .
we find that the @xmath45 term yields the spin - singlet pairing channel .
by contrast , the @xmath46 term contains the spin - triplet pairing channel , and the @xmath47 term mixes the spin - singlet and spin - triplet pairing channels .
specifically , the pairing channels can be written as @xmath48 where the singlet pairing field @xmath49 , the spin - triplet vector field @xmath50 and @xmath51 .
the vector @xmath52 are given by @xmath53 and @xmath54 .
[ table : c4 ] .
the representations of @xmath18 [ cols="^,^,^,^,^",options="header " , ] in addition to superconducting channels , we also include particle - hole chanels and set @xmath55 and @xmath56 with @xmath57 or @xmath41 .
when the hole concentration is @xmath58 , the resulting mean - field hamiltonian @xmath59 can be written as @xmath60 where the lattice constant @xmath27 is set to be 1 , @xmath61-\mu$ ] and @xmath62 with @xmath63 and @xmath64 .
the pairing amplitudes @xmath65 and @xmath66 represent amplitudes for singlet and triplet pairing with corresponding gap functions @xmath67 and @xmath68 respectively .
the allowed representations for gap functions are tabulated in the table [ table : c4 ] . in the absence of the rashba spin - orbit interaction ,
it is known that @xmath3-wave in @xmath69 representation and @xmath14-wave in @xmath70 representation are two superconducting states with lowest energies . in the presence of the rashba spin - orbit interaction
, we find that @xmath71 and @xmath72 representations become the lowest two in energy due to mixing of the singlet and triplet pairings .
for @xmath73 pairing symmetries in @xmath71 representation , we have @xmath74 and @xmath75 . on the other hand , for @xmath22 pairing symmetries in @xmath72 representation , we have @xmath76 and @xmath77 .
the pairing amplitudes are related to the pairing fields by @xmath78 and @xmath79 ( @xmath12 and @xmath80 refer to @xmath71 and @xmath72 representation respectively ) with the order parameters being given by @xmath81 and @xmath82 .
the mean - field parameters are solved using the following self - consistent equations @xmath83 where @xmath84 and @xmath85 are computed with respect to the ground state of eq.([mf ] ) .
note that the triplet pairing can be cast into a gap function as @xmath86 with @xmath87 being the @xmath3-vector . in the @xmath72 representation ,
the gap function @xmath87 vector is not aligned with the internal magnetic field @xmath88 .
we find that the non - aligned @xmath3-vector lower the ground state energy in consistent with previous studies@xcite .
when @xmath46 dominates , we find that the spin - triplet pairing wins over so that @xmath71 representation dominates .
hence as @xmath46 increases , superconducting phases will go through a transition from @xmath22 wave ( @xmath72 ) to @xmath73 wave ( @xmath71 ) . in the @xmath71 representation ,
@xmath87 vector is parallel to @xmath88 vector , the ground state energy is further lowered down in comparison to the @xmath72 representation . wave ( small @xmath46 ) to @xmath73 wave ( large @xmath46 ) . here
@xmath89 and @xmath90 .
unless @xmath46 is very large , for typical strength of rashba interaction , the pairing symmetry is @xmath22 wave . ,
title="fig:",width=264 ] + fig .
[ fig2 ] shows the phase boundary between @xmath22 wave ( @xmath72 ) to @xmath73 wave ( @xmath71 ) for different dopings at a fixed @xmath47 .
the upper left corner of each is @xmath22 wave while the lower right corner is @xmath73 wave .
it is clear that in the strong coupling limit , the pairing symmetry is @xmath73 wave in @xmath71 only when @xmath46 is at least @xmath91 times @xmath92 for @xmath36 . for for typical strength of rashba interaction ,
the pairing symmetry is @xmath22 wave .
[ fig3 ] examines how @xmath3-wave and @xmath2-wave amplitudes change as the rashba interaction @xmath1 increases .
due to that the @xmath46 term provides spin - triplet pairing channel and @xmath93 , the @xmath2-wave amplitude increases quadratically , @xmath94 .
however , for typical strength of @xmath1 , the amplitude of @xmath2-wave is still one order of magnitude less that that of @xmath3-wave .
on the other hand , the amplitude of @xmath3-wave also increases linearly with @xmath1 due to the coupling of @xmath3-wave to @xmath2-wave via the @xmath47 term , which is proportional to @xmath1 . as a result ,
@xmath3-wave component is always larger than that of @xmath2-wave component .
hence for realistic parameter regimes , the resulting superconducting state is still dominated by @xmath3-wave component and the ground state is still in gapless phase .
as an example , we show the boundary between gapped phase and gapless phase for the largest reported value of @xmath1 in fig .
it is seen that for realistic parameters @xmath95 , the superconducting phase is gapless .
-wave and @xmath2-wave versus rashba spin - orbit coupling strength . here
parameters are @xmath96 in units of @xmath97.,title="fig:",width=321 ] + ( @xmath27 is the lattice constant ) . here
@xmath98 is fixed at @xmath99 .
it is seen that for realistic parameters @xmath95 , the superconducting phase is gapless.,width=264 ]
in this section , we investigate the gapless regime of @xmath22 wave .
this is the regime that typical strength of rashba interaction lies in .
it is shown that nodal points split and the resulting superconducting state always supports edge states for any interfaces that are not exactly aligned in @xmath4 or @xmath5 directions .
we start from the general bulk hamiltonian with @xmath22 mixed pairing symmetry and rashba interaction in the nambu basis @xmath100 , where @xmath101 . the @xmath102 matrix form of @xmath103 is given by @xmath104 where the lattice constant @xmath27 is set to be 1 , @xmath105 , the kinetic energy is @xmath106 , and @xmath107 is the singlet - pairing @xmath3-wave gap function . the gap function can be generally combined as @xmath108 with @xmath109 and @xmath110 .
the energy spectra @xmath111 can be solved and are given by @xmath112 where @xmath113 , @xmath114 and @xmath115 .
, open circles are nodal points for @xmath116 and when @xmath117 , nodal points in each quadrant recombine.,width=283 ] in the absence of the rashba spin - orbit interaction and the @xmath2-wave , the hamiltonian @xmath118 supports 4 dirac points determined @xmath119 and @xmath120 . the @xmath121 expansion of @xmath122 about one dirac point @xmath123 can be generally written as @xmath124 where @xmath125 and coordinate axes are along diagonal axes with @xmath126 and @xmath127 . for @xmath123 in the first quadrant , one finds that @xmath128 . in the diagonal coordinates , we have @xmath129 , @xmath130 , and @xmath131 . here
@xmath132 is the pseudospin vector in the particle - hole space . in analogy to the characterization of 3d weyl semimetals@xcite and weyl fermions in ferromagnetic superconductors@xcite , @xmath133 and @xmath134 can be extended to be a complete set of @xmath135 matrices in su(4 ) representation @xmath136 with @xmath137 . if we redefine operations of time - reversal and inversion symmetries as operators with respect to @xmath138 for small @xmath139 , the rashba spin - orbit interaction and @xmath2-wave order parameter break time - reversal symmetry ( operator represented by @xmath140 ) and inversion symmetry ( operator represented by @xmath141 ) . in this case , perturbations due to broken time - reversal and inversion symmetries can be generally expressed in terms of @xmath142 matrices defined by @xmath143 $ ] @xcite .
we find that @xmath144 where @xmath145 , @xmath146 , @xmath147 , and @xmath148 .
the corresponding energy spectrum @xmath111 of eq.([perturb ] ) becomes @xmath149 and is given by @xmath150 where we have defined @xmath151 and @xmath152 .
clearly , when @xmath153 , @xmath149 no longer vanishes .
hence the original nodal points are lifted .
similar to the 3d weyl fermions , the main effect of the rashba interaction and the @xmath2-wave is to split each dirac point into two weyl nodal points in 2d .
specifically , @xmath103 obtained in eq.([perturb ] ) corresponds to the case @xmath154 for 3d weyl semimetal@xcite .
we find that for @xmath155 , @xmath149 vanishes at @xmath156 and @xmath157 ; while for @xmath158 , @xmath149 vanishes at @xmath159 and @xmath160 . the transformation of nodal point structure is illustrated in fig .
[ fig5 ] . for a fixed @xmath1 ,
as @xmath161 increases , a transition of nodal point configuration occurs at @xmath117 before the ground state becomes a full gapped tri superconducting state . the transition of nodal point structure results from competition between @xmath3-wave and @xmath2-wave pairing symmetries . for small @xmath2-wave amplitudes
when @xmath162 , the superconducting state is dominated by the @xmath3-wave symmetry . hence nodal points are determined by nodal lines of @xmath3-wave order parameter , which are along axes of @xmath163 .
the intersections of splitting fermi surfaces by @xmath1 and nodal lines of @xmath3-wave order parameter determines nodal points along @xmath163 . for large @xmath2-wave amplitudes when @xmath116 , nodal points are determined by nodal lines of @xmath2-wave order parameter , which are @xmath164 or @xmath165 . as a result ,
nodal points are along the @xmath166(@xmath167 ) axis in perpendicular to the axes of @xmath168 ( @xmath169 ) .
the nodal points revealed in the last subsection carry winding numbers .
the associated winding numbers with nodes labeled by @xmath170 are shown in fig .
when positions of nodal points change , the distribution of winding numbers also changes .
it is known that positions of nodal points determine edge states . according to refs . and ,
for 3d bulk states , as long as the projection of opposite - winding - number nodal manifolds does not completely overlap in the boundary brillouin zone , there will be gapless surface bound states .
for 2d bulk states , the gapless surface bound states become zero - energy edge states . therefore , for a given edge , the projection of nodal point to the edge determines zero - energy edge states . for ( a ) @xmath162 and ( b ) @xmath116 .
here each nodal point is associated with a branch cut in @xmath171 direction with phase jump being @xmath172 . due to the cancellation of phase jumps , only lines that connect @xmath80 and @xmath12 are branch cuts with non - vanishing phase jumps .
the projection of each line connecting @xmath80 and @xmath12 onto the edge @xmath173 axis gives rise to one zero - energy flat band at the projected range of wavevector @xmath174 .
absolute values of numbers ( @xmath175 or @xmath176 ) shown in the edge @xmath173 axis are numbers of flat - bands in the corresponding @xmath174 range .
it is seen that for realistic rashba strength , @xmath162 , one always gets single zero - energy mode for certain wavevector @xmath174 along the edge @xmath173.,title="fig:",width=158 ] for ( a ) @xmath162 and ( b ) @xmath116 . here
each nodal point is associated with a branch cut in @xmath171 direction with phase jump being @xmath172 . due to the cancellation of phase jumps , only lines that connect @xmath80 and @xmath12 are branch cuts with non - vanishing phase jumps .
the projection of each line connecting @xmath80 and @xmath12 onto the edge @xmath173 axis gives rise to one zero - energy flat band at the projected range of wavevector @xmath174 .
absolute values of numbers ( @xmath175 or @xmath176 ) shown in the edge @xmath173 axis are numbers of flat - bands in the corresponding @xmath174 range .
it is seen that for realistic rashba strength , @xmath162 , one always gets single zero - energy mode for certain wavevector @xmath174 along the edge @xmath173.,title="fig:",width=158 ] fig .
[ fig6 ] illustrates the projection of nodal points to edges . for each nodal point , one associates a branch cut with phase jump being @xmath172 . in fig .
[ fig6 ] , branch cuts are chosen in @xmath171 direction .
due to the cancellation of phase jumps , only lines that connect @xmath80 and @xmath12 are branch cuts with non - vanishing phase jumps .
here the phase jumps for the line connecting @xmath80 to @xmath12 and that for @xmath12 to @xmath80 are opposite in sign . by projecting branch cuts onto the momentum line ( denoted by @xmath174 ) that represents the edge ,
it is clear that for @xmath162 shown in fig .
[ fig6]a , the nodal manifolds for @xmath170 nodes do not overlap completely except when orientations of edges are exactly along @xmath4 or @xmath5 directions . therefore , for @xmath174 in the non - overlapped region , there always exists single zero - energy mode along the edge .
the single zero - energy mode is a majorana mode .
more specifically , the projection of each connection of @xmath80 and @xmath12 to a given edge @xmath173 gives rise to one zero - energy mode for each wavevector @xmath174 along @xmath173 that lies in the non - overlapped region .
it thus results in a flat band at zero - energy@xcite .
therefore , in the overlapped region , there will be two flat - bands ; while in the non - overlapped region , there is only one branch flat - band which is a majorana flat - band . as a result
, we find that except when the edge is exactly along @xmath4 or @xmath5 directions , in general , there always exist majorana modes for any edges .
since real interfaces always possess facets in different orientations , existence of majorana modes appear to be a robust feature . as the amplitude of @xmath2-wave increases until @xmath177 , nodal points switch to be in the @xmath178 direction . as illustrated in fig .
[ fig6]b , the nodal manifolds for @xmath170 nodes do not overlap completely except when orientations of edges are exactly along @xmath6 or @xmath179 directions . therefore , except when the edge is exactly along @xmath6 or @xmath179 directions , in general , there always exists majorana modes for any edges .
note that in the case when @xmath177 , there is a region near @xmath180 for @xmath6 edge in which as we shall show in below , the connection of nodal points with the same winding number leads to dispersive majorana modes . in this subsection
, we shall verify connections of edge states to the nodal structure for the particular orientation of edge @xmath6 .
the @xmath6 edge is known to be the most important edge that exhibits zero - bias peak in the tunneling spectrum due to the existence of andreev bound states at zero energy@xcite .
it is therefore important to examine how edge states change when the rashba spin - orbital interaction is included .
we shall start by partial fourier transforming the hamiltonian along the @xmath6 interface direction .
let @xmath174 be the fourier wavevector along @xmath6 direction .
the hamiltonian can be expressed as @xmath181 and is characterized by an one - dimensional hamiltonian @xmath182 for a given @xmath174 , which is given by @xmath183 where the hopping and pairing matrices @xmath184 and @xmath185 are given by @xmath186 the block matrices @xmath184 are expressed as @xmath187 with @xmath188 and @xmath189 .
the pairing matrices satisfy @xmath190 and @xmath191 with @xmath192 here elements in matrices are given by @xmath193 , @xmath194 , @xmath195 , @xmath196 , @xmath197 , @xmath198 , @xmath199 , @xmath200 with @xmath201 and @xmath202 . edge .
( b)-(e ) evolution of quasi - particle spectrum and edge states for @xmath6 edge of cuprates in the gapless phase when @xmath1 changes . here
the mean - field solution corresponds to @xmath204 with @xmath205 in ( b ) .
edge states are labelled as blue and red solid lines with red lines representing majorana edge states .
it is clear that majorana flat - bands at zero energy exist only for physical accessible region @xmath206 ; while the dispersive majorana modes exist near @xmath180 only for @xmath207.,title="fig:",width=321 ] edge .
( b)-(e ) evolution of quasi - particle spectrum and edge states for @xmath6 edge of cuprates in the gapless phase when @xmath1 changes . here
the mean - field solution corresponds to @xmath204 with @xmath205 in ( b ) .
edge states are labelled as blue and red solid lines with red lines representing majorana edge states .
it is clear that majorana flat - bands at zero energy exist only for physical accessible region @xmath206 ; while the dispersive majorana modes exist near @xmath180 only for @xmath207.,title="fig:",width=321 ] edge in the gapless phase .
here @xmath208 , and @xmath209 with @xmath210 and @xmath10 being holding fixed . as shown in ( a)-(c ) ,
as @xmath211 increases ( @xmath212 , @xmath213 and @xmath214 ) , @xmath170 nodal points from different quadrants move toward each other and annihilate . correspondingly , as shown in ( d)-(f ) , the region with dispersive edge modes around @xmath180 expands and the region ( labelled by @xmath175 ) with fermionic andreev bound states shrinks .
it is clear that dispersive edge modes near @xmath180 evolve into helical majorana modes of the fully gapped @xmath2-wave superconductors.,width=302 ] we shall focus on the case when the superconducting side is semi - infinite . for a given @xmath174 ,
the system becomes one - dimensional and its topology can be characterized by the winding number@xcite .
here we extend the winding number calculation@xcite to any edge that makes an angle @xmath215 clock - wisely with respect to @xmath4 direction .
the winding number @xmath216 is then given by @xmath217 here @xmath218 is the wavevector along the edge , @xmath219 is the wavevector perpendicular to the edge , and the matrix @xmath220 is given by @xmath221 .
@xmath222 and @xmath223 are the real and imaginary parts of the phase factor of @xmath224 respectively .
for a give @xmath174 , if number of sites in perpendicular to @xmath6 edge is @xmath225 , @xmath182 is a @xmath226 matrix and can be exactly diagonalized .
for large @xmath225 , one obtains numerical solutions for semi - infinite superconducting state . in fig .
[ fig7](a ) , we show how the quasi - particle spectrum for the @xmath6 edge changes as the rashba spin - orbit interaction @xmath1 varies . starting from the mean - field solution with @xmath227 shown in fig .
[ fig7](a ) , there are two majorana flat - bands with winding number @xmath228 at @xmath229 .
the decreasing of @xmath1 leads to the reduction of the majorana flat - band regimes as shown fig .
[ fig7](b ) .
in particular , as indicated in fig . [ fig7](c ) and ( d )
, majorana flat band disappears for @xmath230 ( i.e. @xmath231 ) .
instead , we see that dispersive edge states occur near @xmath180 with @xmath230 . to reveal the nature of the dispersive edge states , we fix @xmath1 and varies @xmath161 . as illustrated in fig . [ fig7 ] ( a)-(c ) , when the amplitude @xmath161 increases , @xmath170 nodal points from different quadrants move toward each other and annihilate .
correspondingly , as shown in ( d)-(f ) , the region with dispersive edge modes around @xmath180 expands and the region with fermionic andreev bound states shrinks .
it is clear that dispersive edge modes near @xmath180 evolve into helical majorana modes of the fully gapped @xmath2-wave superconductors . therefore , when @xmath232 , dispersive edge states around @xmath180 emerge and one expects that they are of the same nature as helical majorana modes of the fully gapped @xmath2-wave superconductors . to examine
if these dispersive edge mode are indeed majorana modes , one needs to examine if the quasi - particle obeys @xmath233@xcite . in general , the eigenstate at @xmath31 site to the one - dimensional hamiltonian @xmath182 is represented by a 4-component @xmath234 so that the quasi - particle is given by @xmath235 for majorana modes , one requires @xmath236 , which in turn requires @xmath237 and @xmath238 .
however , due to an uncertainty in defining @xmath239 and @xmath240 up to a @xmath241 global phase factor , in order for @xmath135 to be majorana modes , one requires @xmath242 therefore , @xmath243 is the condition for edge modes to be majorana modes . in fig .
[ fig9 ] , we examine energy dispersions for @xmath6 edge near @xmath180 . as shown in the inset , @xmath243 is satisfied for dispersive edge modes . hence when @xmath232 , the emerging dispersive edge states around @xmath180 are dispersive majorana modes .
edge near @xmath180 . here
solid lines represent bulk states .
filled circles represent edge states .
lower panel : the majorana condition , @xmath244 , is checked by examining @xmath243
. clearly , the dispersive edge states are majorana modes.,width=302 ]
the tunneling spectroscopy is one of means to examine majorana fermions .
a typical experiment is the measurement of the tunneling spectrum for normal metal @xmath245 , insulator @xmath246 and superconductor @xmath247 junction , i.e. , the nis junction , in which zero - energy edge states would appear as a zero - bias conductance peak .
however , the zero - bias conductance peak may also arise from other mechanism@xcite and hence it is not considered as a smoking - gun evidence for majorana fermions . on the other hand ,
the @xmath248 periodicity of supercurrent flowing across the josephson junction is considered as a smoking - gun evidence for existence of majorana fermions . in this section , we shall examine both the spectrum of the nis junction and josephson junction for high @xmath0 cuprates in the presence of rashba spin - orbit interaction .
it will be shown that dispersive majorana modes would exhibit a unique feature as a small plateau in tunneling spectrum near zero bias peak .
furthermore , we find that the flat - band majorana modes always result in 4@xmath7-josephson effect in typical josephson junctions . therefore , by suitable designing junctions involved in the tricrystal experiments , the flux trapped in the center of the tricrystal jumps in unit of two flux quanta , which would be a convincing way to detect majorana fermions .
the well - known zero - bias conductance peak observed in high-@xmath0 cuprates along @xmath6 direction manifests the fermionic andreev bound states at the interface of nis junctions@xcite .
as indicated in the sec .
[ gapless ] c , in the presence of the rashba spin - orbit interaction in high-@xmath0 cuprates , @xmath6 edge also hosts majorana flat bands when @xmath2-wave is small .
the detection of this majorana flat bands is thus ambiguous in the nis planar junctions due to the coexistence of zero - energy flat bands formed by fermionic and majorana edge states .
both of them contribute to the zero - bias peak based in the blonder - tinkham - klapwijk theory @xcite . on the other hand ,
the dispersive majorana fermions occur at finite energy scale and hence its appearance would be unambiguous . for a planar nis junction , after partial fourier transformation along the edge of the junction with @xmath174 being the wavevector along the edge ,
the junction is effectively a one - dimensional system . for a give @xmath174
, the effective one - dimensional hamiltonian for the superconducting side with @xmath6 edge is given by eq.([h1d ] ) , @xmath249 .
following blonder , tinkham , and klapwijk@xcite , the nis junction is formed by connecting @xmath249 with the effective 1d hamiltonian of the normal metal side . if the one dimensional eigenstate @xmath250 at @xmath32 site is represented by a 4-component @xmath251 .
the tight - binding bogoliubov equation can be generally written as @xmath252 where the potential @xmath253 is the tunneling barrier at the interface due to the insulator layer , the superconductor @xmath254 occupies @xmath255 and the normal metal @xmath225 occupies @xmath256 .
matrices @xmath257 and @xmath258 denotes effective nearest neighbor and next nearest neighbor hopping , while @xmath259 represents the effective on - site potential .
they are given by @xmath260 here for the metal side , corresponding matrices are denoted by @xmath261 and @xmath262 .
@xmath263 , @xmath264 , @xmath265 , and @xmath266 are set to zero , while @xmath267 , @xmath268 , @xmath269 , and @xmath270 are set to be hopping amplitudes @xmath271 , @xmath272 , @xmath273 , and @xmath274 .
for the superconducting side , corresponding matrices are denoted by @xmath275 , @xmath276 and @xmath277 with @xmath278 , @xmath279 , @xmath280 , @xmath281 , @xmath282 , @xmath283 , @xmath284 and @xmath285 . and lower panel is for @xmath207 . in the n side , solid lines and dash lines represent spectra of particles and holes respectively .
different colors ( black and blue ) stand for two bands split by the rashba interaction .
black and blue dots with arrows represent incident and reflected particles , while open circles denote the reflected holes . in the s side for @xmath286 ,
gray dots stand for the particle - like and open circles denote the hole - like transmitted modes . for @xmath287 ,
only two nodal points are present in @xmath254 side .
specifically , two evanescent modes represented by dash arrows can be found by solving @xmath288 with @xmath286 . here , @xmath289 is the lowest positive @xmath290 in eq.([ek]).,width=340 ] at a given energy @xmath291 , far away from the interface , the quasi - particle wavefunction @xmath251 satisfies the bulk bogoliubov equation . as illustrated in fig .
[ fig10 ] , there are two incident waves ( indicated as black and blue dots ) in the n side , which will be denoted by @xmath292 and @xmath293 . for reflected waves
, there are 2 bulk quasi - particle wavefunctions indicated by black and blue dots and 2 bulk quasi - hole wavefunctions indicated by open circles .
the corresponding normalized wavefunctions are denoted by @xmath294 , @xmath295 , @xmath296 , and @xmath297 .
similarly , for the s side , there are 2 particle - like and 2 hole - like wavefunctions denoted as @xmath298 , @xmath299 , @xmath300 , and @xmath301 . by using these wavefunctions ,
@xmath302 can be expressed as @xmath303 here indices with @xmath304 and @xmath305 represent quantities of incident waves .
the wavevectors @xmath306 and @xmath307 in the corresponding quantities @xmath308 and @xmath309 are perpendicular to the interface in the @xmath254 side and the @xmath225 side respectively . @xmath306 and @xmath307 must be solved from the bulk energy dispersion @xmath288 . @xmath310 and @xmath311 are corresponding transmission and reflection amplitudes for @xmath312 and @xmath313 . solutions for @xmath255 and @xmath256 with forms given by eq.([psi ] ) are then joined together at the interface @xmath314 . by comparing eq.([bogoliubov ] ) at @xmath315 and @xmath316 with eq.([bogoliubov ] ) at @xmath317 and @xmath318 , we find that the boundary conditions are given by @xmath319 after re - arrangements of the above two boundary conditions , amplitudes @xmath310 and @xmath311 can be solved by the following @xmath320 linear equation @xmath321 where @xmath322 . for a given @xmath174 ,
there is a minimum of bulk energy at @xmath323 as illustrated in fig .
[ fig10 ] . when the energy @xmath291 of an injected particle with @xmath286 , @xmath306 and @xmath307 are both real .
we obtain two transmitted electron - like and two hole - like quasiparticles in the @xmath254 side and two reflected electrons and holes in the @xmath225 side . on the other hand , for @xmath324 , two electron - like and two hole - like in the @xmath254 side become evanescent modes with @xmath325 and @xmath326 being complex numbers .
note that for @xmath327 , even for @xmath286 , since there are only two nodal points , both transmitted and evanescent modes are present in the @xmath254 side .
solving eq .
( [ andreev_eq ] ) , one obtains solutions for @xmath310 and @xmath311 which then determine the transmitted and reflected particle currents given by @xmath328 in the @xmath225 side and similar expression for the @xmath254 side with @xmath261 and @xmath329 replaced by @xmath275 and @xmath330 . hence we obtain the particle current @xmath331 through the tunneling barrier @xmath332 in the @xmath225 side and the reflected particle and hole currents : @xmath333 and @xmath334 .
the normalized tunneling conductance for the nis junction is then given by@xcite @xmath335 where @xmath336 is the ballistic conductance of the @xmath225 side and @xmath337 is defined to be @xmath338/j^e_{i}(e , p_y)$ ] . in fig .
[ fig11 ] , we show the computed tunneling conductance . in the strong interface scattering limit @xmath339 ,
the tunneling conductance is proportional to local density of state @xmath340 at @xmath6 edge@xcite , dominated by edge states . for edge states with the dispersion @xmath341 , the local density state
is @xmath342 as illustrated in the inset of fig .
[ fig11 ] .
clearly , the zero - bias conductance peak in fig .
[ fig11 ] manifests the dispersionless andreev bound states shown in the inset of fig .
[ fig11 ] while the small plateau at finite energy scale reveals the linearly dispersive majorana fermion near @xmath180 . as a consequence , an overall candleholder - like structure of the tunneling conductance reflects the coexistence of andreev flat - band and dispersive majorana fermions at the interface .
cuprate superconductor in the presence of rashba spin - orbit interaction ( @xmath207 ) . here
@xmath291 is measured in terms of @xmath343 with @xmath344 .
the tunneling conductance @xmath345 near @xmath346 exhibits a candleholder - like structure , which reflects the local density of states at the edge in the strong interface scattering limit ( @xmath347 ) .
the insets show the corresponding local density of states and the energy spectrum in which the andreev flat edge modes contributes to the zero - energy divergence while the dispersive edge modes contribute to the plateau.,width=302 ] in this subsection , based on our mean - field solutions , the quasi - particle spectrum and josephson current in sis@xmath8 junctions are investigated .
we shall examine periodicity of josephson current in sis@xmath8 junctions with s and s@xmath8 formed by high @xmath0 cuprates in different edge orientations as illustrated in fig .
[ fig13](a ) .
in particular , we will show that the hybridization of majorana zero - energy flat - bands of s and s@xmath8 leads to @xmath348-periodicity@xcite .
however , we show that this only occurs when there is an overlap between majorana zero - energy flat - bands from @xmath254 and @xmath349 . based on these findings ,
we explore and design the tricrystal junction which can generate a spontaneous vortex with half - flux quanta with jump in unit of two flux quanta as the magnetic field is applied .
this could serve as a way for detecting the elusive majorana fermions in the high-@xmath0 cuprate - based heterostructure .
we start by representing the wavevector along the junction by @xmath174 . for a given @xmath174 , the thin insulating layer @xmath350 and superconducting sides s and s@xmath8
are effectively one - dimensional with coordinates denoted by @xmath40 and their corresponding hamiltonians are generally given by eq.([h1d ] ) , denoted by @xmath351 , @xmath352 , and @xmath353 respectively .
the effective hamiltonian of sis@xmath8 is then given by @xmath354 where the insulator occupies @xmath355 , s occupies @xmath356 , s@xmath8 occupies @xmath357 , and @xmath358 and @xmath359 are the tunneling hamiltonians describing tunneling between superconductors and the thin insulator . to investigate the josephson current , we impose a a @xmath241 phase difference @xmath360 between the pairing potentials in s and s@xmath8 sides by setting the pairing potential @xmath361 for s side and @xmath362 for s@xmath8 side . junction . here the relative angles of @xmath4 directions ( denoted by @xmath27 and @xmath363 axes ) of s and s@xmath8 to the tunneling direction ( @xmath40 axis ) are @xmath215 and @xmath364 respectively .
( b ) comparison of projected @xmath174 along the junction for edge states that would appear in isolated and semi - infinite s and s@xmath8 .
here @xmath365 and red and blue dots are nodal points .
it is seen that @xmath174 s of majorana zero - energy modes ( flat - bands ) from s and s@xmath8 may overlap in the region @xmath366 and result in @xmath248 periodicity in josephson current.,width=321 ] we first analyze edge states of isolated semi - infinite superconducting sides ( s and s@xmath8 ) . following fig .
[ fig6 ] , by projecting nodal structures onto an edge , one obtains possible number of edge modes for each @xmath174 along the edge . in fig .
[ fig12](b ) , we show the comparison of two edges with difference orientations .
it is seen that a junction formed by these two edges may hybridize two majorana zero - energy flat - bands ( e.g. @xmath366 ) , one fermionic andreev flat - band with one majorana zero - energy flat - band ( e.g. @xmath367 ) , or two fermionic andreev flat - bands ( e.g. @xmath368 ) .
as we shall see , only the overlap of zero - energy majorana bound states lead to their hybridization and result in @xmath248 periodicity of josephson current when these two edges form a sis@xmath8 junction@xcite .
if the relative angle of @xmath4 directions of a superconductor s to the tunneling direction ( @xmath40 axis ) is @xmath215 , the relative angle of projected @xmath174 to the @xmath4 direction is @xmath369 .
hence projection of the majorana zero - energy flat - band onto the junction interface is @xmath370 , where @xmath371 are magnitudes of wavevectors for nodal points .
assuming that both s and s@xmath8 have the same doping level , the projection of majorana zero - energy flat - band for s@xmath8 is also determined by @xmath371 and hence the projected majorana zero - energy flat - band is given by @xmath372 .
for doping @xmath373 and @xmath374 , we find that @xmath375 and the majorana zero - energy flat - bands for the sis@xmath8 junction overlap as long as @xmath376/@xmath377 edge , @xmath378/@xmath4 edge and ( ii ) @xmath379 , @xmath380/@xmath381 edge . in fig .
[ fig13](a ) , we show quasi - particle spectrum for the s(@xmath382)-i - s@xmath8(@xmath383 ) junction . here due to finite lengths of s ans s@xmath8 ( @xmath384 ) , in - gap states include edge states from outer edges of s and s@xmath8 . it is seen that there is no majorana zero - energy state ( single branch ) formed by hybridization of edge states from s ans s@xmath8 .
this is verified by checking energies of edge states versus the @xmath241 phase difference @xmath385 shown in fig . [
fig13](b ) .
the period is @xmath386 . )
-i - s@xmath8(@xmath383 ) junction with @xmath387 . here due to finite lengths of s ans
s@xmath8(@xmath388 ) , in - gap states include edge states from outer edges of s and s@xmath8 .
edge states inside the junction are marked by blue ( andreev fermionic modes ) and red ( majorana modes ) lines .
it is seen that single branch majorana flat - band is missing .
( b ) energies versus the phase difference @xmath385 for edge states in the junction for a few @xmath174 s .
the periods are found to be @xmath389 . here
length of s and @xmath349 is @xmath390.,title="fig:",width=170])-i - s@xmath8(@xmath383 ) junction with @xmath387 . here due to finite lengths of s ans
s@xmath8(@xmath388 ) , in - gap states include edge states from outer edges of s and s@xmath8 .
edge states inside the junction are marked by blue ( andreev fermionic modes ) and red ( majorana modes ) lines .
it is seen that single branch majorana flat - band is missing .
( b ) energies versus the phase difference @xmath385 for edge states in the junction for a few @xmath174 s .
the periods are found to be @xmath389 . here
length of s and @xmath349 is @xmath390.,title="fig:",width=170 ] next , we consider the s(@xmath382)-i - s@xmath8(@xmath391 ) junction .
[ fig14](a ) shows the full quasi - particle spectrum .
it is seen that there is a majorana flat - band for @xmath392 , resulting from hybridizing two majorana zero - energy flat - bands from s and s@xmath8 .
in addition , there are in - gap states due to edge states in the junction formed by hybridizing one fermionic andreev flat - band with one majorana zero - energy flat - band or two fermionic andreev flat - bands . in fig .
[ fig14](b ) , we examine energy versus phase difference @xmath385 for edge states in the junction . it is clearly seen that only majorana modes due to hybridization of majorana fermions from s and @xmath349 exhibit @xmath348 periodicity . the @xmath248 period results from conservation of the fermion parity@xcite and the fermion parity is conserved only when two majorana zero - energy modes are hybridized .
we thus verified that only when the majorana zero - energy flat - bands for each edge that form the junction overlap , the supercurrent in the junction shows @xmath248 periodicity . )
-i - s@xmath8(@xmath391 ) junction for @xmath393 . here
length of s and s@xmath8 is @xmath394 .
edge states inside the junction are marked by blue ( andreev fermionic modes ) and red ( majorana modes ) lines .
( b ) energy versus phase difference @xmath385 for edge states in the junction for a few @xmath174 s . here
length of s and @xmath349 is @xmath390 .
plot of @xmath395 represents the spectrum due to the hybridization of majorana fermions from s and @xmath349 .
the red and black solid lines denote different fermion parity @xmath396 .
plots of @xmath397 and @xmath398 stand for the spectrum formed by hybridizing fermionic andreev zero - energy bound states .
the remaining plot for @xmath399 is the spectrum resulted from hybridization of fermionic and majorana zero energy modes .
it is seen that only hybridization of majorana fermions results in @xmath348 periodicity . ,
title="fig:",width=154 ] ) -i - s@xmath8(@xmath391 ) junction for @xmath393 . here
length of s and s@xmath8 is @xmath394 .
edge states inside the junction are marked by blue ( andreev fermionic modes ) and red ( majorana modes ) lines .
( b ) energy versus phase difference @xmath385 for edge states in the junction for a few @xmath174 s . here
length of s and @xmath349 is @xmath390 .
plot of @xmath395 represents the spectrum due to the hybridization of majorana fermions from s and @xmath349 .
the red and black solid lines denote different fermion parity @xmath396 .
plots of @xmath397 and @xmath398 stand for the spectrum formed by hybridizing fermionic andreev zero - energy bound states .
the remaining plot for @xmath399 is the spectrum resulted from hybridization of fermionic and majorana zero energy modes .
it is seen that only hybridization of majorana fermions results in @xmath348 periodicity . , title="fig:",width=166 ] and @xmath400 are the angles between the @xmath27 axis of the upper left crystal ( a ) and right crystal ( b ) and @xmath4 direction of the bottom crystal .
( b ) regions of parameters for @xmath7 rings with majorana modes ( shaded areas with dots ) . here
grey areas are regions for zero ring@xcite .
white areas are regions for @xmath7 rings without majorana modes , and shaded areas with lines are regions for zero rings with majorana modes . for doping @xmath402 and @xmath403 ,
the width @xmath216 for the region with majorana fermions is in the range of @xmath404.,width=226 ] the @xmath248 periodicity implies fluxes of vortices trapped in the ring formed by the sis@xmath8 junction jump in unit of two flux quanta@xcite .
one of configurations that allows one to explore the pairing symmetry and also vortices trapped in high @xmath0 cuprates is the tricrystal experiment with a @xmath7-ring@xcite .
here we examine the periodicity of @xmath241 phase winding phase for the tricrystal configuration shown in fig .
[ fig15](a ) . as illustrated in fig .
[ fig15](a ) , for general tricrystal configurations , one may fix the edge direction of the cuprate in the bottom labelled by c to be in @xmath5 direction . let angles between @xmath27 axes of the upper left crystal ( a ) and right crystal ( b ) and @xmath4 direction of the bottom crystal be @xmath405 and @xmath400 . if the interface between a and b be @xmath173 and the angle between @xmath173 and the @xmath4 direction be @xmath215 , the interface @xmath173 is an s(@xmath406)-i - s@xmath8(@xmath407 ) junction . clearly , since there is no majorana zero - energy mode in the @xmath5 edge , in the tricrystal configuration shown in fig .
[ fig15](a ) , majorana modes can appear only in the interface between a and b. according to eq.([theta ] ) , as long as @xmath408 and @xmath409 , the condition latexmath:[\[\begin{aligned } is satisfied , there are majorana zero - energy modes in the interface between a and b. on the other hand , the supercurrent passing an junction interface between superconductors @xmath31 and @xmath32 is given by @xmath411 ( clean limit ) or @xmath412 ( dirty limit ) @xcite , in which @xmath413 and @xmath414 are angles of the crystallographic axes(such as @xmath4 ) with respect to the junction interface .
for the ring configuration shown in fig .
[ fig15](a ) , in order that there is effectively a @xmath7-junction in the three junctions involved , we require @xmath415 ( clean limit ) or @xmath416 ( dirty limit ) . hence to form a half - flux vortex in the ring , in the clean limit , one requires @xmath417 .
therefore , it is sufficient to require one of angles , @xmath405 , @xmath400 , @xmath406 , and @xmath407 to be greater than @xmath418 and all the others to be less than @xmath418 .
there are two possibilities : @xmath419 and all the other angles are less than @xmath418 . as a result ,
the @xmath7-junction is either the junction between a ( or b ) and c or the junction between a and b. for typical configurations of tricrystal experiments@xcite , one has @xmath401 .
the requirement of a ring to be a @xmath7-ring for the clean and dirty limits becomes the same and is given by @xmath420 on the other hand , eq.([tricrystal ] ) implies @xmath421 , which , when combined with with eq.([tricrystal1 ] ) , determines parameters that allow @xmath7 rings to host majorana fermions . in fig .
[ fig15](b ) , we summarize regions of parameters for @xmath7 rings and zero rings with majorana fermions .
it is seen that there are large parameter regions satisfying the requirement for hosting majorana fermions .
therefore , one expects that the half - flux vortex trapped in a typical @xmath7-ring would jump in unit of two flux quanta in external magnetic fields . due to the coupling of s and s@xmath8 in the sis@xmath8 junction . here
@xmath215 is fixed to be @xmath422 and the energy for the quasi - particle of @xmath254 is chosen at the nodal point with the fourier wavevector @xmath174 .
@xmath423 is the minimum energy of quasi - particles in s@xmath8 with fixed @xmath174 .
@xmath423 generally does not vanish except when s and s@xmath8 are the same or @xmath424 .
( b ) comparison of the enegy - phase relation due to quasi - particles ( solid circle ) and majorana modes ( open circle ) localized at the sis@xmath8 junction .
the josephson current due to quasi - particles is heavily suppressed.,width=302 ] in this subsection , we examine stability of @xmath348 periodicity and effects due to energy dispersion along @xmath23 axis . specifically , we examine whether the @xmath348 periodicity is spoiled by the presence of quasi - particles .
we first note that the @xmath248 periodicity relies on the conservation of fermion pairty . in the presence of quasi - particles ,
the fermion parity is no longer conserved@xcite .
however , if the energies of quasi - particles are gapped , the fermion parity is approximately conserved at low temperatures@xcite . as indicated in fig .
[ fig4 ] , the bulk superconducting states in realistic parameter regime are gapless .
however , the sis@xmath8 junction involves the coupling of the superconducting states in two different orientations .
when two superconducting states couple , since the fourier wavevector @xmath174 that is parallel to the junction is conserved , only quasi - particles with the same @xmath174 couple . due to different orientations of s and s@xmath8
relative to the junction , @xmath174 of nodal points for s and s@xmath8 are not the same . as a result , quasi - particles at nodal points of s ( or s@xmath8 ) couple to quasi - particles at finite energies of s@xmath8 ( or s ) .
hence the resulting quasi - particles residing near the junction are gapful and there is an energy gap , @xmath423 , associated with each orientation of the tunneling junction . in fig.[fig16]a , we examine @xmath423 versus @xmath364 by fixing @xmath426 . here
the energy of the quasi - particle for s side is chosen at the nodal point with the fourier wavevector @xmath174 and @xmath423 is defined to be the minimum energy of quasi - particles in s@xmath8 with fixed @xmath174 .
it is seen that except for special orientations when s and s@xmath8 are the same or @xmath424 , @xmath423 generally does not vanish .
the parity conservation thus holds for quite general orientations of s and s@xmath8 .
in addition to the protection of the pairty conservation , the momentum mismatch also suppresses the quasi - particle contribution .
this is illustrated in fig .
[ fig16]b .
it is seen that @xmath427 for the contribution of quasi - particles is much more flat in comparison to that due to majorana modes .
hence the josephson current ( @xmath428 ) due to quasi - particles is heavily suppressed and leads to little effects .
finally , we discuss effects of the energy dispersion ( @xmath425 ) along @xmath23-axis .
it is generally accepted that high-@xmath0 cuprates are approximately two dimensional with weak dispersion along @xmath23-axis .
therefore , results derived in the above are approximately correct . for high-@xmath0 cuprates with one cuo@xmath429 per unit cell
, the energy dispersion along @xmath23-axis can be approximated by@xcite @xmath430 with @xmath431 .
following the discussion on eq.([spectra ] ) , for a given @xmath425 , nodal points are determined by @xmath432 . in other words , @xmath433 changes the fermi surface . for realistic parameters , @xmath434 , nodal points are along axes @xmath163 on which @xmath435 is satisfied .
hence nodal points have no @xmath425 dispersion .
therefore , all results based on projection of nodal points ( including @xmath248 periodicity ) are the same except that there are more majorana edge modes due to the @xmath425 dispersion . on the other hand , for @xmath436 , the @xmath425 dispersion changes the fermi energy by 10% for a give @xmath425 . as a result , there is no qualitative change on the dispersive majorana modes except for the corresponding 1% change of the range for the existence of the dispersive majorana modes .
in conclusion , we have explored effects of the rashba spin - orbit interaction on the tunneling spectroscopy of high-@xmath0 cuprate superconductors . the mean - field phase diagram in the large - u limit of hubbard model
is established .
it is shown that due to the dzyaloshinskii - moriya and spin dipole - dipole interactions induced by the spin - orbit interaction , the majority regime in the phase diagram is gapless with pairing symmetry of superconductivity being @xmath437-wave .
furthermore , the gap function @xmath3-vector for @xmath2-wave superconductivity is not aligned with the internal magnetic field of the spin - orbit interaction .
as the spin - orbit interaction is turned on , we find that the ground state undergoes a phase transition to a topological gapless phase with each nodal point originated from pure @xmath3-wave being split into two stable nodal points characterized by the symmetry class diii . due to the splitting nodal structure
, zero - energy majorana modes always exist for any interfaces that are not exactly in @xmath4 or @xmath5 directions .
in addition , due to non - aligned @xmath3-vector , as the @xmath2-wave amplitude further increases beyond the rashba spin - orbit interaction , a transition of nodal point configuration occurs .
the ground state is still gapless .
however , the majorana flat band near @xmath180 in @xmath6 edge becomes the dispersive majorana edge modes .
in addition to the mean - field phase diagram , the tunneling conductances for the nis and sis@xmath8 junctions are also computed .
our results indicate that due to the presence of dispersive majorana modes , a small plateau with shoulders near the zero bias peak would be induced in the tunneling spectrum .
this small plateau with shoulders is a result due to the dirac cone of dispersive majorana modes at the edge .
the appearance of the dirac cone at @xmath180 shown in fig .
[ fig11 ] is a manifestation of @xmath2-wave order parameter and results from competition between the @xmath3-wave and @xmath2-wave pairing symmetries .
when the @xmath3-wave dominates for @xmath438 , there are only flat bands at the edge , exhibiting a single zbcp in the tunneling spectrum . only when @xmath439 , the dirac cone near @xmath180 emerges and results in the additional feature that exhibits the feature of a small plateau with shoulders .
the existence of the dirac cone at the edge requires particular symmetries .
general perturbations such as disorders do not respect these symmetries and would not yield the feature of a small plateau with shoulders in the tunneling spectrum@xcite .
hence the corresponding feature can not be generated by disorders .
the feature of a zero - bias conductance peak sitting on a small plateau with shoulders has been frequently observed in tunneling spectrum of @xmath6 edge during the past@xcite .
even though these evidences may not be the unique signature to conclude the existence of dispersive majorana fermions and further evidence from other measurements is in order , they show supportive evidence for dispersive majorana modes and indicate the importance of the spin - orbit interaction in high-@xmath0 cuprates .
in addition , we find that the overlap of dispersionless majorana flat - bands in a josephson junction always results in 4@xmath7-periodicity in josephson effect .
in particular , for typical configurations of @xmath7-ring in tricrystal experiments that can trap a half - flux vortex , we find that as long as the difference of the orientations for @xmath27 axes of two crystals involved in a junction is within @xmath9 , the junction always exhibit 4@xmath7-periodicity in josephson effect . for typical configurations of tricrystal experiments , one has @xmath440 , @xmath441 , and @xmath442@xcite . the edge of crystal b in facing a is along @xmath5 direction .
if the orientation of interfaces is perfect , since there is no edge state in @xmath5 edge , one expects that there is no majorana fermion trapped in the @xmath7-ring .
however , in real experiments , there may exist deviations in orientations of edges .
any deviation of the edge for crystal b from the @xmath5 direction would satisfy eq.([tricrystal ] ) .
this would result in trapped majorana fermions between a and b so that the trapped flux in the @xmath7-ring jumps in unit of two flux quanta in externally applied magnetic fields . this would be consistent with the experimental observation that @xmath443-flux quanta has not been observed in tricrystal experiments@xcite .
therefore , the above analyses indicate that majorana fermions may have been already observed in tunneling experiments on high-@xmath0 cuprates .
while so far in this work we only consider results based on the mean - field theory , we expect our results are robust in the presence of correlation effects as long as the symmetry of the system is not changed .
even though our results show agreement with past experimental observations , definite confirmation of majorana fermions in high-@xmath0 cuprates requires further experimental studies .
nonetheless , our results offer important signatures to look for in future experiments . in particular ,
the setup configuration for tricrystal experiments would offer a unique way to hold majorana fermions in high-@xmath0 cuprates .
all the signatures we find are crucial for successfully searching majorana fermions and are left for future experimental confirmations .
n. reyren , s. thiel , a. d. caviglia , l. f. kourkoutis , g. ham- merl , c. richter , c. w. schneider , t. kopp , a. s. ruetschi , d.jaccard , m. gabay , d. a. muller , j. m. triscone , and j. mannhart , science 317 , 1196 , 2007 .
f. giubileo , a. jossa , f. bobba , a.i .
akimenko , g. malandrino , l.m.s .
perdicaro , i.l .
fragala , and a.m. cucolo , physica c * 367 * , 170(2002 ) ; i. shigeta , f. ichikawa , and t. aomine , physica c * 378 - 381 * , 316(2002 ) | we explore possible signatures for observing majorana fermions in the tunneling spectroscopy of high-@xmath0 superconductors .
it is shown that due to the rashba spin - orbit interaction ( @xmath1 ) generally introduced by contact electrode , in addition to the heisenberg spin exchange interaction , the dzyaloshinskii - moriya and spin dipole - dipole interactions are induced . as a result
, @xmath2-wave superconductivity is induced with the gap function @xmath3-vector being not aligned with the internal magnetic field of the spin - orbit interaction . for typical strength of the rashba interaction , the induced @xmath2-wave is weak .
hence the resulting superconductor is still gapless and is not a topological superconductor .
however , we find that the ground state undergoes a phase transition to a topological gapless phase with each nodal point originated from pure @xmath3-wave being split into two stable nodal points characterized by the symmetry class diii . due to the splitting nodal structure , zero - energy majorana modes always exist for any interfaces that are not exactly in @xmath4 or @xmath5 directions .
hence for general interfaces , existence of majorana modes is a robust feature .
in addition , due to the non - aligned @xmath3-vector , for @xmath6 interfaces in which @xmath3-wave is subdominant to @xmath2-wave , there exist sizable dispersive majorana edge states .
our results indicate that due to the presence of these majorana modes , a small plateau in tunneling spectrum near zero bias peak would be induced .
furthermore , zero - energy majorana modes result in 4@xmath7 periodicity in typical sis@xmath8 junctions with difference in orientations of s and s@xmath8 being within @xmath9 for @xmath10 ev@xmath11 ev . as a result
, it is easy for a @xmath7-ring in tricrystal experiments to hold majorana fermions and exhibit periods of two flux quanta in external magnetic field .
these phenomena may have been already observed in experiments and their connections to experimental results are discussed . |
the biological evolution theory of darwin is well known to all .
the biological evolution can be divided into two different types , micro - evolution and macro - evolution , which focus on the dynamics of evolution at the level of a single population and of all species in ecosystem respectively . up to now , many scaling laws have been found in macroevolutionary data . the distribution @xmath6 of extinction sizes for families @xmath7 decreases with @xmath7 according to a power - law : @xmath8@xcite .
the distribution @xmath9 of genera lifetimes @xmath10 follows @xmath11@xcite .
the distribution of the number of genera @xmath12 with @xmath13 species can be fitted with @xmath14 @xcite . the disappearance of species or genera is similar to radioactive decay , in which the amount of the original element decays exponentially with time .
this phenomenon has been named as red queen effect@xcite .
the extinction events exhibit punctuated patterns@xcite .
the punctuated equilibrium can be interpreted as a metastable state of the dynamics of that single species@xcite .
if an ecosystem is in equilibrium , every species is stable .
however , if one of them fluctuates with time , it will lead to the destabilization of its neighboring species .
thus , the avalanche fluctuation will sweep through the evolution .
the existence of scaling laws in macro - evolution indicates that the dynamics of large - scale evolution is the result of a self - organized critical process@xcite .
based upon this view , bak and sneppen introduced a toy model for species evolution@xcite .
it is defined and simulated as follows : ( i ) @xmath15 species are arranged on a one - dimensional lattice with periodic boundary conditions .
( ii ) random numbers @xmath16 , which are chosen from a uniform distribution between @xmath17 and @xmath18 , @xmath19 , are independently assigned to each site ( species ) .
( iii ) at each time step , the extremal site , i.e. , the one with the smallest random number @xmath20 , and its nearest neighbors are mutated by assigning new random numbers in @xmath21 .
as the system evolves it turns out that the species with fitness significantly above the self - organized threshold @xmath22 ( equal to @xmath23@xcite ) will never mutate , unless its neighbors mutate .
after enough mutations have occurred , the ecosystem reaches a statistically stationary state in which the distribution for the fitness of all species is statistically stationary .
since the bak - sneppen model was defined , its punctuated patterns and avalanche dynamics have been investigated extensively . through observing the fluctuations of the smallest fitness ,
the model shows punctuated equilibrium behavior , i.e. , it self - organizes into a critical state with intermittent coevolutionary avalanches of all size@xcite .
paczuski , maslov and bak studied the @xmath24 avalanche ( pmb avalanche ) , forward avalanche and backward avalanche and derived the exact scaling equations , they also established the relations among the critical exponents@xcite . recently , by monitoring the variations of the average fitness , the @xmath25 avalanche ( lc avalanche ) , is found , and the exact equations and scaling relations are obtained@xcite . as we know , the avalanche dynamics is a kind of macroscopic phenomenon in driven dissipative systems .
the detailed dynamics of these systems sensitively depend on the initial configurations .
but the distribution of avalanches , i.e. , scaling law , does not depend on such details due to the universality of complexity@xcite . in the existing works , analyzing the smallest fitness only directly connects with the feature of individuals , but it does not directly represent the global feature ; discussing the average fitness only directly connects with the global dynamics , but it does not directly manifest the difference in individuals . can we find a new quantity , which can describe the global feature and the individual difference simultaneously ? in this paper , we define such a quantity and observe the corresponding avalanche dynamics with this quantity . in the next section , the standard distribution width of fitness is defined , and its gap equation is obtained . in sec .
iii , the dynamics of avalanche is analyzed in detail , the exact mast equation is derived , and the scaling relations are established among the critical exponents . the last section is a brief summary .
in the bak - sneppen model , the fitness of the @xmath26-th species is denote as @xmath27 @xmath28 , with @xmath15 being the total number of species .
the fitness represents the population or the living capability of the species , large fitness means immense population or great living capability , and vice versa . to describe the average population or average living capability of an ecosystem , li and cai defined the average fitness as@xcite @xmath29 where @xmath30 is the fitness of the @xmath26-th species at time @xmath31 . based upon the average fitness
, we define the standard distribution width of fitness as @xmath32 from the above definition , we know that the standard distribution width of fitness simultaneously manifests both global dynamics and individual difference . at the beginning of the evolution , i.e. , @xmath33 , the fitnesses are uniformly distributed in @xmath34 .
thus for an infinite - size ecosystem @xmath35 , the initial average fitness @xmath36 equals @xmath37 , the initial standard distribution width @xmath38 is @xmath39 .
but for a finite - size ecosystem , @xmath36 and @xmath38 fluctuate around the above values due to the finite size . because of the absence of mutations or updates , @xmath38 does not show the correlation between species , only when the evolution goes on , such correlation is contained in @xmath40 gradually . as time increases
, @xmath41 shows an increasing tendency@xcite , whereas , @xmath40 shows a decreasing tendency ( see the bottom of fig .
1 ) . after a long transience
, i.e. , @xmath42 , the ecosystem self - organizes to a critical state , and , @xmath41 and @xmath40 approach to different thresholds .
when @xmath43 , almost all species have fitnesses above the threshold @xmath44@xcite , and the fitnesses are uniformly distributed in @xmath45 . with the definitions of @xmath41 and @xmath40 , we obtain the relation among the thresholds @xmath46 here , @xmath1 and @xmath47 are thresholds of @xmath40 and @xmath41 respectively .
thus , one gets @xmath48 and @xmath49 . for the finite - size ecosystem
, the self - organized critical thresholds will fluctuate around the above values , and this fluctuation decreases with the increase of @xmath15 .
similar to the definition in refs.@xcite , we define the gap of @xmath40 as its envelope : @xmath50 ) & , & s>0 .
\end{array } \right.\ ] ] the above definition means that the current gap of @xmath40 is the most minimum of all @xmath51 @xmath52 . in the top of fig . 1
, we show the gap @xmath53 as a function of @xmath31 . with the increase of time ,
@xmath54 gradually approaches close to the threshold . by definition ,
when the gap falls to its next lower value , the separate instances are separated by avalanches .
the avalanches correspond to plateaus in @xmath53 during which @xmath55 .
a new avalanche is initiated when the gap falls , and the old avalanche terminates at the same time .
once a avalanche is over it will never affect the behavior of any subsequent avalanche .
when the gap decreases , the probability for @xmath40 to jump above @xmath53 increases , thus longer and longer avalanches happen typically . finally , the ecosystem falls into statistically stationary state .
following refs.@xcite , we derive the gap equation of @xmath53 .
if the gap is changed from @xmath53 to @xmath56 , the average number of avalanches occurred is @xmath57 .
thus , when @xmath58 , the average number of time steps required for increasing the gap from @xmath53 to @xmath56 is @xmath59 , where @xmath60 is the average size of avalanche of the current plateau of the gap .
so the differential equation for @xmath53 is @xmath61 eq.(5 ) describes the relaxation to attractor .
if the law of @xmath62 is obtained , one can derive the law of relaxation to attractor from the above equation .
we know that all self - organized critical systems will exhibit a power - law avalanche dynamics .
this has been confirmed by a lot of models with self - organized criticality@xcite . in this section ,
based on the definition of pmb avalanche@xcite and lc avalanche@xcite , the definition of a new avalanche , @xmath63avalanche , is presented .
then its critical exponents are calculated with numerical simulation and its exact master equation is derived .
lastly , the scaling relations are established .
the definition for @xmath63avalanche is as follows .
similar to those used in refs.@xcite , we introduce an auxiliary parameter @xmath0 , where @xmath64 .
suppose that at time step @xmath65 , the current standard distribution width @xmath66 is less than @xmath0 . if the next standard distribution width @xmath67 is larger than @xmath0 , a creation - annihilation branching process is initiated .
the avalanche will continue to run unless @xmath40 becomes less than @xmath0 .
this means that , at time @xmath31 , if all @xmath68 for @xmath69 , the current avalanche continues . in terms of the above definition ,
the size of an avalanche is the number of time steps of subsequent punctuations above @xmath70 .
if the first appearance of @xmath71 after time @xmath65 occurs at time @xmath72 , the size of the current avalanche is @xmath73 .
clearly , the above definition ensures the hierarchical structure , i.e. , larger avalanches consist of smaller ones . with the decreasing of @xmath0 ,
smaller avalanches combine into larger ones ; if @xmath0 is set as @xmath1 , the infinite - size avalanche appears . in the other case , with the increasing of @xmath0 , larger avalanches split into smaller ones .
thus , the cutoff effects is unavoidable when @xmath0 is not chosen as @xmath1 .
nevertheless , the same scaling laws can be obtained when @xmath0 is very close to the threshold @xmath1 . for the bak - sneppen model with @xmath74 species , choosing @xmath70 as @xmath75
, we find by numerical simulation that the distribution of avalanche size follows the power law @xmath76 where @xmath77 is the probability of avalanche of size @xmath73 , @xmath78 equals @xmath79 ( see fig .
because of the cutoff effects , the average avalanche size varies when @xmath0 changes .
similar to those in refs.@xcite , our numerical results show that the average size of @xmath0 avalanche obeys @xmath80 here , @xmath1 is the threshold , the exponent @xmath5 equals @xmath81 ( see fig .
this law means that the divergence of the average avalanche size satisfies power law .
our numerical simulation also shows that the relaxation to attractor abides @xmath82 here , @xmath31 is the time steps , the exponent @xmath4 equals @xmath83 .
obviously , the value of @xmath4 is very close to @xmath84 .
so far , we have found some scaling laws of the @xmath63avalanche by numerically simulations .
now , we will give some exact results . to describe the cascade process of smaller avalanches combining into larger ones
when the parameter @xmath0 is changed , an exact master equation is derived as follows . denoting the probability of @xmath63avalanche of size @xmath73 as @xmath85 ,
when parameter changes from @xmath0 into @xmath86 , some @xmath63avalanches merge to larger @xmath86 avalanches , and some small avalanches merge into @xmath63avalanche .
thus , at the same time , the probability of a avalanche with a given size flows in and out . because the termination of @xmath63avalanche is uncorrelated , the probability of a @xmath63avalanche merging to @xmath87 avalanche is proportional to @xmath88@xcite .
thus the probability which flows out from @xmath85 can be written as @xmath89 where @xmath90 is a constant .
the probability which flows in @xmath85 is @xmath91 when @xmath0 approaches @xmath1 and @xmath92 approaches zero , we obtain the following master equation @xmath93 as pointed out in the previous section , with the increasing of time , the standard distribution width shows a decreasing tendency .
so , when @xmath0 decreases , the first term on the right side of the master equation reflects the loss of @xmath73-size avalanche , and the second term depicts the gain of @xmath73-size avalanche .
similar to ref.@xcite , after defining a new variable @xmath94 , the master equation becomes @xmath95 through performing a laplacian transformation for @xmath96 , i.e. , @xmath97 thus the master equation reads as @xmath98.\ ] ] when @xmath99 , @xmath100 , this is the normalization of @xmath96 . while @xmath101 , one gets @xmath102 . expanding both sides of eq.(14 ) as taylor series throughout the neighborhood of the point @xmath99
, one can obtain @xmath103 \nonumber \\ & = & \lambda [ 1-\alpha \left\langle s\right\rangle _ { u}+\frac{1}{2!}\alpha ^{2}\left\langle s^{2}\right\rangle _ { u}-\frac{1}{3!}\alpha ^{3}\left\langle s^{3}\right\rangle _ { u}\cdot \cdot \cdot ]
\nonumber \\ & & \times [ \alpha \left\langle s\right\rangle _ { u}-\frac{1}{2!}\alpha ^{2}\left\langle s^{2}\right\rangle _ { u}+\frac{1}{3!}\alpha ^{3}\left\langle s^{3}\right\rangle _ { u}\cdot \cdot \cdot ] , \end{aligned}\ ] ] where @xmath104 . comparing the coefficients of different powers of @xmath105 in eq.(15 ) , one gains an infinite series of exact equations .
the first exact equations can be expressed as @xmath106 replacing @xmath107 as @xmath108 , we obtain the @xmath5 equation @xmath109 noting the scaling law @xmath110 , inserting it into the above equation , one can easily gain @xmath111 by far , we have obtained three critical exponents of the @xmath0 avalanche : @xmath112 , @xmath3and @xmath4 , which describe the behavior of the avalanche size distribution , the average avalanche size , and the relaxation to attractor , respectively . as listed in ref.@xcite
, there are three other critical exponents : @xmath113 , @xmath114 and @xmath115 .
the avalanche dimension @xmath116 is defined by the following scaling relation @xmath117 where @xmath73 is the avalanche size ( temporal duration ) , @xmath118 is the spatial extent .
the @xmath114 and @xmath115 are defined by the following scaling laws of cutoff @xmath119 @xmath120 the symbols @xmath121 and @xmath122 are the cutoff of the avalanche size and that of spatial extent of the avalanche .
although there are six critical exponents , only part of them are independent . after choosing @xmath78 and
@xmath5 as the independent exponents , the other ones can be expressed as functions of them . inserting the scaling law @xmath110 into the gap equation ( eq.(5 ) ) , and noting the scaling law @xmath123 , one can find @xmath124 integrating the equation @xmath125 , it is easy to obtain @xmath126 comparing with @xmath110 , one can get @xmath127 noting that the average number of the covered sites @xmath128 covered by a @xmath0 avalanche scales near the critical point as@xcite @xmath129and @xmath130 , so @xmath131 , integrating the equation @xmath132 gives @xmath133 thus , the avalanche dimension @xmath113 can be written as @xmath134 noting @xmath135 and @xmath136 , one can immediately obtain @xmath137 comparing with @xmath138 , it is easy to know @xmath139 up to now , the scaling relations among the six critical exponents have been successfully established .
the existence of two independent critical exponents means that there are two kinds uncorrelated critical behavior in @xmath70 avalanche .
in conclusion , by defining a new quantity , the standard distribution width of fitness , a new avalanche , @xmath0 avalanche , is observed in the evolution of bak - sneppen model .
we declare this avalanche as a different hierarchy of avalanche , because firstly , the observed quantity is different from those of refs.@xcite and refs.@xcite ; and secondly , the values of the critical exponents are also different from those found in refs.@xcite and refs.@xcite . from the definition of the standard distribution width of fitness , we easily obtain the corresponding gap equation and the self - organized threshold .
apparently , with this definition , the global feature and the difference among individuals are simultaneously described by the standard distribution width .
so , for the evolutionary distribution with changeless or small changed average values , the standard distribution width is a good quantity to observe the evolutionary dynamics . according to the definition of pmb avalanche and lc avalanche ,
the definition of the @xmath0 avalanche is presented . with numerical simulation ,
the critical exponents @xmath78 , @xmath140and @xmath4 are obtained .
then the exact master equation is derived . from the master equation ,
the @xmath5 equation is immediately obtained .
finally , combining all scaling laws , the scaling relations are successfully established among the critical exponents . with these relations
, we find there are only two uncorrelated critical behavior in the @xmath0 avalanche , for there exist only two independent critical exponents .
the work is supported by the national natural science foundations of china and foundations of chinese academy of sciences .
the authors are very grateful the help of prof .
x. cai and dr .
w. li of institute of particle physics in hua - zhong normal university . | we introduce the standard distribution width of fitness to characterize the global and individual features of a ecosystem in the bak - sneppen evolution model . through tracking this quantity in evolution , a different hierarchy of avalanche dynamics ,
@xmath0 avalanche is observed . the corresponding gap equation and
the self - organized threshold @xmath1 are obtained .
the critical exponents @xmath2 @xmath3and @xmath4 , which describe the behavior of the avalanche size distribution , the average avalanche size and the relaxation to attractor , respectively , are calculated with numerical simulation . the exact master equation and @xmath5 equation are derived . and
the scaling relations are established among the critical exponents of this new avalanche . |
since the dawn time the chaos is an indispensable part of human life .
one of the most famous models with chaotic behavior is the logistic map .
its simple analytic form meant that it was used in many scientific disciplines such as biology , cryptography , communication chemical physics and stock market @xcite .
one of the mathematical models of chaos was discovered by mitchell feigenbaum ( 1978 ) @xcite .
feigenbaum considered ordinary difference equations used for example in biology to describe the development of population in its dependence on time .
he discovered that population oscillates in time between stable powers ( fixed points ) , the number of which doubles according to changes in the power of external parameter .
the generalization of the feigenbaum model contains all first - order difference equations @xmath8 .
the condition for the existence of chaos is the single maximum of a function @xmath9 .
feigenbaum also proved that the transition to chaos is described by two universal constans @xmath10 and @xmath11 , later named the feigenbaum numbers .
+ we quote here a few definitions which are basic in the development of this exposition .
* definition 1 .
@xcite , p. 62 ] + let @xmath12\subset\re$ ] , @xmath13 , @xmath14 .
discrete iterations of a function @xmath15 are called functions @xmath16 defined inductively : @xmath17 , @xmath18 , @xmath19 .
* * definition 2 .
[ @xcite , p. 62
] + the sequence @xmath20 , where @xmath21 is called the sequence of iterates of a function @xmath22 generated by point @xmath23 .
* * definition 3 .
@xcite , p. 65 ] + a point @xmath24 , with @xmath25 is called a first - order fixed point of @xmath22 .
a fixed point @xmath26 of a function @xmath22 is locally stable if there exists a neighbourhood @xmath27 such that all sequences of iterates are convergent to @xmath26 for @xmath28 .
+ @xmath29 [ see @xcite , p. 68 ] + if function @xmath22 is differentiable in fixed point @xmath26 and @xmath30 , then @xmath26 is attractive . *
definition 4 .
@xcite , p. 72 ] + let @xmath15 .
we say that @xmath26 is a periodic point of the function @xmath22 , or that @xmath26 generates a cycle of order @xmath31 , if @xmath32 and @xmath33 .
+ @xmath34 @xcite + if function @xmath22 has a cycle of order different to @xmath35 , then @xmath22 is chaotic .
+ @xmath36 [ cf .
@xcite , p. 79 ] + for arbitrary continuous function @xmath15 existence cycles of order @xmath37 involve the existence cycles of order @xmath38 in the following order : + @xmath39 . * * _ _ the logistic map @xcite , @xmath40,\ ; ( p , q)>0,\ ; n=0,1,2, ... \ ] ] is one of the most simple forms of a chaotic process . basically , this map , like any one - dimensional map , is a rule for getting a number from a number .
the parameter @xmath5 is fixed , but if one studies the map for different values of @xmath5 ( up to 4 , else the unit interval is no longer invariant ) it is found that @xmath5 is the catalyst for chaos .
after many iterations @xmath41 reaches some values independent of its starting value 3 regimes : @xmath42 for large @xmath38,@xmath43 for large @xmath38,@xmath44 cyclic behavior , where @xmath45,@xmath46 mostly chaotic , + an interesting thing happens if a value of @xmath5 greater than 3 is chosen .
the map becomes unstable and we get a pitchfork bifurcation with two stable orbits of period two corresponding to the two stable fixed points of the second iteration of @xmath22 . with @xmath5
slightly bigger than 3.54 , the population will oscillate between 8 values , then 16 , 32 , etc .
the lengths of the parameter intervals which yield the same number of oscillations decrease rapidly ; the ratio between the lengths of two successive such bifurcation intervals approaches the @xmath47 @xmath48 .
the period doubling bifurcations come faster and faster ( 8 , 16 , 32 , ... ) , then suddenly break off . beyond a certain point , known as the accumulation point @xmath49
, periodicity gives way to chaos . in the middle of the complexity
, a window suddenly appears with a regular period like 3 or 7 as a result of mode locking .
the 3-period bifurcation occurs at @xmath50 , and period doubling then begin again with cycles of 6,12 , ... and 7,14,28 , ... and then once again break off to chaos .
one - dimensional representations are the simplest arrangements capable to producing a chaotic movement .
let us consider @xmath51 .
we postulate non - linear function : @xmath52 \label{glm}\ ] ] set by difference equation : @xmath53,\ ; n=0,1,2 , .... \label{genlogmap}\ ] ] for the first time the function ( [ glm ] ) has been presented in @xcite .
the maximum of the function ( [ glm ] ) is reached for @xmath54 , because @xmath55 and @xmath56 , @xmath57 .
+ it results that @xmath58=@xmath59 and for @xmath60 where + @xmath61 function @xmath62 is @xmath63)\subset [ 0,1]$ ] . + parameter @xmath64 depending on @xmath2 and @xmath3 accepts the following limit values ( fig.[fig1 ] ) : @xmath65 @xmath66 for the sake of parameters @xmath2 and @xmath3 the function ( 3 ) generates a lot of new functions which can be analyzed using mathematical methods , just like for a logistic map , thus obtaining very interesting results .
+ a)let us assume that @xmath67 and @xmath68 .
then @xmath5 is depend on the value @xmath69 and @xmath70.$ ] + b ) if @xmath71 , then @xmath72 and @xmath64 is independent on @xmath2 , @xmath3 and @xmath73 always - for logistic map , we have @xmath4 and @xmath73 . + c ) if @xmath74 , then @xmath75 , because @xmath76 + d ) if @xmath77 , then @xmath78 , because @xmath79 + e ) now let us consider a case , when @xmath6 and @xmath80
. then we have @xmath81,~~ n=0,1,2, .... \ ] ] and @xmath82.$ ] taking into account the above relationship we can observe change maximum of the function ( 5 ) depending on @xmath3 ( fig.[fig2 ] ) . * a)the property of @xmath83 for @xmath84 , @xmath85 in the periodic region .
+ in the following part of this paper we will show that the feigenbaum model of transition to chaos is correct for representation ( 4 ) for given positive values of parameters @xmath2 and @xmath3 .
therefore , we will study the dynamics of this map in the case of @xmath6 and @xmath7 .
then the representation ( 5 ) assumes the following form : @xmath86,\;\ ; n=0,1,2 , ...
\label{eq12}\ ] ] a function @xmath62 in the point @xmath87 has a local maximum and the value of function in this point is equal to @xmath88 . from this relationship
, we have @xmath63)\subset [ 0,1]$ ] for @xmath89 $ ] .
the trajectories behavior generated by @xmath62 depends on @xmath5 and , therefore , we have to distinguish several intervals of this value : + 1)for @xmath90 $ ] all the sequences of iterates @xmath91 depend on the initial value @xmath92 tend to @xmath93 .
this point is the fixed point of the map for @xmath90 $ ] and according to _ definition 3 it is also the stable point of this map .
+ 2)let @xmath94 .
if @xmath95 then , @xmath96 , @xmath97 , so for @xmath98 function ( 6 ) has two fixed points . using _
theorem 1 we get : @xmath99 , @xmath100 and @xmath101 .
+ therefore , we have that @xmath102 is an unstable fixed point , whereas @xmath97 is an attractive fixed point for @xmath103 .
+ change @xmath5 from 1 to 2 has caused the migration of the attractive point from 0 to @xmath104 .
+ _ conclusion 1 .
+ for @xmath105 there exists a stable fixed point attracting all points . when @xmath106 , then @xmath107 and this point stops being an attractive point .
+ 3)for @xmath108 we will look into the stability of fixed points @xmath62 and @xmath109 of the map ( 5 ) as a function of parameter @xmath5 , because the feigenbaum model is generated by bifurcations connected with iterations of a function .
_ _ _ * let us consider the second iteration of @xmath62 , namely @xmath110 , with fixed points : @xmath96 , @xmath97 , @xmath111 , @xmath112 . because @xmath113 we have @xmath114 , @xmath115 , @xmath116 whereas for @xmath117 fixed points @xmath118 and @xmath119 are attractive points according to condition in _ theorem 1 .
+ _ conclusion 2 .
+ the fixed point @xmath120 of @xmath121 is also the fixed point of @xmath109 as well as of all higher iterations .
+ _ conclusion 3 .
+ if the fixed point of @xmath121 becomes unstable , then it is also the unstable fixed point of @xmath109 and of all next iterations . + from the inequality
@xmath122 we have @xmath123f^{\ ; ' } ( s)|=|f^{\ ; ' } ( s)|>1 $ ] .
+ if @xmath124 , then the fixed points of @xmath125 become repulsive at the same time . following this instability ,
the fourth iteration shows two new pitchfork bifurcations giving cycle @xmath126 order for four attractive fixed points , which is called period duplication . _
_ _ to generalize the above examples we get : + * a)for @xmath127 there exists a stable cycle @xmath128 , whose elements @xmath129 are defined by the following connections : + @xmath130 all points of the cycle of @xmath128 become unstable simultaneously and the bifurcation creates a new stable cycle of @xmath131 order for @xmath132 existing for @xmath127 .
+ * c)for a map ( 6 ) the value of parameter @xmath133 we can describe the following connection : @xmath134 where @xmath135 . + * d)there exists a point of accumulation of an infinite number of the bifurcation of period duplication for the finite value @xmath5 which we appointed as @xmath49 : @xmath136 .
+ taking into account the above , @xmath137,@xmath138 , @xmath139,thus @xmath140 . + * e)for the representation ( 6 ) there exists a constant , whose value corresponds to the feigenbaum constant : @xmath141 + _ proof.let us assume that @xmath142 .
then @xmath143 , @xmath144 and @xmath145 .
+ a table of @xmath131 cycles type and values of @xmath146 is given below .
it is clearly visible that doubling bifurcations come faster and faster ( 8 , 16 , 32 , ... ) and suddenly break off beyond a certain point known as the accumulation point and periodicity gives way to chaos .
_ * * * * * 12.5 cm + * b)the property of @xmath83 for @xmath84 , @xmath85 in chaotic region .
+ * a)for @xmath49 the bifurcation sequence ends with a set of the infinitely numerous points , which is called the set of feigenbaum attraction .
+ * b)if @xmath147 then we observe the irregular dynamics of map ( 6 ) with narrow ranges in which the set of attraction has a periodical character . +
* c)numeric calculations for the representation ( 6 ) show that the largest range is for @xmath148
. then there exists 3-cycle ( fig.[fig3 ] ) , which according to th.2 and th.3 , implicates the chaos as well as the existence of cycles of different orders . +
* d)for the map ( 6 ) at @xmath149 , + * * * * * @xmath150 can be solved by the simple change of variables : @xmath151 then ( 7 ) can be converted into + @xmath152={3\sqrt 3\over 2}\cos{(y_n)}\sin{^2(y_n)},\ ] ] + which has one solution @xmath153}$ ] and + @xmath154 + the invariant density function @xmath155 of @xmath156 can be calculated using the following definition : + @xmath157 using @xmath158 , ( 9 ) becomes : @xmath159 $ ] + i.e. @xmath160{\sqrt{x^2 - 1}-x } \sqrt{-\left(\sqrt{x^2 - 1}-x\right)^{2/3}-\frac{1}{\left(\sqrt{x^2 - 1}-x \right)^{2/3}}+1 } } .\ ] ] the above function is shown in fig.[fig4 ] .
+ the invariant density function of the logistic map for @xmath161 is given by @xmath162 .
+ + * c)ordered chaos - construction of an attractor for + @xmath163 and @xmath164 .
+ + at the end of this considerations we will present a method of the _ ordering of chaos_. a lot of numeric experiments have led to the situation where it is possible to rank chaos in such a way that the successive iterations generated by chaotic dynamic system would reach an attractor of a regular shape .
this method also has it , that from the sequence of several thousand iterations ( for representation @xmath165 where @xmath166 defined recursively ) we draw up the following points @xmath167 on a plane - we construct an attractor . in this way
an attractor comes into being , which is visited irregularly by points of successive iterations ( fig.[fig5 ] ) . + * 13.5 cm * d)bifurcation diagram and lyapunov exponent .
+ + nowadays the lyapunov exponent is applicable in various scientific disciplines such as biology , engineering , bioengineering and informatics @xcite .
a quantitative measure of the sensitive dependence to initial conditions is the lyapunov exponent , which is a measure of the exponential separation of nearby orbits . for discrete time system @xmath168 , for an orbit starting with @xmath92 the lyapunov exponent @xmath169 can be calculated by @xmath170 a positive lyapunov exponent can be considered as an indicator of chaos , whereas negative exponents are associated with regular and periodic behavior of the system .
figure [ fig6 ] shows the bifurcation diagram and lyapunov exponent for the equation ( [ eq12 ] ) .
it is clear that even in chaotic areas ( @xmath147 ) there are many periodic intervals ( @xmath171 ) . *
currently , modeling real - world systems is very popular . a well - known model that takes into account the deterministic chaos is the logistic map .
however , this one - dimensional map can be controlled only by one parameter . in the present contribution
, we postulate a generalization of the classical logistic map which is generally controlled by three parameters .
this could allow to adapt to the conditions prevailing in the modeled system .
as an example , we carry out analytical and quantitative analysis in the case where @xmath6 and @xmath7 both in the periodic and chaotic regime where the lyapunov exponent has positive values .
it turns out that the dynamics of this equation is faster than in the case of the logistic map - the value of the parameter @xmath172 is less than the value of @xmath73 for the logistic map . in the periodic regime
we identified feigenbaum constant @xmath11 which is typical for all dissipative systems ( also for the logistic map ) .
for chaotic area we have found an analytical form of the invariant density function .
for both regimes we have proposed specific form of data representation which allowed to obtain a non - trivial structure of an attractor set . + * footnotes + this work was partially supported by the centre for innovation and transfer of natural sciences and engineering knowledge ( university of rzeszow ) .
* n.k .
pareeka , vinod patidara , k.k .
sud , _ image encryption using chaotic logistic map _ , image and vision computing , 24 , 926 - 934 , ( 2006 ) .
a. ferrettia , n.k .
rahman , _ a study of coupled logistic map and its applications in chemical physics _
, chemical physics , 119 , 275 - 288 , ( 1988 ) . c. pellicer - lostao , r. lpez - ruiz _ a chaotic gas - like model for trading markets _ , journal of computational science , 1 , 2432 , ( 2010 ) . m. andrecut , _ logistic map as a random number generator _ , int . j.
modern phys .
b , 12 , 921930 , ( 1998 ) .
feigenbaum , _ the metric universal properties of period doubling bifurcations and the spectrum for a route to turbulence _ , newyork .
, 330 - 336 , ( 1980 ) .
feigenbaum , _ quantitative universality for a class of non - linear transformations _ , stat .
phys . , 25 - 52 , ( 1978 ) .
thompson , j. b. mcguire , _ asymptotic and essentially singular solutions of the feigenbaum equation _
, j. stat .
51 , 991 - 1007 , ( 1988 ) .
stephenson , y. wang , _ numerical solution of feigenbaum s equation _
notes 15 , 68 - 78 , ( 1990 ) .
pierre - franois verhulst , _ notice sur la loi que la population poursuit dans son accroissement _ , correspondance mathmatique et physique , 113121 , ( 1838 ) .
e. ott , _ chaos in dynamical systems _ , cambridge university press , ( 1993 ) .
schuster , _ deterministic chaos : an introduction , weinheim , ( 1988 ) . _ a. lyapunov , _ problem general de la stabilite du mouvement _ , ann math stud 17 , ( 1949 ) .
a. wolf , j.b .
swift , h.l .
swinney , j.a .
vastano , _ determining lyapunov exponents from a time series _ , physica d , 16 , 285 - 317 , ( 1985 ) .
m.a . van wzk , w. steeb , kluwer academic publishing , ( 1997 ) .
strogatz , _ nonlinear dynamics and chaos : with applications to physics , biology , chemistry , and engineering _ , addison - wesley , new york , ( 1994 ) .
nayfeh , b. balachandran , _ applied nonlinear dynamics : analytical , computational , and experimental methods _ , john wiley and sons , new york ( 1995 ) .
p. bryant , r. brown , h. abarbanel , _ lyapunov exponents from observed time series , physical review letters _ 65 , ( 1990 ) . | motivated by a possibility to optimize modelling of the population evolution we postulate a generalization of the well - know logistic map .
generalized difference equation reads : @xmath0 @xmath1,\;(p , q)>0,\;n=0,1,2, ... $ ] , where the two new parameters @xmath2 and @xmath3 may assume any positive values .
the standard logistic map thus corresponds to the case @xmath4 .
for such a generalized equation we illustrate the character of the transition from regularity to chaos as a function of @xmath5 for the whole spectrum of @xmath2 and @xmath3 parameters . as an example we consider the case for @xmath6 and @xmath7 both in the periodic and chaotic regime .
we focus on the character of the corresponding bifurcation sequence and on the quantitative nature of the resulting attractor as well as its universal attribute ( feigenbaum constant ) . |
the static electrical ( dc ) resistivity is important for the characterization of the plasma state . at high density ,
when the spitzer theory becomes invalid , a consistent many - particle theory is required in order to evaluate the transport coefficients .
the extended ziman @xcite formulation of the electrical resistivity of liquid metals is based on linear response theory , where free electrons in a metal are uniformly accelerated until they collide with an ion and are scattered . in the @xmath0matrix formulation of evans @xcite ,
the resistivity reads @xcite : @xmath1 where @xmath2 is the matter density in @xmath3 , @xmath4 the matter density in g.@xmath3 , @xmath5 the atomic mass in g and @xmath6 the avogadro number . the fermi - dirac distribution and its derivative read respectively : @xmath7,\ ] ] where @xmath8 and @xmath9 is the chemical potential .
the function @xmath10 is defined by an integral : @xmath11 where @xmath12 , @xmath13 being the electron mass .
the quantity @xmath14 is the ionic structure factor and @xmath15 represents the scattering cross - section , @xmath16 and @xmath17 , where @xmath18 , @xmath19 and @xmath20 being the momenta of the conduction electron respectively before and after the scattering event .
therefore , one has : @xmath21 where @xmath22 , defined by @xmath23 , is the scattering amplitude .
in fact , the resistivity takes now a drude - like form @xmath24 with @xmath25=21.74 @xmath26.cm , @xmath27=52.91772083 10@xmath28 cm being the bohr radius .
@xmath29 represents the electron density .
the characteristic time @xmath30 and the relaxation time @xmath31 are such that @xmath32 note that any model which treats the continuum by a thomas - fermi approximation will not give the correct continuum density of states and not correctly incorporate the effects of electron - electron interactions in the electron - ion phase - shifts @xcite .
such a model is not reliable for the evaluation of the scattering cross - section .
the average - atom ( aa ) model we used relies on the following self - consistent calculation .
schrdinger s equation is solved in the pauli approximation @xcite , in which only first - order relativistic corrections are retained : @xmath33-ev_{\mathrm{scf}}+v_{\mathrm{mv}}+v_{\mathrm{d}}+v_{\mathrm{so}}\right]y_{\epsilon , l}(r)=\epsilon\ ; y_{\epsilon , l}(r),\ ] ] where @xmath34 is the mass - velocity term , @xmath35 is the darwin term , and @xmath36 is the spin - orbit contribution .
@xmath37 is the radial part of the wavefunction multiplied by @xmath38 .
the potential splits into two parts @xmath39 , where @xmath40 represents the exchange - correlation part , evaluated in the local - density approximation @xcite , and @xmath41 the electrostatic part . outside the wigner - seitz ( ws ) sphere ,
_ i.e. _ beyond the radius @xmath42 , a bound - electron wavefunction is given by : @xmath43^{1/2},\ ] ] where @xmath44 is a modified spherical bessel function of the third kind and @xmath45 a constant fixed by the boundary condition of the wavefunction at @xmath42 and the normalization .
such a boundary condition will be noted bc2 in sec .
[ sec4 ] , while the usual `` confining '' boundary condition @xmath46 will be referred to as bc1 . beyond @xmath42 ,
a free - electron radial wavefunction is written @xmath47^{1/2}r\;\left[\cos[\delta_{l}(k)]\;j_l(kr)-\sin[\delta_{l}(k)]\;n_l(kr)\right],\ ] ] where @xmath48 and @xmath49 are the spherical bessel functions of the first and second kinds respectively , @xmath50 is the rest - mass energy of the electron , and @xmath51 is the phase shift , given by @xmath52=\frac{k\;j_l'(kr_{ws})-\xi\;j_l(kr_{ws})}{k\;n_l'(kr_{ws})-\xi\;n_l(kr_{ws})},\ ] ] where @xmath53 is the logarithmic derivative of the radial wavefunction @xmath54 .
the calculation of the phase - shifts @xmath51 is painstaking and requires accurate numerical methods @xcite .
one defines , for a fixed chosen @xmath55 , the finite difference @xmath56 then , if @xmath57 ( @xmath58 being an infinitesimal threshold ) , @xmath59 points are added in the interval @xmath60 $ ] , which enables one to define @xmath61 where @xmath62 .
then , for each @xmath63 of the new grid , if @xmath64 , then @xmath65 points are added in the interval @xmath66 $ ] , etc .
the iterative process goes on until the resonance is correctly described .
the radial electron density is given by @xmath67 the new potential @xmath41 , solution of poisson s equation , is inserted in eq .
( [ schr ] ) and the process continues until convergence is reached .
the electron - ion cross - section for momentum transfer reads @xmath68p_l(\chi)\right|^2,\ ] ] where @xmath69 are the legendre polynomials .
it is useful to split the summation over @xmath70 in two parts in eq .
( [ eics ] ) : the first summation running from 0 to @xmath71 ( typically equal to 30 ) and the second one from ( @xmath71 + 1 ) to @xmath72 .
the latter can be approximated by the classical expression of the cross - section of a screened charge ( summed from 0 to @xmath72 ) minus the values for @xmath73 .
assuming , at high electron energies , a classical scattering cross - section for a screening potential @xmath74 , this consists in writing @xcite
@xmath75p_l(\chi)&=&\frac{a}{2k^2}[\sum_{l=0}^{\infty}(2l+1)q_l(\zeta)p_l(\chi)\nonumber\\ & & -\sum_{l=0}^{l_0}(2l+1)q_l(\zeta)p_l(\chi)]\nonumber\\ & = & \frac{a}{2k^2}\left[\frac{1}{\zeta-\chi}-\sum_{l=0}^{l_0}(2l+1)q_l(\zeta)p_l(\chi)\right ] , \end{aligned}\ ] ] with @xmath76 and @xmath77 , where @xmath78 is a screening length .
the quantity @xmath79 represents legendre function , evaluated using recursive relations for low @xmath80 and from an integral representation ( see ref .
@xcite , p. 1017 ,
8.821 ) for @xmath81 1.02 .
one gets @xmath82e^{i\delta_l(k)}-\frac{a}{2k}q_l(\zeta)\right]+\frac{a}{2k^2}\frac{1}{\zeta-\chi}.\ ] ] since @xmath83 @xcite , and since @xmath84 is known , for @xmath70=0 , @xmath85 , @xmath71 , one fits the ratio @xmath86 to @xmath87 in order to determine @xmath88 and then @xmath89 . in that way , @xmath78 is a function of @xmath63
. it can be numerically costly to solve this equation for the numerous values of @xmath63 . in order to speed up the computation
, one can use the asymptotic form @xmath90 $ ] , which leads to @xmath91 $ ] , _ i.e. _ , for @xmath92 , @xmath93 .
the maximum energy of the continuum is taken to be @xmath94 .
therefore , at high temperatures , it can be relevant to add the following correction to the resistivity : @xmath95 .resistivity @xmath96 ( @xmath26.cm ) of al at t=5 ev for different structure factors : py : percus - yevick ( hard sphere ) , dh : debye - hckel , rinker ( ref .
@xcite , eq .
( 15 ) , p. 4211 ) , and bd : bretonnet - derouiche @xcite , structure factor for the one component plasma ( ocp ) . [
cols="^,^,^,^,^",options="header " , ]
rinker @xcite proposed a combined structure factor which has the virtue to approach debye - hckel limit as the core exclusion radius tends to 0 , and the simple hard - sphere limit as the debye radius tends to @xmath72 .
the structure factor is expected to be important at low temperature only ( @xmath97 10 ev according to @xcite ) , but as can be seen on table [ tab1 ] , except for the debye - hckel case , the differences between percus - yevick , bd ( structure factor proposed by bretonnet and derouiche @xcite for the ocp ) and rinker @xcite are quite small even at 5 ev .
table [ tab2 ] shows the impact of the value of the bound wavefunction at @xmath42 ( fifth and sixth columns ) .
the maximum difference can reach 25 @xmath98 . in the case of iron at @xmath99=5 kev and @xmath4=1965 g/@xmath100 ( see table [ tab3 ] ) , the difference between bc1 ( the wavefunction cancels at @xmath42 ) and bc2 ( which allows a bound - electron wavefunction to extend outside the ws sphere , see eq .
( [ jel ] ) ) drops to 1.14 @xmath98 .
our results lie between the results of the models a and b of rinker @xcite . in the following ,
we keep the condition bc2 ( atom in a jellium ) .
we can see in table [ tab2 ] that our values decrease more rapidly with compression than the values of ref .
@xcite ( with or without multiple scattering ) , and are almost always smaller . the comparisons of fig .
[ fig1 ] show that our results are closer to the pseudo - potential results including multiple scattering @xcite than to the scattering - type calculations @xcite .
some other approaches exist ; for instance , kuhlbrodt _ et al . _
@xcite calculated the resistivity within the linear response method of zubarev .
this correlation method is applicable for arbitrary degeneracy of the plasma , _
i.e _ from the low - temperature spitzer regime up to the strongly coupled , degenerate domain where ziman s approach is appropriate .
it seems , however , that reliable resistivity calculations must include a unified description of the plasma composition and the density dependence of the inter - particle interactions @xcite .
resistivity measurements in plasma state are scarce so that the theoretical approach is in many cases the unique tool available in these studies
. the measurement of resistivity along a thermodynamic path going from the melting point under normal conditions to a state corresponding to a compression 0.03 at t=26.5 ev in al has been performed at los alamos by benage _
_ @xcite using the exploded wire method .
figure [ fig2 ] shows that a good agreement is obtained .
the accuracy of the experimental results is generally within a factor of 2 , which is of the same order of the typical discrepancy between theory and experiment .
it is important to keep in mind that expression ( [ inti ] ) neglects multiple - scattering processes , which was shown to lead to a difference of a factor 1.5 in the theoretical values @xcite .
our values are smaller than the results of perrot and dharma - wardana @xcite and the results of the density - response model of kitamura and ichimaru @xcite . in the intermediate density range ,
our results agree fairly well with the analytical model of lee and more @xcite , which was rather unexpected .
the number of relevant bound - electron configurations in a plasma can be immense , especially as the atomic number @xmath101 increases . as a configuration is defined by orbitals occupied by an integer number of bound electrons , a superconfiguration ( sc )
is defined by super - orbitals , which are groups of orbitals close in energy ( _ i.e. _ whose energies differ by less than a small fraction of @xmath102 ) and that are populated in all possible ways consistent with the pauli exclusion principle . for instance
@xmath103 is a sc composed by four super - orbitals .
a configuration is a particular sc in which each super - orbital contains only one sole orbital . a reasonable number , _
i.e. _ , a few hundred , of scs can contain a large number of configurations ( especially for high values of @xmath101 ) .
the value of thermodynamic quantity @xmath104 is given by @xmath105 where @xmath106 is the probability of the configuration @xmath107 and @xmath108 .
a self - consistent calculation similar to the one described in sec .
[ sec2 ] is performed for each sc .
table [ tab2 ] displays the values of the resistivity averaged over the superconfigurations ( third column ) .
a correction is required in the use of @xmath109 instead of @xmath110 since this occurs as @xmath111 in the scattering cross - section .
that is , we need to use @xmath112 rather than @xmath113 . the fluctuations in @xmath110 are intimately related to the fluctuations of the electron and ion densities of the plasma .
assuming that ion fluctuations are too slow to follow the electron - density fluctuations which determine the electronic configurations and electron - scattering processes , we consider the fluctuations in the electron subsystem only .
we have @xmath114 where the finite - temperature interacting electron - gas compressibility @xmath115 can be obtained from the electron - gas response function via the static structure factor since @xmath116 .
concerning the prefactor of the ziman formula , since one has @xmath117 ( jensen s inequality for convex functions ) and assuming that this dependence overcomes the one from the scattering cross - section , the resistivity averaged over scs is expected to be larger than the one obtained from the aa model .
however , it is important to keep in mind the fact that the latter prefactor should not be @xmath118 but rather @xmath119 , which could impact the results .
this is related to rinker s remark @xcite that @xmath120 is in fact @xmath121 where @xmath122 can indeed be identified with @xmath109 on variational grounds ( boltzmann equation ) , which is not the case of @xmath123 , number of charge carriers , involved in the relaxation time .
we presented calculations of the dc electrical resistivity of warm dense plasmas using the generalized ziman formula .
the calculations do not use any model potential , dielectric function or structure data , but proceed via first - principle calculations based on density - functional theory in the local - density approximation .
the electronic structure of an average scatterer is determined solving schdinger s equation .
the calculations require phase shifts which satisfy the finite - temperature friedel sum rule .
different model structure factor have been tested , with two different boundary conditions of the wavefunctions .
we discussed the approximations in the method arising from the use of an average - atom model and compared our results with other existing theoretical models .
our values are in fairly good agreement with experimental data .
the method proposed here provides a self - contained and self - consistent approach to the electrical conductivity of plasmas of arbitrary degeneracy , structure and density .
it enables one to perform calculations beyond the average - atom model , taking into account the fluctuations due to the electron configurations describing different ionization and excitation states . | the exploration of atomic properties of strongly coupled partially degenerate plasmas , also referred to as warm dense matter , is important in astrophysics , since this thermodynamic regime is encountered for instance in jovian planets interior .
one of the most important issues is the need for accurate equations of state and transport coefficients .
the ziman formula has been widely used for the computation of the static ( dc ) electrical resistivity . usually , the calculations are based on the continuum wavefunctions computed in the temperature- and density - dependent self - consistent potential of a fictive atom , representing the average ionization state of the plasma ( average - atom model ) .
we present calculations of the electrical resistivity of a plasma based on the superconfiguration ( sc ) formalism . in this modeling ,
the contributions of all the electronic configurations are taken into account .
it is possible to obtain all the situations between the two limiting cases : detailed configurations ( a super - orbital is a single orbital ) and detailed ions ( all orbitals are gathered in the same super - orbital ) . the ingredients necessary for the calculation are computed in a self - consistent manner for each sc , using a density - functional description of the electrons . electron exchange - correlation is handled in the local - density approximation .
the momentum transfer cross - sections are calculated by using the phase shifts of the continuum electron wavefunctions computed , in the potential of each sc , by the schrdinger equation with relativistic corrections ( pauli approximation ) .
comparisons with experimental data are also presented . |
the spin-@xmath8 @xmath0 chain and its variants are among the most widely used and quoted models in theoretical investigations on spin systems .
this popularity among theorists is due to two facts . on the one hand
, this family of models allows for exact theoretical description of many static as well as dynamic properties(see ref .
- ) , and on the other hand , in some important cases , it provides a good description of real systems @xcite . as an example
, we note that the dimerized chain , with alternating antiferromagnetic bonds , has been studied in relation with the spin - peierls transition@xcite , and it is thought to represent the spin degrees of freedom of organic compounds that undergo the so called _
peierls distortion_. the present state of the art in fabricating low dimensional systems , with the material science technology developed after the synthesis of the superconducting cuprate oxides , may now tailor compounds that reveal a wealth of new magnetic phenomena , including random spin chains , spin ladders , and doped magnetic chains among other systems with exotic behaviors . for theorists ,
these systems are fascinating , with no parallel in classical or three - dimensional physics .
being systems of low dimensionality with low values of the spin , they are dominated by quantum effects@xcite .
compounds such as @xmath9 , @xmath10 ( @xmath11 _ ethyl sulfate _ ) , and @xmath12 are among the quasi - one - dimensional systems , whose low temperature properties are thought to be described by the @xmath0 model .
a great deal of theoretical information has been gathered concerning equilibrium properties , with calculations of quantities such as the specific heat , the magnetic susceptibility , and equal - time spin - spin correlation functions@xcite . as for the dynamic properties ,
the quantities most thoroughly studied are the time - dependent spin - spin correlation functions ( tdcf ) @xmath13 where @xmath14 is the hamiltonian of the system , @xmath15 is an index for the spin component ( @xmath16 ) , and @xmath17 is the equilibrium average@xcite .
they are important for the description of such dynamic phenomena as magnetic resonance , magnetic neutron scattering , spin diffusion , and other relaxation properties@xcite .
it should be noted however , that their application is restricted to situations near the equilibrium state , where the linear response theory is valid@xcite . in particular , for the uniform @xmath0 model , an old calculation for the time dependent autocorrelation @xmath18 showed the absence of spin diffusion in the limit @xmath1 @xcite .
this behavior was thought to be accidental and specific of the @xmath0 model for spin 1/2 ( see , for example , ref . ) .
however , recent research has shown that this surprising property is shared by a whole family of integrable models , and is attributed to the existence of a macroscopic number of conservation laws@xcite .
this remarkable result has been probed in magnetic resonance experiments in some 1d systems@xcite .
we will turn to this point later . in spite of the wide range of their applicability
, we note that tdcf do not give a direct description of the time evolution of the non - equilibrium process@xcite .
this description is achieved through a different method , that we briefly discuss below . this second direction in investigating the dynamics relies on calculation of averages of the type : @xmath19~ , \label{ave}\ ] ] where @xmath20 is the density matrix which satisfies liouville equation @xmath21\ . \label{liouville}\ ] ] averages of such a kind give a direct account of the non - equilibrium evolution of the physical observable @xmath22 , independently of how far the initial state is from the equilibrium state or from a stationary one .
unfortunately , calculations of these quantities are much more involved than the calculation of tdcf ( [ cf ] ) for linear response , and few exact results are known @xcite .
whether the dynamical process is near or far from equilibrium , most works in the literature deal with cases where the initial state is _ spatially uniform_. this premise is assumed in explicit or non - explicit way , and the methods developed to solve the problem are heavily based on it .
in contrast , the study of spatially nonuniform excitations is practically new , in spite of its interest , both theoretical and experimental .
this problem is important for a deeper understanding of dynamical processes in many - particle spin systems with strong exchange interactions .
inhomogeneous initial states can be prepared in real systems by external actions , for instance , strong inhomogeneous magnetic fields or acoustic waves . on the other hand , from the theoretical side , exact results on the dynamics ( as in the case of the @xmath0 model ) ,
can elucidate details of spin - spin relaxation processes in more complicate systems . in the present contribution
, we will adopt the method based on formulae ( [ ave ] ) and ( [ liouville ] ) , and will analyze in detail the long time evolution of the magnetization in three versions of the @xmath0 model .
the initial excitation is always prepared in the form of a spatially inhomogeneous magnetization ( sim ) , and the calculation is done in exact analytic form .
the three variants of the @xmath0 model that we consider are enumerated below : 1 .
isotropic dimerized @xmath0 model with hamiltonian : @xmath23 ; \nonumber\end{aligned}\ ] ] 2 .
isotropic @xmath0 model in a staggered magnetic field with hamiltonian : @xmath24 1 . @xmath0 model with bond alternation without magnetic field : @xmath25 we will refer to them as models i , ii , iii , respectively . in expressions ( [ dimer])-([bond ] ) , @xmath26 @xmath27 is the bohr magneton , @xmath28 is the gyromagnetic ratio , @xmath29 , @xmath30 , @xmath31 are magnetic fields , @xmath32 , @xmath33 , @xmath34 , @xmath35 , @xmath36 are exchange integrals , @xmath37 , and @xmath38 ) , and cyclic boundary conditions are assumed .
the above restriction on @xmath39 , @xmath34 and @xmath32 is not important in our problem and is introduced for convenience . to investigate sim dynamics in the above models , we will use a previously developed method given in ref ..
this method was applied earlier by one of the authors@xcite , to the same study for the isotropic @xmath0 chain in a homogeneous magnetic field , with hamiltonian : @xmath40 we will refer to this model as model iv subsequently .
we remark that highly nontrivial results were obtained in ref .. in particular , it was shown that in the @xmath41 limit , some of the spatially inhomogeneous excitations do not disappear and are still time - dependent
. the time evolution of sim s ( and also their spatial distribution ) is probed through the computation of the fourier components @xmath42 of the magnetization as a function of the @xmath5-wave number .
the calculation yields a relaxation process in the form of a power law @xmath43 where the exponent @xmath44 depends on the initial state . in ref .
it was also studied the anisotropic @xmath0 model in the limit of strong anisotropy .
it was found that the exponent @xmath44 in ( [ power ] ) changes discontinuously at some critical values @xmath7 of the wave vector @xmath5 .
the values @xmath7 depend solely on the parameters of the hamiltonian and are not connected with the preparation of the initial state , nor the particular component of the magnetization that is relaxing .
moreover , in the limit @xmath45 , the relaxation time @xmath46 of ( [ power ] ) diverges to @xmath47 , with the corresponding slowing down of the process .
this phenomenon is at variance with some conventional views for the time evolution of physical quantities in many particle systems , according to which all quantities must be temporal independent in the limit @xmath41 .
the concept of spin temperature , widely used in non - equilibrium magnetic phenomena @xcite , is based on the assumption that the spin - spin relaxation is much faster than the final relaxation to the lattice degrees of freedom .
model i was also preliminary investigated by the some of the authors , and similar conclusions were attained for the dimerized @xmath0 model@xcite , in the sense that a critical - like slowing down of the relaxation process takes place at special points @xmath48 of the @xmath5-space . in other words , for @xmath49 ,
the inverse time scale @xmath50 for some of the oscillating components of sim , goes to zero following the power law @xmath51 .
such behavior is not surprising for the case @xmath52 , for which the corresponding value of the total magnetization @xmath53 is a constant of motion , but is unusual for @xmath54 where the corresponding fourier component @xmath55 is not conserved .
we note that @xmath56 is the only critical point for the uniform isotropic @xmath0 model ( model iv and limit @xmath57 for model i and @xmath58 for model ii ) .
the @xmath41 behavior is dominated by one oscillating component , and this component has no critical properties . the present paper is devoted to elucidate this remarkable difference within a more general context , by the extensive study of a whole family of models .
we suggest that the dissimilar behaviors are due to the presence of a gap in the spectrum of the energy excitations , the uniform isotropic model being gapless .
our paper is organized as follows : in section ii , we outline the main steps in the analytic calculation .
all the models are diagonalized by means of a modified jordan - wigner transformation , which maps the spin model into an equivalent fermion hamiltonian . then
, the average @xmath59 is calculated in exact closed form , and its asymptotic behavior for @xmath60 is obtained after taking the thermodynamic limit @xmath61 . in section iii , we give a detailed analysis of the results , including all models and the long time behavior of the magnetization .
section iv closes the paper with the final discussions .
the quantity of our interest is the time - dependent @xmath62-component of sim @xmath63 ~ , \label{szqav}\ ] ] where @xmath20 is the density matrix of the system , and @xmath64 is the fourier transform of the magnetization . in ( [ szqav ] ) , @xmath5 is the wave vector characterizing the spatial inhomogeneity of the initial state ( @xmath65 @xmath66 ) . to calculate @xmath59
one can use the identity : @xmath67 \ , \ ] ] where @xmath68 is initial density matrix , and @xmath69 is the spin operator in the heisenberg representation @xmath70 to obtain exact results for @xmath59 , for each of the models considered , we follow the steps : 1 .
diagonalization of the corresponding hamiltonian using the fermion representation of the jordan - wigner transformation . at the end
, we get a free - fermion system ; 2 .
transformation of @xmath55 in terms of fermionic operators to get the time dependent @xmath69 ; 3 .
averaging of @xmath69 with the initial density matrix @xmath68 .
the last step depends on the form of initial matrix @xmath68 . as it was shown in ref .
, exact solutions of this problem may be obtained in the case when @xmath68 is a functional of only one component of the spin operator : @xmath71 let us note that such initial state can actually be prepared in real systems at low temperature , with a strong nonhomogeneous magnetic field directed along the coordinate axis @xmath72 .
methods for diagonalization of hamiltonians ( [ dimer])-([bond ] ) are well known ( see for example ref . ) .
minor differences are specific for the version of the model to be solved . in our case
we have employed the following procedure : 1 .
use of the jordan - wigner transformation to change from spin to fermi operators @xmath73 : + @xmath74 + where @xmath75 2 .
introduction of two types of fermi operators for even and odd sites : @xmath76 and of their fourier transforms @xmath77 , @xmath78 : @xmath79d_{m}^{\alpha } , \label{blochop}\ ] ] @xmath80 , with @xmath81 , where we have adopted the compact notation @xmath82 3 . after the above steps , hamiltonians ( [ dimer])-([bond ] )
are transformed into quadratic forms in terms of operators @xmath83 .
diagonalization of these quadratic hamiltonians is standard , but depends on the specific model under consideration . for hamiltonians ( [ dimer ] ) and ( [ stagg ] )
it can be done with the help of the bogolyubov transformation given below : @xmath84 where@xmath85 are new fermi operators , @xmath86 and@xmath87 is a real function of @xmath88 .
we are using the same convention ( [ compact ] ) for @xmath89 . + for hamiltonian ( [ dimer ] ) , the functions @xmath90 are given by : @xmath91 where @xmath92 + for hamiltonian ( [ stagg ] ) , one gets : @xmath93^{1/2}~,\quad v_{k}=[(p_{k}+\omega _ { d})/2p_{k}]^{1/2}\mbox{sign}(\cos k/2 ) \label{coeffstagg}\ ] ] where @xmath94 is real ( independent of @xmath95 ) and @xmath96 4 .
following ref .
, diagonalization of hamiltonian ( [ bond ] ) is achieved through : @xmath97 as a final result , all the three hamiltonians can be represented in the diagonal form @xmath98 \ , \quad \\ ( i=1,2,3 ) \end{array } $ } \label{diagonal}\ ] ] except for constant terms that are not important for the dynamics .
the index @xmath99 refers to models i , ii and iii , respectively .
the dispersion relations of ( [ diagonal ] ) are given by : @xmath100 where @xmath101 note that here and in the following , we neglect , as usually done@xcite , the boundary term of the order @xmath102 which appeared in a chain with cyclic boundary conditions .
the thermodynamic limit @xmath103 will be taken at the end of the calculation .
using formulae ( [ ft ] ) and ( [ jw])-([blochop ] ) , one can easily obtain the following expression @xmath105 where @xmath106 , with definitions common for all models . to obtain the time dependence of @xmath55
, we first express the relation ( [ sq ] ) in terms of the canonical @xmath107 s and @xmath108 s through transformations ( [ bogo ] ) for models i and ii , and ( [ bogosp ] ) for model iii .
then , we insert the time evolution of such operators @xmath109 to get @xmath110\eta _ { p}^{+}\eta _ { q}+(u_{p}u_{q}+v_{p}v_{q}^{*})\exp [ -i\omega _ { i}^{-}(k , q)t]\beta _ { p}^{+}\beta _ { q}+\right . \nonumber \\ & & \left . + ( u_{p}v_{q}^{*}-u_{q}v_{p}^{*})\exp [ i\omega _ { i}^{+}(k , q)t]\eta _ { p}^{+}\beta _ { q}+(u_{q}v_{p}-u_{p}v_{q})\exp [ -i\omega _ { i}^{+}(k , q)t]\beta _
{ p}^{+}\eta _ { q}\right\ } \label{sqt}\end{aligned}\ ] ] for models i and ii , where @xmath111 and functions @xmath112 are defined by formulae ( [ coeff ] ) and ( [ coeffstagg ] ) respectively . in the same way , for model iii , we obtain : @xmath113+\eta _ { -p}\beta _ { q}\exp [ -i(\omega _ { p}^{(3)}+\theta _ { q}^{(3)})t]\right\ } ~. \label{sqtiii}\ ] ] to calculate the time dependent average @xmath59 , we use the identity @xmath114 which means that , equivalently , one can calculate the average of operator @xmath115 in the heisenberg picture in relation to the initial density matrix @xmath68 . from equations ( [ sqt ] ) and ( [ sqtiii ] ) , one sees that this problem reduces to calculation of averages of the type @xmath116 , @xmath117 , ... , including all the combinations of canonical operators .
the calculation is described in detail in appendix [ a ] .
below , we will summarize these results together with the ones obtained in ref . for model iv .
we remark that exact solutions can be obtained for special forms of the initial density matrix @xmath68 .
we assume here that @xmath68 is a functional of only one component of the spin operators , _
i.e. _ , @xmath118 1 . in the case
@xmath119 ( the case @xmath120 is the same by symmetry ) we have @xmath121 for models i , iii , iv , and @xmath122-\cos [ t\omega _ { 2}^{-}(k , q)]\right\ } dk , \label{sqtx}\end{aligned}\ ] ] for model ii , where @xmath123 is the fourier transform of the short - range order correlation of the @xmath124 component .
2 . in the case
@xmath125 , we have different expressions for all the cases
. 1 . for model i
, we get the formula : @xmath126+f^{-}(k , q)\cos [ t\omega _ { 1}^{-}(k , q)]\right\ } dk , \label{sqtzi}\ ] ] with @xmath127 2 . for model
ii , we have the more explicit form : @xmath128+\right .
\\ \left .
+ \left [ 1+\left ( \omega _ { d}^{2}+\omega _ { e}^{2}\cos ( p/2)\cos ( q/2)\right ) /p_{p}p_{q}\right ] \cos [ t\omega _ { 2}^{-}(k , q)]~\right\ } \ . \end{array } \label{sqtzii}\ ] ] 3 . for model iii , the compact expression : @xmath129dk \label{sqtziii}\ ] ] where @xmath130 let us note that at the specific value @xmath131 , the time dependence of @xmath132 is given by the simple formula : @xmath133 where @xmath134 is the bessel function of first kind .
4 . for model iv , it was found in ref .
@xmath135 with @xmath136 .
in this section , we firstly discuss general properties of the time evolution of the magnetization , which can be inferred directly from equations ( [ 22 ] ) and ( [ null])-([sqtziv ] ) . in addition , we will study in detail the asymptotic behavior at long times , in the limit @xmath41 .
we will show that the evolution of sim s displays interesting unusual features , depending on the wave vector @xmath5 characterizing the inhomogeneous initial state .
let us first consider model i in the case @xmath125 for the initial density matrix , since it presents a specific feature that is absent in the other models studied here . from equations ( [ 22 ] ) and ( [ sqtzi])-([def ] ) , one sees that the value of the magnetization @xmath59 at @xmath138 , depends on the initial values of the two fourier components @xmath139 and @xmath140 .
this means that , if only one component of @xmath55 exists at @xmath141 , for example for wave vector @xmath142 , _ i.e. _ @xmath143 then a twofold response develops for @xmath144 , with components @xmath145 and @xmath146 , whose time evolution is described by equation ( [ 22 ] ) , with @xmath147 and @xmath148 , respectively .
in contrast , for models ii - iv , if there is only one initial component @xmath149 , only one component @xmath59 of the same wave vector @xmath150 exists at subsequent times , @xmath144 .
when the initial condition is prepared with @xmath119 , the average @xmath151 vanishes identically for models i , iii and iv , for any @xmath144 .
this result may be understood intuitively , without any calculation , if one notes that the part of the hamiltonian that describes the coupling of spins with the magnetic field , commutes with the other one describing the exchange interactions ( in the case of model iii , this statement is trivial for we only have the exchange part ) . denoting the latter as @xmath152 , we get the time evolution @xmath153 ~ , \label{x1}\ ] ] which is governed by the exchange only . inside the trace in ( [ x1 ] ) , we now perform a unitary transformation consisting in a rotation by @xmath154around the @xmath124 axis . under this operation ,
the spin operators transform as follow : @xmath155 the exchange hamiltonian @xmath152 is invariant under this transformation because it is a sum of products of the form @xmath156 and @xmath157 .
the initial density matrix @xmath68 is also invariant since it only depends on the @xmath158 component of the spins ( @xmath159 ) .
so , the right hand side of equation ( [ x1 ] ) will change its sign , yielding @xmath160
i.e. @xmath161 . in the case of model ii , the exchange part does not commute with the total hamiltonian @xmath162 .
this leads , in general , to a nonvanishing @xmath151 , for @xmath144 , which results proportional to @xmath163 , where @xmath164is the fourier transform of the spin - spin correlation function @xmath165evaluated at the initial condition ( see relation ( [ fourier ] ) ) .
if the initial state is prepared with a single @xmath5-vector in the form @xmath166 , then @xmath167 is the only component that exists for @xmath144 . for @xmath168 ( homogeneous initial state ) , and for any @xmath72 , we get the ( evident ) result : @xmath169 for models i , ii , iv .
this is a direct consequence of the conservation of the total @xmath62-component of the magnetization . in the case of model iii , when @xmath125 , we get the closed result@xmath170@xmath171 where @xmath172 . another important property valid for models i , ii and iv , and for any @xmath173 , is that the evolution of sim s is independent of the value of external field @xmath29 .
this fact is not surprising because the zeeman term of the respective hamiltonians commutes with both , the remaining part describing the exchange interaction and the operator @xmath55 .
this way , the time evolution of sim s is determined by the exchange term only .
special important limits are @xmath174 in model i and @xmath175 in model ii , that map onto model iv . in both of these cases , the corresponding formulae for @xmath59 reduce to ( [ sqtziv ] ) which was obtained in ref . . unfortunately , it is difficult to get more information from formulae ( [ 22 ] ) , ( [ sqtx ] ) , ( [ def ] ) , and ( [ sqtzii ] ) , because they are rather involved .
so , in the following , we will consider the limiting case @xmath60 , that will be studied using the stationary phase method .
according to the stationary phase method @xcite , the long - time evolution of @xmath59 is dominated by the contribution of the stationary points of the functions @xmath176 .
the number of these points depends on the value of the wave vector @xmath5 . for convenience of further discussion ,
let us introduce here the so called _ critical _ values of @xmath177@xmath178 the above values are determined in a standard manner through the stationary phase method , being the locus where stationary points become degenerate@xcite .
let us first examine the role of @xmath179 .
it is easy to show that it has five nondegenerate stationary points in the interval @xmath180
@xmath181 @xmath182 in turn , for @xmath183 and @xmath184 , one gets three nondegenerate points at : @xmath185 exactly at the critical values , we obtain : 1 . for @xmath186 , one nondegenerate point @xmath187 and two degenerate ones @xmath188 ( @xmath189^{\prime \prime } $ ] @xmath190 @xmath189^{\prime \prime \prime } = 0,$ ] @xmath191^{\left ( iv\right ) } \neq 0 $ ] , where derivations are taken with respect to @xmath88 ) ; 2 . for @xmath192 , two nondegenerate points @xmath193 , and one degenerate at @xmath194^{\prime \prime } = [ \omega _ { 1}^{+}(0,q)]^{\prime \prime \prime } = 0,[\omega _ { 1}^{+}(0,q)]^{iv}\neq 0)$ ] ; 3 . for @xmath195 , there is no time evolution , since the magnetization @xmath196 is a constant of the motion , @xmath197 the function @xmath198 vanishes at @xmath195 and @xmath199 .
for any other wave number @xmath5 , it has two nondegenerate stationary points : @xmath200 this finishes the analysis of the stationary points that dominate the long - time behavior of ( [ sqtzi ] ) .
the corresponding asymptotic development of @xmath59 for @xmath1 , can be represented as a sum of several oscillating components with @xmath5-dependent frequencies and amplitudes : @xmath201 @xmath202 \ , \label{series}\ ] ] where @xmath203 are functions of @xmath5 and the exponent @xmath204 assumes the values @xmath8 or @xmath205 ( the latter value will be discussed in detail below ) .
the number of components @xmath206 depends on @xmath5 . for @xmath207(excluding @xmath208 , which is a degenerate point where @xmath209 and @xmath210 coincide ) , there are four components . for @xmath211 and @xmath212
( excluding accidental degeneracies ) , there are only three terms in ( [ sum ] ) .
the frequencies @xmath213 and the inverse of the characteristic times @xmath214 of ( [ series ] ) are given below : @xmath215 we display typical @xmath5-dependences of @xmath213 and @xmath214 in figs.1 and 2 , respectively , for particular values of @xmath216 .
slowing down of the relaxation process occurs at points where the @xmath214 s vanish . at a generic @xmath5-point , the first nonvanishing terms in ( [ series ] ) have the form : @xmath217 \ , \quad ( i=1, ... ,4 ) , \label{36}\ ] ] with the amplitudes @xmath218 showing that the value at @xmath5 depends on the two initial components @xmath219and @xmath220 .
formulae for the real and imaginary parts @xmath221 and @xmath222 , as well as the phases @xmath223 , are given in appendix [ b ] .
the dynamical process described by ( [ sum ] ) and ( [ series ] ) , with the explicit formulae given in ( [ 36 ] ) and appendix [ b ] , is remarkable , since it exhibits long time tails in the relaxation of sim s to the spatially homogeneous state .
in contrast to the exponential relaxation , which is characterized by a single parameter that yields the time scale or the relaxation rate , the power law relaxation given by ( [ series ] ) and ( [ 36 ] ) is characterized by two parameters : @xmath214 , which determines the inverse time scale of the process , and the exponent @xmath224 , which determines the relaxation rate . in general , no conservation laws or long - lived hydrodynamic modes seem to be associated with the above long time tails . in the neighborhood of critical points
, the relaxation of some of the components of ( [ sum ] ) begins to stop , with the corresponding relaxation time @xmath225 diverging , as shown in fig .
exactly in the limit , the corresponding exponent @xmath224 jumps discontinuously from @xmath8 to @xmath205 .
this slowing down of the relaxation process is very similar to the critical slowing down found in phase transition phenomena.@xcite let us summarize below the singular behavior of the relaxation time at special points ( they are displayed in fig .
2 ) : 1 . @xmath226 when @xmath227 , for any @xmath216 ; 2 .
@xmath228 when @xmath229 , for any @xmath216 ; 3 .
@xmath230 when @xmath231 , for any @xmath216 ; 4 .
@xmath232 when @xmath233 , for any @xmath216 .
all the @xmath214 vanish according with the law : @xmath234 we now give the behavior of components of ( [ sum ] ) at the critical points : 1 . in the limit @xmath235 , the components @xmath236 and @xmath237 merge into a single component with @xmath238 : @xmath239 , \label{40}\ ] ] with @xmath240{24}}{8\pi } \gamma ( 1/4)\left [ ( 1+\delta ) \left\langle s_{\pm q_{c_{1}}}^{z}\right\rangle _ { 0}+i\sqrt{% 1-\delta ^{2}}\left\langle s_{\pi
\pm q_{c_{1}}}^{z}\right\rangle _ { 0}\right ] \ , \\ \\ \tau _ { c}^{-1}=3\omega _ { 1}\delta ^{2}/(1-\delta ^{2 } ) , \end{array}\ ] ] where @xmath241 is the gamma function .
the other components follows the law ( [ 36 ] ) with @xmath186 ; 2 .
analogously , for @xmath242 , the components @xmath243 and @xmath237 join together with @xmath238 , and the time dependence : @xmath244 \ , \ ] ] where @xmath245{24}}{8\pi } \gamma ( 1/4)\left [ ( 1-\delta ) \left\langle s_{\pm q_{c_{2}}}^{z}\right\rangle _ { 0}+i\sqrt{% 1-\delta ^{2}}\left\langle s_{\pi \pm q_{c_{2}}}^{z}\right\rangle _ { 0}\right ] \ .\ ] ] the other components follows the law ( [ 36 ] ) with @xmath192 ; 3 .
the component @xmath246 is critical at @xmath195 , with the relaxation time @xmath247 diverging in the vecinity of @xmath56 .
note that the frequency @xmath248 vanishes in the limit @xmath249 .
the criticality of @xmath250 has a different connotation , since it is related to the conservation of the total @xmath251component of the spin @xmath252 which is valid for all the model with axial symmetry along the @xmath251axis .
the divergence of @xmath253 near @xmath254 is connected to the stability of hydrodynamic excitations of long wavelength ; 4 .
note the exception of @xmath255 for the component @xmath256 , because the amplitude @xmath257 goes to zero in this limit .
when applying the same method to get the asymptotic behavior of ( [ sqtzii ] ) for long times , one sees that the work is facilitated by the development of the previous section .
in fact , the frequencies @xmath258and the functions @xmath259 that appeared in ( [ sqtzii ] ) of model ii , can be mapped onto the corresponding @xmath176 and @xmath260 of model i , with the substitution : latexmath:[\[\delta \rightarrow \frac{\sqrt{\omega _ { d}^{2}+\omega _ { e}^{2}}-|\omega _ { d}|% } { \sqrt{\omega _ { d}^{2}+\omega _ { e}^{2}}+|\omega _ { d}|}\ , \qquad \omega _ { 1}\rightarrow \frac{1}{2}{\sqrt{\omega _ { d}^{2}+\omega _
{ e}^{2}}+}\frac{% we use the previous symbols @xmath216 and @xmath262 throughout this section , with the meaning given above in ( [ normalized ] ) . with these variables , the stationary and critical points ,
the inverse characteristic times @xmath214 and the frequencies @xmath263 coincide with those of model i. the expression for @xmath264 has the form ( [ series ] ) with the following values of the first nonvanishing amplitudes @xmath265 ( at a generic @xmath5-point ) : @xmath266 the corresponding phases @xmath267 are given by formulae ( [ phases ] ) .
note that the component with frequency @xmath268 decays to zero as @xmath269 ( exponent @xmath270 and @xmath271 ) , i.e. more rapidly if compared with the similar of model i .
we now summarize the behavior at the critical points : 1 . at the critical point @xmath186 , the components @xmath236 and @xmath237 are degenerate and decay to zero with exponent @xmath238 , following the law ( [ 40 ] ) with @xmath272{24}}{2\pi } \gamma ( 1/4)\frac{\delta } { % 1+\delta } ~;\ ] ] 2 . at @xmath192 ,
the components @xmath273 and @xmath274 merge into a single component with the evolution ( @xmath238 and @xmath271 in ( [ 40 ] ) ) @xmath275 \label{40a}\ ] ] with @xmath276 the long time behavior of ( [ sqtziii ] ) is much simpler than those displayed by models i and ii , for the phase function @xmath277 in ( [ defiii ] ) has only one stationary point at @xmath278 . in the asymptotic limit @xmath41 , it may be represented in the form @xmath279 where @xmath280 is given by formula ( [ defiii ] ) and @xmath281 . as it was mentioned in sec .
[ calculation ] , @xmath283 , for the initial condition @xmath119 , vanishes identically for models i , iii , iv , and only has a nonzero value for model ii . for the latter , the asymptotic expression of @xmath59for @xmath41 , has the form of ( [ series ] ) with the quantities @xmath284 and @xmath225 defined through formulae ( [ timesandfreq ] ) and ( [ normalized ] ) .
the coefficients @xmath285 for the first nonvanishing amplitudes in ( [ series ] ) , at generic @xmath5-points , are : @xmath286 where @xmath287is the correlation function defined in ( [ fourier ] ) .
we remark that the first nonvanishing term for the component with frequency @xmath288 has @xmath270 and @xmath271 , yielding a @xmath269 power law for the time decay .
the corresponding phases are given in ( [ phases ] ) in appendix [ b ] .
we make a summary of the critical behavior at singular points : 1 . at the critical point @xmath186
, the components @xmath236 and @xmath237 merge into one component with time evolution given by ( [ 40 ] ) , with @xmath238 and @xmath289 , with the amplitude @xmath290{24}}{2\pi } \
\gamma ( 1/4)\frac{\sqrt{% \delta } } { 1+\delta } \left ( 1-\delta \mp i\sqrt{1-\delta ^{2}}\right ) \ ; \ ] ] 2 . in analogous form , at @xmath192 the components
@xmath243 and @xmath237 become degenerate with the time evolution of ( [ 40a ] ) , with @xmath238 and @xmath271 , and the amplitude + @xmath291
we have studied the relaxation to the homogeneous state , of an initial excitation which has been prepared with a nonhomogeneous magnetization profile along the magnetic chain ( sim ) .
the time evolution of this excitation is probed through the calculation of the fourier component @xmath42 of the magnetization , which is analyzed as a function of the wave vector @xmath5 .
we use periodic boundary conditions for the spin hamiltonian and take the thermodynamic limit ( @xmath61 ) before studing asymptotic long times . as remarked in ref .
, the order of the limits is very important .
all the models treated here , show long time tails in the relaxation of @xmath292 , which is apparent from the asymptotic study for @xmath1 , developed in the previous sections .
this behavior manifests itself in the form of a power law decay in the long time evolution of sim s .
this slow relaxation is a remarkable result for its own sake and is probably due to the absence of dissipation in the models .
since the systems are isolated at zero temperature , the dynamics is exclusively driven by quantum fluctuations .
our calculation also shows striking differences , when one compares the behaviors of models i and ii on the one hand , with models iii and iv , on the other .
the dissimilarity lies in the presence of critical values @xmath293 for the wave vector @xmath5 , where we get a slowing down of the relaxation process . near critical points , the time scale @xmath225 , giving the relaxation of a specific oscillating component of @xmath292 , diverges with the law @xmath294 , for @xmath295 , reducing the damping of these excitations .
such behavior is not surprising for the case @xmath296 , for which the corresponding value @xmath297 is an integral of the motion , but is unusual for the case @xmath298 when @xmath55 is not conserved . in this sense ,
the criticality at @xmath299 , @xmath300 has different implications in the theory .
the position of the critical values @xmath301 and @xmath300 depends only on parameters of the hamiltonian , and in the limit @xmath302 , the exponent @xmath44 jumps discontinuously , the relaxation rate becoming slower than for excitations with @xmath303 .
this means that the critical components will be the only surviving ones at long enough times .
these long - lived excitations may be pictured as spin density waves ( charge density waves for the fermion model ) with incommensurate wave vector @xmath304 , which are developed in our gapped models as a consequence of the peierls instability@xcite .
in contrast , the dynamics of models iii and iv does not have such anomalies . there is no gap in the spectrum , and in the language of particles , the systems are always metallic and display no critical points ( except the point @xmath56 ) . in the asymptotic regime @xmath41 ,
only one oscillating component exists which displays no critical behavior .
peculiar properties of the energy spectrum seem to determine the critical relaxation phenomena .
in fact , models i and ii display an energy gap between the ground and excited states , whose size is monitored by the parameter @xmath305 ( for model ii through the transformation ( [ normalized ] ) ) .
the existence of critical points is directly related to the presence of the gap , since the latter changes the curvature of the dispersion relation for the energy . if one looks at formula ( [ sqt ] ) for the fourier component @xmath69 of the magnetization , one realizes that processes which contribute with the @xmath306 frequency ( the one that is critical ) come from transitions between both branches of the spectrum ( interband transitions ) , where a particle with momentum @xmath307 is destroyed in a given band , and a particle with momentum @xmath308 is created in the other one , with a net momentum transfer @xmath309 .
so the criticality at @xmath304 is not directly related to the one particle spectrum , but to a big density of states for those interband transitions that exchange the same momentum @xmath309 coherently ( constructive interference ) .
nesting _ condition is attained at @xmath310 , with the onset of a spin density wave for the @xmath53 magnetization .
however , we have found that the above anomalous behavior in the relaxation is not an exclusive property of the peierls transition .
in fact , other _ 1-dim _ gapped spin models , as the @xmath0 chain with anisotropic interactions , present the same slowing down of the relaxation at special points ( ref . treats the ising limit ) . a general connection between this critical behavior and the spectral properties is still lacking and currently under research@xcite .
our results are all exact , with closed analytical forms , from where the asymptotic regime for @xmath1 is obtained .
this has been done using the fermion representation of the jordan - wigner approach for spin chains .
other procedures to solve the same systems are available , the most conspicuous being the bethe ansatz method , which applied to the @xmath0 model might hint at `` hidden '' conservation laws of the several versions of the model treated here . however , it is difficult to conceive a similar analytic calculation of the relaxation properties , using bethe ansatz techniques .
we remark that integrable models displays other unusual dynamical phenomena , with anomalies in the transport properties@xcite . for the heisenberg model ,
the absence of spin diffusion has been probed by bosonization techniques@xcite and numerical calculations@xcite .
presents exact numerical computations for the spin-1/2 @xmath311 chain at @xmath312 , which include , as particular cases , the isotropic heisenberg model and the @xmath0 regime , which are gapless , and the heisenberg - ising model , with a gap in the spectrum and long - range order for the ground state .
numerical data is inconclusive due to finite - size effects , but there are strong hints of a crossover from non - diffusive to diffusive behavior when we go from the gapless region to the gapped one .
concerning the low - temperature properties , a calculation by sachdev and damle@xcite in the gapped region , shows that the diffusive behavior holds in the presence of long - range order .
it is however puzzling to admit that low energy properties of the spectrum may affect the @xmath313 behavior . from a more fundamental point of view
, we have studied the time evolution of a closed quantum many body system ( at zero temperature ) , which is prepared in an arbitrary nonhomogeneous initial state .
the dynamics is solely given by the schrdinger equation for the wave function or by the liouville equation for the density operator ( which includes the more general situation of a mixed ensemble for the initial state ) .
surprisingly , in the thermodynamic limit and for long times ( @xmath103 first than the limit @xmath1 ) , we get irreversibility in the form of a power law relaxation for the magnetization , in spite that the dynamics is unitary .
no internal interactions are present , since all the hamiltonians are reduced to the free - particle form .
the irreversibility can be entirely ascribed to interference effects , that in our calculation are handled using the stationary phase method at asymptotic long times .
due to constructive interference , gapped models develop long - lived collective excitations with the texture of spin density waves ( charge density waves for the fermion versions ) , that persist for longer times .
we believe that those structures are universal features of gapped one - dimensional models .
finally , our simple quantum systems are good candidates for _ aging _ effects , a concept that has been coined to specify nonequilibrium dynamics that depends on the initial conditions and relaxes very `` slowly''@xcite . but a complete characterization of quantum aging requires the calculation of the two - times correlation functions@xcite .
we illustrate our method with the calculation of the averages @xmath314 , ( @xmath315 ) , for models i and ii . after taking
the inverse of the bogolyubov transformation @xmath316 one obtains : @xmath317 then using definition ( [ blochop ] ) and relations @xmath318 where @xmath319 are the spin ladder operators and @xmath320 is the _ string _ operator of the jordan - wigner transformation ( [ jw ] ) , it is possible to reduce the problem to calculation of the following spin averages : @xmath321 all the averages above are taken with the initial density matrix @xmath68 .
if this is assumed to have the form ( [ inirho ] ) , then most averages that involve products of different components of spin operators vanish .
the only nonzero averages are those with products of @xmath322components of spin operators belonging to the same or neighboring lattice sites . due to this property ,
all averages in ( [ spinave ] ) vanish except for @xmath323 .
the results , for all the possible initial conditions @xmath324 , are given below : 1 .
case @xmath119 @xmath325 2 .
case @xmath120 : @xmath326 1 .
case @xmath125 : @xmath327 next , one calculates all the averages contained in the expressions for @xmath328 .
we only give the results for @xmath329 ( the remaining case @xmath120 is identical with @xmath119 ) : 1 .
case @xmath119 : @xmath330<\varepsilon ^{x}(q^{+})>_{0}~ , \\
< \beta _ { p}^{+}\beta _ { q}>_{0}=(u_{p}v_{q}+u_{q}v_{p}^{*})<\varepsilon ^{x}(q^{-})>_{0}+[u_{p}v_{q}e^{-ip}+u_{q}v_{p}^{*}e^{iq}]<\varepsilon ^{x}(q^{+})>_{0 } , \\
< \eta _ { p}^{+}\beta _ { q}>_{0}=(u_{p}u_{q}-v_{p}v_{q})<\varepsilon ^{x}(q^{-})>_{0}+[u_{p}u_{q}e^{iq}-v_{p}v_{q}e^{-ip}]<\varepsilon ^{x}(q^{+})>_{0}~ , \\
< \beta _ { p}^{+}\eta _ { q}>_{0}=(u_{p}u_{q}-v_{p}^{*}v_{q}^{*})<\varepsilon ^{x}(q^{-})>_{0}+[u_{p}u_{q}e^{-ip}-v_{p}^{*}v_{q}^{*}e^{iq}]<\varepsilon ^{x}(q^{+})>_{0}\ , \end{array } \label{ave_x}\ ] ] where @xmath331 1 .
case @xmath125 : + @xmath332 + where @xmath333 using expressions ( [ sqt ] ) and ( [ sqtiii ] ) , together with ( [ ave_x ] ) and ( [ 21 ] ) , and taking the continuum limit when @xmath334 , we get the following expression for @xmath59 : @xmath335+f^{-}(k , q)\cos [ t\omega _ { i}^{-}(k , q)]\}dk , \label{22}\ ] ] where @xmath336 , for the two models treated here . quantities in ( [ 22 ] ) are @xmath337\end{aligned}\ ] ] for @xmath119 , and @xmath338~ , \\
b^{\pm } ( k , q ) & = & \pm 4iu_{p}u_{q}\ \mbox{im}\ ( v_{p}^{*}v_{q})\end{aligned}\ ] ] for @xmath125 . substituting the specific values ( [ coeff ] ) and ( [ coeffstagg ] ) for @xmath339 into formula ( [ 22 ] ) one obtains the corresponding expressions for models
i and ii given in section ii .
in the same way , analogous results may be obtained for model iii .
we give in this appendix , formulae for the amplitudes and phases of the oscillating components in the asymptotic development ( [ series ] ) , to lowest order and for a generic @xmath5-point : j. stolze , a. nppert and g. mller , phys.rev .
b * 52 * , 4319 ( 1995 ) ; m. bhm , h. leschke , physica a * 199 * , 116 ( 1993 ) ; m. bhm , h. leschke , m. henneke , v. s. viswanath , j. stolze , g. mller , phys . rev .
b * 49 * , 417 ( 1994 ) ; k. fabricius , u. lw , and j. stolze , phys . rev .
b * 55 * , 5833 ( 1997 )
. b. m. mccoy , in _ statistical mechanics and field theory _ , edited by v. v. bazhanov and c. j. burden , ( world scientific , singapore , 1995 ) , p. 26 ; h. castella , x. zotos and p. prelovek , phys .
lett . * 74 * , 972 ( 1995 ) .
m. v. fedoryuk , _ asymptotic behavior : integrals and series _
( nauka , moscow , 1984 ) , in russian ; a. erdlyi , _ asymptotic expansions _ ( dover , new york , 1956 ) ; j. j. stamnes , _ waves in focal regions _ ( adam hilger , bristol and boston , 1986 ) .
l. f. cugliandolo and g. lozano , phys .
lett . * 80 * , 4979 ( 1998 ) ; g. m. schtz and s. trimper , europhys . lett . * 47 * , 164 ( 1999 ) ; f. igli and h. rieger , phys .
lett . * 85 * , 3233 ( 2000 ) ; n. pottier and a. mauger , physica a * 282 * , 77 ( 2000 ) . * fig .
1 * frequencies of the asymptotic components of the magnetization in ( [ series ] ) , as functions of the wave number @xmath5 , for several values of the parameter @xmath216 .
the different branches @xmath344 , for @xmath345 , are indicated by the numbers .
we also display the critical points @xmath301 and @xmath300 by solid squares .
note that the @xmath346 branch is tangent to the @xmath209 and @xmath210 branches at @xmath347 and @xmath348 respectively , but it is only defined between critical points , @xmath349 .
the complete @xmath350 curve is shown as a guide for the eye . *
2 * inverse of the relaxation times @xmath351 of the power - law decay of ( [ series ] ) , as functions of the wave number @xmath5 for the same values of the parameter @xmath216 displayed in the previous figure .
the numbers label the different branches , in close correspondence with fig .
solid squares are used to display @xmath352 , that vanishes at the critical points .
the arrows in the upper figure indicate values out of the scale . | exact analytic calculations in spin-1/2 @xmath0 chains , show the presence of long - time tails in the asymptotic dynamics of spatially inhomogeneous excitations .
the decay of inhomogeneities , for @xmath1 , is given in the form of a power law @xmath2 where the relaxation time @xmath3 and the exponent @xmath4 depend on the wave vector @xmath5 , characterizing the spatial modulation of the initial excitation .
we consider several variants of the @xmath0 model ( dimerized , with staggered magnetic field , with bond alternation , and with isotropic and uniform interactions ) , that are grouped into two families , whether the energy spectrum has a gap or not .
once the initial condition is given , the non - equilibrium problem for the magnetization is solved in closed form , without any other assumption . the long - time behavior for @xmath6
can be obtained systematically in a form of an asymptotic series through the stationary phase method .
we found that gapped models show critical behavior with respect to @xmath5 , in the sense that there exist critical values @xmath7 , where the relaxation time @xmath3 diverges and the exponent @xmath4 changes discontinuously . at those points , a slowing down of the relaxation process
is induced , similarly to phenomena occurring near phase transitions .
long - lived excitations are identified as incommensurate spin density waves that emerge in systems undergoing the peierls transition .
in contrast , gapless models do not present the above anomalies as a function of the wave vector @xmath5 . |
spintronics is the field of electronics that uses the spin of the charge carrier for the transport and manipulation of information .
preserving a spin state is fundamentally limited by the spin orbit interaction which provides the mechanism(s ) for spin relaxation .
carbon based material systems promise a weak spin orbit interaction ( i.e. , long spin relaxation times ) due to the low atomic number of carbon .
spin transport investigations have been reported for molecular systems and carbon nanotubes , see @xcite for a review , and more recently for single or multi - layer graphene @xcite . combined with the relatively large carrier mobilities
@xcite , this determines a long spin relaxation length , i.e. a long distance which a spin can travel without loosing its initial orientation . in previous experiments we deduced spin relaxation times of 150
ps ( spin relaxation lengths of 1.5 - 2 @xmath9 m ) at room temperature @xcite , value confirmed by more recent findings @xcite .
these relaxation times are shorter than expected by at least one order of magnitude , pointing to an extrinsic spin relaxation mechanism rather than to an intrinsic one . in this study
, we present all electrical spin valve and spin precession experiments aiming at determining the possible causes for such short spin relaxation times . from the experimental point of view , three major aspects are identified as discussed below . for theoretical investigations of the possible spin relaxation mechanisms we refer to @xcite .
first , in order to combat the conductivity mismatch problem @xcite and realize efficient electrical spin injection and detection in graphene , a thin aluminium oxide layer has been introduced in between the ferromagnetic co injector / detector and graphene @xcite . yet , in some experiments the contact resistances are comparable in magnitude with the graphene square resistance over one spin relaxation length and significant spin relaxation occurs via the contacts .
we quantify this effect by using an one dimensional spin diffusion model .
second , in all our previous experiments the whole graphene layer was covered by the thin aluminum oxide layer , a procedure which simplified the fabrication process . whether this oxide layer , which needs to be only underneath the co electrodes , induces extra spin relaxation was not known . in order to investigate this issue we performed experiments where the aluminium oxide barriers were present only underneath the co electrodes .
the third aspect is whether the low spin relaxation times are due to spin scattering at the edges of the graphene flakes .
therefore , we performed spin transport experiments using narrow structures ( down to 100 nm in width , 15 - 20 times narrower than the spin relaxation length ) .
this paper is organized as follows . in section [
sec : exp ] , after a short introduction regarding the functioning of a four terminal spin valve device , we present fabrication / measurement details and general electrical characteristics of the devices . in section [ sec : thexp ] we discuss modeling and experiments pertaining the spin valve and spin precession experiments .
details of the modeling of spin relaxation via the contacts are presented in the appendix .
in fig . [ fig : expsem ] we show scanning electron micrographs ( sem ) of two graphene flakes contacted by four ferromagnetic co electrodes ( 25 - 40 nm in thickness and different widths of up to 800 nm ) .
all devices were fabricated on an oxidized silicon substrate and have an aluminium oxide layer all over the graphene ( type i devices ) or only underneath the co electrodes ( type ii and type iii devices ) .
the fabrication procedure which defines the device types is described in the next section . + in fig .
[ fig : svclassic]a ) we show a schematic cross section of a lateral graphene type i spin valve device in the non - local measurement geometry @xcite .
spin signal .
note the sharp switching and the symmetry of the spin signal around zero non - local resistance.,title="fig : " ] + for simplicity , we consider the outer electrodes ( f1 and f4 ) as being non - magnetic .
the current set between f2 and f1 creates a spin accumulation ( imbalance ) in graphene at the position of f2 , i.e. the electrochemical potentials for the spin up and spin down ( @xmath10 ) channels split @xcite , fig .
[ fig : svclassic]b ) .
the spin accumulation @xmath11 diffuses away from the injection point ( @xmath12 ) and , in the linear regime , obeys the bloch equation @xmath13 where @xmath14 and @xmath15 represent the spin diffusion constant and the spin relaxation time , respectively .
the term @xmath16 describes the precession of the spin accumulation in an external magnetic field @xmath17 with the larmor frequency @xmath18 , with @xmath19 the gyromagnetic factor and @xmath20 the bohr magneton . in eq .
[ eq : diff ] we used a single relaxation time @xmath15 for reasons discussed later ( see also @xcite ) . the spin accumulation is probed non - locally by the voltage difference @xmath21 between f3 and f4 , with f3 placed at a distance @xmath22 of the order of @xmath23 with respect to the injection point .
depending on its magnetization orientation ( parallel / antiparallel ) with respect to the spin accumulation , electrode f3 is sensitive to the electrochemical potential of either the spin up channel or the spin down one , fig .
[ fig : svclassic]b ) .
the _ spin signal _ is defined as the non - local resistance @xmath24 , where @xmath25 represents the injection current . in the absence of precession @xmath26
the spin accumulation decays exponentially with the distance with the characteristic length @xmath27 , the spin relaxation length . under the assumption of high impedance ferromagnets
one can show that @xmath28 where eq . 1 from @xcite has been adapted for the two dimensional graphene .
@xmath29 represents the graphene square ( sheet ) resistance , @xmath30 is the width of the flake and @xmath31 is the spin injection / detection efficiencies of the ferromagnetic electrodes .
the @xmath32 signs correspond to the parallel ( antiparallel ) orientation of the magnetization of the injector ( f2 ) and the detector ( f3 ) .
the spin valve measurement , see fig .
[ fig : svclassic]c ) , consists in monitoring the non - local resistance as a function of an external magnetic field which manipulates the relative orientation of the injector / detector magnetizations .
the electrodes are engineered to have different widths so that , due to the shape anisotropy , they switch their magnetization orientation at different magnetic fields applied along their easy axis ( the length ) . the _ spin valve signal _
@xmath33 represents the change in resistance when the injector / detector magnetization configuration changes from parallel to antiparallel .
the exponential dependence of the spin valve signal @xmath34 on the injector - detector separation @xmath35 allows for the extraction of the spin relaxation length , the parameter of most interest . in the non - local geometry , due to the separation of the injection and detection circuits ,
magnetoresistance effects related to the charged current flow , such as anisotropic magnetoresistance and hall effects , do not superimpose on the spin signal . theoretically , the non - local resitance is determined only by the spin accumulation injected at @xmath12 which diffused towards the detector ( @xmath22 ) , i.e. , there are no other resistances ( voltages ) on top of the non - local resistance ( voltage ) . in most samples , however , a small background ( frequency and gate voltage dependent ) is measured . in a local measurement geometry ,
the injection and detection circuits share the same pair of electrodes .
the spin signals are superimposed on the graphene resistance ( typically few k@xmath36 ) and are more difficult to detect , especially if the contact or graphene resistance fluctuates .
nevertheless , there are a few local measurements reported @xcite .
the devices were fabricated on a dry oxidized , highly doped ( @xmath37=1m@xmath36 cm ) silicon substrate commercially available .
the oxide was 500 nm or 300 nm thick .
the application of a back gate voltage allows for the control of the charge carrier type and density in the graphene flakes .
the gate electrode was fabricated by etching the back side oxide of the wafer followed by the deposition of a 100 nm thick ti / au layer .
on this substrate we defined a set of ti / au markers using electron beam lithography ( ebl ) .
next , the substrate was cleaned by oxygen plasma in order to remove the polymeric residuals from the ebl step .
then , highly oriented pyrolytic graphite ( hopg from ge advanced ceramics and spi ) was cleaved with a scotch tape and firmly pressed against the substrate .
the tape is pealed off the substrate leaving a random distribution of graphite pieces among which there are also single layer graphene flakes . using an optical microscope ,
the lowest contrast flakes were selected as potential single layer candidates .
after that , the single layers were identified using atomic force microscope ( afm ) measurements .
at the same time their position with respect to the known position of the ti / au markers is noted which later on allowed for a precise positioning of the ferromagnetic electrodes . in a few cases raman spectroscopy
was performed in order to confirm the single layer selection performed by afm @xcite .
the fabrication procedure continued differently for the different types of devices as explained below .
for the type i samples , we continued with the thermal evaporation of a thin layer of al ( 6 ) in an ultra high vacuum ( uhv ) system with the substrate being liquid nitrogen cooled .
after that , the al layer was oxidized for minimum 30 minutes in the load lock of the system in an oxygen atmosphere of about 100 mbar .
next , ferromagnetic electrodes of different widths were defined using ebl followed by electron beam evaporation of co in a high vacuum system . after lift off in hot acetone for typically 10 minutes , the sample was glued on a chip carrier and electrical connections between the sample and the chip carrier were made using ultrasonic wire bonding .
the sample was then placed in a vacuum container and measured . for the type i devices , the graphene flakes are completely covered by the thin aluminium oxide . in fig .
[ fig : expafm]a)-b ) we show afm measurements of a graphene flake after the aluminium deposition . in spite of the cryogenic evaporation of al ,
the oxide layer on graphene is granular in nature .
sem images , not shown here , support this observation .
nm@xmath38 ) after al deposition ( type i device ) .
c ) and d ) height data ( @xmath39m@xmath38 ) after etching and after annealing in ar+h@xmath4 atmosphere ( type iii device).,title="fig : " ] + in order to determine the influence of the alox layer covering graphene on the spin transport properties , devices where the alox layer was deposited only underneath the co electrodes were fabricated ( type ii devices ) . in this case
an ebl step was performed before the sequential deposition / oxidation of al and the thermal evaporation of co in the same uhv system . due to the configuration of the uhv system
the al deposition on the cooled stage was done under an angle of 30@xmath40 with respect to the sample normal .
the deposition was set to take place along the electrode length with a precision of about 1 - 2@xmath40 .
this could potentially result in shadowing effects due to the resist walls , yielding non - uniform deposition of al in the patterned areas .
the type iii devices were fabricated in order to investigate spin transport in narrow graphene ribbons . in this case
all samples have the alox layer only underneath the co contacts ( same as type ii ) .
however , naturally narrow flakes are difficult to spot under an optical microscope . to fabricate devices with widths down to 100 nm
, somewhat larger flakes were selected and then oxygen plasma etched , prior to the definition of the composite alox / co electrodes .
the oxygen plasma etching step often resulted in a contaminated surface .
even though after etching the flakes were annealed in an ar(95)%:h@xmath4(5% ) atmosphere at 350@xmath40 for about two hours , large particles of unknown chemical composition still remain on the surface , see fig .
[ fig : expafm]c)-d )
. the electrical measurements were performed using a standard ac lock - in technique ( @xmath4117 hz ) in the four terminal non - local geometry .
the ac current ( 1 - 20 @xmath9a ) was set between a pair of electrodes and the in phase ac non - local voltage was measured between the other pair of electrodes .
the room temperature measurements were performed with the sample inserted in a vacuum container , base pressure of @xmath42 mbar ( measured close to the pump ) , placed in between the poles of a room temperature electromagnet .
the measurements were done in vacuum in order to reduce the hysteresis of the graphene resistance as a function of the gate voltage and to avoid the degradation of the spin signal as a function of time . in the case of atmospheric measurement conditions ,
the reduction of the spin signal in time ( on a time scale of about 24 hours ) was accompanied by an increase of the contact resistances . the low temperature ( 4.2 k ) measurements were performed in an oxford cryostat with the sample placed in a dip - stick and using a superconducting magnet . ) for type i ( left column ) and type ii ( right column ) devices.,title="fig : " ] + in fig .
[ fig : expprop ] we compare electrical characteristics of graphene and contacts for two devices , one type i and one type ii . as a function of the gate voltage ,
the four terminal measurements of the graphene resistivity ( top panels ) show a maximum which corresponds to the dirac neutrality point where the average charge density is equal to zero . at negative voltages with respect to the dirac point position the conduction takes place via holes , whereas for positive voltages the conduction happens via electrons .
the mobility of all samples presented in this study was in the range 2 - 5@xmath7 @xmath8/vs .
a clear correlation between the carrier mobilities of the type i ( completely covered by aluminium oxide ) and type ii samples ( oxide barriers only underneath the electrodes ) could not be established .
for the type iii devices the carrier mobilities were the lowest ( 2 - 2.5@xmath7 @xmath8/vs ) . in the lower panels of fig .
[ fig : expprop ] we show three - terminal contact resistance measurements as a function of contact area , gate voltage and dc current bias . in general , the contact resistances of type i and type iii devices show a large spread in values and no clear scaling with the contact area . for the type
ii samples , the spread is less and there seems to be some scaling of the contact resistance with the area . for all types of contacts , a small and irregular gate voltage dependence of the contact resistance is recorded .
some low resistance contacts show a similar behavior to graphene , i.e. the contact resistance is higher when the graphene flake is at the charge neutrality point .
most likely , this indicates that part of the graphene layer contributes to the contact resistance . with respect to the dc bias current dependence ,
some high resistance contacts show high values in zero bias conditions whereas low resistance contacts show very weak or no dc bias dependence .
unfortunately , the large spread in the overall behavior does not allow for a quantitative analysis .
we take these observations and the afm measurements ( fig .
[ fig : expafm ] ) as evidence that the contacts characteristics are determined by a random distribution of regions with increased transparency ( thinner oxide layer or pinholes ) in contrast to the ideal case of tunnel barriers .
analyzing fig . [
fig : svclassic]a ) we see that the spins injected by f2 do not only diffuse symmetrically to the left and to the right in graphene , but can also return into the ferromagnetic electrode where they loose their initial orientation very fast ( due to the strong spin orbit coupling in the ferromagnet ) .
therefore , the magnitude of the spin accumulation created in graphene is drastically reduced if the contact resistance is much lower than the graphene resistance over one spin relaxation length .
this effect is known as the conductivity mismatch problem @xcite .
taking into account the spin current drawn by the electrode , see the appendix for details , one can show that @xmath43 where the parameter @xmath44 given by @xmath45 represents spin relaxation due to the finite contact resistances @xmath46 of the injector ( f2 ) and detector ( f3 ) .
in fact , @xmath47 represents the ratio between the contact resistance and the graphene resistance over one spin relaxation length .
[ eq : rnl2 ] is in agreement with @xcite given our assumptions of negligible ferromagnet resistances with respect to the contact and graphene resistances and small polarizations @xmath31 of the injector / detector electrodes . in the limit of high impedance contacts ( @xmath48 ) eq .
[ eq : rnl2 ] reduces to eq .
[ eq : rnl ] as expected and the spin signal decays exponentially with the distance . reducing the value of the relaxation parameter ( of the contact resistance for given @xmath29 and @xmath30 ) the length dependence of the spin signal starts to deviate from the exponential form @xcite . in the case of finite @xmath46 and very long @xmath23 ( @xmath49 ) and short devices ( @xmath50 ) ,
the non - local resistance is @xmath51 .
now , the spin signal is independent of @xmath23 , it depends on the injector detector spacing as @xmath52 .
due to strong contact induced spin relaxation , spin transport is relevant only on the device length scale , the electrodes effectively cut the graphene flake with respect to the spin transport . for two values of @xmath53 m ; b ) injector - detector separation @xmath35 for @xmath54 m.,title="fig : " ] + in fig .
[ fig : rnlvsr]a ) we plotted the variation of the non - local spin valve signal with respect to the relaxation parameter @xmath44 for a device with an injector - detector separation @xmath55 m and two values for @xmath23 , 2 and 10 @xmath9 m , respectively .
the signal is normalized to the value corresponding to infinite contact resistance . for given graphene properties , a variable @xmath44 parameter represents a variable contact resistance @xmath46 .
the full circles in fig .
[ fig : rnlvsr]a ) correspond to @xmath56 m ( i.e. , @xmath57 k@xmath36 for typical values @xmath58 k@xmath36 and @xmath59 m ) .
due to the conductivity mismatch , for @xmath60 m one measures only 25 % of the totaly available signal in the ideal case of high impedance contacts . for @xmath61 m
, only 3.7 % of the ideal signal is measured .
the strong contact induced spin relaxation regime is identified as the region where the signal is proportional to @xmath62 . in fig .
[ fig : rnlvsr]b ) we show the variation of the spin signal with respect to the injector - detector separation for @xmath63 m and three values of the relaxation parameter . for @xmath64 m and @xmath65 the spin signal depends on the injector - detector separation as @xmath52 .
in this section we discuss the general characteristics of the spin valve measurements . in fig .
[ fig : svtypeivsii ] we show spin valve measurements for a type i and a type ii device for three charge density regimes : zero charge density ( dirac neutrality point ) and metallic hole and electron densities of 2 - 3@xmath6610@xmath67@xmath1 .
10@xmath67@xmath1 ) .
the y - scales have the same span for each device .
the resistance levels r1 , r2 , r3 and r4 observed for the device ii.b are due to the magnetic configuration of all four electrodes ( see fig . [
fig : svgt02steps ] and the text ) . ] due to the different contact separations , carrier mobilities and metallic charge densities , a direct comparison between these measurements is not trivial . in general , we observe large signals ( few ohms up to 50 ohms ) for the type i devices , whereas for the type ii the spin signals were a few tenths of an ohm up to a few ohms .
we identify two reasons for this difference .
first , for the type i devices the contact resistances were in the order of 10 to 200 k@xmath36 , whereas for the type ii the contact resistances were always lower , of the order of 1 to 10 k@xmath36 .
as already discussed the low impedance contacts induce significant spin relaxation , which severely reduces the magnitude of the spin signal .
second , the measurements are consistent with relatively high spin injection / detection efficiencies ( @xmath31=10 - 20% ) in the case of high impedance contacts , whereas for low impedance contacts the efficiencies were of the order of 3 - 5% . .
the magnetic configurations of all four ferromagnetic electrodes ( which give the multiple resistance levels r1 , r2 , r3 and r4 ) and the electrode positions are indicated by the arrows.,title="fig : " ] + for the type ii device in fig .
[ fig : svtypeivsii ] we observe four resistance steps which we associate with the magnetization switching of all four electrodes , graphically shown in fig . [ fig : svgt02steps ] .
let us consider that for high positive magnetic field the magnetization of all electrodes points in the `` up '' direction and that f2 injects spin up carriers in graphene .
at f1 , where the current flows from graphene to f1 , spin down injection takes place .
this partially cancels the spin up accumulation created by f2 .
both f3 and f4 electrodes probe the spin up channel and a positive non - local voltage ( resistance ) is measured since f3 is closer to the injectors than f4 , fig .
[ fig : svgt02steps]a ) . ramping down the field towards -50
mt the electrodes switch their magnetization in the reverse order of their width : f1 , f4 , f3 and f2 . when f1 changes its magnetization orientation , both f1 and f2 inject spin up carriers and the total spin accumulation increases causing a positive change in @xmath68 , fig .
[ fig : svgt02steps]b ) .
next , the detector f4 switches , probing the spin down channel , and the non - local voltage difference increases again , fig .
[ fig : svgt02steps]c ) . at around -25
mt , f3 switches its magnetization .
now , both injectors inject spin up carriers whereas the detectors probe the spin down ones which causes @xmath68 to become negative , fig .
[ fig : svgt02steps]d ) .
the last switch is that of f2 .
this configuration ( not shown ) is equivalent to the starting one with all magnetizations pointing down now and the starting level of non - local resistance is recovered .
the four changes in the non - local resistance point out that the spin relaxation length in graphene is at least of the order of the f1-f3 or f4-f2 distances ( @xmath69 m for this sample ) since otherwise the resistance step due to f1 or f4 switching their magnetization would not be observed .
additionally , the spin accumulation can extend underneath all contacts , which means that the aluminum oxide barriers are not too opaque nor too transparent . in the next sections we extract quantitative information regarding the spin relaxation length @xmath23 .
[ eq : rnl2 ] can be used to extract @xmath23 from the length dependence of the spin signal .
ideally , all one has to do is to make a series of devices which should strictly meet the following conditions .
first , the f1 and f4 electrodes do not contribute to the measurement , they are either placed at distances much larger than @xmath23 with respect to f2 and f3 or they are replaced with non - magnetic ones .
second , f2 and f3 should have identical spin injection / detection efficiencies and resistances from device to device .
in addition , all the graphene flakes should have identical transport properties ( same carrier mobility ) and the same width ( in case that enhanced spin relaxation takes place at the device edges ) .
the samples should be processed and measured in identical conditions , since it is not actually known how the substrate could influence the spin relaxation . if all these conditions are met , then the only variable left is the spacing between the injector and the detector electrodes .
this has to be varied systematically , so that the application of eq .
[ eq : rnl2 ] would allow one to extract @xmath23 and @xmath31 .
experimentally some of these conditions can be fairly met ( especially if several devices can be fitted on the same flake ) , except controlling the polarization of the electrodes and the reproducibility of contact resistance which are probably the most important parameters . in spite of these problems
we have been able to show some consistent behavior for samples where the contact resistances were more uniform @xcite , see also fig .
[ fig : t1t2]a ) . and hole / electron densities of 2 - 3@xmath6610@xmath67@xmath1 .
for these samples we calculated an average @xmath44 parameter of 1 - 2@xmath6610@xmath70 m. the spin relaxation lengths are higher in the metallic regime ( subscript @xmath71/@xmath72 for electron / hole conduction ) than at the dirac neutrality point ( subscript @xmath14).,title="fig : " ] + in this paragraph we show how the multiple resistance steps , which are due to ferromagnetic electrodes placed at various distances , can be used for a quantitative estimation of @xmath23 . for the all up magnetization configuration which gives the resistance level @xmath73 ( fig .
[ fig : svtypeivsii ] and fig .
[ fig : svgt02steps]a ) the spin accumulation present at the position of the voltage probe f3 is just the sum of two spin accumulations of opposite sign produced by f2 and f1 which have decayed exponentially with the distance ( f2-f3 and f1-f3 ) .
we neglect here the influence of the electrode f2 on the spin accumulation produced by f1 and detected by f3 .
the same arguments apply for the detector f4 but the exponential decays are for larger distances .
additionally , we note that some measurements ( including this one ) show a gate voltage dependent background resistance ( which we denote as @xmath74 ) .
note that , theoretically , the spin signals @xmath75 and @xmath76 should have identical values but opposite signs , see fig . [
fig : svgt02steps ] b ) and d ) . therefore , we can write @xmath77-[+r_{24}-r_{14}]+r_0 $ ] , where @xmath78 is the resistance level produced by the f2-f3 injector - detector pair which are placed at the distance @xmath79 and so on . applying a similar reasoning for @xmath75,@xmath80 and @xmath76
we obtain a system of four equations with five unknowns @xmath81 and @xmath74 .
in this sample the f1-f3 and f2-f4 distances were virtually equal ( @xmath69 m ) so by taking @xmath82 we can solve the system of equations and extract the length dependence of the spin signal from a single measurement .
the same analysis has been performed for another sample .
the results are plotted in fig . [
fig : ldepswitches ] together with fits according to eq .
[ eq : rnl2 ] .
the fits reveal a spin relaxation length of the order of @xmath83 m , somewhat higher in the metallic regime than for the zero charge density regime . in this section
we discuss the effect of the contact impedance on the spin precession measurements .
let us consider a spin valve device in which the ferromagnetic injector / detector electrodes have their respective magnetizations oriented parallel and a magnetic field @xmath84 , smaller than what is required to pull the magnetization of ferromagnetic electrodes out of plane , is applied perpendicular to the sample plane .
in this situation , the injected spins undergo precession ( hanle effect ) while diffusing from the injector towards the detector , see the schematics of the inset in fig [ fig : thprec2]a ) .
the detector signal oscillates with @xmath84 . ,
@xmath14 and @xmath23 .
only the parallel magnetization of the injector / detector electrodes case is shown for positive values of the perpendicular magnetic field.,title="fig : " ] + the signal is maximum at @xmath85 ( no precession ) , vanishes at fields which correspond to an average spin precession angle of 90 degrees and it is minimum when the spins have precessed an average angle of 180@xmath40 . in fig .
[ fig : thprec2 ] we present several calculated precession signals for various values of the relaxation parameter @xmath44 , the diffusion constant @xmath14 and the spin relaxation time @xmath15 .
all relevant parameters are indicated on the figures . to allow comparison
the signals have been normalized to the maximum value .
we refer to the appendix for modeling details .
next , we analyze the shape of the curves with respect to their width , defined by the 90@xmath40 precession angle , and the amount of overshoot , defined as the ratio between the minimum and the maximum values of the signal ( 180@xmath40 and 0@xmath40 precession ) . in fig .
[ fig : thprec2]a ) we plotted the precession signal for three values of @xmath44 and for fixed @xmath14 and @xmath15 ( values close to the ones corresponding to our samples ) .
with decreasing the relaxation parameter @xmath44 , i.e. decreasing the contact resistance for given graphene properties , the precession curve becomes broader and the overshoot increases in magnitude . of interest is to determine how @xmath86 influences the shape of the precession curve , and whether one can accurately extract the spin relaxation time in the case of very low impedance contacts , for example @xmath87 , which for @xmath58k@xmath36 and @xmath88 m corresponds to @xmath64 m . in fig .
[ fig : thprec2]b ) we show three curves for which @xmath64 m and the values for @xmath14 and @xmath15 are such that they give a constant spin relaxation length of 2 @xmath9 m .
this plot indicates that the overshoot depends only on @xmath23 .
extensive analysis reveals that for a given injector - detector distance @xmath35 and for @xmath89 , the overshoot depends only on @xmath23 and @xmath44 , it does not depend on the individual values of @xmath14 and @xmath15 as long as they produce the same @xmath23 .
therefore , for a given sample ( for which @xmath35 and @xmath44 are known ) the spin relaxation length can be estimated just by quantifying the amount of the overshoot . in figs .
[ fig : thprec2]c ) and d ) we plotted spin precession signals by varying either @xmath14 or @xmath15 while keeping the other parameters constant . for a given @xmath15 ,
the precession curve is wider for a higher @xmath14 .
this can be simply understood by the fact that for a higher @xmath14 ( i.e. a higher carrier mobility ) , a stronger magnetic field is required to induce 90@xmath40 precession since it takes less time for the spins to diffuse the fixed injector / detector distance .
[ fig : thprec2]d ) shows wider precession curves for smaller spin relaxation times . for high relaxation times ,
the shape of the curves is essentially the same .
this can be understood by the fact that most of the injected spins do not relax before reaching the detector and the angle of precession is mainly determined by the time it takes them to arrive at the detector , i.e. the diffusion constant and the device length . for low spin relaxation times ,
the spins following a long diffusive path have a high chance to relax before reaching the detector and , therefore , do not contribute to the measured signal . in this case , the spin signal is mainly determined by the spins following a short diffusive path , and which arrive faster at the detector ( effectively having a higher diffusion constant ) .
the overall behavior is that , for given device characteristics and dimensions , the precession measurements allow accurate extraction of the spin transport properties provided that the contacts do not induce strong spin relaxation , i.e. the relaxation parameter @xmath44 is comparable or higher than the spin relaxation length . in fig .
[ fig : t1t2]a ) we show the length dependence of the spin valve signal for three type i devices made on the same graphene flake measured at 4.2 k in the metallic hole conduction regime @xmath90 @xmath1 . for this sample
the contact resistances were more uniform ( 1 - 2 k@xmath36 ) .
the measurements indicate a spin relaxation length of 1.6 @xmath9 m . + in fig .
[ fig : t1t2]b ) we present a spin precession measurement for a 5 @xmath9 m long device from the same batch of samples .
the fit of the low @xmath91 field part of the precession curve indicates a spin relaxation length of 1.4 @xmath9 m , in very close agreement with the length dependence measurement .
this indicates that the longitudinal and transversal relaxation times ( t1 and t2 ) are similar for these devices @xcite .
therefore , we argue that the precession measurements can be used to determine the spin transport properties ( @xmath15,@xmath23 ) of graphene . the advantage over the difficult to realize length dependence of the spin signal ( due to the irreproducibility of the contact resistances or the spin injection / detection efficiencies ) is obvious : the fabrication and proper characterization of a single device . with increasing the perpendicular magnetic field above 0.5 t the magnetization of the co electrodes is slowly pulled out of plane . for fields above 1.25 t
the relative orientation of the magnetization of the co electrodes is parallel ( both magnetizations are completely out of plane ) .
now , the injected / detected spins are perpendicular to the two dimensional plane of the graphene layer and parallel to the external field ( no precession takes place ) .
we observe that the spin signal is somewhat lower than the zero field signal , when the orientation of the injected / detected spins is parallel with the sample plane .
the difference in the magnitude of the signal corresponds to a relaxation time @xmath6 smaller by @xmath92 % than @xmath5 .
similar room temperature experiments not shown here support these low temperature spin anisotropy measurements .
this indicates that the effective magnetic fields responsible for the spin relaxation lie mostly in the two dimensional plane of graphene . for a full discussion
we refer to @xcite .
however , we note that in the case of rashba or dresselhaus type spin orbit interaction the effective spin orbit fields lie exclusively in the graphene plane and , theoretical calculations indicate that @xmath93 @xcite .
= 20 - 40 k@xmath36 , @xmath94 @xmath8/vs and for a type ii device ( right column ) with @xmath46=1 - 2 k@xmath36 , @xmath95 @xmath8/vs in different conduction regimes .
the carrier concentrations are expressed in units of @xmath96 @xmath1 .
the flake dimensions are indicated on the figure .
the solid red lines through the data points are fits with the parameters indicated on the figures .
@xmath14 , @xmath15 and @xmath23 are expressed in units of @xmath97m@xmath38/s , ps and @xmath9 m respectively.,title="fig : " ] + in fig .
[ fig : precivsii ] we show spin precession measurements for a type i ( contact resistances of 20 - 40 k@xmath36 ) and a type ii sample ( contact resistances of 1 - 2 k@xmath36 ) for similar charge carrier type and density . even though the mobilities and widths of the two devices were different , the spin relaxation lengths are rather similar .
this indicates that the presence of the aluminium oxide layer on top of graphene ( type i sample ) does not influence the spin transport properties significantly .
overall the values of the diffusion constant agree well with the ones extracted from the conductivity measurements ( see @xcite for details ) , whereas the spin relaxation times are of the order of 100 - 200 ps .
only a few devices show relaxation times smaller than 100 ps . in all the measurements we observe that the spin transport properties depend on the gate voltage , i.e. on the charge density .
generally , for the same device we observe that both the spin diffusion constant @xmath14 and the spin relaxation time @xmath15 at the dirac point are approximately a factor of two smaller than at high carrier densities , see fig .
[ fig : precivsii ] . since @xmath98 ( with @xmath99 the fermi velocity and @xmath100 the momentum relaxation time )
, it seems that @xmath101 .
this behavior is expected for the elliott - yafet type spin scattering mechanism .
@xmath1 for four type iii devices of different widths.,title="fig : " ] + in this section
we discuss room temperature spin precession measurements for structures down to 100 nm in width ( type iii devices ) .
these measurements were performed in order to establish whether the short spin relaxation times may be due to enhanced spin flip processes taking place at the edges of the graphene flake .
the fabrication procedure involved an oxygen plasma etching step and it was described in the experimental section .
the values of the contact resistances for this type of etched structures show the poor reproducibility compared with the type i and ii devices .
we attribute this to the unknown residues produced during the etching process which hinders making proper contacts to graphene . in fig .
[ fig : prectypeiii ] we show a set of four precession measurements performed on different devices in the metallic hole conduction regime . in spite of similar charge transport properties , with carrier mobilities of the order to 2 - 2.5@xmath102 @xmath103/vs , the spread in the spin relaxation times is rather large . generally , the spin valve signals at the dirac point were too small to produce a useful precession curve .
for some precession curves there are a few sets of parameters that describe reasonably the experiment . in that case
we fixed the spin diffusion constant to the charge diffusion constant , leaving the spin relaxation time as the only relevant fitting parameter . in the order of increasing the graphene flake width ( 0.10 , 0.24 , 0.39 and 2 @xmath9 m ) the spin relaxation times read 40 , 200 , 154 and 60 ps respectively
no clear scaling behavior is found .
though the spin relaxation time for the narrowest device ( 100 nm ) shows the lowest value of 40 ps , a definitive conclusion with respect to a presumably strong spin relaxation at the device edges can not be drawn since spin relaxation times of 50 - 60 ps were measured for relatively wide flakes as well .
we have successfully realized all electrical injection and detection of spin accumulation in graphene at room temperature and 4.2 k. the conductivity mismatch problem has been partially overcome by the introduction of a thin aluminium oxide layer between the metallic co electrodes and the semiconducting graphene . by proper modeling , we have shown that the spin relaxation via the contacts does not dominate the spin transport properties of our devices . from the length dependence of the spin valve signal and spin precession measurements
, we have shown that the longitudinal and transversal spin relaxation times are similar .
for the 2 - 5@xmath7 @xmath8/vs carrier mobilities of our samples , we found spin relaxation times of 50 - 200 ps .
good agreement has been found between the spin diffusion constant and charge diffusion constant .
the linear dependence of the spin relaxation time on the momentum scattering time indicates that the spin relaxation mechanism is elliot - yafet type . for carrier mobilities a factor of 100 higher than our samples , a spin relaxation length of the order of 100 micrometers
is expected .
the anisotropy in spin relaxation times reveal that the effective magnetic fields due to the spin orbit interaction lie mostly in the two dimensional graphene plane . in the range of carrier mobilities of our samples , the scattering at the devices edges , for structures down to 100 nm in width , and the presence of the aluminum oxide covering the graphene flakes in some experiments do not influence the relaxation times significantly .
we would like to acknowledge b. wolfs , s. bakker , j. baas and j. buurma for technical support .
this work was financed by msc@xmath104 , nanoned , nwo ( via a pionier grant and a top talent grant ) and fom .
we consider the geometry shown in fig .
[ fig : svr ] with f1 and f4 being non - magnetic .
the injector f2 and the detector f3 are placed at the distance @xmath35 . in the frame of the two channel current model the current
is carried by two independent channels , a spin up and a spin down one . in the absence of external magnetic fields
the spin accumulation is constant in the @xmath105 and @xmath106 directions , only the @xmath107 component varies with the distance . to simplify the mathematical description we consider symmetric splitting of the electrochemical potentials of the two spin species and discuss only the spin up channel .
considering parallel orientation of the magnetization of the injector / detector electrodes and the same polarization @xmath31 for the injector / detector electrodes , we write the non - local resistance as @xmath110 where we took into account the symmetric splitting @xmath111 when applying eq . 2 from @xcite .
the constants @xmath112 are determined from the boundary conditions .
the continuity of the spin up electrochemical potential at @xmath113 gives the following two equations the spin relaxation via the contacts is taken into account when writing the spin current conservation equations as explained in the following .
the ferromagnet f2 injects a spin polarized current in graphene via tunnel barriers of resistance r@xmath115 with an efficiency @xmath31 .
the total injected spin current is @xmath116 and the total spin up current available is half , @xmath117 .
the spin up current @xmath118 diffuses to the left and to the right in graphene or flows back into the contact .
the spin up current flowing in graphene is written in the form @xmath119 . here
@xmath71 is the electron charge , @xmath120 and @xmath30 are the conductivity and the width of the graphene flake .
the ( unwanted ) spin up current through the contact is written as @xmath121 .
here we have considered @xmath122 , with @xmath123 the effective resistance the up spins encounter when going back into contact .
the spin current conservation equations at @xmath113 read here we have used the fact that the spin current flowing away from the contact ( the exponential decaying electrochemical potential from the contact point of view ) is positive .
the boundary conditions give a system of four equations with four unknowns from which the constants @xmath112 are extracted .
using @xmath125 , we find the expression for @xmath126 , given by eq .
[ eq : rnl2 ] . in the presence of an external magnetic field applied in the @xmath106 direction the general solutions of the spin up electrochemical potential are of the form @xcite @xmath127 where the @xmath128 signs correspond to the positive ( negative ) @xmath107 direction and @xmath129 and @xmath130 are given by @xmath131 and @xmath132 .
since the electrochemical potential can not increase when @xmath107 approaches @xmath133 , the solutions for @xmath108 ( @xmath134 ) contain only the exponentially decaying terms , i.e. the terms containing @xmath135 for @xmath108 and the terms @xmath136 for @xmath134 .
for @xmath137 the solutions contain six terms so that the back flow of spins due to the presence of the detector is taken into account .
there are in total twelve coefficients ( @xmath138,@xmath139,@xmath140 , @xmath141,@xmath142,@xmath143,@xmath144,@xmath145,@xmath146,@xmath147,@xmath148,@xmath149 ) that describe the solutions .
these are found from the boundary conditions : the continuity of electrochemical potentials and the spin current conservation , laws which we apply for each @xmath150 and @xmath106 component .
the continuity of the electrochemical potentials give in total six equations , three for injection and three for detection .
the spin current conservation laws give another six equations and they are written in a similar manner as it was described above . except the @xmath107 component
@xmath118 all the other spin current components @xmath151 are null .
the spin up electrochemical potential present at the detector ( @xmath107 component ) is given by @xmath152 the boundary conditions give a system of twelve equations with twelve unknowns . for given @xmath15 and @xmath14 we solve the system of equations numerically to find the coefficients @xmath148 and @xmath149 for each value of the magnetic field .
@xmath31 is just a multiplication factor .
we calculate a series of precession curves for a mesh given by different values of @xmath15 and @xmath14 . from this set of curves
we choose the one that shows the smallest deviation from the experiment by the least mean squares method .
the parameters @xmath153 corresponding to this curve represent the best fit .
e. w. hill , a. k. geim , k. novoselov , f. schedin , and p. blake , ieee trans .
* 42 ( 10 )
* , 2694 ( 2006 ) n. tombros , c. jozsa , m. popinciuc , h. t. jonkman , and b. j. van wees , nature * 448 * , 571 ( 2007 ) .
s. cho , y .- f .
chen , and m. s. fuhrer , appl .
. lett . * 91 * , 123105 ( 2007 ) m. nishioka , and a. m. goldman , appl . phys . lett . * 90 * , 252505 ( 2007 ) m. ohishi , m. shiraishi , r. nouchi , t. nozaki , t. shinjo , and y. suzuki , jpn
. j. appl .
* 46 * , no .
25 , l605-l607 ( 2007 ) w. h. wang , k. pi , y. li , y. f. chiang , p. wei , j. shi , and r. k. kawakami , phys .
* 77 * , 020402(r ) ( 2008 ) n. tombros , s. tanabe , a. veligura , c. jozsa , m. popinciuc , h. t. jonkman , and b. j. van wees , phys .
lett . , * 101 * , 046601 ( 2008 ) c. jozsa , m. popinciuc , n. tombros , h. t. jonkman , and b. j. van wees , phys .
lett . , * 100 * , 236603 ( 2008 ) h. goto , a. kanda , t. sato , s. tanaka , y. ootuka , s. odaka , h. miyazaki , k. tsukagoshi , and y. aoyagi , appl .
, * 92 * , 212110 ( 2008 ) c. jozsa , m. popinciuc , n. tombros , h. t. jonkman , and b. j. van wees , phys . rev
. b , * 79 * , 081402(r ) ( 2009 ) . m. shiraishi , m. ohishi , r. nouchi , t. nozaki , t. shinjo , and y. suzuki , arxiv:0810.4592 w. han , w. h. wang , k. pi , k. m. mccreary , w. bao , y. li , f. miao , c. n. lau , and r. k. kawakami , phys . rev
* 102 * , 137205 ( 2009 ) w. han , k. pi , w. bao , k. m. mccreary , y. li , w. h. wang , c. n. lau , and r. k. kawakami , appl .
lett . * 94 * , 222109 ( 2009 ) k. s. novoselov , a. k. geim , s. v. morozov , d. jiang , y. zhang , s. v. dubonos , i. v. grigorieva , and a. a. firsov , science * 306 * , 666 ( 2004 ) .
a. k. geim , and k. s. novoselov , nature mater .
* 6 * , 183 - 191 ( 2007 ) .
k. i. bolotin , k. j. sikes , z. jiang , m. klima , g. fudenberg , j. hone , p. kim , and h. l. stormer , solid state commun . * 146 * , 351 ( 2008 ) c. l kane , and e. j. mele , phys . rev . lett . * 95 * , 226801 ( 2005 ) d. huertas - hernando , f. guinea , and a. brataas , phys . rev .
b * 74 * , 155426 ( 2006 ) h. min , j. e. hill , n. a. sinitsyn , b. r. sahu , l. kleinman , and a. h. macdonald , phys . rev .
b * 74 * , 165310 ( 2006 ) y. yao , f. ye , x - l .
qi , s - c .
zhang , and z. fang , phys .
b * 75 * , 041401(r ) ( 2007 ) d. huertas - hernando , f. guinea , and a. brataas , arxiv:0812.1921v2 ( 2008 ) m. gmitra , s. konschuh , c. ertler , c. ambrosch - draxl , and j. fabian , arxiv:0904.3315v2 ( 2009 ) c. ertler , s. konschuh , m. gmitra , and j. fabian , arxiv:0905.0424v2 ( 2009 ) a. h. castro neto , and f. guinea , phys
103 , 026804 ( 2009 ) g. schmidt , d. ferrand , l. w. molenkamp , a. t. filip , and b. j. van wees , phys .
b * 62 * , r4790 ( 2000 ) e. i. rashba , phys . rev .
b * 62 * , r16267 ( 2000 ) a. fert , and h. jaffres , phys . rev .
b * 64 * , 184420 ( 2001 ) d. l. smith , and r. n. silver , phys . rev .
b , * 64 * , 045323 ( 2001 ) s. takahashi , and s. maekawa , phys . rev .
b , * 67 * , 052409 ( 2003 ) a. van staa , j. wulfhorst , a. vogel , u. merkt , and g. meier , phys .
b * 77 * , 214416 ( 2008 ) t. kimura , and y. otani , phys
lett . * 99 * , 196604 ( 2007 ) m. johnson , and r. h. silsbee , phys .
b , * 37 * , 5312 ( 1988 ) f. j. jedema , h. b. heersche , a. t. filip , j. j. a. baselmans , and b. j. van wees , nature * 416 * , 713 ( 2002 ) p. c. van son , h. van kempen , and p. wyder , phys .
58 * 2271 ( 1987 ) f. j. jedema , m. v. costache , h. b. heersche , j. j. a. baselmans , and b. j. van wees , appl .
. lett . * 81 * , 5162 ( 2002 ) n. tombros , ph.d .
thesis , university of groningen , the netherlands j. fabian , a. matos - abiague , c. ertler , p. stano , and i. uti , acta physica slovaca * 57 * , 565 - 907 ( 2007 ) . | spin transport experiments in graphene , a single layer of carbon atoms , indicate spin relaxation times that are significantly shorter than the theoretical predictions .
we investigate experimentally whether these short spin relaxation times are due to extrinsic factors , such as spin relaxation caused by low impedance contacts , enhanced spin flip processes at the device edges or the presence of an aluminium oxide layer on top of graphene in some samples .
lateral spin valve devices using a field effect transistor geometry allowed for the investigation of the spin relaxation as a function of the charge density , going continuously from metallic hole to electron conduction ( charge densities of @xmath0@xmath1 ) via the dirac charge neutrality point ( @xmath2 ) .
the results are quantitatively described by a one dimensional spin diffusion model where the spin relaxation via the contacts is taken into account .
spin valve experiments for various injector / detector separations and spin precession experiments reveal that the longitudinal ( t@xmath3 ) and the transversal ( t@xmath4 ) relaxation times are similar .
the anisotropy of the spin relaxation times @xmath5 and @xmath6 , when the spins are injected parallel or perpendicular to the graphene plane , indicates that the effective spin orbit fields do not lie exclusively in the two dimensional graphene plane .
furthermore , the proportionality between the spin relaxation time and the momentum relaxation time indicates that the spin relaxation mechanism is of the elliott - yafet type . for carrier mobilities of 2 - 5@xmath7 @xmath8/vs and for graphene flakes of 0.1 - 2 @xmath9 m in width
, we found spin relaxation times of the order of 50 - 200 ps , times which appear not to be determined by the extrinsic factors mentioned above . |
soon after the @xmath9-ray flash , optical and x - ray afterglows of gamma ray bursts ( grb ) are among the brightest sources in the sky at cosmological redshifts .
more than thirty grb redshifts have been measured to date and their distribution ranges from 0.168 ( greiner et al .
2003 ) to 4.5 ( andersen et al . 2000 ) with a median of z @xmath10 ( excluding grb 980425 if at z=0.0085 ) .
follow - up observations of the grb localized by bepposax , the ipn and by the rxte and hete2 satellites , show that tens of minutes after the grb the optical afterglow can be as bright as r=14 - 16 mag ; a few hours later it can still be as bright as r=17 - 19 mag .
an exceptionally bright example is the case of grb 030329 for which the optical afterglow reached r=12.7 mag at 1.5 hours from the grb and it decreased down to r=19 mag after @xmath11 days ( e.g. price et al . 2003 ; stanek et al .
. the x - rays afterglow can be as bright as the crab nebula a few minutes after the grb , while 5 - 8 hours later it can be as bright as a bright agn , i.e. @xmath12mcrabs ( see frontera et al .
2000 , also fiore et al .
2000 for an estimate of the grb afterglows logn - logf in the range 0.5 - 2.0 kev , where fluxes are integrated from minutes up to hours after the grb event ) .
this opens up the perspective for gathering spectra of high quality of sources at cosmological redshifts , provided that the afterglow can be observed in such short time scales .
spectral studies of grb afterglows can provide crucial data on the environments in which grb occurs .
this can give us both important constraints on the nature of the grb progenitor and on the interstellar matter ( ism ) in the grb host galaxies .
candidate grb progenitor include : collapsars ( woosley 1993 , paczynski 1998 ; macfadyen & woosley 1999 ) , supranovae ( vietri & stella 1998 ) , mergers of two neutron stars or a neutron star and a black hole ( e.g. eichler et al .
1989 ; paczynski 1990 ; narayan , paczynski & piran 1992 ; meszaros & rees 1992 ) .
multiwavelength afterglow observations suggest the association of `` long grb '' ( for which the duration of the prompt event peaks at @xmath13s , e.g. norris , scargle & bonnell 2000 ) progenitor with massive stars .
some indications are , among others : i ) the grb locations relative to their host galaxy ( bloom , kulkarni & djorgovski 2002a ) ; ii ) the detection of iron lines in the x - ray afterglow spectrum of five grb ( piro et al .
1999 , 2000 ; yoshida et al .
1999 ; antonelli et al .
2000 ; amati et al .
2000 ) ; iii ) the presence of late re - brightening in the light curves of some afterglows , interpreted as the evidence of an underlying supernova ( e.g. bloom et al .
2002b ) ; iv ) the presence of supernova features emerging in the optical afterglow spectra of grb 030329 ( stanek et al .
2003 , hjorth et al .
2003 ; kawabata et al .
2003 ) and of grb 021211 ( della valle et al . 2003 ) . because of the short lifetimes of massive stars , it is likely that their collapse happens close to their star forming region , that is , close to a dense and dusty environment ( e.g. fryer , woosley & hartman 1999 ; perna & belczynski 2002a )
therefore , one important element in favor of massive star progenitor would be the evidence of strong dust extinction .
indeed , the non detection of the @xmath14 of x - ray afterglows at optical wavelengths ( dark grb ) , favors this hypothesis .
however , the optical spectra of grb afterglows , do not generally show strong reddening ( simon et al .
2001 ; galama & wijers 2001 ) .
moreover , lazzati , covino & ghisellini ( 2002 ) show that even if dark grb are located in the innermost regions of galactic - like molecular clouds , in several cases the corresponding visual extinction is not enough to hide their optical afterglows .
therefore , a more accurate study of the grb environment is needed .
prompt observations of grb afterglows offer a new and distinctive path for the study of the matter in the immediate surroundings of the grb , r@xmath61 - 10 pc ( perna & loeb 1998 ; bttcher et al .
1999 ; perna et al . 2002b ; perna & lazzati 2002c ; fox et al .
2003 ) and in the grb host galaxy ( ciardi & loeb 2000 ; fiore 2001 ; bloom et al . 2002a
; savaglio , fall & fiore 2003 ) x - ray and optical - uv spectroscopy can tell us about gas density and ionization status , metal abundances , dust content and kinematics of the matter in the grb environment .
this can be done by using both emission features and absorption features ( see e.g. kumar & narayan 2003 ; bttcher et al .
1999 ; piro et al .
2000 ; ghisellini et al .
although absorption features are more difficult to study than emission lines , they carry with them unbeatable information , because they probe matter along a single beam , i.e. along the line of sight to the background beacon .
this greatly reduces the complications related to the matter geometry and dynamics , which strongly affects emission line studies .
even when such emission or absorption features are not detected or are not resolved , x - ray spectra can provide useful information from measures of low energy cut - offs due to photoelectric absorption moreover , grb afterglows can be used to study the ism of their host galaxies .
high redshifts galaxies have so far been studied mainly via ` lyman - break galaxies ' ( e.g. steidel et al . 1999 ) , and via galaxies which happen to lie along the line of sight to bright quasars , notably the ` damped lyman - alpha ' systems ( dla , e.g. pettini et al .
1997 , 1999 ) .
however , probably neither of these methods gives an undistorted view of the bulk of the typical high redshifts galaxy .
lyman - break galaxies are characterized by pronounced star - formation and their inferred chemical abundances may be related to these regions rather than being representative of typical high z galaxies ( see e.g. steidel et al .
dla studies are likely to be biased against dusty systems , since these would hide the background quasars in color - based surveys .
furthermore , dla are often identified with dwarf or low surface brightness galaxies and therefore may not be representative of the full high z galaxy population .
on the other hand , grb seem to occur well within the main body of their host galaxies , not in the outer halos ( bloom et al .
2002a ) , therefore , grb afterglows can provide a powerful , independent tool to study the ism of high redshifts galaxies ( castro et al .
2003 , mirabal et al .
2002 , fiore 2001 ; savaglio et al .
the latter authors compared metal column density in grb host galaxies with those of dlas .
they find that while the column densities of iron and magnesium are similar or slightly higher than that in dlas , the column of zinc are significantly higher .
since iron and magnesium , unlike zinc , tend to be depleted in dust , this finding indicates that a significant fraction of these elements is in dust since not observed through absorption lines , thus grb are probably probing denser and dustier regions , than dlas .
we present in this paper the results of a systematic spectral analysis of 13 grb afterglows observed by bepposax in x - rays , and by several ground based telescopes in optical - uv , aimed at measuring and constraining the x - ray absorption along the line of sight as well as the extinction in the optical - uv bands .
the comparison of the results obtained in the two spectral band will provide constraints on the dust to gas ratio of the absorbing matter and on the dust properties , as seen from several hours to a few days from the grb event .
the paper is organized as follows : 2 shows the bepposax x - ray afterglow selected sample ; 3 presents the x - ray data reduction and spectra extraction procedures , along with a discussion on the systematics affecting the column density measurements ; 4 presents the optical photometry of 9 grb afterglows of our sample with a detected optical counterpart ; results are presented and discussed in 5 and 6 respectively .
we selected from the bepposax archive all the x - ray afterglows observed during a period of four years ( from february 1997 to february 2001 ) that have enough counts in the medium energy concentrator spectrometer ( mecs ) to extract a spectrum and to perform a spectral analysis , that is , with a signal to noise ratio @xmath15 in the 2.0 - 10.0 kev .
we also required that the low energy concentrator spectrometer ( lecs ) 0.1 - 2 kev spectra contain at least 20 counts , or that the counts expected in this lecs band , based on simulations performed using the best fit power law model obtained from the mecs spectrum and assuming a column density equal to the galactic value along the line of sight , are also smaller than the above threshold at the @xmath16 confidence level , not to miss bright and highly obscured afterglows .
the number of afterglow observations satisfying these requirements is 13 .
table [ tbl-1 ] summarize the selected grb x - ray afterglows along with the galactic hydrogen column density along the line of sight to each grb , as determined from the coarse , @xmath17 deg resolution , maps of dickey & lockmann ( 1990 ) , the host galaxy redshifts if known , the r.a . and
coordinates and the optical detection .
redshifts measurements are from optical spectroscopy except for grb 000214 ( antonelli et al .
2000 ) , for which a redshifts of @xmath18 was inferred assuming that the observed emission line is the iron @xmath19 line at 6.7 kev ( rest frame ) .
the grb afterglows observations were performed with the bepposax narrow field instruments ( nfi ) , lecs 0.1 - 10 kev ( parmar et al .
1997 ) , mecs 1.3 - 10 kev ( boella et al .
1997 ) , hpgspc 4 - 60 kev ( manzo et al . 1997 ) and pds 13 - 200 kev ( frontera et al .
lecs ( 1 unit ) and mecs ( 3 units ) are imaging gas scintillation proportional counters , the hpgspc is a collimated high pressure gas scintillation proportional counter and the pds consists of four collimated phoswich units .
we report here the analysis of only the lecs and mecs data because grb afterglows are usually many times fainter than the pds and hpgspc internal backgrounds .
after 1997 may 6@xmath20 mecs observations were performed with mecs2 and mecs3 units only because on this date a technical failure caused the switch off of the mecs1 unit .
for all the selected grb sample , data from the mecs2 and mecs3 units were combined together , after gain equalization , to increase the signal to noise ratio . in the case of grb 970228
( that is the only grb in our sample observed before 1997 may ) we used the data from mecs3 only , since the source happened to lie just below the beryllium strongback support of the mecs1 and mecs2 windows ( which basically absorbs all photons below about 5 kev ) .
standard data reduction was performed using the saxdas software package version 2.0 following fiore , guainazzi & grandi ( 1999 ) .
lecs and mecs spectra were extracted from regions of radii between 3 and 8 arcmin , chosen to maximize the count rate in the 0.5 - 2 kev and 2 - 10 kev band respectively .
the choice of the extraction radius depends on the source and background intensity , and on the instrument point spread function , psf .
the 90% encircled energy generally is within 3 - 4 arcmin for the mecs and within 6 - 8 arcmin for the lecs ( fiore et al .
1999 ) .
background spectra were extracted in detector coordinates from high galactic latitude ( @xmath21 deg ) `` blank fields '' ( averaged lecs and mecs pointing of sky regions without strong x - ray sources , for a total exposure of about half a million of seconds in both instruments , fiore et al .
1999 ) using the same source extraction regions .
`` local '' background can not be safely used when analyzing lecs and mecs spectra , because of the relatively broad psf and because the background strongly varies with the position in the detector ( fiore et al .
1999 and chiappetti et al .
we have compared the mean level of the background in the lecs and mecs `` blank fields '' observations to the mean level of the background in the grb observations using source free regions at various positions in the detectors .
the `` local '' mecs background count rate is within 2% than that in the `` blank fields '' . on the other hand ,
the `` local '' lecs background count rate differs from that of the `` blank fields '' by up to 20% in 7 cases . in these cases
the `` blank fields '' background spectrum was multiplied by appropriate correction factors before subtraction from the source spectrum .
usually , the position of grb afterglows is not know with accuracies better than several arcminutes at the moment of the bepposax nfi pointing . on these scales
the vignetting of the telescopes varies significantly ( see e.g. fiore et al .
furthermore , the lecs thin polypropylene window is supported by a mesh of grids of tungsten wires that introduces a complicated pattern of obscuration .
the actual value of obscuration strongly depends on the exact position of the source centroid in detector coordinates .
the spacing of the thicker wire grid is of about 2 arcmin , much less than the uncertainty on the afterglow position .
this means that the afterglow position in detector coordinates with respect to the thick wire grid can not be known ` a priori ' as well as the level of the grid obscuration .
for these two reasons we therefore not used the standard on axis effective area files , but we evaluated the instrument effective area at the afterglow position in detector coordinates ( we used the sax / lecs data analysis software lemat ) .
these ` ad hoc ' effective area file takes correctly into account the energy dependent telescope vignetting and the photons absorbed by the lecs support mesh wires . on axis redistribution matrices , from the december 1999 release
have been used for both lecs and mecs , since these depend very little on the source position in the detectors .
table [ tbl-2 ] gives the lecs and mecs exposure times , optimum extraction radius , background and source count rates .
note that lecs exposure times are usually smaller than mecs ones since it is operated during dark time only .
table [ tbl-3 ] gives the mecs off - axis angles , the time interval between the grb trigger and the nfi pointing , the lecs ` raw ' detector pixel coordinates of the source , at which the effective area has been computed . to get gaussian statistics and
so to ensure the applicability of the @xmath8 test to evaluate the goodness of our fits , lecs and mecs spectra were rebinned to obtain at least 20 counts per energy channel .
we then performed simultaneous lecs and mecs spectral fitting using the bands 0.1(or 0.4 kev , see below ) 4.0 kev and 1.6 - 10.0 kev in the two instruments respectively ( see 5.1 ) .
errors quoted in this paper for the column density and the spectral index are 90% confidence intervals for two parameters of interest ( @xmath22 ) , unless differently specified .
the knowledge of the galactic column density along the line of sight to each burst is crucial for our study .
relatively small errors in the evaluation of the galactic columns would translate in a larger errors on the rest frame absorption , since best fit @xmath2 values scale with redshift as @xmath23 ( the photoelectric cross section scaling with energy roughly as @xmath24 ) .
the values in table [ tbl-1 ] have been obtained interpolating on a grid with spacing of 3 degrees ( dickey & lockmann 1990 ) .
it is well known that the galactic column density can vary on much smaller scales ( elvis , lockman & wilkes 1989 ) .
to evaluate the average magnitude of this variation we compared the galactic column density obtained using the coarse 3 degrees 21 cm emission maps along the line of sights of a sample of agn for which an accurate 21 cm emission measurements were already performed by using a 21 arcmin beam nrao 140 ft telescope at green bank by elvis et al .
we found that the average ratio between the two determinations is close to 1 and that the standard deviation is of 0.16 , independent of the value of the galactic column density .
the highest positive deviation found is of 50% .
we therefore performed for each grb three series of fits , one with the nominal galactic @xmath2 value and the other two by increasing this value by 16% and 50% respectively , to check whether additional extra galactic absorption would still be significantly even in case the galactic column has been underestimated in the dickey & lockmann ( 1990 ) maps .
another tracer of the column density of matter in a given direction is the infrared emission at @xmath25 .
this can therefore be used as a further check of the galactic column density values given in table [ tbl-2 ] .
we used the iras faint source survey to search for any `` cirrus '' contamination along the line of sight of each grb .
we used a 10 to 30 arcmin radius search region to measure the @xmath25 mjy / sr brightness @xmath26 . by adopting @xmath27 mjy / sr 10@xmath28 @xmath4 as typical galactic dust to gas ratio ( reach , heiles & koo 1993 ; heiles ,
reach & kooh 1988 ; de vries , heithausen & thaddeus 1987 ) we estimated the implied @xmath29 values and then compared them with the maps of dickey & lockmann ( 1990 ) .
all the inferred @xmath29 values reveal no absorption excess due to presence of cirrus along the analyzed line of sight .
the background subtraction procedure previously described works well if the galactic column density along the line of sight to the grb , is not much higher ( or lower ) than that in the `` blank fields '' .
otherwise the background low energy spectrum ( e.g. below 0.3 kev ) would be very different from that of the `` blank fields '' .
blank fields are all at high galactic latitude , and the average galactic column density along their line of sight is @xmath30 @xmath4 .
several afterglows in our sample have a galactic column density much higher than this value ( see table [ tbl-2 ] ) . adopting the `` blank field '' background spectrum for those afterglows at galactic @xmath31@xmath4 it would results in an over - subtraction of the background below 0.3 - 0.4 kev and in a consequent overestimate of the intrinsic absorbing column . as an example
, the 0.2 kev flux would decrease by @xmath32 by increasing @xmath2 from @xmath33@xmath4 to @xmath34@xmath4 . at 0.4 - 0.5
kev the flux would decrease by much smaller amount , of the order of 10 - 20% .
we therefore decided to limit the spectral analysis of the afterglows with galactic column densities higher than @xmath34 @xmath4 to energies higher than 0.4 kev .
a constant normalization factor must be introduced in the fitting models in order to take into account the systematic error in the knowledge of the absolute lecs and mecs normalization factor ( fiore et al .
1999 ) . assuming the mecs as reference instrument , the expected factor between lecs and mecs for constant sources is about @xmath35 .
however , grb afterglows are not constant sources : their flux decreases usually as a power law with index @xmath36 in the range @xmath37 ( see e.g. costa et al . 1999 ; piro 2002a ) .
moreover , the non completely simultaneous observations of the lecs and mecs instruments due to the switch - off of the lecs during part of the orbit , introduces a further complication in the determination of the normalization factor .
we have therefore decided to leave completely free the normalization factor in the fits .
we are aware that letting the normalization factor be free to vary introduces an additional degree of freedom in the fit and consequently increases the uncertainty on the derived @xmath2 .
in fact , a truly curved spectral shape toward the lower energies could be mimicked , in spectra of low statistics , by a lower mecs to lecs normalization factor . as an example a factor of @xmath38 lower normalization factor , could mimic a factor of @xmath39 lower @xmath2 .
the @xmath40 lecs - mecs normalization factor confidence intervals are reported in tables [ tbl-5 ] and [ tbl-6 ] .
for 9 of the 13 selected grb with x - ray afterglow an optical afterglow was also discovered ( see table [ tbl-1 ] ) .
we collected the optical and near infrared magnitudes of these nine grb from the literature to study the extinction properties through the comparison of dust extinction models to the afterglow spectral energy distribution ( sed ) .
we selected , among the large number of papers reporting photometric data on a given grb optical afterglow , those works containing the widest possible frequency coverage and the best and most complete documentation .
we list and discuss the references used for each grb afterglow in the appendix section .
observations in different bands are usually performed at different times . since optical afterglows
are sources with fluxes rapidly decaying in time , their magnitudes have been extrapolated to a single specific time @xmath41 .
the @xmath41 was selected to be as close as possible to the first optical observations , to keep the statistical uncertainties on the observed magnitudes small .
the extrapolation of the observations at different bands at @xmath41 was performed by adopting the best fit decay index quoted in the literature .
corrections for the galactic interstellar extinction have been performed , unless explicitly specified , using the iras 100 @xmath42 e(b - v ) maps by schlegel , finkbeiner & davis ( 1998 , sfd ) and we derived the extinction at different wavelengths using the extinction curve parameterization by cardelli , clayton & mathis ( 1989 ) assuming @xmath43 .
table [ tbl-4 ] gives for each grb the magnitude in each photometric bands , extrapolated at @xmath41 ( see notes on table [ tbl-4 ] ) and corrected for galactic extinction .
we conservatively kept the error on the magnitudes to 0.03 mags when the published error is smaller that this value .
grb 000926 is not part of our bepposax sample because its x - ray data has poor statistics ( see section 2 ) .
however , it has relatively high quality chandra data and a complete optical - nir photometry .
furthermore , grb 000926 has been observed at a medium - high resolution with keck esi ( castro et al .
2003 ) and savaglio et al .
( 2003 ) were able to derive information on the metal column densities from this spectrum .
for these reasons , we decided to include it in our analysis .
we take the photometry of grb 000926 ( u , b , v , r , i , j , h and k ) from fynbo et al .
( 2001 ) and the galactic extinction in the direction of this burst from the sfd maps as @xmath44 .
the redshift of grb 000926 is @xmath45 ( castro et al .
2003 ) .
the optical to near infrared seds of the nine grb afterglows are shown in figure [ f1 ] .
in this section we summarize the results of afterglow spectral fitting , performed separately in the x - ray and in the optical - nir energy ranges .
our goal is to estimate the absorption in the soft x - ray energy and extinction in the optical band . at the time of the nfi observations ( on average after @xmath11 hours after the grb event ,
see table [ tbl-3 ] ) , afterglow x - ray spectra are usually well described by simple power laws .
we therefore performed a spectral fit by adopting a simple power law model with two photoelectric absorption component : one representing the absorption from our galaxy and the other representing the additional absorption along the line of sight .
the quality of the data does not allow to fit for the absorber redshift .
we therefore present in table [ tbl-5 ] ( 0.1 - 10 kev band ) and table [ tbl-6 ] ( 0.4 - 10 kev band ) best fit parameters obtained with the additional absorber at the redshift of the grb , or at z=1 when the redshift of the grb is unknown , since grb redshift distribution peaks at about this value ( djorgovski et al .
2001 ) .
tables [ tbl-5 ] and [ tbl-6 ] also give the results of the fits performed fixing the redshift of the additional absorbing column to zero .
figure [ f2 ] shows the best fit additional column density at z=0 against the galactic column .
no trend is evident , supporting the robustness of our analysis .
figure [ f3 ] shows the best fit additional absorbing column density measured at z=0 and at z = z@xmath46 ( or z=1 for the grb with unknown redshift ) as a function of the redshift . since the photoelectric cross section scales with energy roughly as @xmath24 , effective @xmath2 values scale with redshift as @xmath47 .
consequently , fitting the absorption in the grb frame the resulting column is highly magnified with respect to that fitted at z=0 and small errors in the evaluation of the galactic column would translate in a strong trend of the additional column with z toward high @xmath2 values at high z. this is not observed , supporting the robustness of our analysis . to evaluate the statistical significance of a given additional @xmath2 measurement we used the f test ( bevington & robinson 1992 ) .
we found that in two cases , grb 990123 and grb 010222 , the improvement of the @xmath8 including the additional absorber to the fit is highly significant ( figure [ f4 ] ) , probability @xmath48 99.99% . in both cases
the probability remains higher than 99.99% increasing @xmath49 by 16% ( that correspond to 68 % of uncertainty on our estimate of the galactic column along the line of sight , see 3.1.1 ) .
even increasing @xmath49 by 50% ( the highest deviation we expect from our estimates of @xmath49 ) the probability remains higher than 99.99% for grb 010222 and higher than 98 % for grb 990123 . in three other cases ( grb 980703 , grb 990510 and grb 001109 ) the improvement of the @xmath8 including the additional absorber to the fit is marginally significant ( probability of 97% ) . in all these cases increasing @xmath49 by 16% reduces the statistical significance of the @xmath8 improvement below the threshold of 95% .
we conclude that the detection of absorption in addition to the galactic one is robust against statistical and systematic uncertainties in 2 of the 13 grb analyzed ( grb 990123 and grb 010222 ) .
we note that the spectra of grb 990123 and grb 010222 are also those with the best statistics .
however , fitting the spectra of two other grb afterglows with lower but still reasonably good statistics , grb 971214 and grb 970228 , including an additional absorber , does not produce a significant improvement in @xmath8 ( probability of @xmath50 and @xmath51 respectively ) .
for the rest of the sample , additional absorption may be common , although the quality of the data in the majority of the cases does not allow highly significant detection .
assuming that the most of the matter responsible of absorption is at the grb redshift , we found @xmath2 rest frame values in the range @xmath52 @xmath4 ( 90% confidence level ) , similar to the column usually observed toward the disk and the bulge of our galaxy ( see figure [ f3 ] ) .
a more detailed analysis has been performed for this burst which is the second brightest x - ray afterglow of the selected sample and shows a peculiar behavior at low energies ( figure [ f4 ] ) .
we note from table [ tbl-5 ] that our @xmath2 measurement for this burst , is somehow lower , but still consistent with the @xmath2 reported by int zand et al .
( 2001 ) ( @xmath53 @xmath4 ) using the same bepposax data ( 0.3 - 10.0 kev ) . on the other hand ,
our measure is higher than that one reported by bjrnsson et al .
( 2002 ) @xmath54@xmath4 , using chandra observations ( 0.5 - 10.0 kev ) .
we note , however , that the chandra data are affected by pile - up .
this introduces an additional systematic uncertainty in the spectral parameter estimates . the low energy absorption for this burst , may show possible complexities and/or variability . in fact
, the presence of a neutral absorber produces a good fit at energies @xmath55 kev but at lower energies an excess with respect to the best fit model is observed ( figure [ f5 ] ) .
there are at least two possible explanations for these residuals : 1 ) the recovery of the spectrum below the energies of strong absorption edges of ovii ( 0.74/(1+z)=0.30 kev ) , oviii ( 0.87/(1+z)=0.35 kev ) ; 2 ) the spilling of the emission spectrum through a leaky absorber .
we tested both models .
first we fitted to the data a ionized absorber model ( ` absori ' in xspec ) at the redshift of the host galaxy .
we obtained a good fit ( @xmath56 with 110 degrees of freedom , d.o.f . ) with a rest frame column density of @xmath57 @xmath4 and a ionization parameter of @xmath58 ( figure [ f5 ] ) .
the probability for a chance improvement in @xmath8 is 0.3% using the f - test .
second , we fitted a partial covering absorber ( ` zpcfabs ' in xspec ) .
in this case we obtain a similarly good fit ( @xmath59 with 110 d.o.f . ) , a rest frame column density of @xmath60 @xmath4 and a cover fraction of @xmath61 .
the probability for a chance improvement in @xmath8 is 0.4% using the f - test . to check for the presence of a narrow iron k@xmath62 emission line we added to the simple power law model a gaussian component , fixing its width to @xmath63 kev and letting its energy to be free to vary in the range 6.4 - 6.9 kev rest frame .
we found an upper limit to the line intensity of @xmath64 photons s@xmath65 @xmath4 ( 90% confidence level ) and an upper limit to its equivalent width ( ew ) of 0.15 kev .
we compared our bepposax results with other rest frame column density measures that became available , thanks to chandra and xmm data for 9 additional grb at known redshift .
ballantyne et al .
( 2002 ) found a 90% upper limit of @xmath66 @xmath4 at z=1.0 for grb 991216 using 0.4 - 8.0 kev chandra data .
piro et al .
( 2001 ) and ( 2002b ) , estimated for grb 000926 and for grb 000210 , @xmath67 @xmath4 at z=2.04 and @xmath68 @xmath4 at z=0.846 respectively , using chandra _ and _ bepposax data .
note here that harrison et al .
( 2001 ) do not present evidences for significant absorption in addition to the galactic one for grb 000926 at the redshift of the host galaxy using chandra data .
reeves et al .
( 2002 ) and borozdin & trudolyubov ( 2003 ) do not detected any significant absorption above the galactic value in the x - ray spectra of grb 011211 using xmm data .
watson et al . (
2002 ) measured a rest frame @xmath69@xmath4 at @xmath70 for grb 020322 using xmm .
however , the uncertainty on the redshift estimated directly from the x - ray data ( @xmath71 ) is large .
the confidence interval plotted for this burst in figure [ f3 ] includes the uncertainty over z in addition to the statistical uncertainty .
mirabal et al .
( 2002b ) measured @xmath72 @xmath4 for grb 020405 at z=0.690 , using chandra data .
butler et al .
( 2003 ) found no additional absorption for the x - ray afterglow spectra of grb 020813 and grb 021004 using chandra data .
holland et al .
( 2002 ) found no absorption in the vicinity of grb 021004 ( z=2.33 ) with an upper limit of @xmath73@xmath4 using chandra data .
mereghetti et al .
( 2003 ) found @xmath74 @xmath4 at @xmath75 for grb 030227 using xmm data .
on the other hand , the redshift estimated using the x - ray emission lines is @xmath76 ( watson et al .
2003 ) . considering a possible blueshift from the x - ray emitting plasma , flowing at a velocity of @xmath60.1c ( reeves et al . 2002 ; butler et al .
2003 ) , the progenitor redshift should be at @xmath77 .
the rest frame column at this redshift would be @xmath78 @xmath4 .
all these values are in agreement with our rest frame @xmath2 range .
we plot the above @xmath2 determinations along with our bepposax measurements in figure [ f3 ] . according to the standard fireball model
( see sari , piran & narayan 1998 ) at @xmath481 - 2 days from the grb event , the electron population accelerated by shock mechanisms in a power law energy distribution , is expected to radiate in the so called slow cooling regime . in this regime , if the optical range @xmath79 is at frequencies higher than the peak of the emission , the spectrum can be well described by a broken power law with indices @xmath80 for @xmath81 and @xmath82 for @xmath83 , where @xmath84 is the cooling frequency and @xmath85 is the electron energy distribution index . to evaluate the rest frame visual extinction @xmath5
we fitted the optical to near infrared data with the following model : @xmath86 where c is the spectral normalization constant , @xmath87 is the rest frame extinction and z is the grb redshift .
since the optical - nir photometry covers a rather narrow frequency range and the number of independent photometric points for each afterglow is usually small , we could not obtain from the fits strong constraints on _ both _ the emission spectrum and the rest frame extinction .
we estimated the electronic spectral index @xmath85 from each x - ray afterglow spectral index @xmath88 ( tables [ tbl-5 ] and [ tbl-6 ] ) , assuming that the x - ray frequencies @xmath89 are above the cooling frequency @xmath84 ( berger , kulkarni & frail 2003 ) .
on the basis of the @xmath85 measures , we then estimated the optical spectral index according to the fireball model ( sari et al .
1998 ) considering two possible cases , where the optical energy range is below or above the cooling frequency , respectively @xmath90 for @xmath81 or @xmath91 for @xmath83 , or @xmath92 , @xmath93 if the flux is expressed in wavelength ( see tables [ tbl-7 ] and [ tbl-8 ] ) .
we have then estimated the extinction uncertainties taking into account the uncertainties on the estimated optical spectral index @xmath92 and @xmath93 . for grb 000926 , that is not included in our x - ray selected sample , we took the electronic spectral index derived from broad band modeling results published in previous works .
the interpretation of the broad band spectrum for this burst is not univoque .
however , similar values are obtained for the @xmath85 parameter .
piro et al .
( 2001 ) find a best fit model with @xmath94 , from which the authors derive @xmath95 .
harrison et al . ( 2001 ) adopt a different continuum model with @xmath96 and they find a minimum @xmath8 for @xmath97 . due to the uncertainties on the cooling frequency energy regime , we tested for this afterglow both the possible cases , @xmath81 and @xmath83 , assuming @xmath95 .
the model in equation ( 1 ) has been convolved with the transmission functions of the optical and near infrared filters and then converted into a magnitude using the photometric zero points from fukugita , shimasaku & ichikawa ( 1995 ) for the optical range and from bersanelli , bouchet & falomo ( 1991 ) for the nir range .
we then compared our results with the observed optical - nir afterglow seds using a @xmath8 minimization technique .
four different extinction curves have been tested and best fit rest frame visual extinction @xmath5 have been estimated .
we used : _
i ) _ the galactic extinction curve from cardelli et al .
( 1989 ) ( `` g '' ) assuming @xmath43 ; _ ii ) _ the extinction curve of the small magellanic clouds ( `` smc '' ) from pei ( 1992 ) ; _ iii ) _ the extinction curve computed by calzetti , kinney & storchi - bergmann ( 1994 ) and calzetti ( 1997 ) for a sample of local starburst galaxies ( `` c '' ) ; _ iv ) _ the extinction curve computed from simulations by maiolino , marconi & oliva ( 2001a ) based on a dust model skewed toward large grains .
the galactic extinction curve is reasonably well described by a dust model were grain sizes are distributed as a power law , namely @xmath98 , where @xmath99 is the number density of grains with size @xmath100 , @xmath101 and @xmath102 m @xmath103 m ( mathis , rumpl & nordsieck 1977 ) .
the maiolino et al .
( 2001a ) extinction curve q1 used the same index @xmath101 , but a higher maximum grain size @xmath104 m .
all these extinction models are plotted in figure [ f6 ] .
note that the q1 and the c extinction curves are flatter than the rough @xmath105 behavior of the galactic extinction curve .
this gives rise to a more gray extinction .
note also that in the c and in the smc curves , the 2175 hhump is absent , while is quite prominent in the g and q1 curve .
the origin of this 2175 feature is generally interpreted as due to the graphite grains component of the interstellar dust ( mathis 1990 ) . in starburst galaxies
the absence of this feature was interpreted to be the effect of a complex geometry for the dust configuration or to a chemical composition different from that of our galaxy ( calzetti et al .
1994 ) . for the smc extinction curve ,
the absence of the 2175 hump was explained by a different relative abundances of graphite and silicate grains ( pei 1992 ) .
strictly speaking , the calzetti et al .
( 1994 ) extinction curve applies to an extended region rather than to a simple line of sight .
nevertheless we think it is instructive to test it because the above mentioned peculiarities of its shape ( the absence of the strong 2175 hump and the curvature flatter than @xmath105 ) seems particularly well suited to describe the afterglow optical - nir seds . the optical extinction is typically quantified by the extinction in the visual band .
tables [ tbl-7 ] and [ tbl-8 ] give the best fit rest frame visual extinction @xmath5 and the @xmath8 for the two different spectral index values obtained for @xmath81 and @xmath83 and the four extinction models . in all cases the extinction is assumed at the redshift of the grb , or at z=1 when the redshift of the grb is unknown . in the next sections individual cases for which interesting results have been obtained from our analysis
are discussed .
we found that , for these two afterglows , the c extinction curve provides the best fit model to the optical sed for grb 000926 and grb 010222 for @xmath81 ( figure [ f7 ] and table [ tbl-8 ] ) .
the smc extinction curve provides a fit of quality similar to that obtained with the c model for grb 010222 while it does not provide a good fit to grb 000926 .
extinction curves with a strong 2175 hump provide an unacceptable fit for these grb . by comparing the g model with respect to the c model for the case @xmath81 , we found a probability @xmath106 of @xmath107 for grb 000926 and @xmath106 of @xmath108 for grb 010222 .
we note that these two bursts are those with the widest optical - nir frequency coverage , which strongly helps in constraining the extinction model .
lee et al .
( 2001 ) find acceptable fits for grb 010222 assuming the smc extinction curve with @xmath109 for @xmath83 consistent with our measure ( see table 8) .
for the same burst , masetti et al .
( 2001 ) obtain acceptable fit over the nir - optical range using the starburst extinction model , assuming a continuum model different from that one of lee et al .
( 2001 ) , for @xmath81 .
their extinction value is somehow higher but still consistent with the value that we found for the c model with @xmath81 .
bjrnsson et al .
( 2002 ) derive an optical spectral index of @xmath110 for @xmath81 .
we tested the latter spectral index value and we found that it does not affect the result on the c model as the best fitting extinction curve for grb 010222 .
galama et al .
( 2003 ) found a rest frame visual extinction value comparable to the lee et al .
( 2001 ) result by using the host galaxy extinction curve model by reichart ( 2001 ) . for grb
000926 rest frame extinction measured from our analysis are consistent with those of fynbo et al .
( 2001 ) , price et al .
( 2001 ) and harrison et al .
( 2001 ) using the galactic or the smc extinction curves and their derived spectral index .
however , we point out that a better agreement with the data is obtained for this burst with the c extinction curve ( see table 8) .
the extinction curve that provides the best fit to the seds of these grb is the smc model ( figure [ f8 ] ) .
the statistical improvement with respect to the g model for @xmath81 , is @xmath111 for grb 980519 and @xmath112 for grb 971214 ( with no nir points ) , corresponding to a probability @xmath113 and @xmath106 of @xmath114 respectively .
we remind that for these two afterglows , the number of photometric points is smaller than for grb 010222 and grb 000926 and therefore the constraint on the extinction model is less tight .
moreover , the redshift of grb 980519 is unknown ( we fixed it at z=1 ) .
we tested also the redshift as a free parameter .
we found that [email protected] improves the quality of the fit using the g and q1 models , because the 2175 hump moves out of the observed frequency range .
however , the statistical improvement is admittedly low . for grb 971214 we tested the hypothesis that the spectral break observed between the optical and nir bands is due to extinction ( ramaprakash et al .
1998 , see 4.3 ) .
we found that the smc can fit the data ( see figure [ f8 ] ) , with @xmath5 consistent with the extinction value obtained for the same model fitting only the v , r and i points . by comparing the g model with respect to the smc model for @xmath81 , we found a probability @xmath106 of @xmath115 .
grb 970508 is the only case for which the extinction curves including a strong 2175 hump ( g and the q1 ) provide a good fit to the data ( see table [ tbl-7 ] ) .
however , the c and smc models provide only slightly higher @xmath8 .
the redshift of grb 980329 is unknown , puzzling the interpretation of the sharp break between the r and i points observed in its sed ( e.g. palazzi et al .
1998 , reichart et al .
this spectral jump has been explained in terms of ly@xmath116 absorption from neutral hydrogen along the line of sight to a high redshift source ( @xmath117 fruchter et al .
1999b ) . however , yost et al .
( 2002 ) suggested that such a high redshift is not compatible with the grb host galaxy colors .
the sharp break between the r and the i band may be due to the presence of a strong 2175 hump opportunely redshifted .
indeed , although the @xmath8 are still not acceptable , we found a minimum @xmath8 at @xmath118 by using the galactic extinction curve ( figure [ f9 ] ) .
a similar analysis has been performed by lamb , castander & reichart ( 1999 ) on the broad band , from radio to x - ray sed and they obtained a qualitatively similar result .
these values are consistent with the redshift range found by jaunsen et al .
( 2003 ) that excluded at @xmath119 confidence level @xmath120 and @xmath121 .
the latter authors found an host galaxy photometric redshift of z@xmath122 .
we found that , for these afterglows , best fit spectral models are consistent with null extinction for both the two tested spectral indexes . the lack of extinction for these afterglows will put severe constraints on the estimation of the dust to gas ratio in grb environment ( see next section ) , especially the case of grb 990510 , for which a significant @xmath2 column density has been detected . a dedicated spectral analysis of the bright grb 990123 has been previously performed by galama et al .
( 1999 ) , andersen et al .
( 1999 ) and holland et al .
they found spectral index values of @xmath123 ( from optical to x - ray spectrum ) , @xmath124 ( from a weighted mean over temporal scansion of the optical spectrum ) and @xmath125 respectively , and an extinction consistent with zero .
we tested their spectral indexes in our analysis and we obtain similar results
. figures [ f10 ] , [ f11 ] and [ f12 ] compare the best fit additional column densities at the redshift of the grb ( see previous section ) with best fit @xmath5 obtained with the spectral index @xmath92 , corresponding to the case @xmath81 ( see tables [ tbl-7 ] and [ tbl-8 ] ) and assuming three different extinction curves . in the case of galactic extinction curve ( figure [ f10 ] ) , several grb lie well above the @xmath2 versus @xmath5 relationship expected if a galactic dust to gas mass ratio is assumed .
the same analysis was performed with spectral index @xmath93 , corresponding to the case @xmath83 and the obtained best fit extinction values are in this case even lower than in the previous case ( tables [ tbl-7 ] and [ tbl-8 ] ) .
the solid curve in figure [ f10 ] is the relationship estimated by predehl & schmitt ( 1995 ) between @xmath2 and @xmath5 : @xmath126 @xmath4 .
we performed a @xmath8 test to compare the rest frame @xmath2 values ( for those afterglows at known redshift ) to those predicted by the observed extinction @xmath5 based on this relationship , and we obtained @xmath127 for 5 d.o.f .
( probability p(@xmath128 ) of @xmath129 ) .
this behavior has been seen in other cosmic sources like agn as can be seen from the open squares in figure [ f10 ] ( from a compilation of maiolino et al .
2001b and elvis et al .
it appears that grb are similar to agn in their @xmath2/a@xmath130 ratio , although agn show somewhat higher @xmath2 and a@xmath130 .
a good agreement between measured @xmath2 and @xmath5 is obtained using the smc model , since for the smc , the @xmath131 ratio is roughly one order of magnitude higher than the galactic case and is @xmath132 @xmath4 ( weingartner & draine 2001 ) .
we performed a @xmath8 test to compare the rest frame @xmath2 values ( for those afterglows at known redshift ) to those predicted by the observed extinction @xmath5 and we obtained @xmath133 for 5 d.o.f .
( probability p(@xmath128 ) of @xmath134 , figure [ f11 ] ) .
we note , however , that in estimating the column densities from the x - ray absorption , a metallicity lower than the galactic one should have been considered if we hypothize a smc - like environment ( pei 1992 ) . to reproduce the energy cut - off in the x - ray afterglow spectra with a metallicity @xmath135 of the solar one , an hydrogen column density increased by a factor of @xmath136 is required .
we measured the rest frame column densities with a chemical abundance @xmath137 of the solar one and we compared them to those predicted by the observed extinction @xmath5 .
we found that the @xmath8 increases by 24.1 , with @xmath138 .
a better result is obtained using the q1 extinction curve . for this latter model
the @xmath2/@xmath5 ratio is @xmath139 @xmath4 ( maiolino et al .
2001a ) for a solar metallicity , thus providing a self consistent solution .
the @xmath8 obtained by comparing the measured @xmath2 to those predicted by the observed @xmath5 is @xmath140 for 5 d.o.f .
( probability p(@xmath128 ) of @xmath141 , figure [ f12 ] ) .
we note that the q1 model does not usually provide good fits to the observed optical - nir seds , because of the presence of a strong 2175 hump ( see figure 6 ) .
however , it seems that this is the model that can better explain the measured @xmath2/@xmath5 values .
this is due to the flat shape of this extinction curve that gives a more gray extinction than the galactic or smc cases and to the high @xmath131 ratio .
the c extinction curve has a similar shape to q1 ( see figure [ f6 ] ) and no strong 2175 hump .
unfortunately , the @xmath131 ratio has not been estimated for the starburst galaxies , because of the complexity of the geometry of the dust and stars distribution inside these galaxies ( calzetti 2001 ) .
therefore , we can not test the data against the prediction of this model . as discussed in section 5.2 the continuum spectral index adopted in the optical - nir photometry
fits was derived from the best fit x - ray spectral index , assuming the fireball model , and assuming that the x - ray frequencies are above the cooling frequency .
we have performed a selfconsistency check of these assumptions by plotting the afterglow nir to x - ray sed , after correction for dust extinction and gas absorption , along with the best fit x - ray power law model and the assumed optical power law model .
the normalization of the x - ray spectra has been scaled to the same time of the optical - nir observations , using the published decay indices .
we note that in most cases ( but grb 970228 and grb 970508 ) the x - ray observations encompass the times at which the optical - nir observations were performed , see table 3 and 4 , thus minimizing any systematic uncertainty in the scaling of the x - ray spectra .
figure [ f13 ] shows three examples : grb 990123 , grb 990510 and grb 010222 , which are among the brightest of our sample . in these three cases
the optical - nir photometry has been corrected for rest frame extinction using the best fit value relative to the q1 model , or its upper limit when it is consistent with zero .
similar results are obtained using the c model .
note as using these extinction laws the optical and x - ray spectra of grb 010222 are consistent with the fireball model . adopting the g and the smc extinction laws , or no extinction , we find again the result already reported by int zand et al .
figure [ f13 ] show that the nir to x - ray seds are well consistent with the fireball model , and that the cooling frequency is constrained within the uv and the x - ray band .
qualitatively similar results are found also for the other afterglows .
we remark , however , that in latter cases this is not a truly stringent test , given the lower statistics in the x - ray spectra and the consequent larger uncertainties on the x - ray spectral index , ( and therefore also in the allowed range of values for the optical index ) .
we analyzed the x - ray absorption properties of a sample of 13 bright grb afterglows observed by bepposax in the energy range @xmath142 kev .
we found that absorption in addition to the galactic one along the line of sight is highly statistically significant in two cases : grb 990123 and grb 010222 ( probability @xmath0 ) .
these are also the two grb with the best statistics . however , fitting the spectra of two other grb afterglows with lower but still reasonably good statistics ( grb 971214 and grb 970228 ) , including an additional absorber does not produce a significant improvement in @xmath8 ( probability of @xmath50 and @xmath51 respectively ) . in three other cases ( grb 980703 , grb 990510 and grb 001109 ) the presence of an additional absorber is marginally significant ( probability @xmath143 ) .
x - ray absorption may be common , although the quality of present data does not allow us to reach a firm conclusion . in all cases ,
the absorbing columns at the grb redshift are in the range @xmath144 - a few @xmath145 @xmath4 , values similar to the column usually observed toward the disk and the bulge of the galaxy , suggesting that grb afterglows are probing similar environments in the grb host galaxies .
this range is also in agreement with the estimation of the @xmath2 expected if grb were embedded in giant molecular clouds similar to those observed in our galaxy ( reichart & price 2002 ) .
de pasquale et al . ( 2003 ) performed an analysis on the bepposax x - ray afterglows , aimed at studying the nature of dark grb ( for which no or very faint optical counterpart has been discovered ) .
their conclusions from column density measures are generally consistent with those given in this paper .
the rest frame column density estimated for 9 x - ray afterglows observed by xmm ( grb 011211 , grb 020322 and grb 030227 ) and chandra ( grb 991216 , grb 000210 , grb 000926 , grb 020405 , grb 020813 and grb 021004 ) are consistent within the range found for the 13 bepposax grb ( see figure [ f3 ] ) .
the spectrum of grb 010222 suggests the presence of complex absorber or an additive emission component ( figure [ f5 ] ) .
two possible explanations invoke the recovery of the spectrum below the energies of strong ionized absorption edge lines as ovii , oviii or the spilling of the emission spectrum through a leaky absorber .
we tested both models obtaining a significant improvement of the @xmath8 in both cases ( figure [ f5 ] ) .
we found an upper limit to the equivalent width of a line at 6.4/(1+z ) kev of 150(1+z ) ev ( 90% confidence level ) , that is , one order of magnitude smaller than the line equivalent width measured for grb 000214 ( antonelli et al .
2000 ) .
we evaluated the rest frame visual extinction of the nine optical afterglows in our sample .
in addition we analyzed also the optical afterglow of grb 000926 ( see 5.2.1 ) .
we fitted the optical - nir afterglows photometry with a power law model modified at short wavelengths by rest frame extinction ( see equation ( 1 ) in 5.2 ) .
we tested four different extinction curves : _ i ) _ galactic ( `` g '' ) ; _ ii ) _ small magellanic clouds ( `` smc '' ) ; _ iii ) _ the extinction curve computed for a sample of local starburst galaxies ( `` c '' ) ; _ iv ) _ an extinction curve computed from simulations by maiolino et al . ( 2001a ) based on a dust distributions skewed toward large grains ( `` q1 '' ) . in two cases , grb 010222 and grb 000926 , we found that the extinction curve that better fit the data is the c model .
these afterglows have the widest optical - nir frequency coverage in our sample .
this strongly help in constraining the extinction model .
in particular , those extinction curves with a strong 2175 hump ( g and q1 ) are not suitable to model these two afterglows .
in fact , despite q1 has a slope very similar to that of the c model ( see figure [ f6 ] ) , it fails to fit their optical seds due to the presence of a strong 2175 hump that , on the other hand is completely absent in the c curve .
we note that assuming the c or the q1 model , the nir to x - ray sed of grb 010222 is fully consistent with the fireball model , without the need of a strong contribution from inverse compton scattering in the x - ray band ( see eg .
int zand et al .
2001 ) .
the smc model provides good fit to the grb 980519 and grb 971214 data ( by using for the latter both optical and nir data , see 5.2.2 ) .
however , the number of photometric points is smaller than for grb 010222 and grb 000926 and therefore the constraint on the extinction model is less tight .
moreover , the redshift of grb 980519 is unknown and for grb 971214 the interpretation of the spectral break observed between the optical and nir bands is not univoque ( see 4.3 ) . excluding the nir points , both the smc and the c extinction curves provide comparable good fit to grb 971214 .
the small curvatures of the grb 990510 and grb 990123 seds require very low extinction .
no strong constraints on the extinction model can be obtained also in the cases of grb 970228 and grb 980329 because of the few photometric points .
the extinction models tested in our analysis that have a strong 2175 hump are indicated as g and q1 ( see 5.2 and figure [ f6 ] ) .
we found that in all the analyzed optical - nir seds , except grb 970508 , the g and q1 model give bad @xmath8 ( see table [ tbl-8 ] ) .
this 2175 hump does not appear to be a common feature in extra galactic sources ( calzetti et al .
1994 ) and there is not yet a clear and univoque interpretation of its absence . in the case of grbs ,
the absence of the 2175 can be explained by the results obtained by perna et al .
( 2002b ) .
they find that dust destruction of the small graphite grains due to an intense x - ray / uv source , as could be for grb , would produce a significantly suppression of the 2175 hump .
two bursts in our sample are at unknown redshift : grb 980519 and grb 980329 .
we found that for grb 980519 , a lower redshift ( [email protected] ) improves , although not very significantly , the quality of the fit using the g and q1 models , because the 2175 hump moves out of the observed frequency range . for grb 980329 , the jump between the r and i photometric points is marginally explained by the 2175 hump redshifted to the optical range ( [email protected] , see figure [ f9 ] ) .
as discussed in the previous sections , we have evaluated rest frame optical extinction @xmath5 from the optical to near infrared photometry and rest frame equivalent hydrogen column densities @xmath2 from the x - ray data .
we found that several grb lie well above the @xmath2 versus @xmath5 relationship expected for the galactic extinction curve ( figure [ f10 ] ) . in this
regard our results are qualitatively similar to those of galama & wijers ( 2001 ) , although the best fit @xmath2 values are somewhat smaller than those reported by the latter authors , and visual extinction are evaluated here using detailed extinction curves ( while galama & wijers ( 2001 ) assume a linear extinction curve @xmath146 and computed rest frame @xmath2 by using a simple power law scaling of the @xmath2 measured in the observer frame ) .
scenarios that can reconcile the observed x - ray absorption with the lack of strong optical extinction , involve the presence of a dust to gas ratio much lower than the galactic one ( see figure [ f6 ] ) , and/or of a dust grain size distribution skewed toward large grains .
we discuss these two possibilities in turn .
a dust to gas ratio much lower than the galactic one has been tested assuming the smc extinction curve . in fact , for the smc ism it has been inferred a dust content @xmath147 of the galactic one ( pei 1992 ) . a good agreement between the best fit @xmath2 and @xmath5 is obtained using this extinction curve ( figure [ f11 ] ) .
we note , however , that the best fit @xmath2 values have been obtained by assuming galactic metal abundances , while the metallicity of the smc ism is @xmath135 the galactic one ( pei 1992 ) .
if such low metallicity were assumed , best fit @xmath2 values became @xmath148 times higher than those quoted in tables 5 and 6 as plotted in figure [ f11 ] ( dashed line ) , worsening the agreement with the best fit optical extinction .
we can therefore conclude that also for an smc - like ism the observed extinction is significantly lower than predicted from the x - ray column density .
large grains give gray opacity from optical to the infrared ( e.g. kim , martin & hendry 1994 , weingartner & draine 2000 , maiolino et al .
2001a ) , that could explain the small curvature observed in the optical afterglow seds .
a low @xmath149 supports the scenario were large grains are formed by coagulation mechanisms ( e.g. kim & martin 1996 , maiolino et al .
since the grain coagulation rate increases with density as @xmath150 ( draine 1985 ) , the depletion of small grains by coagulation mechanisms is expected to be particularly efficient in high density environments , as the giant molecular clouds which could harbor grb events .
high density gas and dust coagulation is not the only viable mechanism to produce a dust grain size distribution skewed toward large grains .
if the dust to metal ratio is constant , as in the galaxy and the two magellanic clouds ism ( pei 1992 ) , the lack of strong extinction ( dust ) , along with a consistent amount of x - ray absorption ( metal ) , suggest a scenario were dust is destroyed by the intense grb x - ray / uv flux .
dust grains may be heated and evaporated by the intense x - ray and uv radiation field up to @xmath151 ( waxmann & draine 2000 , draine & hao 2002 , fruchter , krolik & rhoads 2001 ) .
perna et al . ( 2002b ) and perna & lazzati ( 2002c ) show that the result of the exposure of dust to the intense grb radiation field can be a grain size distribution flatter that the original one .
the main reason is that dust destruction is more efficient on small grains .
perna et al . ( 2002b ) computed the extinction curve which is obtained if standard galactic dust is exposed to a grb , and found that the dust extinction curve can be very flat , at least for burst lasting more than a few tens of seconds .
we tested a dust model skewed toward large grains using the model q1 adapted from maiolino et al .
( 2001a ) .
the x - ray @xmath2 agree very well with the hydrogen column density expected from the best fit @xmath5 and the visual extinction - reddening ratio appropriate for this model ( see figure [ f12 ] ) .
we note that this model does not imply a metallicity lower than solar , thus providing a self consistent solution , unless the smc - like ism .
however , the q1 curve gives generally worse @xmath8 ( see table 7 and 8) than the smc and c models .
this is mainly due to the presence of a strong 2175 hump in the q1 model .
the c extinction curve has a slope similar to that of the q1 model ( figure [ f6 ] ) , but no 2175 hump , thus providing better fits
. this difference does not affect the rest frame extinction estimates , whose values from both models are indeed comparable ( see table 7 and 8) and suggests that an extinction model with a `` flat '' slope and no 2175 hump would produce both good fits to most observed seds and a nice agreement between @xmath2 and @xmath5 .
in the near future we will have the chance to confirm this intriguing scenario .
_ swift _ will enable us to obtain high quality x - ray spectra of the afterglow as a function of the time .
the _ swift _ optical monitor will allow us to obtain precise and reliable optical - uv photometry .
robotic telescopes like rem , triggered by _
swift _ , will be able to obtain simultaneous near infrared photometry
. therefore it will be possible to obtain precise reddening measures , constraining the correct extinction law , and to follow the evolution of the reddening as a function of the time , that is the smoking gun of a possible grb uv / x - ray flux influence on their environment .
this research has been partially supported by cofin 2001 grants , by cnaa 1999 and 2000 grants and by university of rome `` la sapienza '' support .
we would like to thank roberto maiolino for his precious support to this work , rosalba perna , sandra savaglio , daniela calzetti , nicola masetti for useful discussions and an anonymous referee for corrections that improved the presentation .
bepposax was a program of the italian space agency ( asi ) with participation of the dutch space agency ( nivr ) .
we would like to thank all the members of the bepposax team for the support in performing observations with the satellite .
we also thank an anonymous referee for comments that helped in improving this presentation .
ahn , s. 2000 , apj , 530 , 9 amati , l. , et al .
2000 , science , 290 , 953 andersen , m.i .
1999 , science , 283 , 2075 andersen , m.i .
2000 , a&a , 364 , l54 antonelli , l.a .
2000 , apj , 545 , 39 ballantyne , d.r . , ramirez - ruiz , e. , lazzati , d. & piro , l. 2002 , a&a , 389 , l74 berger , e. , kulkarni , s.r . and frail , d.a . , 2001 , apj , 560 , 652 berger , e. , kulkarni , s.r . and frail , d.a . , 2003 , apj , 590 , 379 bersanelli , m. , bouchet , p. & falomo , r. 1999 , a&a , 252 , 854 beuerman , k. et al . 1999 ,
a&a , 352 , l26 bevington , f.r . & robinson , d.k .
1992 , data reduction and error analysis for the physical sciences , second edition ( wcb ; mc - grow hill ) bjrnsson g. et al .
2002 , apj , 579 , l59 bloom , j.s .
, et al . 1998a , apj , 508 , l21 bloom , j.s . ,
djorgovski , s.g . & kulkarni , s.r .
1998b , apj , 507 , 25 bloom , j.s . , et al . 1999 ,
apj , 518 , l1 bloom , j.s . ,
djorgovski , s.g . &
kulkarni , s.r .
2001 , apj , 554 , 678 bloom , j.s . , kulkarni , s.r . &
djorgovski , s.g .
2002a , aj , 123 , 1111 bloom , j.s .
. 2002b , apj , 572 , l45 boella , g. et al .
1997b , a&as , 122 , 327 bttcher , m. , dermer , c.d . ,
crider , a w. & liang , e.p . 1999 ,
a&a , 343 , 111 borozdin , k.n . & trudolyubov , s. p. 2003 , apj , 583 , l57 butler , n.r .
et al . 2003 , apj , 597 , 1010 calzetti , d. , kinney , a.l . & storchi - bergmann , t. 1994 , apj , 429 , 582 calzetti , d. 1997 , aj , 113 , 1 calzetti , d. 2001 , pasp , 113 , 1449 cardelli , j.a . ,
clayton , g.c . & mathis , j.s .
1989 , apj , 345 , 245 castander , f.j . & lamb , d. 1999 , apj , 523,602 castro , s. et al .
2003 , apj , 586 , 128 castro - tirado , a.j . ,
et al . 1999 ,
apj , 511 , l85 chiappetti l. et al .
1998 , nucl .
b ( proc . supp . ) , 69/103 , 610 ciardi , b. & loeb 2000 , , 540 , 687 costa , e. , et al . 1999 ,
a&as , 138 , 425 cowsik , r .
2001 , bull .
india , 29 , 157 de vries , h.w . , heithausen , a. & thaddeus , p. 1987
, , 319 , 723 de pasquale , m. , et al .
2003 , apj , 592 , 1018 della valle , m. et al .
2003 , a&a , 406 , 33 djorgovski , s.g .
1997 , nature , 387 , 876 djorgovski , s.g .
et al . 1998a , apj , 508 , l17 djorgovski , s.g . ,
kulkarni , s.r . ,
bloom , j.s . ,
goodrich , r. , frail , d.a .
1998b , apj , 508 , l17 djorgovski , s.g . , kulkarni , s.r .
, bloom j.s . & frail , d.a . 1999 ,
289 djorgovski , s.g .
2001 , proc .
`` gamma - ray bursts in the afterglow era : 2nd workshop '' , edited by enrico costa , filippo frontera , and jens hjorth .
berlin heidelberg : springer , 2001 , 218 draine , b.t .
1985 , protostars and planets ii , ed .
d.c.black & m.s.matthews , sp . sc .
( the univ . of arizona pr . ) draine , b.t . , &
hao , l. 2002 , apj , 569 , 780 dickey , j.m . & lockman , f.j .
1990 , ara&a , 28 , 215 eichler , d. , livio . , m. , piran , t. & shramm , d.n .
1989 , nature , 340 , 126 elvis , m. , lockman , f.j . & wilkes , b.j .
1989 , , 97 , 777 elvis , m. , fiore , f. , giommi , p. & padovani ,
p. 1998 , apj , 492 , 91 fiore , f. , guainazzi , m. & grandi , p. 1999 , _
handbook for bepposax nfi spectral analysis _ , ftp://www.sdc.asi.it/pub/sax/doc/software_docs/saxabc_v1.2.ps.gz fiore , f. , nicastro , f. , savaglio , s. , stella , l. & vietri , m. 2000 apjl , 544 , l7 fiore , f. 2001 , proceeding of ` new century of x - ray astronomy ' , yokohama , japan , astro - ph/[0107276 ] fox , d.w .
, et al . 2003 , nature , 422 , 284 frail , d.a .
2002 , apj , 565 , 829 frail , d.a .
2003 , apj , 590 , 992 frontera , f. , costa , e. , dal fiume , d. , feroci , m. , nicastro , l. , orlandini , m. , palazzi , e. & zavattini , g. 1997 , a&as , 122 , 357 frontera , f. , et al .
2000 , apjs , 127 , 59 fruchter , a.s .
et al . 1999a , , 516 , 683 fruchter , a.s .
1999b , , 562 , 683 fruchter , a.s . ,
krolik , j.h .
, & rhoads , j.s .
2001 , apj , 563 , 597 fryer , c.l . ,
woosley , s.e . & hartman , d.h .
1999 , aj , 526 , 152 fukugita , m. , shimasaku , k. & ichikawa , t. 1995 , pasp , 107 , 945 fynbo j.p.u .
2001 , a&a , 373 , 796 galama , t.j .
1997 , nature , 387 , 479 galama , t.j .
et al . 1998a , apj , 497 , 13 galama , t.j . ,
wijers , r.a.m.j . , bremer , m. , groot , p.j . ,
strom , r.g . ,
kouveliotou , c. & van paradijs , j. 1998b , apj , 500 , 97 galama , t.j . , et al .
1999 , nature , 398 , 394 galama , t.j .
2000 , apj , 536 , 185 galama , t.j . , & wijers , a.m.j .
2001 , apjl , 549 , l209 galama , t.j .
2003 , apj , 587 , 135 garcia , m.r .
1998 apj , 500 , l105 ghisellini , g. , lazzati , d. , rossi , e. & rees , m.j .
2002 , a&a , 389 , l33 gonzalez , r.a . ,
fruchter , a.s .
& dirsch , b. 1999 , apj , 515 , 69 gorosabel , j. et al .
1998 , a&a , 335 , l5 greiner , j. et al .
, cgn circ .
2020 guarnieri , a. et al .
1997 , a&a , 328,l13 halpern , j.p . , thorstensen , j.r . ,
helfan , d.j . & costa , e. 1998 , nature , 393 , 41 halpern , j.p . , kemp , j. & piran .
t. 1999 , apj , 517 , l105 harrison , f.a .
, et al . 1999 ,
apj , 523 , l121 harrison , f.a . , et al .
2001 , apj , 559 , 123 heiles , c. , reach , w.t . & koo , b .-
1988 , , 332 , 313 henden , a.a .
1998 , gcn circ .
131 hjorth , j. et al .
2003 , nature , 423 , 847 holland , s . ,
bjrnsson , j . ,
hjorth , j .
thomsen , b .
2000 a&a , 364 , 467 holland , s. et al .
2001 , a&a , 371 , 52 holland , s. et al .
2002 , aj , 124 , 639 jha , s. et al .
2001 , apj , 554 , l155 jaunsen , a.o .
2001 , apj , 546 , 127 jaunsen , a.o .
2003 , a&a , 402 , 125 int zand , j.j.m .
2001 , apj , 559 , 710 kawabata , k.s . ,
2003 , apj , 593 , 19 kim , s .- h . , martin , p.g .
& hendry , p.d .
1994 , apj , 422 , 164 kim , s .- h . & martin , p.g .
1996 , apj , 462 , 296 kulkarni , s.r .
, et al . 1998 , nature , 393 , 35 kulkarni , s.r .
, et al . 1999 , nature , 398 , 389 kumar , p. & narayan , r. 2003 , apj , 584 , 895 lamb , d.q . ,
castander , f.j . & reichart , d.e . 1999 ,
a&ass , 138 , 479 laureijs , r.j .
1989 , ph.d.thesis , groningen university laureijs , r.j . ,
helou , g. & clark , f.o . _ the first symposioum on the infrared cirrus and diffuse interstellar clouds _
( edscutri , r . & latter , w . ) 133 - 136 ( conf
58 , astron .
pacific . ,
provo , ut , 1994 ) lazzati , d. , covino , s. & ghisellini , g. 2002a , mnras , 330 , 583 lee b.c .
2001 , apj , 561 , 183 macfayden , a. & woosley , s.e .
1999 , apj 524,262 maiolino , r. , marconi , a. , & oliva , e. 2001a , a&a , 365 , 37 maiolino , r. , marconi , a. , salvati , m. , risaliti , g. , severgnini , p. , oliva , e. , la franca , f. & vanzi , l. 2001b , a&a , 365 , 28 manzo , g. , giarrusso , s. , santangelo , a. , ciralli , f. , fazio , g. , piraino , s. & segreto , a. 1997 , a&as , 122 , 341 masetti , n.,et al .
2001 , a&a , 374 , 382 mathis , j. s. , rumpl , w. & nordsieck , k. h. 1977 , apj , 217 , 425 mathis , j.s .
1990 , ara&a , 28 , 37 mereghetti , s. et al .
2003 , apj , 590 , 73 meszaros , p. & rees , m.j .
1992 , apj , 397 , 570 mirabal , n. et al . 2002a , apj , 578 , 818 mirabal , n. , paerels , f. & halpern , j.p .
2002b , apj , 587 , 128 mirabal , n. et al . 2002c ,
apj , 578 , 818 narayan , r. , paczynski , b. & piran , t. 1992 , apj , 395 , l83 norris , j . , scargle , j . &
bonnell , j. 2000 , baas , 32 , 1244 odewahn , s.c . , et al .
1998 , apj , 509 , 5 paczynski , b. 1998 , apjl , 494 , l45 paczynski , b. 1990 , apj , 348 , 485 palazzi , e. et al . 1998 , a&a , 336 , l95 parmar a. , et al .
1997 , a&as , 122 , 309 pedersen , h. et al .
1998 , gcn circ .
142 pedichini , f. et al .
1997 , a&a , 327 , l36 pei , y.c .
1992 , apj , 395 , 130 perna , r. & loeb , a. 1998 , apj , 501 , 467 perna , r. & belczynski , k. 2002a , apj , 570 , 252 perna , r. , lazzati , d. & fiore , f. 2002b , apj , 585 , 775 perna , r. & lazzati , d. 2002c , apj , 580 , 261 pettini , m. , smith , l.j . , king , d.l .
& hunstead , w. 1997 , apj , 486 , 665 pettini , m. , ellison , s.l . ,
steidel , c.c . & bowen , d.v . 1999 ,
apj , 510 , 576 piro , l. et al . 1999 ,
apjl , 514 , l73 piro , l. , et al .
2000 , science , 290 , 955 piro , l. , et al .
2001 , apj , 558 , 442 piro , l. 2002a , proc .
`` gamma - ray bursts in the afterglow era : 3nd workshop '' , in press piro , l. , et al . 2002b , apj , 577 , 680 price , p.a .
2001 , apj , 549 , l7 price , p.a .
2003 , nature , 423 , 844 predehl , p. & schmitt , j.h.m.m .
1995 , a&a , 293 , 889 ramaprakash , a.n . , et al . , 1998 , nature , 393 , 43 reach , w.t . , heiles , c. & koo , b .- c .
1993 , , 412 , 127 reeves , j.n .
2002 , nature , 416 , 512 reichart , d. e. 1997 , apj , 485 , l57 reichart , d. e. 1998 , apj , 495 , 99 reichart , d. e. 1999a , apj , 521 , 111 reichart , d. e. , et al . 1999b , apj , 517 , 692 reichart , d. e. 2001 , apj , 553 , 235 reichart , d. e. & price , p.a .
2002 , apj , 565 , 174 rhoads , j. et al .
1998 , gcn circ .
144 rowan - robinson , m. et al .
1991 , mnras , 249 , 729 sahu , k. et al .
1997 , nature , 387,476 sari , r. , piran , t. & narayan , r. 1998 , apjl , 197 , l17 savage , b.d . &
mathis , j.s .
1979 , ara&a , 17 , 73 savaglio , s. , fall , m. , & fiore , f. 2003 , apj , 585 , 638 schlegel , d.j . ,
finkbeiner , d.p . &
davis , m. 1998 , , 500 , 525 simon , v. , hudec , r. , pizzichini , g. & masetti , n. 2001 , a&a , 377 , 450 stark , a.a .
1992 , apjs , 79 , 77 sokolov , v.v . ,
kopylov , a.i . ,
zharikov , s.v . ,
feroci , m. , nicastro , l. & palazzi , e. 1998a , a&a , 334 , 117 sokolov , v.v .
et al . 1998b , gcn circ .
147 stanek , k.z .
, garnavich , p.m . ,
kaluzny , j . ,
pych , w . ,
thomson , i .
1999 , apj , 522 , 39 stanek , k.z .
2003 , apj , 591l , 17 steidel , c. et al . 1999 ,
apj , 519 , 1 vietri , m. & stella , l. 1998 , , 507 , l45 vreeswijk , p.m. , et al . 1999 ,
apj , 523 , 171 vreeswijk , p.m. , et al .
2000 , apj , 546 , 672 watson , d. , reeves , j.n . , osborne , j.p . ,
tedds , j.a . ,
obrien , p.t . , tomas , l. & ehle , m. 2002 , a&a , 395 , l41 watson , d. , reeves , j.n . , hjorth , j. , jakobsson , p. & pedersen , k. 2003 , apj , 595 , 29 waxman , e. , & draine , b. t. 2000 , apj , 537 , 796 weingartner . j.c .
& draine , b.t .
2000 , aas , 197 , 4207 weingartner . j.c . & draine , b.t .
2001 , apj , 548 , 296 wijers , r.a.m.j . & galama , t.j . 1999 ,
apj , 523 , 177 woosley , s. e. 1993 , , 405 , 273 yoshida , a. et al . , 1999 , a&as , 138 , 433 yost , s.a .
2002 , apj , 577 , 155 zapatero osorio , m.r .
1998 , iau circ .
* grb 970228 * : a large number of papers report photometric observations of the optical afterglow of grb 970228 ( e.g. guarnieri et al . 1997 , sahu et al .
1997 , galama et al .
1997 , reichart 1997 , pedichini et al .
1997 , garcia et al .
1998 , castander & lamb 1999 , fruchter et al .
1999a , reichart 1999a , galama et al . 2000 ) .
the latter paper presents a complete and exhaustive summary of the available optical observations , and therefore we refer to this paper for this afterglow .
the galactic extinction in the direction of grb 970228 is @xmath152 , from the sfd maps .
castander & lamb ( 1999 ) estimate a galactic extinction of @xmath153 combining the results of four different methods for measuring the extinction ( also see fruchter et al .
1999a , and gonzalez , fruchter & dirsch 1999 ) .
we used both these values and we verified that their difference does not affect significantly the estimation of the rest frame extinction . * grb 970508 * : galama et al .
( 1998a ) collect observations over the entire optical band , from u to i , for this afterglow , and therefore we referred to this paper in this case .
different authors have reported different estimates of the galactic extinction in the direction of grb 970508 .
galama et al .
( 1998a ) report @xmath154 ( laureijs 1989 ) , while djorgovski et al .
( 1997 ) report @xmath155 ( laureijs , helou & clark 1994 ) , both using the @xmath25 cirrus flux maps . reichart ( 1998 ) reports @xmath156 , using the rowan - robinson et al .
( 1991 ) maps ( also see sokolov et al . 1998a , galama et al . 1998b , bloom , djorgovski & kulkarni , 1998a , castro - tirado & gorosabel 1999 ) .
we adopt here the value found by the two latter authors .
* grb 971214 * : we refer to the ramprakash et al .
( 1998 ) work that report k - band magnitudes and collect previous v , r , i , j observations ( gorosabel et al .
1998 , odewahn et al .
1998 , wijers & galama 1999 , ahn 2000 ) for this burst .
the galactic extinction in the direction of this burst is negligible ( halpern et al . 1998 ) .
a strong spectral curvature is evident in figure [ f1 ] for this afterglow .
wijers & galama ( 1999 ) interprets this feature as due to a genuine break in the emission spectrum , expected in the blast wave model when the cooling frequency is in the optical range ( galama & wijers 2001 exclude for this reason the nir data from their analysis , when trying to determine the rest frame extinction fitting the sed with a simple power law ) . on the other hand , halpern et al .
( 1998 ) and ramaprakash et al .
( 1998 ) interpret the spectral curvature in terms of extinction by intervening dust .
they point out that a simple @xmath157 dust extinction law does not fit properly the data and that a more complex extinction law should be considered . in the
following we consider both possibilities . *
grb 980329 * : we take the photometric data from the work of reichart et al .
( 1999b ) which present an extended and complete summary of all the observations of this burst ( but see also palazzi et al .
the galactic reddening in the direction of grb 980329 is estimated as e(b - v)=0.073 mag using the sfd maps . * grb 980519 *
: we take the photometric data from the work of jaunsen et al .
halpern et al .
( 1999 ) report two different values for the galactic reddening : e(b - v)=0.267 using the sfd maps and e(b - v)=0.348 using the 21 cm stark et al .
( 1992 ) maps and assuming a standard @xmath2 to @xmath158 conversion ( savage & mathis 1979 ) .
jaunsen et al .
( 2001 ) find e(b - v)=0.24 using an alternative method , similar to the sfd determination .
we adopt here the latter value .
* grb 980703 * : the first observations of this afterglow were performed in the nir band some @xmath630 hours after the grb event .
later optical observations are strongly affected by the host galaxy emission .
we considered here only the nir points from the work of vreeswijk et al .
( 1999 ) which summarize most previous observations ( zapatero osorio et al .
1998 , rhoads et al .
1998 , henden et al .
1998 , bloom et al .
1998a , pedersen et al .
1998 , djorgovski et al .
1998a , sokolov et al .
1998b , castro - tirado et al . 1999 , but also see berger , kuulkarni & frail 2001 , holland et al .
2001 , frail et al .
the galactic extinction in the direction of grb 980703 is @xmath159 , using the sfd maps .
the optical and infrared observations of this afterglow , ( bloom et al .
1999 , holland et al .
2000 ) , are well summarized by galama et al .
( 1999 ) and by kulkarni et al .
( 1999 ) and therefore we took the photometric data from these two works .
the galactic extinction in the direction of grb 990123 is @xmath160 , using sfd maps . *
grb 990510 * : we took the b , v , r , i photometry from the work of stanek et al .
( 1999 ) ( but also see harrison et al .
1999 , beuerman et al .
1999 and holland et al .
2000 ) , which gives a complete summary of the optical observation on this burst .
these authors report magnitudes already corrected for the galactic reddening e(b - v)=0.203 , estimated using the sfd maps .
* grb 010222 * : masetti et al .
( 2001 ) present a complete summary of the optical to nir photometry ( lee et al .
2001 , cowsik et al .
2001 , jha et al .
2001 , also see frail et al .
2002 , mirabal et al .
2002a , bjrnsson et al .
2002 , galama et al .
2003 ) . a recent work from frail et al .
( 2002 ) rules out that the k - band emission is originated from the burst or its afterglow and concludes that it is instead due to the emission of a high redshift starburst galaxy .
we therefore excluded the k band point in our analysis .
the galactic extinction in the direction of this burst is @xmath161=0.07 , using the sfd maps .
lccccc grb & @xmath29 & redshift & r.a . & dec .
& ot + & @xmath162 @xmath4 & & h m s & @xmath163 @xmath164 @xmath165 & + 970228 & 16.0 & @xmath166 & 15 28 0.9 & + 19 37 02 & y + 970508 & 5.2 & @xmath167 & 06 53 46.7 & + 79 16 02 & y + 971214 & 1.6 & @xmath168 & 11 56 25 & + 65 13 11 & y + 980329 & 9.4 & - & 07 02 36 & + 38 50 24 & y + 980519 & 1.7 & - & 23 22 22 & + 77 16 06 & y + 980703 & 5.8 & @xmath169 & 23 59 6.8 & + 08 35 45 & y + 990123 & 2.1 & @xmath170 & 15 25 30 & + 44 45 14 & y + 990510 & 9.4 & @xmath171 & 13 38 11.2 & -80 30 02 & y + 990704 & 3.1 & - & 12 19 27.3 & -03 50 22 & n + 000210 & 2.5 & @xmath172 & 01 59 09.8 & -40 39 15 & n + 000214 & 5.8 & 0.47@xmath173 & 18 54 27 & -66 27 30 & n + 001109 & 4.2 & - & 18 30 07.8 & + 55 17 56 & n + 010222 & 1.6 & @xmath174 & 14 52 12.0 & + 43 01 02 & y + lcccccccc & & + grb & exposure@xmath175 & r@xmath176 & bkg@xmath177 & source@xmath178 & exposure@xmath175 & r@xmath176 & bkg@xmath177 & source@xmath178 + & ks & arcmin & @xmath179 & @xmath179 & ks & arcmin & @xmath179 & @xmath179 + 970228 & 5.5 & @xmath180 & @xmath181 & @xmath182 & 14.3 & @xmath183 & @xmath184 & @xmath185 + 970508 & 17.2 & @xmath180 & @xmath186 & @xmath187 & 28.2 & @xmath183 & @xmath188 & @xmath189 + 971214 & 45.5 & @xmath180 & @xmath190 & @xmath191 & 101.1 & @xmath183 & @xmath181 & @xmath192 + 980329 & 24.5 & @xmath183 & @xmath193 & @xmath194 & 63.8 & @xmath183 & @xmath181 & @xmath195 + 980519 & 22.7 & @xmath196 & @xmath197 & @xmath198 & 77.8 & @xmath183 & @xmath181 & @xmath199 + 980703 & 16.3 & @xmath196 & @xmath200 & @xmath201 & 39.1 & @xmath196 & @xmath202 & @xmath203 + 990123 & 15.1 & @xmath196 & @xmath197 & @xmath204 & 45.2 & @xmath183 & @xmath181 & @xmath205 + 990510 & 31.2 &
@xmath196 & @xmath200 & @xmath206 & 67.9 & @xmath183 & @xmath181 & @xmath207 + 990704 & 13.4 & @xmath196 & @xmath197 & @xmath208 & 37.0 & @xmath183 & @xmath181 & @xmath209 + 000210 & 15.7 & @xmath180 & @xmath210 & @xmath211 & 44.4 & @xmath183 & @xmath181 & @xmath212 + 000214 & 14.8 & @xmath196 & @xmath197 & @xmath213 & 50.8 & @xmath183 & @xmath181 & @xmath214 + 001109 & 19.2 & @xmath196 & @xmath197 & @xmath215 & 85.4 & @xmath183 & @xmath216 & @xmath192 + 010222 & 50.3 & @xmath196 & @xmath197 & @xmath217 & 88.4 & @xmath183 & @xmath181 & @xmath218 + lcccc & mecs & & & lecs + grb & off - axis@xmath175 & @xmath219&@xmath220 & raw coord.@xmath177 + & arcmin & h & h & pixels + 970228 & @xmath221 & 8.0&17.0 & 118.4,155.0 + 970508 & @xmath222 & 5.7&15.7 & 112.9,116.6 + 971214 & @xmath223 & 6.7&60.8 & 132.6,123.5 + 980329 & @xmath224 & 7.0&48.5 & 134.2,126.9 + 980519 & @xmath225 & 9.7&34.2 & 135.6,135.8 + 980703 & @xmath225 & 22.3&45.8&132.1,119.8 + 990123 & @xmath225 & 5.8&32.2 & 130.1,125.3 + 990510 & @xmath224 & 8.0&52.3 & 139.2,128.5 + 990704 & @xmath224 & 8.0&29.9 & 134.1,119.4 + 000210 & @xmath224 & 7.2&38.9 & 131.8,133.7 + 000214 & @xmath225 & 12.0&40.9&138.0,119.2 + 001109 & @xmath225 & 16.5&38.3&129.4,124.1 + 010222 & @xmath226 & 8.0&65.0 & 128.0,131.9 + lcccccccc grb(ut ) & u & b & v & r & i & j & h & k + 970228.12@xmath227&-&- & @xmath228 & @xmath229 & @xmath230 & - & - & - + 970508.90@xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & - & - & - + 971214.97@xmath237&- & - & @xmath238 & @xmath239 & @xmath240 & @xmath241 & - & @xmath242 + 980329.15@xmath243 & - & - & - & @xmath244&@xmath245 & @xmath246 & - & @xmath247 + 980519.51@xmath248 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & @xmath253 & - & - & - + 980703.18@xmath254&- & - & - & @xmath255 & @xmath256 & - & - & @xmath257 + 990123.41@xmath258&@xmath259 & @xmath260 & @xmath261 & @xmath262 & @xmath263 & - & @xmath264 & @xmath265 + 990510.37@xmath266&-&@xmath267 & @xmath268 & @xmath269&@xmath270&- & - & - + 000926.99@xmath271&@xmath272 & @xmath273 & @xmath274 & @xmath275 & @xmath276 & @xmath277 & @xmath278 & @xmath279 + 010222.31@xmath280 & @xmath281 & @xmath282 & @xmath283 & @xmath284&@xmath285 & @xmath286 & host gal ? & host gal ?
+ @xmath227 from galama et al .
( 2000 ) , @xmath41=ut 28.99 ( @xmath41=extrapolation time , see 4 . in this case
this is at + 20.9h from the gamma - ray event ) ; @xmath231galama et al . ( 1998a ) , @xmath41=ut 10.98 ( + 49.8h ) ; @xmath237ramaprakash et al . (
1998 ) , @xmath41=ut 15.51 ( + 12.9h ) ; @xmath243reichart et al .
( 1999b ) , @xmath41=ut 29.99 ( + 19.9h ) ; @xmath248jaunsen et al .
( 2001 ) , @xmath41=ut 19.96 ( + 10.7h ) ; @xmath254vreeswijk et al .
( 1999 ) , @xmath41=ut 4.4 ( + 29.2h ) ; @xmath258galama et al .
( 1999 ) , @xmath41=ut 24.65 ( + 29.8h ) ; @xmath266stanek et al .
( 1999 ) , @xmath41=ut 11.26 ( + 21.4h ) ; @xmath280masetti et al .
( 2001 ) , @xmath41= ut 23.28 ( + 23.4h ) ; @xmath271fynbo et al .
( 2001 ) , @xmath41=ut 29.67 ( + 64.3h ) .
lcccccccc grb & lecs - mecs & @xmath2 & z & @xmath287 & @xmath8(dof ) & p(f ) + & constant & @xmath288 @xmath4 & & & & % + 971214 & 0.7 - 1.4 & @xmath289 & 3.42 & @xmath290 & 34.12 ( 42 ) & 93.3 + & & 0 fixed & 3.42 & @xmath291 & 36.99 ( 43 ) & + & & @xmath292 & 0 & @xmath293 & 34.79 ( 42 ) & + 990123 & 0.57 - 0.69 & @xmath294 & 1.60 & @xmath295 & 55.94 ( 66 ) & @xmath296 + & & 0 fixed & 1.60 & @xmath297 & 70.44 ( 67 ) & + & & @xmath298 & 0 & @xmath295 & 55.73 ( 66 ) & + 990704 & 0.6 - 1.0 & @xmath299 & 1 & @xmath300 & 17.36 ( 16 ) & 92.5 + & & 0 fixed & 1 & @xmath301 & 21.29 ( 17 ) & + & & 1.4@xmath302 & 0 & @xmath303 & 17.24 ( 16 ) & + 000210 & 0.9 - 2.2 & 8.0@xmath302 & 0.846 & @xmath304 & 12.63 ( 13 ) & 89.9 + & & 0 fixed & 0.846 & @xmath305 & 15.65 ( 14 ) & + & & 1.8@xmath302 & 0 & @xmath306 & 12.77 ( 13 ) & + 001109 & 0.4 - 1.2 & @xmath307 & 1 & @xmath308 & 29.09 ( 27 ) & 97.0 + & & 0 fixed & 1 & @xmath309 & 34.70 ( 28 ) & + & & @xmath310 & 0 & @xmath311 & 29.10 ( 27 ) & + 010222 & 1.21 - 1.37 & @xmath312 & 1.477 & @xmath295 & 99.80 ( 111 ) & @xmath296 + & & 0 fixed & 1.477 & @xmath313 & 146.11 ( 112 ) & + & & @xmath314 & 0 & @xmath295 & 104.61 ( 111 ) & + lcccccccc grb & mecs - lecs & @xmath2 & z & @xmath287 & @xmath8(dof ) & p(f ) + & constant & @xmath288 @xmath4 & ( fixed ) & & & % + 970228 & 0.6 - 1.0 & @xmath315 & 0.695 & @xmath316 & 6.90 ( 10 ) & 90.8 + & & 0 fixed & 0.695 & @xmath317 & 9.29 ( 11 ) & + & & @xmath318 & 0 & @xmath316 & 7.00 ( 10 ) & + 970508 & 0.4 - 1.0 & @xmath319 & 0.835 & @xmath320 & 7.38 ( 12 ) & 90.7 + & & 0 fixed & 0.835 & @xmath321 & 9.43 ( 13 ) & + & & @xmath322 & 0 & @xmath293 & 7.43 ( 12 ) & + 980329 & 0.6 - 0.9 & @xmath323 & 1 & @xmath324 & 15.75 ( 17 ) & - + & & 0 fixed & 1 & @xmath325 & 15.80 ( 18 ) & + & & @xmath326 & 0 & @xmath327 & 15.70 ( 17 ) & + 980519 & 0.7 - 0.9 & @xmath328 & 1 & @xmath329 & 15.95 ( 8) & 50.3 + & & 0 fixed & 1 & @xmath330 & 16.96 ( 9 ) & + & & @xmath331 & 0 & @xmath332 & 15.91 ( 8) & + 980703 & 0.4 - 1.0 & @xmath333 & 0.966 & @xmath334 & 23.30 ( 20 ) & 97.5 + & & 0 fixed & 0.966 & @xmath335 & 30.12 ( 21 ) & + & & @xmath336 & 0 & @xmath337 & 24.45 ( 20 ) & + 990510 & 0.6 - 0.8 & @xmath338 & 1.62 & @xmath339 & 75.13 ( 65 ) & 98.3 + & & 0 fixed & 1.62 & @xmath340 & 83.11 ( 66 ) & + & & @xmath341 & 0 & @xmath339 & 76.43 ( 65 ) & + 000214 & 0.3 - 0.7 & @xmath342 & 0.47 & @xmath343 & 42.03 ( 20 ) & - + & & 0 fixed & 0.47 & @xmath343 & 41.96(21 ) & + & & @xmath344 & 0 & @xmath343 & 42.03(20 ) & + lccccccccc grb & models & & + & & @xmath92 & @xmath5 & @xmath8 & @xmath93 & @xmath5 & @xmath8 + 990123 & g & @xmath345 & @xmath346 & 2.9/5 & @xmath347 & @xmath348 & 45.2/5 + z=1.60 & smc & @xmath345 & @xmath349 & 3.1/5 & @xmath347 & @xmath350 & 45.2/5 + @xmath351 & q1 & @xmath345 & @xmath352 & 2.9/5 & @xmath347 & @xmath353 & 45.2/5 + & c & @xmath345 & @xmath352 & 2.6/5 & @xmath347 & @xmath353 & 45.2/5 + 990510 & g & @xmath354 & @xmath355 & 22.6/2 & @xmath356&@xmath357 & 62.9/2 + z=1.62 & smc & @xmath354 & @xmath355 & 22.6/2 & @xmath356&@xmath353 & 64.1/2 + @xmath358 & q1 & @xmath354 & @xmath350 & 22.6/2 & @xmath356&@xmath353 & 62.9/2 + & c & @xmath354 & @xmath353 & 22.6/2 & @xmath356&@xmath353 & 62.9/2 + 000926 & g & @xmath359 & @xmath360 & 26.9/6 & @xmath361 & @xmath362 & 22.6/6 + z=2.04 & smc & @xmath359 & @xmath363 & 30.4/6 & @xmath361 & @xmath364 & 22.4/6 + @xmath95 & q1 & @xmath359 & @xmath365 & 23.1/6 & @xmath361 & @xmath366 & 21.1/6 + & c & @xmath359 & @xmath367 & 16.7/6 & @xmath361 & @xmath368 & 19.6/6 + 010222 & g & @xmath369 & @xmath370 & 11.2/4 & @xmath371&@xmath350&7.7/4 + z=1.477 & smc & @xmath369 & @xmath372 & 2.9/4 & @xmath371&@xmath373 & 7.6/4 + @xmath374 & q1 & @xmath369 & @xmath375 & 10.8/4 & @xmath371&@xmath349 & 7.7/4 + & c & @xmath369 & @xmath376 & 2.4/4 & @xmath371&@xmath357&7.7/4 + | we studied the x - ray and optical absorption properties of 13 gamma ray burst afterglows observed by bepposax .
we found that x - ray absorption in addition to the galactic one along the line of sight is highly statistically significant in the two cases with the best statistics ( probability @xmath0 ) . in three other cases
the presence of x - ray absorption is marginally significant ( probability @xmath1 ) .
measured rest frame absorbing column densities of hydrogen , @xmath2 , range from 0.1 to 10.0 @xmath3 @xmath4 ( at 90% confidence level ) assuming a solar metal abundance .
x - ray absorption may be common , although the quality of present data does not allow us to reach a firm conclusion .
we found that the rest frame column densities derived from xmm and chrandra data as quoted in the literature are in good agreement with the bepposax estimated rest frame @xmath2 range , supporting our result . for the same grb afterglow sample we evaluated the rest frame visual extinction @xmath5 .
we fitted the optical - nir afterglow photometry with a power law model corrected at short wavelengths by four different extinction curves . by comparing x - ray absorptions and optical extinction
, we found that if a galactic - like dust grain size distribution is assumed , a dust to gas ratio lower than the one observed in the galaxy is required by the data .
a dust to gas ratio @xmath6 1/10 than the galactic one , as in the small magellanic cloud ( smc ) environment , has been tested using the smc extinction curve , which produces good agreement between the best fit @xmath2 and @xmath5 .
we note , however , that the best fit @xmath2 values have been obtained by assuming solar metal abundances , while the metallicity of the smc ism is @xmath6 1/8 the solar one ( pei 1992 ) .
if such low metallicity were assumed , the best fit @xmath2 values would be higher by a factor of @xmath7 , providing a significant increase of the @xmath8 .
alternative scenarios to explain simultaneously the optical and x - ray data involve dust with grain size distributions biased toward large grains .
possible mechanisms that can bring to such a grain size distribution are discussed . |
the origin of cosmic rays ( crs ) , in particular above the knee energy @xmath4 and at ultra - high energies ( uhes ; @xmath5 ) , is still uncertain , but the discovery of sub - pev neutrinos by icecube @xcite may provide clues to this origin . given that the distribution of the arrival directions is consistent with isotropy , most events are likely to come from extragalactic pev accelerators ( * ? ? ? * and references therein ) , although a fraction of them may be attributed to galactic sources ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
relevant constraints on the parent cr acceleration and the isotropic - neutrino - background ( inb ) production were given by @xcite .
they showed that if @xmath6 interactions are responsible for the inb , the parent cr spectrum must have a hard spectral index @xmath7 in order for the by - product gamma rays not to overshoot the observed isotropic gamma - ray background ( igb ) @xcite .
conversely , a significant fraction of the inb and the igb may be attributed to a single type of source with a sufficiently hard spectral index . for this , the required local cr injection rate below the knee energies per energy decade is @xmath8^{-1 } \
\rm erg \ mpc^{-3 } \ yr^{-1}$ ] , where @xmath9 is the effective hadronuclear optical depth at the source .
importantly , the above injection rate is comparable to that required for the uhecr sources @xmath10 , or equivalently , the observed flux of the inb is consistent with the waxman - bahcall limit @xcite with @xmath11 .
this may indicate that uhecrs also originate in the same type of sources ( e.g. , * ? ? ?
note , however , that if indeed the parent cr spectrum of the inb is as hard as @xmath1 and the maximum energy is in the uhe range , @xmath9 may have to become low at @xmath12 in order for the by - product neutrinos to be consistent with the non - detection of super - pev neutrinos by icecube @xcite . in summary
, a significant fraction of the inb , the igb , and possibly also the uhecrs , can be consistently attributed to a single type of source based on a @xmath6 scenario if * the sources have the right local cr injection rate of @xmath13 , * the sources accelerate crs up to uhes with a hard spectral index of @xmath7 , and * the crs with @xmath14 lose a fraction of their energy via hadronuclear reactions , i.e. , @xmath15 , ( hereafter , we use @xmath16 ) .
individual local source identifications using e.g. , hawc @xcite and cta @xcite are the keys to testing the scenarios . here , we investigate galaxy merger shocks ( gmss ) as such a possibility . in the concordance cosmology , a good fraction of the massive galaxies experience a major merger and several minor mergers in a cosmological timescale . during the direct encounters between two galaxies ,
their cold - gas components are shocked with sufficiently large mach numbers to enable an efficient cr acceleration .
below , we show that gmss can meet the conditions ( ii ) and ( iii ) , and the anticipated cr injection rate can be a good fraction of that required by condition ( i ) for a cr acceleration efficiency of @xmath17 .
first , let us estimate the intrinsic energy budget of a galaxy merger .
hereafter , we mainly consider major mergers between massive galaxies with a stellar mass of @xmath18 , which likely dominate the total energy budget .
the energy dissipated by a gms can be estimated as @xmath19 , or @xmath20 where @xmath21 is the shocked gas mass and @xmath22 is the shock velocity . at least up to @xmath23 , the galaxy gas mass is typically @xmath2 of the stellar mass at the high - mass end with @xmath24 ( e.g. * ? ? ?
the shock velocity is essentially the relative velocity between the merging galaxies , which can range up from @xmath25 at a pericenter distance of @xmath26 in the point particle approximation .
the shock dissipation occurs with a dynamical timescale , @xmath27 and the energy dissipation rate per merger is @xmath28 the local major merger rate of massive galaxies has been estimated as @xmath29 ( e.g. , * ? ? ?
* ; * ? ? ?
equivalently , an @xmath30 of massive galaxies experience major mergers in a cosmological time given that the mean density is @xmath31 @xcite .
the possible cr injection rate by gmss can be estimated as @xmath32 , or @xmath33 where @xmath34 is the cr acceleration efficiency .
next , we consider the diffusive shock acceleration ( dsa ) mechanism ( e.g. , * ? ? ?
* ) at gmss . for this ,
we consider first the characteristics of gms ( see fig .
[ fig1 ] ) . the mach number of gmss can be estimated as @xmath35 , which is much larger than unity for the cold gas component with @xmath36 . here
, we use @xmath37 for ionized gas ( see below ) . on the other hand , the temperature downstream of the shock can be estimated as @xmath38 .
such a high downstream temperature plays an important role in characterizing gmss , first of all because the radiative cooling becomes relevant in this region . at @xmath39 ,
the main cooling channel is provided by metal line emissions and free - free emission ( e.g. , * ? ? ?
using an approximate form of the cooling function for a solar metallicity interstellar medium ( ism ) , the cooling timescale can be estimated as @xcite @xmath40 where @xmath41 is the preshock gas density . from eqs .
( [ eq : t_dyn ] ) and ( [ eq : t_rad ] ) , one can see that for a shock velocity @xmath42 , the downstream plasma cools within a dynamical timescale , i.e. , the gms shocks are radiative . in this case , a larger compression ratio than the @xmath43 ratio of strong adiabatic shocks is realized beyond the radiative region , which is @xmath44 from the shock surface .
such highly compressed regions are plausible sites for violent star formation during the galaxy merger @xcite .
second , the uv photons from the radiative zone can significantly ionize the upstream ism .
@xcite showed that this occurs for shock velocities @xmath45 . in this case
, one can naturally expect alfv@xmath46n - wave turbulence behind the gms , which is crucial for the dsa mechanism .
large - scale magnetic fields in interacting galaxies are observed to be @xmath47 @xcite .
one can also expect an amplification of the magnetic fields at least locally around the shocks as @xmath48 , or @xmath49 where @xmath50 is the amplification factor . in the dsa mechanism , the resultant cr spectral index , @xmath51 , where @xmath52 , is characterized by the compression ratio as @xmath53 in the test - particle approximation .
as discussed above , the compression ratio can be much larger than unity , @xmath54 , comparing the upstream and the far downstream beyond the radiative zone .
on the other hand , the size of the acceleration region in the downstream can be estimated as @xmath55 here , @xmath56 is the diffusion coefficient in the bohm limit , @xmath57 is the larmor radius , and @xmath58 is the electric charge of the crs . from eqs .
( [ eq : l_rad ] ) and ( [ eq : l_acc ] ) , the crs in the gms acceleration reside in the region where @xmath59 for @xmath60 , and thus , the cr spectral index can be @xmath1 . the maximum energy possible for crs is given by @xmath61 , where @xmath62 is the acceleration timescale .
this gives @xmath63 , or @xmath64 here , @xmath65 depends on the effective compression ratio and the magnetic field configuration , etc . for an @xmath66 parallel shock in the bohm limit , @xmath67 .
thus , one can expect @xmath68 for proton crs ( @xmath69 ) at gmss .
if highly efficient magnetic - field amplification , @xmath70 , is the case at gmss , the maximum energy could be further boosted up to uhes for , e.g. , iron crs ( @xmath71 ) , which may be consistent with the possible transition in the auger uhecr composition reported by @xcite ( but see also @xcite ) .
note , however , that such uhecr fluxes can be suppressed by dissociation of the nucleus during the acceleration @xcite .
also , even proton crs can be accelerated up to uhes with larger gms velocities as possible in galaxy collisions , @xmath72 , where the shocks are adiabatic . however , the size of the acceleration region of proton crs with @xmath73 can be larger than the width of the merging galaxy @xmath74 ( see eq .
[ eq : l_acc ] ) .
then , a fraction of such crs escape from the acceleration region , which also suppresses the uhecr flux .
next , let us discuss the energy loss of the crs , which we have so far neglected , focusing on proton crs in this section .
when a gms completes sweeping the galaxies , the high - pressure radiative region begins to cool down almost adiabatically , and so do the crs trapped in this region .
this occurs on a dynamical timescale , @xmath75 ( eq.([eq : t_dyn ] ) ) . on the other hand ,
the energy loss timescale due to @xmath6 interactions can be estimated as @xmath76 where @xmath77 and @xmath78 for @xmath79 , and the cross section evolves only logarithmically with cr energy .
the effective hadronuclear optical depth can be estimated as @xmath80 , or @xmath81 thus , crs predominately lose their energy via @xmath6 interactions .
note that other energy loss processes , such as synchrotron , are irrelevant at least up to @xmath82 . in @xmath6 interactions ,
the charged and neutral pions are produced with a ratio of @xmath83 .
the charged pions finally decay into three neutrinos and one positron with roughly the same energy , and the neutral pions decay into two gamma rays .
the flavor is totally mixed via neutrino oscillation for extragalactic sources .
the spectral index of both the neutrinos and gamma rays is approximately the same as that of the parent protons , i.e. , @xmath84 in our case .
the high energy cutoff is @xmath85 for neutrinos , and twice larger for gamma rays , although gamma rays above @xmath86 from @xmath87 are attenuated by the extragalactic background light .
first , let us estimate the neutrino flux per flavor from a local gms , that is , @xmath88 ( \xi_{cr}l_{gms}/{\cal c } ) ( 1/4 \pi d_l{}^2)$ ] , or @xmath89 here , @xmath90 is the luminosity distance to the source , and @xmath91 ( or @xmath92 for @xmath93 ) is the bolometric collection factor .
we take @xmath94 and @xmath95 for the estimate . with the above parameter
set , one can only expect a small muon event rate , @xmath96 , using icecube .
thus , individual local gms source identifications through high - energy neutrinos will be difficult in the coming years .
on the other hand , the gamma - ray flux from a local gms can be estimated as @xmath97 , or @xmath98 detecting such a flux is marginally difficult for _ fermi _ and current air cherenkov telescopes , but can be detectable by cta up to @xmath99 with a sufficiently long observation time .
candidate sources are discussed in the next section .
finally , let us estimate the diffuse flux from cosmological gmss .
generally , the inb flux can be estimated as ( e.g. , * ? ? ?
* ) @xmath100(\varepsilon_{cr } q_{\varepsilon_{cr}}).\ ] ] where @xmath101 is the hubble timescale with cosmological parameters @xmath102 , @xmath103 , and @xmath104 , @xmath105 accounts for cosmological evolution of the source , and @xmath106 . from eqs ( [ eq : inj_rate ] ) and ( [ eq : nu_diffuse_flux ] ) , one obtains @xmath107 which can be @xmath3 of the flux observed by icecube for @xmath108 ( see fig .
[ fig2 ] ) .
note that the observed local value , @xmath109 still has a factor of a few uncertainty , and the merger rate can evolve with redshift as @xmath110 with @xmath111 ( e.g. , * ? ? ?
such uncertainties in galaxy evolution can be absorbed in the @xmath112 factor .
a possible pev cutoff is consistent with the conservative estimate of the maximum cr energy , but the maximum neutrino energy can be larger if the uhecr acceleration is typically the case at gmss .
our arguments regarding the cr acceleration and the @xmath6 energy loss can be also applied to gmss in minor mergers .
although the energy dissipation in a minor merger is a factor of a few smaller , the event rate is a factor of a few larger than that of a major merger . given this ( or a higher cr acceleration efficiency , @xmath113 ) ,
the observed inb could be totally attributed to cosmological gmss .
the igb flux can again be tightly connected to the inb flux as @xmath114 . following the same reasoning as in @xcite , we see here that a fraction @xmath30 of the observed igb at @xmath115 can be attributed to gmss , for a slope of @xmath84
. corresponds to a case with @xmath70 and @xmath116 ( see eq .
[ eq : e_max ] ) .
the solid and dashed lines represent the inb and the atmospheric background flux observed by icecube , respectively.,width=340 ]
as we showed above , given the concordance cosmology and a cr acceleration efficiency of @xmath117 , gmss can provide at least @xmath3 of the observed inb .
this scenario can be indirectly tested by detecting gamma rays from the local merging galaxies using cta . within the cta detection horizon , @xmath99
, one can expect @xmath118 such systems .
two of the interesting candidates are the taffy " colliding galaxy pairs ugc 12914/5 and ugc 813/6 @xcite , located at @xmath119 and @xmath120 , respectively . in both cases , gas components of @xmath121 collided with a velocity of @xmath122 a few @xmath123 ago .
the bridge region between the galaxy pairs filled with shocked gases , and shows strong synchrotron emission which may come from non - thermal electrons accelerated by the gms @xcite .
another interesting candidate is vv 114 @xcite , which is a gas - rich merging galaxy pair with a core separation of @xmath124 , located at @xmath125 from the earth @xcite .
the galaxy interaction may have already triggered the starburst and active galactic nucleus ( agn ) activity in this system @xcite .
if the acceleration in gmss is efficient , we can again expect @xmath126 local sources within the gzk radius @xmath127 , which might form hotspots in the uhecr sky is not likely in our scenario .
here we use the gzk radius as a benchmark distance . ] .
the number density of gmss is smaller than the estimated number of sources required inside the gzk radius @xmath128 @xcite , but the flux from the merging galaxies could be @xmath17 of the observed one .
such uhecrs from local gmss can arrive at the earth on overlapping time windows with neutrinos and gamma rays for an intergalactic field of @xmath129 ( e.g. , * ? ? ?
finally , we should note that the gms scenario is competitive with , but not easy to separate from , other scenarios .
a fraction of starburst and agn activities are triggered by galaxy mergers , as observed in vv 114 .
these latter activities can also inject a sufficient amount of crs ( e.g. , * ? ? ?
an identification in terms of morphology is possible based on extensive computational simulations of merger histories . a more promising way to discriminate between the contributions from gmss and
those of starbursts may be to select for samples with relatively low far - infrared / radio flux ratios , which is observed in taffy galaxies @xcite .
, f. c. 1998 , in american institute of physics conference series , vol .
433 , workshop on observing giant cosmic ray air showers from > 10(20 ) ev particles from space , ed .
j. f. krizmanic , j. f. ormes , & r. e. streitmatter , 3741 | we investigate the shock acceleration of particles in massive galaxy mergers or collisions , and show that cosmic rays ( crs ) can be accelerated up to the second knee energy @xmath0 eev and possibly beyond , with a hard spectral index of @xmath1 .
such crs lose their energy via hadronuclear interactions within a dynamical timescale of the merger shock , producing gamma rays and neutrinos as a by - product .
if @xmath2 of the shock dissipated energy goes into cr acceleration , some local merging galaxies will produce gamma - ray counterparts detectable by cta .
also , based on the concordance cosmology , where a good fraction of the massive galaxies experience a major merger in a cosmological timescale , the neutrino counterparts can constitute @xmath3 of the isotropic background detected by icecube . |
in the main text , we show numerical calculations of an edge - state wave packet which is scattered by an electrostatic puddle and extract the time - dependent probability density @xmath3 of the wave packet being localized in the puddle .
we model the electronic structure of a hgte / cdte quantum well using the bernevig - hughes - zhang hamiltonian @xcite , @xmath31 with spin - subblock hamiltonians where @xmath33 and @xmath34 .
intrinsic spin - orbit coupling of the dresselhaus - type is included by the parameter @xmath35 in eq .
( [ bhzhamilton ] ) , which mixes the two spin blocks @xcite .
we use an electrostatic potential @xmath36 to model circular and stadium shaped puddles and confine the states by the potential @xmath37 to get a quantum spin hall edge state at the system boundary . for the calculations presented in the manuscript ,
the potential strength of the puddle was set to @xmath38 leading to hole - like states . as initial state , we create a gaussian edge - state wave packet @xmath39 , which is assembled in reciprocal space along the boundary . throughout the manuscript
, we use a gaussian with a width in position space of @xmath40 . also , we include only states propagating towards the electrostatic puddle , leading to a strongly spin - polarized wave packet .
we calculate the propagation of the wavefunction @xmath41 using a propagator based on an expansion of the time - evolution operator in chebyshev polynomials @xcite . during the propagation of the wave packet @xmath41
, we integrate @xmath42 over the puddle region resulting in the time - dependent probability density @xmath3 . in order to avoid effects of mesoscopic fluctuations
, we average over 20 different configurations , which differ in a random impurity potential with an amplitude of @xmath43 and a wall distortion of @xmath44 . in the following section
, we will show how dephasing is included in the numerical calculations presented in fig .
[ figdecohnumerics ] . the dephasing algorithm described here
is inspired by the concept of einselection ( `` environment - induced superselection '' ) pioneered by w. zurek @xcite , which is a mechanism proposed to understand the influence of dephasing on quantum systems and , more generally , to explain the quantum - to - classical transition . according to einselection
, the interaction of an open system with its environment leads to decoherence , which causes a decay of the quantum states of the system into an incoherent mixture of so - called pointer states .
this strongly suppresses quantum interference effects between different pointer states on the time scale of the dephasing time @xmath22 .
the character of the set of pointer states , the pointer basis , heavily depends on the environment and the coupling to it .
for example , for very weak coupling to the environment , the pointer basis coincides with the set of energy eigenstates of the system .
however , in the case of an intermediate system - environment coupling based on a local interaction , e. g. , in the case of electron - phonon or electron - electron coupling , it is expected to be a set of states that is localized in phase space , i. e. , in position and momentum . as in the puddle lifetime calculations
, we again employ a numerical time evolution based on a single pure state , i. e. , we do not use a representation in terms of density matrices , which would drastically increase the computational effort .
still , we incorporate the dephasing - induced interference suppression by occasionally ( partially ) randomizing the phases of the components of the state vector in a representation that tries to faithfully mimic a decomposition in terms of the pointer basis .
this randomization is done at fixed event times @xmath45 , which are sampled from an exponential distribution with time constant @xmath46 , i. e. , we assume the events to be fully uncorrelated ( poisson process ) . in between these events , the propagation is done fully coherently using the propagator based on a polynomial expansion mentioned above .
the decomposition and the subsequent randomization is done in the following way : at the time of an event @xmath45 , we extract a set of pseudo eigenstates , @xmath47 at the energies @xmath48 using a short - time propagation of the wave packet @xmath41 around the time @xmath45 .
these states fulfill the requirement that they are both local in energy as well as in position space ( as they are extracted from a propagation over a finite time interval ) .
the degree of localization in position space can by tuned by the propagation time @xmath49 , and a matching set of energies , this approach yields the expected pointer basis in the limit of very weak coupling to the environment : the exact energy eigenstates of the system . ]
which we choose @xmath50 .
we use these states @xmath51 to spectrally decompose the current state before the event @xmath52 , as sketched in fig .
[ figdecoposition ] .
for this , we first orthogonalize them with a gram - schmidt process , leading to the set @xmath53
. then , we calculate the weights @xmath54 , which can be used to express the current state as the decomposition @xmath55 since the set of 11 states is not sufficient to describe the wave packet @xmath52 exactly , we also allow for account a small residual part @xmath56 . in this decomposed representation
, the dephasing can be easily added by modifying the set of amplitudes to @xmath57 with the random numbers @xmath58 $ ] .
subsequently , we create the new wave packet @xmath59 which is used in the following propagation . with this dephasing algorithm , we are able to calculate the time evolution in arbitrary mesoscopic systems .
it roughly conserves the energy , the spatial extent , as well as the total spin of the wave packet , which is what one would expect from the interaction with a spin - unpolarized ( non - magnetic ) environment .
the strength of a single dephasing event can be tuned by the degree of randomization of the phase in each event , i. e. , by the parameter @xmath60 .
this also makes the regime of weak but frequent dephasing accessible ( compared to the regime of rare and very strong dephasing which allows a simpler model treatment discussed in the first part of the manuscript ) . however , to compare with the experimentally accessible dephasing time @xmath22 , which is the time for full phase coherence loss , one has to find the relation between the time constant of the events @xmath46 and the time for full dephasing @xmath22 .
this can be done in the following way : the average correlation of a state in the pointer basis @xmath61 with itself after @xmath62 events is given by , @xmath63 with the chance for @xmath62 events occurring after time @xmath64 in a poisson process , @xmath65 one can evaluate the average decay of autocorrelation after time @xmath64 , @xmath66\nonumber \\ = \ , & \exp\left[-\frac{t}{\tau_{\mathrm{e}}}\left(1-\frac{\sin\left(\pi q\right)}{\pi q}\right)\right]=\exp\left[-\frac{t}{\tau_{\phi}}\right],\end{aligned}\ ] ] with @xmath67 yielding the desired relation between decoherence time @xmath22 and the time constant of the events @xmath46 .
to model the frequent but weak dephasing regime , we use @xmath68 , thus , @xmath69 . in fig .
[ figdecohnumerics ] , we apply the dephasing scheme on three different electrostatic puddle shapes to show how the coupling between the puddle and the edge states affects the total spin randomization of the puddle in presence of dephasing . a video of a sample propagation can be also found online @xcite . | to understand the seemingly absent temperature dependence in the conductance of two - dimensional topological insulator edge states , we perform a numerical study which identifies the quantitative influence of the combined effect of dephasing and elastic scattering in charge puddles close to the edges .
we show that this mechanism may be responsible for the experimental signatures in hgte / cdte quantum wells if the puddles in the samples are large and weakly coupled to the sample edges .
we propose experiments on artificial puddles which allow to verify this hypothesis and to extract the real dephasing time scale using our predictions .
in addition , we present a new method to include the effect of dephasing in wave - packet - time - evolution algorithms .
the discovery of two - dimensional topological insulators , 2d - tis , stirred up hope for potential applications using their ability to host reflectionless one - dimensional transport . in practice , however , the transport along the edges of the state - of - the - art realizations of 2d - tis exhibits a length - dependent resistance which increases above the expected quantized value in samples significantly larger than @xmath0
. moreover , most notably , the edge state resistance seems to exhibit no observable temperature dependence in a wide temperature range ( @xmath1 ) @xcite .
this seems to be hard to reconcile with the notion that the backscattering on 2d - ti edges is due to inelastic processes .
still , this behavior has been been frequently observed in various recent experiments on 2d - tis , both based on hgte / cdte @xcite and inas / gasb samples @xcite .
understanding this apparently generic feature is important for the field , as it might show up ways to improve the edge transport in existing material systems . as coherent elastic backscattering
is symmetry forbidden at 2d - ti edges @xcite , there have been theory studies considering alternative backscattering mechanisms . besides the possibility for elastic backscattering by magnetic impurities which explicitly break time - reversal symmetry @xcite , they mainly focused on inelastic backscattering by coulomb interaction or phonons on the edge @xcite as well as coulomb backscattering in quantum dots which are expected to appear naturally along the edges of the currently used material systems ( hgte / cdte and inas / gasb quantum wells ) due to trapped charges at the gate insulator interface @xcite .
in addition , there have been a few proposals in which the inelastic interaction is mainly held responsible for breaking the electron phase coherence while the backscattering is then attributed to elastic scattering . for a dephasing process which does not explicitly flip the spin
, it was found that one does not expect backscattering along a clean edge of a 2d - ti ribbon , as backscattering requires a full spin flip in this setup @xcite .
this changes if a quantum dot in which the extended states are naturally spin mixed due to the spin - orbit coupling is present along or close to the edge . in this setup , even a decoherence mechanism which conserves
the average spin may lead to backscattering , as has been qualitatively shown in ref .
where dephasing was mimicked by adding lattice sites with an imaginary self energy . in this publication
, we also address dephasing - induced backscattering . however , given the fact that the interesting @xmath2-independent edge state resistance has been consistently observed in various systems and material classes , we do not attempt to assign it to a specific source for decoherence along with a corresponding specific @xmath2-dependence of the dephasing / decoherence time . instead , assuming that edge channels are coupled to nearby puddles due to spatial charge inhomogeneities @xcite we present a more general study which is aiming at a quantitative description of backscattering through the interplay of dephasing and spin mixing in such quantum dots .
this includes to consider puddle dwell times and couplings to the puddles , which are calculated for realistic model hamiltonians for the case of the hgte / cdte material system .
thereby , we extract a range of dephasing times that would be compatible with the measured temperature - independent resistance , and which is a function of the puddle dwell time .
this dependence of the conductance on the puddle properties could be checked in future experiments with artificially created puddles . for our calculations
, we use the bernevig - hughes - zhang ( bhz ) hamiltonian @xcite discretized on a square grid .
intrinsic spin - orbit coupling of the dresselhaus type is included , which mixes the two spin blocks .
we study the time evolution of a wave packet which is initially localized on the edge of a 2d - ti and approaches a nearby puddle defined by a local electrostatic potential close to the edge , cf .
fig [ figgeometry ] , employing a method which has been shown to be well suited to describe the edge - state dynamics in hgte / cdte 2d - tis @xcite .
the detailed calculation setup is described in detail in the appendix . from the time evolution of the wave packet , we calculate the time - dependent probability density @xmath3 to find the charge carrier in the puddle , which includes edge - puddle - coupling and lifetime effects .
impurity - configuration - averaged results of such calculations for various puddle sizes are shown in fig .
[ figpuddleonedge ] . as can be seen from the short - time dynamics , fig .
[ figpuddleonedge](a ) , the wave packet enters the cavity on a time scale of a picosecond . a small fraction of the density
, @xmath4 depending on the puddle size , exits the puddle after a ballistic traversal . however , most of the density stays in the puddle and decays only slowly with an almost constant outflow , visible as a linear density decay for intermediate times , cf .
fig .
[ figpuddleonedge](b ) .
the shown density is scaled to the total density of the wave packet and the fact that values close to one are reached reflects the forbidden backreflection on the clean edge : most of the density has to couple into the puddle in this geometry .
as soon as the absolute density in the puddle falls below @xmath5 , the time dependence of the density changes into a power law .
this is best seen in fig .
[ figpuddleonedge](c ) , which shows the negative time derivative of the density in a log - log plot . here
, one clearly recognizes the linear density decay as a plateau with superimposed oscillations , which can possibly be attributed to ballistic orbits in the puddle .
this turns over into a @xmath6-power - law decay for long times as indicated by the dashed lines . and implies @xmath7 for the density decay . for the @xmath8 sized puddle
, the best - fitting exponent would be slightly smaller , rather like @xmath9 , i.e. , @xmath10 .
] such power - law behavior is known @xcite and similar results have already been obtained in wave - packet simulations on other material systems @xcite . with only a few additional assumptions ,
the knowledge of the dwell time of the electrons in the quantum dot will allow us to estimate the effects of dephasing on its transport properties .
the first one is that the decoherence disturbs the electron phase evolution but does not explicitly lead to spin flips , which is quite realistic for setups without magnetic fields and a very low density of magnetic impurities .
due to the strict spin - momentum locking of the ti edge states , this spin conservation implies that the dephasing process will not lead to backscattering as long as the electrons are propagating along the edges . in a quantum dot however ,
the spin is fully mixed already after a very short time , due to the combined effect of impurity scattering and intrinsic spin - orbit coupling ( see fig . [ figgeometry ] ) .
thus , the efficiency of backscattering due to dephasing will depend on the relation of the dephasing timescale and the electron lifetime in the puddle . to make this quantitative , we assume that decoherence is statistically occuring with independent events .
we first additionally assume that each event leads to full phase loss ( an assumption that we will later refine ) .
thus , the electron density being in the puddle at the event will leave it in a random direction in the subsequent dynamics . then
, the average reflection can be estimated by @xmath11\nonumber \\ & & -\frac{1}{2\tau_{\phi}^{2}\rho_\mathrm{max}}\int_{-\infty}^{t_{\mathrm{max}}}\ ! \ ! \ ! \ ! dt_{1}\int_{t_{\mathrm{max}}}^{\infty}\ ! \ ! \ ! \ !
dt_{2}\rho(t_{1})\rho(t_{2})e^{-\frac{t_{2}-t_{1}}{\tau_{\phi}}},\label{eq : bsformula}\end{aligned}\ ] ] where @xmath12 is the mean time between dephasing events and @xmath13 is the time at which the density @xmath3 in the puddle attains its maximum @xmath14 .
equation ( [ eq : bsformula ] ) makes use of the piecewise monotonous structure of @xmath3 and can be understood as follows : suppose that the first dephasing event occurs at a time @xmath15 with the density already decaying monotonously .
as each event is assumed to lead to full phase randomization , all subsequent events will not matter as the propagation in the puddle is already fully random .
hence , the knowledge of the first event after reaching @xmath16 is enough to determine the total backscattering probability for the time window of the decay . it can be calculated as an expectation value of @xmath3 with an exponential distribution for the dephasing event , describing the mean waiting time in a poisson process .
this main contribution enters eq .
( [ eq : bsformula ] ) as the second term . the first term in eq .
( [ eq : bsformula ] ) describes the backscattering during the ( monotonous ) rise of @xmath3 . here
, the argument can be reversed as the latest event will determine the amount of backscattered density in the time window of the rise .
the last term in eq .
( [ eq : bsformula ] ) takes care of the double counting that occurs for events both on the rising and the falling edge of @xmath3 . for the shape of the density curves obtained from the numerical calculations
, one can simplify the above expression .
since the rising edge is very short compared to the decay , it suffices to consider only the second term of eq .
( [ eq : bsformula ] ) .
in addition , we showed above that the subsequent decay is approximately linear .
the power law decay can be omitted as long times are exponentially suppressed in the integral .
we then find for the average transmission , using @xmath17 $ ] , @xmath18 here @xmath19 is the time after which the puddle would be empty assuming a steady linear decay .
the values for @xmath19 which were extracted from the linear fits to @xmath3 are plotted in a log - log plot in the inset of fig .
[ figpuddleonedge](b ) .
one finds that @xmath19 approximately scales like @xmath20 . ) , while the dotted lines show the result from eq .
( [ eq : simpl_linear ] ) , assuming a linear decay . ] using these formulae , we can estimate the average transmission probability for a single puddle of size @xmath21 for a given decoherence time @xmath22 as shown in fig .
[ figdecohfrommodel ] .
as anticipated , we find good agreement between the results from the full model , eq .
( [ eq : bsformula ] ) , and the approximation with a linear fit , eq .
( [ eq : simpl_linear ] ) . both show a characteristic sigmoidal behavior with a saturation at @xmath23 for small @xmath22 . for the smaller puddles , the extracted value for @xmath16
is underestimated for the linear fit , see fig .
[ figpuddleonedge](a ) .
more generally , the value of @xmath16 is not only given by the puddle size , but also by the coupling of the puddle to the edge and it would therefore be strongly dependent on the distance of the puddle from the edge . this coupling will also influence the dephasing - time scale below which one observes the nearly constant backscattering probability : with worse coupling , there will be less backscattering per puddle but the saturation will already be reached earlier .
the extracted cutoff times therefore refer to the `` perfect coupling limit '' meaning that they may be interpreted as a lower bound for the cutoff time .
so far , the dephasing only entered the model a posteriori but was not included in the calculation of the electron dynamics .
we found a way to include it in the wave - packet dynamics calculation using an algorithm which is inspired by the concept of einselection @xcite .
it treats dephasing in a rather general fashion and allows not only to vary the strength of single dephasing events but also to study the influence of dephasing on the wave - packet dynamics itself . for details of the implementation
, we refer to the appendix .
the underlying interaction responsible for the dephasing is left unspecified and the dephasing time remains as a parameter .
also , the implementation may not be applicable for all kinds of environments but we think that it captures the important effects and expect similar results for other implementations . ) . ]
our results for the total transmission of dynamical dephasing calculations are shown as symbols in fig .
[ figdecohnumerics](d ) . here
, the dephasing was chosen to be less effective such that one needs on average 2.75 events to achieve full dephasing ( the time scale of the dephasing time axis corresponds to full dephasing ) .
the dotted lines show a fit to the data using eq .
( [ eq : simpl_linear ] ) which expectedly agrees well for small dephasing times . in this limit
, the time between the events is small compared to the dynamics of the system , thus many small events lead to the same net result as one strong event . for longer dephasing times , however , in our refined calculation using partial dephasing , it comes into play that the dynamics of the system will be strongly influenced by the dephasing as this effectively opens another exit channel for the puddle thus decreasing the electron lifetime .
this can be nicely seen in the plots of the time - dependent density for different dephasing times in fig .
[ figdecohnumerics](a c ) . in the long dephasing time limit , many frequent weak events lead to more backscattering than one rare strong event , as the second exit channel ( backscattering ) will be ( partially ) opened already at an earlier time .
fig .
[ figdecohnumerics ] also shows the influence of different puddle geometries on the transmission and the system dynamics .
the results match the expectation that puddles further away from the edge will lead to less backscattering in the strong dephasing limit but will also saturate below a higher dephasing time cutoff . with this
, we can now relate to the experimental results on 2d - ti samples .
so far , the measured resistance does not show any observable temperature dependence in the experimentally accessible @xmath2-range ( @xmath24@xmath25 ) @xcite .
this would be in line with our findings if the dephasing times in the experimental samples were below the cutoff time even at the lowest temperature .
then , an increase of the temperature ( and a connected reduction of the dephasing time ) would not lead to additional backscattering . to make a specific example , if the backscattering in the sample was mainly caused by @xmath26 puddles which are well coupled to the edges
, this would imply that the dephasing time should be shorter than @xmath27 .
for a rough comparison , from low - temperature magnetoconductance measurements on 2d hgte quantum well samples , dephasing times of @xmath28 were extracted @xcite .
thus , given that the coupling to the puddles in the real samples is likely to be smaller , it might indeed be that the lack of observed temperature dependence is mainly due to the inherently short dephasing times in these materials .
we only consider puddles of a depth of @xmath29 as we are limited to the range of validity of the bhz - hamiltonian .
deeper puddles , similar to larger puddles , would also shift the curves toward longer dephasing times and decrease the temperature above which the conductance is expected to be temperature independent . the puddle depth and
the connected change of electron density may also influence @xmath22 , however , at least in the 2d - limit , recent experiments show controversial results whether an increasing density leads to an increased @xcite or a decreased @xcite dephasing time .
note that , in line with experiments , our model predicts reproducible conductance fluctuations which are similar to ucfs as the coupling to the puddles is expected to be depending on gate voltage and magnetic field .
however , the conductance is not @xmath2-dependent as long as the dephasing time is below the cutoff . to make a definite statement
, one would have to experimentally check this hypothesis which could be done with experiments on artificially created puddles .
as there are experimental samples which show the quantization , it seems to be possible to create puddle - free samples in which one could artificially introduce puddles of varying size using a local gate .
this should allow for experimentally reproducing the calculated sigmoidal curve and extract the temperature - dependent dephasing time . to summarize , we quantitatively studied the backscattering of 2d - ti edge states due to the interplay of dephasing and dynamical scattering in charge puddles .
our results suggest that the seemingly absent temperature dependence could be due to the saturation of backscattering at dephasing times which are short compared to the puddle lifetimes .
one should be able to verify this hypothesis quantitatively with experiments on small ( @xmath30 ) artificial puddles , which , according to our study , should show a detectable temperature dependence at the experimentally available temperatures and from which one could extract the actual dephasing time . in passing ,
we developed a scheme to implement dephasing into wave - packet - time - evolution algorithms which is general and could be used in a wide range of scenarios .
it is described in detail in the appendix . for the charge puddles in 2d - tis
, it would also allow for the explicit inclusion of other sources of backscattering , like external magnetic fields , which we recently showed to have a strong influence on the backscattering due to puddles @xcite .
this work was supported by dfg spp 1666 and the enb graduate school `` topological insulators '' .
we thank l. molenkamp and m. wimmer for fruitful discussions .
35ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physrevlett.113.026804 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.125305 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.165309 [ * * , ( ) ] link:\doibase 10.1038/nmat3682 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.096802 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.226801 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.256803 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.236402 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.206803 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.106401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.73.045322 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.256804 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.235304 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.121106 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.086602 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.156402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.075118 [ * * , ( ) ] , link:\doibase 10.1103/physrevb.89.235136 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.216402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.115309 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.036803 [ * * , ( ) ] link:\doibase 10.1126/science.1174736 [ * * , ( ) ] link:\doibase 10.1103/physrevx.3.021003 [ * * , ( ) ] link:\doibase 10.1126/science.1133734 [ * * , ( ) ] link:\doibase 10.1143/jpsj.77.031007 [ * * , ( ) ] link:\doibase 10.1088/0031 - 8949/82/03/038101 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.086803 [ * * , ( ) ] http://journals.aps.org/pra/abstract/10.1103/physreva.45.701 [ * * , ( ) ] link:\doibase 10.1103/physrevb.67.085320 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.75.715 [ * * , ( ) ] link:\doibase 10.1134/s0021364012230099 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.235312 [ * * , ( ) ] link:\doibase 10.1088/2053 - 1583/2/2/024005 [ * * , ( ) ] link:\doibase 10.1063/1.448136 [ * * , ( ) ] |
coalescing binaries are important and copious sources of gravitational waves for the projected interferometric gravitational wave detection experiments ligo , virgo and lisa .
the requisite signal templates challenge theoreticians to predict the behaviour of such systems under gravitational radiation backreaction .
much progress has been made in this direction mainly by perturbative approaches . in the initial epoch ,
the separation @xmath8 of the two bodies forming the binary is large relative to the schwarzschild radius and the relative motion can be considered slow .
this enables one to employ the post - newtonian expansion parameter @xmath9 @xcite .
alternatively , with the system of units chosen appropriately , the inverse @xmath10 of the speed of light can be used as an expansion parameter@xcite .
computations of instantaneous radiation losses of energy , momentum and angular momentum up to ( post@xmath11-newtonian order , including the spin - orbit and spin - spin interaction terms , were given by kidder@xcite .
the equations of motion of spinning bodies , disregarding radiation backreaction , were first considered by barker and oconnell@xcite , then by thorne and hartle@xcite .
the relevant equations can be derived from a generalized lagrangian depending on relative position , velocity and acceleration .
the two - body problem is reduced to a one - body problem by eliminating the center of mass .
the masses @xmath12 and @xmath13 still appear in the lagrangian as reminders of the initial two - body character of the problem .
apostolatos , cutler , sussman and thorne@xcite describe the evolution of the spin and orbital angular momentum vectors in the presence of radiation backreaction .
they supplement the equations of motion with the leading - order averaged gravitational radiation loss terms , computed by peters and mathews@xcite . in the test particle
limit the leading spin terms from the averaged losses in energy , magnitude of the orbital angular momentum and spin projection of the orbital angular momentum , respectively has been computed by ryan@xcite for generic orbits .
generalizing his earlier work on circular orbits@xcite and shibata s work on equatorial orbits@xcite , ryan obtains the first - order spin corrections to the peters equations of radiation - induced change in the orbit parameters . in a previous paper@xcite , to be referred to as * i * , we completed the description of the test particle by giving the averaged losses in terms of unambigous conserved quantities and by computing the radiation induced change of the remaining orbit elements .
our description relies on the use of eulerian angle variables , a new parametrization of the orbit and averaging by the residue theorem .
the computation of finite - mass effects can be carried out by choosing a suitable spin supplementary condition ( _ ssc _ ) .
three convenient choices of _ ssc _ are discussed by kidder@xcite .
a treatment of the nonradiative evolution using a noncovariant _ ssc _ ( a2b ) of @xcite can be found in wex s work@xcite , and some radiative losses in rieth and schfer@xcite .
the finite mass effects in the spin terms of averaged losses , using the covariant _ ssc _ , were considered previously by kidder , will and wiseman@xcite and by kidder@xcite .
they computed the energy loss and the angular momentum loss for the case of circular orbits .
it is the purpose of the present paper to obtain , for generic orbits , the averaged losses of energy , magnitude of the orbital angular momentum and spin projection of the orbital angular momentum .
we employ the second order lagrangian suitable for the description of the motion of comparable mass bodies in the presence of spin , following ref .
we want to describe finite - mass effects in binaries consisting of a compact body captured by another , massive , spinning body .
the latter is exemplified by a spinning black hole or neutron star . in this way
we generalize the picture of the evolution of a system consisting of a black hole and a test particle , which we considered in * i*. the equations of motion and the spin precession equations valid to the ( post@xmath14-newtonian order are reviewed in the second section .
the constants of motion are the energy @xmath4 and the total angular momentum @xmath15 .
since our goal is to get the radiation losses up to the leading terms in the spin , we approximately picture the spin precession by a rigid paralelogram with sides @xmath1 and @xmath16 , rotating about the total angular momentum vector @xmath15 . in this approach
the magnitude @xmath1 and the spin projection @xmath17 of the orbital angular momentum are constants of the motion . from their expressions
an uncoupled radial equation can be found . in the third section
, we determine the turning points .
we then introduce two parameters , the eccentric anomaly and the true anomaly. they have the respective properties , and in the no - spin limit reduce to , the eccentric anomaly and true anomaly of kepler orbits .
the parametrization of the orbit by the eccentric anomaly enables one to compute the orbital period defined as twice the period of time elapsed between consecutive turning points . in the fourth section ,
we compute the instantaneous radiation losses . for this purpose
, we make the assumption widely used in the literature that the radius @xmath18 of the spinning body is comparable with its mass @xmath12
. then @xmath19 and @xmath20 .
thus the spin is of @xmath21 order smaller than the orbital angular momentum . under the assumption that the rotational velocity @xmath22 is large , apostolatos , cutler , sussman and thorne@xcite estimated the @xmath16-changing piece of the radiation - reaction torque ( the radiation loss of the spin ) to be of order @xmath23 .
this is smaller by two full post - newtonian orders than the leading term in@xcite @xmath24 which is of order @xmath25 , and by one half order than the spin - orbit terms given by kidder@xcite which are of order @xmath26 . at the end of the section we present an estimate of the order of magnitude of the radiation loss of the spin for smaller values of @xmath22 .
we then find that , to the order we are interested in , we may interpret the instantaneous total angular momentum loss as consisting entirely of the loss in the orbital angular momentum .
these considerations enable us to compute the loss of magnitude and spin projection of the orbital angular momentum . however the latter will still contain radiative spin loss terms , which are evaluated from the burke - thorne potential .
fortunately , introducing suitable new angular variables , all losses can be expressed in terms of the true anomaly parameter alone . in the fifth section we average these expressions by using the residue theorem .
we find that the spin - loss terms give no averaged contribution to the losses .
the finite - mass backreaction effects are of the same order as the other spin effects .
the results obtained in the lense - thirring case emerge as smooth limits , although the test particle case is outside the framework of this paper .
the losses are given both in terms of conserved quantities and in terms of ellipse parameters @xmath6 and the angle @xmath3 between the spin vector @xmath0 and the orbital angular momentum @xmath27 .
we give also the changes of the geometric orbit elements @xmath6 and @xmath3 due to radiation backreaction .
this is carried out in the sixth section . as in acst
, we separate the motion of the spin and orbital angular momentum vectors in their momentary planes from the motion of the plane . our approximation in describing the radiation effects on the former motion up to linear terms in the spin . since we eliminate the post - newtonian parameter @xmath28 from the losses at the very beginning of the averaging process , we keep the gravitational constant @xmath29 and the speed of light @xmath30 in our formulae .
our post - newtonian parameter is @xmath31 .
the explicit @xmath30 dependence allows an easier bookkeeping in the post - newtonian expansions .
we consider the motion of a bound two - body system with masses @xmath12 and @xmath13 and spins @xmath32 and @xmath33 , respectively . considering only the leading - order spin - orbit coupling , and adopting the spin supplementary condition of@xcite ,
the lagrangian is : @xmath34 where @xmath35 is the relative distance , @xmath36 is the relative velocity , and the perturbation term due to rotation effects is @xmath37 here @xmath38 is the reduced mass and @xmath39 the total mass of the system , @xmath40 the lagrangian depends on the relative acceleration @xmath41 .
the application of a generalized , second - order variational formalism yields the equations of motion : @xmath42 -(4 + 3\eta){\bf v}\times { \bf s } + 3(2+\eta){\dot r\over r}{\bf r}\times { \bf s } \right\}\ , \label{acc}\ ] ] and the momenta : @xmath43 @xmath44 the energy @xmath4 and the total angular momentum @xmath15 are constants of motion : @xmath45 @xmath46 here the orbital angular momentum @xmath2 is @xmath47 with the newtonian orbital angular momentum @xmath48 and the spin - orbit term @xmath49 -{\eta\over 2}\left[{\bf v}\times ( { \bf v}\times{\bf s})\right ] \right\}\ .\ ] ] formally the spin terms arising in these expressions are of order @xmath50 .
since the spin is of order @xmath51 , the terms containing the spin in the lagrangian , the acceleration , the energy and the orbital angular momentum are of order @xmath52 higher than the corresponding newtonian terms . the spin precession equation@xcite is : @xmath53 hence the magnitude @xmath16 of the spin is constant .
note that up to this point the description is valid up to ( post)@xmath54-newtonian order .
since we are interested in the leading spin terms of the gravitational radiation backreaction losses , we may replace @xmath55 by inserting terms of higher order in @xmath16 .
this is allowed , as we are neglecting terms of order @xmath56 in the precessional angular velocity , which are @xmath57 order higher then the spin terms that we have kept .
( we shall return to the order - of - magnitude estimates at the end of this section . ) from the conservation of total angular momentum ( [ j ] ) and the ( [ sprec ] ) precession of @xmath0 about @xmath15 , the vector @xmath2 is found to precess about @xmath15 : @xmath58 combining ( [ eq : sprec ] ) ( with @xmath15 instead of @xmath59 ) and ( [ eq : lprec ] ) it follows that the angle @xmath3 subtended by @xmath0 and @xmath2 is constant . as shown in appendix a , the angle @xmath3 is constant to order @xmath57 .
thus we have the picture of a paralelogram of constant sides @xmath1 and @xmath16 rigidly rotating about its diagonal @xmath15 ( fig.1 ) .
we choose the cartesian coordinates @xmath60 of the reduced - mass particle with origin at the center of mass such that @xmath15 points along the @xmath61 axis@xcite . in this approximation
we can choose the conserved quantities characterizing the orbital motion as : the energy @xmath4 ( due to the time - independent nature of the lagrangian ) , the magnitude @xmath1 of @xmath2 ( the vector @xmath2 undergoes a pure rotation in the order we are considering ) and @xmath17 , the spin - oriented component of @xmath2 ( _ cf .
_ appendix a ) : @xmath62 this description is close to the one in the lense - thirring approximation .
the inclusion of finite mass does not destroy the basic features of the formalism developed in the test particle limit . in computing the expressions of the constants of motion ,
@xmath63 was replaced by @xmath64 , which is compatible with the order we are considering . using polar coordinates , @xmath65 , @xmath66 , @xmath67 ,
the constants of the motion are : @xmath68 } -{g m \mu\over r } + \eta { gl_ss\over c^2r^3 } \ , \label{e}\\ \label{eq : l2 } l^2&=&{\mu^2}r^4(\dot\theta^2+\sin^2\theta\ \dot\varphi^2 ) -4{g\mu
l_ss\over c^2r } + { 2\eta\over c^2m}el_ss \ , \label{l2 } \\ \label{eq : ls } l_s&=&{\bf l}\cdot { { \bf s}\over s}\ = \
l\cos{\kappa } \ . \label{ls}\end{aligned}\ ] ] from the first expression of the energy , @xmath28 can be expressed in terms of constants of motion and @xmath8 : @xmath69 from the second expression of the energy in ( [ eq : e ] ) and from ( [ eq : l2 ] ) , the equation for the radial motion is found to decouple from the angular degrees of freedom : @xmath70 we conclude this section by considering some order - of - magnitude estimates in the approximation which has led to the picture of a rigidly rotating paralelogram . the neglected time - dependent terms in @xmath71 and @xmath17
are @xmath72 and @xmath73 , respectively .
these are of order @xmath57 and @xmath74 smaller than the terms we keep . in the radial equation ( [ eq : radial ] ) ,
as declared , we keep terms linear in the spin and drop terms of order @xmath57 smaller .
in this section , we employ eq .
( [ eq : radial ] ) for determining the turning points and then finding convenient parametrizations of the orbit .
we describe the procedure only in its outlines , as a detailed exposition was previously given for the lense - thirring case in * i*. the condition for turning poits @xmath75 yields a cubic equation . by looking for roots slightly different from the roots known for the nonspinning case , one unphysical root and the turning points @xmath76
are found : @xmath77\ .
\label{turning}\ ] ] here @xmath78 is the length of the runge - lenz vector to the zeroth order in the spin : @xmath79 as in the lense - thirring case , we define the eccentric anomaly parametrization of the kepler orbits , @xmath80\cos\xi\ , \label{xi}\ ] ] such that @xmath76 is at @xmath81 and @xmath82 , respectively .
the parameter derivative @xmath83 follows from ( [ xi ] ) . expressing @xmath84 from ( [ radial ] ) and linearizing in @xmath16 , the expression for @xmath85
is found : @xmath86 integration of ( [ txi ] ) from 0 to @xmath87 gives the orbital period : @xmath88 note that the expression for the period has the same functional form as in the lense - thirring case .
the eccentric anomaly parametrization @xmath89 was useful in computing the orbital period , but it leads to unnecessarily complicated expressions when instantaneous losses need to be averaged . for the latter purpose
we introduce the true anomaly parametrization @xmath90 requiring that it has the following properties : @xmath91 where @xmath92 are constants .
property ( b ) generalizes kepler s second law for the area
. the unique parametrization satisfying both ( a ) and ( b ) is : @xmath93 in the same way as @xmath85 [ eq . ( [ txi ] ) ]
was obtained , @xmath94 is found to have the form @xmath95 \ . \label{tchi}\ ] ] the true anomaly parametrization ( [ chi ] ) will be used for the integration over one period of all instantaneous losses . in the next section ,
all of these losses will be expressed in the form @xmath96 .
they contain no other @xmath97 dependent factor in the denominator than @xmath98 , where @xmath99 is positive integer .
time integration can be replaced by parameter integration ; @xmath100 where @xmath94 is given by ( [ eq : tchi ] ) .
here we encounter an advantage of property ( b ) of our parametrization : the @xmath97 dependence of the denominator of the integrand is especially simple , @xmath101 . as in the lense - thirring case ,
the integrals are evaluated by computing the residues enclosed in the circle @xmath102 .
we find , as the second advantageous feature of the parametrization that there is only one pole , at @xmath103 .
averaging over one period is achieved by dividing the result by the period @xmath104 [ eq.([period ] ) ] .
in this section , we obtain the instantaneous losses in the energy @xmath4 , magnitude @xmath1 and spin projection @xmath105 of the orbital angular momentum .
we then find suitable angular variables such that all losses can be rewritten in terms of the true anomaly parameter @xmath97 . from the radiative power and the instantaneous total angular momentum losses of the binary system given by kidder
, we keep the newtonian and spin - orbit terms : @xmath106 \\
\label{eq : djdt } { d{\bf j}\over dt}=&-&{8\over 5}{g^2m\mu\over c^5r^3 } { \bf l}_n\bigl(-3\dot{r}^2 + 2v^2 + 2{gm\over r}\bigr ) -{4\over 5}{g^2\mu^2\over c^7r^3}\bigl\
{ -{2\over 3}{gm\over r}(\dot{r}^2-v^2)(1-\eta){\bf s } -\dot{r}{gm\over 3r^2 } { \bf r}\times({\bf v}\times{\bf s})(7 + 5\eta ) \nonumber\\ & + & { gm\over r^3}{\bf r}\times({\bf r}\times{\bf s } ) \bigl[6\dot{r}^2-{17\over 3}v^2 + 2{gm\over r } + \eta\bigl(9\dot{r}^2 - 8v^2-{2\over 3}{gm\over r}\bigr)\bigr ] \nonumber\\ & + & { \dot{r}\over r}{\bf v}\times({\bf r}\times{\bf s } ) \bigl[-30\dot{r}^2 + 24v^2+{29\over 3}{gm\over r } + 5\eta\bigl(-5\dot{r}^2 + 4v^2+{5\over 3}{gm\over r}\bigr)\bigr]\\ & + & { \bf v}\times({\bf v}\times{\bf s } ) \bigl[18\dot{r}^2 - 12v^2-{23\over 3}{gm\over r } + \eta\bigl(18\dot{r}^2-{35\over 3}v^2 - 9{gm\over r}\bigr)\bigr ] \nonumber\\ & + & { { \bf l}_n\over \mu^2r^2}({\bf l}_n\cdot { \bf s } ) \bigl[30\dot{r}^2 - 18v^2-{92\over 3}{gm\over r } + \eta\bigl(35\dot{r}^2 - 19v^2-{71\over 3}{gm\over r}\bigr)\bigr ] \bigr\ } \nonumber\end{aligned}\ ] ] we want to compute the instantaneous radiative losses of the constants of motion . as we argued in the introduction and
will show in detail later in this section , the total angular momentum loss can be taken equal to the orbital angular momentum loss , @xmath107 .
the loss in the magnitude of the orbital angular momentum is given by @xmath108 the loss in @xmath17 is given as : @xmath109 although the radiation - induced change in the spin @xmath0 is @xmath21 order smaller than the spin - orbit part of the change in the orbital angular momentum , due to the vectors @xmath110 , the second and third terms are of comparable order with the first term .
we will evaluate the terms depending on the structure of the spinning body later in this section .
the first term is found simply by multiplying ( [ eq : djdt ] ) by @xmath111 .
the next step is to find variables in which these radiative losses can be expressed solely in terms of the parameter @xmath97 and constants of the motion .
the polar coordinates @xmath112 are not suitable for this purpose : @xmath113 is not a monotonous function of time , thus the root of @xmath114 can not be extracted unambiguously .
we introduce new variables conveniently characterizing the separation vector @xmath115 , the spin vector @xmath0 and the orbital angular momentum vector @xmath2 .
the relative coordinates can be expressed in terms of the time dependent euler angles @xmath116 as : @xmath117 here @xmath118 is the polar angle and @xmath119 the azimuthal angle of the newtonian orbital momentum @xmath120 ( fig .
1 ) . the variable @xmath121 is the angle subtended by the momentary position vector and the node line ( the intersection of the the plane orthogonal to the @xmath120 with the plane orthogonal to the total angular momentum @xmath15 ) . from the leading terms of the equations of motion ( cf .
appendix of * i * ) it follows that @xmath121 is a monotonous function of @xmath122 .
thus we extract the square root of the equation for @xmath123 , choosing the positive root . in the absence of spin
, the orbital plane is at rest , and @xmath121 is the usual polar angle , satisfying the area law .
thus the time derivatives of the angles are given by @xmath124 we will not need the explicit expressions of the first - order ( in @xmath16 ) terms @xmath125 and @xmath126 for parameterizing the instantaneous losses
. rather it will be sufficient for us to find how they are interrelated .
first , by computing the square @xmath127 of the newtonian orbital angular momentum in two different ways : @xmath128 where @xmath129 are the spin terms from ( [ l2 ] ) , the following relation is found : @xmath130 second , we compute the @xmath61 component of @xmath131 in two independent ways : @xmath132 linearizing the second equality of ( [ lnz ] ) in the spin , we obtain @xmath133 when the relations ( [ eq : rel1 ] ) and ( [ eq : rel2 ] ) are inserted in the instantaneous losses of energy , magnitude and spin projection of the orbital angular momentum , all terms with @xmath134 will cancel .
the paralelogram spanned by the spin and orbital angular momentum is next described in terms of the new variables ( fig.1 ) .
we denote the polar and azimuthal angles of @xmath0 and @xmath2 by @xmath135 and @xmath136 . during the first period when @xmath15 lies on the @xmath61 axis , @xmath137 .
similarly as with @xmath2 and @xmath131 , the azimuthal and polar angles differ only by first order terms in the spin . as a result ,
no dependence on the azimuthal angles @xmath138 and @xmath139 remains in the losses . as we mentioned in the introduction ,
the spin is of @xmath21 order smaller than the orbital angular momentum .
this implies that the side @xmath16 of the paralelogram shrinks . as the angles @xmath140 and @xmath141 are present in the losses only in the spin - terms , we need only their zeroth order expressions : @xmath142 the nonradiative evolution of the angles @xmath143 and @xmath118 is discussed in appendix b. from ( [ v2 ] ) and ( [ radial ] ) @xmath28 and @xmath144 can be replaced by expressions containing solely @xmath8 and constants .
making use of the foregoing observations , we obtain the instantaneous losses of energy , magnitude and spin component of the orbital angular momentum : @xmath145 \nonumber\\ { dl\over dt } = & + & { 8g^2ml\over 5c^5\mu r^5}\left(2\mu er^2 - 3l^2\right ) \nonumber\\ & + & { 8g^2s\cos\kappa\over 15c^7\mu^2r^7 } \bigl\{12gm\mu^3er^3 + 3\mu r^2\left(6el^2 + g^2m^2\mu^3\right)-11gm\mu^2l^2 r \nonumber\\ & & \qquad\qquad\quad + \eta\left[2\mu^2e^2r^4 + 12gm\mu^3er^3 + 3\mu r^2(5el^2 + g^2m^2\mu^3)-5gm\mu^2l^2 r+15l^4\right]\bigr\ } \\ { dl_s\over dt } = & + & { 8g^2ml\cos\kappa\over 5c^5\mu r^5}\left(2\mu
er^2 - 3l^2\right ) + { 2g^2s\over 15c^7\mu^2r^7}\biggl\ { 48 gm\mu^3er^3 + 12\mu r^2\left(6 el^2 + g^2m^2\mu^3\right ) - 44 gm\mu^2l^2 r \nonumber\\ & & \qquad\quad 4 \eta\bigl[2 \mu^2e^2r^4 + 12 gm\mu^3er^3 + 3\mu r^2\left(5 el^2 + g^2m^2\mu^3\right)- 5 gm\mu^2l^2 r + 15 l^4\bigr ] \nonumber\\ & + & \sin^2\kappa \bigl\{-24 gm\mu^3er^3 + 6\mu r^2\left(6 el^2 - g^2m^2\mu^3\right ) + 72 gm\mu^2l^2 r - 90 l^4 \nonumber\\ & & \qquad\quad + \eta\bigl[-4 \mu^2e^2r^4 - 24 gm\mu^3er^3 + 6\mu r^2\left(5 el^2 - g^2m^2\mu^3\right ) + 49 gm\mu^2l^2 r - 135 l^4\bigr]\bigr\ } \nonumber\\ & - & \mu lr\dot r\sin^2\kappa\sin 2\psi \bigl\{36 \mu er^2 + 12 gm\mu^2r-18 l^2 + \eta\left(34 \mu er^2 + 12 gm\mu^2r-15l^2\right)\bigr\ } \nonumber\\ & + & \sin^2\kappa\cos 2\psi \bigl\ { 24gm\mu^3er^3 + 6\mu r^2\left(6 el^2 + g^2m^2\mu^3\right ) + 6 gm\mu^2l^2 r + 18 l^4 \nonumber\\ & & \qquad\qquad\qquad + \eta\bigl[4 \mu^2e^2r^4 + 24 gm\mu^3er^3 + 2\mu r^2(13 el^2 + 3g^2m^2\mu^3)+3 gm\mu^2l^2 r+15 l^4 \bigr]\bigr\}\biggr\ } \nonumber\\ & + & \left({{\bf l}\over s}\cdot{d{\bf s}\over dt}\right ) -\left({{\bf l}\over s}\cdot{{\bf s}\over s}\right ) \left({{\bf s}\over s}\cdot{d{\bf s}\over dt}\right ) \ .\nonumber \end{aligned}\ ] ] now we proceed to evaluate the two radiative spin - loss terms appearing in the @xmath17-loss .
we need these terms to the newtonian order and we compute them from the radiation - reaction potential .
to lowest order , this is the burke - thorne potential @xcite @xmath146 where @xmath147 is the fifth time derivative of the system s quadrupole - moment tensor and @xmath148 are coordinates centered on the spinning body .
the torque @xmath149 is the integral of the backreaction force over the volume of the body : @xmath150 the torque induces the radiative change in the spin .
the volume integral can be expressed with the tensor of inertia @xmath151 : @xmath152 for an axisymmetrically spinning body the tensor of inertia is @xmath153 in the system of principal axes with @xmath0 on the @xmath61 axis . the magnitude of the spin is @xmath154 , where @xmath155 is the angular velocity of the body .
thus @xmath156 is perpendicular to the spin @xmath0 . here
the quantity @xmath157 is the deviation from sphericity and @xmath158 the spin direction .
when the body is rapidly rotating ( @xmath159 ) , it is centrifugally flattened ( @xmath160 is nonnegligible ) and the estimate of acst @xcite for @xmath161 holds . for slow rotation , @xmath162 , the deviation from sphericity is @xmath160 is small . in both cases , the approximation above , equating @xmath24 with @xmath163 is valid . from ( [ sdirdot ] ) , the last term in the @xmath17-loss vanishes , and the other term containing the radiative spin loss is @xmath164 the losses in the energy and in the magnitude of angular momentum contain constants of the motion and they depend only on @xmath90 .
these expressions will take the suitable form for averaging , after inserting the parametrization ( [ eq : chi ] ) .
the loss @xmath165 depends also on the zeroth order expressions of the angle variable @xmath121 and of @xmath166 , which are : @xmath167 with these expressions , the task of parametrizing is completed : the loss in @xmath17 can also be expressed in terms of the true anomaly parameter @xmath97 .
the averaged losses are computed by the residue theorem , passing to the complex variable @xmath102 .
the average of the term ( [ spinlossterm ] ) vanishes , @xmath168 the averaged losses in the constants of motion are : @xmath169 \nonumber\\ \bigl<{dl\over dt}\bigr>= & -&{4g^2m(-2e\mu)^{3/2}\over 5c^5l^4}(14el^2 + 15g^2m^2\mu^3 ) \nonumber\\ & + & { g^2(-2e\mu)^{3/2}s\cos\kappa\over 15c^7l^7 } \bigl[1188e^2l^4 + 6756g^2m^2\mu^3el^2 + 5345g^4m^4\mu^6\label{averlossl } \\ & & \qquad\qquad\qquad\qquad\quad + \eta(772e^2l^4 + 4476g^2m^2\mu^3el^2 + 3665g^4m^4\mu^6)\bigr ] \nonumber\\ \bigl<{dl_s\over dt}\bigr>= & -&{4g^2m(-2e\mu)^{3/2}\cos\kappa\over 5\,c^5l^4 } ( 14el^2 + 15g^2m^2\mu^3 ) + { g^2(-2e\mu)^{3/2}s\over 15\,c^7l^7 } \nonumber\\ & \bigl\ { & \bigl[1188e^2l^4 + 6756g^2m^2\mu^3el^2 + 5345g^4m^4\mu^6 + \eta(772e^2l^4 + 4476g^2m^2\mu^3el^2 + 3665g^4m^4\mu^6)\bigr ] \nonumber \\ & -&{\sin^2\kappa\over 2 } \bigl[3516e^2l^4 + 17676g^2m^2\mu^3el^2 + 12975g^4m^4\mu^6 \nonumber\\ & & \qquad\quad + \eta(2428e^2l^4 + 12216g^2m^2\mu^3el^2 + 9125g^4m^4\mu^6)\bigr ] \label{averlossls}\\ & + & { \sin^2\kappa\cos(2\psi_0 ) } \bigl[(26el^2 + 33g^2m^2\mu^3)(12el^2 + 6g^2m^2\mu^3 ) \nonumber\\ & & \qquad\qquad\qquad\quad + \eta
( 119el^2 + 156g^2m^2\mu^3)(2el^2+g^2m^2\mu^3)\bigr]\bigr\ } \ .
\nonumber\end{aligned}\ ] ] the expressions ( [ averlosse ] ) and ( [ averlossl ] ) agree with the available results of @xcite for the averaged power and @xmath170 , despite the differences encountered in the higher - order terms of the definition of the energy and angular momentum .
the averaged loss in @xmath171 contains terms proportional to @xmath172 .
ryan@xcite argues that in most cases such terms are averaged out by the schwarzschild precession .
as in * i * , we define the generalized semimajor axis and eccentricity imposing the conditions : @xmath173 solving the system ( [ turning ] ) and ( [ rmin ] ) , one has : @xmath174 \ , \nonumber\\ 1-e^2=-{2el^2\over g^2m^2\mu^3}\bigl[1&+&{8l_ss\over c^2ml^4 } ( el^2+g^2m^2\mu^3)+{2\eta
l_ss\over c^2ml^4}(el^2 + 2g^2m^2\mu^3)\bigr]\ .\end{aligned}\ ] ] from these relations one can express the constants of motion in terms of the orbit parameters : @xmath175 inserting ( [ el2 ] ) in the equations ( [ averloss ] ) we get the averaged radiation losses in terms of @xmath176 and @xmath3 : @xmath177 \nonumber\\ \bigl<{dl\over dt}\bigr>= & -&{4g^{7/2}m^{5/2}\mu^2\over 5c^5a^{7/2}(1-e^2)^2}(7e^2 + 8 ) \nonumber\\ & + & { g^4m^2\mu^2s\cos\kappa\over 15c^7a^5(1-e^2)^{7/2 } } \bigl[549e^4 + 1428e^2 + 488+\eta ( 403e^4 + 1366e^2 + 456)\bigr ] \\
\bigl<{dl_s\over dt}\bigr>= & -&{4g^{7/2}m^{5/2}\mu^2\cos\kappa\over 5c^5a^{7/2}(1-e^2)^2}(7e^2 + 8 ) + { g^4m^2\mu^2s\over 30c^7a^5(1-e^2)^{7/2 } } \nonumber\\ & \bigl\ { & 2\bigl[549e^4 + 1428e^2 + 488 + \eta ( 403e^4 + 1366e^2 + 456)\bigr ] \nonumber\\ & -&\sin^2\kappa \bigl[1383e^4 + 4368e^2 + 1464+\eta ( 1027e^4 + 3922e^2 + 1296)\bigr ] \nonumber\\ & + & \sin^2\kappa\cos(2\psi_0 ) \,e^2\bigl[156e^2 + 240+\eta ( 119e^2 + 193)\bigr]\bigr\ } % + \bigl<{{\bf l}\over s}\cdot{d{\bf s}\over dt}\bigr > \ .\nonumber\end{aligned}\ ] ] the losses of @xmath178 and @xmath179 follow by taking the time derivatives of ( [ ls ] ) and of ( [ a1me2 ] ) : @xmath180\bigr\}\nonumber\\ \bigl<{da\over dt}\bigr>= & -&{2g^3m^2\mu\over 15c^5a^3(1-e^2)^{7/2}}(37e^4 + 292e^2 + 96 ) \\ & + & { g^{7/2}m^{3/2}\mu s\cos\kappa\over 15c^7a^{9/2}(1-e^2)^5 } \bigl[363e^6 + 3510e^4 + 7936e^2 + 2128+\eta ( 291e^6 + 4224e^4 + 7924e^2 + 1680)\bigr ] \nonumber\\ \bigl<{de\over dt}\bigr>= & -&{g^3m^2\mu\over 15c^5a^4(1-e^2)^{5/2}}\ , e ( 121e^2 + 304 ) \nonumber\\ & + & { g^{7/2}m^{3/2}\mu s\cos\kappa\over 30c^7a^{11/2}(1-e^2)^4}\ , e
\bigl[1313e^4 + 5592e^2 + 7032+\eta ( 1097e^4 + 6822e^2 + 6200)\bigr]\ .
\nonumber\end{aligned}\ ] ] from eq .
( [ aekaloss ] ) , the angle @xmath3 increases .
hence we find that the vectors @xmath181 and @xmath32 of a finite - mass binary system tend to antialign as in * i*. substituting @xmath182 for a test particle with negligible mass , we reproduce the lense - thirring losses of * i*. the evolution of the angles @xmath143 and @xmath118 under radiation backreaction is determined by @xmath170 .
this computation , relegated to appendix b , involves higher - order terms in @xmath16 which are , however of ( post)@xmath183-order .
the computation of finite - mass effects has proved to be much more difficult than the corresponding description of a test particle . we have been helped by several fortunate circumstances in overcoming the obstacles in computing the radiative losses up to the leading spin terms .
first , in this approximation , one can still decouple a radial equation of motion .
furthermore , a generalization of the true anomaly parametrization @xmath97 , found in * i * for the lense - thirring case , exists with the inclusion of the finite mass contributions . although the dynamics of the system is much more complicated , we succeeded in describing all quantities entering in the radiative losses in terms of the parameter @xmath97 .
we were able to separately describe the relative evolution of the spin and orbital angular momentum vectors , up to first order in spin , under radiation backreaction . in the loss of the spin projection of the orbital angular momentum two terms , containing the radiative spin loss ,
one of them is known to vanish ( acst ) .
we computed the other term , and have shown that it averages out to nothing .
we did not consider no - spin post - newtonian effects here .
in the equations of motion , we did not keep terms of higher order than @xmath57 , which , however , are comparable or even larger than the radiative losses . they are needed for a complete characterization of the orbit , but they do not contribute to the radiative losses in the order to which we are computing them . in the test - particle ( @xmath182 ) and
constant - spin case our results agree with those of * i*. the lense - thirring description implies @xmath184 .
these pictures arise as different limits of a more general description in which both @xmath16 and @xmath1 are arbitrary .
such a description is subject to further investigation as it will go beyond considering only the linear terms in @xmath16 .
this work has been supported by otka no .
t17176 and d23744 grants .
the algebraic package reduce was used for checking our computations .
we thank the referee for amendments on the manuscript .
in this appendix , we consider the nonradiative evolution of @xmath17 and the angle @xmath3 to linear order in the spin , and we show that these changes are smaller than order-@xmath57 . from the conservation of the total angular momentum @xmath15 and the spin precession equation ( [ eq : sprec ] ) ,
the time derivative of @xmath185 has the form : @xmath186 -\bigl({{\bf r\cdot}}\frac{{\bf s}}{s}\bigr ) \bigl({{\bf v\cdot}}\frac{{\bf s}}{s}\bigr ) \bigl(\frac{2+\eta}{r}+\frac{\eta v^2}{2gm}\bigr)\biggr\}\ .\ ] ] the multiplicative factor of @xmath16 needs to be evaluated only to newtonian order .
we write the non - radiative change in @xmath17 in terms of the radial and euler - angle coordinates ( [ eq : xyz ] ) : @xmath187 \ .\ ] ] after parameterizing by @xmath97 , the averaged expression over one period is : @xmath188 this expression is of order @xmath189 , comparable with the leading radiative loss in @xmath1 given by peters and mathews , corresponding to 2.5pn relative order .
however we pursue our description to the 1.5pn relative order .
if further terms were considered , e.g. @xmath190 , they would generate additional terms in @xmath2 , as well as in @xmath191 .
it is simple to check that these terms would be of the same order as ( [ eq : lsdot ] ) .
thus the change in @xmath17 is negligible in our approximation .
similarly , the change in @xmath192 is negligible , as follows from @xmath193 .
in this appendix , we discuss the evolution , both in the absence of radiation and under radiation backreaction , of the angles @xmath194 where @xmath195 and @xmath196 are the magnitude of the total angular momentum @xmath15 and @xmath131 , respectively .
an expansion of the expressions on the right hand sides of ( [ eq : defio ] ) to 1.5pn order yields @xmath197 by this relation , both the radiative and the nonradiative changes of @xmath198 and @xmath199 are given in terms of the corresponding changes in @xmath3 , @xmath16 and @xmath1 : @xmath200 \nonumber\\ & & -\left(\frac{s}{l}\right)^3 \biggl[\bigl(1 - 3\cos^2\kappa\bigr)\delta\cos\kappa + 3\cos\kappa\sin^2\kappa \biggl(\frac{\delta s}{s}-\frac{\delta l}{l}\biggr)\biggr]\ .\end{aligned}\ ] ] the first and second terms on the right hand side are of 1pn order and of 1.5pn order , respectively .
we have shown in appendix a that the nonradiative evolution of the angle @xmath3 is negligible in our 1.5pn approximation . because @xmath1 and @xmath16 are also constants , @xmath201 under the nonradiative evolution .
finally we turn our attention to the radiative evolution of the angles @xmath143 and @xmath118 . the spin term correction in ( [ averlossaeka ] ) to the leading order radiative loss of @xmath1 and the loss ( [ aekaloss ] ) of the angle @xmath3 are of 1.5pn order , as they originate in the spin - orbit term @xmath202 .
thus these terms do not contribute to the losses ( [ eq : deioion ] ) .
similarly the loss in the magnitude of the spin @xmath16 was seen to vanish by ( [ sdirdot ] ) .
hence the only loss contributing to the radiative changes in @xmath143 and @xmath118 is the leading term in the loss of @xmath1 : @xmath203 as the factors multiplying @xmath204 are constant during a period of revolution , the averaged rates can be simply obtained by use of the leading terms in eq .
( [ averlossl ] ) for @xmath170 . 99 l.kidder , c.will and a.wiseman , phys . rev . *
d47 * , 4183 ( 1993 ) l.kidder , phys . rev .
* d52 * , 821 ( 1995 ) l.blanchet , phys .
* d55 * , 714 ( 1997 ) l.blanchet , t.damour and b.iyer , phys .
* d51 * , 5360 ( 1995 ) b.m.barker and r.f.oconnell , gen.relativ.gravit . * 11 * , 149 ( 1979 ) k.s.thorne and j.hartle , phys .
* d31 * , 1815 ( 1985 ) t.a.apostolatos , c.cutler , g.j.sussman and k.s.thorne , * d49 * , 6274 ( 1994 ) p.c.peters and s.mathews , phys .
rev . * 131 * , 435 ( 1963 ) p.c.peters , phys .
rev . * 136 * , b1224 ( 1964 ) f.ryan , phys .
rev . * d53 * , 3064 ( 1996 ) f.ryan , phys
* d52 * , r3159 ( 1995 ) m.shibata , phys .
* d50 * , 6297 ( 1994 ) l.gergely , z.perjs and m.vasth , paper * i * , ( 1997 ) n. wex , class . quantum grav .
* 12 * , 983 ( 1995 ) r. rieth and g. schfer , class .
quantum grav .
* 14 * , 2357 ( 1997 ) we adopt the convention that the derivatives with respect to the time parameter @xmath122 occurring in the equations of motion ( sec.ii ) are denoted by an overdot but we reserve the notation @xmath205 for time derivatives due to radiation backreaction .
the gravitational backreaction will slowly rotate out the vector @xmath195 from the @xmath61 direction . in our perturbative treatment , this effect , as well as the changes in the other constants of motion are small during the period of averaging , thus neglected .
w.l.burke , j. math .
* 12 * , 401 ( 1971 ) | a convenient formalism for averaging the losses produced by gravitational radiation backreaction over one orbital period was developed in an earlier paper . in the present paper
we generalize this formalism to include the case of a closed system composed from two bodies of comparable masses , one of them having the spin @xmath0 .
we employ the equations of motion given by _ barker _ and _ oconnell _ , where terms up to linear order in the spin ( the spin - orbit interaction terms ) are kept . to obtain the radiative losses up to terms linear in the spin ,
the equations of motion are taken to the same order .
then the magnitude @xmath1 of the angular momentum @xmath2 , the angle @xmath3 subtended by @xmath0 and @xmath2 and the energy @xmath4 are conserved .
the analysis of the radial motion leads to a new parametrization of the orbit . from the instantaneous gravitational radiation losses computed by _
kidder _ the leading terms and the spin - orbit terms are taken .
following _ apostolatos , cutler , sussman _ and _ thorne _ ( acst ) , the evolution of the vectors @xmath0 and @xmath2 in the momentary plane spanned by these vectors is separated from the evolution of the plane in space .
the radiation - induced change in the spin is smaller than the leading - order spin terms in the momentary angular momentum loss .
this enables us to compute the averaged losses in the constants of motion @xmath4 , @xmath1 and @xmath5 . in the latter , the radiative spin loss terms average to zero .
an alternative description using the orbital elements @xmath6 and @xmath3 is given .
the finite mass effects contribute terms , comparable in magnitude , to the basic , test - particle spin terms in the averaged losses .
# 1@xmath7#1 epsf |
protein ubiquitin ( ub ) continues to attract the attention of researchers because there exist many processes in living systems where it plays the vital role .
usually , ub presents in the form of a polyubiquitin chain that is conjugated to other proteins .
different ub linkages lead to different biological functions . in case of lys48-c and n - c linkages polyubiquitin chain serves as a signal for degradation proteins @xcite , whereas in the lys63-c case it plays completely different functions , including dna repair , polysome stability and endocytosis @xcite .
the folding properties of ub have been studied in detail by experiments @xcite as well as by simulations @xcite .
the unfolding of ub under thermal fluctuations was investigated experimentally by cordier and grzesiek @xcite and by chung _
@xcite , and studied theoretically @xcite .
there is strong evidence that thermal unfolding pathways are reverse of the folding ones .
the mechanical unfolding of ub has been studied by several groups @xcite .
it was found that the unfolding may proceed through rare intermediates , but the overall behavior remains two - state , although a three - state scenario was also reported @xcite .
the mechanical unfolding pathways are very different from the thermal pathways , and depend on what terminal is kept fixed @xcite . recently fernandez and coworkers @xcite have applied the force - clamp technique to probe refolding of ub under quench force , @xmath1 , which is smaller than the equilibrium critical force separating the folded and unfolded states .
this kind of experiment has a number of advantages over the standard ones . in the force - clamp technique
, one can control starting conformations which are well prepared by applying the large initial force of several hundreds of pn .
moreover , since the quench force slows down the folding process , it is easier to monitor refolding pathways .
however , this begs the important question as to whether the experiments with one end of the protein anchored probes the same folding pathways as a free - end protein .
recently , using a simple go - like model , it has been shown that fixing the n - terminal of ub changes its folding pathways @xcite .
if it is so , the force - clamp technique in which the n - terminal is anchored is not useful for prediction of folding pathways of the free - end ub .
when one studies thermodynamics of a large system like multi - domain ubiquitin the problem of slow dynamics occurs , due to the rough free energy landscape .
this problem might be remedied using the standard replica exchange method in the temperature space in the absence of external force @xcite as well as in the presence of it @xcite . however ,
if one wants to construct the force - temperature phase diagram , then this approach becomes inconvenient because one has to collect data at different values of forces .
moreover , the external force increases unfolding barriers and a system may get trapped in some local minima . in order to have better sampling for a system subject to external force we propose a new replica exchange method in which the exchange is carried not in the temperature space but in the force space , i.e. the exchange between different force values .
this procedure would help the system to escape from local minima efficiently . in this paper
we address three topics .
first , we develop a new version of the replica exchange method to study thermodynamics of a large system under the force .
the basic idea is that for a given temperature we perform simulation at different values of force and the exchange between them is carried out according to the metropolis rule .
this new approach has been employed to obtain the force - temperature phase diagram of the three - domain ubiquitin . within our choice of force replicas it speeds up
computation about four times compared to the conventional simulation .
second , question we ask is under what conditions the force - clamp technique still gives the same folding pathways as for the free - end ub .
third , we determine the temperature dependence of the unfolding force , the unfolding barriers and the location of the transition state of the single ub and of the three - domain ub ( the three - domain ub will be also referred to as trimer ) .
because the study of refolding of 76-residue ub by all - atom simulations is beyond present computational facilities , the go modeling is an appropriate choice .
the go - like model @xcite was proved @xcite to give not only qualitative but also quasi - quantitative agreement with existing refolding and unfolding experiments .
therefore we will use it to study the folding and unfolding of a single and multi - domain ub . in agreement with an earlier study @xcite
, we show that fixing n - terminal of the single ub changes its folding pathways .
our new finding is that anchoring c - terminal leaves them unchanged .
more importantly , we have found that for the three - domain ub with either end fixed , each domain follows the same folding pathways as for the free - end single domain .
therefore , to probe the folding pathways of ub by the force - clamp technique one can either use the single domain with c - terminal fixed , or several domains with either end fixed . in order to check
if the effect of fixing one terminus is valid for other proteins , we have studied the titin domain i27 .
it turns out that the fixation of one end of a polypeptide chain does not change the refolding pathways of i27 .
therefore the force - clamp can always predict the refolding pathways of the single as well as multi - domain i27 .
our study suggests that the effect of the end fixation is not universal for all proteins , and the force - clamp spectroscopy should be applied with caution .
the third part of this paper was inspired by the recent pulling experiments of yang _
_ @xcite . performing the experiments in the temperature interval between 278 and 318 k
, they found that the unfolding force ( maximum force in the force - extension profile ) , @xmath2 , of ub depends on temperature linearly .
in addition , the corresponding slopes of the linear behavior have been found to be independent of pulling velocities .
an interesting question that arises is if the linear dependence of @xmath2 on @xmath3 is valid for this particular region , or it holds for the whole temperature interval . using the same go model @xcite , we can reproduce the experimental results of yang _
@xcite on the quasi - quantitative level .
more importantly , we have shown that for the entire temperature interval the dependence is not linear , because a protein is not an entropic spring in the temperature regime studied . as a part of the third topic
, we have studied the effect of multi - domain construction and linkage on the location of the transition state along the end - to - end distance reaction coordinate , @xmath4 .
it is found that the multi - domain construction has a minor effect on @xmath4 but , in agreement with the experiments @xcite , the lys48-c linkage has the strong effect on it . using the microscopic theory for unfolding dynamics @xcite , we have determined the unfolding barrier for ub . to summarize , in this article we have obtained the following novel results .
we have proposed a new replica exchange method , in which the exchange is carried out between different external forces at a given temperature .
it is shown that the force - clamp technique probes the same folding pathways as the free - end ub , if one keeps the c - terminal fixed or uses the multi - domain ub with what ever terminal anchored .
contrary to the @xmath5-protein ubiquitin , anchoring one end has a minor effect on refolding pathways of the titin domain i27 .
we attribute this to its @xmath6-sandwich structure , which is mechanically more stable .
if ub is pulled at lys48 and c then , in the bell approximation , we obtained the parameter @xmath7 0.61 nm .
this estimate agrees well with the experimental result of fernandez group @xcite .
we propose to fit simulation data with different theoretical schemes to determine the distance between the transition state and the native state , @xmath4 , for biomolecules .
our estimate for unfolding barrier of ub is in reasonable agreement with the experiments .
it has been demonstrated that the temperature dependence of the unfolding force is linear for some narrow interval , but that this dependence should be nonlinear in general .
_ three - domain model of ubiquitin .
_ since the native conformation of poly - ubiquitin is not available yet , we have to construct it for go modeling .
the native conformation of single ub , which has five @xmath6-strands ( s1 - s5 ) and one helix ( a ) , was taken from the pdb ( pi : 1ubq , fig .
[ native_structure]a ) .
we assume that residues @xmath8 and @xmath9 are in native contact if @xmath10 is less than a cutoff distance @xmath11 taken to be @xmath12 , where @xmath10 is the distance between the residues in the native conformation .
then the single ub has 99 native contacts .
we constructed the three - domain ub ( fig .
[ native_structure]b ) by translating one unit by the distance @xmath13 along the end - to - end vector , and slightly rotating it to have 9 inter - domain contacts ( about 10% of the intra - domain contacts ) . _ the go - like modeling_. the energy of the go - like model for the single as well as multi - domain ub is as follows @xcite @xmath14 + k_{\phi}^{(3 ) } [ 1 - \cos 3(\phi_i - \phi_{0i } ) ] \ } \nonumber \\ & + & \sum_{i >
j-3}^{nc } \epsilon_h \left [ 5\left ( \frac{r_{0ij}}{r_{ij } } \right)^{12 } - 6 \left ( \frac{r_{0ij}}{r_{ij}}\right)^{10}\right ] + \sum_{i > j-3}^{nnc } \epsilon_h \left(\frac{c}{r_{ij}}\right)^{12 } .
\label{hamiltonian}\end{aligned}\ ] ] here @xmath15 , @xmath16 is the distance between beads @xmath8 and @xmath17 , @xmath18 is the bond angle between bonds @xmath19 and @xmath8 , @xmath20 is the dihedral angle around the @xmath8th bond and @xmath21 is the distance between the @xmath8th and @xmath9th residues .
subscripts `` 0 '' , `` nc '' and `` nnc '' refer to the native conformation , native contacts and non - native contacts , respectively . the first harmonic term in eq . ( [ hamiltonian ] ) accounts for chain connectivity and the second term represents the bond angle potential .
the potential for the dihedral angle degrees of freedom is given by the third term in eq .
( [ hamiltonian ] ) .
the interaction energy between residues that are separated by at least 3 beads is given by 10 - 12 lennard - jones potential . a soft sphere repulsive potential ( the fourth term in eq .
[ hamiltonian ] ) disfavors the formation of non - native contacts .
we choose @xmath22 , @xmath23 , and @xmath24 , where @xmath25 is the characteristic hydrogen bond energy and @xmath26 .
since for the trimer @xmath27 ( see below ) , which is very close to @xmath28 obtained for the single ub by the same model @xcite , we will use the same energy unit @xmath29 as in our previous work .
this unit was obtained by equating the simulated value of @xmath30 to the experimental @xmath31 @xcite .
the force unit is then @xmath32 = \epsilon_h/\aa \ , = 68.0 $ ] pn .
most of our simulations were performed at temperature @xmath33 k = @xmath34 .
the dynamics of the system is obtained by integrating the following langevin equation @xcite @xmath35 where @xmath36 is the mass of a bead , @xmath37 is the friction coefficient , @xmath38 .
the random force @xmath39 is a white noise , i.e. @xmath40 .
it should be noted that the folding thermodynamics does not depend on the enviroment viscosity ( or on @xmath37 ) but the folding kinetics depends on it .
most of our simulations ( if not stated otherwise ) were performed at the friction @xmath41 , where the folding is fast . here
@xmath42 ps . the equations of motion were integrated using the velocity form of the verlet algorithm @xcite with the time step @xmath43 . in order to check the robustness of our predictions for refolding pathways ,
limited computations were carried out for the friction @xmath44 which is believed to correspond to the viscosity of water @xcite ) . in this overdamped limit
we use the euler method for integration and the time step @xmath45 . in the constant force simulations , we add an energy @xmath46 to the total energy of the system ( eq . [ hamiltonian ] ) , where @xmath47 is the end - to - end distance and @xmath48 is the force applied to the both termini or to one of them .
we define the unfolding time , @xmath49 , as the average of first passage times to reach a rod conformation .
different trajectories start from the same native conformation but , with different random number seeds . in order to get the reasonable estimate for @xmath49 , for each value of @xmath48
we have generated 30 - 50 trajectories . in order to probe folding pathways , for @xmath8-th trajectory
we introduce the progressive variable @xmath50 , where @xmath51 is the unfolding time @xcite .
then one can average the fraction of native contacts over many trajectories in a unique time window @xmath52 and monitor the folding sequencing with the help of the progressive variable @xmath53 .
in the constant velocity force simulation , we fix the n - terminal and pull the c - terminal by force @xmath54 , where @xmath55 is the displacement of the pulled atom from its original position @xcite and the spring constant @xmath56 is set to be the same as in eq .
( [ hamiltonian ] ) .
the pulling direction was chosen along the vector from fixed atom to pulled atom .
the pulling speeds are set equal @xmath57 nm / s and 4.55 @xmath58 nm / s which are about 5 - 6 orders of magnitude faster than those used in experiments @xcite .
the equilibration of long peptides at low temperatures is a computationally expensive job . in order to speed up computation of thermodynamic quantities
we extend the standard replica exchange method ( with replicas at different temperatues ) developed for spin @xcite and peptide systems @xcite to the case when the replica exchange is performed between states with different values of the external force @xmath59 .
suppose for a given temperature we have @xmath60 replicas @xmath61 , where @xmath62 denotes coordinates and velocities of residues .
then the statistical sum of the extended ensemble is @xmath63 the total distribution function has the following form @xmath64 for a markov process the detailed balance condition reads as : @xmath65 where @xmath66 is the rate of transition @xmath67 . using @xmath68 and eq .
[ markov_eq ] we obtain @xmath69}{\exp[-\beta(h_0(x_m ) - \vec{f}_m\vec{r}_m ) - \beta(h_0(x_n ) - \vec{f}_n\vec{r}_n ) ] } \ ; = \ ; \exp(-\delta),\end{aligned}\ ] ] with @xmath70 this gives us the following metropolis rule for accepting or rejecting the exchange between replicas @xmath71 and @xmath72 : @xmath73 since the three - domain ub is rather long peptide ( 228 residues ) , we apply the replica exchange method to obtain its temperature - force phase diagram .
we have performed two sets of the re simulations . in the first set we fixed @xmath74 and the re
is carried out in the standard temperature replica space @xcite , where 12 values of @xmath3 were chosen in the interval @xmath75 $ ] in such a way that the re acceptance ratio was 15 - 33% .
this procedure speeds up the equilibration of our system nearly ten - fold compared to the standard computation without the use of re . in the second
set , the re simulation was performed in the force replica space at @xmath76 using the metropolis rule given by eq .
[ metropolis_eq ] .
we have also used 12 replicas with different values of @xmath48 in the interval @xmath77 to get the acceptance ratio about 12% . even for this modest acceptance rate
our new re scheme accelerates the equilibration of the three - domain ubiquitin about four - fold .
one can expect better performance by increasing the number of replicas .
however , within our computational facilities we were restricted to parallel runs on 12 processors for 12 replicas .
the system was equilibrated during first 10@xmath78 , after which histograms for the energy , the native contacts and end - to - end distances were collected for @xmath79 . for each replica
, we have generated 25 independent trajectories for thermal averaging . using the data from two sets of the re simulations and the extended reweighting technique @xcite in the temperature and force space @xcite we obtained the @xmath80 phase diagram and the thermodynamic quantities of the trimer .
the @xmath80 phase diagram ( fig . [ diagram]a ) was obtained by monitoring the probability of being in the native state , @xmath81 , as a function of @xmath3 and @xmath48 , where @xmath81 is defined as an averaged fraction of native contacts ( see ref . for more details ) . the folding - unfolding transition ( the yellow region )
is sharp in the low temperature region , but it becomes less cooperative ( the fuzzy transition region is wider ) as @xmath3 increases . the folding temperature in the absence of force ( peak of @xmath82 or @xmath83 in fig .
[ diagram]b ) is equal @xmath27 which is a bit lower than @xmath28 of the single ub @xcite .
this reflects the fact the folding of the trimer is less cooperative compared to the monomer due to a small number of native contacts between domains .
one can ascertain this by calculating the cooperativity index , @xmath84 @xcite for the denaturation transition .
> from simulation data for @xmath83 presented in fig .
[ diagram]b , we obtain @xmath85 which is indeed lower than @xmath86 for the single ub @xcite obtained by the same go model . according to our previous estimate @xcite ,
the experimental value @xmath87 is considerably higher than the go value .
the underestimation of @xmath88 is not only a shortcoming of the off - lattice go model @xcite but also a common problem of much more sophisticated force fields in all - atom models @xcite .
rigid lattice models with side chains give better results for the cooperativity @xcite but they are less realistic than the off - lattice ones .
although the present go model does not provide the realistic estimate for cooperativity , it still mimics the experimental fact , that folding of a multi - domain protein remains cooperative observed for not only ub but also other proteins .
[ diagram]c shows the free energy as a function of native contacts at @xmath89 .
the folding / unfolding barrier is rather low ( @xmath90 1 kcal / mol ) , and is comparable with the case of single ubiquitin @xcite . the low barrier is probably an artifact of the simple go modeling .
the double minimum structure suggests that the trimer is a two - state folder .
in order to study refolding under small quenched force we follow the same protocol as in the experiments @xcite .
first , a large force ( @xmath91 pn ) is applied to both termini to prepare the initial stretched conformations .
this force is then released , but a weak quench force , @xmath1 , is applied to study the refolding process .
the refolding of a single ub was studied @xcite in the presence or absence of the quench force . fixing
the n - terminal was found to change the refolding pathways of the free - end ub @xcite , but the effect of anchoring the c - terminal has not been studied yet . here
we study this problem in detail , monitoring the time dependence of native contacts of secondary structures ( fig .
[ single_ub_pathways_fig ] ) .
since the quench force increases the folding time but leaves the folding pathways unchanged , we present only the results for @xmath92 ( fig .
[ single_ub_pathways_fig ] ) .
interestingly , the fixed c - terminal and free - end cases have the identical folding sequencing @xmath93 this is reverse of the unfolding pathway under thermal fluctuations @xcite . as discussed in detail by li _
@xcite , eq .
( [ free - end_pathways_eq ] ) partially agrees with the folding @xcite and unfolding @xcite experiments , and simulations @xcite .
our new finding here is that keeping the c - terminal fixed does not change the folding pathways .
one should keep in mind that the dominant pathway given by eq .
[ free - end_pathways_eq ] is valid in the statistical sense .
it occurs in about 52% and 58% of events for the free end and c - anchored cases ( fig .
[ single_ub_pathways_fig]d ) , respectively . the probability of observing an alternate pathway ( @xmath94 is @xmath95 % and 36 % for these two cases ( fig . [
single_ub_pathways_fig]d ) .
the difference between these two pathways is only in sequencing of s1 and s3 .
other pathways , denoted in green , are also possible but they are rather minor . in the case when the n - terminal is fixed ( fig .
[ single_ub_pathways_fig ] ) we have the following sequencing @xmath96 which is , in agreement with ref . , different from the free - end case .
we present folding pathways as the sequencing of secondary structures , making comparison with experiments easier than an approach based on the time evolution of individual contacts @xcite .
the main pathway ( eq . [ fixedn_pathways_eq ] ) occurs in @xmath97 % of events ( fig .
[ single_ub_pathways_fig]d ) , while the competing sequencing @xmath98 ( 28 % ) and other minor pathways are also possible .
> from eq .
[ free - end_pathways_eq ] and [ fixedn_pathways_eq ] it follows that the force - clamp technique can probe the folding pathways of ub if one anchores the c - terminal but not the n - one . in order to check the robustness of our prediction for refolding pathways ( eqs . [ free - end_pathways_eq ] and [ fixedn_pathways_eq ] ) , obtained for the friction @xmath99
, we have performed simulations for the water friction @xmath100 ( see ii : _ model _ above ) .
our results ( not shown ) demonstrate that although the folding time is about twenty times longer compared to the @xmath99 case , the pathways remain the same .
thus , within the framework of go modeling , the effect of the n - terminus fixation on refolding pathways of ub is not an artifact of fast dynamics , occuring for both large and small friction .
it would be very interesting to verify our prediction using more sophisticated models .
this question is left for future studies .
the time dependence of the total number of native contacts , @xmath101 , @xmath47 and the gyration radius , @xmath102 , is presented in fig .
[ q_rnc_rg_trimer_fig ] for the trimer .
the folding time @xmath103 553 ns and 936 ns for the free end and n - fixed cases , respectively .
the fact that anchoring one end slows down refolding by a factor of nearly 2 implies that diffusion - collision processes @xcite play an important role in the ub folding .
namely , as follows from the diffusion - collision model , the time required for formation contacts is inversely proportional to the diffusion coefficient , @xmath104 , of a pair of spherical units .
if one of them is idle , @xmath104 is halved and the time needed to form contacts increases accordingly .
the similar effect for unfolding was observed in our recent work @xcite .
> from the bi - exponential fitting , we obtain two time scales for collapsing ( @xmath105 ) and compaction ( @xmath106 ) where @xmath107 . for @xmath47 ,
e.g. , @xmath108 ns and @xmath109 ns if two ends are free , and @xmath110 ns and @xmath111 ns for the fixed - n case .
similar results hold for the time evolution of @xmath102 .
thus , the collapse is much faster than the barrier limited folding process . monitoring the time evolution of the end - to - end extension and of the number of native contacts , one can show ( results not shown ) that the refolding of the trimer is staircase - like as observed in the simulations @xcite and the experiments @xcite .
[ trimer_pathways_detail_fig ] shows the dependence of the number of native contacts of the secondary structures of each domain on @xmath53 for three situations : both termini are free and one or the other of them is fixed . in each of these cases
the folding pathways of three domains are identical .
interestingly , they are the same , as given by eq .
[ free - end_pathways_eq ] , regardless of we keep one end fixed or not . as evident from fig .
[ trimer_prfpw_fig ] , although the dominant pathway is the same for three cases its probabilities are different .
it is equal 68% , 44% and 43% for the c - fixed , free - end and n - fixed cases , respectively . for the last two cases ,
the competing pathway s@xmath112 s@xmath113 a @xmath114 s@xmath115 s@xmath116 s@xmath117 has a reasonably high probability of @xmath90 40% .
the irrelevance of one - end fixation for refolding pathways of a multi - domain ub may be understood as follows .
recall that applying the low quenched force to both termini does not change folding pathways of single ub @xcite .
so in the three - domain case , with the n - end of the first domain fixed , both termini of the first and second domains are effectively subjected to external force , and their pathways should remain the same as in the free - end case .
the n - terminal of the third domain is tethered to the second domain but this would have much weaker effect compared to the case when it is anchored to a surface .
thus this unit has almost free ends and its pathways remain unchanged .
overall , the `` boundary '' effect gets weaker as the number of domains becomes bigger . in order to check this argument ,
we have performed simulations for the two - domain ub .
it turns out that the sequencing is roughly the same as in fig .
[ trimer_pathways_detail_fig ] , but the common tendency is less pronounced ( results not shown ) compared to the trimer case .
thus we predict that the force - clamp technique can probe folding pathways of free ub if one uses either the single domain with the c - terminus anchored , or the multi - domain construction .
although fixing one end of the trimer does not influence folding pathways of individual secondary structures , it affects the folding sequencing of individual domains ( fig .
[ trimer_pathways_overall_fig ] ) .
we have the following sequencing @xmath118 , @xmath119 and @xmath120 for the free - end , n - terminal fixed and c - terminal fixed , respectively .
these scenarios are supported by typical snapshots shown in fig .
[ trimer_pathways_overall_fig ] .
it should be noted that the domain at the free end folds first in all of three cases in statistical sense ( also true for the two - domain case ) . as follows from the bottom of fig .
[ trimer_pathways_overall_fig ] , if two ends are free then each of them folds first in about 40 out of 100 observations .
the middle unit may fold first , but with much lower probability of about 15% .
this value remains almost unchanged when one of the ends is anchored , and the probability that the non - fixed unit folds increases to @xmath121% .
as shown by experiments @xcite and simulations @xcite , one can use the dependence of refolding times , @xmath122 , on the quenched force to find the distance between the denaturated state and transition state , @xmath123 , along the end - to - end distance reaction coordinate .
namely , in the bell approximation @xcite @xmath124 , where @xmath125 is the folding time in the absence of the external force . then from fig .
[ refold_trimer_fig ] we obtain @xmath126 nm for the three - domain ub . within
the error bars this value coincides with @xmath127 nm obtained by the same go model for the single ub @xcite , and with the experimental result @xmath128 nm @xcite .
our results suggest that the multi - domain structure leaves @xmath123 almost unchanged .
we now ask if the effect of fixing one end on refolding pathway , observed for ub , is also valid for other proteins ? to answer this question , we study the single domain i27 from the muscle protein titin . we choose this protein as a good candidate from the conceptual point of view because its @xmath6-sandwich structure ( see fig .
[ titin_str_ref_pathways_fig]a ) is very different from @xmath5-structure of ub .
moreover , because i27 is subject to mechanical stress under physiological conditions @xcite , it is instructive to study refolding from extended conformations generated by force .
there have been extensive unfolding ( see recent review @xcite for references ) and refolding @xcite studies on this system , but the effect of one - end fixation on folding sequencing of individual secondary structures have not been considered either theoretically or experimentally . as follows from fig .
[ titin_str_ref_pathways_fig]b , if two ends are free then strands a , b and e fold at nearly the same rate .
the pathways of the n - fixed and c - fixed cases are identical , and they are almost the same as in the free end case except that the strand a seems to fold after b and e. thus , keeping the n - terminus fixed has much weaker effect on the folding sequencing as compared to the single ub .
overall the effect of anchoring one terminus has a little effect on the refolding pathways of i27 , and we have the following common sequencing @xmath129 for all three cases .
the probability of observing this main pathways varies between 70 and 78% ( fig .
[ titin_str_ref_pathways_fig]e ) . the second pathway , d @xmath114 ( a , a,b , e , g ) @xmath114 ( f , c ) , has considerably lower probability .
other minor routes to the folded state are also possible .
because the multi - domain construction weakens this effect , we expect that the force - clamp spectroscopy can probe refolding pathways for a single and poly - i27 .
more importantly , our study reveals that the influence of fixation on refolding pathways may depend on the native topology of proteins .
in experiments one usually uses the bell formula @xcite @xmath131 to extract @xmath4 for two - state proteins from the force dependence of unfolding times .
[ bell_eq ] is valid if the location of the transition state does not move under external force .
recently , assuming that @xmath4 depends on external force and using the kramers theory , dudko _ et al .
_ have tried to go beyond the bell approximation and proposed the following force dependence for the unfolding time @xcite @xmath132\rbrace . \label{dudko_eq}\ ] ] here @xmath130 is the unfolding barrier and @xmath133 and 2/3 for the cusp @xcite and linear - cubic free energy surface@xcite , respectively . note that @xmath134 corresponds to the phenomenological bell theory ( eq . [ bell_eq ] ) .
one important consequence , following from eq .
[ dudko_eq ] , is that one can apply this technique to estimate not only @xmath4 but also @xmath135 for @xmath133 and 2/3 .
this will be done in this section for the single ub and the trimer .
single ub * : using the bell approximation , we have already estimated @xmath136 @xcite which is consistent with the experimental data @xmath137 @xcite . with the help of an all - atom simulation li _
@xcite have shown that @xmath4 does depend on @xmath48 . at low forces , where the bell approximation is valid @xcite , they obtained @xmath138 , which is noticeably higher than our and the experimental value .
presumably , this is due to the fact that these authors computed @xmath4 from equilibrium data , but their sampling was not good enough for such a long protein as ub .
we now use eq .
[ dudko_eq ] with @xmath139 and @xmath133 to compute @xmath4 and @xmath130 .
the regions , where the @xmath139 and @xmath133 fits works well , are wider than that for the bell scenario ( fig .
[ refold_unfold_vs_force_fig ] ) .
however these fits can not to cover the entire force interval .
the values of @xmath140 and @xmath130 obtained from the fitting procedure are listed in table 1 . in accord with ref .
all of these quantities increase with decreasing @xmath141 . in our opinion ,
the microscopic theory ( @xmath139 and @xmath133 ) gives too high a value for @xmath4 compared to its typical experimental value @xcite .
however , the latter was calculated from fitting experimental data to the bell formula , and it is not clear how much the microscopic theory would change the result . in order
to estimate the unfolding barrier of ub from the available experimental data and compare it with our theoretical estimate , we use the following formula @xmath142 where @xmath143 denotes the unfolding time in the absence of force and @xmath144 is a typical unfolding prefactor .
since @xmath144 for unfolding is not known , we use the typical value for folding @xmath145s @xcite . using @xmath146 s @xcite and eq.[unfbarrier_eq ]
we obtain @xmath147 which is in reasonable agreement with our result @xmath148 , followed from the microscopic fit with @xmath133 . using the gb / sa contimuum solvation model @xcite and the charmm27 force field @xcite li and makarov @xcite obtained a much higher value @xmath149 kcal / mol @xmath150 .
again , the large departure from the experimental result may be related to poor sampling or to the force filed they used . *
the effect of linkage on @xmath4 for single ub * one of the most interesting experimental results of carrion - vazquez _
et al._@xcite is that pulling ub at different positions changes @xmath4 drastically .
namely , if the force is applied at the c - terminal and lys48 , then in the bell approximation @xmath151 , which is about two and half times larger than the case when the termini n and c are pulled . using the all - atom model li and makarov @xcite have shown that @xmath4 is much larger than 10 .
thus , a theoretical reliable estimate for @xmath4 of lys48-c ub is not available .
our aim is to compute @xmath4 employing the present go - like model @xcite as it is successful in predicting @xmath4 for the n - c ub .
[ refold_unfold_vs_force_fig ] shows the force dependence of unfolding time of the fragment lys48-c when the force is applied to lys48 and c - terminus .
the unfolding time is defined as the averaged time to stretch this fragment . from the linear fit ( @xmath134 in fig .
[ refold_unfold_vs_force_fig ] ) at low forces we obtain @xmath152 nm which is in good agreement with the experiment @xcite . the go model is suitable for estimating @xmath4 for not only ub , but also for other proteins @xcite because the unfolding is mainly governed by the native topology .
the fact that @xmath4 for the linkage lys48-c is larger than that of the n - c ub may be understood using our recent observation @xcite that it anti - correlates with the contact order ( co ) @xcite . defining contact formation between any two amino acids ( @xmath153 ) as occuring when the distance between the centers of mass of side chains @xmath154 (see also @xmath155_@xmath156 )
, we obtain co equal 0.075 and 0.15 for the lys48-c and n - c ub , respectively .
thus , @xmath4 of the lys48-c linkage is larger than that of the n - c case because its co is smaller .
this result suggests that the anti - correlation between @xmath4 and co may hold not only when proteins are pulled at termini @xcite , but also when the force is applied to different positions .
note that the linker ( not linkage ) effect on @xmath4 has been studied for protein l @xcite .
it seems that this effect is less pronounced compared the effect caused by changing pulling direction studied here .
we have carried out the microscopic fit for @xmath157 and @xmath158 ( fig .
[ refold_unfold_vs_force_fig ] ) . as in the n - c ub case
, @xmath4 is larger than its bell value .
however the linkage at lys48 has a little effect on the activation energy @xmath130 ( table 1 ) .
determination of @xmath4 for the three - domain ubiquitin * since the trimer is a two - state folder ( fig .
[ diagram]c ) , one can determine its averaged distance between the native state and transition state , @xmath4 , along the end - to - end distance reaction coordinate using kinetic theory @xcite .
we now ask if the multi - domain structure of ub changes @xmath4 . as in the single ub case @xcite
, there exists a critical force @xmath159pn separating the low force and high force regimes ( fig .
[ refold_unfold_vs_force_fig ] ) . in the high force region , where the unfolding barrier disappears
, the unfolding time depends on @xmath48 linearly ( fitting curve not shown ) as predicted theoretically by evans and ritchie @xcite . in the bell approximation , from the linear fit ( fig .
[ refold_unfold_vs_force_fig ] ) we obtain @xmath160 0.24 nm which is exactly the same as for the single ub @xcite .
the multi - domain construction of ub does not affect @xmath4 , but this may be not the case for other proteins ( m.s .
li , unpublished results ) .
the values of @xmath140 and @xmath130 , extracted from the nonlinear fit ( fig .
[ refold_unfold_vs_force_fig ] ) , are presented in table 1 .
for both @xmath133 and @xmath139 , @xmath130 is a bit lower than that for the single ub . in the bell approximation ,
the value of @xmath4 is the same for the single and three - domain ub but it is no longer valid for the @xmath139 and @xmath133 cases
. it would be interesting to perform experiments to check this result and to see the effect of multiple domain structure on the free energy landscape . recently , using the improved temperature control technique to perform the pulling experiments for the single ub , yang _
_ @xcite have found that the unfolding force depends on @xmath3 linearly for 278 k @xmath161 318 k , and the slope of linear behavior does not depend on pulling speeds .
our goal is to see if the present go model can reproduce this result at least qualitatively , and more importantly , to check whether the linear dependence holds for the whole temperature interval where @xmath162 .
the pulling simulations have been carried at two speeds following the protocol described in _ ii : model_. fig .
[ fmax_t_fig]a shows the force - extension profile of the single ub for @xmath163 and 318 k at the pulling speed @xmath164 nm / s .
the peak is lowered as @xmath3 increases because thermal fluctuations promote the unfolding of the system .
in addition the peak moves toward a lower extension .
this fact is also understandable , because at higher @xmath3 a protein can unfold at lower extensions due to thermal fluctuations . for @xmath165
k , e.g. , the maximum force is located at the extension of @xmath166 nm , which corresponds to the plateau observed in the time dependence of the end - to - end distance under constant force @xcite .
one can show that , in agreement with chyan _
@xcite , at this maximum the extension between strands s@xmath167 and s@xmath117 is @xmath90 0.25 nm . beyond the maximum , all of the native contacts between strands s@xmath167 and s@xmath117 are broken . at this stage ,
the chain ends are almost stretched out , but the rest of the polypeptide chain remains native - like .
the temperature dependence of the unfolding force , @xmath168 , is shown in fig .
[ fmax_t_fig]b for 278 k @xmath169 318 k , and for two pulling speeds .
the experimental results of yang _
are also presented for comparison .
clearly , in agreement with experiments @xcite linear behavior is observed and the corresponding slopes do not depend on @xmath170 . using the fit @xmath171 we obtain the ratio between the simulation and experimental slopes @xmath172 .
thus , the go model gives a weaker temperature dependence compared to the experiments .
given the simplicity of this model , the agreement between theory and experiment should be considered reasonable , but it would be interesting to check if a fuller accounting of non - native contacts and environment can improve our results . as evident from fig .
[ fmax_t_fig]c , the dependence of @xmath168 on @xmath3 ceases to be linear for the whole temperature interval .
the nonlinear temperature dependence of @xmath168 may be understood qualitatively using the simple theory of evans and k. ritchie @xcite .
assuming the bell eq.[bell_eq ] for the unfolding time and a linearly ramped force @xmath173 ( see ii : _ model _ ) the unfolding force is given by @xmath174 .
a more complicated microscopic expression for @xmath168 is provided in ref . .
since @xmath175 is temperature dependent ( @xmath4 also displays a weak temperature dependence @xcite ) , the resulting dependence should be nonlinear .
this result can also be understood by noting that the temperatures considered here are low enough so that we are not in the entropic limit , where the linear dependence would be valid for the worm - like model @xcite .
the arrow in fig .
[ fmax_t_fig]c separates two regimes of the @xmath3-dependence of @xmath168 .
the crossover takes place roughly in the temperature interval where the temperature dependence of the equilibrium critical force changes the slope ( fig .
[ diagram ] ) . at low temperatures ,
thermal fluctuations are weak and the temperature dependence of @xmath168 is weaker compared to the high temperature regime .
thus the linear dependence observed in the experiments of yang _
@xcite is valid , but only in the narrow @xmath3-interval .
we have adapted the standard temperature re method to the case when the force replicas are considered at a fixed temperature .
one can extend the re method to cover both temperature and force replicas , as has been done for all - atom simulations @xcite where pressure is used instead of force .
one caveat of the force re method is that the acceptance depends on the end - to - end distance ( eq . [ delta_eq ] and [ metropolis_eq ] ) , and becomes inefficient for long proteins .
we can overcome this by increasing the number of replicas , but this will increase cpu time substantially .
thus , the question of improving the force re approach for long biomolecules remains open .
it has been clearly shown that the secondary structures have the same folding pathways for all domains , and this is probably the reason why poly - ub folds cooperatively . the folding sequencing of individual domains depends on whether one end is kept fixed or not .
we predict that the force - clamp technique gives the same folding pathways for individual secondary structures as the free - end case , if one anchores either the c - terminal of the single ub , or performs the experiments on the poly - domain ub .
it would be very interesting to verify our prediction experimentally .
another exciting challenge is to see if the force - clamp technique can probe the folding sequencing of other biomolecules using a multi - domain construction .
we have demonstrated that the anchoring of one terminal has a minor effect on refolding pathways of i27 .
this is probably due to the rigidity of its @xmath6-sandwich structure
. it would be interesting to check if this conclusion holds for other @xmath6-proteins .
the question of to what extent the force - clamp technique is useful for predicting pathways for @xmath176-proteins is left for future studies .
we have shown that , in agreement with the experiment of carrion - vazquez _ et al . _
@xcite , the lys48-c linkage changes @xmath4 drastically . from the point of view of biological function
, the linkage lys63-c is very important , but the study of its mechanical properties is not as interesting as the lys48-c because this fragment is almost stretched out in the native state .
similar to titin and rna , @xmath4 and @xmath130 are sensitive to the parameter @xmath141 , and one has to be very careful in the interpretation of experimental data . when comparing theoretical results with experiments , the same fitting procedure should be used .
the weak inter - domain interaction has a negligible effect on @xmath4 in the bell approximation , but changes become substantial in the nonlinear approximation ( @xmath157 and @xmath158 ) .
the validity of this observation for other biomolecules requires further investigation . using the microscopic fitting with @xmath133
we have obtained the reasonable value for @xmath130 for the single ub .
this result suggests that the go modeling is useful for estimating unfolding barriers of proteins .
finally , we have reproduced an experiment @xcite of the linear temperature dependence of unfolding force of ub on the quasi - quantitative level .
moreover , we have shown that for the whole temperature region the dependence of @xmath168 on @xmath3 is nonlinear , and the observed linear dependence is valid only for a narrow temperature interval .
this behavior should be common for all proteins because it reflects the fact that the entropic limit is not applicable to all temperatures .
msl thanks d.k .
klimov , g. morrison , and d. thirumalai for very useful discussions .
this work was supported by the polish kbn grant no 1p03b01827 , the national science council ( taipei ) under grant no .
nsc 96 - 2911-m 001 - 003-my3 , academia sinica ( taipei ) under grant no .
as-95-tp - a07 , and national center for theoretical sciences in taiwan .
@xmath177 & 1/2 & 2/3 & 1 & 1/2 & 2/3 & 1 & 1/2 & 2/3 & 1 + @xmath178&13200&1289&9.1&4627&2304&157&1814&756&47 + @xmath179 & 7.92&5.86&2.4&12.35&10.59&6.1&6.21&5.09&2.4 + @xmath180&17.39&14.22&-&15.90&13.94&-&13.49&11.64&- + table 1 .
dependence of @xmath4 on fitting procedures for the three - domain ub and lys48-c .
@xmath134 corresponds to the phenomenological bell approximation ( eq . [ bell_eq ] ) . @xmath133 and 2/3 refer to the microscopic theory ( eq . [ dudko_eq ] ) . for comparison
we show also the data for the single ub which are taken from our previous work @xcite . for the single and three - domain ub the force
is applied to both termini . * figure 1 . * ( a ) native state conformation of ub taken from the pdb ( pdb i d : 1ubq ) .
there are five @xmath6-strands : s1 ( 2 - 6 ) , s2 ( 12 - 16 ) , s3 ( 41 - 45 ) , s4 ( 48 - 49 ) and s5 ( 65 - 71 ) , and helix a ( 23 - 34 ) .
( b ) the native conformation of the three - domain ub was designed as described in _ ii : model_. there are 18 inter- and 297 intra - domain native contacts .
5 mm * figure 2 . * ( a ) the @xmath80 phase diagram obtained by the extended replica exchange and histogram method .
the force is applied to termini n and c. the color code for @xmath181 is given on the right .
blue corresponds to the folded state , while red indicates the unfolded state .
the vertical dashed line denotes to @xmath182 k , at which most of simulations have been performed .
( b ) temperature dependence of the specific heat @xmath183 ( right axis ) and @xmath83 ( left axis ) at @xmath74 .
their peaks coincide at @xmath89 .
( c ) the dependence of the free energy of the trimer on the total number of native contacts @xmath101 at @xmath89 .
* figure 3 . *
the dependence of native contacts of @xmath6-strands and the helix a on the progressive variable @xmath53 when the n - terminal is fixed ( a ) , both ends are free ( b ) , and c - terminal is fixed ( c ) .
the results are averaged over 200 trajectories .
( d ) the probability of refolding pathways in three cases .
each value is written on top of the histograms . * figure 4 . * ( a ) the time dependence of @xmath101 , @xmath47 and @xmath102 at @xmath33 k for the free end case .
( b ) the same as in ( a ) but for the n - fixed case .
the red line is a bi - exponential fit @xmath184 .
results for the c - fixed case are similar to the @xmath185-fixed case , and are not shown .
* figure 5 . * the same as in fig . 3 but for the trimer .
the numbers 1 , 2 and 3 refer to the first , second and third domain .
the last row represents the results averaged over three domains .
the fractions of native contacts of each secondary structure are averaged over 100 trajectories .
* figure 7 . * the dependence of the total number of native contacts on @xmath53 for the first ( green ) , second ( red ) and third ( blue ) domains .
typical snapshots of the initial , middle and final conformations for three cases when both two ends are free or one of them is fixed .
the effect of anchoring one terminus on the folding sequencing of domains is clearly evident . in the bottom
we show the probability of refolding pathways for three cases .
its value is written on the top of histograms .
* figure 8 . * the dependence of folding times of the trimer on @xmath1 at @xmath33 k. @xmath186 was computed as the median first passage times of 30 - 50 trajectories for each value of @xmath1 .
> from the linear fit @xmath187 , we obtained the average distance between the unfolded and transition states @xmath126 nm . * figure 9 . * ( a ) native state conformation of ig27 domain of titin(pdb i d : 1tit ) .
there are 8 @xmath6-strands : a ( 4 - 7 ) , a ( 11 - 15 ) , b ( 18 - 25 ) , c ( 32 - 36 ) , d ( 47 - 52 ) , e ( 55 - 61 ) , f(69 - 75 ) and g ( 78 - 88 ) .
the dependence of native contacts of different @xmath6-strands on the progressive variable @xmath53 for the case when two ends are free ( b ) , the n - terminus is fixed ( c ) and the c - terminal is anchored ( d ) .
( e ) the probability of observing refolding pathways for three cases .
each value is written on top of the histograms .
* figure 10 . *
the semi - log plot for the force dependence of unfolding times at @xmath33 k. crosses and squares refer the the single ub and the trimer with the force applied to n- and c - terminal , respectively .
circles refer to the single ub with the force applied to lys48 and c - terminal . depending on @xmath48 , 30 - 50 folding events were using for averaging . in the bell approximation ,
if the n- and c - terminal of the trimer are pulled then we have the linear fit @xmath188 ( black line ) and the distance between the native and transition states , @xmath7 0.24 nm .
the same value of @xmath4 was obtained for the single ub @xcite . in the case when we pull at lys48 and c - terminal of single ub the linear fit ( black line ) at low forces is @xmath189 and @xmath190 nm . the correlation level of fitting is about 0.99 .
the red and blue curves correspond to the fits with @xmath157 and @xmath158 , respectively ( eq . [ dudko_eq ] ) .
* figure 11 . * ( a ) the force - extension profile obtained at @xmath33 k ( black ) and 318 k ( red ) at the pulling speed @xmath164 nm / s .
@xmath168 is located at the extension @xmath191 nm and 0.6 nm for @xmath33 k and 318 k , respectively .
the results are averaged over 50 independent trajectories .
( b ) the dependence of @xmath192 on temperature for two values of @xmath141 .
the experimental data are taken from ref .
for comparison .
the linear fits for the simulations are @xmath193 and @xmath194 . for the experimental sets we have @xmath195 and @xmath196 .
( c ) the dependence temperature of @xmath192 for the whole temperature region and two values of @xmath141 .
the arrow marks the crossover between two nearly linear regimes . | we have developed a new extended replica exchange method to study thermodynamics of a system in the presence of external force .
our idea is based on the exchange between different force replicas to accelerate the equilibrium process .
this new approach was applied to obtain the force - temperature phase diagram and other thermodynamical quantities of the three - domain ubiquitin . using the c@xmath0-go model and the langevin dynamics
, we have shown that the refolding pathways of single ubiquitin depend on which terminus is fixed .
if the n - end is fixed then the folding pathways are different compared to the case when both termini are free , but fixing the c - terminal does not change them .
surprisingly , we have found that the anchoring terminal does not affect the pathways of individual secondary structures of three - domain ubiquitin , indicating the important role of the multi - domain construction .
therefore , force - clamp experiments , in which one end of a protein is kept fixed , can probe the refolding pathways of a single free - end ubiquitin if one uses either the poly - ubiquitin or a single domain with the c - terminus anchored .
however , it is shown that anchoring one end does not affect refolding pathways of the titin domain i27 , and the force - clamp spectroscopy is always capable to predict folding sequencing of this protein .
we have obtained the reasonable estimate for unfolding barrier of ubiqutin , using the microscopic theory for the dependence of unfolding time on the external force .
the linkage between residue lys48 and the c - terminal of ubiquitin is found to have the dramatic effect on the location of the transition state along the end - to - end distance reaction coordinate , but the multi - domain construction leaves the transition state almost unchanged .
we have found that the maximum force in the force - extension profile from constant velocity force pulling simulations depends on temperature nonlinearly .
however , for some narrow temperature interval this dependence becomes linear , as have been observed in recent experiments . |
hydrodynamic turbulence remains a basically unsolved problem of modern physics .
this is especially noticeable as the fundamentals seem to be fairly easy the navier - stokes - equations for the velocity and pressure fields @xmath0 and @xmath1 , @xmath2 with the additional constraint @xmath3 express the conservation of momentum in a classical , newtonian , incompressible fluid of viscosity @xmath4 and density @xmath5 .
laminar flows are reproduced very accurately ; in the turbulent regime , it still is an open question how the universal characteristics of a flow , the scaling exponents @xmath6 of the structure functions @xmath7 of order @xmath1 , defined by @xmath8 can be extracted from the basic equations .
intermittency is reflected by exponents @xmath6 that differ from those expected by dimensional analysis .
+ monte - carlo - simulations enable us to analyze turbulent flow patterns in detail , to gain direct insight into the formation of localized structures and their behaviour , and to measure observables like the scaling exponents .
+ instead of working with the full navier - stokes - eqs . , we decided to elaborate the methods using burgers eq . @xmath9 which can be interpreted as the flow equation for a fully compressible fluid . a finite viscosity @xmath4 and energy dissipation @xmath10
provide a dissipation length scale @xmath11 corresponding to the kolmogorov - scale in navier - stokes - turbulence : @xmath12 besides being of interest of its own ( e.g. in cosmology ) , working with eq .
( [ dh : eqburg ] ) has a number of technical advantages : * burgers eq .
is local , while the incompressibility condition acts as a nonlocal interaction in navier - stokes - turbulence . * the fundamental solutions to burgers eq .
are well - known ; in the limit of vanishing viscosity ( hopf - eq . ) , these form singular shocks as seen in fig .
( [ dh : fig1 ] ) .
the dissipation - scale provides an uv - regularization of the shock structures .
+ as a function of @xmath13 in periodic boundary conditions at constant @xmath14.,width=216 ] * a huge variety of analytical methods have been applied to burgers eq . , giving results that can directly be compared to numerical measurements ; and the origin of intermittency is well understood .
for the moment , we restrict our work to 1-dimensional burgulence which already shows intermittent statistics . a random force
@xmath15 has to be introduced into eq .
( [ dh : eqburg ] ) so that solutions are statistically homogenous in time : @xmath16\doteq \partial_{t}v + v \frac{\partial}{\partial x}v - \nu \frac{\partial^2}{\partial x^2 } v = f,\ ] ] where we model the stochastic force to be gaussian distributed with zero mean , energy injection rate @xmath10 and correlation length @xmath17 : @xmath18 we then expect intermittent statistics to be found within the inertial subrange @xmath19 .
we can further identify a characteristic velocity at the injection scale , @xmath20 , and a reynolds - number @xmath21 the path integral is introduced via the standard martin - siggia - rose - formalism , giving for the generating functional @xmath22 & = & \int \mathcal{d}v \ , \mathcal{d}f \ , \delta(\mathbf{b}[v ] - f )
\exp \left(-\frac{1}{2 } \int f \chi f + \int jv\right ) \\ & = & \int \mathcal{d}v \
, e^{-s[v ; j]},\end{aligned}\ ] ] with the action @xmath23 = \int dx dt dx ' dt ' \ , \left(\mathbf{b}[v(x , t)]\chi(x , x ' ; t , t ' ) \mathbf{b}[v(x',t ' ) ] + j(x , t)v(x , t)\right).\ ] ] beginning from an equivalent path integral , it has been shown that intermittent statistics of burgers eq . can be understood in terms of instanton solutions ( @xcite ) .
we discretized the above action onto a rectangular lattice with @xmath24 sites in space- , and @xmath25 sites in time - direction .
derivatives have been written in a symmetric ( stratanovich- ) prescription .
lattice spacings will be denoted @xmath26 and @xmath27 , respectively .
we mainly used a heat bath algorithm on single nodes . to indentify the lattice parameters with the constants of the continuum theory , we first notice that the viscosity has to be defined as @xmath28 in which we define the arbitrary constant @xmath29
the so - defined @xmath4 gives us @xmath30 according to eq .
( [ dh : egre ] ) .
we further find that the dissipation length @xmath11 is related to the correlation length @xmath17 simply by @xmath31 in any practical application , it is sufficient to define the @xmath32 and @xmath30 , to chose @xmath24 and @xmath17 , and to calculate from that @xmath25 and @xmath10 . stability considerations lead to further constraints for the lattice size , as will be explained in the following subsection .
as would be expected , the stability of the simulations over a large number of monte - carlo - steps is a big issue , due to the shock - like solutions of eq .
( [ dh : eqburg ] ) . indeed , if certain restrictions to the lattice parameters are not taken care of rigorously , the simulation terminates sooner or later due to divergencies . + it is interesting to notice that the occurence of instabilities in our mc - simulations is related to the ( non - trivial ) existence of the dissipation scale @xmath11 .
we found that to obtain stable simulations , @xmath11 has to be resolved on the lattice : @xmath33 .
unstable simultations occur otherwise we observed that the overall energy of the configurations accumulate in the smallest scale @xmath26 , causing the configuration to separate into two sub - configurations , of which one looks as the expected solution to the hopf - eq . , while the other grows beyond any limit , eventually breaking the simulation . +
as long as the dissipation length is resolved , the simulations are stable . having performed several millions of mc - steps , no further instabilities occured .
if length scales are measured in units of the system size , this translates into a constraint involving the reynolds - number : @xmath34 which enforces huge lattices for high @xmath30 as @xmath35 .
we simulated systems of different sizes , from @xmath36 to larger lattices of the same viscosity , as @xmath37 and @xmath38 , and also different viscosities , like @xmath39 or @xmath40 .
the reynolds - number ranged from @xmath41 to @xmath42 , not respecting the above conditions for stability .
this may seem surprising , but we could see that interesting information could also be extracted from the physical sub - configurations .
stable runs have been obtained for @xmath43 .
+ long runs of several millions of mc - steps , on a single node , are realistic for small lattices like @xmath44 or @xmath37 only . for larger lattices , a parallelized version of the code will be needed .
+ we obtained the following results : * thermalization and autocorrelation times are very long , up to the order of @xmath45 mc - steps . * in the stable runs , after thermalization , the typical shock solutions of the burgers eq . form and can be observed moving and interacting with each other , see fig .
( [ dh : fig2 ] ) . + of a @xmath46-lattice ; the plot shows a slize of constant @xmath14 , periodic boundary conditions in @xmath13 .
two shock - like structures are clearly visible-,width=216 ] * in the unstable runs before occurence of the instability , configurations resemble the kink - solutions of the hopf - eq .. * the disctinction between stable and unstable simulations can directly be related to the existence of a dissipation length scale which is either bigger ( stable ) or smaller ( unstable simulations ) than the lattice spacing .
we have shown how to perform stable mc - simulations of stochastic partial differential eq . , like burgers or the hopf - eq .. the lattice versions of the theories can directly be identified with their continuum counterparts , and , as long as certain constraints on the lattice size are respected , unlimited numbers of configurations produced . direct insight into the structures leading to intermittency and ,
thus , multiscaling , can be obtained .
especially , we want to point out that the existence of a dissipation length scale can be observed . + as next steps , complete statistics will be made ; especially the scaling exponents of the structure functions have to be compared with the analytic results .
later , we will proceed by analyzing the incompressible navier - stokes - eqs ..
we especially want to thank gernot mnster , westflische wilhelms - universitt mnster , for his support in numerous discussions .
99999999 e. balkovsky , g. falkovich , i. kolokolov , and v. lebedev , _ phys .
lett _ * 78 * , 1452 ( 1997 ) j. bec , k. khanin , _ arxiv:0704.1611v1 [ nlin.cd ] _ ( 2007 ) d. hochberg , c. molina - pars , j. prez - mercader , and m. visser , _ phys.rev .
e _ * 60 * , 6343 - 6360 ( 1999 ) | we investigate the behaviour of stochastic differential equations , especially burgers eq . , by means of monte - carlo - techniques . by analysis of the produced configurations ,
we show that direct and often intuitive insight into the fundamentals of the solutions to the underlying equation , like shock wave formation , intermittency and chaotic dynamics , can be obtained .
we also demonstrate that very natural constraints for the lattice parameters are sufficient to ensure stable calculations for unlimited numbers of monte - carlo - steps . |
the rieman zeta function is @xmath0 where @xmath1 , @xmath2 , is a complex number .
it is known that the zeta function can be continued analytically to the whole complex plane except for @xmath3 where the function has a pole .
the zeta function has trivial zeroes at even negative integers .
it does not have zeroes for @xmath4 and the only zeroes for @xmath5 are the trivial ones .
the nontrivial zeroes lie in the strip @xmath6 , see e.g. [ 1 ] .
let @xmath7 be the set of all primes larger than one .
let @xmath1 , @xmath2 and @xmath8 .
we define @xmath9 by @xmath10 as @xmath11 does not have zeroes for @xmath12 and @xmath13 is analytic in the whole complex plane except for @xmath3 , the function @xmath9 is analytic when @xmath8 except for at @xmath3 .
if @xmath14 the function @xmath9 has the absolutely convergent series expression @xmath15 where the numbers @xmath16 are the primes in ( 2 ) .
we generalize the equation @xmath17 in lemma 1 .
let @xmath18 , @xmath19 , @xmath1 , @xmath2 , @xmath14 , and @xmath20 .
we define absolutely convergent products @xmath21 @xmath22 @xmath23 then @xmath24 @xmath25 @xmath26 the functions ( 6 ) and ( 7 ) are given in a form where the highest orders of @xmath27 are directly seen . simplifying ( 6 ) @xmath28 from ( 7 ) @xmath29 directly multiplying the absolutely convergent products @xmath30 @xmath31 at @xmath32 the term @xmath33 and if @xmath34 then @xmath35 . the other claims follow .
lemma 1 is proved .
@xmath36 lemma 2 .
let @xmath18 , @xmath19 , @xmath1 , @xmath2 , and @xmath8 .
there exists @xmath37 such that if @xmath38 then the following claims hold : the product ( 5 ) is absolutely convergent for @xmath8 .
@xmath39 defined by ( 5 ) is finite and nonzero and @xmath40 is finite and nonzero at @xmath32 .
the product ( 7 ) has an analytic continuation for @xmath8 and @xmath41 is finite at @xmath32 .
the product ( 6 ) has an analytic continuation for @xmath8 in all points except for an @xmath42-neighborhood of any zero of @xmath13 .
let @xmath37 be chosen smaller than @xmath43 and @xmath38 .
then @xmath44 .
this means that the product ( 5 ) is absolutely convergent .
as it is an absolutely convergent infinite product , it does not converge to zero .
thus , @xmath39 and @xmath40 are analytic and therefore finite for @xmath45 . then also @xmath46
. we can take the logarithm of @xmath41 when @xmath47 since the function is not zero . in order to analytically continue ( 7 ) ,
let us write @xmath48 where @xmath49 @xmath50 @xmath51 the sum @xmath52 is obtained by expanding the logarithm into a series .
the sum @xmath52 converges absolutely for @xmath53 and the analytic continuation of @xmath54 yields the analytic continuation for @xmath41 for @xmath8 , @xmath38 .
this is shown as follows .
the first term to the right in @xmath52 converges absolutely since it has the power @xmath55 for @xmath16 .
the two other terms converge absolutely because for any @xmath56 , @xmath57 , holds @xmath58 inserting @xmath59 and @xmath60 shows absolute convergence of the terms for @xmath53 .
the equation ( 8) gives an analytic continuation for @xmath61 : @xmath62 is analytic when @xmath41 and @xmath39 are analytic and nonzero .
@xmath39 is nonzero when @xmath8 and @xmath38 .
however , @xmath41 can be zero .
we have already shown that @xmath41 differs from @xmath54 only by a nonzero function .
let @xmath63 be a zero of @xmath13 .
the zeroes of @xmath13 are isolated , if @xmath42 is selected sufficiently small , the points when @xmath64 are in an @xmath42-neighborhood of @xmath63 .
lemma 2 is proved .
@xmath36 notice that in a neighborhood of a zero @xmath63 of @xmath13 the function @xmath61 is singular always when @xmath64 .
@xmath61 is not continuous at @xmath65 because @xmath66 but @xmath67 is unbounded for some arbitrarily small @xmath68 .
thus , if a zero @xmath63 of @xmath13 exists , @xmath61 does not behave nicely .
this does not mean that @xmath61 actually has such discontinuities since we can not assume to know that a zero of @xmath13 exists . let us define the finite products @xmath69 @xmath70 and the infinite products that are absolutely convergent when @xmath47 , @xmath71 : @xmath72 for any @xmath73 holds @xmath74 lemma 3 . if @xmath75 for @xmath76 , @xmath77 , @xmath78 , then @xmath79 and @xmath80 for every @xmath81 .
let @xmath63 be a point where @xmath75 and @xmath82 . in the equation
@xmath83 @xmath84 and @xmath85 are finite products
. they are nonzero at @xmath65 and continuous at @xmath32 as functions of @xmath68 .
@xmath86 is zero at @xmath65 . if we can vary @xmath68 slightly while keeping @xmath87 at @xmath63 and find such an @xmath68 , @xmath38 ,
that @xmath88 then the product in ( 9 ) is not zero , i.e. @xmath89 . thus , it is not possible to vary @xmath68 in this way .
it follows that all partial derivatives of @xmath86 with respect to @xmath68 vanish at @xmath65 : @xmath90 lemma 3 is proved .
let @xmath75 for @xmath63 such that @xmath91 .
let us assume that @xmath92 for every @xmath93 . then for @xmath81 @xmath94 let @xmath18 , @xmath19 , @xmath1 , @xmath2 , and @xmath8
let @xmath37 be chosen smaller than @xmath43 and @xmath38 . then @xmath44 .
we also assume that @xmath37 is chosen so small that @xmath61 and @xmath41 have the analytic continuations from lemma 2 .
let @xmath95 , @xmath96 , be any continuously differentiable functions of @xmath87 and @xmath68 in an open neighborhood of @xmath97 and let the infinite product @xmath98 be absolutely convergent , then @xmath99 @xmath100 for @xmath41 as in ( 7 ) and @xmath101 we get from ( 10 ) @xmath102 let us write @xmath103 and @xmath104 @xmath105 then @xmath106 is absolutely convergent if @xmath53 and @xmath107 is absolutely convergent if @xmath101 . for @xmath108
holds @xmath109 we can continue @xmath107 analytically to @xmath110 by ( 11 ) , except to points where @xmath111 .
thus , if @xmath112 and @xmath38 , then holds @xmath113 likewise , let @xmath101 and @xmath114 .
by ( 10 ) , for @xmath108 holds @xmath115 where @xmath116 @xmath117 the function @xmath118 converges absolutely if @xmath8 while @xmath119 converges absolutely if @xmath14 .
we can continue @xmath119 analytically to @xmath110 by ( 12 ) , except to points where @xmath120 .
thus , if @xmath112 and @xmath38 , then holds @xmath121 we notice that @xmath122 .
the function @xmath119 has a pole at @xmath63 but @xmath41 is differentiable at @xmath123 , so we can define the value of the product @xmath124 at @xmath125 by first cancelling the pole of @xmath119 by the zero of @xmath13 at @xmath125 .
we denote the function @xmath124 at @xmath125 by @xmath126 . in a similar way we can define @xmath127 at @xmath123 . at @xmath65
we get @xmath128 since @xmath106 is analytic and therefore finite at @xmath65 and @xmath129 . by the assumption
@xmath130 then @xmath131 by the assumption @xmath132 for @xmath133 .
the convergent part @xmath106 and all its partial derivatives are finite at @xmath65 and vanish when multiplied by a zero .
thus @xmath134 the convergent part @xmath118 and all its derivatives are finite at @xmath135 and vanish when multiplied by a zero .
thus , if @xmath136 is shown by induction for @xmath137 , then @xmath138 let us assume @xmath136 is shown by induction for @xmath139
. then @xmath140 @xmath141 by the assumption .
lemma 4 is proved .
let us assume that the following condition holds : if @xmath75 for @xmath63 such that @xmath91 , then @xmath142 for @xmath81
. under this condition holds : if @xmath1 , @xmath2 , @xmath6 , and @xmath120 then @xmath143 .
riemann showed that @xmath144 thus , if there exists a zero @xmath76 of @xmath13 with @xmath145 then there exists a zero of @xmath13 at a symmetric point in @xmath146 .
therefore we only need to look at the strip @xmath147 .
let us assume there exists @xmath76 with @xmath75 and @xmath148 . by the assumption @xmath149 for every @xmath81 .
as @xmath13 is analytic if @xmath150 , it has a converging taylor series in some neighborhood of every point @xmath150 .
we can analytically continue @xmath13 along the @xmath151 line starting from @xmath152 .
since @xmath75 and all derivatives with respect to @xmath87 vanish , @xmath120 for every point @xmath87 to which the function is continued .
thus @xmath153 for some @xmath14 which is a contradiction .
therefore no such @xmath63 exists .
lemma 5 proves the riemann hypothesis if we can show the following conjecture .
conjecture 1 .
if @xmath75 for @xmath76 , @xmath77 , @xmath78 , then @xmath154 and @xmath92 for every @xmath81 . _
argument : _ let us assume that @xmath63 is a zero of @xmath13 and @xmath155 .
the proof of lemma 3 works for @xmath41 in the same way as for @xmath86 if we show that @xmath61 is continuous at @xmath123 .
in fact , if the proof of lemma 3 fails for @xmath41 , then @xmath66 and @xmath156 for an arbitrarily small @xmath157 .
let @xmath158 be very large .
if @xmath159 , then @xmath160 .
@xmath9 is analytic and thus bounded for @xmath161
. we can select @xmath125 and consider @xmath67 as a function of @xmath68 .
if @xmath158 is large , the function is analytic .
it can be continued to smaller values of @xmath158 , except to points when it becomes infinite , by analytic continuation .
especially , the analytic continuation is continuous .
thus , @xmath67 is continuous at @xmath32 since @xmath162 is not infinite .
the argument is finished .
if the argument for conjecture 1 is valid , then the riemann hypothesis holds . the proof is elementary and uses known methods , though functions of several complex variables may not be common in this context .
the proof is , however , very much in line with the original thoughts that the riemann hypothesis is a problem of complex analysis and not of number theory , and that the hypothesis is a consequence of the convergence of the harmonic series , not of deep , and to us unknown , properties of prime numbers . just as the pole of @xmath13 at @xmath3 and nonexistence of zeroes or poles for @xmath47 is a consequence of the harmonic series , in conjecture 1 , the zeroes at @xmath163 and the riemann hypothesis are a consequence of convergence of @xmath39 .
it is not correct to claim that the proof only uses the increasing property of primes .
the regions of convergence are those of the harmonic series .
the proof in fact uses all information we know of all primes : that their products make up all integers and that the harmonic series converges on a certain area . | the paper formulates a condition such that if the condition holds the riemann hypothesis is true .
# 1#2 # 1 # 2 .
# 3 riemann zeta function , riemann hypothesis , complex analysis . |
the photometric works on globular clusters ( gcs ) , very often , do not include the @xmath12 band observations .
the @xmath12band , however , is very important since it provides a comparatively long base colour index @xmath6 , which removes the degeneracy in hb colours for @xmath13 and is also the most sensitive temperature parameter for early type stars in the @xmath0 system ( buser & kurucz 1978 ) .
this colour index is furthermore useful to obtain stellar atmospheric parameters using theoretical models ( e.g. kurucz 1992 ) .
taking in mind the good resolution in the rc
focus ( f/8 , scale 12.86 arcsecmm@xmath14 ) of the 2-m telescope of the natl .
rozhen , some years ago , we started a program to study the central parts of some globular clusters . here
are presented the results concerning m5@xmath1ngc5904 . in this case , our aim was to produce @xmath0 magnitudes for a large , statistically significant sample of stars at the level of the horizontal branch and above and to investigate some puzzling cmd features .
we draw special attention to stars occupying the blue horizontal branch ( bhb ) .
the distribution of these stars poses a number of questions that are still open and need further investigation as e.g. : the presence of a gap in the bhb ; the ultraviolet deficiency of some bhb stars ; the real location of the bluest hb stars on the @xmath15 diagram . in the following
we present the observational material and reduction procedures ( sect.2 ) .
the @xmath3diagrams are analysed in sect .
the main topics in this section are the presence of a gap , the @xmath3-bright star candidates and the discrepancy between the observed bhb and the zero - age horizontal branch ( zahb ) model .
the two colour diagram features are discussed in sect .
this includes the observed ultraviolet deficiency of bhb stars and the ultraviolet excess of the rgb stars .
an assessment of the cluster metallicity and reddening is also given .
the results obtained are summarised and discussed in sect .
the present investigation of m5 is based on photographic plates and ccd frames taken of the central part of the cluster . to achieve better resolution of stellar images
all observations were obtained with short - time exposures .
the average seeing , measured at fwhm of the stellar image , is @xmath16 arcsec . on each plate - 5@xmath12,6@xmath17 and 5@xmath18 - nine overlaping regions , covering up to 5.5 arcmin away from the cluster center ,
were scanned with a joyce loebl microdensitometer .
the data were smoothed by means of a wavelet transform and subsequently converted into relative intensities using calibration curves specific for each plate ( for details see markov 1994 ) .
the digital images were analysed with originally developed reduction procedures . to check the reliability of these procedures part of the observations
were processed via daophot ( stetson 1987 ) a detailed comparison of the results is given in markov et al .
the efficiency of the wavelet transform in localising stellar images in crowded fields is described in borissova et al .
a similar approach had been previously employed by aurire & coupinot ( 1989 ) to obtain improved photometry for stars in dense globular clusters fields . in order to obtain more precise photometric calibration , supplementary
3@xmath18 and 3@xmath17 ccd observations of the m5 core were taken at march 16 , 1997 with a sbig st-6 camera ( pixelsize = 0.32 arcsec ) .
seeing conditions for the night were stable with fwhm@xmath19 . the standard field in m92 of christian et al .
( 1985 ) was observed to calibrate the program images .
the transformation of the instrumental magnitudes to the standard @xmath12 system was performed by means of 24 stars from the arp s ( 1962 ) photoelectric standard sequence in the field of m5 .
this sample covers wide range of colours : @xmath20 , @xmath21 and @xmath22 .
the bhb stars are of special interest in our investigation and separate transformations to the standard system have been employed for them . in this colour region 7
arp s photoelectric and 3 secondary photometric standard stars , taken from richer and fahlman ( 1987 ) , were used .
the calibration into @xmath23 standard system was performed in two steps .
first , we used the ccd observations to calibrate @xmath24 instrumental magnitudes with respect to the primary standard stars in m92 ( christian et al .
then , 86 stars common for photographic and ccd measurements were used as a local secondary standard to calibrate all other program stars .
the mean internal errors are in the range of @xmath25 . the total error including the uncertainties in the photoelectric data is @xmath26 in the stellar magnitudes and @xmath27 in the colour indices .
our photometry can be compared with earlier works e.g. : 1 . _ the ubv ccd photometry of von braun et al .
( 1998)_. for 18 stars in common , the mean differences between our data and those of von braun et al .
( 1998 ) are : @xmath28 ; @xmath29 and @xmath30 .
obviously , the two magnitude sets match each other within the error
. a systematic difference of about 0.06 mag was found for the u magnitudes of 5 agb stars ( in the sense von braun s data being brighter ) .
these stars were not included in the above assessments .
the photographic photometry of bcf_. the mean difference in the zero point ( @xmath31 ) , based on 182 common stars , is 0.01 mag .
there seems to be a small colour difference of 0.03 mag for bhb stars ( our magnitudes being redder ) .
_ the cmd mean lines of sandquist et al .
( 1996)_. figure 1 shows the @xmath32 diagram on which the sandquist et al .
( 1996 ) fiducial lines are overlaid for comparison .
figure 2a shows the @xmath33 diagram for m5 .
two features of this diagram are worth noting : the bhb in this plane is not as steep as that in @xmath32 ( fig .
1 ) ; the bhb stars form an almost linear sequence with a discontinuity at @xmath34 and @xmath35 .
the latter is illustrated in the histogram ( fig .
2a ) . to test statistical significance of this bhb feature , the @xmath36 co - ordinate deduced from ferraro et al .
( 1992 ) and dixon et al .
( 1996 ) is used .
@xmath36 is a measure of the star position along the hb in the @xmath37 diagram as marked by its projection on the hb ridgeline .
the @xmath36 parameter grows from the red extreme ( re ) to the blue end of the hb .
the zero point for @xmath36 is set at @xmath38 .
the hb ridgeline is determined as a non - linear fit to the sample of 160 hb stars supplemented with 68 rr lyr variables .
their colours are assumed to be randomly distributed inside the instability strip boundaries @xmath39 with a mean magnitude @xmath40 .
the scale ratios used is @xmath41 and @xmath42 .
a histogram of the @xmath36 distribution is shown in fig . 3 and the cumulative distribution in fig .
the gap is clearly recognised at @xmath43 .
following hawarden ( 1971 ) and newell ( 1973 ) , we tested the significance of the gap by means of @xmath44 statistics applied to the cumulative distribution and found that the gap is statistically significant at the 99.93 per cent level .
however , catelan et al . ( 1998 ) have recently argued that the hawarden - newell technique substantially overestimates the probability that the gap is real .
this result together with the fact that only 80 per cent ( 228 stars ) of all observed hb stars have been involved in our analysis make the reality of the gap not convincing enough .
figure 5 shows the theoretical zahb ( dorman , rood , oconel 1993 , oxygen - enhanced model , @xmath45 ; [ fe / h]=-1.48 ; [ o / fe]=0.63 ; @xmath46 ) overlaid on the observed hb .
squares denote stars with spectroscopically derived ( crocker , rood & oconnel 1988 ) .
the compatibility of both sets is satisfactory .
fig 5 . shows that although practically all hb stars lie above the theoretical sequence the discrepancy become marked above @xmath47 .
grundahl , catelan , landsman , stetson & andersen ( 1999 ) observed a similar effect ( called @xmath8jump ) in 14 gc ( including m5 ) from @xmath48 photometry .
these authors suggested that ... the @xmath8jump is a ubiquitous feature , intrinsic to _ all _ hb stars hotter than @xmath49. we note the following points regarding the @xmath50 diagram ( fig .
2b ) : 1 .
the stars of the upper part of the rgb are as bright as the bhb stars .
the @xmath12 magnitudes of the brightest rgb stars increase as the stars become redder
the agb is clearly distinguished from the rgb being , in average , more than 0.20 mag brighter .
note that the brightest agb stars have an astrometric membership probability @xmath51 per cent ( rees 1993 ; spassova & michnevski 1981 ) .
the star on the agb tip , denoted with full square , is zng2 @xmath3star ( zinn et al .
+ ccccccc @xmath52 & identification & u & v & b - v & status & ref . + + 1 & zng1 & & & & & + 2 & zng2 & 14.90 & 12.01 & 1.55 & m & 1 + 3 & zng3 , iii-24 & ..... & 13.51 & 0.69 & f & 1 + 4 & zng4 , iv-78 & 14.44 & 13.62 & 0.78 & f & 1 + 5 & zng5 , iv-65 & 14.49 & 14.05 & 0.51 & f & 1 + 6 & zng6 , var-41 & & & & m & 1 + 7 & zng7 , var 25 & & & & m & 1 + 8 & iii-37 & 15.04 & 14.79 & 0.02 & m & 2,3 + 9 & b169 & 15.09 & 14.61 & 0.38 & m & 3 + 10 & our 62 & 15.09 & 15.27 & -0.02 & & 4 + 11 & i-17 & 15.18 & 14.96 & 0.04 & & 4 + 12 & i-44 & 15.21 & 15.07 & 0.07 & m & 4 + 13 & ii-22 & 15.18 & 15.23 & 0.01 & & 4 + 14 & & & 14.56 & 0.42 & & 4 + 15 & & & 14.18 & 0.60 & & 4 + + + \1 .
zinn et al . ( 1972 ) ; 2 .
cudworth ( 1979 ) ; 3 . rees ( 1993 ) ; 4 .
this study 3 .
the group of the so called @xmath3bright stars is clearly separated from the principle sequences .
they are indicated with full squares in fig .
2 . six of the seven @xmath3bright stars from zinn et al .
( 1972 ) are measured in our study .
three of them are field stars , and one is variable 41 ( sawyer hogg 1973 ) .
the star zng7 is probably variable 25 . on the frames it appears blended with the rgb star iii140 ( buonanno et al .
our @xmath12photometry confirms that the cluster member iii37 ( cudworth 1979 ) is an @xmath3bright star : although both magnitudes and colours show big dispersion in all studies its position on the classical @xmath32 diagram is above the blue hb .
rees ( 1993 ) noted as other possible @xmath3bright stars : iii48 ( arp 1962 ) , b169 and b200 ( barnard 1931 ) .
our results confirm this status only for b169 .
we suspected four new possibles @xmath3bright stars : i17 , i44 , ii22 and the star numbered 62 in our catalogue .
another two stars were added to the list according to their position on the _ v , b v _ diagram .
these two stars , with magnitudes and colours : @xmath53 , @xmath54 and @xmath55 , @xmath56 , were detected in the central part of the cluster .
thus , 15 @xmath3bright stars candidates are found by different authors in m5 .
table 1 lists @xmath3bright stars including some that probably belong to the so - called blue nose ( @xmath57field , @xmath58cluster member ) .
the @xmath5 relation can be used to determine the foreground reddening towards m5 .
our intrinsic colours were derived assuming reddening e@xmath59=0.03 mag , with a slope e@xmath60/ e@xmath59=0.72 . for 38 bhb stars with _ b
[email protected] we estimate e@[email protected] .
the spectroscopically determined and @xmath63 for 11 hb stars in m5 ( crocker et al .
1988 ) , also allow us to determine the cluster reddening .
the @xmath64 values for these stars were derived using corresponding relations found in kurucz ( 1992 ) . in this way we obtained @xmath65 .
figure 6 shows the two colour diagram for rgb ( left panel , crosses ) and agb ( right panel , triangles ) for stars with _ v _ magnitudes above the level of the hb ( @xmath66 ) . in the both panels , the unreddened intrinsic two colour line for 1 normal ( l.c .
iii ) giants ( buser & kurucz 1992 ) is drawn ( solid line ) .
the two types of stars show the well known ultraviolet excess @xmath67(_u b_)@xmath68 .
this excess , measured at ( _ b v_)@xmath69=1.0 , is a sensitive indicator for the metal abundance ( wallerstein & helfer 1966 ; sandage 1970 ) . in our case @xmath70(_u
b_)@xmath71 for all stars with ( _ b
v_)@xmath72[[email protected] ] were averaged .
the mean excess @xmath70(_u
b_)@xmath74 is [email protected] . using the buser & kurucz s relation between @xmath70(_u b_)@xmath74 and [ m / h ] we obtained [ fe / h]@xmath75 .
the quoted error is derived only from the formal error propagation in the photometric and model data .
the m5 metallicity assessments found in the literature show a wide range ( briley et al .
1992 , their table 3 ) .
our value is in an agreement with that found in zinn & west ( 1984 ) , @xmath76 , but is inconsistent with the more recent high - resolution spectroscopic measurements carried out by carreta & bragaglia ( 1998 ) , @xmath77 ; carreta & gratton ( 1997 ) , @xmath78 , and sneden , kraft , prosser and langer ( 1992 ) @xmath79 . while we could not offer a certain answer on this topic , we would like to note that borissova et al .
( 1999 ) pointed out a remarkable similarity of the upper red giant branches of m5 and ngc6229 .
for this reason only , one might expect both clusters to share the same metallicity .
the reliability of the value derived photometrically for ngc6229 in borissova et al .
( 1999 ) , [ fe / h]@xmath80 , ( based on the zinn & west scale ) is confirmed by wachter , wallerstein , brown and oke ( 1998 ) .
these authors obtained the global metallicity of the cluster,[m / h ] @xmath80 , from a direct medium - resolution spectroscopy of cluster giants .
the agb stars ( within the same colour range ) show a smaller ultraviolet excess in comparison with the rgb stars .
this is demonstrated in fig .
6 ( right panel ) , where the mean position of the rgb stars is marked with a dashed line .
the difference in the @xmath3excesses for rgb and agb stars is obvious .
the implied difference in @xmath81 is 0.10 mag at @xmath82 .
a smaller @xmath3 excess of the agb stars has been found practically in all globular clusters with available @xmath0 measurements .
figure 7 illustrates the hb in the colour colour plane based on 161 stars ( 114 bhb and 47 rhb ) .
a dashed line shows the empirical 1 line .
a solid line indicates the model zahb transformed to @xmath83 .
this two colour diagram reveals the following features : * stars with @xmath84 fit very well the empirical 1 relation and zahb . these are stars occupying the region in the @xmath85 plane , where differences in _
g _ and metal abundance have negligible effect on the observed colours .
* the cool bhb stars , appearing just to the blue side of the rr lyrae instability strip , show a small but definite ultraviolet deficiency with respect to the intrinsic two colour line ( allen 1973 ) .
the mean deficiency measured for 18 stars in the colour interval @xmath86 is @xmath87 .
this effect was first found by arp ( 1962 ) and is based on only 5 bhb stars with photoelectric measurements .
the same effect was also observed in some other clusters : ngc 1904 ( stetson & harris 1977 ; kravtsov et al .
1997 ) , ngc 3201 ( lee 1977 ) , ngc 6723 ( menzies 1974 ) , ngc 5139 ( newell et al .
1969b ) , ngc 6397 ( newell et al .
1969a ) , m 15 ( sandage 1970 ) , ngc 1841 ( alcaino et al . 1996 ) . to quantify the effect
, we reanalysed available _
_ data for bhb stars in some of these clusters using the e@xmath59 values found in harris ( 1997 ) .
table 2 summarises the results .
+ lllll + ngc & @xmath88 ( u - b ) & n & e@xmath59 & reference + + 1904 & 0.162 @xmath89 & 20 & 0.01 & stetson et.al .
, 1977 + 3201 & @xmath90 & 26 & 0.21 & lee , 1977 + 5139 & @xmath91 & 6 & 0.12 & newell et.al . ,
1969b + 5904 & @xmath92 & 18 & 0.03 & this study + 6397 & @xmath93 & 8 & 0.18 & newell et .
, 1969a + 6723 & @xmath94 & 12 & 0.05 & menzies , 1974 + + + the third column shows the number of stars used .
however , extensive photoelectric observations of bhb stars in other globular clusters give somewhat conflicting results .
data for m3 , m13 and m92 ( sandage 1970 ) place such stars essentially on the standard 1 two colour line .
although the possibility of a small systematic error in our @xmath12 magnitudes can not be totally neglected , the observed effect in the m5 two colour diagram may be at least partly real .
the different values of the ultraviolet deficiency , toghether with it absolute absence in some cases , points to a probable connection of this observable with the cluster parameters .
the ultraviolet deficiency of the bhb stars in respect of the canonical zahb sequence is smaller ( see fig .
although the scatter of the data is large mainly due to the low quality of the photographic observations
the deviation from the model line is systematic and appears to be real . from the kuruzc s ( 1992 ) computations
it is evident that @xmath95 colour is a gravity indicator for stars with @xmath96 ( @xmath97 ) .
this point is illustrated in fig .
7 , where the @xmath98 and @xmath99 model loci are plotted .
it is clearly seen , that a satisfactory overlapping of the observed and theoretical sequences is achieved involving low gravity ( @xmath100 ) atmospheric models transformed to the @xmath101 plane .
+ the suggestion of arp ( 1962 ) that ... fainter bhb stars , as they get very blue have an increasingly large ultraviolet deficiency relatively to 1 line. is not confirmed .
* the rhb stars do not show any ultraviolet deficiency or ultraviolet excess with respect to the standard 1 colour colour line .
it has been known ( sandage 1969 , 1970 ) that the sequence of metal - deficient rhb stars follows an empirical 1 relation .
sandage ( 1970 ) gives an explanation in terms of the fortuitous cancellation of the effects of a low metal abundance and low surface gravity .
@xmath0 photographic and @xmath24 ccd observations in the central part ( @xmath102 arcmin ) of the globular cluster @xmath103 are presented . the @xmath3 colour - magnitude and two - colour diagrams for stars down to @xmath104 mag
are analysed .
the main results of our analysis can be summarised as follows : 1 .
we confirm the presence of a gap in the bhb stellar distribution of m5 announced by brocato at al .
( 1995 ) and drissen & shara ( 1998 ) . in the @xmath105 diagram , the gap is visible at @xmath106 and @xmath35 .
although the high statistical significance of this bhb feature , we consider the question of its reality still uncertain and will discuss it elsewhere ( baev et al . , in press ) .
the ultraviolet excess observed for rg stars , @xmath70 ( _ u
b_)@[email protected] , is used to estimate the metallicity of the cluster . the value obtained in our study , @xmath108
, confirms that derived by zinn & west ( 1984 ) , @xmath76 .
6 stars are suspected as new @xmath3bright stars .
the @xmath0 data for these stars are summarised in table 1 .
4 . in the @xmath109 diagram
we find that the hottest bhb stars tend to lie above the canonical zahb at a temperature threshold @xmath110 .
this resembles the effect observed for 14 globular clusters by grundahl et al .
( 1998,1999 ) and called @xmath8jump. in our investigation , the bhb sample is considerably richer than those used by gclsa .
this result on the other hand demonstrates the ability of johnson s @xmath0 to reproduce an effect observed by means of the str@xmath9mgren s photometry .
ultraviolet deficiency is detected for cool bhb stars .
based on kuruzc s ( 1992 ) atmospheric models , we interprete this effect as an indication for lower surface gravity for these stars . having in mind that @xmath8-jump demonstrated in gclsa is intimately connected to low gravities ( @xmath63jump )
a conclusion could be drawn that two groups of bhb stars in m5 tend to show surface gravities lower than those predicted by existing zahb models . finally , we would like to add few remarks concerning the nature of the ultraviolet deficiency and the jump .
gclsa argued that a stellar atmosphere effect radiative levitation of metals rather than helium mixing , is the primary cause of the @xmath8- and @xmath63-jump phenomenon .
one of their arguments against helium mixing is that the size of the jump and its location do not depend on a cluster s individuality .
collecting data for ultraviolet deficiency in clusters with available @xmath0 data , we found that the magnitude of the deficiency varies from cluster to cluster ( table 2 ) while in some clusters it is not observed at all .
these finds together with the results of gclsa allow us to suggest that the jump and the deficiency more likely have different nature .
in addition , taking in view certain evidences for mixing processes in some m5 red giant stars ( based on spectral observations referenced in gclsa ) , we suppose that the ultraviolet deficiency might be a photometric signature of helium mixing .
this draws a possible direction for a future ultraviolet deficiency investigation : to obtain a homogeneous set of ubv data for a large number of gcs and to study the relation of the cluster parameters to the characteristics of the deficiency ( if any ) .
the authors are grateful to s.moehler , f.grudahl and m.catelan for helpful information and discussions .
we would also like to thank the anonymous referee for very careful and detailed review of our paper .
+ this research was supported by the national sci . found .
grant under contract no .
f-604/1996 with the ministry of education and sciences . + this research has made use of nasa s astrophysics data system abstract service .
[ acknowledgements ] allen c.w . , 1973 , astrophysical quantities , 3rd ed . , the athlone press , london alcaino , g. ; liller , w. ; alvarado , f. ; kravtsov , v. ; ipatov , a. ; samus , n. ; smirnov , o. , 1996 , aj , 112 , 2020a arp h. , 1962 , apj , 135 , 311 auriere m. , & coupinot g. , 1989 , in proceedings of the 1-st eso / st - ecf data analysis workshop , eds p.g.grosbol , f.murtagh and r.h.warmels , p. 101
baev p.v . ,
markov h.s . ,
spassova n.m . ,
2000 , ( subbmitted in mnras ) barnard e.e . , 1931 ,
yerkes obs .
bohlin r.c . , cornett r.h.,hill j.k .
, smith a.m. , stecher t.p .
, 1985 , apj , 292 , 687 borissova j. , catelan m. , spassova n. , sweigart a.v . , 1997 , aj , 113 , 692 borissova j. , catelan m. , ferraro f.r .
, spassova n. , buonanno r. , iannicola g. , richtler t. , sweigart a.v . , 1999 ,
a&a , 343 , 813 briley m.m , smith g.h . , oke j.b .
, hesser j.e . , 1992 ,
apj , 387 , 612 brocato e. , castellsni v. , ripepi v.,1995 , aj , 109 , 1670 buser r. , kurucz r.i . , 1978 ,
a&a , 70 , 555 buser r. , kurucz r.i . , 1992 , a&a , 264 , 557 buonanno r. , corsi c.e . , fusi pecci f. , 1981 , mnras . , 196 , 435 , ( bcf ) carretta e. , bragaglia a. , 1998 , a&a , 329 , 937 carreta e. , gratton r.g . , 1997 , a&as 121 , 95 catelan m. , borissova j. , sweigart a.v . , spassova n. , 1998 , apj , 494 , 265 christian , c . , adams m. , barnes j.v .
, hayes d.s . , siegel m. , bulcher h. , mould j.r .
, 1985 , pasp , 97 , 363 crocker d. a. , rood r.t . , oconnel r.w . , 1988 , apj , 332 , 236 , ( cro ) crocker d.a . , rood r.t . , 1985 , in `` hb and uv - bright stars '' , ed.philip , p.107 cudworth k.m . , 1979 ,
aj , 84 , 1866 dixon w. v. d. , davidsen a.f . , dorman b. & ferguson h. c. , 1996 , aj , 111 , 1936 dorman , b. , rood , r. t. , & oconnell , r. w. 1993 , apj , 419,596 drissen l. , shara m.m . , 1998 , aj , 115 , 725 ferraro f.r . ,
fusi pecci f. , buonanno r. , 1992 , mnrs , 256 , 376 grundahl f. , vandenberg d.a . , andersen m.i .. 1998 , apjl , 500 , l179 grundahl f. , catelan m. , landsman w.b . , stetson p.b . ,
andersen m.i . 1999 ,
apj , 524 , 242 ( gclsa ) harris w.e . , 1997 ,
catalogue of parameters for mw globular clusters hawarden t.g . , 1971 , observatory , 91 , 78 kravtsov v. , ipatov a. , samus n. , smirnov o. , alcaino g. , alvarado f. , 1997 , a&as , 125 , 1 kurucz r.l .
, 1992 , iau symp . 149 , 225 lee s .- w.,1977 , a&as , 28 , 409 markov h.s . , 1994 , a&a , 291 , 1027 markov h. , valtchev t. , borisova j. , golev v. , 1997 , a&ass , 122 , 193 menzies j. , 1974 , mnras , 168 , 177 newell e.b .
, rodgers a.w . , searl i. , 1969a , apj , 156 , 597 newell e.b . , rodgers a.w . , searl i. , 1969b , apj , 158 , 699 newell e.b . , 1973 ,
apjs , 26 , 37 rees r.f . , 1993 , aj , 106 , 1524 richer h.b . &
fahlman g.g . , 1987 , apj , 316 , 189 sandage a. , 1969 , apj , 157 , 515 sandage a. , 1970 , apj , 162 , 841 sandquist e.l . , bolte m. , stetson p.b . , hesser j.e . , 1996 ,
apj , 470 , 910 sawyer hogg , h. , 1973 , publ .
david dunlop obs . , 3 , # 6 sneden ch . ,
kraft r.p .
, prosser ch .
f. , langer g.e .
, 1992 , aj , 104 , 2121 spassova n.m . , michnevski n. , 1981 , conpt .
rendus del acad .
, 36 , 83 stetson p.b . ,
harris w.e . , 1977 , aj , 82 , 954 stetson p.b . , 1987 ,
pasp 99 , 191 wachter s. , wallerstein g. , brown j.a . , oke j.b .
, 1998 , pasp 110 , 821 wallerstein g. , helfer h.l , 1966 , aj , 71 , 350 von braun k. , chiboucas k. , minske j.k . , salgado j.f .
, 1998 , pasp , 110 , 810 zinn r.j . , newell e.b . , gibson g.b . , 1972 ,
a&a , 18 , 390 zinn r.j . , west m. , 1984 , apj , 55 , 45 fig .
1 @xmath4 diagram for stars measured in the annulus with radius @xmath2 arcmin .
2 . the @xmath3 colour - magnitude diagrams . in both diagrams ,
the place of the suspected gap is indicated by a bar ; the @xmath3bright stars are marked with squares . in the panel
a ) , above the bhb , is shown the histogram of the stellar counts found in the corresponding colour bins
3 . the @xmath36 distribution histogram
4 . the @xmath36 cumulative distribution function . + fig .
comparison between the observed bhb ( grey points ) and the zahb ( solid line ) .
note the general discrepancy between observations and theory as well as the more rapidly rising discrepancy for colours below @xmath111 .
the squares mark the positions of m5 bhb stars with spectroscopically measured and @xmath63 ( crocker et al . 1988 ) .
+ fig . 6 . the two colour diagrams for rgb ( left panel ) and agb ( right panel ) stars .
the standard pop .
i line ( allen 1973 , solid line ) is overlaid on both diagrams . in the right pannel , the mean position of the rgb stars is marked with a dashed line .
two - colour diagram for hb stars .
the empirical 1 line is marked with a dashed line .
zahb is overlaid also ( bold solid line ) .
two - colour sequences taken from the kurucz ( 1992 ) atmospheric models for two different values of @xmath63 are marked with filled stars and squares . | we present stellar photometry in the @xmath0 passbands for the globular cluster m5@xmath1ngc5904 .
the observations , short - exposured photographic plates and ccd frames , were obtained in the rc - focus of the 2 m telescope of the natl .
astron .
obs .
rozhen. all stars in an annulus with radius @xmath2 arcmin were measured .
we show that the @xmath3 cmds describe different evolutionary stages in a better manner than the classical @xmath4 diagram .
we use hb stars , with known spectroscopic to check the validity of the colour zero - point . a review of all known @xmath3-bright star candidates in m5 is made and some of their parameters are catalogued .
six new stars of this kind are suspected on the basis of their position on the cmd .
new assessment of the cluster reddening and metallicity is done using the @xmath5 diagram .
we find [ fe / h]=1.38 , which confirms the zinn & west ( 1984 ) value contrasting with recent spectroscopic estimates . in an effort to clarify the question of the gap in the bhb stellar distribution and to investigate some other peculiarities , we use the relatively long - base colour index @xmath6 . a comparison of the observed @xmath7 distribution of hb stars with a canonical zahb model ( dorman et al .
1993 ) reveals that the hottest stars rise above the model line .
this is similar to the @xmath8jump found in the str@xmath9mgren photometry ( grundahl et al .
1998 , 1999 ) .
18 bhb stars with ( @xmath10 are used to estimate their ultraviolet deficiency . it is shown that low gravity ( @xmath11 ) kurucz s atmospheric models fit well the observed distribution of these stars along the two - colour diagram .
= = = = = = = = # 1 # 1 # 1 # 1 @mathgroup@group @mathgroup@normal@groupeurmn @mathgroup@bold@groupeurbn @mathgroup@group @mathgroup@normal@groupmsamn @mathgroup@bold@groupmsamn = `` 019 = ' ' 016 = `` 040 = ' ' 336 = " 33e = = = = = = = = # 1 # 1 # 1 # 1 = = = = = = = = 1pop .
i [ firstpage ] globular clusters : general globular clusters : individual ( m5@xmath1ngc5904 ) stars : evolution stars : horizontal branch
stars : population ii . |
flavour violation in the lepton sector ( lfv ) has been firmly established in neutrino oscillations . a direct evidence of conversion of electron neutrinos into a combination of muon and tau neutrinos is provided by solar neutrino oscillations . in a three neutrino framework
atmospheric neutrino data can only be explained by the dominant oscillation of muon neutrinos into tau neutrinos .
it is natural to expect that lfv takes place , at least at some level , in other processes such as those involving charged leptons .
flavour violating decays of charged leptons , strictly forbidden in the standard model ( sm ) , are indeed allowed as soon as neutrino mass terms are considered . if neutrino masses are the only source of lfv , the effects are too small to be detected , but in most extensions of the sm in which new particles and new interactions with a characteristic scale @xmath19 are included , the presence of new sources of flavour violation , in particular in the lepton sector , is a generic feature . in a low - energy description
, the corresponding effects can be parametrized by higher - dimensional operators describing flavour - violating rare decays of the charged leptons .
the dominant terms are represented by dimension six operators , suppressed by two powers of @xmath19 .
the present bounds on the branching ratios @xcite of these decays set stringent limits on combinations of the scale @xmath19 and the coefficients of the involved operators .
typically , for coefficients of order one , the existing bounds require a large scale @xmath19 , several orders of magnitude larger than the tev scale .
conversely , to allow for new physics at the tev scale , coefficients much smaller than one are required , which might suggest the presence of a flavour symmetry .
new physics at the tev scale supplemented by a flavour symmetry provides an interesting framework .
flavour symmetries have been invoked to describe the observed pattern of lepton masses and mixing angles @xcite , but quite often this approach is limited to a fit of the existing data , with very few new testable predictions @xcite .
specific relations among lfv processes are usually consequences of flavour symmetries and of their pattern of symmetry breaking .
this allows to get independent information on the flavour symmetry in the charged lepton sector .
moreover , if @xmath19 is sufficiently small , new particles might be produced and detected at the lhc , with features that could additionally confirm or reject the assumed symmetry pattern . in @xcite we have recently analyzed in an effective lagrangian approach lfv processes within a class of models possessing a flavour symmetry @xmath20 .
models of this class @xcite automatically reproduce nearly tri - bimaximal ( tb ) lepton mixing at the leading order ( lo ) via a vacuum alignment mechanism . in this type of models @xcite , corrections to tb mixing are generically proportional to the symmetry breaking parameter @xmath4 , especially the reactor mixing angle @xmath5 is of the order of @xmath4 .
the parameter @xmath4 is expected to vary between a few per mil and a few percent . in particular
we have evaluated the normalized branching ratios @xmath21 for the lfv transitions @xmath22 : r_ij= . in @xcite
we found that the generic expectation for the ratios @xmath21 is : r_ij = w_ij u^2 [ lfv ] where @xmath23 is the fine structure constant , @xmath24 is the fermi constant and @xmath25 are dimensionless parameters that can not be predicted within the effective lagrangian approach , but are expected to be of order one .
given the range of @xmath4 , it turns out that the present bound on @xmath0 requires the scale @xmath19 to be larger than about 10 tev .
if the underlying fundamental theory is weakly interacting , this translates into a lower bound of about @xmath26 on the typical mass of the new particles . at variance with this generic estimate ,
it was also observed that , if the flavour symmetry is realized in a supersymmetric ( susy ) model , a cancellation might take place in the amplitudes for lfv transitions .
an argument was given suggesting that in this case the ratios @xmath27 are of the form : r^susy_ij= [ lfvsusy ] where @xmath19 now corresponds to the supersymmetry breaking scale , @xmath28 ( @xmath29 ) are the charged lepton masses and @xmath30 are unknown dimensionless quantities of order one .
therefore in a supersymmetric context the predicted rates are more suppressed , allowing for new physics close to the tev scale , without conflicting with the present bounds .
since the cancellation expected in the supersymmetric case considerably changes the conclusion that can be derived from the existing bound on @xmath0 , it is important to perform a direct computation of the branching ratios within an explicit supersymmetric model incorporating the flavour symmetry @xmath3 . in the present paper
we consider the @xmath16 realization proposed in @xcite in which susy breaking effects were ignored and we extend it to a more realistic model , by adding a full set of susy breaking terms consistent with the flavour symmetry .
we assume that the breaking of supersymmetry occurs at a scale higher than or comparable to the flavour scale , simulated in our effective lagrangian by a cutoff , so that at energies close to the cutoff scale we have non - universal boundary conditions for the soft susy breaking terms , dictated by the flavour symmetry .
depending on the assumed mechanism of susy breaking we may have boundary conditions different from these , possibly enforced at a smaller energy scale .
for this reason , our approach maximizes the possible effects on lfv processes . through a detailed calculation of the slepton mass matrices in the physical basis ,
in which kinetic terms are canonical and leptons are in their mass eigenbasis , and evaluation of @xmath21 in the mass insertion ( mi ) approximation we find that the behaviour of @xmath21 , expected from the susy variant of the effective lagrangian approach , given in eq .
( [ lfvsusy ] ) is violated by a single , flavour independent contribution .
the correct scaling of @xmath21 is the one of eq .
( [ lfv ] ) , with a universal constant @xmath25 .
this implies @xmath31 at the lo in @xmath4 .
we identify the source of violation of the expected behaviour in a contribution to the right - left ( rl ) block of the slepton mass matrix , associated to the sector necessary to maintain the correct breaking of the flavour symmetry @xmath16 .
we also enumerate the conditions under which such a contribution is absent and the behavior in eq .
( [ lfvsusy ] ) is recovered , though we could not find a dynamical explanation to justify the realization of these conditions in our model . as shown in the mi approximation the coefficient @xmath25 of the leading contribution in eq .
( [ lfv ] ) is numerically small in a large region of the parameter space , where the sub - leading contributions play an important role .
we then provide a numerical study of @xmath21 by using full one - loop expressions and explore the parameter space of the model .
thereby , we assume a supergravity ( sugra ) framework with a common mass scale @xmath9 for soft sfermion and higgs masses and a common mass @xmath10 for gauginos at high energies . applying the current mega bound on @xmath32
we find that small values of @xmath4 or @xmath8 are favoured for @xmath9 and @xmath10 below @xmath11 gev .
employing the foreseen bound coming from the meg experiment constrains the model severely allowing only for small @xmath4 and @xmath8 for @xmath33 gev or requires larger @xmath9 and @xmath10 reducing the prospects for detection of sparticles at lhc .
furthermore , it turns out to be rather unnatural to reconcile the values of superparticle masses necessary to account for the measured deviation @xmath34 in @xmath35 , the muon anomalous magnetic moment , from the sm value with the present bound on @xmath32 . in our model values of @xmath36
smaller than @xmath37 are favoured .
the paper is structured as follows : in section 2 we review the basic features of the susy @xmath16 model and calculate kinetic terms and mass matrices of leptons and sleptons in an expansion in the symmetry breaking parameters . in section 3
we compute the slepton mass matrices in the physical basis .
furthermore , we estimate the possible effects on the slepton masses coming from renormalization group ( rg ) running .
section 4 is dedicated to the study of the quantities @xmath21 in the mi approximation , while in section 5 we discuss @xmath32 in a more quantitative way in the sugra context and comment on the size of the measured deviation in the muon anomalous magnetic moment @xmath35 in our model .
finally , we conclude in section 6 .
appendix a contains details about the group theory of @xmath16 , appendix b details of the calculation of the slepton mass matrices in the physical basis and in appendix c conventions and formulae , used in the numerical study of section 5 , are given .
in this section , we first discuss the main features and predictions of the lepton mixing in the class of @xmath16 realizations to which the present susy model belongs
. then we will present the khler and the superpotential of our model from which kinetic and mass terms are derived for leptons and sleptons .
the flavour symmetry @xmath16 can give rise to tb mixing in the lepton sector , if it is broken in a specific way , as discussed in @xcite . the group
@xmath16 is the group of even permutations of four objects which is isomorphic to the tetrahedral group .
it has 12 elements and four irreducible representations : three inequivalent singlets , denoted by @xmath38 , @xmath39 and @xmath40 and a triplet @xmath41 .
@xmath16 is generated by two generators @xmath42 and @xmath43 which fulfill the following relations @xcite : s^2=(st)^3=t^3=1 .
[ @xmath16 ] a set of generators @xmath42 and @xmath43 for all representations as well as kronecker products and clebsch gordan coefficients are given in appendix a. the specific breaking pattern of @xmath16 which leads to tb mixing for leptons requires that a @xmath45 subgroup , called @xmath46 , is preserved in the charged lepton sector , whereas a @xmath47 group , @xmath48 , remains unbroken in the neutrino sector .
the @xmath45 group is generated by the element @xmath43 and is left intact , if a scalar triplet @xmath49 acquires a vacuum expectation value ( vev ) of the form : ( 1,0,0 ) .
[ unozero ] breaking @xmath16 down to @xmath48 is achieved by a triplet which gets a vev : ( 1,1,1 ) .
[ unotre ] apart from @xmath16 , being responsible for tb mixing , the theory is invariant under additional flavour symmetries so that the full symmetry is : g_f = a_4z_3_fn_r .
[ flavgroup ] the cyclic symmetry @xmath45 ( not to be confused with the @xmath45 subgroup of @xmath16 preserved at lo in the charged lepton sector ) is necessary in order to avoid large mixing effects between the flavons that give masses to the charged leptons and those giving masses to neutrinos .
the mass hierarchy among the charged leptons is explained through the @xmath50 factor @xcite .
the second @xmath51 factor , @xmath52 , is a continuous @xmath53-symmetry containing @xmath53-parity as a subgroup . the flavour symmetry @xmath54 is broken by the chiral superfields @xmath55 , @xmath56 , @xmath57 and @xmath58 whose transformation properties under @xmath54 are shown in table 1 together with those of the lepton supermultiplets @xmath59 , @xmath60 , @xmath61 , @xmath62 and of the electroweak doublets @xmath63 . introducing
the driving fields @xmath64 , @xmath65 and @xmath66 , see also table 1 , we can write down the following superpotential : transforms in the same way under the symmetries of the model as the field @xmath57 .
however , it does not acquire a vev at lo and thus is not relevant for our discussion here .
for further details see @xcite . ]
@xmath67 where dots denote sub - leading non - renormalizable corrections . by @xmath68
we denote the contraction to an @xmath16 invariant . in the limit of unbroken supersymmetry all @xmath69 terms vanish . from these conditions
we can derive the vacuum alignment of the flavons @xmath55 , @xmath56 and @xmath57 : & = & ( u,0,0)+(c u^2,c u^2,c u^2)+o(u^3 ) + [ vevs ] & = & c_b(u , u , u)+o(u^2 ) + & = & c_a u+o(u^2)where @xmath70 are complex numbers with absolute value of order one and @xmath4 is one of the two small symmetry breaking parameters in the theory .
the scale @xmath71 is the cutoff scale associated to the flavour symmetry .
its value is expected to be around the scale of grand unification @xmath72 gev .
we display the sub - leading @xmath73 corrections , which arise from non - renormalizable terms in the superpotential @xmath74 and which have been studied in detail in @xcite , only for the field @xmath55 , since only these are relevant in the following analysis , whereas those of the fields @xmath57 and @xmath56 are important in the computation of neutrino masses and lepton mixings .
we remark that in the susy limit the vevs of the driving fields @xmath64 , @xmath65 and @xmath66 are zero .
this is however in general no longer true , if we include soft susy breaking terms into the flavon potential , as we shall see in the next section and as has been discussed in @xcite . for the froggatt - nielsen ( fn ) field @xmath58 to acquire a vev
, we assume that the symmetry @xmath50 is gauged such that @xmath58 gets its vev through a @xmath75 term .
the corresponding potential is of the form : @xmath76 where @xmath77 is the gauge coupling constant of @xmath50 and @xmath78 denotes the contribution of the fayet - iliopoulos ( fi ) term .
dots in eq.([dterm ] ) represent e.g. terms involving the right - handed charged leptons @xmath60 and @xmath61 which are charged under @xmath50 .
these terms are relevant in the calculation of the different contributions to the soft mass terms of right - right ( rr ) type , as we shall see below .
@xmath79 leads in the susy limit to : @xmath80 which we parametrize as : @xmath81 with @xmath82 being the second small symmetry breaking parameter in our model . both @xmath4 and @xmath82
can also be in general complex , but through field redefinitions of the supermultiplets @xmath55 and @xmath58 , they can be made real and positive .
@xmath83 \text{field } & l & e^c & \mu^c & \tau^c & h_{u , d } & \phit & \phis & \xi & \theta_{fn } & \varphi^t_0 & \varphi^s_0 & \xi_0\\[10pt ] \hline & & & & & & & & & & & & \\[-9pt ] a_4 & 3 & 1 & \onepp & \onep & 1 & 3 & 3 & 1 & 1 & 3 & 3 & 1\\[3pt ] \hline & & & & & & & & & & & & \\[-9pt ] \zt & \om & \om^2 & \om^2 & \om^2 & 1 & 1 & \om & \om & 1 & 1 & \om & \om\\[3pt ] \hline & & & & & & & & & & & & \\[-9pt ] \uu_{fn } & 0 & 2 & 1 & 0 & 0 & 0 & 0 & 0 & -1 & 0 & 0 & 0\\[3pt ] \hline & & & & & & & & & & & & \\[-9pt ] \uu_{r } & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 2 & 2 & 2\\[3pt ] \hline \end{array } $ ] 0.2 cm 0.2 cm at the lo , the mass matrices of charged leptons and light neutrinos have the following form : m_l ( ccc y_e t^2 & 0 & 0 + 0 & y_t & 0 + 0 & 0 & y _ ) u , [ yf ] and m _ ( ccc a+2 b/3 & -b/3 & -b/3 + -b/3 & 2b/3 & a - b/3 + -b/3 & a - b/3 & 2 b/3 ) u , [ y ] with @xmath84 , @xmath85 , @xmath86 , @xmath87 and @xmath88 being complex numbers with absolute values of order one . as one can see the relative hierarchy among the charged lepton masses is given by the parameter @xmath82 for which we take : t0.05 . [ tbound ]
the neutrino mass matrix is diagonalized by : u_tb^t m_u_tb ( a+b , a ,-
a+b ) u , where @xmath89 , up to phases , is the tb mixing matrix@xcite : u_tb= ( ccc & 1/ & 0 + -1/ & 1/ & -1/ + -1/ & 1/ & + 1/ ) . [ utb ] the allowed range of the parameter @xmath4 is determined by the requirement that sub - leading corrections which perturb the lo result are not too large and by the requirement that the @xmath15 yukawa coupling @xmath86 does not become too large . the first requirement results in an upper bound on @xmath4 of about @xmath90 , which mainly comes from the fact that the solar mixing angle should remain in its @xmath91 range @xcite .
the second one gives a lower bound which we estimate as : u= 0.01 [ tanb&u&yt ] where @xmath92 gev and @xmath8 is the ratio between the vevs of the neutral spin zero components of @xmath93 and @xmath94 , the two doublets responsible for electroweak symmetry breaking .
for the @xmath15 lepton we use its pole mass @xmath95 @xcite .
requesting @xmath96 we find a lower limit on @xmath4 close to the upper bound @xmath90 for @xmath97 , whereas @xmath98 can not be too small @xcite . here
we take @xmath98 as the smallest allowed value . ]
gives as lower limit @xmath99 .
obviously , these limits depend on the largest allowed value of @xmath100 , as well as on whether we identify @xmath101 with the pole mass or with the @xmath15 mass renormalized at some reference scale , such as the scale of grand unification .
we choose as maximal range : 0.007 u 0.05 , [ ubound ] which shrinks when @xmath8 is increased from 2 to 15 . concerning the relative size of the two symmetry breaking parameters we note that @xmath102 holds for all values of @xmath4 .
0.5 cm 0.5 cm we analyse the lagrangian of the model : & = & d^2d^2 ( , e^2 v z)+ + & + & , [ leel ] where @xmath103 is the khler potential , a real gauge - invariant function of the chiral superfields @xmath104 and their conjugates , of dimensionality ( mass)@xmath105 ; @xmath106 is the superpotential , an analytic gauge - invariant function of the chiral superfields , of dimensionality ( mass)@xmath107 ; @xmath108 is the gauge kinetic function , a dimensionless analytic gauge - invariant function ; @xmath109 is the lie - algebra valued vector supermultiplet , describing the gauge fields and their superpartners .
finally @xmath110 is the chiral superfield describing , together with the function @xmath108 , the kinetic terms of gauge bosons and their superpartners .
each of the terms on the right - hand side can be written in an expansion in powers of the flavon fields .
since we have two independent symmetry breaking parameters , @xmath4 and @xmath82 , see eqs .
( [ vevs ] ) and eq .
( [ vevtheta ] ) , we consider a double expansion of @xmath111 in powers of @xmath4 and @xmath82 . in this expansion
we keep terms up to the second order in @xmath4 , i.e. terms quadratic in the fields @xmath112 and @xmath57 .
the expansion in the parameter @xmath82 , responsible for the breaking of the froggatt - nielsen u(1)@xmath113 symmetry , is stopped at the first non - trivial order , that is by allowing as many powers of the field @xmath58 as necessary in order to obtain non - vanishing values for all entries of the matrices describing lepton masses as well as for the entries of the matrices describing kinetic terms and slepton masses .. these contribute to the diagonal elements of the kinetic terms and the slepton masses . in the khler potential for the left - handed fields they can be safely neglected , since the lo correction is of @xmath114 . in the right - handed sector , they contribute at the same order as the terms arising through a double flavon insertion . ] finally , second order corrections in @xmath4 also arise from the sub - leading terms of the vev @xmath115 , eq .
( [ vevs ] ) , and are included in our estimates .
the soft susy breaking terms are generated from the susy lagrangian by promoting all coupling constants , such as yukawa couplings , couplings in the flavon superpotential and couplings in the khler potential , to superfields with constant @xmath116 and @xmath117 components @xcite . through this
we derive subsequently the soft masses @xmath118 and @xmath119 from the khler potential .
one contribution to @xmath120 , which we call @xmath121 in the following , arises from the yukawa couplings present in the superpotential @xmath122 .
important contributions to slepton masses originate from the modification of the vevs of flavons and driving fields due to susy breaking effects .
a detailed study of the vevs of these fields and their dependence on the soft susy breaking parameters is presented in @xcite and we summarize the main results here . when soft susy breaking terms are included into the flavon potential , the vevs in eq .
( [ vevs ] ) receive additional contributions of order @xmath9 , completely negligible compared to @xmath123 . at the same time , the driving fields @xmath64 , @xmath65 and @xmath66 develop a vev of the size of the soft susy breaking scale @xmath9 .
an equivalent statement is that the auxiliary fields of the flavons acquire a vev at the lo of the size of @xmath124 .
especially , for the auxiliary fields contained in the flavon supermultiplet @xmath55 we have @xcite : @xmath125 where @xmath126 , @xmath127 and @xmath128 are in general complex numbers with absolute value of order one .
the parameter @xmath126 vanishes in the special case of universal soft mass terms in the flavon potential .
when different from zero , the vevs of the auxiliary components of the flavon supermultiplet @xmath55 generate another contribution to the soft masses of rl - type , which we denote as @xmath129 .
this contribution is analogous to the one which has been found before in the supergravity context and which can have a considerable effect on the size of the branching ratio of radiative leptonic decays , as shown in @xcite . indeed , as we shall see below , in the global susy model under consideration the leading dependence of the normalized branching ratios @xmath21 on @xmath4 is dominated by @xmath129 .
we remark that the vevs in eq .
( [ vevsauxphit ] ) and those of the corresponding flavon field @xmath55 in eq .
( [ vevs ] ) have a similar structure but they are not proportional , in general .
this is due to the different coefficients @xmath130 , @xmath131 and @xmath128 , @xmath127 , which can be qualitatively understood as follows : the coefficients @xmath130 , @xmath131 mainly depend on a set of parameters that remain in the susy limit and receive completely negligible corrections from the susy breaking terms . on the contrary @xmath132 vanishes in the susy limit , to all orders in @xmath4 , and @xmath128 , @xmath127 crucially depend on the set of parameters describing the susy breaking .
we will see that , if @xmath130 and @xmath128 accidentally coincide ( up to complex conjugation ) , a cancellation in the leading behaviour of @xmath21 takes place .
similarly , the vev of the fn field @xmath58 becomes shifted , when soft susy breaking terms are included into the potential , so that : @xmath133 with @xmath134 being an order one number , holds .
this will lead to a contribution @xmath135 to the soft masses of rr - type , since only the right - handed charged leptons @xmath60 and @xmath61 are charged under @xmath50 . apart from these there are supersymmetric contributions to @xmath136 and @xmath137 from @xmath69 and @xmath75 terms , @xmath138 and @xmath139 , as well as a contribution to @xmath140 coming from the @xmath69 term of @xmath94 , called @xmath141 in the following . in our notation a chiral superfield and its @xmath53-parity even component
are denoted by the same letter .
the @xmath53-parity odd component is indicated by a tilde in the following and the conjugate ( anti - chiral ) superfield is denoted by a bar . the expansion of the khler potential can be written as : k = k^(0)+k^(1)+k^(2)+ ... in our model the khler potential deviates from the canonical form , @xmath142 , due to the contributions of non - renormalizable terms , invariant under both the gauge and the flavour symmetries .
such contributions are sub - leading in the @xmath143 expansion , but in principle they are non - negligible , since the redefinitions required to arrive at canonically normalized fields can also affect lepton and slepton mass matrices . the lo term in @xmath4
is given by : k^(0)=k_0_i=1 ^ 3|l_i l_i+ _
i=1 ^ 3 |l^c_i l^c_i+k_u h_u^2+k_d h_d^2 + k_fn _ fn ^2 + ... [ kappa0 ] where @xmath144 and dots stand for additional contributions related to the flavon sector .
the quantities @xmath145 , @xmath146 , @xmath147 , @xmath148 , @xmath149 and the corresponding ones in the following formulae , are treated as superfields with a constant @xmath117 component , as is explained above .
concerning the effect of the superfield @xmath58 on the khler potential @xmath150 of the supermultiplets @xmath151 and @xmath152 note that for our purposes we can neglect such terms , with the exception of the right - handed sector , where terms up to the second order in @xmath82 have to be taken into account . at the first order in @xmath4 we have : k^(1)= ( _ t ( |l l)_s)+ ( _ t ( |l l)_a)+ ( _ t ( |l l)_s)+ ( _ t ( |l l)_a)+h.c .
@xmath153 denotes an invariant under @xmath16 , while @xmath154 and @xmath155 stand for the @xmath39 and @xmath40 singlets .
finally two @xmath16 triplets , @xmath87 and @xmath88 , can be combined into the symmetric triplet @xmath156 or the anti - symmetric one @xmath157 .
the su(2 ) singlet fields @xmath158 are not affected by the first - order correction @xmath159 . at the second order in @xmath4 we have a richer structure : k^(2)=k_l^(2)+k_r^(2 ) , with the labels @xmath160 and @xmath53 referring to lepton doublets @xmath59 and singlets @xmath158 , respectively . for lepton doublets
we find : k_l^(2)= _
i=1 ^ 7 ( x_i |l l ) [ here ] where @xmath161 is the list of @xmath45-invariant operators , bilinear in the flavon superfields @xmath112 and @xmath57 and their conjugates , x=\ { , _ t^2,(_t)^2,_t_t,_s _ s,_s,_s } , [ xlist ] and each quantity @xmath162 represents a list of parameters since there can be different non - equivalent ways of combining @xmath163 with @xmath164 to form an @xmath16-invariant .
there are also obvious relations among the coefficients @xmath162 to guarantee that @xmath165 is real .
note that the sum in eq .
( [ here ] ) runs over all bilinears which can couple to form @xmath16-invariants . whether they lead to a non - trivial contribution to the relevant terms
depends on the vevs of the flavons . for lepton singlets
, we can distinguish a diagonal contribution and a non - diagonal one : k_r^(2)=[k_r^(2)]_d+[k_r^(2)]_nd [ k_r^(2)]_d= _ i=1 ^ 5 _ nd&= & _
i=2 ^ 5 + [ here2 ] & + & _
i=2 ^ 5 + & + & _
i=2 ^ 5 .due to the structure of the flavon vevs only the term with @xmath166 gives a non - vanishing contribution in the sum in eq .
( [ here2 ] ) . in order to generate a set of supersymmetry breaking soft mass terms , the quantities : k_i=\{k_0,(k^c_0)_i,(^c_0)_i , k_u , d , k_fn , k_s , k_a , k_s , k_a , k_j,(k^c_e)_l,(k^c_)_l,(k^c_)_l,(k^c_e)_l,(k^c_e)_l,(k^c_)_l } are treated as superfields with a constant @xmath117 component : k_i = p_i+q_i m^2_susy [ deck ] where @xmath167 and @xmath168 do not depend on the grassmann variables . and/or @xmath169 , including also terms proportional to @xmath116 and to @xmath170 .
these terms can be absorbed in our parametrization after a suitable redefinition of the parameters .
] the quantities @xmath167 and @xmath168 are parameters with absolute values of order one . in particular , it is not restrictive to choose ( see eq .
( [ kappa0 ] ) ) : k_0=1+q_0 m_susy^2 , ( k_0^c)_i=1+(q_0^c)_i m_susy^2 , [ kudexp ] k_u , d=1+q_u , d m_susy^2 , k_fn=1+q_fn m_susy^2 . when the flavon fields acquire a vev according to the pattern shown in eqs .
( [ vevs ] ) and eq .
( [ vevtheta ] ) , the khler potential @xmath171 gives rise to non - canonical kinetic terms for lepton and slepton fields of the following form : ( @xmath172 ) denotes the left - handed ( right - handed ) component of the electron field .
for instance , in terms of the four - component spinor @xmath173 , the bilinears @xmath174 and @xmath175 correspond to @xmath176 and @xmath177 [ @xmath178 respectively .
we take @xmath179 , @xmath180 , @xmath181 , @xmath182 and @xmath183 , where @xmath184 are the @xmath185 pauli matrices . here
the four - component matrix @xmath186 is in the chiral basis , where the 2@xmath1872 blocks along the diagonal vanish , the upper - right block is given by @xmath188 and the lower - left block is equal to @xmath189 . ] _
kin&=&i k_ij|l_i |^d_l_j+i k^c_ij|l^c_i & + & k_ij_i d__j+k^c_ij _ i d_^c_j [ kt ] the matrices @xmath190 and @xmath191 are given by : k= ( ccc 1 + 2 t_1 u & t_4 u^2 & t_5 u^2 + _ 4 u^2 & 1-(t_1+t_2 ) u & t_6 u^2 + _ 5 u^2 & _ 6 u^2 & 1-(t_1-t_2 ) u ) , [ t ] 0.5 cm k^c= ( ccc 1+t^c_1 u^2+t^c_1 t^2 & t^c_4 u^2 t & t^c_5 u^2 t^2 + _ 4 u^2 t & 1+t^c_2 u^2+t^c_2 t^2 & t^c_6 u^2 t + _ 5 u^2 t^2 & _ 6 u^2 t & 1+t^c_3 u^2+t^c_3 t^2 ) .
[ tc ] the coefficients @xmath192 , @xmath193 and @xmath194 are real , while the remaining coefficients are complex .
as one can see , the corrections to the kinetic terms which render these non - canonical are small , at most at order @xmath114 and @xmath195 .
they have to be taken into account when calculating lepton and slepton masses . the coefficients @xmath196 and @xmath197 are linearly related to the parameters @xmath167 introduced before , see eq .
( [ deck ] ) .
such a relation is not particularly significant and in the rest of this paper we will treat @xmath196 and @xmath197 as input parameters , with absolute values of order one .
notice that we have neglected possible sub - leading contributions to the kinetic terms of @xmath63 , and to the flavons themselves and @xmath58 , which have no impact on the present analysis .
note further that through the expansion of @xmath148 as given in eq .
( [ kudexp ] ) also the soft masses @xmath198 for the higgs doublets @xmath93 and @xmath94 are generated of size @xmath199 .
similarly , the expansion of @xmath149 gives rise to a soft mass term for @xmath58 of size @xmath200 .
we continue with the discussion of the superpotential : w = w_l+w_+w_d+w_h+ ... there is a part responsible for the charged lepton masses : w_l = w_l^(1)+w_l^(2)+ ... the leading term in the @xmath4 expansion is : w_l^(1)=
_ fn^2e^c h_d ( _
( _ t l) + ^c h_d ( _
t l ) .
[ yl1 ] at the next order in @xmath4 we find : w_l^(2)= _
fn^2e^c h_d ( _ t^2 l ) + _
fn^c h_d ( _ t^2 l) + ^c h_d ( _ t^2 l ) . [ yl2 ] to generate both the yukawa interactions and the soft mass contribution @xmath121 , we regard the quantities @xmath201 and @xmath202 as constant superfields , of the type : x_f = y_f - z_f m_susy , x_f = y_f - z_f m_susy ( f = e , , ) , [ xdecomp ] where the coefficients @xmath203 , @xmath204 , @xmath205 and @xmath206 have absolute values of order one . from eqs .
( [ yl1 ] ) and ( [ yl2 ] ) , after the breaking of the flavour and the electroweak symmetries , we find the following mass matrix for the charged leptons : @xmath207 note that the matrix @xmath208 is shown in the basis in which the kinetic terms for the lepton fields are given by eqs .
( [ kt],[t],[tc ] ) . at the lo in the @xmath4 expansion
, @xmath208 matches the matrix shown in eq .
( [ yf ] ) .
the off - diagonal elements of @xmath208 , all proportional to @xmath130 , originate from the sub - leading contributions to the vev of the @xmath55 multiplet . similarly , the superpotential giving rise to neutrino masses can be expanded as : w_=w_^(1)+w_^(2)+ ... with the lo terms : w_^(1)= ( h_u l h_u
l ) + ( _ s h_u l h_u l ) , [ here3 ] where @xmath209 is the scale at which lepton number is violated .
also in this case @xmath210 are constant superfields : x_a , b = y_a , b+z_a , b m_susy .
the terms in @xmath211 lead to the form of @xmath212 as in eq .
( [ y ] ) with @xmath213 and @xmath214 .
the next term in the expansion @xmath215 is not relevant for our discussion here , however gives rise to deviations of relative order @xmath4 from tb mixing .
the form of @xmath74 , responsible for the vacuum alignment of the flavon fields , has already been displayed above in eq .
( [ wd ] ) .
finally , the term @xmath216 is associated with the @xmath217 parameter : w_h = h_u h_d .
[ muterm ] this term explicitly breaks the ( continuous ) @xmath52 symmetry of the model , while preserving @xmath53-parity .
the @xmath217 term might originate from the khler potential of the theory @xcite , after susy breaking : d^2 d^2 |(|x h_u h_d + h.c . ) where @xmath218 is some large scale as , for instance , the planck scale and @xmath161 is a chiral superfield , whose @xmath69 component @xmath219 develops a vev .
this gives rise to @xmath220 . in our model
we simply assume the existence of the term in eq .
( [ muterm ] ) in the superpotential and allow for an explicit breaking of the @xmath52 symmetry controlled by the parameter @xmath217 .
the soft susy breaking term @xmath221 can then arise from the @xmath217 term by considering @xmath217 as superfield @xmath222 .
a source of the @xmath217 term within our model are terms which involve one driving field , a certain number of flavons and the two higgs doublets @xmath93 and @xmath94 .
the lowest order term in the superpotential @xmath122 , allowed by all symmetries of the model , is : @xmath223 this term generates a contribution to the @xmath217 term of the order of @xmath224 , when the driving fields acquire vevs through the inclusion of soft susy breaking terms into the flavon potential @xcite .
the size of such a term is expected to be @xmath225 gev for @xmath226 . from the khler potential and the superpotential we can read off the slepton masses
; they can be parametrized as follows : -l_m _ i , j=1 ^ 3 + _
i , j=1 ^ 3 |_i ( m_ll^2)_ij _ j where @xmath227 and @xmath228 are hermitian matrices and @xmath229 . in the sneutrino sector
only the block @xmath230 is present .
we neglect contributions to the sneutrino masses associated to @xmath231 .
each of these blocks receives several contributions : m^2_(e,)ll&=&(m^2_(e , ) ll)_k+(m^2_(e , ) ll)_f+(m^2_(e , ) ll)_d , + m^2_err&=&(m^2_err)_k+(m^2_err)_f+(m^2_err)_d+(m^2_err)_d , fn .
the contribution to the slepton masses from the susy breaking terms in the khler potential is given by : ( m^2_ell)_k=(m^2_ll)_k= ( ccc n_0 + 2 n_1 u & n_4 u^2 & n_5 u^2 + _ 4 u^2 & n_0-(n_1+n_2 ) u & n_6 u^2 + _ 5 u^2 & _ 6 u^2 & n_0-(n_1-n_2 ) u ) m^2_susy , [ m2ll ] 0.5 cm ( m^2_err)_k= ( ccc n^c_1 & n^c_4 u^2 t & n^c_5 u^2 t^2 + _ 4 u^2 t&n^c_2 & n^c_6 u^2 t + _ 5 u^2 t^2 & _ 6 u^2 t & n^c_3 ) m^2_susy , [ m2rr ] where the coefficients are complex , except for @xmath232 and @xmath233 . and @xmath233 are positive in order to have positive definite square - masses , to avoid electric - charge breaking minima and further sources of electroweak symmetry breaking . ]
the coefficients @xmath234 and @xmath235 are linearly related to the parameters @xmath168 introduced in eq .
( [ deck ] ) .
again , such a relation is not particularly significant and in the rest of this paper we will treat @xmath234 and @xmath235 as input parameters , with absolute values of order one .
the susy contribution from the @xmath69 terms is completely negligible for sneutrinos , i.e. @xmath236 . for charged sleptons @xmath237 and @xmath238 read : ( m^2_ell)_f = m_l^ ( k^c)^-1 m_l , ( m^2_err)_f = m_l ( k^-1)^t m_l^ , where @xmath208 is the charged lepton mass matrix and @xmath190 , @xmath191 are the matrices specifying the kinetic terms , eqs . ( [ t ] ) and ( [ tc ] ) .
the susy @xmath75 term contribution is : ( m^2_ell)_d&=&(-+^2_w ) 2 m_z^2 k , + ( m^2_ll)_d&=&(+ ) 2 m_z^2 k , + ( m^2_err)_d&=&(-1 ) ^2_w 2 m_z^2 ( k^c)^t , where @xmath239 is the @xmath240 mass and @xmath241 is the weinberg angle .
notice again the presence of the matrices @xmath190 and @xmath191 . for the right - handed charged leptons we find an additional @xmath75 term contribution stemming from the fact that @xmath58 , @xmath60 and @xmath61 are charged under the froggatt - nielsen symmetry @xmath50 , which we assume to be gauged .
the relevant contribution of the @xmath50 group to the scalar potential ( through a @xmath75 term ) is : v_d , fn&= & ( m_fi^2 + g_fn q_fn^i z_i)^2 + q_fn m_susy^2 _ fn ^2 + & = & g_fn^2 c_m_susy^2 ( 2||^2+||^2+t_4^c u^2 t | + _ 4 u^2 t | ) + ... where @xmath242 stands for the fn charge of the scalar field @xmath243 , and in the second line we have displayed only the leading contribution to the terms quadratic in the matter fields .
one can check that this contribution is of a similar form and size as the one originating from the khler potential , @xmath119 .
thus , we can simply absorb it by redefining the -anyway- unknown coefficients @xmath244 , @xmath245 parametrizing @xmath119 in eq .
( [ m2rr ] ) .
concerning the block @xmath246 , this receives three contributions : m^2_erl=(m^2_erl)_1+(m^2_erl)_2 + ( m^2_erl)_3 .
the first one originates from the superpotential , eqs .
( [ yl1 ] ) and ( [ yl2 ] ) , and is proportional to the parameters @xmath204 and @xmath206 of the decomposition in eq .
( [ xdecomp ] ) : @xmath247 with a_1 =( ccc z_e t^2 u+(z_e+c z_e ) t^2 u^2 & c z_e t^2 u^2 & c z_e t^2 u^2 + c z_t u^2 & z_t u + ( z_+ c z _ ) t u^2 & c z_t u^2 + c z_u^2 & c z_u^2 & z_u+(z_+ c z _ ) u^2 ) .
[ rlnsusy ] the second one is related to the fact that the auxiliary fields of the flavon supermultiplet @xmath55 acquire non - vanishing vevs of the form as shown in eq .
( [ vevsauxphit ] ) , when soft susy breaking terms are included into the flavon potential .
formally the contribution can be written as : @xmath248 so that the soft mass matrix @xmath249 reads : @xmath250 with a_2 = | ( ccc y_e t^2 u+(2 y_e+ y_e ) t^2 u^2 & y_e t^2 u^2 & y_e t^2 u^2 + y_t u^2 & y_t u + ( 2 y_+ y _ ) t u^2 & y_t u^2 + y_u^2 & y_u^2 & y_u+(2 y_+ y _ ) u^2 ) .
[ rlnsusy2 ] the third contribution is proportional to the charged lepton mass matrix and is enhanced in the large @xmath8 regime : ( m^2_erl)_3=-| m_l .
[ rlsusy ] we remark that the origins of the second and the third contribution are quite similar , since both arise from the auxiliary component of a superfield : the second one is attributed to the auxiliary component of the flavon superfield @xmath55 , while the third one originates from the auxiliary component associated to the higgs doublet @xmath94 .
in this section we first discuss the results for the slepton masses in the physical basis and comment on results found in the literature . in the second part , we give an estimate of the renormalization group effects on the slepton masses and show in the leading logarithmic approximation that these effects can be neglected or absorbed into the parametrization of the soft mass terms .
all matrices above are given in the basis in which the kinetic terms of the slepton and lepton fields are non - canonical .
to derive the physical masses and the unitary transformations that enter our computation , we have to go into a basis in which kinetic terms are canonical , for both , slepton and lepton , fields .
subsequently , we diagonalize the mass matrix of the charged leptons via a biunitary transformation . to avoid flavour - violating gaugino - lepton - slepton vertices in this intermediate step , we perform the same transformation on both fermion and scalar components of the involved chiral superfields .
this procedure , described in detail in appendix b , gives us the physical slepton mass matrices @xmath251 , @xmath252 and @xmath253 . the results shown here are obtained under the assumption that all parameters of the model are real .
the analytical expressions for the slepton mass matrices in the physical basis contain the first non - vanishing order in each of the matrix elements .
we start with the left - left ( ll ) block .
the contribution from the soft breaking terms is common to charged sleptons and sneutrinos and reads : @xmath254 @xmath255 the supersymmetric @xmath69 and @xmath75 term contributions are given by : @xmath256 and @xmath257 with @xmath258 being the mass matrix for the charged leptons in the same basis , i.e. diagonal and with canonically normalized kinetic terms .
the supersymmetric @xmath75 term contributions are proportional to the unit matrix .
notice that in the physical basis all susy contributions are diagonal in flavour space . both , the @xmath69 and the @xmath75 term , contributions are small compared to that coming from the khler potential .
the relative suppression is of order @xmath259 and @xmath260 , respectively , which do not exceed the per cent level for typical values of @xmath9 around 1 tev .
note also that the susy part is the only one that distinguishes between charged sleptons and sneutrinos .
the dominant mass matrix , @xmath261 , has a structure which is very similar to that of the corresponding matrix in the original basis , i.e. all matrix elements in the two bases are of the same order in the @xmath143 expansion .
they only differ for coefficients of order one . finally , coming to the rl block of the mass matrix for charged sleptons
, we find : @xmath272 c \
, \dd\frac{(z_\mu y_\mu^\prime - z_\mu^\prime y_\mu)}{y_\mu^2 } \ , m_\mu u^2 & \dd\frac{z_\mu}{y_\mu } m_\mu & 2 c \,\dd\frac{(z_\mu y_\tau - z_\tau y_\mu)}{y_\mu y_\tau } m_\mu u\\[0.1 in ] c \ , \dd\frac{(z_\tau y_\tau ^\prime - z_\tau ^\prime y_\tau)}{y_\tau^2 } \ , m_\tau u^2 & c \ , \dd\frac{(z_\tau y_\tau ^\prime - z_\tau ^\prime y_\tau)}{y_\tau^2 } \ , m_\tau u^2 & \dd\frac{z_\tau}{y_\tau } \ , m_\tau \end{array } \right ) \ , m_{susy } \ ; , \ ] ] @xmath273 ( c_f -c ) \ , m_\mu u & m_\mu & ( c_f -c ) \ , m_\mu u\\[0.1 in ] ( c_f -c ) \ , m_\tau u & ( c_f -c ) \ , m_\tau u & m_\tau \end{array } \right ) \ , m_{susy } \ ; , \ ] ] and [ eq : m_rl_hat_3 ] ( ^2_erl)_3=- _ l .
the matrix @xmath274 , which is the sum of these three contributions , does not depend on the parameters describing the non - canonical kinetic terms through @xmath190 and @xmath191 .
an important feature of @xmath275 is that the elements below the diagonal are suppressed by a factor @xmath4 compared to the corresponding elements of @xmath276 , i.e. before the transformations for canonical normalization of the kinetic terms and diagonalization of the charged lepton mass matrix are applied .
however , this does not happen for the second contribution associated to the non - vanishing vevs of the auxiliary components of the supermultiplet @xmath55 so that the elements of @xmath277 are still of the same order in the expansion parameters @xmath82 and @xmath4 as those of the matrix @xmath278 . nevertheless there are cases in which this contribution can be suppressed . in the first case
the vevs of the auxiliary fields contained in the supermultiplet @xmath55 vanish , i.e. the parameter @xmath126 is zero , due to the fact that the soft susy breaking terms in the flavon potential are ( assumed to be ) universal , that is equal to the terms of the superpotential @xmath74 up to an overall proportionality constant @xcite .
the second possibility arises , if the vevs of the auxiliary fields can be completely aligned with those of the flavon @xmath55 at lo as well as nlo , such that @xmath128 becomes equal to @xmath130 . in both cases the off - diagonal elements of @xmath277
are further suppressed than shown in eq .
( [ eq : m_rl_hat_2 ] ) .
we emphasize this fact here , since it turns out that the suppression of the off - diagonal elements below the diagonal as it occurs in the case of @xmath279 is relevant for the actual size of the leading behaviour of the normalized branching ratios @xmath21 with respect to the expansion in @xmath4 . as we shall see in section 4 , in a general case @xmath280 holds , whereas , if the contribution in eq .
( [ eq : m_rl_hat_2 ] ) vanishes or is also suppressed , @xmath21 is proportional to @xmath13 .
the contribution @xmath281 is diagonal in flavour space . concerning the possible size of this contribution , note that @xmath282 is the relative magnitude of the non - vanishing elements of @xmath283 with respect to the corresponding ones in @xmath284
. notice finally that the ( 31 ) and ( 32 ) element of @xmath274 coincide .
we can compare our results with those found in @xcite , where the slepton mass matrices for a model possessing the same flavour symmetry were also estimated in a similar framework .
the main difference between the two setups is that in our model susy is a softly broken global symmetry , whereas in @xcite the model has been embedded into sugra .
we agree on the structure of the matrix @xmath285 ( @xmath286 in the notation of @xcite ) .
concerning the matrix @xmath263 ( @xmath287 in @xcite ) , we see that the off - diagonal matrix elements of @xmath287 are all of order @xmath6 , whereas we find that those of @xmath263 are of order @xmath4 .
such a discrepancy has a minor impact on the estimate of the rates for the radiative transitions , since , as we shall see in the next section , the rr block gives a subdominant contribution .
it is interesting to note that also in the sugra context analyzed in @xcite the vevs of the auxiliary components of the flavon supermultiplets give rise to @xmath288 , which corresponds to our @xmath289 , with similar implication on the rate of @xmath0 .
here we briefly comment on possible effects of the running from the scale @xmath290 at which the sfermion masses originate in our effective theory , down to the low energy scale , at which the amplitudes of lfv transitions are evaluated .
the renormalization group equations for the soft mass terms @xmath291 , @xmath292 and @xmath293/(v \cos \beta)$ ] , denoting @xmath294 , are @xcite : @xmath295 where @xmath296 are the gauge couplings and @xmath297 . ] of su(2)@xmath298 , @xmath299 the corresponding gaugino mass terms , @xmath300 are the yukawa matrices for charged leptons and down quarks , @xmath301/(v \cos \beta)$ ] and : @xmath302 the matrix @xmath303 coincides with @xmath304 and has the same evolution . for squarks
we have introduced soft mass terms analogous to those previously discussed for sleptons . to estimate the corrections to the slepton masses induced by the renormalization group evolution we adopt the leading logarithmic approximation and substitute each of the running quantities with their initial conditions at the scale @xmath305 in eqs .
( [ rge ] ) . in particular , for the matrices @xmath306 , @xmath307 and @xmath308 , these initial conditions are given in eqs .
( [ eq : m_ll_hat ] ) , ( [ eq : m_rr_hat ] ) and ( [ eq : m_rl_hat ] ) , ( [ eq : m_rl_hat_2 ] ) , respectively . in this approximation one
easily sees that the largest corrections to the matrices @xmath306 and @xmath307 come from electroweak gauge interactions and are proportional to the identity matrix in flavour space .
due to the negative sign of the dominant contribution these diagonal elements increase by evolving the mass matrices from the cutoff scale down to the electroweak scale .
this effect is taken into account in our numerical study presented in section 5 . and @xmath309 .
] each off - diagonal element of @xmath306 receives at most a relative correction of order : u ( ) , while those to @xmath307 are even more suppressed , i.e. u^2 ( ) .
all such contributions can be safely neglected .
the matrix @xmath308 gets a first correction by an overall multiplicative factor that can be absorbed , for instance , by a common rescaling of the parameters , and a second correction of the type @xmath310 , which can be absorbed by the redefinition @xmath311 , where @xmath190 and @xmath312 are constants . in our numerical study
these effects are treated in this way .
finally , additional corrections to the off - diagonal elements of @xmath308 are negligible .
we can conclude that the corrections induced by the rg running are either negligible or could be absorbed in our parametrization .
thus , in our model the soft mass terms of sleptons are completely controlled by the flavour symmetry and by its spontaneous breaking .
we recall that our model does not contain right - handed neutrinos and ignores the dynamics above the scale @xmath305 .
notice that the yukawa couplings of the charged leptons , @xmath313 , remain diagonal during the evolution . in a see - saw version
, one should also include the effects of the running from the cutoff scale down to the right - handed neutrino mass scale(s ) and the corresponding threshold effects . in this case
the previous conclusions might change , since we expect for generic order one contributions from rg running that they enter the amplitudes of the branching ratios at the level of @xmath314 , due to the loop suppression , which is ( roughly ) equal to a contribution of order @xmath6 in the amplitude in our context .
we can now evaluate the normalized branching ratios @xmath21 for the lfv transitions @xmath0 , @xmath1 and @xmath2 . in this section
we establish the leading dependence of the quantities @xmath21 on the symmetry breaking parameter @xmath4 .
we recall that in the class of models based on the flavour symmetry group @xmath3 an estimate based on an effective lagrangian approach suggests that @xmath21 generically scales as @xmath6 , which can be reconciled with the present bound on @xmath32 only if the scale @xmath19 of new physics is sufficiently large , above 10 tev .
the effective lagrangian approach also indicates that , under certain conditions , in the susy case a cancellation might take place , and @xmath21 might scale as a combination of two terms , one proportional to @xmath13 and one proportional to @xmath315 , as shown in eq .
( [ lfvsusy ] ) , thus allowing for a substantially smaller scale of new physics , in the range of @xmath316 tev . when @xmath32 is dominated by a one - loop amplitude with virtual particles of mass @xmath9 , @xmath19 and @xmath9 are roughly related by @xmath317 and a given lower bound on @xmath19 corresponds to a lower bound on @xmath9 one order of magnitude smaller .
here we determine the actual leading order behaviour of @xmath21 by expressing the result as a power series in the parameter @xmath4 . in order to do so
it is useful to first analyse the predictions in the so - called mass insertion ( mi ) approximation , where we have a full control of the results in its analytic form .
a more complete discussion based on one - loop results can be found in section 5 .
for the case at hand , the mi approximation consists in expanding the amplitudes in powers of the off - diagonal elements of the slepton mass matrices , normalized to their average mass . from the expression of the mass matrices of the previous section we see that in our case such an expansion amounts to an expansion in the parameters @xmath4 and @xmath82 , which we can directly compare with eq .
( [ lfvsusy ] ) . a common value in the diagonal entries of both ll and rr blocks is assumed and we consequently set @xmath318 and also @xmath319 in this section , so that the average mass becomes @xmath9 . on the contrary ,
no assumptions have been made for the trilinear soft terms , which keep the expression as in eqs .
( [ eq : m_rl_hat]-[eq : m_rl_hat_3 ] ) .
concerning chargino and neutralino mass matrices , they carry a dependence on the vector boson masses @xmath320 through off - diagonal matrix elements .
such a dependence is not neglected in this approximation , but only the leading order term of an expansion in @xmath320 over the relevant susy mass combination is kept . at the same time , to be consistent , we have to neglect the supersymmetric contributions of @xmath321 and @xmath322 and therefore @xmath321 and @xmath322 coincide . using these simplifications , the ratios
@xmath21 can be expressed as : r_ij= ( a_l^ij ^2+a_r^ij ^2 ) .
[ rij ] at the lo , the amplitudes @xmath323 and @xmath324 are given by : a_l^ij&=&a_ll ( _ ij)_ll + a_rl ( _ ij)_rl + a_r^ij&=&a_rr ( _ ij)_rr + a_lr ( _ ij)_lr [ alar ] where @xmath325 @xmath326 are dimensionless functions of the ratios @xmath327 , @xmath328 and of @xmath329 .
their typical size is one tenth of @xmath330 , @xmath331 being the su(2)@xmath332 gauge coupling constant . in our conventions their explicit expression
is given by : a_ll&= & + a_rl&=&^2_w 2 f_2n(a_1 ) + a_rr&=&^2_w + a_lr&=&^2_w 2 f_2n(a_1)[micoefficients ] where @xmath333 , @xmath334 and @xmath335 .
the functions @xmath336 and @xmath337 , slightly different from those in @xcite , are given by : f_1n(x)&=&(-17 x^3 + 9 x^2 + 9 x-1 + 6 x^2(x+3 ) x)/(24(1-x)^5 ) + f_2n(x)&=&(-5 x^2 + 4 x+1 + 2x(x+2)x)/(4(1-x)^4 ) + f_3n(x)&=&(1 + 9x -9x^2-x^3 + 6x(x+1 ) x)/(2(1-x)^5 ) + f_1c(x)&=&(-x^3 - 9x^2 + 9x+1 + 6x(x+1 ) x)/(6(1-x)^5 ) + f_2c(x)&=&(-x^2 - 4 x+5 + 2(2x+1)x)/(2(1-x)^4 ) .[mifunctions ] @xmath338 ij & w^{ll}_{ij } & w^{rl}_{ij } & w^{rr}_{ij } & w^{lr}_{ij}\\[10pt ] \hline & & & & \\[-9pt ] \mu e & \hat{n}_4 & \zeta ( c_f - c ) & 0 & 2 \ , \dd\frac{(z_e y_\mu - z_\mu y_e ) } { y_e y_\mu}\,~c + \zeta ( c_f - c)\\[3pt ] \hline & & & & \\[-9pt ] \tau e & \hat{n}_5 & \zeta ( c_f - c ) & 0 & 2 \ , \dd\frac{(z_e y_\tau - z_\tau y_e)}{y_e y_\tau } ~c+\zeta ( c_f - c)\ \\[3pt ] \hline & & & & \\[-9pt ] \tau \mu & \hat{n}_6 & \zeta ( c_f - c ) & 0 & 2 \,\dd\frac{(z_\mu y_\tau - z_\tau y_\mu)}{y_\mu y_\tau}~c+\zeta ( c_f - c ) \\[3pt ] \hline \end{array } $ ] notice that @xmath325 do neither depend on @xmath4 nor on the fermion masses @xmath339 .
finally , @xmath340 parametrize the mis and are defined as : ( _ ij)_cc= . from the mass matrices of the previous section , we find ( @xmath341 ) : ll ( _ ij)_ll = w^ll_ij u^2 , & ( _ ij)_rl= ( w^rl_ij u + w^rl_ij
u^2 ) + ( _ ij)_rr = w^rr_ij u , & ( _ ij)_lr = w^lr_ij u . [ deltas ] where for the mass insertion @xmath342 we have also displayed the nlo contributions , in order to better compare our results with those of the effective lagrangian approach .
the explicit expression for the lo coefficients @xmath343 are listed in table 2 .
also the nlo coefficients @xmath344 are dimensionless combinations of order one parameters . by substituting the mass insertions of eq .
( [ deltas ] ) into the amplitudes @xmath345 of eq .
( [ alar ] ) and by using eq .
( [ rij ] ) , we get : r^susy_ij= [ rsusy ] with @xmath346 and w^(0)_ij&= & a_rl w^rl_ij , + w^(1)_ij&= & ( a_ll w^ll_ij+a_rl
w^rl_ij ) , + w^(2)_ij&= & ( a_rr w^rr_ij+a_lr w^lr_ij ) .
[ wcoeff ] the behaviour displayed in eq .
( [ rsusy ] ) differs from the one expected on the basis of the effective lagrangian approach in the susy case , eq .
( [ lfvsusy ] ) .
this is due to the presence of the term @xmath347 .
assuming @xmath348 we recover what is expected from the effective lagrangian approach in the susy case , whereas when @xmath349 does not vanish , the lo behaviour matches the prediction of the effective lagrangian approach in the generic , non - supersymmetric case , eq .
( [ lfv ] ) .
as shown in table 2 , the coefficient @xmath349 is universal , namely it is independent from the flavour indices and it vanishes in two cases : * @xmath350 , which reflects the alignment of the vevs of the scalar and auxiliary components of the flavon supermultiplet @xmath55 , see eqs .
( [ vevs ] ) and ( [ vevsauxphit ] ) . * @xmath351 which can be realized by special choices of the soft susy breaking terms in the flavon sector , i.e. the assumption of universal soft susy breaking terms in the flavon potential . in our model none of these possibilities is natural , see @xcite , and both require a tuning of the underlying parameters . if @xmath348 , the result expected from the effective lagrangian approach in the susy case is obtained in a non - trivial way
indeed , it is a consequence of a cancellation taking place when going from the lagrangian to the physical basis . in particular , for @xmath348 the mi @xmath342 scales as @xmath352 and not as @xmath353 as we might naively guess by looking at eq .
( [ rlnsusy ] ) . as
a consequence @xmath27 scales as @xmath13 and not as @xmath6 for @xmath354 . in the general case when @xmath349 is non - vanishing , the dominant contribution to @xmath355 regarding the expansion in @xmath4 is flavour independent and , at the lo in the @xmath4 expansion , we predict @xmath31 , at variance with the predictions of most of the other models , where , for instance , @xmath356 can be much smaller than one @xcite .
if @xmath349 is non - vanishing , it is interesting to analyze the relative weight of the leading and sub - leading contributions to @xmath21 . for this purpose we list in table [ table_acc ] the expressions and the numerical values of the functions @xmath325 , in the limit @xmath357 .
@xmath358 a_{ll}&\dd\frac{1}{240}\frac{g^2}{16 \pi^2}\left[1 - 3\left(1 + 4\left(\dfrac{z_i}{y_i}+\zeta\right)\right)\tan^2\theta_w+4 \big(4 + 5\tan^2\theta_w\big)\tan\beta\right]&+(2.0\div16.3 ) \\ & & \\[3pt ] \hline & & \\[3pt ] a_{rl}=a_{lr}&\dd\frac{1}{12}\frac{g^2}{16 \pi^2}\tan^2\theta_w&0.30 \\ & & \\[3pt ] \hline & & \\[3pt ] a_{rr}&\dd\frac{1}{60}\frac{g^2}{16 \pi^2}\tan^2\theta_w\left[-3 - 3\left(\dfrac{z_i}{y_i}+\zeta\right)-\tan\beta\right]&-(0.5\div1.3 ) \\ & & \\[3pt ] \hline \end{array } $ ] as one can see , in this limit the dominant coefficient is @xmath359 , which is larger than @xmath360 by a factor @xmath361 , and larger than @xmath362 by a factor @xmath363 , depending on @xmath364 . assuming coefficients @xmath365 of order one in eqs .
( [ wcoeff ] ) , we see that the most important contributions in the amplitudes for the considered processes are @xmath366 and @xmath367 . the ratio between the sub - leading and the leading one is @xmath368 .
when @xmath4 is close to its lower bound , which in our model requires a small value of @xmath8 , the leading contribution clearly dominates over the sub - leading one .
however , for @xmath4 close to @xmath90 , which allows to consider larger values of @xmath369 , the non - leading contribution can be as large as the leading one and can even dominate over it .
the transition between the two regimes occurs towards larger values of @xmath4 .
the numerical dominance of the coefficient @xmath370 has also another consequence : for vanishing @xmath349 , @xmath21 is dominated by the contributions of @xmath371 , whose values are not universal , but expected to be of the same order of magnitude for all channels .
thus even when @xmath348 , we predict @xmath372 .
even if to illustrate our results explicitly we have taken the special limit @xmath357 , a numerical study confirms that they remain approximately correct when more generic regions of the parameter space of the model are considered .
why the results of the effective lagrangian approach in the susy case do not apply in the model under consideration ?
it is worth to summarize the main assumptions underlying the effective lagrangian approach in the susy case in @xcite : * the only sources of chirality flip are either fermion masses or sfermion masses of rl - type . *
both , fermion masses and sfermion masses of rl - type , up to the order @xmath6 , are dominated by the insertions of @xmath55 or @xmath373 in the relevant operators . as a consequence
also the operators describing dipole transitions , which have the same chiral structure as the charged lepton mass terms , are dominated by the insertions of @xmath55 or @xmath373 , up to the order @xmath6 .
for instance one such operator is : e^c h_d _ f^ ( l)+ ... [ old ] and similarly for @xmath61 and @xmath62 .
dots denote additional contributions such as those coming from the insertion of @xmath373 . here
@xmath374 is the electric charge .
indeed , in our model condition 2 is violated .
the explicit susy model considered here contains another set of fields which was not present in the effective lagrangian framework , namely the driving fields @xmath375 .
the driving fields represent an important ingredient of our model , since they are directly responsible for the vacuum alignment of the flavon fields . in the effective lagrangian approach
the alignment was postulated , without referring to a specific dynamical mechanism to generate it , and the driving fields were not included among the relevant degrees of freedom . in our model
@xmath375 has no direct coupling to matter fields .
such a coupling arises indirectly through the mediation of the flavon @xmath55 , which is coupled to matter via interactions suppressed by @xmath376 and to @xmath375 via interactions proportional to e.g. a large scale @xmath377 : w = x_e e^c h_d ( l ) + m_t ( _ 0^t _ t)+ ... @xmath377 is a mass scale of the order of the vev of @xmath55 , @xmath378 and dots denote further contributions .
it can be shown that the vev of @xmath375 is proportional to @xmath9 @xcite .
when the auxiliary fields of @xmath55 are eliminated slepton masses of rl - type receive a contribution from the insertion of @xmath379 : y_e ^c h_d ( m_t ) + ... similar terms for @xmath380 and @xmath381 are also present .
therefore dipole operators with the same type of insertions are expected in the low - energy limit .
up to loop factors : e^c h_d _ f^ ( m_t l)+ ... [ new ] since @xmath375 has a dominant vev of order @xmath9 and @xmath382 , this operator is similar in size to the operator of eq .
( [ old ] ) .
a similar contribution arises , if the coupling of two flavons @xmath55 to the driving field @xmath383 is taken into account instead of the term @xmath384 .
operators of this type were not included in the effective lagrangian approach , simply because the driving field @xmath383 was absent .
the set of operators of type as in eq .
( [ new ] ) and eq .
( [ old ] ) are aligned in flavour space only to the lo . at the nlo
such an alignment fails and this produces a non - vanishing term @xmath349 .
notice that the insertion of @xmath375 both in the slepton mass and in the dipole operators occurs through a combination of the type @xmath385 .
if we did not account for the large coupling of order @xmath377 , the mere insertion of @xmath386 in the mass and in the dipole operators would largely underestimate the effect .
it would be interesting to investigate the low - energy structure of dipole operators in a complete susy model in which the driving fields are absent and the alignment of flavon vacua is realized via some alternative mechanism .
this would allow to verify whether the failure of the effective lagrangian approach is entirely due to the presence of the driving fields or other subtleties occur when performing the low - energy limit .
in this section we perform a numerical study of the normalized branching ratios @xmath21 and of the deviation @xmath34 of the anomalous magnetic moment of the muon from the sm value .
we use the full one - loop results for the branching ratios of the radiative decays as well as for @xmath34 .
these can be found in @xcite and are displayed in appendix c for convenience . as discussed in the preceding sections , in our model the flavour symmetry @xmath387 constrains not only the mass matrices of leptons , but also those of sfermions .
these are given at the high energy scale @xmath388 , which we assume to be close to @xmath72 gev , the susy grand unification scale .
the flavour symmetry does not fix the soft susy mass scale @xmath9 .
it also does not constrain the parameters involved in the gaugino as well as the higgs(ino ) sector .
these are fixed by our choice of a sugra framework in which @xmath9 is the common soft mass scale for all scalar particles ( in the literature usually denoted as @xmath389 ) and @xmath10 the common mass scale of the gauginos .
thus , at the scale @xmath388 we have is not relevant in our analysis . ] m_1 ( _ l ) = m_2 ( _ l ) = m_1/2 .
effects of rg running lead at low energies ( at the scale @xmath390 of the @xmath391 mass ) to the following masses for gauginos m_1(m_w)m_1(_l)m_2(m_w)m_2(_l ) , where @xmath392 ( @xmath393 ) and according to gauge coupling unification at @xmath394 , @xmath395 . concerning the effects of the rg running on the soft mass terms , as we have seen in section 3.2 these are small or can be absorbed into our parametrization of the soft mass terms .
thus , in the contributions @xmath284 to the rl block we take @xmath9 as input parameter .
nevertheless , we explicitly take into account the rg effect on the average mass scale of the ll block , @xmath396 , and in the rr block , @xmath397 , rcl m_l^2(m_w ) & & m_l^2(_l)+0.5m_2 ^ 2(_l)+0.04m_1 ^ 2(_l ) m_susy^2 + 0.54 m_1/2 ^ 2 , + m_r^2(m_w ) & & m_r^2(_l)+0.15m_1 ^ 2(_l ) m_susy^2 + 0.15 m_1/2 ^ 2 .
the parameter @xmath217 is fixed through the requirement of correct electroweak symmetry breaking ||^2=-m_z^2 .
[ defmu ] since in the sugra framework the soft higgs mass parameters are also given by @xmath9 at the high energy scale , @xmath398 , the relation in eq .
( [ defmu ] ) reads at low energies determined by @xmath9 , @xmath10 and @xmath8 up to its sign .
we recall that in our model the low energy parameter @xmath8 is related to the size of the expansion parameter @xmath4 , the mass of the @xmath15 lepton and the @xmath15 yukawa coupling @xmath86 , as shown in eq .
( [ tanb&u&yt ] ) . for this reason
, requiring @xmath399 constrains @xmath8 to lie in the range @xmath400 .
as already commented , the lower bound @xmath401 is almost excluded experimentally , since such low values of @xmath8 usually lead to a mass for the lightest higgs below the lep2 bound of @xmath402 gev @xcite .
gev @xcite . ] in our numerical analysis the parameters are the following : the two independent mass scales @xmath9 and @xmath10 , the sign of the parameter @xmath217 and the parameters of the slepton mass matrices shown in section 3.1 in the physical basis .
we recall that the results of section 3.1 have been obtained under the assumption that the parameters are real and we keep working under the same assumption here .
we also assume that the parameters on the diagonal of the slepton mass matrices @xmath403 and @xmath263 , @xmath404 and @xmath233 , are positive in order to favour positive definite square - masses , to avoid electric - charge breaking minima and further sources of electroweak symmetry breaking .
the absolute value of the @xmath405 parameters is varied between @xmath406 and @xmath407 .
we will chose some representative values for @xmath4 in the allowed range @xmath408 .
the other expansion parameter @xmath82 is fixed to be @xmath90 . in the analysis of the normalized branching ratios @xmath21 we fix @xmath8 and @xmath4 and then we derive the yukawa couplings @xmath84 , @xmath85 and @xmath86 .
when discussing the anomalous magnetic moment of the muon instead we vary @xmath86 between @xmath409 and @xmath41 and calculate @xmath8 by using eq .
( [ tanb&u&yt ] ) . having determined @xmath8 ,
the yukawa couplings @xmath84 and @xmath85 can be computed .
the allowed region of the parameter space is determined by performing several tests .
we check whether the mass of the lightest chargino is above 100 gev @xcite , whether the lightest neutralino is lighter than the lightest charged slepton , whether the lower bounds for the charged slepton masses are obeyed @xcite and whether the masses of all sleptons are positive .
the constraint on the mass of the lightest chargino implies a lower bound on @xmath10 which slightly depends on the sign of @xmath217 . in our plots for @xmath21 we also show the results for points of the parameter space that do not respect the chargino mass bound . for low values of @xmath9 , e.g. @xmath410 gev , the requirement that the lightest neutralino is lighter than the lightest charged slepton is equivalent to the requirement that the parameters in the diagonal entries of the slepton mass matrices @xmath403 and @xmath263 are larger than one . for larger values of @xmath9 , e.g. @xmath411 gev
, this requirement does not affect our analysis anymore .
we note that masses of charginos and neutralinos are essentially independent from the @xmath405 parameters of the slepton mass matrices and thus their masses fulfill with very good accuracy ( better for larger @xmath10 ) m _ ^0_1 0.4 m_1/2 , m _ ^0_2
m _ ^-_1 0.8 m_1/2 , m _ ^0_3 m _ ^0_4 m _
^-_2 || . for the slepton masses
we find certain ranges which depend on our choice of the @xmath405 parameters .
we first discuss the results for the branching ratio of the decay @xmath7 .
this branching ratio is severely constrained by the result of the mega experiment @xcite r_e br(e ) < 1.2 10 ^ -11 and will be even more constrained by the on - going meg experiment @xcite which will probe the regime r_ebr(e ) 10 ^ -13 .
we explore the parameter space of the model by considering two different values of the expansion parameter @xmath4 , @xmath412 and @xmath413 , two different values of @xmath8 , @xmath98 and @xmath97 , as well as two different values of the mass scale @xmath9 , @xmath410 gev and @xmath411 gev .
we show our results in scatter plots in figure [ scatterbr ] choosing @xmath10 to be @xmath414 gev .
all plots shown in figure [ scatterbr ] are generated for @xmath415 .
as one can see from figure [ scatterbr](a ) , for very low @xmath98 , small @xmath412 , small @xmath410 gev the experimental upper limit from the mega experiment on @xmath416 can be passed in almost all parameter space of our model for values of @xmath10 as small as @xmath417 gev .
for @xmath410 gev and @xmath418 gev the sparticle masses are rather light : the lightest neutralino has a mass of @xmath419 gev , the lightest chargino of @xmath420 gev , the masses of the right - handed ( charged ) sleptons vary between @xmath419 and @xmath421 gev and the masses of the left - handed sleptons are in the range @xmath422 gev .
thus , especially the right - handed sleptons are expected to be detected at lhc . in a model also including quarks ( and hence squarks ) we find for the squarks that they can have masses @xmath423 gev and gluinos with masses of about @xmath11 gev , all accessible at lhc . to pass the prospective bound coming from the meg experiment in a sizable portion of parameter space of our model @xmath10 has to be chosen larger , @xmath424 gev .
then , however , the masses of the sleptons might only be detected at lhc in case of right - handed sleptons .
as one can see , values of @xmath425 gev are excluded due to the constraint on the lightest chargino mass . studying the same value of @xmath8 and @xmath4 , but taking @xmath9 to be as large as @xmath11 gev ,
we can see from figure [ scatterbr](b ) that now the bound set by the mega experiment on @xmath416 is respected in the whole parameter space of our model for all values of @xmath10 .
also the foreseen limit of the meg experiment can only exclude a smaller part of the parameter space of the model for all values of @xmath10 . in this setup , the prospects for detecting susy particles at lhc are the best for gauginos due to the possible low value of @xmath10 .
the slepton masses are expected to be roughly @xmath9 and thus too large to allow for a detection at lhc .
increasing the value of the expansion parameter @xmath4 from @xmath412 to @xmath413 , as done in figure [ scatterbr](c ) and [ scatterbr](d ) , increases also the branching ratio of the decay @xmath7 by approximately two orders of magnitude , since the different contributions to the branching ratio scale at least with @xmath6 , as analyzed in section 4 . for this reason for low values of @xmath410 gev
, @xmath10 has to take values @xmath424 gev in order for the result of @xmath416 to be compatible with the mega bound at least in some portion of the parameter space of our model .
for the point @xmath426 the sparticle spectrum is characterized as follows : the lightest neutralino has a mass of @xmath427 gev , the lightest chargino of @xmath428 gev , right - handed sleptons between @xmath429 and @xmath420 gev and left - handed sleptons generally above @xmath430 gev . for this reason there still exists the possibility to detect right - handed sleptons at lhc . concerning gluinos and squarks these
are expected to have masses between @xmath11 and @xmath431 gev so that they also can be detected at lhc . as one can see from figure [ scatterbr](c )
the on - going meg experiment can probe nearly the whole parameter space of the model for @xmath98 , @xmath413 and @xmath410 gev for values of @xmath414 gev .
increasing the parameter @xmath9 to @xmath11 gev shows that applying the existing bound on @xmath416 of @xmath432 can not exclude small values of @xmath10
. the situation changes , if the expected bound from the meg experiment is employed , because then values of @xmath10 smaller than @xmath11 gev become disfavoured .
finally , we show in figure [ scatterbr](e ) and [ scatterbr](f ) the results obtained for @xmath97 . we remind that this value is the largest possible one of @xmath8 in our model . requiring that the @xmath15 yukawa coupling does not become too large entails that @xmath97 fixes the expansion parameter @xmath4 to take a value close to its upper limit , @xmath413 .
the value of @xmath433 is thus enhanced through @xmath8 as well as @xmath4 .
this is clearly shown in figure [ scatterbr](e ) and [ scatterbr](f ) , because for a low value of @xmath410 gev already the mega bound practically excludes almost the whole parameter space of our model for all values of @xmath414 gev .
increasing the mass parameter @xmath9 to @xmath11 gev slightly improves the situation , because now there exists a marginal probability to pass the mega bound .
again , however , the meg experiment can probe all parameter space of our model for @xmath414 gev .
thus , for @xmath434 gev and @xmath414 gev the parameter space of our model is already severely constrained for moderate values of @xmath8 which entail large @xmath435 by the bound coming from the mega collaboration , but surely will be conclusively probed by the meg experiment .
choosing @xmath436 hardly affects the results presented here apart from slightly decreasing the lower bound on @xmath10 coming from the chargino mass bound .
thus , all statements made also apply for @xmath436 . in summary ,
the current bound on @xmath416 prefers regions in the parameter space of our model with small @xmath4 or small @xmath8 , as long as the sugra mass parameters should be chosen smaller than @xmath11 gev .
the foreseen meg bound strongly favours regions in which @xmath4 is small for @xmath9 and @xmath10 being not too large .
the fact that smaller values of @xmath4 are preferred has consequences also for the expectations of the detection prospects for the reactor mixing angle @xmath5 , because this angle scales with @xmath4 : it might thus not be possible to detect @xmath5 with the reactor and neutrino beam experiments under preparation @xcite .
concerning the radiative @xmath15 decays , @xmath437 and @xmath438 , the result found in the mi approximation that the branching ratios of these decays are of the same order of magnitude as @xmath416 is essentially confirmed in a numerical analysis . due to the random parameters differences up to two orders of magnitude
are expected and found , especially for the case of larger @xmath8 .
however , it is still highly improbable that the decays @xmath437 and @xmath438 could be detected at a superb factory , assuming a prospective limit of @xmath439 @xcite .
as is well known , the value found for the anomalous magnetic moment of the muon @xcite a_^exp=116592080(63)10 ^ -11 shows a 3.4 @xmath440 deviation a_=a_^exp - a_^sm=+302(88)10 ^ -11 [ da ] from the value expected in the sm @xcite a_^sm=116591778(61)10 ^ -11 .
similar discrepancies have been reported in @xcite .
thus , it might be interesting to consider the case in which this deviation is attributed to the presence of susy particles with masses of a few hundred gev .
the one - loop contribution to the anomalous magnetic moment of the muon in supersymmetric extensions of the sm has been studied by several authors @xcite .
we study the compatibility between the requirement that @xmath34 is explained by the exchange of relatively light susy particles and the experimental upper limit on @xmath416 coming from the mega experiment .
we choose again two different values of @xmath4 , @xmath412 and @xmath413 . to better explore the parameter space we vary the @xmath15 yukawa coupling between @xmath409 and @xmath41 and
fix the value of @xmath8 through the relation given in eq .
( [ tanb&u&yt ] ) . as a consequence in the plot for @xmath412 ,
see figure [ scatterbrg2](a ) , @xmath441 holds and for @xmath413 @xmath8 takes values @xmath442 .
similarly , @xmath9 and @xmath10 are chosen to lie in intervals @xmath443 $ ] and @xmath444 $ ] for @xmath412 and @xmath413 , respectively .
the different choice of intervals is due to the fact that values of a few hundred gev for @xmath9 and @xmath10 are disfavoured by the existing limit on @xmath416 when @xmath413 .
as one can clearly see from figure [ scatterbrg2 ] , in almost the whole parameter space of our model it is not natural to reproduce the observed deviation of the muon anomalous magnetic moment and at the same time to respect the existing bound on the branching ratio of @xmath0 .
this kind of incompatibility is well known in supersymmetric theories , because the explanation of the @xmath445 discrepancy necessitates small values of @xmath9 and @xmath10 and larger values of @xmath8 , which in turn enhance the branching ratio of the radiative lfv decays .
thus , we have to conclude that either there exist further sources of contributions to the anomalous magnetic moment of the muon beyond those present in our model , or - as is also discussed in the literature - the theoretical value @xmath446 found in the sm is closer to @xmath447 so that the eventual discrepancy becomes less than @xmath37 , a value which could well be explained in our model .
while awaiting the start of lhc and the first exploration of the tev scale , the solution to the hierarchy problem offered by susy still represents a very appealing option , with several interesting consequences , such as a successful gauge coupling unification , a viable particle candidate for dark matter and a possible explanation of the observed discrepancy in the anomalous magnetic moment of the muon .
nevertheless the existence of new states carrying flavour indices at the tev scale is a serious problem in model building , given the success of the sm in describing all known phenomena involving flavour transitions .
the lepton sector makes no exception . the flavour conversion observed in neutrino oscillations
might be related to potentially large effects in rare decays of the charged leptons .
a naive dimensional estimate of these effects , assuming new physics at the tev scale , requires a large suppression in the relevant amplitudes , to pass the existing bounds .
a possible explanation for such a suppression is the presence of a flavour symmetry , which is independently motivated by the hierarchy in the charged lepton mass spectrum and by the nearly tb mixing pattern of the lepton mixing matrix .
here we have considered a susy model invariant under the flavour symmetry group @xmath3 , originally proposed in order to describe lepton masses and mixing angles .
we have extended the original model by including susy breaking terms consistent with all symmetry requirements .
our model is an effective theory , valid at energy scales below a cutoff @xmath448 , where we have derived the spectrum of susy particles , in the slepton sector , under the assumption that the susy breaking scale is larger than @xmath449 .
it provides an example of a model in which the slepton mass matrices at the scale @xmath449 are not universal .
left - handed sleptons are approximately universal , with a small departure from universality controlled by @xmath4 , the flavour symmetry breaking parameter .
right - handed sleptons have soft masses of the same order , but the relative difference among them is expected to be of order one .
off - diagonal elements in both sectors , as well as in the rl block , are small and have a characteristic pattern in terms of powers of @xmath4 .
this structure is maintained by the effects coming from the rg running from @xmath448 down to the electroweak scale .
the symmetry breaking parameter @xmath4 lies in a restricted range around few per cent and has a size similar to the reactor mixing angle @xmath5 .
we have exploited the knowledge of the slepton mass matrices to compute the normalized branching ratios @xmath21 for the transitions @xmath0 , @xmath1 and @xmath2 . at variance with other models based on flavour symmetries we found @xmath372 and , given the present limit on @xmath32 , these rare @xmath15 decays are practically unobservable in our model . on a more theoretical side
, the scaling @xmath450 , found in the mi approximation , violates an expectation based on an effective lagrangian approach , which suggested @xmath451 in the limit of massless final charged lepton .
we have identified the source of such a violation in a single , flavour independent , contribution of the rl block of the slepton mass matrix .
such a contribution originates from the vevs of the auxiliary components of the flavon supermultiplets .
we have classified the conditions under which this universal contribution is absent . in a numerical analysis of @xmath32
we found that already the current bound from the mega experiment requires the parameter @xmath4 to be small or @xmath8 to be small for susy mass parameters @xmath9 and @xmath10 below @xmath11 gev , to guarantee detection of sparticles at lhc . applying
the prospective meg bound tightens the parameter space of our model even more to small @xmath4 and @xmath8 or requires mass parameters @xmath9 and @xmath10 above @xmath11 gev .
furthermore , we showed that the deviation of the experimentally observed value of the magnetic moment of the muon from the sm one can not be naturally explained in our framework , for @xmath416 below the current bound .
the maximal value of @xmath36 in our model is around @xmath37 for @xmath452 .
we thank lorenzo calibbi , stefano rigolin and fabio zwirner for many useful discussions .
we recognize that this work has been partly supported by the european commission under contracts pitn - ga-2009 - 237920 ( unilhc ) and mrtn - ct-2006 - 035505 .
the group @xmath16 is generated by two elements @xmath42 and @xmath43 obeying the relations@xcite : s^2=(st)^3=t^3=1 .
it has three independent one - dimensional representations , @xmath38 , @xmath39 and @xmath40 and one three - dimensional representation @xmath41 .
the one - dimensional representations are given by : lll 1&s=1&t=1 , +
1&s=1&t = e^i 4 /3 ^ 2 , + 1&s=1&t = e^i 2/3 .
+ [ s@xmath16 ] the three - dimensional representation , in a basis where the generator @xmath43 is diagonal , is given by : t= ( ccc 1&0&0 + 0&^2&0 + 0&0 & ) , s= ( ccc -1&2&22&-1&22&2&-1 ) .
[ st ] the multiplication rule for triplet representations is the following : 33=1 + 1+1+3_s+3_a if we denote by : a=(a_1,a_2,a_3 ) , b=(b_1,b_2,b_3 ) two triplets , the singlets contained in their product are given by : llll 1&(ab)&=&(a_1 b_1+a_2 b_3+a_3 b_2 ) + 1&(ab)&=&(a_3 b_3+a_1 b_2+a_2 b_1 ) + 1&(ab)&=&(a_2 b_2+a_1 b_3+a_3 b_1 ) [ dec2 ] the two triplets can be separated into a symmetric and an antisymmetric part : 3_s(ab)_s&=&(2 a_1 b_1-a_2 b_3-a_3 b_2,2 a_3 b_3-a_1 b_2-a_2 b_1,2 a_2 b_2-a_1 b_3-a_3 b_1)3_a(ab)_a&=&(a_2 b_3-a_3 b_2,a_1 b_2-a_2 b_1,a_3 b_1-a_1
b_3 ) [ dec3 ] moreover , if @xmath130 , @xmath453 and @xmath454 are singlets transforming as @xmath38 , @xmath39 and @xmath40 , and @xmath455 is a triplet , then the products @xmath456 , @xmath457 and @xmath458 are triplets explicitly given by @xmath459 , @xmath460 and @xmath461 , respectively .
note that due to the choice of complex representation matrices for the real representation 3 the conjugate @xmath462 of @xmath463 does not transform as @xmath41 , but rather @xmath464 transforms as triplet under @xmath16 .
the reason for this is that @xmath465 and @xmath466 where @xmath467 is the matrix which exchanges the second and third row and column .
we perform the following transformations on the superfields : the kinetic terms are canonically normalized , i.e. we find the basis in which the matrices @xmath190 and @xmath468 , shown in eqs .
( [ t],[tc ] ) , are unit matrices , and in a second step the charged lepton mass matrix @xmath208 , eq .
( [ yl_subleading ] ) , is diagonalized . as explained in the main text , we perform these transformations not only on the lepton fields , but also on the sleptons in order to ensure that the gaugino - lepton - slepton vertices do not violate flavour at this stage . to diagonalize the hermitian matrices @xmath190 and
@xmath468 we apply the unitary transformations @xmath391 and @xmath469 : @xmath470 normalizing @xmath190 and @xmath191 requires a rescaling of the fields via the real ( diagonal ) matrices @xmath53 and @xmath471 : @xmath472 the superfields @xmath59 and @xmath158 ( @xmath144 as above ) are expressed as : @xmath473 so that the kinetic terms for leptons and sleptons are in their the canonical form , i.e. eq . ( [ kt ] ) becomes : @xmath474 the mass matrices for leptons and sleptons in this basis read : @xmath475 and stands for both , the mass matrix of the left - handed charged sleptons and the one of the sneutrinos . ] @xmath476 we diagonalize the resulting mass matrix @xmath477 of the charged leptons by the usual bi - unitary transformation : @xmath478 and arrive at the mass eigenbasis @xmath479 and @xmath480 : @xmath481 finally , the slepton mass matrices @xmath482 , @xmath483 and @xmath484 are given as : @xmath485 \ , \tilde{l}^{\prime\prime } \equiv \tilde{l^c } ^{\prime\prime } \hat{m}_{rl}^2 \ , \tilde{l}^{\prime\prime}~~,\\ & & \bar{\tilde{l } } \ , m_{ll}^{2 } \ ,
\tilde{l } = \bar{\tilde{l}}^{\prime\prime } [ v^{\dagger } r w^{\dagger } \
, m_{ll}^2 \ , w r v ] \ , \tilde{l}^{\prime\prime } \equiv \bar{\tilde{l}}^{\prime\prime } \hat{m}_{ll}^2 \ , \tilde{l}^{\prime\prime}~~,\\ \nonumber & & \tilde{l^c } \
, m_{rr}^{2 } \ , \bar{\tilde{l^c } } = \tilde{l^{c \ , \prime\prime } } [ u^{t } r^c ( w^c)^{t } m_{rr}^2 \ , w^{c \ , \star } r^{c } \ , u^{\star } ] \ , \bar{\tilde{l^{c \ , \prime\prime } } } \equiv \tilde{l^{c \ , \prime\prime } } \hat{m}_{rr}^2 \ , \bar{\tilde{l^{c \ , \prime\prime}}}~~.\end{aligned}\ ] ] as we assume for the actual calculation of @xmath486 , @xmath487 and @xmath488 that all couplings involved are real , the matrices @xmath391 , @xmath489 , @xmath467 and @xmath109 turn out to be orthogonal instead of unitary .
furthermore , we express the small parameter @xmath4 in terms of @xmath82 as @xmath490 with @xmath491 according to eq .
( [ tbound ] ) and eq .
( [ ubound ] ) .
we then can do the calculation in just one expansion parameter @xmath82 . and
@xmath4 are recovered in the final result by replacing @xmath492 with @xmath493 . ] in the course of the calculation we pose the following requirements : the kinetic terms are canonically normalized up to and including @xmath494 , @xmath477 is diagonal also up to and including @xmath494 , and the matrices @xmath391 , @xmath489 , @xmath467 and @xmath109 are orthogonal up to the same order .
these calculations have been checked by using two independent methods .
in this part we fix the notations and we report the formulae for @xmath21 and for @xmath495 which we have used in section 5 .
the main references are @xcite and we follow these . ) .
this matrix is diagonalized by @xmath185 rotation matrices @xmath496 and @xmath497 as : o_r m_c o_l^t = diag ( m _ ^-_1 , m _
^-_2 ) , where the diagonalizing matrices connect mass and interaction eigenstates in the following way : ( ) [ mn ] we can diagonalize @xmath501 by a rotation matrix @xmath502 : o_n m_n o_n^t = diag ( m _ ^0_1 , m _ ^0_2 , m _ ^0_3 , m _
^0_4 ) , where @xmath502 connects mass and interaction eigenstates in the following way : ( ) where @xmath506 and @xmath507 are in general hermitian matrices and @xmath508 .
we diagonalize the mass matrix @xmath509 by a @xmath510 rotation matrix @xmath511 as : u^_e^2u^t = m_^2 where the mass eigenstates are : _ x = u^_x , i_i+u^_x , i+3|^c_i with masses @xmath512 ( @xmath513 ) .
analogously , the sneutrino mass matrix is diagonalized by : u^ _ ll^2u^
t = m_^2 where the mass eigenstates are : _ x = u^_x , i_i with masses @xmath514 ( @xmath515 ) .
+ + the normalized branching ratios , @xmath21 , for the lfv transitions @xmath516 are : r_ij==(|a_2^l|^2+|a_2^r|^2 ) and the decay rates are given by : ( l_il_j_i_j)=m_i^5 , ( l_il_j)=m_i^5(|a_2^l|^2+|a_2^r|^2 ) .
each coefficient @xmath517 can be written as a sum of two terms : a_2^l , r = a_2^(n)l , r+a_2^(c)l , r , where @xmath518 and @xmath519 stand for the contributions from the neutralino and from the chargino loops , respectively .
these coefficients are explicitly given by : a_2^(n)l = and @xmath520 with @xmath521 , and a_2^(c)l =- and @xmath522 with @xmath523 .
+ the terms @xmath524 and @xmath525 and the loop functions @xmath526 and @xmath527 read as follows : @xmath528 u^{\tilde{l}}_{x , i}+\dfrac{m_i}{m_w\cos{\beta}}(o_n)_{a,3}u^{\tilde{l}}_{x , i+3}\right\}\label{functions : niaxr}\\ [ -0.3cm]\nonumber\\ & & n_{iax}^r=-\dfrac{g_2}{\sqrt{2}}\left\{\dfrac{m_i}{m_w\cos{\beta}}(o_n)_{a,3}u^{\tilde{l}}_{x , i}+ 2(o_n)_{a,1}\tan{\theta_w}u^{\tilde{l}}_{x , i+3}\right\}\label{functions : niaxl}\\ [ -0.3cm]\nonumber\\ & & c_{iax}^l =- g_2(o_r)_{a,1}u^{\tilde{\nu}}_{x , i}\label{functions : ciaxr}\\ [ -0.3cm]\nonumber\\ & & c_{iax}^r = g_2\dfrac{m_i}{\sqrt{2}m_w\cos{\beta}}(o_l)_{a,2}u^{\tilde{\nu}}_{x , i}\label{functions : ciaxl}\end{aligned}\ ] ] and @xmath529\nonumber\\ & & g_{2n}(x_{ax})=\dfrac{1}{(1-x_{ax})^3}\left(1-x_{ax}^2 + 2x_{ax}\ln{x_{ax}}\right)\label{functions : g2n}\\ [ -0.3cm]\nonumber\\ & & g_{1c}(x_{ax})=\dfrac{1}{6(1-x_{ax})^4}\left(2 + 3x_{ax}-6x_{ax}^2+x_{ax}^3 + 6x_{ax}\ln{x_{ax}}\right)\label{functions : g1c}\\ [ -0.3cm]\nonumber\\ & & g_{2c}(x_{ax})=\dfrac{1}{(1-x_{ax})^3}\left(-3 + 4x_{ax}-x_{ax}^2 - 2\ln{x_{ax}}\right)\label{functions : g2c}\;.\end{aligned}\ ] ] + we note that the functions @xmath530 and @xmath531 , displayed in section 4 , are related to the loop functions @xmath526 and @xmath527 , mostly through taking the first derivative .
the deviation of the muon anomalous magnetic moment , @xmath495 , due to susy contributions can be written as : a__i^susy= with g__i^(n , c)=g__i^(n , c)l+g__i^(n , c)r .
these terms are explicitly given by : g__i^(c)l = and @xmath532 with @xmath533 , and g__i^(n)l =- and @xmath534 with @xmath521 .
the terms @xmath524 and @xmath525 and the functions @xmath526 and @xmath527 have been already introduced in eqs .
( [ functions : niaxr]-[functions : ciaxl ] ) and in eqs .
( [ functions : g1n]-[functions : g2c ] ) . for a review , see m. raidal _ et al .
_ , eur .
j. c * 57 * ( 2008 ) 13 [ arxiv:0801.1826 [ hep - ph ] ] .
a. strumia and f. vissani , arxiv : hep - ph/0606054 ; m. c. gonzalez - garcia and m. maltoni , phys .
* 460 * ( 2008 )
1 [ arxiv:0704.1800 [ hep - ph ] ] ; phys . lett .
b * 663 * ( 2008 ) 405 [ arxiv:0802.3699 [ hep - ph ] ] ; g. l. fogli _ et al . _ , nucl . phys .
* 168 * ( 2007 ) 341 ; a. bandyopadhyay , s. choubey , s. goswami , s. t. petcov and d. p. roy , arxiv:0804.4857 [ hep - ph ] ; g. l. fogli , e. lisi , a. marrone , a. palazzo and a. m. rotunno , phys .
* 101 * ( 2008 ) 141801 [ arxiv:0806.2649 [ hep - ph ] ] ; g. l. fogli , e. lisi , a. marrone , a. palazzo and a. m. rotunno , arxiv:0809.2936 [ hep - ph ] .
f. feruglio , c. hagedorn , y. lin and l. merlo , nucl .
b * 809 * ( 2009 ) 218 [ arxiv:0807.3160 [ hep - ph ] ] ; f. feruglio , c. hagedorn , y. lin and l. merlo , in the proceedings of no - ve 2008 , 4th international workshop on neutrino oscillations in venice ( venice , italy , 2008 ) , university of padua publication ( papergraf edition , padua , 2008 ) , 29 - 43 [ arxiv:0808.0812 [ hep - ph ] ] . s. f. king and m. malinsky , phys . lett .
b * 645 * ( 2007 ) 351 [ arxiv : hep - ph/0610250 ] ; f. feruglio , c. hagedorn , y. lin and l. merlo , nucl .
b * 775 * ( 2007 ) 120 [ arxiv : hep - ph/0702194 ] ; m. c. chen and k. t. mahanthappa , phys . lett .
b * 652 * ( 2007 ) 34 [ arxiv:0705.0714 [ hep - ph ] ] ; p. h. frampton and t. w. kephart , jhep * 0709 * ( 2007 ) 110 [ arxiv:0706.1186 [ hep - ph ] ] ; g. j. ding , phys .
d * 78 * ( 2008 ) 036011 [ arxiv:0803.2278 [ hep - ph ] ] ; p. h. frampton and s. matsuzaki , phys .
b * 679 * ( 2009 ) 347 [ arxiv:0902.1140 [ hep - ph ] ] .
e. ma and g. rajasekaran , phys .
d * 64 * ( 2001 ) 113012 [ arxiv : hep - ph/0106291 ] ; k. s. babu , e. ma and j. w. f. valle , phys .
b * 552 * ( 2003 ) 207 [ arxiv : hep - ph/0206292 ] ; m. hirsch , j. c. romao , s. skadhauge , j. w. f. valle and a. villanova del moral , arxiv : hep - ph/0312244 ; phys .
d * 69 * ( 2004 ) 093006 [ arxiv : hep - ph/0312265 ] ; e. ma , mod . phys . lett .
a * 17 * ( 2002 ) 627 [ arxiv : hep - ph/0203238 ] ; phys .
d * 70 * ( 2004 ) 031901 [ arxiv : hep - ph/0404199 ] ; new j. phys .
* 6 * ( 2004 ) 104 [ arxiv : hep - ph/0405152 ] ; arxiv : hep - ph/0409075 ; phys . rev .
d * 72 * ( 2005 ) 037301 [ arxiv : hep - ph/0505209 ] ; mod .
lett . a * 20 * ( 2005 ) 2601 [ arxiv : hep - ph/0508099 ] ; phys .
d * 73 * ( 2006 ) 057304 [ arxiv : hep - ph/0511133 ] ; mod .
a * 21 * ( 2006 ) 2931 [ arxiv : hep - ph/0607190 ] ; mod .
. lett .
a * 22 * ( 2007 ) 101 [ arxiv : hep - ph/0610342 ] ; s. l. chen , m. frigerio and e. ma , nucl .
b * 724 * ( 2005 ) 423 [ arxiv : hep - ph/0504181 ] ; k. s. babu and x. g. he , arxiv : hep - ph/0507217 ; a. zee , phys .
b * 630 * ( 2005 ) 58 [ arxiv : hep - ph/0508278 ] ; x. g. he , y. y. keum and r. r. volkas , jhep * 0604 * ( 2006 ) 039 [ arxiv : hep - ph/0601001 ] ; b. adhikary , b. brahmachari , a. ghosal , e. ma and m. k. parida , phys . lett .
b * 638 * ( 2006 ) 345 [ arxiv : hep - ph/0603059 ] ; l. lavoura and h. kuhbock , mod .
a * 22 * ( 2007 ) 181 [ arxiv : hep - ph/0610050 ] ; s. f. king and m. malinsky , phys .
b * 645 * ( 2007 ) 351 [ arxiv : hep - ph/0610250 ] ; s. morisi , m. picariello and e. torrente - lujan , phys .
d * 75 * ( 2007 ) 075015 [ arxiv : hep - ph/0702034 ] ; f. yin , phys .
d * 75 * ( 2007 ) 073010 [ arxiv:0704.3827 [ hep - ph ] ] ; f. bazzocchi , s. kaneko and s. morisi , jhep * 0803 * ( 2008 ) 063 [ arxiv:0707.3032 [ hep - ph ] ] ; f. bazzocchi , s. morisi and m. picariello , phys .
b * 659 * ( 2008 ) 628 [ arxiv:0710.2928 [ hep - ph ] ] ; m. honda and m. tanimoto , prog .
theor . phys .
* 119 * ( 2008 ) 583 [ arxiv:0801.0181 [ hep - ph ] ] ; b. brahmachari , s. choubey and m. mitra , phys . rev .
d * 77 * ( 2008 ) 073008 [ erratum - ibid .
d * 77 * ( 2008 ) 119901 ] [ arxiv:0801.3554 [ hep - ph ] ] ; f. bazzocchi , s. morisi , m. picariello and e. torrente - lujan , j. phys .
g * 36 * ( 2009 ) 015002 [ arxiv:0802.1693 [ hep - ph ] ] ; b. adhikary and a. ghosal , phys .
d * 78 * ( 2008 ) 073007 [ arxiv:0803.3582 [ hep - ph ] ] ; m. hirsch , s. morisi and j. w. f. valle , phys . rev .
d * 78 * ( 2008 ) 093007 [ arxiv:0804.1521 [ hep - ph ] ] ; y. lin , nucl . phys .
b * 813 * ( 2009 ) 91 [ arxiv:0804.2867 [ hep - ph ] ] ; c. csaki , c. delaunay , c. grojean and y. grossman , jhep * 0810 * , 055 ( 2008 ) [ arxiv:0806.0356 [ hep - ph ] ] ; p.
h. frampton and s. matsuzaki , arxiv:0806.4592 [ hep - ph ] ; w. grimus and l. lavoura , jhep * 0904 * , 013 ( 2009 ) [ arxiv:0811.4766 [ hep - ph ] ] ; f. bazzocchi , m. frigerio and s. morisi , phys .
d * 78 * , 116018 ( 2008 ) [ arxiv:0809.3573 [ hep - ph ] ] ; f. bazzocchi and s. morisi , arxiv:0811.0345 [ hep - ph ] ; s. morisi , phys .
d * 79 * ( 2009 ) 033008 [ arxiv:0901.1080 [ hep - ph ] ] ; p. ciafaloni , m. picariello , e. torrente - lujan and a. urbano , phys . rev .
d * 79 * , 116010 ( 2009 ) [ arxiv:0901.2236 [ hep - ph ] ] ; f. bazzocchi , l. merlo and s. morisi , phys .
d * 80 * ( 2009 ) 053003 [ arxiv:0902.2849 [ hep - ph ] ] ; m. c. chen and s. f. king , jhep * 0906 * , 072 ( 2009 ) [ arxiv:0903.0125 [ hep - ph ] ] ; y. lin , phys .
d * 80 * ( 2009 ) 076011 [ arxiv:0903.0831 [ hep - ph ] ] ; g. altarelli and d. meloni , j. phys .
g * 36 * , 085005 ( 2009 ) [ arxiv:0905.0620 [ hep - ph ] ] ; m. hirsch , s. morisi and j. w. f. valle , phys .
b * 679 * ( 2009 ) 454 [ arxiv:0905.3056 [ hep - ph ] ] . c.
i. low and r. r. volkas , phys .
d * 68 * ( 2003 ) 033007 [ arxiv : hep - ph/0305243 ] ; s. f. king , jhep * 0508 * ( 2005 ) 105 [ arxiv : hep - ph/0506297 ] ; i. de medeiros varzielas and g. g. ross , nucl .
b * 733 * ( 2006 ) 31 [ arxiv : hep - ph/0507176 ] ; j. matias and c. p. burgess , jhep * 0509 * ( 2005 ) 052 [ arxiv : hep - ph/0508156 ] ; e. ma , phys .
b * 632 * ( 2006 ) 352 [ arxiv : hep - ph/0508231 ] ; europhys .
* 79 * ( 2007 ) 61001 [ arxiv : hep - ph/0701016 ] ; phys . lett .
b * 660 * ( 2008 ) 505 [ arxiv:0709.0507 [ hep - ph ] ] ; s. luo and z. z. xing , phys . lett .
b * 632 * ( 2006 ) 341 [ arxiv : hep - ph/0509065 ] ; i. de medeiros varzielas , s. f. king and g. g. ross , phys .
b * 644 * ( 2007 ) 153 [ arxiv : hep - ph/0512313 ] ; phys .
b * 648 * ( 2007 ) 201 [ arxiv : hep - ph/0607045 ] ; n. haba , a. watanabe and k. yoshioka , phys .
* 97 * ( 2006 ) 041601 [ arxiv : hep - ph/0603116 ] ; p. kovtun and a. zee , phys . lett .
b * 640 * ( 2006 ) 37 [ arxiv : hep - ph/0604169 ] ; z. z. xing , h. zhang and s. zhou , phys .
b * 641 * ( 2006 ) 189 [ arxiv : hep - ph/0607091 ] ; s. f. king and m. malinsky , jhep * 0611 * ( 2006 ) 071 [ arxiv : hep - ph/0608021 ] ; c. s. lam , phys .
d * 74 * ( 2006 ) 113004 [ arxiv : hep - ph/0611017 ] ; phys .
b * 656 * ( 2007 ) 193 [ arxiv:0708.3665 [ hep - ph ] ] ; int . j. mod .
a * 23 * ( 2008 ) 3371 [ arxiv:0711.3795 [ hep - ph ] ] ; c. luhn , s. nasri and p. ramond , phys .
b * 652 * ( 2007 ) 27 [ arxiv:0706.2341 [ hep - ph ] ] ; y. koide , arxiv:0707.0899 [ hep - ph ] ; s. antusch , s. f. king and m. malinsky , jhep * 0806 * ( 2008 ) 068 [ arxiv:0708.1282 [ hep - ph ] ] ; i. de medeiros varzielas , arxiv:0804.0015 [ hep - ph ] . n. haba , r. takahashi , m. tanimoto and k. yoshioka , phys . rev .
d * 78 * ( 2008 ) 113002 [ arxiv:0804.4055 [ hep - ph ] ] ; s. w. li and b. q. ma , phys .
d * 77 * ( 2008 ) 093005 [ arxiv:0806.0670 [ hep - ph ] ] .
h. ishimori , y. shimizu and m. tanimoto , arxiv:0812.5031 [ hep - ph ] ; f. bazzocchi , l. merlo and s. morisi , nucl .
b * 816 * ( 2009 ) 204 [ arxiv:0901.2086 [ hep - ph ] ] .
g. altarelli , f. feruglio and l. merlo , jhep * 0905 * ( 2009 ) 020 [ arxiv:0903.1940 [ hep - ph ] ] .
y. lin , nucl .
b * 824 * ( 2010 ) 95 [ arxiv:0905.3534 [ hep - ph ] ] .
m. hamermesh , _ group theory and its application to physical problems _ , reading , mass.:addison - wesley ( 1962 ) 509 p ;
j. f. cornwell , _ group theory in physics : an introduction _ , san diego , usa : academic ( 1997 ) 349 p ; d. e. littlewood , _ the theory of group characters ( and matrix representations of groups ) _ , oxford university press ( 1958 ) 310 p ; a. d. thomas and g. v. wood , _ group tables _ , shiva publishing limited .
f. feruglio , c. hagedorn and l. merlo , arxiv:0910.4058 [ hep - ph ] .
p. f. harrison , d. h. perkins and w. g. scott , phys .
b * 530 * ( 2002 ) 167 [ arxiv : hep - ph/0202074 ] ; p. f. harrison and w. g. scott , phys .
b * 535 * ( 2002 ) 163 [ arxiv : hep - ph/0203209 ] ; z. z. xing , phys .
b * 533 * ( 2002 ) 85 [ arxiv : hep - ph/0204049 ] ; p. f. harrison and w. g. scott , phys .
b * 547 * ( 2002 ) 219 [ arxiv : hep - ph/0210197 ] ; phys .
b * 557 * ( 2003 ) 76 [ arxiv : hep - ph/0302025 ] ; arxiv : hep - ph/0402006 ; phys . lett .
b * 594 * ( 2004 ) 324 [ arxiv : hep - ph/0403278 ] . c. amsler
_ et al . _
[ particle data group ] , phys .
b * 667 * ( 2008 ) 1 .
s. schael _ et al . _
[ aleph collaboration and delphi collaboration and l3 collaboration and opal collaborations and lep working group for higgs boson searches ] , eur .
j. c * 47 * ( 2006 ) 547 [ arxiv : hep - ex/0602042 ] .
m. a. luty , arxiv : hep - th/0509029 . g. g. ross and o. vives , phys .
d * 67 * ( 2003 ) 095013 [ arxiv : hep - ph/0211279 ] .
g. g. ross , l. velasco - sevilla and o. vives , nucl .
b * 692 * ( 2004 ) 50 [ arxiv : hep - ph/0401064 ] ; s. antusch , s. f. king , m. malinsky and g. g. ross , phys .
b * 670 * ( 2009 ) 383 [ arxiv:0807.5047 [ hep - ph ] ] .
g. f. giudice and a. masiero , phys .
b * 206 * ( 1988 ) 480 .
h. ishimori , t. kobayashi , y. omura and m. tanimoto , jhep * 0812 * ( 2008 ) 082 [ arxiv:0807.4625 [ hep - ph ] ] ; a. hayakawa , h. ishimori , y. shimizu and m. tanimoto , phys .
b * 680 * ( 2009 ) 334 [ arxiv:0904.3820 [ hep - ph ] ] .
s. p. martin and m. t. vaughn , phys .
d * 50 * ( 1994 ) 2282 [ erratum - ibid .
d * 78 * ( 2008 ) 039903 ] [ arxiv : hep - ph/9311340 ] ; j. hisano , t. moroi , k. tobe and m. yamaguchi , phys .
d * 53 * ( 1996 ) 2442 [ arxiv : hep - ph/9510309 ] .
p. paradisi , jhep * 0510 * ( 2005 ) 006 [ arxiv : hep - ph/0505046 ] ; m. ciuchini , a. masiero , p. paradisi , l. silvestrini , s. k. vempati and o. vives , nucl .
b * 783 * ( 2007 ) 112 [ arxiv : hep - ph/0702144 ]
. v. cirigliano , b. grinstein , g. isidori and m. b. wise , nucl .
b * 728 * ( 2005 ) 121 [ arxiv : hep - ph/0507001 ] ; nucl .
b * 763 * ( 2007 ) 35 [ arxiv : hep - ph/0608123 ] ; s. davidson and f. palorini , phys .
b * 642 * ( 2006 ) 72 [ arxiv : hep - ph/0607329 ] .
k. hamaguchi , m. kakizaki and m. yamaguchi , phys .
d * 68 * ( 2003 ) 056007 [ arxiv : hep - ph/0212172 ] ; w. f. chang and j. n. ng , phys .
d * 71 * ( 2005 ) 053003 [ arxiv : hep - ph/0501161 ] ; a. mondragon , m. mondragon and e. peinado , phys .
d * 76 * ( 2007 ) 076003 [ arxiv:0706.0354 [ hep - ph ] ] ; s. antusch , s. f. king and m. malinsky , jhep * 0806 * ( 2008 ) 068 [ arxiv:0708.1282 [ hep - ph ] ] ; s. antusch and s. f. king , phys .
b * 659 * ( 2008 ) 640 [ arxiv:0709.0666 [ hep - ph ] ] ; l. calibbi , j. jones - perez and o. vives , phys . rev .
d * 78 * ( 2008 ) 075007 [ arxiv:0804.4620 [ hep - ph ] ] ; l. calibbi , a. faccia , a. masiero and s. k. vempati , phys . rev .
d * 74 * ( 2006 ) 116002 [ arxiv : hep - ph/0605139 ] ; r. dermisek , m. harada and s. raby , phys .
d * 74 * ( 2006 ) 035011 [ arxiv : hep - ph/0606055 ] .
t. moroi , phys .
d * 53 * ( 1996 ) 6565 [ erratum - ibid .
d * 56 * ( 1997 ) 4424 ] [ arxiv : hep - ph/9512396 ] ; s. p. martin and j. d. wells , phys .
d * 64 * ( 2001 ) 035003 [ arxiv : hep - ph/0103067 ] ; d. stockinger , j. phys .
g * 34 * ( 2007 ) r45 [ arxiv : hep - ph/0609168 ] ; a. czarnecki and w. j. marciano , phys .
d * 64 * ( 2001 ) 013014 [ arxiv : hep - ph/0102122 ] .
j. hisano , t. moroi , k. tobe and m. yamaguchi , phys .
d * 53 * ( 1996 ) 2442 [ arxiv : hep - ph/9510309 ] ; t. fukuyama , a. ilakovac and t. kikuchi , eur .
j. c * 56 * ( 2008 ) 125 [ arxiv : hep - ph/0506295 ] .
e. arganda and m. j. herrero , phys .
d * 73 * ( 2006 ) 055003 [ arxiv : hep - ph/0510405 ] .
j. hisano , t. moroi , k. tobe , m. yamaguchi and t. yanagida , phys .
b * 357 * ( 1995 ) 579 [ arxiv : hep - ph/9501407 ] ; j. hisano and k. tobe , phys .
b * 510 * ( 2001 ) 197 [ arxiv : hep - ph/0102315 ] ; g. isidori , f. mescia , p. paradisi and d. temes , phys .
d * 75 * ( 2007 ) 115019 [ arxiv : hep - ph/0703035 ] ; m. endo and t. shindou , arxiv:0805.0996 [ hep - ph ] .
r. barate _
[ lep working group for higgs boson searches and aleph collaboration and delphi collaboration and l3 collaboration and opal collaboration ] , phys .
b * 565 * ( 2003 ) 61 [ arxiv : hep - ex/0306033 ] .
lep higgs working group and aleph collaboration and delphi collaboration and l3 collaboration and opal collaboration , arxiv : hep - ex/0107030 .
m. l. brooks _ et al . _
[ mega collaboration ] , phys .
* 83 * ( 1999 ) 1521 [ arxiv : hep - ex/9905013 ] .
a. maki , aip conf .
* 981 * ( 2008 ) 363 .
t. schwetz , aip conf .
proc . * 981 * ( 2008 ) 8 [ arxiv:0710.5027 [ hep - ph ] ] . j. burguet - castell , d. casper , e. couce , j. j. gomez - cadenas and p. hernandez , nucl . phys .
b * 725 * ( 2005 ) 306 [ arxiv : hep - ph/0503021 ] ; p. huber , j. kopp , m. lindner , m. rolinec and w. winter , jhep * 0605 * ( 2006 ) 072 [ arxiv : hep - ph/0601266 ] ; j. e. campagne , m. maltoni , m. mezzetto and t. schwetz , jhep * 0704 * ( 2007 ) 003 [ arxiv : hep - ph/0603172 ] ; p. huber , m. lindner , m. rolinec and w. winter , phys .
d * 74 * ( 2006 ) 073003 [ arxiv : hep - ph/0606119 ] .
m. bona _ et al .
_ , arxiv:0709.0451 [ hep - ex ] ; a. g. akeroyd _ et al . _ [ superkekb physics working group ] , arxiv : hep - ex/0406071 .
m. passera , w. j. marciano and a. sirlin , phys .
d * 78 * ( 2008 ) 013009 [ arxiv:0804.1142 [ hep - ph ] ] ; m. passera , w. j. marciano and a. sirlin , aip conf . proc .
* 1078 * ( 2009 ) 378 [ arxiv:0809.4062 [ hep - ph ] ] . | we compute the branching ratios for @xmath0 , @xmath1 and @xmath2 in a supersymmetric model invariant under the flavour symmetry group @xmath3 , in which near tri - bimaximal lepton mixing is naturally predicted . at leading order in the small symmetry breaking parameter @xmath4 , which is of the same order as the reactor mixing angle @xmath5
, we find that the branching ratios generically scale as @xmath6 . applying the current bound on
the branching ratio of @xmath7 shows that small values of @xmath4 or @xmath8 are preferred in the model for mass parameters @xmath9 and @xmath10 smaller than @xmath11 gev .
the bound expected from the on - going meg experiment will provide a severe constraint on the parameter space of the model either enforcing @xmath12 and small @xmath8 or @xmath9 and @xmath10 above @xmath11 gev . in the special case of
universal soft supersymmetry breaking terms in the flavon sector a cancellation takes place in the amplitudes and the branching ratios scale as @xmath13 , allowing for smaller slepton masses .
the branching ratios for @xmath14 and @xmath2 are predicted to be of the same order as the one for @xmath0 , which precludes the possibility of observing these @xmath15 decays in the near future .
+ 2.5 cm * lepton flavour violation in a supersymmetric model + .3 cm with @xmath16 flavour symmetry * 0.5 cm ferruccio feruglio @xmath17 , claudia hagedorn @xmath18 , + .2 cm yin lin @xmath17 and luca merlo @xmath17 + .2 cm @xmath17 dipartimento di fisica ` g. galilei ' , universit di padova + infn , sezione di padova , via marzolo 8 , i-35131 padua , italy + .1 cm @xmath18 sissa , scuola internazionale superiore di studi avanzati + via beirut 2 - 4 , i-34014 trieste , italy + and + infn , sezione di trieste , italy 0.2 cm 2truecm |
the dynamical processes governing massive star formation are much less understood than is the case for low - mass stars @xcite . the various observational `` appearances '' of low - mass star formation ( t tauri stars ,
fu orionis stars , infrared class i - iii sources , molecular outflows , ... ) have been linked into a single , if rough , general evolutionary scenario @xcite .
in contrast , no clear evolutionary sequence has been established for high - mass stars .
objects such as the bn object in orion are highly embedded and emit the bulk of their luminosity in the mid- and far - infrared .
ultracompact ( uc ) regions are small ( @xmath24 pc ) sources of free - free emission at radio wavelengths @xcite .
`` hot core''-type objects have bright molecular line emission at submillimeter wavelengths , which indicate temperatures of several @xmath25 k and high abundances of saturated carbon - bearing molecules ( e.g. , blake et al .
they usually lack detectable radio emission , which might be due to `` quenching '' by material accreting at rates @xmath26 yr@xmath7 @xcite .
the distinction between uc regions and hot cores is not always clear , and the two are often found located very close to each other @xcite . as a step towards understanding the evolution of massive ysos , this paper presents models for a sample of fourteen such objects .
the formation of high - mass stars is invisible at optical wavelengths because of the high opacity of the surrounding material , so that reliable age estimates for massive protostellar objects are difficult to obtain
. such estimates for massive protostars have traditionally come from the morphology of the radio continuum emission @xcite , but this method applies only to relatively evolved objects .
more recently , the size to velocity ratio or dynamical time scale of molecular outflows has been used as a measure of age , but the derived age depends strongly on the adopted dynamical model @xcite .
the chemical composition of the material is a potentially powerful clock , but difficult to calibrate @xcite .
the density and temperature structure of the circumstellar material are also expected to change as the central star develops . for the density , power laws @xmath27
are predicted by many theoretical models .
before collapse begins , low - mass cores may relax to a power law , with @xmath28 for thermal support @xcite , and @xmath29 for turbulent support @xcite .
once collapse begins , the density distribution tends toward @xmath30 @xcite .
the temperature structure @xmath31 is determined by the luminosity , the dust opacities , and @xmath32 .
since the response of @xmath31 to changes in luminosity is rapid , we use the observed luminosity to determine @xmath31 . besides giving information on the dynamics of star formation , a good model of the physical structure of objects
is a prerequisite for determining their chemical composition , which then by itself can be used as a powerful additional evolutionary indicator .
the physical conditions around newly - born massive stars also reveal some of the influence that such stars have on their surroundings , which is of interest for the effects of star formation on large ( @xmath33 1 pc ) scales . the physical structure of massive yso envelopes has been studied based on dust continuum observations by , e.g. , @xcite , @xcite , @xcite and @xcite .
such studies are sensitive to the column density and the temperature of the envelope , but not to density itself . a major source of uncertainty for all dust models
is the choice of optical properties of the grains , which limits the accuracy of derived envelope masses to a factor of 2 @xcite .
molecular rotational lines are direct probes of the density and are also sensitive to temperature . @xcite and @xcite investigated the structure of massive star - forming regions using lines of cs , and other molecules with large dipole moments . however , in these studies , the column density ( or the cloud mass ) was not independently constrained .
doing this requires knowledge of the molecular abundance , information which is generally unavailable except in the case of co , which is relatively inert and much more abundant than most molecules .
still , part of the co may be frozen out onto grain surfaces in the cold regions far from embedded stars .
this paper uses observations of both the dust and the molecular gas to constrain the physical structure of the envelopes of a sample of massive young stars .
the study builds on the method of analysis developed in a previous paper for one source , gl 2591 , by @xcite .
the column density is inferred from continuum observations , and the density from molecular lines , while the temperature is calculated self - consistently from the luminosity , dust properties , and @xmath32 .
we use the derived density and temperature profiles to characterize the early evolution of massive stars and their surroundings .
the fourteen sources studied in this paper are introduced in table 1 , which lists the source names , positions , distances , luminosities , radio continuum flux densities and associated iras point source catalog ( psc ) entries .
all sources have been well studied before at radio and infrared wavelengths .
they were selected to be luminous ( @xmath34 ) and visible from the northern hemisphere .
the distances to some of the sources are quite uncertain ; in particular , for the sources in the cygnus x region , we use a fiducial distance of 1 kpc @xcite , for ease of scaling once better distances are available .
the main sample consists of nine deeply embedded massive young stars , which were additionally selected for mid - infrared brightness ( @xmath35 jy at @xmath36 ) , allowing comparison to existing absorption line studies .
a combined analysis of emission and absorption lines gives powerful constraints on the physical structure and geometry of the system , and allows a nearly complete inventory of the chemical composition of both the solid and the gas phases .
these nine sources have been extensively studied in the infrared , both from the ground @xcite and with the infrared space observatory @xcite . as a comparison sample ,
five luminous embedded young stars which are weak in the mid - infrared are studied , whose surroundings contain a `` hot core '' and/or an ultracompact region .
the full sample contains sources with luminosities between @xmath37 and @xmath38 and distances between @xmath39 and @xmath40 kpc .
the radio flux densities are at the @xmath41mjy level for most sources , which is orders of magnitude lower than expected from an region photoionized by a star with the same luminosity .
the radio emission from the bright mid - infrared sources has a spectral index @xmath42 in the range @xmath43 and may arise in a spherical wind for which @xmath44 .
the principal difference between our main sample and that of other studies of massive young stars ( e.g. , @xcite ) is the selection on mid - infrared brightness , which could introduce an orientation bias , in the sense that low - density cavities such as produced by molecular outflows may be preferentially directed towards us .
alternatively , the high mid - infrared brightness may be an evolutionary effect .
depending on the relative importance of different effects , the mid - infrared bright sources in the main sample could be younger or older than the sources in the comparison sample .
we will discuss this point in more detail in section [ sec : evol ] .
the sources were selected on isolated location within @xmath45 at infrared wavelengths , so that the heating is dominated by the central star . however , while gl 2591 and gl 490 are examples of objects forming in isolation , ngc 6334 , s 140 , ngc 7538 and w 3 are regions where low - mass stars are also born , as revealed by near - infrared images @xcite . in addition , s 140 contains three massive objects separated by @xmath46 , which contribute about equally to the far - infrared luminosity @xcite .
for the other sources , this information is not available . in the case of w 3 irs5
, radio continuum observations @xcite suggest the presence of seven early b - type stars within @xmath47 , a very large number in view of the average mass function and stellar density in young clusters @xcite . as long as the power source is much smaller than our beam
, its multiplicity does not matter for our analysis . in several cases ,
high - resolution continuum observations reveal the presence of an ( ultra-)compact region next to the source of interest .
since luminosity and/or mass are dominated by the infrared source , our models of centrally heated sources are still reasonable approximations .
spectroscopy of selected molecular lines in the @xmath48 , @xmath49 and @xmath50 ghz bands was performed with the 15-m james clerk maxwell telescope ( jcmt ) on mauna kea , hawaii during various runs in 19951997 . in total ,
15 - 20 lines were observed per source .
the antenna has an approximately gaussian main beam of fwhm @xmath51 at 230 ghz , @xmath52 at 345 ghz , and @xmath53 at 490 ghz .
detailed technical information about the jcmt and its receivers and spectrometer can be found on - line at @xmath54http://www.jach.hawaii.edu / jcmt / home.html@xmath55 .
receivers a2 , b3i and c2 were used as front ends at 230 , 345 and 490 ghz , respectively .
the digital autocorrelation spectrometer served as backend , with continuous calibration and natural weighting employed .
values for the main beam efficiency , @xmath56 , determined by the jcmt staff from observations of mars and jupiter , were @xmath57 , @xmath58 and @xmath59 at 230 , 345 and 490 ghz for the 1995 data , and @xmath60 , @xmath61 and @xmath59 for 1996 and 1997 .
absolute calibration should be correct to @xmath62 , except for data in the 230 ghz band from may 1996 , which have an uncertainty of @xmath63 due to technical problems with receiver a2 .
pointing was checked every 2 hours during the observing and was always found to be within @xmath64 .
integration times are 3040 minutes per frequency setting , resulting in rms noise levels in @xmath65 per @xmath66 khz channel ranging from @xmath67 mk at 230 ghz to @xmath68 mk at 490 ghz . to subtract the atmospheric and instrumental background , a reference position @xmath69 east was observed using the chopping secondary mirror . for the c@xmath0o @xmath70 line ,
we also position - switched using an @xmath71 offset , which increased the line fluxes typically by 15% .
the contribution by extended emission in other lines should be less since they all need higher densities for the excitation . where two measurements of c@xmath0o @xmath72 are available , the position - switched data will be used .
observations of the @xmath70 and @xmath73 and cs @xmath74 lines were carried out with the 30-m telescope of the institut de radio astronomie millimtrique ( iram ) on pico veleta , spain , on january 28 - 30 , 1999 .
receivers b100 , c150 and b230 were used at 96 , 145 and 245 ghz , respectively , tuned single sideband .
the autocorrelator was used as the backend , which covers a bandwidth of @xmath75 at a resolution of @xmath76 .
the fwhm beam sizes are @xmath77 , @xmath78 and @xmath79 ; the forward efficiencies were @xmath80 , @xmath81 and @xmath82 and the beam efficiencies @xmath83 , @xmath84 and @xmath85 .
the data were calibrated onto the @xmath65 scale by multiplying @xmath86 by the ratio of the forward and beam efficiencies . as a calibration check ,
observations of gl 2591 and ngc 7538 irs1 were compared to results from @xcite and found to agree to @xmath87% .
the data on gl 2591 have been presented in @xcite .
the observations on gl 490 are given in schreyer et al .
( 2000 , in prep . ) .
most of the data on w 3 ( ) and w 3 irs5 are taken from @xcite .
in addition , we used data from the surveys by @xcite , @xcite , @xcite , and from the observations of @xcite for gl 2136 , @xcite for s140 irs1 , @xcite for iras 20126 , @xcite and @xcite for dr 21 ( oh ) , dartois ( 1998 , ph.d .
thesis ) for gl 7009s and olmi & cesaroni ( 1999 , a&a , in press ) for w 28 a2 .
care was taken not to include data at positions @xmath88 away from those in table [ sam.tab ] .
for the sources gl 7009s , iras 20126 + 4104 and w 28 a2 , no data are available
. line parameters were measured by fitting a gaussian profile , where the free parameters were the total line flux @xmath89 , the fwhm line width and the center velocity .
the measured line fluxes are collected in table [ t : lineflux ] ; table [ vlsr.tab ] gives for each source the central lsr velocity and the fwhm line width .
the values are the averages of the c@xmath0o and lines ( 4 lines in most cases ) , which are assumed to be optically thin .
the error bars reflect the spread among the lines , which for all sources exceeds the error in the individual measurements for both quantities .
none of the cs and h@xmath19co line profiles show obvious evidence of self - absorption .
maps of the cs @xmath90 and/or @xmath91 and/or c@xmath92o @xmath72 lines were obtained in 1998 july and december and 1999 july with the @xmath93-m antenna of the caltech submillimeter observatory ( cso ) .
the backends were the facility 500 and 50 mhz bandwidth acousto - optical spectrometers ( aos ) .
the pointing is accurate to within @xmath47 .
beam sizes and main beam efficiencies were @xmath94 and 0.66 at 230 ghz and @xmath95 and 0.61 at 345 ghz . in july 1998 ,
submillimeter continuum maps at @xmath96 of w 33a , gl 2136 , s 140 , gl 490 , gl 2591 and w 3 irs5 were made using the submillimeter high angular resolution camera ( sharc ) @xcite on the cso .
the cso has a beam of size @xmath97 at this wavelength .
the weather was good , with zenith optical depths at @xmath96 of @xmath98 . from observations of uranus
, the gain was measured to be @xmath99 jy / v .
maps at @xmath100 ghz of gl 2136 , ngc 7538 irs1 , ngc 7538 irs9 and w 33a were obtained with the six - element interferometer of the owens valley radio observatory ( ovro ) .
the ovro interferometer consists of six 10.4 m antennas on north - south and east - west baselines .
a detailed technical description of the instrument is given in @xcite .
data were collected in 1997 - 1999 in several array configurations at frequencies of @xmath101 , @xmath102 and @xmath103 ghz .
the two sources in ngc 7538 were observed in the compact and extended array configurations , while for the southern sources gl 2136 and w 33a , a hybrid configuration with long north - south and short east - west baselines was also used to improve the beam shape .
baseline lengths range from the shadowing limit out to @xmath104 at @xmath101 ghz and out to @xmath105 at @xmath48 ghz , corresponding to spatial frequencies of @xmath106 to @xmath107 rad@xmath7 .
the antenna gains and phases were monitored with short integrations of nearby quasars : bl lac for ngc 7538 irs1 and irs9 , and nrao 530 for w 33a and gl 2136 .
passband corrections were derived using a local signal generator and by fitting first order polynomials to data on 3c273 .
flux calibration is based on snapshots of uranus and neptune , and is estimated to be accurate to @xmath67 % at @xmath108 ghz and to @xmath109 % at @xmath48 ghz .
simultaneous continuum observations in the @xmath48 ghz window were also performed , but only in the cases of w 33a and ngc 7538 irs1 did the weather allow useable data to be taken .
in addition , observations of molecular lines at @xmath100 ghz were carried out with the ovro digital correlator , the results of which will be presented elsewhere .
the cs line profiles are presented in figure [ f : jcmt_res ] .
in addition to a strong , single - peaked line core , which is also detected in isotopic lines ( , ) and which has an approximately gaussian shape , high - velocity wings are clearly detected .
these wings are more prominent in @xmath74 than in higher-@xmath110 lines , are also detected in the strongest lines , and must arise on small scales , since they are stronger in the iram 30 m beam than in the jcmt beam .
the wings are not detected in the w 3 sources , for which we do not have @xmath111 data .
since the sources studied in this paper are known to drive co outflows , it is natural to associate the cs and wings with dense gas in these outflows . in all bright mid - infrared sources studied in this paper except w 33a ,
blueshifted infrared absorption of co is detected at similar velocities as in co rotational line emission @xcite , and in the cases of w 3 irs5 and gl 2591 up to much higher velocities , implying that the outflow is directed along the line of sight . in realistic outflow models with a finite opening angle , a mixture of blue- and redshifted emission is expected on both sides of the driving source , which should be visible in submillimeter emission .
however , the wings seen on the cs and lines are much stronger at blueshifted than at redshifted velocities , and for some sources no redshifted wing emission is detected at all . since the line asymmetry is stronger in the iram 30 m spectrum of cs @xmath74 than in the jcmt profile of the same line , the asymmetry must arise on scales of @xmath112 . since the redshifted outflow lobe lies in the background
, we suggest that the lack of redshifted co and cs wing emission is due to obscuration . since the outflow lies at a velocity offset , the absorption must be continuous absorption by dust . the column density required to do
this is @xmath113 on a @xmath112 scale , corresponding to a visual extinction of @xmath114 magnitudes .
estimates of the mean temperature and density of the gas can be obtained by comparing observed line ratios of specific molecules with non - lte models which include radiative trapping .
examples of this approach can be found in @xcite and @xcite ; it is similar to the `` large velocity gradient '' method used by @xcite and @xcite
. the observed line ratios of are well - suited to constrain the density .
we have calculated synthetic line ratios for a range of temperatures and densities , using the rate coefficients by s. green ( priv .
cf.@xcite ) .
comparison with the data indicates densities of @xmath115 to @xmath116 , but different line ratios generally give inconsistent answers for the same source , indicating the presence of density inhomogeneities , probably in the form of a gradient since lines observed with several telescopes are systematically brighter when observed with a smaller beam .
the measured ratios of the @xmath117/@xmath118 and the @xmath119/@xmath117 lines are particularly sensitive to temperature . unlike , these line ratios were measured in the same beam , or even in the same spectrum , improving the calibration between pairs of lines , although their filling factors might still be different .
model calculations using collisional rate coefficients by @xcite give temperatures of @xmath120 k for s 140 irs1 to @xmath121 k for w 3 ( ) and ngc 6334 irs1 . again
, the two ratios usually do not yield the same temperature for the same source , indicating the need for models with a varying temperature .
in general , the modeling indicates somewhat higher temperatures and densities for the weak mid - infrared sources . in section [ s : models ] , we will see that this effect is due to a steeper density gradient in these sources .
the co column densities derived from c@xmath0o , using the temperatures and densities found above , are listed in column 3 of table [ vlsr.tab ] .
a plot of these values against the column densities derived by @xcite from infrared absorption observations is shown in figure [ f : coldens ] . abundance ratios of @xmath122c/@xmath123c = 60 and @xmath124o/@xmath0o = 2500 @xcite are assumed .
the two measurements agree to a factor of 2 for all sources , and often much better .
this result implies that the circumstellar envelopes of our objects have to first order a spherical shape , with the infrared source in the center ( see also section [ sec : struc ] ) .
the average absorption column density is @xmath125% of the emission value , which is higher than half as expected in a uniform model .
beam dilution in the emission data could explain this difference , in which case the sources are centrally condensed .
the cs and c@xmath92o maps are presented in figure [ f : lmaps ] ; figure [ f : cmaps ] shows the sharc @xmath126 maps and maps of ngc 7538 at @xmath127 , obtained with the submillimeter common - user bolometer array ( scuba ) on the jcmt , provided by g. sandell ( 1998 , private communication ) .
the maps appear compact but slightly extended . in a few cases ,
the map peak is offset from the center position , but this offset falls within the @xmath64 pointing uncertainty .
in the case of cs , this suggests that the infrared sources coincide with local density maxima in the surrounding molecular cloud .
the c@xmath92o maps are sometimes peaked at the same position as cs , which implies a maximum of the column density , but more often , these maps are not strongly peaked and have a much lower dynamic range than the cs maps .
the map of ngc 7538 shows a chaotic structure , rather than clear peaks .
column 4 of table [ vlsr.tab ] lists the diameter of the cs @xmath111 emission , measured from the maps as the point where the brightness has dropped to 50% of the maximum .
these numbers will be used in the next section to constrain the radii of the models .
the values span a fairly narrow range : @xmath128 and do not show an obvious correlation with distance , suggesting that the cores do not have a single intrinsic linear size .
instead , it appears that the molecular gas has a scale - free density structure , such as can be described by a power law .
this conclusion is supported by the fact that the cs diameters are somewhat ( @xmath23% ) larger than the beam size , which is also what power law models predict .
the maps do not show a cutoff in the emission , such as would be produced if the star - forming cores had a distinct edge .
rather , the brightness keeps going down until the detection limit .
maps of a wider field and with a higher dynamic range than presented here may reveal if there is such an edge or if at large radii , the density gradient flattens out and the core merges into a surrounding molecular cloud . some objects , such as w 3 irs5 , dr 21 ( oh ) , ngc 6334 irs1 and the ngc 7538 sources , are clearly part of a more extended giant molecular cloud .
however , the focus of this paper is on the @xmath45 region around the massive young stars .
figure [ fig : ovromaps ] presents maps created from the ovro data by gridding and fourier transforming the visibility data with uniform weighting .
deconvolution with the clean algorithm and self - calibration of the @xmath129 phases based on the brightest clean components helped to improve the image quality .
table [ posflux.tab ] lists the positions and flux densities of the detected sources .
the maps typically show a single compact source at or near phase center , within the @xmath130 positional error of the infrared positions from table [ sam.tab ] .
the extended @xmath131 ghz emission north of ngc 7538 irs1 is the region surrounding ngc 7538 irs2 , which has also been detected at lower frequencies with the vla ( e.g. , @xcite ) . in the case of
w 33a , two sources are detected , neither of which coincides with the ir position of @xcite .
the brightest source , mm1 , coincides however with a vla detection @xcite and with the infrared position by @xcite , who also report a second @xmath132 source @xmath133 south of w 33a , which may be related to mm2 .
it is unknown if these two sources are physically associated ; if they are , they may represent a young massive binary star at a separation of @xmath134 or @xmath135 au . this separation is larger than the @xmath136 au found by @xcite for w 3 ( ) and by @xcite for a sample of low - mass young stellar objects , but comparable to that found by @xcite and @xcite for dr 21 ( oh ) .
the flux densities found here for ngc 7538 irs1 are in agreement with those by @xcite and @xcite .
the spectral index , measured from 107 to 230 ghz , is @xmath137 .
the spectral indices of the sources in w 33a are @xmath138 ( mm1 ) and @xmath139 ( mm2 ) .
the values of the spectral indices for w 33a rule out optically thin emission from either ionized gas or from dust .
they are , however , consistent with black body emission , probably from a compact dust source . in paper 1 ,
ovro observations of gl 2591 were presented which gave similar results ; we refer the reader to that paper for a detailed discussion .
motivated by the results of the excitation analysis ( [ s : est_phys ] ) and of the cs maps ( [ s : res_maps ] ) , we will proceed by developing spherically symmetric power law models .
the modeling procedure follows @xcite to which we refer for details .
the dust emission from the sources was modeled using the one - dimensional diffusion code by @xcite .
a density structure of the form @xmath140 was used , with @xmath141 in the range @xmath142 .
the fiducial radius @xmath143 was set to be half the diameter of the cs emission ( @xmath144 in table [ vlsr.tab ] ) , converted to au in table [ t : m_gas ] .
the density @xmath145 at this radius was derived for each @xmath141 by matching the submillimeter dust emission ; in the next section , we will constrain the value of @xmath141 by modeling molecular lines . to avoid edge effects in modeling the full range of emission
, we used outer radii in our models that are twice @xmath143 .
the inner radius was arbitrarily set to 1/300 of the outer radius , small enough not to influence the calculated brightness or temperature profile , as verified by test calculations for w 33a and gl 2591 .
dust opacities appropriate for star - forming regions were taken from @xcite ( their model 5 ) and converted to absorption cross - sections using a grain mass density of @xmath146 g and a radius of @xmath147 .
this radius is close to the median of more realistic size distributions , so that the calculated dust temperatures represent the bulk of the dust ( cf .
dust properties are known to vary from one region to the next @xcite and are likely to change inside our sources as well , due to the changing temperature ( @xcite ; see [ s : ice ] ) .
maps at several far - infrared wavelengths at @xmath112 resolution would allow us to disentangle temperature and dust opacity variations , but awaiting such data , we assume a constant grain opacity . with the luminosity , the distance and the size of the source fixed , was used as the only free parameter to match the observed submillimeter continuum emission .
figure [ f : dusto ] shows the synthetic continuum spectra compared with observations .
most submillimeter data are from this work and from @xcite ; the sources of the additional data are listed in the caption .
the models have been constructed to fit the submillimeter ( @xmath148 ) data , and are seen to reproduce the observed emission at @xmath149 for every source , i.e. up to optical depths of a few .
the shorter - wavelength emission is in general not matched , although care was taken to include only photometry in small ( @xmath112 ) beams to avoid a contribution from reflection nebulosities .
the high brightness of gl 2591 at @xmath150 was attributed by @xcite to the evaporation of ice mantles close to the star , which decreases the @xmath150 optical depth by 30% @xcite .
similar effects play a role for the sources presented here , as illustrated for w 33a in the figure . the emission at @xmath151 is not expected to be well reproduced in these models , since the high optical depth makes it very sensitive to deviations from the assumed spherical shape .
the fact that all sources require less extinction at short wavelengths is discussed quantitatively in [ sec : struc ] .
this point is illustrated by the fact that the column densities derived in this paper are much higher than those found by @xcite by fitting the near- to far - infrared spectra of a sample of ultracompact regions , which have similar submillimeter flux densities and lie at similar distances .
the calculated temperatures follow an @xmath152 profile at distances greater than @xmath153 au from the star , with the absolute temperature scale set by the luminosity . inside this radius ,
the envelope is optically thick to the photons carrying most of the energy , and the temperature gradient is steeper than @xmath152 , with the absolute scale set by the extinction , which acts as a blanket .
hence , for a given luminosity , a lower extinction leads to a higher temperature at a certain radius .
this is the case for our models with shallow density gradients . for a given column density
, a shallower density gradient implies a lower extinction ( in a pencil beam ) , because a larger fraction of the beam is filled with warm dust .
we therefore expect , like @xcite and @xcite , that bright mid - infrared sources have shallow density gradients .
however , it is seen from figure [ f : dusto ] that even the models with @xmath154 fail to fit the near - infrared part of the spectrum .
this is not due to our selection on brightness in the mid - infrared since the data of the comparison samples are not matched either .
the reason for the discrepancy must be sought in ( small ) geometrical effects . to determine the slope of the density gradient @xmath141
, the c@xmath0o , cs , and line radiation from the 14 sources was modeled with the monte carlo code written by @xcite .
the models consist of 30 spherical shells spaced logarithmically within the same radii as the dust models .
the same density structures were used as in the dust models , and the temperature structure was taken from the dust continuum models for the same value of @xmath141 .
the intrinsic line profile was assumed to be a gaussian with the measured fwhm ( table [ vlsr.tab ] ) .
initially , molecular abundances were assumed to be constant throughout the model .
the gas column density ( or equivalently or the value of @xmath145 ) for each @xmath141 was also taken from the dust models , assuming a gas to dust mass ratio of 100 .
this assumption was tested for gl 2591 against c@xmath0o observations by @xcite because this source has negligible depletion of co ( see also
[ s : ice ] ) . with
the column density fixed , the density profile can be obtained by modeling the emission lines of the high - critical density molecules cs and . after solving for the molecular excitation ,
velocity - integrated intensities have been calculated in the appropriate beam for each observation . in many cases ,
a line was observed in several beams , which in our models acts as a simple substitute for simulating map data .
comparison to the observations proceeds by minimizing the quantity @xmath155 where @xmath156 and @xmath157 are the observed and synthetic line fluxes and the sum is taken over all @xmath158 observed lines .
a grid of models was run for each molecule by varying the density law exponent @xmath141 and the molecular abundance .
the success of any fit procedure depends critically on a good understanding of the error budget ( @xmath159 ) .
most line fluxes were assigned a 30% error , but the following lines have a larger uncertainty of 50% : ( i ) lines in the 460 - 490 ghz band , because of the more uncertain calibration ; ( ii ) lines measured in beams of @xmath160 diameter , which may suffer from confusion ; and ( iii ) the @xmath161 lines of cs and , which may contain a contribution from the surrounding cloud .
the following observations were found to be more than a factor of 2 off any of our models , and therefore discarded in the @xmath162 calculation : cs @xmath90 and @xmath91 in gl 490 and cs @xmath91 in w 33a from @xcite , cs @xmath163 and @xmath164 in s140 irs1 from @xcite , @xmath165 in iras 20126 + 4104 from @xcite and @xmath166 in w 28 a2 from @xcite .
the results of the emission line models are presented in figure [ f : chi ] .
the fit quality parameter @xmath162 is plotted as a function of the density law exponent @xmath141 and the cs abundance with respect to .
the elongation of the @xmath162 contours indicates that the quality of the fit depends to first order on the cs abundance and to second order on the density profile .
the parameters of the best fitting models are summarized in table [ t : m_gas ] .
it is seen from table [ t : m_gas ] that for the main sample , @xmath167 , while the other sources have @xmath168 .
this result is consistent with the result from the dust models ( [ s : mod_dust ] ) , namely that for a given source , the model with lowest @xmath141 has the highest mid - infrared flux .
thus , the mid - infrared brightness of the main sample is not a pure orientation effect .
however , since the spherical dust models fail to reproduce the near - infrared emission of all sources , deviations from spherical symmetry are important .
this conclusion is supported by synthetic cs line profiles from our power law models , which are self - absorbed for all but the least massive sources .
the effect of deviations from spherical symmetry on line profiles and near - infrared emission was discussed for gl 2591 in @xcite .
decreasing the central column density by a factor of a few can increase the mid - infrared emission and decrease the self - absorption substantially .
the best - fit cs abundance is @xmath16 on average , with source - to - source variations of a factor of 3 for most sources .
the hot cores and the ultracompact region have the highest cs abundances : @xmath18 . the low cs abundance in gl 7009s , @xmath169 ,
is probably due to freezing out of the molecules onto the dust grains , as suggested by the large column densities of several ice species towards this source @xcite .
figure [ f : chi ] also lists the number of cs and lines observed .
the sources for which the most lines have been observed also have the tightest constraints on @xmath141 . for less well - observed sources ,
our results may be influenced by sampling effects .
in particular , if only a small range of molecular energy levels is available , the value of @xmath141 may be biased , or a gradient may be hard to detect .
we have checked for such effects by calculating @xmath170 , the average upper j - level of the data set .
most sources have @xmath171 , which implies that the observable range of critical densities has been evenly sampled .
however , for gl 7009s , and iras 20126 , for which fewer high - excitation line data exist , @xmath172 , and the model results are therefore less robust .
the value @xmath173 found for the hot cores is higher than for the other sources , and we have investigated if this result could be an artifact of our assumption of a constant molecular abundance .
hot cores have high abundances of many molecules including saturated organics , which are thought to be due to freshly evaporated grain mantles . in this picture
, the radius where the temperature reaches 90 k is an important threshold because water ice , the main ingredient of interstellar grain mantles , evaporates .
it is possible that the abundance of cs is enhanced above this temperature as well . to test this idea
, a model has been run with the luminosity and the of w 3 ( ) , but with @xmath30 .
the cs abundance is @xmath174 in the outer parts , but enhanced where the dust temperature exceeds @xmath175
k. a match to the data of equal quality as with the best constant - abundance model ( @xmath28 , cs / h@xmath176 : table [ t : m_gas ] ) was obtained by enhancing the cs abundance in the inner region by a factor of 10 .
this result indicates that the hot cores may have a density structure similar to that of the main sample , if plausible variations of cs/ with radius are considered .
scuba maps of other `` hot core''-type sources by hatchell et al .
( 2000 , submitted to a&a ) indeed suggest values of @xmath177 .
the density structures derived from the cs excitation will now be tested by using them to model the radial profiles of the submillimeter dust emission .
this emission is optically thin and hence measures the column density distribution , given the calculated temperature structure .
the points in figure [ fig : dust_prof ] are averages of slices along the north - south and east - west directions through the images shown in figure [ f : cmaps ] . to reduce the noise ,
the data have been folded about the image maximum . in the case of w 3 irs5
, the local maximum in the sharc map that corresponds to the infrared position was used to center the slices , and the western direction was not included in the scans because of confusing extended emission clearly visible in figure [ f : cmaps ] .
superposed are slices through model images for various values of @xmath141 as dotted lines , while the solid line corresponds to the value of @xmath141 derived in
[ s : mod_gas ] from the cs excitation .
the model profiles ( in a @xmath97 beam ) were calculated with a code kindly supplied by l. mundy .
the models are exactly those used to produce fig .
[ f : chi ] , except that for w 3 irs5 , gl 490 and ngc 7538 irs9 , the outer radii were increased to @xmath178 to avoid edge effects . in these cases ,
the radial profile of the dust emission was quite extended .
figure [ fig : dust_prof ] demonstrates that the dust and gas tracers agree very well on the best value of @xmath141 , implying that the volume density and column density distributions are consistent .
this result indicates that the dust and gas are well - mixed and that the structure of the envelopes is fairly homogeneous and not very clumpy .
the only exception to this rule is s 140 irs1 , for which the dust gives @xmath29 while the gas gives @xmath30 .
although an inhomogeneous ( clumpy ) structure would cause a discrepancy in this direction , it is hard to see why only one of our sources would have such a structure .
clumps in s 140 have been proposed by e.g. @xcite , but in a much larger ( @xmath41arcmin - sized ) region outside the dense star - forming core . more likely to be important here
is an elevated temperature caused by the nearby sources irs2 and irs3 at @xmath46 offsets , which have luminosities similar to that of irs1 @xcite , and by external heating by the nearby ionization front @xcite .
so s 140 is unique in our sample in that it is the only source where our assumption of one central heating source breaks down .
although the data on gl 2591 are well matched by the power law on average , the slope of the data is shallower inside a radius of @xmath179 , and drop more steeply outside this radius .
this suggests a variation in the temperature or the density gradient with radius .
since the effect is more pronounced towards shorter wavelengths , a variation of the temperature gradient with radius seems more plausible .
an inner region of roughly constant temperature , such as caused by a central cavity with little or no extinction , may reproduce the data , but such models fall outside the scope of this paper .
we have modeled single - dish continuum and line data with power law models @xmath140 , and obtained good solutions ( @xmath180 ) with @xmath181 for the bright mid - infrared sources and @xmath182 for the other sources , although the highest value @xmath28 found for two `` hot cores '' may be an overestimate due to an enhanced cs abundance close to the center .
this section investigates the validity of the assumptions that went into the models and discusses possible implications of their results .
the models developed in section [ s : models ] assume that the clouds are spherical and homogeneous .
@xcite observed the @xmath183 band at @xmath184 in absorption towards the bright mid - infrared sources .
the excitation of the quiescent ( non - outflowing ) gas rules out single - temperature models , but the data are well fitted by a model with two discrete temperatures . the column density ratio of these `` cold '' and `` hot '' components varies from @xmath6 to @xmath185 .
the physical origin of such a two - temperature structure is not clear , however .
we have calculated the column densities in each rotational state of up to @xmath186 from our power - law models , and the results are compared to the data of mitchell et al . in figure
[ fig : ir_abs ] . the low - resolution data for gl 7009s from @xcite may contain a contribution from outflowing gas , and are hence regarded as upper limits .
the models have been scaled to the observed total column density , which is always within a factor of 2 from the column density measured by our submillimeter emission data ( [ s : est_phys ] ) .
the good match ( maximum deviation of a factor of @xmath187 ) to the infrared data , which sample a pencil beam toward the infrared source , indicates that a spherically symmetric model is a reasonable first - order description of these sources .
together with the evidence from the cs emission line profiles and the near - infrared continuum emission , we estimate the deviations from spherical geometry as a decrease in optical depth by a factor of @xmath13 over an @xmath112 area . the assumption of a homogeneous density structure is tested in figure [ fig : mvir ] , where is compared to the mass derived from the virial theorem inside the same radius .
the latter method assumes that the cloud s gravitation just balances its pressure as measured by the line width @xmath188 ( table [ vlsr.tab ] ) . if the gas is clumpy , the virial mass is smaller than the power law mass by a fraction @xmath189 , with the volume filling factor @xmath190 .
the ratio @xmath191 has a mean value of 2.77 and a standard deviation of 1.66 .
the only source for which @xmath192 is gl 7009s , for which the model results are uncertain . while there is no evidence for clumping within the envelopes , other authors have found that regions of massive star formation are usually clumped on a much larger scale . it may not be a coincidence that the sizes derived for the clumps in s 140 by @xcite and in m 17 sw by @xcite , @xmath76 pc , are similar to the sizes of the envelopes found here .
it is however outside the scope of this paper to investigate a possible evolutionary connection between molecular cloud clumps and the envelopes of massive young stars .
our assumed values of the gas to dust mass ratio and the dust opacity at submillimeter wavelengths were tested by modeling c@xmath0o @xmath193 and @xmath194 observations .
the best - fit values for the abundance of co are listed in column 6 of table [ t : m_gas ] .
the derived abundances are @xmath195% of the value of @xmath196 measured by @xcite towards the warm cloud ngc 2024 .
our co/ value for gl 2591 is in good agreement with the value of @xmath197 measured by c. kulesa ( 1999 , priv .
comm . ) , while in the case of gl 490 , our value is a factor of 7 lower .
the co abundance measured in emission is expected to be lower than in absorption , because the emission data are more sensitive to extended cold gas . when measured in infrared absorption , which samples warm material close to the star , the column density ratio of solid to gaseous co never exceeds unity towards massive young stars @xcite .
although source - to - source variations in grain opacity can not be ruled out , a more likely explanation of the spread in co abundances is that different amounts of co are frozen out on grain surfaces .
this occurs at temperatures @xmath198 k in the case of pure co @xcite , and up to @xmath199 k for co- mixtures . in both cases ,
a correlation of the abundance with temperature is expected .
figure [ fig : coh2 ] shows a plot of the co abundance derived from c@xmath0o observations versus the mass - weighted temperature , defined as @xmath200 ( see table [ t : m_gas ] ) .
the two quantities are seen to be correlated , which we interpret as the effect of depletion of co at low temperatures .
these results strongly suggest that depletion and thermal desorption are the processes controlling the abundance of co in the gas phase , after chemical reactions lock up almost all of the available carbon .
this situation is in contrast with that of cs , for which the abundance shows no clear trend with @xmath201 .
this difference presumably reflects the much greater chemical inertness of co compared to cs , and the lower evaporation temperature of co compared with cs . in the case of gl 7009s ,
the column density observed in absorption by @xcite is similar to that in our model .
this implies that along this line of sight , co is much less depleted than cs , for which a depletion by a factor @xmath202 was found compared to the other sources in our sample ( section [ s : mod_gas ] ) .
again , this can be understood from the difference in evaporation temperature for these two molecules .
the large column of co ice towards , e.g. , w 3 irs5 , is not predicted by our model .
we suggest that this star - forming core is surrounded by an extended dense cloud which also produces the self - reversed co emission profiles .
sources like gl 2591 and gl 2136 do not seem to have such a `` skin '' , or they have at most a translucent one .
@xcite found that the submillimeter opacity of dust grains increases by @xmath203% when the temperature rises from 24 to 300 k. this effect would decrease the masses of our sources and increase the inferred abundances of co , and qualitatively produce the same trend as seen in figure [ fig : coh2 ] .
however , over the applicable temperature range , @xmath204 k , an effect of only @xmath205% should occur , which is much less than the factor of 3 increase in co abundance that we find .
although the grain opacity is likely to vary within our sources , they do not influence the conclusion that the co abundance is controlled by freezing and sublimation . in this section ,
the envelope models derived in [ s : models ] will be compared to the ovro continuum data . since the images presented in figure [ fig : ovromaps ] show only compact , circularly symmetric structure , a simple one - dimensional analysis in the fourier plane is sufficient .
the points in figure [ fig : cvis ] are the ovro data , averaged in annuli around the source in the _ uv _ plane . in the case of
w 33a , the source mm2 was subtracted from the data before averaging .
the error bars represent the @xmath207 spread between the data points in each bin , and do not include the overall calibration error . superposed on the data points
are model curves , derived by calculating the dust emission from the best - fit models ( table [ t : m_gas ] ) and fourier transforming the result . only at the lowest spatial frequencies ,
the data match the model curves within the calibration error .
most data points lie well above the model curves , and the observed amplitudes do not drop with increasing spatial frequency , as the models do .
these differences suggest that the emission detected with ovro is not related to the envelopes seen in the single - dish data , but due to compact structure of size @xmath208 .
the same conclusion was reached for gl 2591 by @xcite . towards the highest
observed spatial frequencies , @xmath209 rad@xmath7 , corresponding to @xmath210 au at 1 kpc , the compact emission is @xmath23 times stronger than that of the envelope model .
this factor is a lower limit to the column density contrast between the two components , since the compact emission is probably optically thick , as indicated by the spectral indices .
combined with the envelope column densities in table [ vlsr.tab ] , this result suggests that the compact emission seen with ovro can explain the asymmetry observed in the cs line profiles ( [ s : res : spectra ] ) .
further constraints on the nature of the compact emission can be obtained from the flux densities ( table [ posflux.tab ] ) , which imply brightness temperatures of @xmath211 k for ngc 7538 irs1 and @xmath6 k for the other sources .
these values imply that the emission is beam diluted by a factor of @xmath212 , since the physical temperature of the compact structure must be at least that of the surrounding envelope , which acts as an oven ( see @xcite ) . assuming a temperature of @xmath213 k for the compact dust , a lower limit to the mass of @xmath214
is obtained . in the case of ngc 7538 irs1 , where free - free and dust emission both contribute at these frequencies
@xcite , the physical temperature may be @xmath35 k , and the beam dilution correspondingly more . for the other sources ,
the compact emission is most likely due to dust , perhaps in the form of a dense shell or of a circumstellar disk .
the implied radius of the emitting region , @xmath215 au at @xmath216 kpc , seems small to hide an entire outflow lobe , but presumably , the dense part of the flow which emits in high-@xmath110 cs lines is confined to its center .
this section discusses possible origins of the spread in @xmath141 within our sample .
first , the value of @xmath141 may be related to other physical parameters , in particular luminosity or envelope mass .
the structure of the material surrounding pre - main sequence stars of low ( @xmath41 solar ) and intermediate ( @xmath217 solar ) mass has received considerable attention in the recent literature : e.g. @xcite .
these studies generally find @xmath218 inside radii of 0.1 pc , masses of @xmath21910 and mean densities of @xmath220 .
of these properties , only the radii are similar to those of the objects studied here ; the masses and mean densities are at least two orders of magnitude smaller .
the density gradients are significantly steeper than found in this paper , but the gradients for intermediate - mass stars are also lower than those for low - mass stars @xcite .
indeed , in an earlier study of regions of high - mass star formation , @xcite suggested that more massive star - forming regions tend to have flatter density distributions .
however , figs . [
fig : alpha]a and b show that within our sample , @xmath141 does not appear correlated with source luminosity or envelope mass . we did not find a relation with any other physical parameters either , such as turbulent pressure as measured by the line width .
power laws have been proposed for the density structures of the envelopes of young stars , with the index depending on the dominant term in the pressure .
if support against gravitational collapse is primarily due to thermal pressure , a value of @xmath12 is expected , while if the dominant pressure is of nonthermal origin , the density structure should follow an @xmath221 law . in between these two cases ,
a continuum of solutions can be constructed @xcite .
our observations may thus indicate that the cores where massive stars form differ from those where low - mass stars form in that they are supported against collapse by a different mechanism . within this interpretation
, the spread in @xmath141 may be due to source evolution from an ( @xmath29 ) logatrope to a collapsing region ( @xmath30 ) . despite having similar radio and infrared properties
our sources show different degrees of dispersal and warming - up of their envelopes .
this is shown by @xcite and @xcite for a source sample similar to ours based on several tracers : the far - infrared color , the fraction of crystalline co@xmath19 ice , the gas / solid ratios of co , co@xmath19 and h@xmath19o , and the excitation temperature of co. these quantities all trace the envelope temperature , but all in a different way : some trace the dust , others the gas , and some trace small scales and others large scales . as shown by boogert et al . , the temperature indicators correlate with each other .
of particular interest is the fraction of crystalline co@xmath19 ice , because crystallization is an irreversible process , so that this indicator measures the maximum temperature , while the others measure the current temperature .
hence , the temperature variations are not random fluctuations ( like in fu orionis objects ) , but reflect a systematic increase with time of the temperature of the envelope as a whole . figure [ fig : evol]a plots one of these indicators , the far - infrared color , versus the ratio of to stellar mass , measured as @xmath222 .
this quantity is equivalent to the ratio of submillimeter to bolometric luminosity , often used to trace the evolution of young low - mass objects @xcite .
the color @xmath223 is the ratio of the flux densities at @xmath224 and @xmath25 , both observed with iso - lws in an @xmath225 beam , which is large enough to cover the entire envelope .
these data were provided by a. boogert ( 1999 , priv .
the plotted quantities are seen to be correlated , with higher temperatures ( traced by a larger @xmath223 ) corresponding to lower values of @xmath226 .
thus , the temperature variations are indeed due to dispersal of the circumstellar material and can be used to trace evolution .
interestingly , @xmath141 is not seen to be correlated with far - infrared color ( fig .
[ fig : evol]b ) , and we conclude that on the time scale of the dispersal of the envelopes of massive young stars , the density structure of the envelopes does not change significantly .
the masses of the compact sources detected by us and by @xcite and @xcite do not show a relation with /@xmath222 either , although the uncertainty in these masses is rather high .
in addition to @xmath226 , we have considered two `` derivative '' evolutionary indicators .
the first is the bolometric temperature , introduced for low - mass objects by @xcite . as the envelopes in those systems are dispersed ( probably by the bipolar outflows ) , their bolometric temperatures rise because the far - infrared emission decreases while the near - infrared emission increases @xcite .
this process makes increase monotonically from @xmath227 k to @xmath228 k on a time scale of @xmath229 yr . for all but two of our objects , values of based on the spectral energy distributions of fig .
[ f : dusto ] , are @xmath230 k. this small spread suggests that all the sources are at similar evolutionary stages ; may also work less well than @xmath223 because near- and mid - infrared emission have a dependence on source orientation .
another potential evolutionary indicator is the radio continuum flux , which for a given luminosity and distance is expected to increase with time as the dusty envelope is cleared away and the lyman continuum photons can all go into ionizing the gas without being absorbed by dust .
we have normalized the radio continuum flux densities in table [ sam.tab ] to the value expected for an h ii region ionized by a main sequence star of the same luminosity and at the same distance as the source , following @xcite .
however , figs .
[ fig : evol]c and d show that neither the radio continuum emission nor the bolometric temperature correlates well with /@xmath222 , suggesting these are less useful evolutionary indicators for very young massive objects than @xmath223 .
possibly the radio emission arises in a wind ( [ s : samp ] ) , or the stellar surface temperatures are significantly below main sequence values .
we conclude that dispersal of the envelopes underlies the evolution of both high - mass and low - mass objects , but that the observational appearance is very different .
the dispersal process can be understood in more detail by recalling that the brightest mid - infrared sources in our sample are the weakest at radio wavelengths and vice versa ( table [ sam.tab ] ) .
this anticorrelation suggests that the erosion of the envelope proceeds from the inside out .
massive stars are efficient at removing circumstellar dust by their ionizing uv radiation and by their strong winds . when the innermost , hottest region of dust disappears , the mid - infrared emission will decrease , but the radio continuum will increase because a larger region can be ionized .
the compact dust component detected with ovro is optically thick in the bright mid - infrared sources but thin in the others , which trend may also trace dispersal of hot dust .
the far - infrared emission will not be affected since it arises on larger scales .
these trends are consistent with the properties of our sample , although other processes may play a role as well .
we have presented maps and spectra of 14 regions of massive star formation in continuum and molecular lines at submillimeter wavelengths .
the data are used to develop models of the physical structure of the circumstellar envelopes on @xmath2 au scales .
the column density is derived from the submillimeter continuum flux densities , and the temperature structure is calculated self - consistently using the size of the cs emission and the sources luminosities and distances from the literature .
the density structure is constrained with emission lines of cs , and .
the following main conclusions are found : \1 .
the physical structure of the envelopes of deeply embedded massive young stars is characterized by radii of @xmath231 au , masses of @xmath232 inside these radii and visual extinctions of @xmath233 magnitudes .
temperatures increase from @xmath234 k at the outer edge to several @xmath25 k at @xmath235 au from the star ; densities increase from @xmath236 to @xmath237 .
the slope of the density gradient @xmath141 is @xmath238 , significantly smaller than the value of @xmath239 usually found for low - mass objects .
this difference in density structure may be due to nonthermal pressure resisting gravitational collapse , while thermal pressure dominates in lower - mass objects .
an @xmath12 density structure is found here for two hot cores , but this is an upper limit since the cs abundance may be enhanced in the warm gas close to the star .
the shapes of the envelopes deviate somewhat from spherical , as shown by modeling infrared absorption line data of , the near - infrared continuum and the cs line profiles
. the data are consistent with a decrease of the optical depth by a factor of @xmath13 in the central @xmath112 area .
the brightness of our main sample at mid - infrared wavelengths is partly due to this optical depth effect , but also to the flatter density structure .
inhomogeneities on top of the power law structure are small , since the masses obtained by integrating the power law models agree with masses found from the virial theorem to a factor of 3 .
the values of @xmath141 found from cs are verified by model fits to the maps of dust emission which trace the column density distribution .
the only exception is s 140 irs1 , probably because our assumption of a single central heating source is invalid in this case .
modeling of c@xmath0o emission lines shows that @xmath240% of the co gas is depleted onto dust , assuming a dust opacity of @xmath6 g@xmath7 at @xmath5 , as found previously for one of our sources , gl 2591 @xcite .
the derived co abundances correlate well with the mass - weighted temperature in our models .
this result suggests that in these sources , freeze - out and thermal desorption control the gas - phase abundance of co. the cs abundance is @xmath16 on average , ranging from @xmath169 in the cold source gl 7009s to @xmath241 in the `` hot core''-type objects .
dense outflowing gas is seen in the cs and h@xmath19co line wings , but much more so at blue- than at redshifted velocities .
this asymmetry may be due to absorption by dense , optically thick material within 10@xmath20 from the star .
the opacity of the power law models is a factor of @xmath23 below that required for this absorption .
however , interferometric continuum observations at @xmath21 show compact emission , probably from an @xmath22 diameter optically thick dust component .
the emission is a factor of @xmath242 stronger than expected for the envelopes seen in the single - dish data , and may be due to a dense shell or a circumstellar disk .
this component may explain the asymmetric cs and line profiles .
the evolution of these high - mass sources is traced by the overall temperature ( measured by the far - infrared color ) , which increases systematically with decreasing ratio of envelope mass to stellar mass .
the observed anticorrelation of near - infrared and radio continuum emission suggests that the disruption of the envelope proceeds from the inside out .
conventional tracers of the evolution of low - mass objects do not change much over this narrow age range .
we conclude that the evolution of high - mass and low - mass envelopes have the same underlying mechanism ( envelope dispersal ) , despite their different observational properties .
the authors are grateful to the staffs of the ovro , jcmt , cso and iram 30 m telescopes , in particular remo tilanus and fred baas at the jcmt , gilles niccolini at iram and leonardo testi , christine wilson and debra shepherd at ovro .
we are also indebted to michiel hogerheijde for developing the monte carlo code , to gran sandell for providing the scuba maps of ngc 7538 and for his constructive referee report , to lee mundy for providing the code to model dust emission profiles , to yancy shirley who reduced the sharc data , to adwin boogert for providing the iso - lws data and to craig kulesa for sending his co data .
this work also benefitted from discussions with malcolm walmsley , ed churchwell and xander tielens .
this research is supported by nwo grant 614.41.003 .
gab gratefully acknowledges support provided by nasa ( grants nag5 - 2297 , -4813 ) .
nje gratefully acknowledges support from the state of texas and nasa grant nag5 - 7203 .
we also wish to thank the people who maintain the bibliographic databases at cds ( strasbourg ) and ads ( harvard ) , of which we have made extensive use .
, t. k. , menten , k. m. , wyrowski , f. , & schilke , p. 1999 , in star formation 1999 , proceedings of star formation 1999 , held in nagoya , japan , june 21 - 25 , 1999 , editor : t. nakamoto , nobeyama radio observatory , p. 183 - 184 , 183184 , e. f. , helmich , f. p. , schutte , w. a. , ehrenfreund , p. , lahuis , f. , boogert , a. c. a. , tielens , a. g. g. m. , de graauw , t. , gerakines , p. a. , & whittet , d. c. b. 1998 , in asp conf .
ser . 132 : star formation with the infrared space observatory , 54 observed continuum spectra ( _ open circles _ ) and models for several values of @xmath141 .
the symbol legend is in the top left panel . for w 33a , a model using grains without ice mantles is also shown .
data are taken from @xcite for w 3 irs5 ; @xcite for gl 490 ; @xcite , @xcite , @xcite , @xcite , @xcite and @xcite for w 33a ; @xcite , @xcite , van dishoeck ( priv .
comm . ) and @xcite for gl 2136 ; @xcite and @xcite for gl 7009s ; @xcite , @xcite and @xcite for s 140 irs1 , @xcite for ngc 7538 irs1 ; @xcite and g. sandell ( priv . comm . ) for ngc 7538 irs9 ; @xcite and @xcite for iras 20126 + 4104 ; @xcite , @xcite , @xcite and @xcite for dr 21 ( oh ) ; @xcite , @xcite and @xcite for w 3 ( ) , @xcite and @xcite for ngc 6334 irs1 ; and @xcite , @xcite and @xcite for w 28 a2 . &
& & & + + & & & & + 02219 + 6152 & w 3 irs5 & @xmath243 @xmath244 @xmath245 & @xmath246 @xmath247 @xmath248 & @xmath249 & @xmath132 & 10.7 & 1,2,21 + 03236 + 5836 & gl 490 & @xmath250 @xmath251 @xmath252 & @xmath253 @xmath254 @xmath255 & @xmath256 & @xmath257 & 3.2 & 3,4,22 + 181171753 & w 33a & @xmath258 @xmath259 @xmath260 & @xmath261 @xmath262 @xmath263 & @xmath264 & @xmath40 & 1.9 & 5,23 + 181961331 & gl 2136 & @xmath258 @xmath265 @xmath266 & @xmath267 @xmath268 @xmath269 & @xmath270 & @xmath271 & & 6 + 183160602 & gl 7009s & @xmath258 @xmath272 @xmath273 & @xmath274 @xmath275 @xmath276 & @xmath277 & @xmath187 & 2.7 & 7 + 20275 + 4001 & gl 2591 & @xmath278 @xmath279 @xmath280 & @xmath281 @xmath282 @xmath52 & @xmath283 & @xmath257 & 0.4 & 8,9,24 + 22176 + 6303 & s 140 irs 1 & @xmath284 @xmath285 @xmath286 & @xmath287 @xmath288 @xmath289 & @xmath283 & @xmath39 & 5.8 & 10,11,25 + 23116 + 6111 & ngc 7538 irs 1 & @xmath290 @xmath259 @xmath291 & @xmath246 @xmath292 @xmath293 & @xmath294 & @xmath295 & 111 & 12,13,26 + 23118 + 6110 & ngc 7538 irs 9 & @xmath290 @xmath259 @xmath296 & @xmath246 @xmath297 @xmath298 & @xmath299 & @xmath295 & @xmath300 & 12,13,23 + & & & & + + & & & & + 20126 + 4104 & iras 20126 + 4104 & @xmath278 @xmath301 @xmath302 & @xmath303 @xmath304 @xmath95 & @xmath305 & @xmath257 & @xmath306 & 14,9,27 + & dr 21 ( oh ) & @xmath278 @xmath307 @xmath308 & @xmath309 @xmath310 @xmath53 & @xmath311 & @xmath257 & @xmath312 & 15,9,28 + 02232 + 6138 & w 3 ( ) & @xmath243 @xmath251 @xmath313 & @xmath246 @xmath314 @xmath315 & @xmath316 & @xmath132 & 1.5 & 16,2,29 + 171753544 & ngc 6334 irs 1 & @xmath317 @xmath285 @xmath318 & @xmath319 @xmath320 @xmath321 & @xmath322 & @xmath323 & 1200 & 17,18,30 + 175742403 & w 28 a2 & @xmath317 @xmath324 @xmath325 & @xmath326 @xmath288 @xmath327 & @xmath294 & @xmath328 & 2100 & 19,20,31 + & ( = g5.89 - 0.39 ) & & & & & + + & & & & & & & & & + c@xmath0o 21 & 12.2 & & 11.3&6.8 & 6.4 & 8.7 & 3.8 & 26.5 & + 32 & 16.6 & 4.9 & 15.7&9.9 & 10.8&17.9&4.5 & 57.9&15.1 + cs 54 & & 16.5&36.2&11.2&30.2&54.8&20.6 & & + 76 & 25.2 & 8.0&17.4&4.7 & 22.3&56.0&20.5&162.&83.7 + 109 & 16.1 & & & & & 34.4&3.3 & & 78.4 + c@xmath329s 54 & 3.7 & 1.1 & 5.6&1.1 & 2.7 & 7.0 & 2.4 & 56.5&20.7 + 76&2.5 & 0.6 & 5.3 & @xmath330&2.1&7.9 & 3.1 & 64.8 & 25.8 + 109&2.2 & & & & & 4.6 & @xmath331&&31.6 + @xmath332&8.5 & 4.4 & 13.1 & 4.4&13.7&27.0&14.2&63.2&20.2 + @xmath333&19.0&&15.0 & 6.3&19.2&27.0&13.6&63.0&41.2 + @xmath334&3.6 & 0.7 & 3.5 & 1.6&3.5 & 7.1 & 3.7 & 26.2&6.9 + @xmath335&5.2 & & 10.9 & 4.3&8.7 & 18.2&9.5 & 82.9&31.1 + @xmath336&14.0&&20.0 & 7.4&20.5&28.8&&105.&36.2 + @xmath337&2.7 & 1.6 & 4.5 & 2.2&3.7 & 7.9 & 4.5 & 52.9&19.5 + @xmath338&2.9 & 1.5 & 5.4&1.5 & 3.6 & 8.8 & 4.5 & 48.1&33.4 + @xmath339&2.6 & 1.2 & 4.9&2.6 & 3.3 & 9.9 & 4.2 & 50.6&24.7 + @xmath340&0.9&@xmath341&1.5&@xmath330&@xmath300&2.4&1.3&29.4&8.3 + @xmath342&21.9&&&&7.3 & 15.8&10.8&&24.3 + & & & & & & & & & + + & & & & & & & & & + cs 54 & & & 80.7 & 22.5 & 37.7 & 79.0 & 37.2 & & + c@xmath329s 21 & & & 7.6 & 1.8 & 3.6 & & 4.1 & & + c@xmath329s 32 & & & & 2.4 & 4.3 & & 5.4 & & + & & & & & & & & & + + & & & & & & & & & + cs 54 & 21.5 & & 20.3 & 8.3 & 30.5 & 54.4 & 18.2 & 78.6 & 33.5 + cs 76 & 20.0 & 4.9 & 10.5 & 6.6 & 20.7 & 50.7 & @xmath343 & 103.6 & 64.2 + c@xmath329s 76 & & & & & & & & & 15.8 + w 3 irs5 & -38.4 ( 0.3 ) & 2.7 ( 1.1 ) & 3.7 & 55 + gl 490 & -13.3 ( 0.2 ) & 3.3 ( 0.6 ) & 0.8 & 38 + w 33a & + 37.5 ( 0.9 ) & 5.4 ( 0.5 ) & 3.1 & 37 + gl 2136 & + 22.8 ( 0.1 ) & 3.1 ( 0.4 ) & 1.4 & 36 + gl 7009s & + 40.3 ( 0.6 ) & 5.0 ( 1.0 ) & & + gl 2591 & -5.5 ( 0.2 ) & 3.3 ( 0.6 ) & 3.4 & 52 + s 140 irs1 & -7.0 ( 0.2 ) & 3.3 ( 0.3 ) & 1.4 & 64 + ngc 7538 irs1 & -57.4 ( 0.5 ) & 4.1 ( 1.4 ) & 3.9 & 52 + ngc 7538 irs9 & -57.2 ( 0.3 ) & 4.1 ( 0.5 ) & 1.0 & 47 + & & & & + iras 20126 & -3.7 ( 0.2 ) & 3.2 ( 0.2 ) & 0.8 & 40 + dr 21 ( oh ) & -3.1 ( 0.2 ) & 4.5 ( 0.2 ) & 15.6 & 56 + w 3 ( ) & -47.6 ( 0.6 ) & 5.8 ( 0.6 ) & 8.5 & 41 + ngc 6334 irs1 & -7.4 ( 0.2 ) & 5.3 ( 0.3 ) & 16.9 & 50 + w 28 a2 & + 10 & 5.7 & & 38 + gl 2136 & 86 & 18:19:36.62 ( 0.02 ) & -13:31:43.8 ( 0.3 ) & 61 & @xmath344 +
w 33a mm1 & 86 & 18:11:44.22 ( 0.02 ) & -17:52:58.2 ( 0.2 ) & 24 & @xmath345 + & 106 & & & 54 & @xmath346 + & 233 & & & 190 & @xmath347 + w 33a mm2 & 86 & 18:11:43.91 ( 0.02 ) & -17:52:59.8 ( 0.2 ) & 38 & @xmath345 + & 106 & & & 30 & @xmath346 + & 233 & & & 140 & @xmath347 + ngc 7538 irs1 & 86 & 23:11:36.64 ( 0.02 ) & + 61:11:49.8 ( 0.2 ) & 1500 & @xmath348 + & 107 & & & 1800 & @xmath349 + & 115 & & & 2000 & @xmath350 + & 230 & & & 3500 & @xmath351 + ngc 7538 irs9 & 107 & 23:11:52.90 ( 0.07 ) & + 61:10:59.2 ( 0.5 ) & 43 & @xmath352 + & 115 & & & 95 & @xmath349 + w 3 irs5 & 60 & 1.5 & @xmath353 & 262 & 355 & 33 & 1.6 & 5 . & 3 . + gl 490 & 19 & 1.5 & @xmath354 & 30 & 168 & 19 & 0.4 & 1 . & 1 .
+ w 33a & 74 & 1.0 & @xmath355 & 1089 & 1984 & 20 & 0.5 & 5 . & 4
. + gl 2136 & 35 & 1.25&@xmath356 & 72 & 316 & 28 & 1.2 & 4 . & 8 .
+ gl 7009s & 91 & 0.5 & @xmath357 & 3955 & 2218 & 21 & & 0.4 - 0.8 & + gl 2591 & 27 & 1.0 & @xmath358 & 44 & 268 & 28 & 1.5 & 10 . & 4 .
+ s 140 irs1 & 29 & 1.5 & @xmath356 & 46 & 128 & 27 & 1.0 & 5 . & 5 .
+ ngc 7538 irs1 & 72 & 1.0 & @xmath359 & 815 & 1128 & 25 & 0.6 & 10.&10 .
+ ngc 7538 irs9 & 66 & 1.0 & @xmath360 & 430 & 1016 & 20 & 0.3 & 10.&10 . +
iras 20126 & 39 & 1.75&@xmath132 & 84 & 294 & 27 & 0.5 & 3 .
& + dr 21 ( oh ) & 29 & 1.75&@xmath361 & 350 & 429 & 15 & 0.8 & 3 . & 5 . + w 3 ( h@xmath19o ) & 45 & 2.0 & @xmath359 & 385 & 932 & 26 & 1.0 & 10 . & 3 .
+ ngc 6334 irs1 & 43 & 2.0 & @xmath362 & 382 & 725 & 35 & 1.5 & 12 . & 7 . + w 28 a2 & 37 & 1.5 & @xmath363 & 292 & 982 & 42 & 1.7 & 20 . & + | the physical structure of the envelopes around a sample of fourteen massive young stars is investigated using maps and spectra in submillimeter continuum and lines of c@xmath0o , cs , and .
nine of the sources are highly embedded luminous ( @xmath1 ) young stellar objects which are bright near - infrared sources , but weak in radio continuum ; the other objects are similar but not bright in the near - infrared , and contain `` hot core''-type objects and/or ultracompact regions .
the data are used to constrain the temperature and density structure of the circumstellar envelopes on @xmath2 au scales , to investigate the relation between the different objects and to search for evolutionary effects .
the total column densities and the temperature profiles are obtained by fitting self - consistent dust models to submillimeter photometry .
the calculated temperatures range from 300 to 1000 k at @xmath3 au and from 10 to 30 k at @xmath4 au from the star .
visual extinctions are a few hundred to a few thousand magnitudes , assuming a grain opacity at @xmath5 , of @xmath6 g@xmath7 of dust , as derived earlier for one of our sources
. the mid - infrared data are consistent with a @xmath8% decrease of the opacity at higher temperatures , caused by the evaporation of the ice mantles .
the cs , and data as well as the submillimeter dust emission maps indicate density gradients @xmath9 . assuming a constant cs abundance throughout the envelope , values of @xmath10 are found , significantly flatter than the @xmath11 generally found for low - mass objects .
this flattening may indicate that in massive young stellar objects , nonthermal pressure is more important for the support against gravitational collapse , while thermal pressure dominates for low - mass sources .
we find @xmath12 for two hot core - type sources , but regard this as an upper limit since in these objects , the cs abundance may be enhanced in the warm gas close to the star
. the assumption of spherical symmetry is tested by modeling infrared absorption line data of , cs emission line profiles and near - infrared continuum .
there is a distinct , but small deviation from spherical symmetry : the data are consistent with a decrease of the optical depth by a factor of @xmath13 in the central @xmath14 .
the homogeneity of the envelopes is verified by the good agreement of the total masses in the power law models with the virial masses .
modeling of c@xmath0o emission shows that @xmath15% of the co is frozen out onto the dust .
the co abundances show a clear correlation with temperature , as expected if the abundance is controlled by freeze - out and thermal desorption .
the cs abundance is @xmath16 on average , ranging from @xmath17 in the cold source gl 7009s to @xmath18 in the two `` hot core''-type sources .
dense outflowing gas is seen in the cs and h@xmath19co line wings ; the predominance of blueshifted emission suggests the presence of dense , optically thick material within 10@xmath20 of the center .
interferometric continuum observations at @xmath21 show compact emission , probably from an @xmath22 diameter , optically thick dust component , such as a dense shell or a disk .
the emission is a factor of @xmath23 stronger than expected for the envelopes seen in the single - dish data , so that this component may be opaque enough to explain the asymmetric cs and line profiles .
the evolution of the sources is traced by the overall temperature ( measured by the far - infrared color ) , which increases systematically with decreasing ratio of envelope mass to stellar mass .
the observed anticorrelation of near - infrared and radio continuum emission suggests that the erosion of the envelope proceeds from the inside out .
conventional tracers of the evolution of low - mass objects do not change much over this narrow age range . |
a large number of empirical studies has shown that financial time series exhibit statistical features strongly departing from the gaussian behavior .
this finding dates back to the work of mandelbrot @xcite , whose attention was mainly focused in recognizing realizations of stable processes , and to the analysis of fama @xcite concerning the long tailed nature of the dow jones industrial average single components . since these fundamental contributions the modeling of financial returns has considerably grown , and very heterogeneous models , able to reproduce the degree of asymmetry and the excess of kurtosis of the measured distributions , have been proposed .
a non exhaustive list includes approaches developing from specific distributional assumptions , as it is the case of the lvy flights @xcite , the generalized student-@xmath0 or tsallis distributions @xcite and the exponential one @xcite .
past empirical analysis have also proved the existence of non trivial scalings of higher order correlations between returns at different times , pointing toward the existence of a secondary stochastic process , as fundamental as that of the price , governing the volatility of returns .
many effective mechanisms allowing to reproduce the observed correlation structures , where the stochastic nature of the volatility plays a central role , were proposed .
discrete time models include arch - garch processes @xcite , and multifractal models @xcite , inspired by the cascades originally introduced by kolmogorov in the context of turbulent flows . as far as continuous time approaches are concerned , fractional brownian motion and stochastic volatility models have been extensively analyzed . for a review of the latter approach we suggest @xcite and remind the reader to the discussion in section [ sec : model ] .
focusing on the continuous time stochastic volatility framework , in this work we aim at reproducing many of the above mentioned facts which are generally accepted as universal evidences , shared among different markets in different times .
the structure of the paper is the following . after introducing a general class of stochastic models driving the evolution of the volatility , in section [ sec : model ] we concentrate on a linear one able to reproduce an inverse gamma distribution in the long run . in section
[ sec : fattails ] we detail the derivation of the moments of the probability density function @xmath1 of the returns over the time lag @xmath0 , taking into account explicitly the time at which the @xmath2 process has started and deriving rigorously the stationary limit of the volatility .
we describe the mechanism through which the power law distribution of @xmath3 induces fat tails on @xmath1 for all the finite time lags . in sections [ sec :
lev ] and [ sec : autoc ] we derive the analytical expressions of the leverage correlation and the volatility autocorrelation functions respectively . in section [ sec :
estimation ] we propose a systematic methodology for estimating the model parameters , and we apply it to the time series of the daily returns of the standard & poor 500 index . the relevant conclusions , along with possible perspectives , will be summarized in section [ sec : concl ] .
we consider a model where the asset price @xmath4 is a function of the stochastic centred log - return @xmath5 and @xmath6 is a constant drift coefficient .
we assume that @xmath5 can be modeled with the following stochastic differential equation ( sde ) @xmath7 where @xmath8 is the instantaneous volatility of the price and @xmath9 is the increment of a standard wiener process . since @xmath10 , from the above assumption we have that @xmath11 and @xmath12 for all @xmath0 . in the context of stochastic volatility models ( svms )
the instantaneous volatility is assumed to be a function of an underlying driving process @xmath13 , i.e. @xmath14 .
typically , the dynamics chosen for @xmath13 corresponds to a particular case of the following general multiplicative diffusion process @xmath15 with suitable constraints on the parameters , in order to ensure the well definiteness of the process .
moreover , the two standard wiener processes @xmath16 are possibly correlated @xmath17 with @xmath18 $ ] , which is necessary to account for skewness effects and for the return - volatility correlation . for instance , in the stein - stein model @xcite the volatility is linear , @xmath19 , and @xmath13 follows a mean reverting ornstein - uhlenbeck dynamics corresponding to @xmath20 , @xmath21 , @xmath22 . under the same @xmath2 dynamics but with @xmath23
we obtain the exponential ornstein - uhlenbeck model @xcite . in the heston model
@xcite @xmath24 and @xmath13 evolves according to a cox - ingersoll - ross dynamics , stemming from by taking @xmath20 , @xmath21 with @xmath25 .
finally , in the hull - white model the volatility has the same functional dependence as in heston , but @xmath13 has a log - normal ( non mean reverting ) dynamics corresponding to @xmath26 and @xmath27 . in the econophysics literature
several studies have been devoted to asses the statistical properties of the volatility ( see for instance chapter 7 in @xcite and @xcite ) , especially its distribution , and it has been recognized that the instantaneous volatiliy , measured by suitable proxies , distributes in good agreement with a log - normal or an inverse gamma law , the best fit being obtained with the latter @xcite which is able to better capture the heavy tail of the empirical distribution .
none of the previously cited models feature an inverse gamma probability density function ( pdf ) for @xmath8 , even though this distribution has been considered previously in different contexts .
for instance , the inverse gamma was introduced in the context of an arch - like evolution of the variance in @xcite , and in the statistical modeling of financial data the marginalization of normally distributed returns conditionally on inverse gamma variance was widely exploited since it generates generalized student-@xmath0 distributions ( see @xcite ) . however , as clarified by the empirical analysis performed in @xcite , where intra - day returns are used to estimate a proxy for the daily volatility , an inverse gamma pdf for @xmath28 leads to an overweighting of the tail region . here
we afford the statistical characterization of the simplest linear svm able to account for this stylized fact about the volatility .
the process has been extensively studied and characterized in @xcite where exact solutions for the moments of the associated pdf have been obtained allowing to study its relaxation modes toward a stationary distribution , if any . in particular ,
when @xmath20 and @xmath29 , with @xmath30 , the process has indeed an inverse gamma stationary distribution , whose support is @xmath31 as long as @xmath21 .
thereby we consider the following svm @xmath32 where @xmath33 , @xmath34 may be a fixed constant or randomly sampled , and the constant factor in the expression of the instantaneous volatility @xmath35 has been added for later convenience . as explained in @xcite the stationary pdf of @xmath8 is @xmath36 where the shape parameter @xmath37 and the scale parameter @xmath38 are given by @xmath39
a major point to be discussed before presenting a detailed derivation of our results is the different role played by the initial time conditions for the @xmath40 and @xmath2 processes .
since @xmath5 represents the detrended logarithmic increment of the price over the time lag @xmath0 , it can be directly measured from real time series , and in a natural way we can assume as starting point for this process the spot time @xmath41 . on the other hand , the secondary process can not be observed directly but some of its statistical properties have been measured by means of suitable proxies . in particular , for intra - day frequencies there is no clear evidence of mean reversion , that is the high frequency volatility is very close to its asymptotic value @xcite . in order to capture this evidence ,
we assume that the process @xmath2 , driving the returns from @xmath42 to @xmath0 , started in the past at @xmath43 and we will perform the limit @xmath44 at the end .
the assumption of stationarity for the @xmath8 process in allows also to consider the returns @xmath45 as identically distributed and uncorrelated , even though not independent variables , by virtue of the _ i.i.d . _
property of the wiener increments .
the structure of the model allows to compute the moments of the pdf of @xmath5 at all times @xmath0 recursively .
application of the it lemma to the function @xmath46 readily provides @xmath47 and the same lemma proves that the correlation functions between @xmath40 and @xmath2 satisfy the following differential equation @xmath48 where we defined @xmath49 , @xmath50 for every @xmath51 , and @xmath52 .
the previous equation is a linear ordinary differential equation ( ode ) for every @xmath53 and @xmath54 , which can be solved recursively starting from the lowest order of @xmath53 and @xmath54 . ] , and whose solution involves integration of the moments @xmath55 of the @xmath2 process . for every @xmath56 and every time @xmath0 the latter can be expressed as a linear superposition of exponential functions @xmath57 } \ , .
\label{eq : ymoms}\ ] ] the explicit expressions of the coefficients in the above expansion can be computed as explained in @xcite , and it turns out that @xmath58 involves the values @xmath59 for @xmath60 , while @xmath61 does not .
this implies that whenever the constants @xmath62 are all negative , the only term surviving in the limit @xmath44 is @xmath61 and the process looses every information about the distribution of @xmath34 .
it is worth noticing that , even though the moments @xmath63 are homogeneous functions of time , when @xmath64 is finite this is not true for the solution of eq . which is obtained by integration from @xmath42 to @xmath0 , with boundary condition @xmath65 for every @xmath66 from the moments @xmath67 .
] . from the analysis of eq .
it can be verified that the moments of @xmath40 can be expressed always as a superposition of exponential functions of the starting time of the volatility as follows @xmath68 the coefficients @xmath69 depend on the time lag @xmath0 and , more precisely , by virtue of the linearity of the odes , they correspond to a combination of exponential terms weighted by polinomial functions of @xmath0 . in appendix
[ app : x3 ] we report the explicit expressions of the coefficients @xmath70 for the cases @xmath71 and @xmath72 , from which it can be readily verified that the skewness of the pdf converges to zero asymptotically for @xmath73
. a messy calculation would show that an analogous behavior holds for kurtosis .
thus the scaling of the lowest order moments is in full agreement with the one of the empirical distributions over long time horizons @xcite .
when @xmath0 is finite the coefficients @xmath69 are finite quantities themselves , and all the relevant information about the behavior of @xmath74 in the stationary limit of @xmath2 is retained by the @xmath64-exponentials in eq . .
two cases are possible here : if all the @xmath62 are negative ( @xmath75 ) , @xmath74 is finite in the stationary limit @xmath76 , otherwise it diverges for every @xmath77 with @xmath78 , no cancellation of the divergent terms can take place in the limit @xmath44 .
] indicating the emergence of fat tails in the pdf of @xmath5 .
the latter case applies when @xmath79 , as can be checked directly from the definition of @xmath80 . since @xmath81 , when @xmath82 , the divergence of @xmath74 implies the divergence of all the higher order moments represents an exception since , due to symmetry arguments , all the odd moments vanish identically . ] .
the same condition is responsible for the divergence of the moments @xmath83 of the volatility for @xmath84 ( see eq . ) in agreement with the fact that the stationary distribution of the volatility is an inverse gamma distribution with tail index @xmath37 . here
we see at work a mechanism in which the power law tail of the stationary distribution of the volatility induces fat tails in the return distribution for every time lag @xmath0 , and its scaling for large @xmath85 is compatible with a power law assumption @xmath86 this is in agreement with empirical studies about the distribution of returns over daily or intra - day time scales @xcite , and from the previous considerations we are able to constraint the tail index in the following range @xmath87 where @xmath88 is the largest integer satisfying @xmath89 . of the second and third moment of @xmath40 at @xmath90 , for @xmath91 , @xmath92 , @xmath93 and @xmath94 , @xmath95 .
yearly units ( 1 yr = 250 trading days ) . ] as an example , in fig .
[ fig : t0scaling ] it is shown the scaling of @xmath96 and of the absolute value of @xmath97 as a function of the starting time of the volatility , for @xmath90 and for a choice of the parameters corresponding to @xmath95 .
for this value of the ratio the tail index of the return distribution is @xmath98 and consequently the third moment of the stationary distribution of the volatility diverges as @xmath64 becomes more and more negative , while @xmath96 approaches its finite stationary value .
for the linear model the leverage , measuring the correlation between returns and volatility , can be computed exactly .
since the squared increment @xmath99 provides an estimation of the instantaneous volatility , it can be defined through the following function @xmath100 empirically , for arbitrary @xmath0 , @xmath101 is found to be exponentially decaying for positive @xmath102 and approximately zero otherwise , meaning that a correlation exists between past returns and the volatility in the future and not _
vice versa_. empirical analysis shows that it is a short range correlation ; more precisely , the decay time of @xmath103 is found to be of approximately @xmath104 days for u.s .
stocks and even smaller , about 10 days , for indexes @xcite .
the numerator can be rewritten as @xmath105 expressing the wiener increment as @xmath106 , where @xmath107 is a gaussian noise with zero mean and @xmath108 variance .
novikov theorem @xcite allows to compute the expectation involving @xmath109 , giving us @xmath110 } \right\rangle}\,,\end{gathered}\ ] ] where we defined @xmath111 , we took into account the correlation structure and we used the following expression of the functional derivative of @xmath2 @xmath112 } \,,\end{gathered}\ ] ] with the heaviside step function @xmath113 defined as zero if @xmath114 and one otherwise .
the expectation @xmath115 } \right\rangle}$ ] satisfies an integral volterra equation of the second kind , whose derivation is detailed in appendix [ app : leverage ] , and the final expression of the leverage correlation reads @xmath116 \times \right.\\ \left.\exp{\left[\left ( 2 a + \frac{3}{2 } c \right)\tau\right ] } -\frac{b}{a+c } \,\mu_2(t ) \exp{\left[\left ( a+\frac{c}{2 } \tau \right)\right]}\right\ } \ , , \label{eq : lev3}\end{gathered}\ ] ] which inherits the explicit dependence on @xmath0 from the moments of @xmath2 . in order to compare the previous expression with real data , following the discussion at the beginning of section [ sec : fattails ]
, we take the limit @xmath44 , so that we can replace @xmath117 and @xmath118 with their asymptotic values , whose general expression , valid for @xmath119 , is @xmath120 substitution in eq .
reveals that the first term vanishes , and the leverage correlation reduces to @xmath121 where the leverage decay time reads @xmath122 so , the model correctly forecasts the exponential decay of @xmath123 and its vanishing for negative correlation times .
the volatility autocorrelation provides an estimate of how much the volatility at time @xmath124 depends on the value it had at time @xmath0 and it is usually defined as @xmath125\ , { \mathrm{var}}[dx_{t+\tau}^2]}}\ , .
\label{eq : acorrdef}\ ] ] it is a well known stylized fact @xcite that @xmath126 decays with multiple time scales and in particular , it shows a long range memory effect , vanishing over a time scale of the order of few years for stock indexes . for the model under investigation ,
the volatility autocorrelation can be computed exactly too .
recalling again the novikov theorem and the fact that @xmath127 , the correlation entering the numerator of becomes @xmath128 } \,dw_{1,t } \right\rangle } \,dt^2 \,,\end{gathered}\ ] ] but , due to the presence of @xmath9 , the second term results to be of order @xmath129 and therefore it can be discarded .
the exact expression of the autocorrelation function @xmath130 can be obtained as explained in appendix [ app : autocorr ] , leaving us with @xmath131+\exp{\left[(a+c)\tau\right ] } \left [ \mu_4(t ) \!+\ ! \frac{2b}{a+c } \mu_3(t ) \!-\ !
\mu_2(t)\left ( \mu_2(t ) \!+\ !
\frac{2b}{a+c } \mu_1(t ) \right ) \right ] \right\}\,,\ ] ] where the denominator of eq . has been approximated with @xmath132=c^2\left [ 3\mu_4(t)-\mu_2(t)^2 \right]\,dt^2 $ ] in view of the stationary limit for @xmath2 .
after replacing the moments @xmath83 with their asymptotic expressions we end with @xmath133 \ , , \label{eq : acorrfinal}\ ] ] where the coefficients read @xmath134 and we also defined the two volatility autocorrelation time scales as @xmath135 at this point it is crucial to notice that in deriving eq .
and eq . we assumed implicitly that the moments of @xmath13 up to the order @xmath136 do converge asymptotically .
recalling the expression of the shape parameter @xmath37 in , we see this assumption imposes @xmath137 which has to be interpreted as a consistency relation for the model .
this constraint imposes the following strict ordering between the time scales of the model @xmath138 where the second inequality for @xmath139 follows from the convergence of third moment of @xmath13 which requires @xmath140 .
the expression obtained for @xmath126 fails to capture the persistence of this correlation identified in several analysis reviewed in @xcite .
the lacking of power law scaling would not be , in principle , a serious drawback as far as one of the two time scales involved in was sufficiently long .
however , the ordering , which is peculiar of the considered model , makes these scales too close each other and the volatility autocorrelation to decay as fast as @xmath141 , an undesired feature shared with other models , such as the stein - stein one .
the persistence of @xmath126 can be accounted for introducing a non linear volatility , as it is for the exponential ornstein - uhlenbeck model @xcite , or coupling a third stochastic equation driving the dynamics of the long run value of @xmath13 as in @xcite .
a further possibility to induce a non exponential time decay would be to consider a non linear drift term for the dynamics of @xmath13 , even though the analytical tractability of the present model will not be preserved .
now we provide a systematic methodology for estimating the model parameters , which are the constants @xmath142 , @xmath143 , @xmath144 entering the dynamics of @xmath13 , plus the correlation coefficient @xmath145 .
we perform the estimation over the standard & poor 500 ( s&p500 ) index daily returns from 1970 to 2010 , approximating @xmath45 with @xmath146 @xmath147 where @xmath148 ( one trading day ) . taking into account that @xmath9 is independent of @xmath8 and that @xmath149 is distributed accordingly to a folded normal law , the following relations hold for the model @xmath150 the constants @xmath151 and @xmath152 can be measured directly from the data , providing us an estimation of the ratio @xmath153 through the relation @xmath154 the value of these quantities extracted from the series of the daily returns of the s&p500 index
are reported in table [ tab : spotvalues ] .
it is crucial to observe that the value obtained for the ratio @xmath155 is compatible with the constraint , supporting the consistency of our model and the convergence of the volatility autocorrelation .
moreover , the same ratio provides an estimate of @xmath156 ( see eq . )
implying for the order of the highest converging moment a value @xmath157 ; consequently , relation indicates the following range for the tail index of @xmath158 @xmath159 the leverage correlation provides a way to obtain the two further relations needed to fix the four free parameters of the models .
indeed , a two parameters fit of the function @xmath160 gives estimates for the time scale @xmath161 and for the limit @xmath162 @xmath163 with the results reported in table [ tab : fitlev ] and fig .
[ fig : levfit ] . cccl estimators & & + a & & 0.1457 & @xmath164 + b & & 0.0295 & @xmath165 + c & & 0.0107 & @xmath166 + @xmath155 & & 1.7895 & + and @xmath167 . ] in particular , the value obtained for the leverage time scale , @xmath168 , and for its amplitude @xmath169 are consistent with those quoted in past analysis of different stock indexes such as the dow jones industrial average @xcite , and confirm the short range nature of this effect .
ccrl estimators & & + @xmath161 & & 0.0864 & yr + @xmath167 & & -30.9515 & + at this point all the parameters can be recovered through the following relations @xmath170^{-1 } \\ a & = c\ , d \\
b & = -\frac{a+c}{\sqrt{c } } \,\frac{c}{b } \\
\label{eq : paramsfinalrho } \rho & = - \frac { b \,(a+c)}{a \,(2 a + c ) } \,{\mathcal{l}}(0^+ ) \,.\end{aligned}\ ] ] ccrl parameter & & + @xmath142 & & @xmath171 & @xmath165 + @xmath143 & & @xmath172 & @xmath165 + @xmath144 & & @xmath173 & @xmath165 + @xmath145 & & @xmath174 & + the final results , reported in table [ tab : parsfinal ] , show a negative correlation coefficient , in agreement with the known leftward asymmetry of daily return distributions .
moreover , our calibration provides for the relaxation time of the volatility process a finite value @xmath175 , implying that , from a practical point of view , the limit @xmath44 is equivalent to @xmath176 .
the fitted values of @xmath161 and @xmath167 provides a good description of real data , as shown in fig .
[ fig : levfit ] ; on the other hand , fig .
[ fig : acorrth ] shows that the theoretical volatility autocorrelation for the estimated values of the parameters , eq . , does not capture the long range persistence of the empirical volatility , as expected from the constraints , while it describes correctly the exponential decay for small values of @xmath102 .
finally , it is important to compare the return pdf predicted by the model with the data sample from which the model parameters were estimated . since we model the return dynamics for increasing @xmath0 , it is even more important to asses to which extent the diffusion process is able to capture the scaling properties of the empirical distribution over different time horizons . at this aim , with the parameters fixed from the daily s&p500 series , we reconstruct the theoretical pdfs simulating the process at different time scales , @xmath177 , and we compare them with the corresponding empirical distributions obtained aggregating the daily returns .
this comparison is shown in fig .
[ fig : linpdf ] and fig .
[ fig : linlogpdf ] .
the daily distribution is very well reproduced by the theoretical pdf , which is able to fully capture the leptokurtic nature of the daily data .
the plots also confirm that the diffusive dynamics , once the parameters have been fixed at the daily scale , follows closely the evolution of the empirical curves for larger @xmath0 .
in particular , it captures the progressive convergence in the central region to a distribution with vanishing skewness and kurtosis .
in this work , we have introduced a class of svms where the volatility is driven by the general process with multiplicative noise analyzed in detail in @xcite .
more specifically , we focused on the set of parameters resulting in an inverse gamma stationary distribution for the @xmath8 process .
we provided an analytical characterization of the moments of the return distribution , revealing the role played by the power law behavior of the inverse gamma in the emergence of fat tails .
nevertheless , even though the highest order moments of @xmath40 diverge for every time lag , the analytical expressions we obtained reveal the vanishing of both the skewness and the kurtosis , in agreement with the normality of returns for long horizons .
as far as the estimation procedure is concerned , it is worth noticing that we do not exploit directly the statistical properties of the instantaneous volatility which is an hidden process , but on the contrary we infer the inverse gamma parameters from well established robust stylized facts holding at the daily scale .
indeed our model correctly predicts zero autocorrelation for the returns , and the short range exponential decay of the leverage .
the persistence of the volatility autocorrelation over yearly horizons is not captured , and in this perspective we would like to explore the possibility of coupling a third sde in the same spirit of @xcite .
moreover , we expect that relaxing the time homogeneity of the processes , as done in @xcite , we may induce time scalings more general than the exponential one .
we also expect the proposed dynamics to be a good candidate to describe the price and volatility dynamics even at higher frequencies .
this belief is supported by the empirical analysis discussed in the literature @xcite concerning the statistical properties of volatility proxies for intra - day data .
a further perspective would be to explore possible ways to characterize analytically the pdf associated to the process or its characteristic function .
this task requires to solve the fokker - planck equation for the pdf or its equivalent version in the fourier space , analogously to what has been done in @xcite for the heston case .
such a result would also allow for an application of the model in the context of market risk evaluation , possibly exploiting efficient fourier methodologies such as those proposed in @xcite .
here we report the explicit expressions of the coefficients @xmath70 entering the expansion of the moments of @xmath5 for the cases @xmath71 and @xmath72 . they were used to plot the analytical curves in fig .
[ fig : t0scaling ] .
where the coefficients @xmath180 and @xmath181 , entering the expansion of the moments of @xmath13 , read
after expressing @xmath182 in terms of its integral solution form @xmath0 to @xmath124 , the function @xmath183 can be rewritten in the form taking into account that for @xmath184 we can always split @xmath185 as the function @xmath183 becomes where we changed the variable of integrations to @xmath186 .
since the process @xmath187 is normally distributed with zero mean and variance @xmath188 , and recalling the expression of the gaussian characteristic function , @xmath189 , we can write application of the novikov theorem also gives where we expressed the wiener variation in terms of a gaussian white noise @xmath190 as @xmath191 . replacing the previous expressions in eq .
we conclude that @xmath183 has to satisfy which is a volterra equation of the second kind , whose solution leads to eq . .
with reference to the model , the cross correlation @xmath193 can be computed exactly . provided to express @xmath194 as integral solution from @xmath0 to @xmath124
it is straightforward to check that @xmath193 satisfies the following equation which is an ode provided that the correlation @xmath195 has been computed at the lower order @xmath196 . in particular , for the case @xmath197 , we need the following correlation whose substitution in eq . provides the solution we wish to acknowledge the anonymous referee for fruitful comments .
we sincerely thank guido montagna for useful discussions and comments about this work , and we also gratefully acknowledge the continuous support over past years of oreste nicrosini . | in this work we afford the statistical characterization of a linear stochastic volatility model featuring inverse gamma stationary distribution for the instantaneous volatility .
we detail the derivation of the moments of the return distribution , revealing the role of the inverse gamma law in the emergence of fat tails , and of the relevant correlation functions . we also propose a systematic methodology for estimating the parameters , and we describe the empirical analysis of the standard & poor 500 index daily returns , confirming the ability of the model to capture many of the established stylized fact as well as the scaling properties of empirical distributions over different time horizons . |
it is important to understand the mechanism of quark confinement in qcd .
we expect that color electric flux is squeezed by the dual meissner effect due to color magnetic monopole condensation@xcite .
it is well known that such a situation occurs in compact u(1 ) lattice gauge theory(lgt)@xcite . in qcd
, thooft conjectured that the role is played by the monopoles in abelian projected qcd which can be regarded as an abelian theory with abelian charges and magnetic monopoles@xcite .
we choose the maximally abelian ( ma ) gauge @xcite in which diagonal components of all link variables are maximized as much as possible by the gauge transformation .
gauge - fixed link variables are decomposed into a product of two matrices @xmath2 where @xmath3 , @xmath4 and @xmath5 are a gauge - fixing matrix , an off - diagonal and a diagonal matrix , respectively .
monte carlo simulations in this gauge have shown interesting results , i.e. , abelian dominance @xcite and interesting monopole behaviors @xcite .
these results show that the monopoles play an important role in qcd confinement , but evidence of the monopole condensation in the confinement phase is not yet obtained . in this report
, we show the ( dual ) meissner effect occurs in the presence of external charges also in qcd@xcite . in the continuum , the assumptions of the abelian dominance and the monopole condensation lead us to a dual form of the abelian ginzburg - landau ( g - l ) theory@xcite .
so we study the g - l type equations between the abelian electric field and the magnetic monopole currents in the presence of a static quark - antiquark pair in su(2 ) and su(3 ) , following the method developed by singh et al.@xcite .
our purpose is to apply the method to su(3 ) lgt as well as su(2 ) and to investigate scaling behaviors of the penetration length @xmath6 and the coherence length @xmath7 .
we used the wilson action and simulations are performed on @xmath8 lattice in su(2 ) ( @xmath9 and @xmath10 ) and on @xmath11 lattice in su(3 ) ( @xmath12 and @xmath13 ) .
measurements are done every 30 sweeps and we adopted totally 400 configurations in su(2 ) and 500 configurations in su(3 ) . for thermalization , we discarded inital 1500 sweeps .
abelian link variables @xmath5 are extracted after the gauge - fixing is done in the ma gauge .
the average of an observable @xmath14 is computed as @xmath15 where @xmath16 is an abelian wilson loop .
we used @xmath17 and @xmath18 in su(2 ) and @xmath17 in su(3 ) .
the electric field operator is defined by @xmath19 where @xmath20 is an abelian plaquette operator .
monopole currents are defined following degrand - toussaint@xcite .
we investigate the dual g - l equations@xcite : @xmath21 @xmath22 @xmath23 the order parameter @xmath24 is given by the following approximate solution of the g - l equation @xmath25 where we adopted @xmath26 .
then we get the following extended london equation : @xmath27 where @xmath28 is an integer and @xmath29 is the quantized electric flux at the origin @xmath30 .
to reproduce the @xmath31data , we choose the following function for the azimuthal component of the monopole current : @xmath32 where @xmath33 and @xmath34 are parameters .
the form ensures that @xmath35 vanishes linearly at @xmath30 and exponentialy as @xmath36 .
then , substituting the current into eq.[er ] , we examine the electric field data except the @xmath30 point and determine both the penetration and the coherence lengths .
the typical data of the curl of the monopole currents and the electric field are plotted in fig.1 ( @xmath1 ) .
if the extreme type-2 dual superconductor is the case , @xmath37 . then , the data with @xmath38 and @xmath39 can not satisfy eq.[er ] .
hence , there must be a region where @xmath40 , i.e. , we need both @xmath6 and @xmath7 also in the su(3 ) case .
the values of the both lengths obtained are plotted in fig.2 .
the fluxoid @xmath29 evaluated from the fluxoid relation is consistent with the net flux through the entire lattice in su(2 ) and su(3 ) .
the g - l parameter @xmath41 are shown in fig.3 .
the values are close to the borderline of the type-1 and type-2 dual superconductor .
it seems that the data of the large wilson loop show the scaling behaviors .
epsf = 0.5 = 0.5 [ fig : curl3 ] = 0.5 = 0.5 = 0.5 [ fig : scale1 ] = 0.5 = 0.5 [ fig : scale3 ]
we have a direct evidence of the dual meissner effect in @xmath0 and @xmath1 qcd . 1 . the color electric field and the magnetic monopole current satisfy the extended dual london equation in @xmath0 @xmath42 and @xmath1 @xmath43 lattice
qcd . the monopole current plays the role of the dual `` diamagnetic current '' squeezing the electric flux .
all @xmath6 and @xmath7 data show that the @xmath0 and @xmath1 vacua are both close to the borderline of the type-1 and type-2 dual superconductor .
so , it is expected that the vortex - vortex interactions are small .
it seems that the g - l parameter scales near the borderline in the case of the large wilson loop .
therefore , scalar and axial - vector glueball masses are nearly degenerate : @xmath44 4 .
it is interesting to extend this analysis to the study of the extended monopoles@xcite of the type-2 , since large monopole loops are seen to be important @xcite .
99 g.thooft,_high energy physics _ , ed.a.zichichi ( editorice compositori , bologna , 1975 ) . s. mandelstam , phys . rep .
23c ( 1976 ) 245 .
t.banks et al . , nucl .
b129 ( 1977 ) 493 .
g. thooft , nucl .
b190 ( 1981 ) 455 .
kronfeld et al . , phys .
lett . 198b ( 1987 ) 516 , + a.s .
kronfeld et al . , nucl.phys .
b293 ( 1987 ) 461 .
t. suzuki and i. yotsuyanagi , phys .
d42 ( 1990 ) 4257 .
s. hioki et al .
272b ( 1991 ) 326 .
t. suzuki , nucl .
b(proc . suppl . ) 30 ( 1993 ) 176 .
s.kitahara et al . , nucl .
b(proc . suppl . ) 30 ( 1993 ) 557 .
v. singh et al . , lsu preprint lsuhep- 1- 92 ( 1992 ) , + r.w .
haymaker , nucl .
b(proc . suppl . ) 30 ( 1993 ) 568 .
p. cea and l. cosmai , nucl .
suppl . ) 30 ( 1993 ) 572 .
t. suzuki , prog .
81 ( 1988 ) 929 .
+ s. maedan and t. suzuki , prog .
81 ( 1989 ) 229 .
+ t. suzuki , prog .
81 ( 1989 ) 752 .
+ s. maedan et al . , prog .
( 1990 ) 130 .
+ s. kamizawa et al . , nucl phys .
b389 ( 1993 ) 563 .
+ h. monden et al . , phys .
b294 ( 1992 ) 100 .
degrand and d. toussaint , phys .
d22 ( 1980 ) 2478 .
ivanenko et al . ,
phys . lett .
252b ( 1990 ) 631 .
h.shiba and t.suzuki , kanazawa university , report no .
kanazawa 93 - 09 , 1993 .
h.shiba and t.suzuki , kanazawa university , report no .
kanazawa 93 - 10 , 1993 . | after making an abelian projection in the maximally abelian gauge , we measure the distribution of abelian electric flux and monopole currents around an abelian wilson loop in @xmath0 and @xmath1 qcd . the ( dual ) meissner effect is observed clearly .
the vacua in the confinement phases of @xmath0 and @xmath1 are both at around the border between type-1 and type-2 ( dual ) superconductor . |
let @xmath4 be a cubic hypersurface , given as the zero locus of a cubic form @xmath5 $ ] .
this paper is concerned with the problem of determining when the set of rational points @xmath3 on @xmath0 is non - empty .
there is a well - known conjecture that @xmath6 as soon as @xmath7 .
in fact , for non - singular cubic hypersurfaces , it is expected that the hasse principle holds as soon as @xmath8 .
this states that in order for @xmath3 to be non - empty it is necessary and sufficient that @xmath9 is non - empty for every prime @xmath10 . for a large class of possibly singular cubic hypersurfaces @xmath11
, salberger has calculated the brauer group @xmath12 associated to a projective non - singular model @xmath13 of @xmath0 . a detailed proof of this calculation is provided by colliot - thlne in the appendix to this paper .
as a consequence of this investigation one has the following prediction .
let @xmath4 be a cubic hypersurface defined over @xmath2 which is not a cone , with @xmath8 and singular locus which is empty or of codimension at least @xmath14 in @xmath0 .
then the hasse principle holds for the locus of non - singular points on @xmath0 .
let us now consider some of the progress that has been made towards this conjecture .
when @xmath15 is diagonal it follows from the work of baker @xcite that @xmath0 has @xmath2-rational points as soon as @xmath16 . at the opposite end of the spectrum , when absolutely no assumptions are made about the shape of @xmath15 ,
a lot of work has been invested in producing a reasonable lower bound for the number of variables needed to ensure that @xmath6 . building on work of davenport @xcite
, heath - brown @xcite has recently shown that @xmath17 variables are enough to secure this property for an arbitrary cubic hypersurface defined over @xmath2 . in the light of this body of work
it is very natural to try and evince intermediate results in which the existence of rational points is guaranteed for cubic hypersurfaces in fewer than @xmath18 variables when mild assumptions are made about the structure of the hypersurface .
it is precisely this point of view that is the focus of the present investigation .
let @xmath19 denote the singular locus of @xmath0 , as a projective subvariety of @xmath0 .
when @xmath20 $ ] is non - singular , so that @xmath19 is empty , it follows from work of hooley @xcite that the hasse principle holds for @xmath0 provided that @xmath21 . as is well - known , the local conditions are automatic
when @xmath22 , and so @xmath3 is non - empty for non - singular @xmath0 provided that @xmath22 .
this fact was first proved by heath - brown @xcite .
when @xmath19 has dimension @xmath23 , joint work of the author @xcite with heath - brown shows that @xmath3 is non - empty provided that @xmath24 we will make use of this result shortly .
let @xmath25 be a positive integer .
we will say that an integral cubic form @xmath15 in @xmath26 variables `` splits off an @xmath27-form '' if there exist non - zero cubic forms @xmath28 with integer coefficients so that @xmath29 identically in @xmath30 .
we will merely say that @xmath15 `` splits off a form '' if @xmath15 splits off an @xmath27-form for some @xmath31 .
the following is our main result .
[ main ] let @xmath11 be a hypersurface defined by a cubic form that splits off a form , with @xmath32
. then @xmath6 .
the essential content of theorem [ main ] is that we can save @xmath33 variable in the result of heath - brown @xcite when the underlying cubic form splits off a form .
it should be stressed that the existence of a single @xmath2-rational point on @xmath0 is enough to demonstrate the @xmath2-unirationality of @xmath0 , and so the zariski density of @xmath3 in @xmath0 , when @xmath0 is geometrically integral and not a cone .
variants of this result have been known for a long time ( cf @xcite ) . in the generality
with which we have stated the result , it appears in the work of colliot - thlne and salberger ( * ? ? ?
* proposition 1.3 ) and in that of kollr @xcite .
our work has implications for the problem of determining when an arbitrary cubic form @xmath20 $ ] represents all non - zero @xmath34 , using rational values for the variables .
when this property holds we say that @xmath15 `` captures @xmath35 '' .
recall that a cubic form is said to be degenerate if the corresponding cubic hypersurface is a cone .
fowler @xcite has shown that any non - degenerate cubic form that represents @xmath36 automatically captures @xmath35 provided only that @xmath37 . hence it suffices to fix attention on those forms that do not represent zero non - trivially .
on multiplying through by denominators it will clearly suffice to establish that cubic forms of the shape @xmath38 represent zero non - trivially , with @xmath39 an arbitrary non - zero integer .
but this form splits off a @xmath33-form , and so is handled by theorem [ main ] . in this way we deduce the following result .
let @xmath20 $ ] be a non - degenerate cubic form , with @xmath40 .
then @xmath15 captures @xmath35 .
this result should be compared with the work of heath - brown @xcite , which implies that @xmath32 variables suffice .
it follows from the work of hooley @xcite that this may be improved to @xmath41 when @xmath15 is non - singular . as indicated in (
* appendix 1 ) the latter lower bound is probably the correct one for arbitrary cubic forms , since always has non - trivial @xmath10-adic zeros for @xmath26 in this range .
let @xmath42 . when @xmath11 is a hypersurface defined by a cubic form that splits off a form , we are able to handle fewer variables when appropriate assumptions are made about one of the forms .
if @xmath0 is a cone then we will see in lemma [ lem : p1.1 ] that @xmath6 .
if , on the other hand , @xmath0 is not a cone let us consider the effect of supposing that the underlying cubic form splits off a non - singular @xmath27-form @xmath43 .
if @xmath44 then @xmath0 is itself non - singular and we automatically have @xmath6 when @xmath22 . if @xmath45 then the residual form @xmath46 defines a projective cubic hypersurface of dimension @xmath47 , and as such has singular locus of dimension at most @xmath48 .
but then it follows that @xmath0 has singular locus of dimension at most @xmath48 .
thus we may deduce from @xcite that @xmath6 provided that @xmath49 .
we record this observation in the following result .
[ t : goat ] let @xmath11 be a hypersurface defined by a cubic form that splits off a non - singular @xmath27-form , with @xmath50 and @xmath22
. then @xmath6 .
it would be interesting to reduce the range of @xmath27 needed to ensure the validity of theorem [ t : goat ] .
ours is not the first attempt to better understand the arithmetic of cubic hypersurfaces that split off a form .
indeed , in davenport s @xcite treatment of cubic forms in @xmath51 variables , a fundamental ingredient in the treatment of certain bilinear equations is a separate analysis of those forms that split into two .
in further work , colliot - thlne and salberger @xcite have shown that the hasse principle holds for any cubic hypersurface in @xmath52 that contains a set of three conjugate singular points , provided only that @xmath42 .
given a cubic extension @xmath53 of @xmath2 , define the corresponding norm form @xmath54 where @xmath55 is a basis of @xmath53 as a vector space over @xmath2 .
in view of the fact that the local conditions are automatically satisfied for cubic forms in at least @xmath56 variables , we observe the following easy consequence . [ main ] let @xmath11 be a hypersurface defined by a cubic form that splits off a norm form , with @xmath22 .
then @xmath6 .
it turns out that theorem [ main ] will play a useful rle in dispatching some of the cases that arise in the proof of theorem [ main ] . following the strategy of birch , davenport and lewis @xcite
, it would however be straightforward to adapt the proof of theorem [ main ] to retrieve theorem [ main ] .
an obvious further line of enquiry would be to investigate cubic hypersurfaces that split off two forms , by which we mean that the corresponding cubic form can be written as @xmath57 identically in @xmath30 , for appropriate @xmath58 . with the extra structure apparent in such hypersurfaces one would like to determine the most general conditions possible under which the conjectured value of @xmath22 variables suffices to ensure the existence of @xmath2-rational points .
one of the remarkable features of our argument is the breadth of tools that it draws upon .
the underlying machinery is the hardy littlewood circle method , and we certainly take advantage of many of the contributions to the theory of polynomial cubic exponential sums that have been made during the last fifty years .
these are detailed in
[ sec : exp_sums ]
. a further component of our work involves a detailed analysis of the case in which one of the forms that splits off in theorem [ main ] is singular and has a relatively small number of variables . to deal with this scenario it pays to reflect upon the classification of singular cubic hypersurfaces
this is a very old topic in algebraic geometry , and can be traced back to the pioneering work of cayley @xcite and schlfli @xcite .
all of the necessary information will be collected together in [ s : geom ] .
the final ingredient in our work comprises good upper bounds for the number of @xmath2-rational points of bounded height on auxiliary cubic hypersurfaces .
the estimates that we will take advantage of are presented in [ sec : upper ] .
when it is applicable , the hardy littlewood circle method allows us to show that @xmath59 for a given cubic hypersurface @xmath60 by evaluating asymptotically the number of @xmath2-rational points of bounded height on @xmath0 .
it is a well - known but intriguing feature of the method that one can achieve such precise information by first establishing weaker upper bounds for the growth rate of @xmath2-rational points on appropriate auxiliary varieties .
in fact , we will show in lemma [ lem:8 ] that the @xmath2-rational points on a non - singular cubic hypersurface @xmath11 satisfy the growth bound @xmath61 for any @xmath62 , provided that @xmath63 . here
@xmath64 is the usual exponential height function .
this should be compared with the manin conjecture @xcite which predicts that the exponent of @xmath65 should be @xmath66 as soon as @xmath67 . throughout our
work @xmath68 will denote the set of positive integers . for any @xmath69
, we will follow common convention and write @xmath70 and @xmath71 .
the parameter @xmath72 will always denote a very small positive real number .
we will use @xmath73 to denote the norm @xmath74 of a vector @xmath75 , whereas @xmath76 will be reserved for the usual euclidean norm @xmath77 .
all of the implied constants that appear in this work will be allowed to depend upon the coefficients of the cubic forms under consideration and the parameter @xmath78 .
any further dependence will be explicitly indicated by appropriate subscripts .
it is a pleasure to thank professor colliot - thlne , professor heath - brown , professor salberger and professor wooley for several useful discussions . while working on this paper the author
was supported by epsrc grant number ` ep / e053262/1 ` .
the proof of theorem [ main ] will depend intimately on the dimension of the hypersurfaces defined by the constituent cubic forms , and the nature of their singularities
. a key step will be to determine conditions on this singular locus under which the hypersurface automatically has rational points .
let @xmath20 $ ] be an arbitrary cubic form , which we assume takes the shape @xmath79 in which the coefficients @xmath80 are symmetric in the indices @xmath81 .
define the @xmath82 matrix @xmath83 with @xmath84-entry @xmath85 we will say that the cubic form @xmath15 is `` good '' if for any @xmath86 and any @xmath78 we have the upper bound @xmath87 for each integer @xmath88 .
a crucial step in davenport s @xcite treatment of general cubic forms is a proof of the fact that forms that fail to be good automatically possess non - trivial integer solutions for `` geometric reasons '' .
our approach has a similar flavour , although the underlying arguments will be more obviously geometric .
assume throughout this section that @xmath89 and @xmath60 is a hypersurface defined by a cubic form @xmath20 $ ] .
a lot of the facts that we will record are classical .
suppose for the moment that @xmath15 is not absolutely irreducible .
then either it has a linear factor @xmath90 defined over @xmath2 , or it is a product @xmath91 of three linear factors that are conjugate over @xmath92 . by considering the equation @xmath93 in the former case , we deduce that @xmath59 . in the latter case ,
we arrive at the same conclusion when @xmath42 , by considering the system of equations @xmath94 .
when @xmath95 and @xmath15 is a product of three conjugate factors we deduce that @xmath0 has precisely three conjugate singular points .
when @xmath37 and @xmath0 is defined by an absolutely irreducible cubic form , but is a cone , we note that the space of vertices on @xmath0 must be a linear space globally defined over @xmath2 .
thus @xmath6 in this case too .
we have therefore established the following simple result .
[ lem : p1.1 ] let @xmath96 . if @xmath0 is not geometrically integral , or if @xmath0 is a cone , then @xmath6 .
when @xmath95 the same conclusion holds unless @xmath0 contains precisely three conjugate singular points .
recall that a cubic hypersurface @xmath0 is said to be non - singular if over @xmath97 the only solution to the system of equations @xmath98 has @xmath99
. henceforth we will be predominantly interested in singular cubic forms , and then only in the cases @xmath100 and @xmath101 .
let @xmath102 be a field .
it has been conjectured by cassels and swinnerton - dyer that any cubic hypersurface @xmath103 defined over @xmath102 that contains a @xmath102-rational @xmath36-cycle of degree coprime to @xmath104 , automatically has a @xmath102-rational point .
the case @xmath95 goes back to poincar .
when the singular locus is non - empty , the case @xmath105 can be deduced from the work of skolem @xcite .
a comprehensive discussion of the arithmetic of singular cubic surfaces can be found in the work of coray and tsfasman @xcite .
coray @xcite has established the conjecture for all local fields and , in a subsequent investigation ( * ? ? ?
* proposition 3.6 ) , has also dispatched the case in which @xmath106 and the @xmath36-cycle is made up of double points .
suppose first that @xmath95 , so that @xmath107 defines a cubic curve , which we assume to be geometrically integral and not a cone .
when @xmath0 is singular it contains exactly one singular point , which must therefore be defined over @xmath2 .
once combined with lemma [ lem : p1.1 ] we arrive at the following result .
[ lem : curve ] let @xmath95 and suppose that @xmath108
. then 1 .
@xmath0 is non - singular ; or 2 .
@xmath0 contains precisely three conjugate singular points . in case ( ii ) of lemma [ lem : curve ]
one concludes that the underlying cubic form can be written as a norm form , for appropriate @xmath109 , where @xmath53 is the cubic number field obtained by adjoining one of the singularities .
we turn now to the case @xmath105 of cubic surfaces @xmath110 , which we suppose to be geometrically integral and not equal to a cone .
suppose that @xmath0 is singular .
the classification of such cubic surfaces can be traced back to cayley @xcite and schlfli @xcite , but we will employ the modern treatment found in the work of bruce and wall @xcite .
in particular the singular locus of @xmath0 is either a single line , in which case @xmath0 is ruled , or else it contains @xmath111 isolated singularities and these are all rational double points .
it follows that @xmath6 unless @xmath112 and the three singular points are conjugate to each other over a cubic extension of @xmath2 . in this final case
, skolem @xcite showed that @xmath15 can be written as @xmath113 for appropriate coefficients @xmath109 and @xmath114 , where @xmath53 is the cubic number field obtained by adjoining one of the singularities to @xmath2 . in terms of the classification over @xmath92 according to singularity type , the only possibility here is that @xmath0 has singularity type @xmath115 for @xmath116 or @xmath117 , since the action of @xmath118 preserves the singularity type . bringing this all together ,
we have therefore established the following analogue of lemma [ lem : curve ] . [ lem : surface ] let @xmath105 and suppose that @xmath108
. then 1 .
@xmath0 is non - singular ; or 2 .
@xmath0 contains precisely three conjugate double points .
we now try to construct a version of lemmas [ lem : curve ] and [ lem : surface ] for the case @xmath106 .
let @xmath119 denote the singular locus of @xmath120 , a variety of dimension at most @xmath117 . as usual
we assume that @xmath0 is geometrically integral and not a cone .
we analyse @xmath13 by considering the intersection of @xmath0 with a generic hyperplane @xmath121 . in particular
the hyperplane section @xmath122 is a geometrically integral cubic surface which is not a cone ( see ( * ? ? ?
* proposition 18.10 ) , for example ) . in taking @xmath123 to be defined over @xmath2
, we may further assume that @xmath124 is defined over @xmath2 .
any @xmath2-rational point on @xmath124 visibly produces a @xmath2-rational point on @xmath0 .
let @xmath125 denote the singular locus of @xmath124 .
then the classification of cubic surfaces implies that @xmath126 is either empty or it is a union of @xmath127 points or it is a line .
when @xmath126 is non - empty it follows from lemma [ lem : surface ] that either @xmath128 or else @xmath126 is finite , with @xmath129 and the three points being conjugate to each other over @xmath92
. now an application of bertini s theorem ( in the form given by harris ( * ? ? ?
* theorem @xmath130 ) , for example ) shows that @xmath131 when @xmath124 is non - singular it therefore follows that @xmath132 is empty for generic @xmath121 , whence @xmath13 must be finite . in the alternative case , when @xmath124 is singular , we may conclude that @xmath133 for generic @xmath121 , whence the maximal component of @xmath13 is a cubic curve .
let us examine further the possibility that the singular locus @xmath13 of @xmath0 has dimension @xmath33 , and that it contains a cubic curve @xmath134 as its component of maximal dimension . clearly @xmath134 is defined over @xmath2 .
furthermore , we may conclude from bzout s theorem that the line connecting any two points of @xmath134 must be contained in @xmath0 , since each such point is a singularity of @xmath0 .
if @xmath134 is reducible over @xmath2 then there are two basic possibilities : either it is a union of lines or it is a union of a conic and a line . in the latter case
@xmath134 contains a line defined over @xmath2 and it trivially follows that @xmath135 .
the former case fragments into a number of subcases : either it is a union of @xmath104 concurrent lines , or it contains a pair of skew lines , or it is a union of @xmath104 coplanar lines , or it contains a repeated line . the second case is impossible since then the join of the two skew lines defines a @xmath104-plane that would also be contained in @xmath0 , contradicting the fact that @xmath0 is geometrically irreducible .
it follows from consideration of the galois action on @xmath134 that @xmath135 in every case apart from the one in which @xmath134 is a union of @xmath104 coplanar lines .
if @xmath134 is geometrically irreducible then it can not be a twisted cubic since then the secant variety @xmath136 would be contained in @xmath0 .
our argument so far has shown that either @xmath135 for trivial reasons , or else @xmath134 is a cubic plane curve that is either geometrically irreducible or a union of @xmath104 distinct lines .
the plane @xmath65 containing @xmath134 is defined over @xmath2 and , after carrying out a linear change of variables , we may take @xmath137 as its defining equations .
but then it follows that the cubic form defining @xmath0 can be written @xmath138 for appropriate quadratic forms @xmath139 defined over @xmath140 . with this notation
one sees that @xmath13 is the locus of solutions to the system of equations @xmath141 in @xmath142 .
it is now clear that the component @xmath134 of @xmath13 of maximal dimension can not be a cubic plane curve of the two remaining types .
it remains to deal with the case in which the singular locus @xmath13 of @xmath0 is finite and globally defined over @xmath2 .
as shown by c. segre @xcite , we have @xmath143 , the extremal case of @xmath56 singular points being achieved by the so - called segre threefold . since @xmath0 is assumed not to be a cone so we may assume that all of the singularities are double points . indeed any singularity with multiplicity exceeding @xmath117 must be a vertex for @xmath0 .
in fact , when @xmath144 it is known ( * ? ? ?
* lemma 2.2 ) that all the singularities are actually nodal .
appealing to coray s partial resolution of the cassels swinnerton - dyer conjecture for threefolds , we are now ready to record our analogue of lemmas [ lem : curve ] and [ lem : surface ] . [ lem:3fold ] let @xmath106 and suppose that @xmath108
. then 1 .
@xmath0 is non - singular ; or 2 .
@xmath0 is a geometrically integral cubic hypersurface whose singular locus contains precisely @xmath145 double points , with @xmath146 . in the second case of lemma [ lem:3fold ] it follows from @xcite and @xcite that the hasse principle holds for @xmath0 when @xmath112 or @xmath147 .
our investigation would be made easier if we were also in possession of this fact when @xmath148 .
lacking this , all that we actually require from part ( ii ) of lemma [ lem:3fold ] is that the singular locus should be finite . in his survey of open problems in diophantine geometry ,
lewis @xcite reports on unpublished joint work with blass , which would appear to give lemma [ lem:3fold ] .
however , in the absence of subsequent elucidation , we have chosen to present our own proof of this result .
let @xmath20 $ ] be an arbitrary cubic form , assumed to take the shape .
our work in this section centres upon various properties of the cubic exponential sums @xmath149 for a suitable family of weights @xmath150 on @xmath151 , and cubic forms that are always either good ( in the sense of the previous section ) or the hypersurface they define has finite ( possibly empty ) singular locus .
specifically , we will collect together some general upper bounds for @xmath152 , some estimates for suitable moments of @xmath152 and some asymptotic formulae for @xmath152 when suitable assumptions are made about how @xmath153 can be approximated by rational numbers .
all of these estimates will depend on the parameter @xmath65 which should be thought of as tending to infinity .
we must begin by saying a few words about the weight functions that we will be working with .
let @xmath154 such that @xmath155 .
when @xmath156 we let @xmath157 be certain vectors , which we think of as being fixed , but whose nature will be determined later . similarly we let @xmath158 .
all of the estimates in our work will be allowed to depend upon the choice of @xmath159 and @xmath160 .
define @xmath161 , via @xmath162 where we have written @xmath163 .
let @xmath164 .
then @xmath165 is exactly the weight function introduced by heath - brown in @xcite .
note that @xmath166 so that @xmath167 for any @xmath168
. next we let @xmath169 denote the characteristic function @xmath170 where @xmath171 .
each weight @xmath150 appearing in our work will either be of the form @xmath172 or @xmath173 or @xmath174 , depending on context . to help distinguish which estimates are valid for which choice of weight function , let us denote by @xmath175 the set of non - negative weight functions on @xmath151 that are of the shape , and let @xmath176 denote the corresponding set of weight functions on @xmath151 of the type .
we let @xmath177 denote the set of mixed functions @xmath174 , with @xmath178 for @xmath179 .
in particular @xmath180 for @xmath179 . in the definition of these sets
the precise value of @xmath159 or @xmath160 is immaterial , unless explicitly indicated otherwise , and the corresponding implied constants will always be allowed to depend on these quantities in any way .
we are now ready to record the upper bounds for @xmath152 that feature in our investigation .
[ lem : b ] let @xmath78 , let @xmath181 and assume that @xmath20 $ ] is a good cubic form .
let @xmath182 such that @xmath183 and @xmath184 .
then if @xmath185 we have @xmath186 if furthermore @xmath187 , then we have @xmath188 this is the essential content of the investigation of davenport @xcite into cubic forms in @xmath51 variables .
the bounds are derived in a more succinct manner in heath - brown @xcite .
the fact that we are working with exponential sums that are differently weighted makes no difference to the validity of the argument , and the reader may wish to consult @xcite , where the necessary modifications can be found in the setting of quartic forms .
define the complete exponential sum @xmath189 for any coprime integers @xmath190 such that @xmath191 .
it can easily be deduced from the proof of lemma [ lem : b ] that @xmath192 for any @xmath78 , under the assumption that the cubic form is good .
the following improvement is due to heath - brown @xcite .
[ lem : paris ] let @xmath78 and assume that @xmath20 $ ] is a good cubic form .
then we have @xmath193 .
we now come to the real workhorse in our argument .
given @xmath194 and @xmath195 we define @xmath196 the following result provides an upper bound for this quantity .
[ lem : d ] let @xmath78 , let @xmath181 and assume that @xmath20 $ ] is a good cubic form .
let @xmath194 , with @xmath197 and @xmath198 .
then for any @xmath199 $ ] and
any @xmath200\cap { \mathbb{z}}$ ] we have @xmath201 where @xmath202 it is clear that @xmath203 hence it follows from hlder s inequality that @xmath204 on employing heath - brown s estimate for @xmath205 , which follows from ( * ? ? ?
( 4.5 ) and ( 5.1 ) ) , we therefore deduce that @xmath206 as in the deduction of lemma [ lem : b ] , the fact that we are working with differently weighted exponential sums makes no difference to the final outcome of the argument . using the fact that @xmath197 and @xmath198 , with @xmath207
, it easily follows that the term involving @xmath208 contributes @xmath209 since @xmath210 .
this completes the proof of the lemma .
lemma [ lem : d ] is based on an averaged version of van der corput s method and comprises the key innovation in the work of heath - brown @xcite already alluded to .
although we have presented it in the context of denominators @xmath211 and values of @xmath212 restricted to dyadic intervals , the general result consists of a bound for @xmath213 , where the integral is taken over a certain set of minor arcs . for cubic forms in few variables we will have better results available .
when @xmath214 and @xmath215 , hua s inequality ( * ? ? ?
* lemma 3.2 ) implies that @xmath216 for any @xmath217 .
the following result is due to wooley @xcite , and generalises this to binary forms .
[ lem : hua_2 ] let @xmath78 , let @xmath218 and let @xmath219 $ ] be a binary cubic form , not of the shape @xmath220 , for integers @xmath221 .
then we have @xmath222 for any @xmath217 .
our next selection of results concern the approximation of @xmath152 on a certain set of arcs in the interval @xmath223 $ ] . for given @xmath224 ,
define @xmath225 to be set of @xmath226 $ ] for which there exists @xmath182 such that @xmath227 and @xmath184 , with @xmath228.\ ] ] the major arcs in our work will be a subset of these , but it will be useful to maintain a certain degree of generality .
when dealing with cubic forms whose singular locus is very small , we have rather good control over the approximation of @xmath152 on the arcs @xmath229 , provided that we work with the class of smooth weight functions @xmath230 .
recall the definition of @xmath231 from and let @xmath232 for @xmath233
. we will need to work with the familiar quantity @xmath234 concerning which we have the following result .
[ lem : f ] let @xmath78 and @xmath37 .
assume that @xmath5 $ ] is a good cubic form defining a projective hypersurface that is not a cone , with singular locus of dimension @xmath235 .
let @xmath224 such that @xmath236 and @xmath237 , and let @xmath238
. then there exists @xmath239 such that @xmath240 furthermore , if @xmath241 and @xmath242 , then we have @xmath243 the proof of this result is based on the investigation carried out by heath - brown @xcite into non - singular cubic forms in @xmath56 variables
. one of the key ingredients in his approach is the poisson summation formula , and it is this part of the argument that we plan to take advantage of .
we begin by choosing @xmath244 to be a point at which the matrix of second derivatives of @xmath15 has full rank at @xmath245 .
the existence of such a point follows from the work of hooley ( * ? ? ?
* lemma 26 ) . with this choice of @xmath245
we now select @xmath150 to be the weight function in , which belongs to @xmath230 .
let @xmath246 for any @xmath247 , and let @xmath248 .
then ( * ? ? ?
* lemma 8) yields @xmath249 where @xmath250 and furthermore , @xmath251 for any @xmath252 . the main difference between what we have recorded here and the statement of ( * ? ? ?
* lemma 8) is that our definition of @xmath152 does not involve a summation over @xmath39 .
this deviation makes no difference to the final outcome .
note that once the existence of a suitable point @xmath245 is established for the definition of the weight function , the manipulations involving the exponential integral remain valid even when @xmath15 is singular .
the summation over @xmath253 in our upper bound for @xmath254 implies in particular @xmath255 . since @xmath236
we automatically have @xmath256 .
hence the condition @xmath255 implies that @xmath257 and @xmath258 putting everything together it follows that @xmath259 where @xmath260 we will show that @xmath261 for @xmath235 . before doing so
let us see how this suffices to complete the proof of the first part of the lemma .
recalling from above that @xmath262 has order of magnitude @xmath263 , and employing , we deduce that @xmath264 if @xmath265 then the first term here is @xmath266 , since @xmath267 .
alternatively , if @xmath268 , then the first term is @xmath269 this establishes the first part of the lemma subject to . to establish we return to the manipulations in @xcite
things are simplified slightly by no longer needing to keep track of the dependence on @xmath15 in each implied constant .
in particular we may take @xmath270 throughout .
the sum @xmath271 satisfies a basic multiplicativity property , as recorded in ( * ? ? ?
* lemma 10 ) .
write @xmath272 , where @xmath273 in particular @xmath274 and we deduce from ( * ? ? ? * lemmas 7 , 10 and 11 ) that @xmath275 here , if @xmath83 denotes the matrix of second derivatives of @xmath276 , then @xmath277 is the number of @xmath278 modulo @xmath27 such that @xmath279 . recalling the notation of ( *
* lemma 12 ) , in which we take @xmath280 and @xmath281 , it follows that there is an absolute constant @xmath282 such that @xmath283 we would now like a version of ( * ? ? ?
* lemma @xmath51 ) which applies to singular forms as well .
we claim that @xmath284 this relies completely on first establishing suitable analogues of ( * ? ? ?
* lemmas 13 and 14 ) .
a little thought reveals that in the present setting we have @xmath285 for any @xmath286 and @xmath287 .
here we have used the fact that @xmath15 is good to bound the number of @xmath288 such that @xmath289 , rather than using ( * ? ? ?
* lemma 2 ) , as there .
furthermore , we have @xmath290 when @xmath291 it follows from the latter bound and an application of cauchy s inequality that @xmath292 , which is acceptable for . in the alternative case ,
when @xmath293 , the necessary modifications to the proof of ( * ? ? ?
* lemma @xmath51 ) are straightforward and we omit full details here
. we may now insert into the preceeding estimate for @xmath294 to conclude that @xmath295 if @xmath296 then this is clearly satisfacory for . alternatively ,
if @xmath297 then we can only enlarge our bound for @xmath294 if we replace @xmath262 by @xmath298 .
but then @xmath294 is easily seen to be bounded by in this case too .
this therefore completes the proof of .
our final task is to establish the second part of the lemma .
since @xmath15 is good , we may combine lemma [ lem : paris ] with to deduce that @xmath299 let us write @xmath300 for convenience .
it therefore follows that @xmath301 since @xmath241 and @xmath242 .
we remark that when @xmath302 it seems likely that an even sharper error term is available in lemma [ lem : f ] through a more careful analysis of the complete exponential sums @xmath271 , when @xmath211 is prime .
it follows from ( * ? ? ?
* lemma 28 ) that the form @xmath15 is automatically good when the corresponding hypersurface has at most isolated singularities and these are suitably mild . in the setting of @xmath33-dimensional exponential sums
, we have even better control over @xmath152 on the arcs @xmath229 .
let @xmath303 for some non - zero coefficient @xmath304 .
then for any @xmath182 such that @xmath227 and @xmath184 , and any @xmath305 , the standard major arc analysis would provide an estimate of the shape @xmath306 where @xmath307 is given by .
our final result in this section improves on this substantially , and is readily derived from the book of vaughan @xcite .
[ lem : e ] let @xmath78 , let @xmath214 and let @xmath308 .
let @xmath224 with @xmath309 .
then for any @xmath238 we have @xmath310 furthermore , if @xmath311 , then we have @xmath312
let @xmath103 be a cubic hypersurface , not equal to a cone , that is defined by an absolutely irreducible cubic form @xmath313 $ ] . for @xmath62 , let @xmath314 according to the conjecture of manin @xcite one expects @xmath315 for some constant @xmath316 as soon as @xmath317 is non - singular and @xmath8 .
when @xmath317 is not necessarily non - singular , or the number of variables is small , there is the dimension growth conjecture due to heath - brown .
this predicts that @xmath318 and has received a great deal of attention in recent years .
let @xmath319 denote the projective dimension of the singular locus of @xmath0 .
the dimension growth conjecture has been established by the author @xcite when @xmath320 .
the following result , which may of independent interest , shows that one can do better than if larger values of @xmath26 are permitted .
[ lem:8 ] we have @xmath321 when @xmath322 . our proof of the lemma is based on the approach developed in @xcite . arguing with hyperplane sections , as in @xcite
, we see that it will suffice to show that there is an absolute constant @xmath323 such that @xmath324 for any weight function @xmath325 belonging to the class of weight functions described at the start of @xcite , any @xmath326 , and any cubic polynomial @xmath327 $ ] such that @xmath41 and the cubic part @xmath328 is non - singular .
here we recall that @xmath329 , where @xmath330 denotes the height of a polynomial @xmath331 .
the bulk of @xcite goes through verbatim , and we are left with reevaluating the estimation of @xmath332 and @xmath333 in @xcite and @xcite , respectively . beginning with the former , we note from ( * ? ? ?
* eq . ( 5.5 ) ) that this breaks into an estimation of @xmath334 and @xmath335 .
the first of these is estimated as @xmath336 .
both of the exponents of @xmath65 are clearly at most @xmath337 when @xmath41 , as required for . turning to @xmath335 ,
one easily traces through the argument , finding that @xmath338 where @xmath339 here the first term ( resp .
second term , sum of the final two terms ) corresponds to the case @xmath340 ( resp .
@xmath341 , @xmath342 ) .
a modest pause for thought reveals that all of these exponents are satisfactory when @xmath41 .
we now turn to the estimation of @xmath343 in @xcite , which is again written as a sum @xmath344 . beginning with @xmath345
, we easily observe that @xmath346 this time with @xmath347 this therefore shows that @xmath348 , as required for . turning to @xmath349
, we will need to modify the argument slightly . on noting that @xmath350
, we easily deduce that @xmath351 where we have set @xmath352 the first and second terms here are satisfactory for @xmath41 . moreover the third term is clearly satisfactory for @xmath353 , on taking @xmath354 . to handle the contribution from the final term when @xmath355 , it will be convenient to recall that @xmath262 has order of magnitude @xmath356 when @xmath357 and @xmath358 when @xmath359 .
suppose first that @xmath360 .
when @xmath357 we deduce that @xmath361 on taking @xmath362 this is clearly satisfactory for @xmath41 .
when @xmath359 we easily deduce that the same bound holds on taking @xmath262 to be of size @xmath358 in the definition of @xmath363 .
suppose now that @xmath364 and @xmath357 .
if @xmath365 , then it is not hard to see that @xmath366 using .
this is satisfactory for @xmath41 . alternatively , if @xmath367 , then one finds that @xmath368 using it easily follows that @xmath369 since @xmath370 and @xmath364 this is clearly satisfactory when @xmath371 . if instead @xmath21 then we deduce that @xmath372 on taking @xmath373 rather than . since @xmath364 we easily conclude that @xmath374 in this case too . finally , when @xmath364 and @xmath359 , we see that @xmath375 using
this therefore concludes the proof of the lemma .
it is clear from the proof of lemma [ lem:8 ] that one actually achieves an estimate of the shape @xmath376 for a constant @xmath323 , when @xmath322 .
it seems likely that one can push the analysis further , obtaining @xmath377 for @xmath378 , as predicted by manin .
a key step in our argument involves generating good estimates for the moments @xmath379 where @xmath152 is the cubic exponential sum , for an appropriate weight @xmath380 . by the orthogonality of the exponential function we have @xmath381 it is clear that there exists a constant @xmath382 depending on @xmath150 such that the overall contribution to @xmath383 from @xmath384 such that @xmath385 is @xmath386 , if @xmath65 is taken to be sufficiently large .
hence it follows that @xmath387 when @xmath15 is a non - singular form in @xmath26 variables it is obvious that @xmath388 is a non - singular form in @xmath389 variables , defining a hypersurface of dimension @xmath390 . when @xmath15 has a finite non - empty singular locus it is not hard to see that @xmath388 has singular locus of dimension @xmath33 .
the following result now flows very easily from and lemma [ lem:8 ] .
[ lem : c ] let @xmath78 and let @xmath37 .
assume that @xmath20 $ ] is a cubic form defining a projective hypersurface whose singular locus has dimension @xmath391 .
then we have @xmath392
in this section we establish theorem [ main ] .
let @xmath393 such that @xmath394 it will be convenient to write @xmath395 and @xmath396 .
we henceforth fix our attention on cubic forms of the shape @xmath397 with @xmath398 $ ] and @xmath399 $ ] . in what follows
we may always suppose that @xmath400 is non - degenerate , by which we mean that it is not equivalent over @xmath140 to a cubic form in fewer variables , since such forms have obvious non - zero integral solutions .
recall the definition of `` good '' cubic forms from [ s : geom ] .
it follows from @xcite that either @xmath401 is good or else the cubic hypersurface @xmath402 has a rational point .
the same is true for the cubic forms @xmath46 and @xmath403 . since the existence of a rational point on any of these hypersurfaces is enough to ensure that @xmath6 in the statement of theorem [ main ] , so we may proceed under the assumption that @xmath404 and @xmath403 are all good .
let @xmath405 , as introduced in
[ sec : exp_sums ] .
when @xmath401 satisfies the hypotheses in lemma [ lem : f ] we will assume that @xmath406 is the weight function constructed there .
our argument revolves around establishing an asymptotic formula for the sum @xmath407 as @xmath408 . as is usual in applications of the hardy littlewood circle method ,
the starting point is the simple identity @xmath409 where @xmath410 for @xmath179
. it will be convenient to define @xmath411 where @xmath174 .
then @xmath412 is a cubic exponential sum of the sort introduced in . in the usual way one divides the interval @xmath223 $ ] into a set of major arcs and minor arcs . for major arcs
we will take the union of intervals @xmath413,\ ] ] which is equal to @xmath414 in the notation of .
the corresponding set of minor arcs is defined modulo @xmath33 as @xmath415\setminus { \mathfrak{m}}$ ] . here
@xmath416 is an arbitrary small parameter .
it turns out the choice @xmath417 is acceptable .
we may deduce from lemma 15.4 and
1618 in @xcite ( see also ( * ? ? ?
* lemma 2.1 ) ) that @xmath418 where @xmath419 are both absolutely convergent .
here , the absolute convergence of @xmath420 follows from lemma [ lem : paris ] , and we have written @xmath421 for @xmath179 . since @xmath420 is
absolutely convergent and @xmath400 is non - degenerate , it follows from standard arguments ( see ( * ? ? ?
* lemma 7.3 ) , for example ) that @xmath422 .
the treatment of the singular integral is routine and we omit giving the details here , all of which can be supplied by consulting @xcite and @xcite . assuming that neither @xmath401 nor @xmath46 has a linear factor defined over @xmath2 it is possible to choose @xmath423 in the definition of @xmath174 , so that each @xmath424 is a non - singular real solution to @xmath425 . on selecting a sufficiently small value of @xmath158 in the definition of @xmath173 we
can then ensure @xmath426 .
the case in which @xmath401 or @xmath46 does factorise over @xmath2 clearly enables us to deduce the statement of theorem [ main ] very easily . in order to conclude the proof of theorem [ main ] it remains to show that the overall contribution from the minor arcs @xmath427 is satisfactory
this is where the bulk of our work lies and we will find it necessary to undertake a lengthy case by case analysis to handle the different values of @xmath428 and @xmath429 . in doing so it will suffice to handle the case @xmath430 , the case @xmath431 being taken care of by @xcite . without loss of generality
we assume henceforth that @xmath432 .
let @xmath433 and let @xmath434 . by dirichlet
s approximation theorem we may find coprime integers @xmath435 such that @xmath436 and @xmath437 .
the value of @xmath438 should satisfy @xmath439 and is chosen to optimise the final stages of the argument .
the obvious approach involves applying estimates for each individual exponential sum @xmath440 and @xmath441 for @xmath434 , before then deriving an estimate for the integral over the full set of minor arcs .
while we have rather good control over these sums when @xmath428 and @xmath429 are both large , the case in which one of @xmath428 or @xmath429 is small presents more of an obstacle .
instead we apply hlder s inequality to deduce that @xmath442 for any @xmath443 such that @xmath444 .
this will allow us to separate out the behaviour of the exponential sums @xmath440 and @xmath441 on the minor arcs .
before embarking on the case by case analysis alluded to above , it will save needless repetition if we give some reasonably general estimates here that can be applied in various contexts .
our principal means for dealing with small values of @xmath428 relies on taking the inequality @xmath445 in .
this will in turn be estimated as @xmath446 for an appropriate @xmath447 , whence a typical scenario entails studying @xmath448 for @xmath443 such that @xmath444 and certain subsets @xmath449 .
we will always assume that @xmath450 .
let @xmath451 and let @xmath433 .
there exist coprime integers @xmath452 such that @xmath437 .
an argument based on dyadic summation reveals that @xmath453 where @xmath454 is given by , and the maximum is over the possible sign changes and @xmath455 such that @xmath456 furthermore , @xmath455 should satisfy whatever conditions are appropriate to ensure we are dealing with points on @xmath457 . in particular , since @xmath458 the inequalities @xmath459 and @xmath460 can not both hold simultaneously .
let @xmath461 be given .
define @xmath462 and @xmath463 let @xmath464 recall that our task is to show that @xmath465 when @xmath466 .
the following result provides us with easily checked conditions on @xmath461 and @xmath429 under which @xmath467 makes a satisfactory contribution .
[ lem : v<2 ] let @xmath468 .
assume that @xmath469 and @xmath470 define @xmath471 to be the set of @xmath434 for which there exist coprime integers @xmath472 such that @xmath473 then @xmath474 for any @xmath458 , provided that @xmath475 and @xmath476 we will commence under the assumption that @xmath477 , saving the restriction @xmath478 until later in the argument .
it is clear that @xmath479 .
let us consider the consequences of applying lemma [ lem : d ] in our estimate for @xmath480 . throughout the proof of lemma [ lem : v<2 ] we will denote @xmath481 by @xmath482 and we will set @xmath483
. we may deduce from lemma [ lem : d ] and that @xmath484 where @xmath485 and @xmath317 are as in the statement of the lemma and @xmath486\cap { \mathbb{z}}$ ] is arbitrary .
furthermore the maximum is over @xmath455 such that holds with any choice of @xmath433 that we care to choose .
we write @xmath487 , with @xmath488 in particular one easily checks that @xmath489 if holds and @xmath78 is sufficiently small .
it follows from that @xmath490 since @xmath491 .
hence the term involving @xmath492 contributes @xmath493 , which is satisfactory .
let us now turn to the contribution from the term involving @xmath317 in our estimate for @xmath494 .
define @xmath495 then our investigation will be optimised by taking @xmath496 if we can show that @xmath497 with this choice of @xmath123 then we will have @xmath498 since @xmath499 .
we deduce from that @xmath500 in either case the final exponent of @xmath65 is less than 1 , by . hence @xmath123 is an integer in the interval @xmath501 $ ] and it remains to show that @xmath497 with this choice of @xmath123 .
recall the definition of @xmath317 from lemma [ lem : d ] .
suppose first that @xmath502 , with @xmath503 given by .
then @xmath504 and it follows that @xmath505 by .
it follows from our expression for @xmath506 that this is @xmath386 . turning to the third term in the definition of @xmath317
, we see that @xmath507 the exponent of @xmath508 in the second term is @xmath509 , which is negative since @xmath510 and @xmath491 .
hence this quantity is @xmath386 provided that @xmath511 , with @xmath512 with @xmath513 given by and @xmath145 given by . here , as is customary
, we have assumed that @xmath72 is sufficiently small .
one also checks that taking @xmath511 is enough to ensure that the first term is @xmath386 , in view of the lower bound for @xmath514 in .
suppose now that @xmath515 .
then @xmath516 and it follows that @xmath517 since @xmath487 .
it follows from and that this is @xmath386 .
thus the second term makes a satisfactory contribution in @xmath317 . turning to the third term
, we find that @xmath518 we now make the assumption @xmath519 .
hence the overall contribution from the second term is @xmath386 provided that @xmath520 , with @xmath521 with @xmath522 given by and @xmath145 given by .
assuming we note that if @xmath520 and @xmath523 then we would have a point on the major arcs if @xmath72 is sufficiently small in terms of @xmath524 , which we have seen to be impossible .
hence the inequality @xmath520 is also enough to ensure that the first term is @xmath386 .
when @xmath525 the exponent of @xmath508 is zero in the above and we will have an overall contribution of @xmath386 unless @xmath526 we will return to this case shortly .
recall the definitions of @xmath527 .
it follows from the inequality @xmath528 that @xmath529 .
we now employ the assumption on the size of @xmath530 . combining the above
we conclude that there is an overall contribution of @xmath531 to @xmath494 from all of the relevant values of @xmath455 , apart from those which satisfy the inequalities @xmath532 but then the relevant point is forced to lie in the set @xmath471 that was defined in the statement of the lemma .
this is impossible , and so completes the proof of lemma [ lem : v<2 ] .
our next result deals with the corresponding case in which @xmath533 . in this setting becomes @xmath534 define @xmath535 and @xmath536 for any @xmath102 and @xmath26 , and recall the definition of @xmath145 .
then we have the following result .
[ lem : v=2 ] let @xmath533 .
assume that @xmath469 .
define @xmath471 to be the set of @xmath434 for which there exist coprime integers @xmath472 such that @xmath537 then @xmath538 for any @xmath458 , provided that holds and @xmath539 we continue to write @xmath540 and @xmath483 throughout the proof , in order to improve the appearance of our expressions .
our starting point is the proof of lemma [ lem : v<2 ] , which on passing to dyadic intervals via , shows that @xmath541 unless @xmath542 , in the notation of , and the inequality holds for @xmath543 .
this much is valid subject to .
we now consider the effect of applying lemma [ lem : b ] in when @xmath543 and @xmath508 are in the remaining ranges , with @xmath533 .
this gives @xmath544 taking @xmath542 and recalling the assumed inequality for @xmath543 we see that the first term here is @xmath545 where @xmath546 is given by .
according to this contribution is satisfactory . turning to the second term in the above estimate for @xmath547
, we will have @xmath548 as an upper bound for this quantity provided that @xmath549 , with @xmath550 since @xmath491 by assumption .
our investigation has therefore allowed us to handle all @xmath153 apart from those for which @xmath551 and @xmath552 , where @xmath553 is given by .
such points are forced to lie on the set of arcs defined in @xmath471 .
this therefore completes the proof of lemma [ lem : v=2 ] .
the ideal scenario is when we can apply lemma [ lem : v=2 ] with @xmath554 and we will find this is possible for certain ranges of @xmath555 such that @xmath430 . when this comes to pass it follows from , that @xmath556 and @xmath557 , and furthermore , @xmath558 .
one easily checks that the conditions in and are satisfied for @xmath469 .
finally we note that @xmath559 in the statement of lemma [ lem : v=2 ] since clearly any element of @xmath471 is forced to lie on the major arcs .
we may conclude as follows .
[ lem : v=2 ] assume that @xmath469 .
then we have @xmath560 we are now ready to apply this collection of estimates in our case by case analysis of the minor arc integral @xmath561 in .
we will assume that @xmath563 throughout this section .
one of the ingredients in our treatment of this case is the use of `` pruning ''
. we will find it convenient to sort the minor arcs into subsets @xmath564 recall the definition of @xmath145 .
we define @xmath565 to be the set of @xmath434 for which there exists @xmath566 such that @xmath567 and @xmath184 , with @xmath568 we denote by @xmath569 the corresponding set of @xmath570 with the property that whenever holds with @xmath184 and @xmath571 , then @xmath572 we will write @xmath573 for the overall contribution to @xmath561 from integrating over the set @xmath574 , for @xmath575 .
our task is to show that @xmath576 for each @xmath577 . to handle the case @xmath578 we begin as in and with @xmath579 .
it easily follows that @xmath580 on interpreting the integral as a sum over the solutions of the equation @xmath581 , with @xmath582 , and applying standard estimates for the divisor function .
hence we have @xmath583 in the notation of . when @xmath579 , @xmath584 and @xmath585 we have @xmath586 in and .
in particular , and are satisfied .
now it is easily to see that @xmath587 , where @xmath471 is as in the statement of lemma [ lem : v<2 ] , since for @xmath588 we have @xmath589 it therefore follows from lemma [ lem : v<2 ] that @xmath590 , as required .
turning to the case @xmath116 , we begin as above with the observation that @xmath591 this time we appeal to lemma [ lem : b ] . on observing that @xmath592 for any @xmath570
, we deduce from the second part of this result that @xmath593 where the maximum is over all @xmath194 such that @xmath594 taking the lower bounds for @xmath508 and @xmath543 that emerge from these inequalities therefore implies that @xmath595 on recalling that @xmath596 from and @xmath78 is arbitrary .
the key idea in our treatment of @xmath597 is to take advantage of the fact that we have rather good control of the @xmath33-dimensional exponential sum @xmath440 on suitable sets of `` major arcs '' .
recall the definition of @xmath229 .
we will take @xmath598 , whence we may deduce from lemma [ lem : e ] that @xmath599 for any @xmath600 , where @xmath601 is given by .
it follows that @xmath602 say .
we will show that @xmath603 and @xmath604 are both @xmath531 .
let us begin by analysing the first term in this bound .
now it follows from the second part of lemma [ lem : e ] and hlder s inequality that @xmath605 in the notation of .
a straightforward application of lemma [ lem : v<2 ] reveals that @xmath606 , where @xmath607 is the set of @xmath570 for which there exists @xmath566 such that @xmath608 and @xmath184 , with @xmath609 to estimate @xmath610 we employ the second part of lemma [ lem : b ] in much the same way that we did in our analysis of @xmath611 .
this implies that @xmath612 where the maximum is over all @xmath194 such that @xmath613 with the inequalities @xmath459 and @xmath460 not both holding simultaneously .
taking the lower bound for @xmath508 we obtain the contribution @xmath614 this is @xmath531 if @xmath615 . if on the other hand @xmath616 we must automatically have @xmath617 , whence we still obtain a satisfactory contribution .
this completes the treatment of @xmath610 , and so that of @xmath603 .
we now turn to the contribution from @xmath604 in . breaking the ranges for @xmath211 and @xmath618 into dyadic intervals as usual , and applying the second part of lemma [ lem : b ]
, we have @xmath619 where the maximum is over all @xmath455 such that @xmath620 with the inequalities @xmath459 and @xmath460 not both holding simultaneously . if @xmath621 then this is @xmath622 whereas if on the other hand @xmath623 , then the same basic conclusion holds .
once taken all together , this therefore completes the treatment of the minor arcs when @xmath624
. we will continue to assume that @xmath563 throughout this section . in what follows
we may assume that @xmath401 does not take the shape @xmath220 , for integers @xmath221 , since otherwise the resolution of theorem [ main ] is trivial .
recall the manipulations in and .
taking @xmath579 it follows from lemma [ lem : hua_2 ] that the latter inequality is bounded by @xmath626 .
thus we are led to estimate @xmath627 , as given by .
we clearly have @xmath628 in and .
one easily checks that the conditions , and in the statement of lemma [ lem : v<2 ] are satisfied with our choice of @xmath461 and @xmath429 .
recall the definition of @xmath145 . define @xmath471 to be the set of @xmath434 for which there exist coprime integers @xmath472
such that @xmath629 then we may conclude from lemma [ lem : v<2 ] that @xmath630 it remains to deal with the contribution from @xmath588 .
we will use lemma [ lem : b ] to handle this remaining range .
now it follows that @xmath631 recalling the definition of @xmath632 , we therefore deduce from the second part of lemma [ lem : b ] with @xmath633 that @xmath634 where the integral is over @xmath635 and the sum is over @xmath636 , with the inequalities @xmath637 and @xmath638 not both holding simultaneously .
the contribution from @xmath639 is therefore @xmath640 which is satisfactory .
taking @xmath641 , we see that the corresponding contribution from @xmath642 is @xmath643 this too is satisfactory , and so completes our analysis of the case @xmath644 . according to lemma [ lem : curve ]
we may proceed under the assumption that either @xmath401 is non - singular or else our cubic form @xmath15 splits off a ternary norm form . in the latter case theorem [ main ]
readily ensures that @xmath6 , and so we may focus our efforts on the case @xmath401 is non - singular .
we will assume that @xmath646 , throughout this section .
our argument relies upon the same notion of pruning that was put to good effect in [ s:1.12 ] .
let us define @xmath647 to be the set of @xmath434 for which there exists @xmath566 such that @xmath648 and @xmath184 , with @xmath649 it will be convenient to refer to the set @xmath650 as the set of `` proper minor arcs '' , and @xmath647 will be the set of `` improper minor arcs '' .
let us write @xmath651 and @xmath652 for the corresponding contributions to @xmath561 .
we begin by estimating @xmath651 .
taking @xmath533 in and , and applying lemma [ lem : c ] , we deduce that @xmath653 . when @xmath533 , @xmath654 and @xmath655 we have @xmath656 in , and .
now it is easily to see that @xmath657 , where @xmath471 is as in the statement of lemma [ lem : v=2 ] , since for @xmath588 we have @xmath658 where @xmath145 is given by . on observing that @xmath659 satisfies , and
that the inequalities in are trivially satisfied , it therefore follows from lemma [ lem : v=2 ] that @xmath660 .
we now turn to the argument needed to control the overall contribution to @xmath561 from the improper minor arcs , which we denote by @xmath652 .
we select @xmath661 in the definition of @xmath229 .
it now follows from taking @xmath95 and @xmath662 in lemma [ lem : f ] that @xmath663 for any @xmath664 , where @xmath601 is given by . hence @xmath665 say . our goal is to show that @xmath666 .
we begin by handling the contribution from @xmath604 . using dyadic summation it follows from lemma [ lem : b ] that @xmath667 where the maximum is over @xmath455 dictated by the definition of @xmath647 .
this is plainly satisfactory and so completes our treatment of @xmath604 .
we now turn to an upper bound for @xmath603 .
combining hlder s inequality with the second part of lemma [ lem : f ] gives @xmath668 in the notation of .
let us dissect @xmath647 into @xmath669 , where @xmath670 is the set of @xmath671 for which there exist coprime integers @xmath472 such that @xmath672 and @xmath673 , and @xmath674 .
an application of lemma [ lem : b ] yields @xmath675 which is satisfactory . turning to @xmath676
, we note that when @xmath579 and @xmath677 one has @xmath678 in and .
since , and are evidently satisfied in lemma [ lem : v<2 ] , a modest pause for thought reveals that @xmath679 , as required .
we follow the strategy of the preceding section . according to lemma [ lem : surface ]
we may assume that either @xmath401 is non - singular or else the surface @xmath402 contains precisely @xmath104 conjugate double points .
in the latter case implies that our cubic form @xmath15 can be written as @xmath681 for appropriate coefficients @xmath109 and @xmath114 , and where @xmath53 is a certain cubic number field . setting @xmath682 we arrive at a cubic form in @xmath683 variables which is exactly of the type considered in theorem [ main ] .
hence @xmath6 in this case , and so we are free to proceed under the assumption that @xmath401 is non - singular . throughout this section
we will take @xmath174 as our weight function , with @xmath684 and @xmath685 .
let @xmath647 be the set of @xmath434 for which there exists @xmath566 such that @xmath686 and @xmath184 , with @xmath687 as previously we define @xmath650 to be the proper minor arcs and @xmath647 to be the improper minor arcs , with the same notation @xmath688 for the corresponding contributions to @xmath561 .
we begin by estimating @xmath651 .
taking @xmath533 in and , and applying lemma [ lem : c ] , we find that @xmath689
. when @xmath533 , @xmath690 and @xmath691 we have @xmath692 in , and .
furthermore , it is easily checked that @xmath657 , where @xmath471 is as in the statement of lemma [ lem : v=2 ] . on observing that and are satisfied
, it therefore follows from lemma [ lem : v=2 ] that @xmath660 .
it remains to estimate @xmath652 , for which we select @xmath693 in the definition of @xmath694 .
taking @xmath105 and @xmath662 in lemma [ lem : f ] therefore gives @xmath695 for any @xmath664 .
it now follows that @xmath696 say .
we begin by handling the contribution from @xmath604 . using dyadic summation it follows from lemma [ lem : b ] that @xmath697 where the maximum is over @xmath455 such that @xmath698 this is plainly satisfactory and so completes our treatment of @xmath604 .
we now turn to an upper bound for @xmath603 .
since @xmath401 is assumed to be good as well as non - singular , we may apply the second part of lemma [ lem : f ] with @xmath699 and @xmath105 to conclude that @xmath700 as in the case @xmath645 we write @xmath701 , where now @xmath670 is the set of @xmath671 for which there exist coprime integers @xmath472 such that @xmath702 and @xmath703 , and @xmath674 .
it follows from lemma [ lem : b ] that @xmath704 to handle @xmath705 we note that when @xmath706 and @xmath677 we have @xmath707 in and .
lemma [ lem : v<2 ] easily gives @xmath708 , as required .
this completes the treatment of the improper minor arcs when @xmath709 .
an application of lemma [ lem:3fold ] reveals that we are free to assume that @xmath401 defines a projective cubic hypersurface whose singular locus is either empty or finite . throughout this section
we will take @xmath711 as our weight function .
we let the improper minor arcs @xmath647 be the set of @xmath434 for which there exists @xmath566 such that @xmath712 and @xmath184 , with @xmath713 with @xmath145 given by , and we let @xmath650 be the proper minor arcs .
as above , let @xmath688 denote the corresponding contributions to @xmath561
. taking @xmath533 in and , and applying lemma [ lem : c ] with @xmath106 and @xmath267 , we find that @xmath714 . when @xmath533 , @xmath715 and @xmath716 we have @xmath717 in , and . furthermore , it is easily checked that @xmath657 , where @xmath471 is as in the statement of lemma [ lem : v=2 ] . on observing that and are satisfied
, it therefore follows from lemma [ lem : v=2 ] that @xmath660 .
it remains to estimate @xmath652 , for which we select @xmath718 in the definition of @xmath694 .
taking @xmath106 and @xmath719 in lemma [ lem : f ] therefore gives @xmath720 for any @xmath664 .
it now follows that @xmath721 say .
we begin by handling the contribution from @xmath604 . using dyadic summation it follows from lemma [ lem : b ] that @xmath722 where the maximum is over @xmath455 such that @xmath723 this is plainly satisfactory and so completes our treatment of @xmath604 .
we now turn to an upper bound for @xmath603 .
since @xmath401 is assumed to be good , we may apply the second part of lemma [ lem : f ] with @xmath724 and @xmath106 to conclude that @xmath725 , in the notation of .
when @xmath726 , @xmath677 and @xmath716 we have @xmath727 in and .
furthermore one easily checks that @xmath728 satisfies the inequalities in .
it therefore follows from lemma [ lem : v<2 ] that @xmath729 , where @xmath730 is the set of @xmath671 for which there exist coprime integers @xmath731 such that @xmath732 note that @xmath733 . to handle @xmath734
we appeal to lemma [ lem : b ] , deducing that @xmath735 on taking @xmath736 . here
the maximum is over the relevant @xmath455 determined by , with the inequalities @xmath459 and @xmath460 not both holding simultaneously .
now either @xmath737 and this estimate is satisfactory , or else @xmath738 and it follows that we may actually take the lower bound @xmath739 in the second term , giving instead @xmath740 this too is satisfactory , and so completes the proof that @xmath741 , thereby completing the treatment of the minor arcs in the case @xmath742 .
we now come to the final case that we need to analyse in our proof of theorem [ main ] .
we take @xmath744 , and seek an estimate for @xmath745 note that the integral is now taken over the set of minor arcs , rather than the entire interval @xmath223 $ ] as in . as usual
we assume that @xmath401 and @xmath46 are good .
we will show that @xmath746 once in place , we may take @xmath533 in and to conclude that @xmath747 whence the desired conclusion is given by lemma [ lem : v=2 ] . it remains to establish .
let us consider the consequences of applying lemma [ lem : d ] to estimate @xmath748 , following the general approach in the proof of lemmas [ lem : v<2 ] and [ lem : v=2 ] .
we have @xmath749 where @xmath485 and @xmath317 are as in the statement of lemma [ lem : d ] and @xmath486\cap { \mathbb{z}}$ ] is arbitrary . furthermore the maximum is over @xmath455 such that holds with any choice of @xmath433 that we care to choose .
we will take @xmath750 .
in our deduction of it will be convenient to allow the value of @xmath78 to take different values at different parts of the argument .
we define @xmath751 and take @xmath752 if we can show that @xmath753 with this choice of @xmath123 then will follow .
it is clear that @xmath123 is an integer in the interval @xmath501 $ ] .
suppose first that @xmath502 .
then @xmath504 and it follows that @xmath754 by and the fact that @xmath750 .
the third term in @xmath317 is @xmath755 which is satisfactory .
suppose now that @xmath515 .
then @xmath516 and it follows that @xmath756 turning to the third term in @xmath317 , we find that @xmath757 which is also satisfactory .
this therefore completes the proof of , and so our treatment of the case @xmath758 . 99 r.c .
baker , diagonal cubic equations , ii , _ acta arith . _ * 53 * ( 1989 ) , 217250 .
birch , h. davenport and d.j .
lewis , the addition of norm forms .
_ mathematika _ * 9 * ( 1962 ) , 7582 . t.d . browning , counting rational points on cubic hypersurfaces . _ mathematika _ * 54 * ( 2007 ) , 93112 .
browning and d.r .
heath - brown , integral points on cubic hypersurfaces . _
analytic number theory : essays in honour of klaus roth _ , cambridge university press , to appear .
browning and d.r .
heath - brown , rational points on quartic hypersurfaces .
_ j. reine angew .
_ , to appear . , on the classification of cubic surfaces
_ j. london math .
* 19 * ( 1979 ) , 245256 .
a. cayley , a memoir on cubic surfaces .
* 159 * ( 1869 ) , 231326 .
colliot - thlne and p. salberger , arithmetic on some singular cubic hypersurfaces .
london math .
* 58 * ( 1989 ) , 519549 .
colliot - thlne , j .- j .
sansuc and p. swinnerton - dyer , intersections of two quadrics and chtelet surfaces .
j. reine angew .
_ * 374 * ( 1987 ) , 72168 .
d. f. coray , algebraic points on cubic hypersurfaces .
_ acta arith . _ * 30 * ( 1976 ) , 267296 .
d. f. coray , cubic hypersurfaces and a result of hermite . _ duke math
. j. _ * 54 * ( 1987 ) , 657670 .
d. f. coray and m. a. tsfasman , arithmetic on singular del pezzo surfaces .
london math .
soc . _ * 57 * ( 1988 ) , 2587 .
d. f. coray , d.j .
lewis , n. shepherd - barron and p. swinnerton - dyer , cubic threefolds with six double points .
_ number theory in progress , vol .
1 ( zakopane - kocielisko , 1997 ) _ , 6374 , de gruyter , berlin , 1999 .
h. davenport , cubic forms in thirty - two variables .
ser . a _ * 251 * ( 1959 ) , 193232 . h. davenport , cubic forms in sixteen variables . _
london . ser . a _ * 272 * ( 1963 ) , 285303 . h. davenport , _ analytic methods for diophantine equations and diophantine inequalities . _
2nd ed . , cup , 2005 .
j. fowler , a note on cubic equations .
camb . philos .
* 58 * ( 1962 ) , 165169 .
j. franke , y.i .
manin and y. tschinkel , rational points of bounded height on fano varieties . _
_ * 95 * ( 1989 ) , 421435 .
j. harris , _ algebraic geometry _
, springer - verlag , 1992 .
heath - brown , cubic forms in ten variables .
* 47 * ( 1983 ) , 225257 .
heath - brown , rational points on cubic hypersurfaces . _ invent .
* 170 * ( 2007 ) , 199230 . c. hooley , on nonary cubic forms .
_ j. reine angew .
_ * 386 * ( 1988 ) , 3298 . c. hooley , on nonary cubic forms .
j. reine angew .
_ * 415 * ( 1991 ) , 95165 .
j. kollr , unirationality of cubic hypersurfaces .
jussieu _ * 1 * ( 2002 ) , 467476 . d.j .
lewis , diophantine problems : solved and unsolved . _ number theory and applications _ , 103121 , kluwer , dordrecht , 1989 .
manin , _ cubic forms_. 2nd ed .
, north - holland mathematical library * 4 * , north - holland publishing co. , 1986 .
l. schlfli , on the distribution of surfaces of the third order into species .
* 153 * ( 1864 ) , 193247 .
b. segre , on arithmetical properties of singular cubic surfaces .
_ j. london math .
* 19 * ( 1944 ) , 8491 . c. segre , suella varieta cubica con dieci punti doppi dello spazio a quattro dimensioni .
_ atta della r. accademia della scienae di torino _ * 22 * ( 188687 ) , 791801 .
t. skolem , einige bemerkungen ber die auffindung der rationalen punkte auf gewissen algebraischen gebilden . _
* 63 * ( 1955 ) , 295312 .
vaughan , _ the hardy littlewood method _ , 2nd ed . , cup , 1997 .
t. wooley , on weyl s inequality , hua s lemma and exponential sums over binary forms . _ duke math
. j. _ * 100 * ( 1999 ) , 373423 .
groupe de brauer non ramifi des hypersurfaces cubiques + singulires ( daprs p. salberger ) en rponse une question de r. heath - brown , p. salberger en 2006 a indiqu les grandes lignes de la dmonstration de lnonc suivant , qui tend un rsultat connu dans le cas lisse ( @xcite ) .
nous donnons le dtail de la dmonstration .
on utilise les notations usuelles dans ce domaine .
pour @xmath0 un schma on note @xmath759 son groupe de picard et @xmath760 son groupe de brauer .
pour les proprits usuelles de ces groupes , nous renvoyons le lecteur @xcite .
soit @xmath102 un corps de caractristique zro , @xmath761 une clture algbrique , @xmath762 le groupe de galois de @xmath761 sur @xmath102 .
soit @xmath763 une intersection complte gomtriquement intgre de dimension au moins @xmath104 .
supposons le lieu singulier vide ou de codimension au moins gale @xmath14 dans @xmath0 .
alors pour tout @xmath102-modle projectif
et lisse @xmath13 de @xmath0 , les anneaux locaux de @xmath0 en codimension au moins 3 sont rguliers , donc factoriels ( thorme dauslander - buchsbaum , voir ( * ? ? ?
* ; * ? ? ?
* thm . 3.13 ) ) .
comme @xmath0 est une intersection complte , un thorme de grothendieck ( * ? ? ?
* ; * ? ? ?
* cor . 3.14 ) , ex - conjecture de samuel )
implique que tous les anneaux locaux de @xmath0 sont factoriels .
ainsi les diviseurs de weil sur @xmath0 sont tous des diviseurs de cartier .
ceci implique que pour tout ouvert @xmath768 la flche de restriction @xmath769 est surjective .
cette flche est aussi injective .
soit en effet @xmath770 un diviseur sur @xmath0 qui est le diviseur dune fonction rationnelle @xmath771 sur @xmath772 .
comme le complmentaire de @xmath772 dans @xmath0 est de codimension au moins 2 et que sur @xmath0 diviseurs de weil et diviseurs de cartier concident , on conclut que @xmath770 est le diviseur de @xmath771 sur @xmath0 .
le mme argument montre que toute fonction rationnelle sur @xmath0 dfinie et inversible sur @xmath772 est dfinie et inversible sur @xmath0 , et comme @xmath0 est projectif et gomtriquement intgre , toute telle fonction est une constante , elle appartient @xmath773 .
lhypothse sur la codimension du lieu singulier est gomtrique , elle vaut pour @xmath774 pour toute extension @xmath775 de corps , par exemple @xmath776 . les mmes conclusions sappliquent donc @xmath777 . en particulier la flche de restriction @xmath778 est un isomorphisme .
par ailleurs , le corollaire 3.7 de @xcite montre que la flche de restriction @xmath779 qui envoie la classe de @xmath780 sur la classe du faisceau inversible @xmath781 est un isomorphisme .
il en est de mme de @xmath782 , et laction du groupe de galois sur @xmath783 est triviale .
en conclusion , sous les hypothses du thorme , le module galoisien @xmath784 est le module galoisien trivial @xmath765 et lon a @xmath785^*$ ] , o @xmath786 $ ] est lanneau des fonctions dfinies sur @xmath787 .
largument ci - dessus montre aussi que lapplication naturelle @xmath788 est un isomorphisme .
dune suite exacte bien connue ( cf .
@xcite ) on dduit que la flche naturelle @xmath789 $ ] est un isomorphisme .
par des arguments standards sur le groupe de brauer ( puret et injection par passage dune varit lisse un ouvert ) un tel nonc implique le mme nonc pour toute @xmath102-varit projective et lisse @xmath102-birationnelle @xmath0 : cest lnonc ( c ) .
soit @xmath790 une @xmath102-compactification lisse de @xmath772 ( le thorme de hironaka assure lexistence dune telle compactification ) . on a alors la suite exacte de modules
galoisiens @xmath791 o le groupe @xmath792 est le module de permutation sur les points de codimension 1 de @xmath793 en dehors de @xmath787 , et le zro gauche tient au fait que lon a @xmath794^*$ ] .
la suite de modules galoisiens ci - dessus est scinde , car tout groupe @xmath795 valeurs dans un module de permutation est nul ( lemme de shapiro ) .
ainsi @xmath796 est la somme directe de deux modules
de permutation , et est donc un module de permutation .
il en rsulte que pour tout autre modle projectif et lisse @xmath797 , le module galoisien @xmath798 est stablement de permutation ( ( * ? ? ?
2.a.1 on p. 461 ) ) .
lnonc ( b ) en rsulte .
soit @xmath763 une hypersurface cubique gomtriquement intgre de dimension au moins @xmath104 qui nest pas un cne .
supposons le lieu singulier vide ou de codimension au moins gale @xmath14 dans @xmath0 .
alors pour tout modle projectif et lisse @xmath13 de @xmath0 , on a @xmath799 .
si lhypersurface @xmath0 est lisse , on a @xmath801 comme il est tabli dans @xcite sans restriction sur le degr de lhypersurface .
par linvariance birationnelle du groupe de brauer pour les varits projectives et lisses ceci implique @xmath800 .
si lhypersurface cubique @xmath0 est singulire , comme elle nest pas un cne , en utilisant les droites passant par un @xmath761-point singulier on obtient une quivalence birationnelle de @xmath777 avec lespace projectif @xmath802 , do nt le groupe de brauer est nul .
par linvariance birationnelle du groupe de brauer pour les varits projectives et lisses ceci implique @xmath800 .
la condition que la codimension du lieu singulier est au moins gale 4 est ncessaire .
dans @xcite on trouve des hypersurfaces cubiques gomtriquement intgres non coniques dans @xmath803 , do nt le lieu singulier est un ensemble fini non vide de points , et qui admettent un modle projectif et lisse @xmath13 avec @xmath804 .
lorsque @xmath102 est un corps de nombres , le corollaire permet de conjecturer que , sous les hypothses donnes , le principe de hasse et lapproximation faible valent pour le lieu lisse de lhypersurface cubique @xmath0 .
a. grothendieck , cohomologie locale des faisceaux cohrents et thormes de lefschetz locaux et globaux ( sga 2 ) .
_ sminaire de gomtrie algbrique du bois marie , 1962 .
augment dun expos de michle raynaud _ , documents mathmatiques ( paris ) * 4 * , socit mathmatique de france , 2005 . | let @xmath0 be a projective cubic hypersurface of dimension @xmath1 or more , which is defined over @xmath2 . we show that @xmath3 is non - empty
provided that the cubic form defining @xmath0 can be written as the sum of two forms that share no common variables . |
the 2-dimensional comb @xmath3 is maybe the simplest example of inhomogeneous graph .
it is obtained from @xmath4 by removing all horizontal edges off the @xmath5-axis ( see figure [ fig:2comb ] ) .
many features of the simple random walk on this graph has been matter of former investigations .
local limit theorems were first obtained by weiss and havlin @xcite and then extended to higher dimensions by gerl @xcite and cassi and regina @xcite .
more recently , krishnapur and peres @xcite have shown that on @xmath3 two independent walkers meet only finitely many times almost surely .
this result , together with the space - time asymptotic estimates obtained in @xcite for the @xmath0-step transition probabilities , suggests that the walker spends most of the time on some tooth of the comb , that is moving along the vertical direction . indeed in (
* section 10 ) , it has been remarked that , if @xmath6 goes to zero with a certain speed , then @xmath7 .
moreover the results in @xcite imply that there are no sub - gaussian estimate of the transition probabilities on @xmath3 .
such estimates have been found on many graphs : by jones @xcite on the @xmath8-dimensional sierpiski graph , by barlow and bass @xcite on the graphical sierpiski carpet and on rather general graphs by grigoryan and telcs ( @xcite ) .
these estimates involve three exponents which are usually associated to infinite graphs : the _ spectral dimension _ @xmath9
( which is by definition twice the exponent of @xmath10 in local limit theorems ) , the _ fractal dimension _ @xmath11
( which is the growth exponent ) and the _ walk dimension _ @xmath12 . these dimensions are in typical cases linked by the so - called _ einstein relation _ : @xmath13 ( see telcs @xcite ) .
the first two dimensions are known for @xmath3 : @xmath14 and @xmath15 . in this paper
we compute @xmath16 , thus showing that the relation does not hold for this graph . in order to point out the different behaviour of the random walk along the two directions we analyze the asymptotic behaviour of the expected value of the distance from the origin reached by the walker in @xmath0 steps .
different concepts of distance are considered : position after @xmath0 steps , maximal deviation and maximal span .
section [ sec : prelim ] is devoted to these definitions , and to the necessary preliminaries , such as the definitions of random walk on a graph and of generating function .
the expression of the generating function of the transition probabilities of the simple random walk on @xmath3 is recalled . in sections [ sec : distances ] , [ sec : maxdev ] and [ sec : maxspan ] we prove the asymptotic estimates of the expected value of the distance of the walk from the origin after @xmath0 steps , of its maximal deviation from the origin and of its maximal span respectively .
the proofs are based on a darboux type transfer theorem : we refer to ( * ? ? ?
* corollary 2 ) , but one may also refer to @xcite and to the hardy - littlewood - karamata theorem ( see for instance @xcite ) . this theorem ( as far as we are concerned ) claims that if @xmath17 and @xmath18 is analytic in some domain , with the exception of @xmath19 , then @xmath20 the aim of our computation is then to determine an explicit expression of the generating functions of the sequence of expected values of the random variables we are interested in .
this is done employing the combinatorial methods used by panny and prodinger in @xcite .
we also refer to that paper for a comparison between our results and the analogous results for the simple random walk on @xmath21 .
theorems [ th : distx ] , [ th : disty ] , [ th : mdx ] , [ th : mdy ] , [ th : msx ] and [ th : msy ] prove that the expected values of the distances along the horizontal direction are all of order @xmath1 , and along the vertical direction they are of order @xmath2 ( the exact constants are determined ) . since the notions of maximal deviation in @xmath0 steps and of first exit time from a @xmath22-ball are closely related , we determine the walk dimension of @xmath3 in section [ sec : maxdev ]
. in section [ sec : scaling ] we deal with the limit of the process obtained dividing by @xmath1 and by @xmath2 respectively the continuous time interpolation of the horizontal and vertical projections of the position after @xmath0 steps . as one would expect the limit of the vertical component
is the brownian motion , while the limit of the horizontal component is less obvious ( it is a brownian motion indexed by the local time at 0 of the vertical component ) .
this scaling limit is determined in theorem [ th : scaling ] .
finally , section [ sec : fr ] is devoted to a discussion of the results , remarks and open questions .
the simple random walk on a graph is a sequence of random variables @xmath23 , where @xmath24 represent the position of the walker at time @xmath0 , such that if @xmath5 and @xmath25 are vertices which are neighbours then @xmath26 where @xmath27 is the number of neighbours of @xmath5 , otherwise @xmath28 . in particular on @xmath3
the non - zero transition probabilities @xmath29 are equal to 1/4 if @xmath5 is on the horizontal axis , and they are equal to 1/2 otherwise . given @xmath30 , let @xmath31 be the @xmath0-step transition probability from @xmath5 to @xmath25 .
recall that the generating function of a sequence @xmath32 is the power series @xmath33 ; by definition the green function associated to the random walk on a graph @xmath34 is the family of generating functions of the sequences @xmath35 , @xmath36 , that is @xmath37 the green function , with @xmath38 , on @xmath3 can be written explicitly as ( see @xcite ) @xmath39 where @xmath40 we refer to ( * ? ? ?
* section 1.1 ) for more details on the random walks on graphs , transition probabilities and generating functions .
when we consider the walk up to time @xmath0 , different concepts of `` distance '' arise .
we consider two equivalent norms on @xmath3 : for a given vertex @xmath41 define @xmath42 note that @xmath43 is the usual distance on the graph . in the following section we deal not only with the asymptotic behaviour of @xmath44 $ ] and @xmath45 $ ] ( we use the notation @xmath46 ) , but also with the asymptotics of the expected value of other random variables , which represent the ( horizontal and vertical ) maximal deviation and the span of the walk . 1 .
the maximal deviations in @xmath0 steps are defined as @xmath47 2 .
the maximal spans in @xmath0 steps are defined as @xmath48
[ th : distx ] @xmath49\stackrel{n\to\infty}\sim\frac 1{2^{3/4}\gamma(5/4)}\,n^{1/4}.\ ] ] since for @xmath50 , @xmath51 it is clear that ( exchanging the order of summation ) @xmath52 z^n= 2\sum_{k=1}^\infty k\sum_{l\in{\mathbb z } } g\big((0,0),(k , l)|z\big).\ ] ] by elementary computation one obtains @xmath53
z^n & = g(z)\,\frac{f_1(z)}{(1-f_1(z))^2}\,\frac 1{1-f_2(z)}\\ & \stackrel{z{\rightarrow}1 ^ -}\sim \frac 1{2^{3/4}(1-z)^{5/4}}. \end{split}\ ] ] thus , applying ( * ? ? ?
* corollary 2 ) we obtain the thesis .
[ th : disty ] @xmath54\stackrel{n\to\infty}\sim \sqrt{\frac 2\pi}\ , n^{1/2}.\ ] ] as in the proof of the previous theorem , for @xmath55 , we write @xmath56 and @xmath57 z^n = 2\sum_{l=1}^\infty l\sum_{k\in{\mathbb z } } g\big((0,0),(k , l)|z\big),\ ] ] thus @xmath58 z^n & = g(z)\,\frac{1+f_1(z)}{1-f_1(z ) } \ , \frac{f_2(z)}{(1-f_2(z))^2}\\ & \stackrel{z{\rightarrow}1 ^ -}\sim\frac 1{2^{1/2}(1-z)^{3/2}}. \end{split}\ ] ] apply ( * ? ? ?
* corollary 2 ) , recalling that @xmath59 , to conclude .
@xmath60 $ ] and @xmath61 $ ] are both asymptotic , as @xmath0 goes to infinity , to @xmath62 .
in order to compute the generating function of @xmath63\}_{n\ge0}$ ] , we first need an expression for another generating function . [
th : pdnx ] let @xmath64 , @xmath65 .
the generating function of the sequence @xmath66 is @xmath67 where @xmath68 is the generating function of the number of paths on @xmath21 of length @xmath0 , from @xmath69 to @xmath70 with maximal deviation less or equal to @xmath71 .
note that the paths we are interested in have no bound on the vertical excursions .
thus we may decompose the walk into its horizontal and vertical components , and consider each horizontal step as a vertical excursion ( whose length might be zero ) coming back to the origin plus a step along the horizontal direction . keeping this decomposition in mind ,
it is clear that the generating function of the sequence @xmath72 is @xmath73 where @xmath74 is the generating function of the probabilities of the @xmath0-step excursions along one single tooth of the comb ( that is paths in figure [ fig : vertcomp ] which do not use the loop at zero ) , from @xmath75 to @xmath76 .
moreover we must admit a final excursion from @xmath77 to @xmath78 for some @xmath79 , that is we must multiply by @xmath80 and the generating function of @xmath66 is @xmath81 using ( * ? ? ?
* lemma 1.13 ) and reversibility ( see ( * ? ? ?
* section 1.2.a ) ) , it is not difficult to compute @xmath82 the thesis is obtained noting that @xmath83 [ th : genfundxv ] @xmath84z^n= 2\,\frac{1 + 6v^2+v^4}{(1-v)^4}\sum_{h\ge1}\frac{v^h}{1+v^{2h}},\ ] ] where @xmath85 is such that @xmath86 let @xmath68 be as in lemma [ th : pdnx ] , and put @xmath87
. then the generating function of @xmath88 is @xmath89 thus we may write the generating function of @xmath90 $ ] as @xmath91 an explicit expression for @xmath92 has been determined by panny and prodinger in ( * ? ? ?
* theorem 2.2 ) : @xmath93 where @xmath94 . by the definition of @xmath95
it is clear that the relation between @xmath96 and @xmath85 is set by equation .
then it is just a matter of computation to obtain @xmath97 whence , substituting in , the thesis .
[ th : vasint ] @xmath98 the proof is quite standard , we report it here for completeness . put @xmath99 , @xmath100 and consider @xmath101 the mellin transform of @xmath102 is : @xmath103 where @xmath104 is the riemann zeta function .
the knowledge of the behaviour of @xmath105 in a neighbourhood of 1 , will give us the behaviour of @xmath106 in a neighbourhood of 0 .
substitute @xmath107 to obtain @xmath108 where @xmath109 , and @xmath110 is the gamma function .
since as @xmath111 @xmath112 we are left with the computation of the asymptotic behaviour of @xmath113 .
we may write @xmath114 where @xmath115 is the hurwitz zeta function . thus using the expansion of @xmath116 for @xmath117 close to @xmath118 ( see ( * ? ? ? * formula 13.21 ) ) we obtain @xmath119 hence we get @xmath120 applying ( * ? ? ?
* theorem 1 , p.115 ) , @xmath121 which , substituting @xmath122 , gives the thesis .
[ th : mdx ] @xmath123\stackrel{n\to\infty}\sim \frac{2^{-7/4}\pi}{\gamma(5/4)}\,n^{1/4}.\ ] ] by proposition [ th : genfundxv ] and lemma [ th : vasint ] , it is clear that @xmath124z^n\stackrel{v{\rightarrow}1 ^ -}{\sim } \frac{4\pi}{(1-v)^5}.\ ] ] choosing the solution @xmath85 of equation which is smaller than 1 when @xmath96 is smaller than 1 , and substituting it in we obtain @xmath84z^n \stackrel{z{\rightarrow}1 ^ -}{\sim } \frac { 2^{-7/4}\pi}{(1-z)^{5/4}},\ ] ] and ( by ( * ? ? ?
* corollary 2 ) ) the thesis .
we recall a result which can be found in ( * ? ? ? * theorem 2.1 ) , which is useful in the sequel .
[ th : ai ] let @xmath125 be the @xmath126 matrix @xmath127\ ] ] and let @xmath128 be its determinant .
then @xmath129 where @xmath130 for the sake of simplicity , in the sequel we write @xmath131 and @xmath132 instead of @xmath125 and @xmath128 respectively .
let @xmath64 , @xmath65 .
the generating function of the sequence @xmath133 is @xmath134 consider the absolute value of the vertical projection of the random walk as the random walk on the non negative integers with one - step transition probabilites described in figure [ fig : absvert ] . note that @xmath135 , @xmath136 , is determined by the linear system : @xmath137 \end{array } \end{array } \right ] \left[\begin{array}{c } \widehat\psi_{h,0}(z ) \\
\vdots \\\vdots \\ \widehat\psi_{h , h}(z ) \\
\end{array}\right ] = \left[\begin{array}{c } 1 \\ 0 \\ \vdots \\ 0 \\ \end{array}\right].\ ] ] using cramer s rule ( compare with ( * ? ? ?
* theorem 2.1 ) ) , we obtain the thesis .
the generating function of @xmath138\}_{n\ge0}$ ] is ( written as a function of @xmath85 ) @xmath139 from and we get @xmath140 whence , summing @xmath70 from @xmath69 to @xmath71 , the proof is complete .
proposition [ th : genfundyv ] , lemma [ th : vasint2 ] and theorem [ th : mdy ] are the analogs of proposition [ th : genfundxv ] , lemma [ th : vasint ] and theorem [ th : mdx ] respectively .
therefore , their proofs are omitted .
[ th : genfundyv ] let @xmath85 be such that holds .
then @xmath141z^n= \frac{(1+v^2)(1+v)}{(1-v)^2}\sum_{h\ge1 } \frac{v^{h}}{(1+v^{2h+1})}.\ ] ] [ th : vasint2 ] @xmath142 [ th : mdy ] @xmath143\stackrel{n\to\infty}\sim \sqrt{\frac\pi2}\ , n^{1/2}.\ ] ] both @xmath144 $ ] and @xmath145 $ ] , as @xmath0 goes to infinity , are asymptotic to @xmath146 .
apply the results of theorem [ th : mdx ] and theorem [ th : mdy ] to the following inequalities @xmath147 the maximal deviation of the walk in @xmath0 steps is linked to the first exit time from a ball of radius @xmath22 . indeed
if we put @xmath148 , where @xmath149 is the ball of radius @xmath22 centered in @xmath150 , then @xmath151 clearly the radius of the ball and @xmath152 must be computed with respect to the same norm on the graph .
we write @xmath153 and @xmath154 for the exit times with respect to the two norms we defined in section [ sec : prelim ] .
recall now that given the simple random walk on a graph , if @xmath155 $ ] is of order @xmath156 , then by definition @xmath157 is the walk dimension of the graph . usually , the norm with the respect to which the radius is computed is @xmath158 , but as we will show , we may equivalently consider @xmath159 .
therefore we are now interested in the asymptotic behaviour of @xmath160 $ ] .
[ th : tninfty ] @xmath160\stackrel{n\to\infty}\sim n^2 $ ] .
we write @xmath161=\sum_{k\ge 0 } { \mathbb p}(t_n^\infty > k),\ ] ] that is , @xmath160 $ ] is equal to @xmath162 , where @xmath163 is the generating function of the sequence @xmath164 .
let us observe that @xmath165 .
we claim that @xmath166 where @xmath85 is such that holds and @xmath167 .
then since for @xmath168 ( choosing the solution @xmath169 which is bounded in some neighbourhood of @xmath170 ) @xmath171 we get that @xmath172 we are left with the proof of equation .
we proceed as in lemma [ th : pdnx ] , that is we separately consider the two components of the walk .
let @xmath173 be the generating function of the probabilites of the @xmath174-step excursions along one single tooth , ending at a height @xmath70 or @xmath175 , with maximal deviation bounded by @xmath0 ( see figure [ fig : absvert ] ) .
put @xmath176 .
then @xmath177 where @xmath178 is the generating function we already used in proposition [ th : genfundxv ] .
by ( * ? ? ?
* lemma 1.13 ) @xmath179 where @xmath180 are the generating functions of the @xmath174-step `` purely vertical '' excursions , from @xmath181 to @xmath75 , with maximal deviation bounded by @xmath0 .
the @xmath182 s are the solutions of the linear system : @xmath183 = \left[\begin{array}{c } z/2 \\ 0 \\ \vdots \\ 0 \\ \end{array}\right ] .
\label{eq : ls}\ ] ] then , with the substitution ( compare with @xcite ) , we get @xmath184 thus , @xmath185 . for the final excursion we must compute @xmath173 , for @xmath186 .
this can be done writing and solving @xmath0 linear systems in the spirit of to obtain @xmath187 @xmath188\stackrel{n\to\infty}\sim n^2.\ ] ] define @xmath189 as the first exit time from the rectangle @xmath190 , and note that @xmath191 .
a calculation similar to that of proposition [ th : tninfty ] , shows that @xmath192\stackrel{n\to\infty } \sim ( 1-\alpha)^2n^2 $ ] . since @xmath157 is an arbitrary positive constant , we are done .
the walk dimension of @xmath3 is equal to @xmath8 .
[ th : msx ] @xmath193\stackrel{n\to\infty}\sim \frac{2^{1/4}}{\gamma(5/4)}n^{1/4}$ ] .
let @xmath194 .
then it is clear that @xmath195=2{\mathbb e}[m^x_n]$ ] .
our first aim is to compute the generating function of @xmath196 , which we denote by @xmath197 .
then @xmath198 where @xmath199 is the generating function of the probabilities of the @xmath0-step paths such that @xmath200 for @xmath201 , @xmath202 and @xmath203 , while @xmath204 was defined and computed in lemma [ th : pdnx ] .
let us note that @xmath205 , where @xmath206 is the generating function of the number of the @xmath0-step paths on @xmath21 which stay in the interval between @xmath207 and @xmath71 , and end at @xmath70 .
the functions @xmath208 are determined by the linear system used in ( * ? ? ?
* theorem 4.1 ) to determine @xmath209 .
thus @xmath210 then we put @xmath211 and we obtain that ( compare with the proof of proposition [ th : genfundxv ] ) @xmath212 then @xmath213z^n&= \frac{1 + 6v^2+v^4}{(1-v)^4}\sum_{h\ge0}v^{h+1 } \stackrel{z{\rightarrow}1 ^ -}\sim\frac{1}{2^{3/4}(1-z)^{5/4 } } , \end{split}\ ] ] whence @xmath214\stackrel{n\to\infty}\sim 2^{-3/4}n^{1/4}/\gamma(5/4)$ ] and we are done .
[ th : msy ] @xmath215\stackrel{n\to\infty}\sim \sqrt{\frac8\pi}\,n^{1/2}. $ ] let @xmath216 , then @xmath215=2{\mathbb e}[m^y_n]$ ] . denote by @xmath217 the generating function of @xmath218
. then @xmath219 where @xmath220 is the generating function of the probabilities of the @xmath0-step paths ( see figure [ fig : vertcomp ] ) , such that @xmath221 for @xmath201 , @xmath222 .
the functions @xmath220 are determined by the linear system ( where only the non - zero terms are displayed ) @xmath223 & \begin{array}{c } \ \\ \ \\ -z/4 \\
\end{array } & \\ -z/2 & ( 1-z/2 ) & -z/2 \\ & \begin{array}{c } -z/4 \\ \ \\ \ \\
\end{array } & \left [ \begin{array}{lcr } \ & \ & \ \\ \ & \mathbf a_{k-1 } & \ \\ \ & \ & \ \\ \end{array}\right ] \end{array}\right ] \left[\begin{array}{c } \widehat \psi_{h , k;h}(z ) \\ \vdots \\ \widehat \psi_{h , k;1}(z ) \\ \widehat \psi_{h , k;0}(z ) \\ \widehat \psi_{h , k;-1}(z ) \\ \vdots \\ \widehat \psi_{h , k ;-
k}(z ) \end{array}\right ] = \left[\begin{array}{c } 0\\ \vdots \\ 0 \\ 1 \\ 0 \\ \vdots \\ 0 \\ \end{array}\right].\ ] ] denote by @xmath224 the determinant of the previous matrix . for @xmath225 elementary computation lead to @xmath226
the interest in @xmath227 originates in that @xmath228z^n=\sum_{h\ge0}\left ( \frac1{1-z}-\widehat\psi_h(z)\right).\ ] ] by lemma [ th : ai ] it is not difficult to prove that for @xmath225 ( and @xmath85 such that holds ) @xmath229 noting that @xmath230 $ ] we get that as @xmath231 ( and @xmath232 ) @xmath233z^n\sim\frac 2{(1-v)^3}\sim \frac 1{\sqrt 2(1-z)^{3/2}}.\ ] ] of course we have to prove that the terms in corresponding to @xmath234 are negligible , but this follows from elementary computation . by (
* corollary 2 ) we deduce that @xmath235\stackrel{n\to\infty}\sim\sqrt{\frac 2\pi}\,n^{1/2}$ ] and we are done .
in the preceding sections we have seen that the expected values of the distances ( with various meanings of this word ) reached in @xmath0 steps are of order @xmath1 for the horizontal direction and of order @xmath2 for the vertical direction . these results lead us to a natural question : what is the asymptotic behaviour of the process where the horizontal component of the position after @xmath0 steps is divided by @xmath1 and the vertical component is divided by @xmath2 ?
of course we have to make this question more precise . in order to study the scaling of the process we choose a suitable realization for the sequence @xmath23 : let @xmath236 be a sequence of random variables representing a simple random walk on @xmath21 , and @xmath237 be a sequence representing the random walk on @xmath21 moving according to figure 2 .
choose @xmath34 and @xmath238 to be independent and let @xmath239 a.s .. moreover , let @xmath240 be the number of loops performed by @xmath238 up to time @xmath22 , that is @xmath241 clearly , @xmath242 is a realization of the position of the simple random walker on @xmath3 at time @xmath0 .
we are now able to define , by linear interpolation of the three discrete time processes @xmath243 and @xmath244 , a continuous time process @xmath245 .
[ th : scaling ] @xmath246n},\frac{y_{nt}}{\sqrt n}\right)_{t\ge0 } \stackrel{law}\longrightarrow \left(w_{l^0_t(b)},b_t\right)_{t\ge0},\ ] ] where @xmath247 and @xmath248 are two independent brownian motions and @xmath249 is the local time at 0 of @xmath248.[th : scallim ] the theorem will be a consequence of proposition [ th : yl ] and proposition [ th : xl ] .
we introduce the following notion of convergence of stochastic processes ( see definition 2.2 of @xcite ) .
a sequence of @xmath250-valued stochastic processes @xmath251 converges to a process @xmath252 in probability uniformly on compact intervals if for all @xmath253 , as @xmath254 @xmath255 where @xmath256 is a norm on @xmath250 ( for instance @xmath158 )
. we will briefly write @xmath257 since u.p
. convergence of a vector is equivalent to u.p .
convergence of its components and implies convergence in distribution , in order to prove theorem [ th : scallim ] it will suffice to prove that each component in u.p . converges to the corresponding limit .
the main idea is that @xmath238 and @xmath244 are not much different from , respectively , a simple random walk @xmath258 on @xmath21 and the process @xmath259 which counts the visits of @xmath258 to 0 .
there is a natural correspondence between @xmath238 and @xmath258 , so let us define @xmath258 and some other auxiliary variables which will be needed in the sequel
. given @xmath238 , let @xmath70 and @xmath260 be the processes which respectively count its returns to 0 ( not including the loops and counting time 0 as the first `` return '' ) and the time spent not looping at 0 .
namely , let @xmath261 , and @xmath262 and @xmath263 for @xmath264 .
clearly @xmath265 .
moreover , note that for @xmath264 , @xmath266 where @xmath267 is a suitable sequence of iid random variables with geometric distribution of parameter 1/2 . now define a simple random walk @xmath258 on @xmath21 by @xmath268 . then @xmath269 ( @xmath259 counts the visits at 0 or , equivalently , the returns to 0 ) .
we note that @xmath270 and @xmath271 .
we first prove a property of @xmath244 .
[ th : limln ] @xmath272 the main ideas of the proof are the facts that @xmath273 and that @xmath274 is not much different from @xmath0 .
indeed one can easily prove the first fact ( for the distribution of @xmath275 see ( * ? ? ?
* chapter iii , exercise 10 ) ) . by
, the thesis is a consequence of @xmath276 by the strong law of large numbers and slutzky s theorem , @xmath277 then will follow once we show that @xmath278 . indeed note that @xmath279 where @xmath280 is the number of visits to 0 of @xmath258 between time @xmath274 and time @xmath0 .
let @xmath281 be the time @xmath258 first visits 0 after time @xmath274 , and for any @xmath282 let @xmath283 be the number of visits to 0 before time @xmath284 of the random walk @xmath285 . clearly @xmath283 is independent of @xmath286 and has the same distribution of @xmath287 .
then @xmath288 where @xmath289 and @xmath290 is equal to zero if @xmath291 .
we are left with the proof that @xmath292 . since @xmath293
, it suffices to prove that @xmath294 .
fix @xmath295 : since @xmath296 we have that @xmath297 now fix @xmath298 and choose @xmath299 such that @xmath300 for all @xmath301 ( this is possible by the law of large numbers using the facts that @xmath302 and @xmath303 are independent and @xmath304 ) . then @xmath305 since @xmath306
we are done .
[ th : yl ] @xmath307 consider the processes @xmath258 and @xmath259 defined before lemma [ th : limln ] and by interpolation define the sequence of two - dimensional continuous time processes @xmath308 .
then by theorem 3.1 of @xcite we have that @xmath309 to prove our statement , it suffices to show that these two properties hold : @xmath310 note that @xmath258 is the sum of iid increments @xmath311 such that @xmath312 , hence @xmath313 where @xmath314 ( and @xmath315 if @xmath316 ) .
thus @xmath317 .
then @xmath318 follows from @xmath319 indeed @xmath320 and if we denote by @xmath321 and by @xmath322 , clearly @xmath323 and @xmath324 are identically distributed , and @xmath325 hence to prove it suffices to show that @xmath326 the distribution of @xmath323 is well known ( see ( * ? ? ? * chapter iii.7 ) ) , and it is easy to show that @xmath327 .
noting that @xmath328 is independent of @xmath329 , we have that @xmath330 by lemma [ th : limln ] , for any positive @xmath331 and @xmath332 there exist @xmath157 and @xmath333 such that @xmath334 for all @xmath335 . on the other hand for any given @xmath336 @xmath337{{\lfloor nt\rfloor } } } > \frac{{\varepsilon}\sqrt[4]n}{\sqrt\alpha\sqrt[4 ] t}\right ) \stackrel{n\to\infty}{\longrightarrow}0,\ ] ] whence @xmath318 is proven .
now , let us address to @xmath338 .
we first note that we may consider a mapping between the number of steps taken by @xmath258 and the ones taken by @xmath238 .
indeed when @xmath258 has taken @xmath339 steps , then @xmath238 has taken @xmath340 steps ( that is , if @xmath341 we decide to count for @xmath238 all the loops it performs after this last return to 0 ) .
let us write @xmath342 we prove that @xmath343 .
indeed @xmath344 , thus @xmath345 now choose @xmath298 .
by independence of @xmath259 and @xmath346 we have that the probability that @xmath347 is larger than @xmath348 is bounded by @xmath349the first term is smaller than @xmath350 if @xmath157 and @xmath0 are sufficiently large . as for the second term , it is clearly less or equal to @xmath351 observe that , by the law of large numbers , for any positive @xmath331 and @xmath350 there exists @xmath352 such that for all @xmath353 @xmath354 hence is less or equal to @xmath355 the first term clearly tends to 0 as @xmath0 grows to infinity , while the second term is not larger than @xmath356 but if @xmath357 for all @xmath353 , then @xmath358 which is smaller than @xmath348 if @xmath331 is sufficiently small .
this proves that @xmath343 .
we now prove that @xmath359 . indeed @xmath360 using and the definitions of @xmath302 and @xmath303 thereafter , we have that @xmath361 , and for any positive @xmath331 , @xmath362 choose @xmath298 .
let us show that @xmath363 if @xmath331 is sufficiently small and @xmath0 sufficiently large .
indeed @xmath364 and by independence of @xmath365 and @xmath366 , @xmath367 choose @xmath157 such that @xmath368 for @xmath0 sufficiently large , and then @xmath331 sufficiently small such that @xmath369 .
now keep @xmath331 fixed . in order to prove that @xmath370 for @xmath0 sufficiently large
, we observe that for any positive @xmath371 there exists @xmath372 such that for all @xmath373 , @xmath374 is larger than @xmath375 . then for all @xmath0 such that @xmath376 , @xmath377 with probability larger than @xmath375 .
hence , as in we have that , for @xmath157 and @xmath0 sufficiently large and @xmath371 sufficiently small
. @xmath378
[ th : xl ] @xmath379n}\right)_{t\ge0 } \stackrel{u.p.}\longrightarrow \left(w_{l^0_t(b)}\right)_{t\ge0}.\ ] ] clearly the statement may be rephrased as @xmath380 note that ( writing @xmath381 instead of @xmath382 ) @xmath383 we have to show that @xmath318 and @xmath338 @xmath384 : it suffices to prove the statement for @xmath385 and @xmath386 instead of @xmath387 and @xmath388 respectively .
represent @xmath389 , where @xmath390 is an iid family such that @xmath391 , then @xmath392 where @xmath393 and @xmath394 .
note that @xmath395 , indeed @xmath396 and both these summands tends u.p . to 0
( the first by proposition [ th : yl ] ) . in order to show that @xmath397 it suffices to prove that @xmath398 now put @xmath399 and recall that @xmath400 and @xmath401 are independent :
this last probability may be written as @xmath402 for all positive @xmath157 , @xmath403 can be made arbitrarily small if @xmath0 is picked large enough , while @xmath404 thus if @xmath298 is fixed , by choosing @xmath157 sufficiently small we get @xmath405 if @xmath0 is sufficiently large .
this proves @xmath406 .
now note that @xmath407 and @xmath381 are independent , so we may think of @xmath408 as defined on a product probability space @xmath409 ( that is @xmath410 ) . for any fixed @xmath411
we have that @xmath412 indeed it is known ( see @xcite ) that @xmath413 , and since for any fixed @xmath411 , as @xmath414 , @xmath415 , @xmath416 whence .
thus , putting @xmath417 we have that @xmath418 , by fubini and the dominated convergence theorem we get @xmath419 and we are done .
the results of the previous sections show that the random walker on @xmath3 spends most of the time walking along the vertical direction .
one would ask to what extent the resemblance between the simple random walk on @xmath21 and the vertical component of the simple random walk on @xmath3 is apparent .
the answer is that , to leading order , the expected values of the distances after @xmath0 steps of these two walks are indistinguishable . indeed
if we denote by @xmath407 the position at time @xmath0 of the walker on @xmath21 , one could easily compute @xmath420\stackrel{n\to\infty}\sim \sqrt{\frac\pi2}\,n^{1/2},\ ] ] which coincides with the estimate of theorem [ th : disty ] .
moreover , a comparison between theorem [ th : mdy ] and ( * ? ? ?
* theorem 2.14 ) and between theorem [ th : msy ] and ( * ? ? ?
* theorem 3.4 ) shows the same coincidence for the estimates of the maximal deviation and the maximal span respectively .
the inhomogeneity of @xmath3 results in the difference between the behaviour of the horizontal and vertical components of the walk . indeed while for any positive integer @xmath421 the expected value of the distance from the origin of the simple random walker on @xmath422 , after @xmath0 steps , is of order @xmath2 ( this may be easily proven via a conditioning argument , with respect to the proportions of time spent along each of the @xmath421 main directions ) , the expected value of the horizontal distance on @xmath3 is of order @xmath1 , that is on this direction we have a subdiffusive behaviour .
of course this is due to the delay observed on the @xmath5-axis while the walker explores the teeth of the comb ( recall that the simple random walk on @xmath21 , although recurrent , is zero - recurrent , that is the expected value of the first return time to the origin is infinite ) .
the difference between the two projections of the walk is remarkable also when one properly rescales the process as we did in section [ sec : scaling ] .
indeed the space where the `` rescaled walker '' lives is @xmath423 , endowed with the topology for which any path between two points @xmath424 and @xmath425 must necessarily include the three edges @xmath424@xmath426 , @xmath426@xmath427 and @xmath427@xmath425 .
this reflects on the limiting process : the horizontal component may change only when the vertical one passes through 0 .
in fact , as we proved in theorem [ th : scaling ] , @xmath428 converges to a standard brownian motion @xmath248 , ( the walk on the vertical direction is , to leading order , unaffected by the bias of the random holding time at zero ) , while @xmath429 converges to a brownian motion whose clock is the local time at zero of @xmath248 .
the fact that the walker on @xmath3 essentially behaves like the walker on @xmath21 makes it clear that @xmath430 must be 2 , as we proved by combinatorial methods in section [ sec : maxdev ] ( note that @xmath431 for all integer @xmath432 ) . indeed it is very likely that the walker will exit the ball of radius @xmath0 moving along some tooth of @xmath3 .
this disproves the einstein relation between @xmath433 , @xmath434 and @xmath435 .
the failure of this relation in this case is due to the inhomogeneity of this particular graph . indeed for strongly inhomogeneous graphs
the growth exponent does not give an accurate description of the `` fractal properties '' of the graph ( the assignment of the same @xmath11 of @xmath4 to @xmath3 disregards the topology of the two structures ) .
it seems that defining @xmath11 as the growth exponent of the graph makes sense only for homogeneous graphs ( like @xmath422 ) or self - similar graphs , which have a clear fractal nature ( like the sierpiski graph ) .
it is our opinion that the study of the sense in which a graph has a fractal nature and what is the proper definition of its fractal dimension should require further investigations .
i feel particularly indebted to peter grabner and helmut prodinger for raising the questions discussed in this paper and for the stimulating discussions during which they suggested the techniques used in sections [ sec : distances ] , [ sec : maxdev ] and [ sec : maxspan ] .
i would also like to thank jean - francois le gall for suggesting the limit of the rescaled process and giving me some precious hints .
barlow and r.f .
bass , ( 1999 ) random walks on graphical sierpiski carpets . _
random walks and discrete potential theory ( cortona , 1997 ) _ , 2655 , _ sympos .
_ , xxxix , cambridge univ . press , cambridge . a.s .
cherny , a.n .
shiryaev and m. yor , ( 2002 ) limit behaviour of the `` horizontal - vertical '' random walk and some extensions of the donsker - prokhorov invariance principle .
_ theory probab .
appl . _ * 47 * no.3 , 377394 .
whittaker and g.n .
watson , ( 1962 ) _ a course of modern analysis .
an introduction to the general theory of infinite processes and of analytic functions : with an account of the principal transcendental functions . _ fourth edition .
cambridge univ . press , new york . | we analyze the differences between the horizontal and the vertical component of the simple random walk on the 2-dimensional comb .
in particular we evaluate by combinatorial methods the asymptotic behaviour of the expected value of the distance from the origin , the maximal deviation and the maximal span in @xmath0 steps , proving that for all these quantities the order is @xmath1 for the horizontal projection and @xmath2 for the vertical one ( the exact constants are determined ) .
then we rescale the two projections of the random walk dividing by @xmath1 and @xmath2 the horizontal and vertical ones , respectively .
the limit process is obtained . as a corollary of the estimate of the expected value of the maximal deviation ,
the walk dimension is determined , showing that the einstein relation between the fractal , spectral and walk dimensions does not hold on the comb .
+ + * keywords : * random walk , maximal excursion , generating function , comb , brownian motion + * ams 2000 subject classification : * 60j10 , 05a15 , 60j65 |
in noncentrosymmetric magnetic crystals , chiral asymmetry of the exchange coupling that arises due to the crystallographic handedness leads to the formation of long - period modulations of the magnetization with a fixed rotation sense @xcite .
phenomenologically , these _ chiral _ dzyaloshinskii - moriya ( dm ) interactions are described by energy contributions linear in the first spatial derivatives of the magnetization @xcite @xmath0 where @xmath1 is a unit vector along the magnetization @xmath2 , @xmath3 and @xmath4 .
theoretical analysis shows that invariants of type ( [ form ] ) stabilize different types of _ homochiral _ modulations that propagate in one direction ( _ helices _ ) and two directions ( _ skyrmion lattices _ ) over a broad range of thermodynamical parameters @xcite . during the last four decades
, homochiral long - period modulations have been reported in many of noncentrosymmetric ferro- and antiferromagnets , although only in a form of simple helical waves ( see bibliography in ref . ) . in bulk chiral magnets complex multidimensional chiral modulations including specific skyrmionic textures
are reported to exist only in close vicinity to the neel temperatures of mnsi and other cubic helimagnets as so called _ precursor states _ ( e.g. , ref . and bibliography in ref . ) .
the synthesis of nanolayers of cubic helimagnets @xcite and ultra - thin films of common magnetic metals with induced dm interactions @xcite radically changed the situation . during last few years a large variety of skyrmion - lattice states and nontrivial helical modulations have been discovered in free - standing films and cubic helimagnetic epilayers @xcite and fe nanolayers @xcite .
a sufficiently strong magnetocrystalline anisotropy or / and an applied magnetic field will destroy chiral modulated states and establish homogeneously magnetized phases as thermodynamically stable states @xcite .
in such `` saturated '' phases , the dm interactions stabilize two- and three - dimensional localized states ( _ chiral _ skyrmions ) @xcite .
solutions for axisymmetric 2d chiral skyrmions have been investigated in a number of papers @xcite , numerical solutions for 3d isolated chiral skyrmions ( also known as _ hopfions _ ) are derived in ref . .
isolated chiral 2d skyrmions have been recently observed in cubic helimagnet films @xcite and fepd nanolayers @xcite .
these axisymmetric solitonic states of nanoscale size have attracted a lot of attention both in fundamental nonlinear physics and as promising objects for different spintronic applications @xcite .
it was found recently that the solutions for skyrmions and helicoids in thin layers of cubic helimagnets are inhomogeneous along the film thickness : their structure can be thought as a superposition of common in - plane modulations and specific _ twisted _ modulations propagating in the perpendicular direction @xcite .
these twisted distortions drastically change the energetics of skyrmion and helicoid phases and stabilize them in a broad range of applied magnetic fields .
experimental investigations of mnsi / si(111 ) epilayers in strong in - plane magnetic fields have led to the discovery of twisted chiral modulations in a form of localized surface states @xcite . these observations and supporting theoretical analysis show that the surface twists persist at arbitrarily high fields and prevent a total saturation of the magnetization .
evidence for this type of surface modulation can also be found in calculations of chiral modulations in confined cubic and uniaxial helimagnets @xcite . in this paper
we present experimental and theoretical investigations of surface modulation instabilities ( twisted states ) that arise in mnsi / si(111 ) epilayers .
we show that the occurrence of such surface chiral twists is a general phenomenon that exists in both bulk and confined noncentrosymmetric magnets .
we develop a consistent theory of surface twists and investigate their influence on modulated phases in chiral magnets .
within the phenomenological theory introduced by dzyaloshinskii @xcite , the magnetic energy density of a non - centrosymmetric ferromagnet can be written as @xcite @xmath5 where @xmath6 is the exchange stiffness constant , and @xmath7 @xmath8 .
the energy density @xmath9 consists of magnetic interactions independent of spatial derivatives of the order parameter and includes zeeman and magnetocrystalline anisotropy ( @xmath10 ) energies as main contributions , @xmath11 ( @xmath12 is the external magnetic field ) .
the dzyaloshinskii - moriya ( dm ) energy functional @xmath13 contains the so called _ lifshitz invariants _ , which are a combination of the differential forms ( [ form ] ) compatible with the crystal symmetry @xcite .
particularly , for noncentrosymmetric uniaxial ferromagnets belonging to crystal classes @xmath14 and @xmath15 ( @xmath16 = 3 , 4 , 6 ) , such as multiferroics b@xmath17a@xmath18feo@xmath19 ( a = ca , sr , pb , ba ) with space group @xmath20 @xcite and helimagnets cscucl@xmath19 ( p6@xmath2122 ) @xcite and cr@xmath22 nb s@xmath23 ( p6@xmath1922 ) @xcite , the dm energy functionals have the following form:@xcite @xmath24 when the two dzyaloshinskii constants are equal , @xmath25 , the functional @xmath13 ( [ lifshitz2 ] ) becomes , @xmath26 which describes the dm energy of cubic helimagnets with a b20-structure ( space group p2@xmath213 ) such as mnsi and fege @xcite .
the equilibrium states of a confined noncentrosymmetric ferromagnet are derived by minimization of the energy functional @xmath27 with corresponding boundary conditions @xcite .
the first integral over the sample volume includes the energy contributions ( [ density ] ) , and the second integral over the surface comprises surface - related energies with a surface density @xmath28 .
here we consider a bulk noncentrosymmetric ferromagnet with chiral modulations suppressed by an applied magnetic field or a uniaxial magnetocrystalline anisotropy .
we show that lifshitz invariants @xmath29 ( eq . ( [ form ] ) ) create surface modulations in an otherwise uniform magnetic state . as a model we consider a semi - infinite chiral ferromagnet that occupies @xmath30 with an applied magnetic field along the @xmath31axis .
we first discuss 1-d modulations where the angle of the magnetization with respect to the @xmath31axis , @xmath32 , remains homogeneous in the @xmath33 planes .
the magnetization twists about the @xmath34 semiaxis from @xmath35 at the surface to @xmath36 in the depth of the sample . depending on the form of the lifshitz invariants , the modulations will either be cycloidal , @xmath37 for uniaxial ferromagnets with @xmath14 symmetry ( eq . [ lifshitz1 ] ) or helicoidal , @xmath38 for uniaxial ferromagnets with @xmath15 symmetry ( eq . [ lifshitz2 ] ) and for cubic helimagnets as shown in fig .
[ twist0](a ) and ( b ) respectively .
the energy functional for isolated one - dimensional surface modulations can then be written as , @xmath39 } _ { w(\theta , \theta_x)}dx + w^{(s ) } ( \theta)|_{x = 0 } , \label{energy2}\end{aligned}\ ] ] with the boundary conditions @xmath40 , @xmath41 and @xmath42 ( @xmath43 )
. the potential @xmath44 represents the energy contributions that do not depend on spatial derivatives , such as bulk anisotropies and is defined as @xmath45 , the difference between energy densities of the modulated state ( @xmath46 ) and that of the homogeneous state @xmath47 .
the boundary energy is given by @xmath48 .
first we consider solutions in a cubic helimagnet that neglect magnetic anisotropies , which are typically weak in these systems . in a bulk ( infinite )
cubic helimagnet , the dm interaction induces single harmonic chiral modulations propagating along the applied magnetic field with wave vector @xmath49 ( the _ cone _ phase ) @xcite .
the helix period , @xmath50 and saturation field , @xmath51 are expressed via basic magnetic parameters , @xcite @xmath52 for @xmath53 , a helimagnet is in the saturated state with @xmath54 . in this case
minimization of functional ( [ energy2 ] ) with @xmath56 ( free boundary conditions ) and a subsequent optimization of the mean energy with respect of parameter @xmath57 yield the following analytical solutions for the chiral surface states in a cubic helimagnet @xmath58 , \ \ \ \label{sol1}\end{aligned}\ ] ] where @xmath59 and @xmath60 for positive dm constants ( @xmath61 ) in functional ( [ energy2 ] ) , twisted states have clockwise modulations , and an opposite rotation sense for @xmath62 ( fig .
[ twist ] ) .
the spatial extent of these surface states can be characterized by a _ penetration length _ defined as the point where the tangent at @xmath63 intersects the @xmath64-axis ( fig .
[ twist ] ) , @xmath65 of a uniaxial noncentrosymmetric ferromagnets with @xmath66 ( @xmath67 ) symmetry ( a ) , and a cubic helimagnet or a uniaxial ferromagnets with @xmath68 ( @xmath69 ) symmetry ( b ) , in applied fields @xmath70 .
the angle @xmath57 is the angles of magnetization with respect to the field direction at the surface of the magnet .
[ twist0 ] ] describes localized chiral twists near lateral surfaces of a isotropic helimagnet at in - plane magnetic field @xmath71 ( eq .
( [ sol1 ] ) ) .
inset shows @xmath57 as a function of the applied field for different values of uniaxial surface anisotropy ( @xmath72 ) ( solutions to eq .
( [ theta02 ] ) ) . [ twist ] ] figure [ twist ] shows a typical distribution of the magnetization in surface chiral twist states , as given by eq .
( [ sol1 ] ) . at @xmath71
the deviation of the magnetization on the surface reaches value @xmath73 at @xmath71 and slowly decreases in increasing magnetic field ( inset in fig .
[ twist ] ) ) . in a general case of functional ( [ energy2 ] )
the euler equation yields the first integral and the solution for a penetration length , @xmath74 inserting solutions ( [ sol0 ] ) into ( [ energy2 ] ) leads to the following expression for the equilibrium energy @xmath75 optimizing ( [ energy3 ] ) with respect to @xmath57 leads to the equation @xmath76 for cycloidal modulations , there is a demagnetizing field that takes the form of a uniaxial anisotropy that must be included in the potential @xmath44 , whereas the helicoidal modulations in this geometry have no stray magnetic field . for the case of helicoidal modulations and when @xmath56 eqs .
( [ sol0 ] ) , ( [ theta0 ] ) yield the solutions @xmath77 note that in this case @xmath78 is equal to the wave vector of helical modulations propagating in bulk cubic helimagnets , @xmath79 ( [ period ] ) .
finite values of spatial derivatives of the magnetization on the surface of saturated chiral ferromagnets signify their instabilities with respect to chiral surface modulations . by using the first integral of functional ( [ energy2 ] ) , eq .
( [ theta0 ] ) can be transformed into the following form @xmath80_{x = 0 } = 0 , \label{boundary}\end{aligned}\ ] ] which coincides with the regular boundary condition @xmath81 = @xmath82 for functional ( [ energy2 ] ) .
the influence of boundary condition ( [ boundary ] ) on vortex and skyrmion states in magnetic nanodots has been investigated in refs . .
chiral symmetry breaking and a modification of electronic and chemical properties at surfaces and interfaces of magnetic nanolayers induce specific magnetic interactions responsible for a number of phenomena unknown in bulk counterparts of the same magnetic materials @xcite .
these induced interactions either penetrate into the depth of nanolayers and are described by different forms of volume energy contributions or are localized near the surface and described by surface energy contributions @xcite .
the latter usually include surface / interface uniaxial anisotropy energy , @xmath83 , and/or the energy of exchange coupling with adjacent magnetic layers ( so called unidirectional anisotropy ) , @xmath84 @xmath85 in chiral magnets , surface contributions of type ( [ surfaceenergy ] ) influence the parameters of chiral twists .
particularly , for @xmath86 , eq . ( [ theta0 ] ) for the equilibrium @xmath57 reads @xmath87 where @xmath88 .
the solutions of eq .
( [ theta02 ] ) plotted as functions @xmath89 for different values of @xmath72 ( inset of fig .
[ twist ] ) show that easy - plane anisotropy ( @xmath90 ) extends the twisted modulations while easy - axis anisotropy ( @xmath91 ) suppresses them .
note that @xmath57 preserves finite values even in a case of strong easy - axis surface anisotropy ( @xmath92^{-1 } < < 1 $ ] ) .
twisted states also arise in the more common achiral magnets and other condensed matter systems as a result of a competition between surface and volume interactions .
initially , twisted states have been identified in thin layers of nematics confined between two parallel plates ( _ freedericksz _ transition @xcite ) .
later , similar surface - induced distortions have been observed in many classes of liquid crystals , and now this phenomenon forms the physical basis for liquid - crystal display technologies @xcite .
theoretical investigations of magnetic twisted states have been started by goto et al .
@xcite and have been developed in a number of other works ( see e.g. ) including twists in the magnetization near the surface of bulk ferromagnets @xcite .
twisted states in centrosymmetric confined ferro- and antiferromagnets have an arbitrary rotation sense and arise only under the condition of specific relations between surface and volumes magnetic anisotropies @xcite .
currently , there is only indirect experimental evidence for the existence of these surface states in centrosymmetric magnets.@xcite in the depth of bulk noncentrosymmetric ferromagnets skyrmions are homogeneous along their axes ( @xmath93-axis in this paper ) and are described by axisymmetric solutions of type @xmath94 , @xmath95 @xcite ( particularly , @xmath96 in cubic helimagnets and in uniaxial ferromagnets with @xmath68 symmetry ) . near the sample surfaces lifshitz invariants with gradients along the skyrmion axis impose twisted distortions . in cubic helimagnets these distortions can be written as @xmath97 , @xmath98 , and are derived by minization of energy functional @xmath99 in common magnetically soft materials , magneto - dipole interactions play an important role @xcite .
however , the effects of spatially inhomogeneous stray - fields in noncentrosymmetric magnets are weakened by the dm interactions @xcite , and are ignored here . in bulk helimagnets and thick layers , with thickness @xmath100 much larger than the penetration depth @xmath101 given by eq .
( [ lambda ] ) , the twisted modulations exist only in a narrow vicinity of the surface , and can be considered as localized states .
the expansion of the skyrmion energy leads to the energy functional @xmath102 for the twisted states @xmath103 , where @xmath104 @xmath105 , @xmath106 $ ] and @xmath107 are the solutions for skyrmions in the depth of the sample , where @xmath108
. energy density ( [ skyrmiontwist1 ] ) has the same functional form as that for isolated one - dimensional twists , eq .
( [ energy ] ) , and thus surface twists are described by solutions of type ( [ sol1 ] ) .
similar chiral surface twists arise in helicoids . in a thin helimagnet layer , with @xmath100 of the order of @xmath109 ,
twisted modulations exist across the depth of the film .
the solutions for skyrmions and helicoids in thin cubic helimagnet films have been investigated in ref . .
it was found that twisted modulations substantially modify the energetics of chiral modulations and stabilize helicoids and skyrmion lattices in a broad range of applied fields @xcite .
the interaction between homochiral localized modulations arising in noncentrosymmetric magnets has a repulsive character .
consider a semi - infinite sample with the surface at @xmath110 , and an isolated skyrmion with its axis along the @xmath31axis .
when the skyrmion and a localized surface twist are separated by a large distance @xmath111 , the interaction potential includes the dm energy of a surface twist as the main contribution to @xmath112 @xcite .
the expansion of eq .
( [ sol1 ] ) for @xmath113 yields @xmath114 } \label{interaction}\end{aligned}\ ] ] recent numerical simulations have demonstrated that chiral skyrmions are easily induced , transported and controlled by electric currents and applied magnetic fields in narrow strips of helimagnets @xcite . in such a track ,
the repulsive forces imposed by surface twists create a lateral confining potential for moving skyrmions along the center of the strip ( @xmath110 ) , @xmath115/(2 w_{\mathrm{int } } ( r))$ ] for @xmath116 ( fig .
[ track ] ) .
in bulk cubic helimagnets , at low temperature the magnetic structure crosses two transition fields on its way to saturation . the first transition at a field @xmath117 corresponds to a rotation of the propagation vector in the direction of the field and a canting of the spins into a conical state@xcite .
this is followed by a second - order collapse of the conical phase into a field induced ferromagnetic state above a field @xmath118 .
the situation for epitaxial cubic helimagnetic thin films is altered by strain induced uniaxial anisotropy and the breaking of symmetry caused by the presence of the interfaces . in the case of mnsi
thin films , the ground state is a helix with an out - of - plane propagation vector@xcite . at low temperatures ,
the uniaxial anisotropy prevents a reorientation of the helix in magnetic fields applied in - plane , and the propagation vector remains fixed along the film normal .
the in - plane field produces helicoids with a discrete number of turns , and the magnetic state evolves by a set of first - order transitions between these states.@xcite the final state is not the uniform ferromagnetic state envisioned in bulk crystals , but rather is a ferromagnetic state with surface twists , as demonstrated by the polarized neutron reflectometry curves in ref . , that show no signs of saturation . to compare with the analytical expressions in this paper , we conducted squid magnetometry of a series of samples that range between 11.5 nm and 40 nm .
samples were grown by molecular beam epitaxy on si(111 ) substrates as described in ref . , and
measurements were conducted in a quantum design mpms xl5 squid magnetometer .
the helical wavelength @xmath119 nm is obtained from polarized neutron reflectometry ( pnr ) and squid analysis.@xcite squid magnetometry of mnsi / si(111 ) is complicated by the large diamagnetic contribution from the si substrate .
one typically determines the substrate contributions from the high - field portion of the @xmath120-@xmath121 loops , far above the saturation field .
however , in the case of mnsi , the high - field susceptibility is non - zero,@xcite which makes it impossible to separate the film s contribution from that of the substrate with this method . in order to determine the high - field susceptibility of our mnsi films
, we performed pnr on a @xmath122 nm mnsi film at a temperature of @xmath123 k , below the curie temperature of @xmath124 k. pnr provides a depth dependent measurement of the magnetization of the film without any contribution from the diamagnetic response of the substrate .
we conducted measurements with 0.237 nm neutrons on the d3 reflectometer at the canadian neutron beam centre mounted with the m5 superconducting magnet cryostat.@xcite an fe / si supermirror and a mezei - type precession spin flipper produced a neutron beam with a spin polarization in excess of 95% . to avoid attenuating the pnr signal , we conducted the experiments without an analyzer , since our previous measurements show that the spin - flip signal is negligible due to the cancelling effects of the bi - chiral magnetic domains.@xcite from the measured spin - up @xmath125 , and spin - down , @xmath126 , reflectivities , we calculated the spin - asymmetery @xmath127 .
the first peak in @xmath128 at low scattering vectors @xmath129 nm provides an excellent measure of the average magnetization across the film . in order to extract the magnetization , we fitted the measurement of @xmath128 in the inset of fig .
[ fig : mh ] to simulations calculated with simulreflec .
the calculated non - spin - flip reflectivities , @xmath130 and @xmath131 , are corrected for non - idealities in the beam polarization before they are compared with experiments .
the corrections , @xmath132 also account for the absence of the analyzer during the reflectivity experiments . the flipping ratio , @xmath133 , is obtained by measuring the unreflected beam intensities @xmath134 and @xmath135 that pass through both the polarizer and analyzer , where the subscripts denote the polarization of the neutron beam after passing through the polarizer and analyzer respectively .
the flipping ratio varies between 20 to 33 over the range of fields measured .
the structural parameters in the simulations are obtained from the x - ray reflectometry and pnr measurements presented in ref . .
the magnetization profile is calculated from eq .
( [ sol1 ] ) by accounting for the surface twists at both interfaces .
the sole fitting parameter to the pnr data in the inset of fig .
[ fig : mh ] is the value of the average magnetization , which is plotted for various fields in the main figure .
the value for the substrate susceptibility subtracted from the squid measurements is adjusted to bring the magnetometry measurements into agreement with the pnr data , as shown by the open circles in fig .
[ fig : mh ] .
since squid measurements show that the si susceptibility is independent of temperature below @xmath136 k , we were then able to accurately remove the substrate contribution for all measured temperatures .
the high - field susceptibility of the mnsi film at @xmath123 k is @xmath137 ka / m / t , which is in good agreement with @xmath138 ka / m / t for bulk mnsi ( estimated from ref . ) at the same reduced temperature .
we therefore assume that the high - field susceptibility is the same for all thicknesses presented in this paper and use this value to remove the substrate susceptibility from the other samples . from a @xmath139-nm @xmath140 thick mnsi film measured at @xmath123
k in fields @xmath141 t ( red triangles ) , 2.2 t ( black circles ) and 3.2 t ( blue squares ) .
the lines are fits to the data described in the text .
the magnetization obtained from the fits is plotted in the main figure ( filled points ) along with the substrate - subtracted squid magnetometry data ( open points ) .
the line is a guide to the eye .
all error bars are @xmath142 . ] measured by squid magnetometry ( filled points ) for ( a ) @xmath143 and 5 k in a @xmath144 film , where @xmath145 nm is the helical wavelength .
the estimated values for @xmath51 are 0.49 , 0.75 , 0.80 and 0.82 t respectively , and the high - field magnetization is @xmath146 ka / m .
the large open triangular points superimposed on the @xmath123 k data correspond to pnr measurements .
the red lines in ( a ) and ( b ) are calculations given by eq .
[ eq : deltam ] .
figure ( b ) shows @xmath147 for four films with thicknesses from @xmath148 to 2.85 @xmath50 , where the curves are vertically offset for clarity .
the red lines are calculations given by eq .
[ eq : deltam ] , where @xmath149 t in order of increasing thickness .
the blue curves show exact numerical calculations of the twisted ferromagnetic state that account for the finite size , as described in ref . .
] in the limit of large field , @xmath53 , and large thickness , @xmath150 , the solution given by eq .
( [ sol1 ] ) provides the field dependence of the magnetization in the film , @xmath151 where @xmath152 is the reduced magnetization @xmath153 . to compare with theory , the saturation magnetization @xmath154 is determined by the value of @xmath155 that gives a linear @xmath156 dependence on @xmath157 at fields @xmath53 .
the values of @xmath51 are obtained using the procedure described in ref . .
the plots of @xmath156 vs @xmath157 in fig .
[ fig : mh2 ] show linear regimes for @xmath158 that correspond to the twisted ferromagnetic state , and a more complex @xmath120-@xmath121 dependence for @xmath159 that correspond to regions with a more complex magnetic texture .
figure [ fig : mh2 ] ( a ) shows that there is a clear departure of the data from the behavior predicted by eq .
( [ eq : deltam ] ) .
however , as the temperature is dropped , the @xmath160 approaches the expected @xmath161 dependence at @xmath162 k. the fact that the magnitude of the discrepancy grows with temperature suggests that spin - fluctuations may be dominating the higher temperature behavior .
they would explain the change in slope of @xmath163 vs @xmath157 since the magnetization is expected to have a @xmath164 field dependence due to fluctuations.@xcite however , even at @xmath162 k , the data in fig .
[ fig : mh2 ] ( a ) is offset from the theoretical curve .
this offset can be eliminated by an increase in @xmath51 .
the presence of twisted surface states required a modification to the method that we use to extract values for @xmath51 , which assumes that as the field is lowered , the system transitions continuously from a saturated state into a conical state@xcite .
this is true for out - of - plane fields , where the saturation field , @xmath165 for in - plane fields , however , a ferromagnetic state with surface twists exists in place of a conical phase and there is no in - plane saturation field,@xmath166 .
if we use the minimum in @xmath167 , as done in refs . , to identify an effective @xmath166 , we obtain estimates for @xmath51 that are approximately 15% lower than required to fit the discrete helicoidal transition fields in ref . .
we use these values as crude estimates of @xmath51 .
the theory for twisted states presented in section [ sec:1dtheory ] provides , in principle , a method of extracting values for @xmath51 from in - plane high magnetic field measurements of @xmath168 .
we used eq .
( [ eq : deltam ] ) to fit the magnetization curves by treating @xmath51 and @xmath155 as fitting parameters .
the results for films from @xmath169 nm to @xmath170 nm shown in fig .
[ fig : mh2](b ) demonstrate that eq .
( [ eq : deltam ] ) fits the data well , but unfortunately we find that the fits yield @xmath51 values that are inconsistent with previous measurements .
for example , we find @xmath171 t for the @xmath172 film , which is too large given @xmath173 t and eq .
( [ hc2])@xcite , and is inconsistent with fits to the helicoidal state for in - plane magnetic fields@xcite . therefore the values for @xmath51 obtained by using eq .
( [ eq : deltam ] ) can only be taken as effective characteristic fields , @xmath174 , which include other factors such as surface anisotropies or extrinsic effects that affect the saturation of the twisted state .
these results highlight the need for independent measurements of @xmath175 so that more accurate values of @xmath51 can be achieved . in fig .
[ fig : mh2 ] ( b ) , we compare the analytical expression ( red ) with the exact numerical calculations ( blue ) . for the films @xmath176 ,
the structure remains in a twisted ferromagnetic state down to the lowest fields , and is well described by the numerical calculation over the entire field range .
as expected , the analytical expression is in good agreement with the calculation for @xmath53 . in this figure
, we also find good agreement between the numerical calculation and the data for the @xmath177 film above @xmath53 .
the departure of the data from the model at lower fields reflects a transition to a helicoidal state at @xmath178 , below which the twisted state is a metastable state .
the difference between the analytical ( red ) and numerical ( blue ) curves is very small due to the fact that the twists at the two interfaces interact very little with each other at this film thickness . for larger film thicknesses
there is almost no difference , and so the numerical calculations have been left out of the figure .
experimental observations of mnsi / si ( 111 ) cubic helimagnet epilayers and theoretical calculations within the standard model ( [ density ] ) show that the formation of chiral surface twists is a general phenomenon attributed to all saturated noncentrosymmetric magnetic crystals . in bulk and confined chiral magnets
for which the thickness @xmath100 is much larger than _ penetration length _
@xmath109 ( [ sol0 ] ) , the surface twists that arise are strongly localized near individual surfaces and do not interact with each other . in thin layers of chiral magnets ( @xmath179 ) surface twists in the saturated phases create bound states @xcite , and helicoids and skyrmions exist as a superposition of common in - plane modulations and conical modulations along the film thickness .
contrary to achiral twisted phases in common ( centrosymmetric ) magnets , which are predicted to exist only for specific relations between material parameters of magnetic materials @xcite , chiral surface twists exist in broad range of thermodynamic parameters and can be of practical interest for spintronics applications @xcite .
we would like to thank filipp rybakov for helpful discussions and eric karhu for technical assistance .
acknowledges financial support by dfg through grant no .
bo 4160/1 - 1 .
t.l.m . and m.n.w .
acknowledge support from nserc .
the research presented herein is made possible by a reflectometer jointly funded by canada foundation for innovation ( cfi ) , ontario innovation trust ( oit ) , ontario research fund ( orf ) , and the national research council canada ( nrc ) .
57ifxundefined [ 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 * * , ( ) , http://www.jetp.ac.ru/cgi-bin/e/index/e/68/1/p101?a=list [ * * , ( ) ] , \doibase doi : 10.1016/0304 - 8853(94)90046 - 9 [ * * , ( ) ] http://stacks.iop.org/0038-5670/27/i=11/a=r04 [ * * , ( ) ] link:\doibase 10.1103/physrevb.66.214410 [ * * , ( ) ] http://stacks.iop.org/1742-6596/303/i=1/a=012105 [ * * , ( ) ] link:\doibase feb 13 [ * * , ( ) ] http://link.aps.org/doi/10.1103/physrevlett.102.197202 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.127203 [ * * , ( ) ] http://stacks.iop.org/0953-8984/24/i=29/a=294204 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.077207 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.220406 [ * * , ( ) ] link:\doibase 10.1038/nature09124 [ * * , ( ) ] link:\doibase doi 10.1103/physrevb.82.184417 [ * * , ( ) ] http://dx.doi.org/10.1038/nmat1646 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.144420 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.267201 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.214420 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys2045 [ * * , ( ) ] link:\doibase 10.1126/science.1240573 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.052403 [ * * , ( ) ] http://stacks.iop.org/0022-3727/44/i=39/a=392001 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.094424 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.094411 [ * * , ( ) ] http://ezproxy.library.dal.ca/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=aph&an=55369153&site=ehost-live [ * * , ( ) ] , http://dx.doi.org/10.1038/nnano.2013.29 [ * * , ( ) ] http://stacks.iop.org/0295-5075/101/i=3/a=37001 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.184422 [ * * , ( ) ] http://dx.doi.org/10.1038/nnano.2013.176 [ * * , ( ) ] http://dx.doi.org/10.1038/nnano.2013.210 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2836802 [ * * , ( ) ] link:\doibase 10.1103/physrevb.65.052402 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.107202 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.224405 [ * * , ( ) ] http://dx.doi.org/ [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1080/00018739300101524 [ * * , ( ) ] in @noop _ _ , ( , , ) http://www.sciencedirect.com/science/article/pii/s0304885301008460 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1063/1.1714613 [ * * , ( ) ] http://www.sciencedirect.com/science/article/pii/030488539291263s [ * * , ( ) ] link:\doibase 10.1103/physrevb.68.012407 [ * * , ( ) ] link:\doibase 10.1103/physrevb.39.12306 [ * * , ( ) ] link:\doibase 10.1103/physrevb.42.6568 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.71.444 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.020415 [ * * , ( ) ] @noop _ _ ( , , ) @noop * * , ( ) , http://dx.doi.org/10.1088/0022-3719/14/31/016 [ * * , ( ) ] link:\doibase dec [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.060404 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.094429 [ * * , ( ) ] http://dx.doi.org/ [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2130666 [ * * , ( ) ] http://stacks.iop.org/0953-8984/21/i=29/a=296003 [ * * , ( ) ] http://stacks.iop.org/0953-8984/11/i=39/a=315 [ * * , ( ) ] | in epitaxial mnsi / si(111 ) films , the in - plane magnetization saturation is never reached due to the formation of specific surface chiral modulations with the propagation direction perpendicular to the film surfaces [ wilson et al .
phys . rev .
b 88 , 214420 ( 2013 ) ] . in this paper
we show that the occurrence of such _ chiral surface twists _ is a general effect attributed to all bulk and confined magnetic crystals lacking inversion symmetry .
we present experimental investigations of this phenomenon in nanolayers of mnsi / si(111 ) supported by detailed theoretical analysis within the standard phenomenological model . in magnetic nanolayers with intrinsic or induced chirality ,
such surface induced instabilities become sizeable effects and play a crucial role in the formation of skyrmion lattices and other nontrivial chiral modulations . |
the standard model ( sm ) describes many phenomena up to the energies that can be reached today .
however , some problems like mass hierarchy and the number of fermion generations can not be addressed by the sm .
many models , which include new fermions and interactions , have been developed to deal with the mentioned shortcomings of the sm .
the string inspired @xmath2 model is a well motivated one which includes extra gauge bosons and new fermions assigned to the 27-dimensional representation @xcite . in the search for
extensions of the sm the new heavy leptons play an important role .
many analysis have been done for the production of heavy leptons at future @xmath3 @xcite , at hadron @xcite and also at @xmath4 colliders @xcite .
the experimental upper bounds for the heavy lepton masses were found to be 44 gev by opal @xcite , 46 gev by aleph @xcite and 90 gev by h1 @xcite collaborations .
this means that , heavy leptons could be detected at future high energy colliders .
we have previously considered the single production of new heavy leptons via string inspired @xmath0 model in @xmath4 collisions @xcite . in this study , we consider both the single and the pair production of heavy leptons separately by using the same model and taking into account the signal and background events at future linear colliders
. the main parameters of these collider options were taken from refs . @xcite and displayed in table [ tab : table1 ] .
the single production of heavy leptons @xmath5 , in @xmath3 collisions occur through the @xmath6 and @xmath7 channel processes @xmath8 caused by the flavor changing neutral current ( fcnc ) lagrangian : @xmath9 where @xmath10 are the mixing angles between right handed components of the ordinary and new heavy charged leptons .
the order of the mixings of the ordinary and heavy leptons are known to be @xmath11 , which comes from low energy phenomenological calculations and the high precision measurements of the @xmath12 properties at linear colliders @xcite .
we use the parameter @xmath13 to denote the mixing angles and take @xmath14 as an upper value in the numerical calculations .
the differential production cross section takes the form , @xmath15\left[(t - m_z^2)^2+m_z^2\gamma_z^2\right]}\right.\nonumber\\ & & \left .
-\frac{(m^2-s - t)(s+t)(a_e - v_e)^2+t(m^2-t)(a_e+v_e)^2 } { ( s - m_z^2)^2+m_z^2\gamma_z^2}\right.\nonumber\\ & & \left.+ \frac{\left[\left[(s+t)^2+s^2-(2s+t)m^2\right](a_e^2+v_e^2)+2t(m^2 - 2s - t)a_ev_e\right]}{(t - m_z^2)^2+m_z^2\gamma_z^2}\right]\end{aligned}\ ] ] where @xmath16 and @xmath17 , @xmath18 is the decay width and @xmath19 is the mass of @xmath12 boson , @xmath6 and @xmath7 are mandelstam variables . in fig .
[ fig : f1 ] , we display the total cross sections as function of the heavy lepton masses , for the three center of mass energies of the proposed options .
after their production , the heavy leptons will decay via the neutral current process @xmath20 , where @xmath21 .
the branching ratios for these processes would be around @xmath22 for each channel . in tables [ tab : t2 ] , [ tab : t3 ] and [ tab : t4 ]
, we presented the single production cross sections ( @xmath23 ) , signal and background cross sections depending on the heavy lepton mass @xmath24 , for 0.5 , 1 and 3 tev energy @xmath3 colliders , respectively .
the branching ratios @xmath25 and @xmath26 refer to @xmath27 and @xmath28 . the significance of signal and background is defined as @xmath29 ,
here @xmath30 and @xmath31 are the signal and background number of events .
the total decay widths of the heavy leptons are given in the last column of the tables .
as seen from these tables , the @xmath29 values are higher than five , which is enough for observability , up to the center of mass energies of the colliders .
single production of heavy lepton is feasible up to the center of mass energies of the @xmath3 colliders even with smaller mixing coupling values .
for example , if we take at least 10 signal events and @xmath32 as discovery criteria , the ilc with @xmath330.5 tev can probe mixing values of @xmath34 for 350 gev leptons .
the same couplings can be probed at @xmath331 and @xmath333 tev for even greater masses such as 800 and 2750 gev .
we applied a cut of @xmath35 gev in order to form the signal and reduce the background for the sm background process @xmath36 .
[ fig : f2 ] shows the @xmath37 distributions at three different linear colliders . in figs .
[ fig : f3 ] , [ fig : f4 ] and [ fig : f5 ] , we give the invariant mass distributions @xmath38 with cut @xmath39 gev at @xmath330.5 , 1 and 3 tev , respectively . figs .
[ fig : f3 ] and [ fig : f4 ] have an increasing character for a cut 10 gev interestingly , but lose this character for higher cuts .
for instance , all of the three distributions in figs .
[ fig : f3]-[fig : f5 ] have decreasing characters with jacobian peaks around 150 - 200 gev for a cut value of 80 gev .
pair production of heavy leptons in @xmath2 occur through the @xmath7-channel flavor changing neutral current process @xmath40 , and the differential cross section for this process is given by , @xmath41}\left[(m^2-t)^2m^4 + 4m_z^2m^4s+4(s+t - m^2)^2m_z^4\right].\ ] ] the total cross sections as functions of heavy lepton masses @xmath24 , are displayed in fig .
[ fig : f6 ] . signal and background cross sections depending again on the heavy lepton masses , are presented in tables [ tab : t5 ] , [ tab : t6 ] and [ tab : t7 ] at 0.5 , 1 and 3 tev , respectively . for the pair production of heavy leptons at linear colliders
we expect of order of @xmath42 signal events for 1250 gev leptons for the coupling value of @xmath43 . on the other hand ,
the lower limit of the coupling which can be probed by pair production at linear colliders is found to be 0.05 .
we applied an initial cut on the electron and jet transverse momentum @xmath44 gev for the signal and background analysis .
[ fig : f7 ] shows the @xmath37 distributions of the background at the colliders .
the distribution of invariant mass @xmath45 is presented in figs .
[ fig : f8 ] , [ fig : f9 ] and [ fig : f10 ] at @xmath330.5 , 1 and 3 , respectively .
we have used the high energy package comphep for calculations of background cross sections reported in this study @xcite .
this study proves that linear colliders can test the existence of single and pair production of heavy leptons .
the production of a single heavy lepton is more relevant than the pair production .
namely , in the case of @xmath1 tev option , we expect 257 single events for @xmath43 and 26 single events for @xmath34 for 2750 gev ( which is the upper bound ) leptons . in the case of pair production , for @xmath43
we expect @xmath42 events for 1250 gev ( upper value ) leptons , while no pair event can be observed for @xmath34 , since the pair production cross section is suppressed by the fourth power of mixing couplings .
j. l. hewett and t. g. rizzo , phys .
rept . *
183 * , 193 ( 1989 ) .
f. m. l. almeida , j. h. lopes , j. a. martins simoes and c. m. porto , phys .
d * 44 * , 2836 ( 1991 ) .
f. m. l. almeida , j. h. lopes , j. a. martins simoes , p. p. queiroz filho and a. j. ramalho , phys .
d * 51 * , 5990 ( 1995 ) f. m. l. almeida , y. a. coutinho , j. a. martins simoes and m. a. b. do vale , phys .
d * 63 * , 075005 ( 2001 ) f. m. l. almeida , y. a. coutinho , j. a. martins simoes , s. wulck and m. a. b. do vale , eur .
j. c * 30 * , 327 ( 2003 ) p. h. frampton , d. ng , m. sher and y. yuan , phys .
d * 48 * , 3128 ( 1993 ) y. a. coutinho , j. a. martins simoes , c. m. porto and p. p. queiroz filho , phys .
d * 57 * , 6975 ( 1998 ) .
j. e. cieza montalvo and p. p.
de queiroz filho , phys .
d * 66 * , 055003 ( 2002 ) .
t. g. rizzo , phys .
b * 188 * , 95 ( 1987 ) .
a. t. alan , a. t. tasci and o. cakir , acta phys .
b * 35 * , 2199 ( 2004 ) .
m. z. akrawy _ et al . _
[ opal collaboration ] , phys .
b * 240 * , 250 ( 1990 ) .
d. decamp _ et al . _
[ aleph collaboration ] , phys .
rept . * 216 * , 253 ( 1992 ) .
t. ahmed _ et al . _
[ h1 collaboration ] , phys .
b * 340 * , 205 ( 1994 ) .
e. accomando
_ et al . _
[ clic physics working group ] , arxiv : hep - ph/0412251 .
international linear collider - technical review report , ilc - trc/2003 report ( 2003 ) , http://www.slac.stanford.edu/xorg/ilc-trc/2002/index.html p. bamert , c. p. burgess and i. maksymyk , phys .
b * 356 * , 282 ( 1995 ) f. m. l. almeida , y. a. coutinho , j. a. martins simoes and m. a. b. do vale , phys .
d * 62 * , 075004 ( 2000 ) .
e. nardi , e. roulet and d. tommasini , phys .
b * 327 * , 319 ( 1994 ) a. pukhov _ et al .
_ , `` comphep : a package for evaluation of feynman diagrams and integration over multi - particle phase space . '' , arxiv : hep - ph/9908288 . | we investigate the single and pair production of new heavy leptons via string inspired @xmath0 model at future linear colliders .
signal and corresponding backgrounds for these leptons are studied .
we have found that single production of heavy leptons is more relevant than that of pair production , as expected . for a maximal mixing value of 0.1 ,
the upper mass limits of 2750 gev in the single case and 1250 gev in the pair production case are obtained at @xmath1 tev collider option . |
as early as 1985 , bhalerao and liu @xcite performed a coupled - channel analysis of the @xmath6 , @xmath7 , and @xmath8 reactions in the close - to - threshold region , and discovered that the interaction between the nucleon and the @xmath0 meson is attractive .
this finding inspired haider and liu to postulate the existence of @xmath0-mesic nuclei @xcite , in which the neutral @xmath0 meson might be bound with nucleons via the strong interaction .
the existence of @xmath0-mesic bound states would allow to investigate the interaction of the @xmath0 meson and nucleons inside a nuclear matter .
moreover , it would provide information about the @xmath9 resonance @xcite and about the @xmath0 meson properties in a nuclear matter @xcite , as well as about contribution of the flavour singlet component of the quark - gluon wave function of the @xmath0 meson @xcite . according to theoretical considerations
, the formation of the @xmath0-mesic nucleus can only take place if the real part of the @xmath0-nucleus scattering length is negative ( attractive nature of the interaction ) , and the magnitude of the real part is greater than the magnitude of the imaginary part @xcite : @xmath10 + a wide range of possible values of the s - wave @xmath0n scattering length , from @xmath11=(0.27 + 0.22i ) fm up to @xmath11=(1.05 + 0.27i ) fm , calculated for hadronic- and photoproduction of the @xmath0 meson has not excluded the formation of @xmath0-nucleus bound states for a light nuclei as @xmath12 , t @xcite and even for deuteron @xcite .
those bound states have been searched for in many experiments @xcite , however none of them gave empirical confirmation of their existence .
there are only a promising experimental observations which might be interpreted as of @xmath0-mesic nuclei .
for example , experimental observations which might suggest the possibility of the existence of the @xmath13-@xmath0 bound system were found by @xcite , @xcite , anke @xcite and @xcite collaborations .
the data analysis of the close to threshold spes-4 and spes-2 measurements of the total cross led to the determination of the @xmath14 scattering length which , within the experimental errors , fulfill the condition given by eq . [ eq : eq1 ] . in the experiments conducted at the cosy
synchrotron the momentum ramping technique of the deuteron beam was used that allows to reduce the systematic uncertainties @xcite .
the beam was accelerated slowly and linearly in time , from excess energy of up to in case of the anke experiment @xcite , while during the experiment @xcite the momentum was varied in the range corresponding to the excess energy from to .
both collaborations performed the measurement of the excitation function and differential cross section of the reaction close to the kinematical threshold .
data taken by the cosy-11 group were used to search for a signal of a @xmath13-@xmath0 bound state below the @xmath0 production threshold of the @xmath15 and @xmath16 @xmath17 reactions @xcite , while the measurements above the threshold enabled the study of the forward - backward asymmetries of the differential cross sections and the extraction of the @xmath0-@xmath13 scattering length .
the data of both groups @xcite show a variation in the phase of the s - wave amplitude in the near - threshold region , consistent with possible existence of a bound state @xcite .
the first direct experimental indications of light @xmath0-nucleus bound states were observed in @xmath0 photoproduction @xmath18@xmath19 which was carried out by the taps collaboration @xcite .
the measurements of the excitation functions for two ranges of the relative angle between @xmath20 and proton were carried out .
it appeared that a difference between those excitation curves in the center - of - mass frame revealed an enhancement just below the threshold of the @xmath21 @xmath13-@xmath0 reaction which was interpreted as a possible signature of a @xmath13-@xmath0 bound state where @xmath0 meson captured by one of nucleons inside helium forms an intermediate @xmath22 resonance which decays into @xmath20-@xmath23 pair .
however , it was shown later that the result may be an artefact due to the strong influence of the resonances on the shape of the excitation function @xcite .
therefore , a new high - statictics exclusive measurement is necessary .
the measurement of the @xmath1-@xmath0 bound states is performed with high precision at the cosy accelerator with the wasa detection system .
signatures of the nuclei are searched for in the excitation function of specific decay channels of the system , formed in deuteron - deuteron collision @xcite .
the measurement is performed for beam momenta varying continously around the @xmath0 meson production threshold .
the beam ramping technique allows to reduce the systematic uncertainities .
the of the bound system should manifest itself as a resonance structure in the excitation curve of eg .
@xmath4 reaction below the @xmath24 @xmath1-@xmath0 reaction threshold .
the kinematics of the reaction is schematically presented in fig .
[ fig1 ] .
the deuteron beam - deuteron target collision leads to fusion into a @xmath1 nucleus in a bound state with the @xmath0 meson .
the @xmath0 meson can be absorbed by one of the nucleons inside the helium and may propagate in the nucleus via consecutive excitation of nucleons to the @xmath25 state @xcite until the resonance decays into a pion - proton pair emitted by the nucleus @xcite .
the relative angle between p and @xmath26 is equal to @xmath27 in the @xmath28 reference frame and it is smeared by about @xmath29 in the center - of - mass frame due to the fermi motion of the nucleons inside the helium nucleus .
-@xmath30 production and decay.[fig1],width=585,height=283 ] in june 2008 , the search for the @xmath1-@xmath0 bound state was performed by measuring the excitation function of the @xmath24 @xmath4 reaction near the @xmath0 production threshold . during the experimental run the momentum of the deuteron beam
was varied continuously within each acceleration cycle from 2.185 gev / c to 2.400 gev / c , crossing the kinematic threshold for the @xmath0 production in the @xmath24 @xmath31 reaction at 2.336 gev / c .
this range of beam momenta corresponds to a variation of @xmath1-@xmath0 excess energy from -51.4 mev to 22 mev .
the excitation function was determined after applying cuts on the p and @xmath26 kinetic energy distribution and the @xmath32 opening angle in the cm system @xcite .
the result is shown in fig .
[ fig2 ] .
@xmath4 reaction obtained by normalizing events selected in individual excess energy intervals by corresponding integrated luminosities .
the solid line represents a fit with second order polynomial combined with a breit - wigner function with fixed binding energy and a width equal to -10 and 10 mev , respectively .
the dotted line corresponds to the contribution from the second order polynomial in the performed fit.[fig2],width=340,height=226 ] the relative normalization of the @xmath24 @xmath4 excitation function was based on quasi - elastic proton - proton scattering .
the excitation function does not show a structure which could be interpreted as a resonance originating from the decay of an @xmath0-mesic @xmath1 . in the experiment , in november 2010
, two channels were measured : @xmath2-@xmath3 @xmath4 and @xmath2-@xmath3 @xmath33 @xmath34 .
the measurement was performed with a beam momentum ramping from 2.127gev / c to 2.422gev / c , corresponding to the range of the excess energy . for both reactions
the geometrical acceptance of the detector as a function of excess energy q was determined in simulations .
the acceptance is presented in fig .
[ reakcja_1 ] for different widths of the presumed bound state and using the av18 model description of the nucleon momentum distribution inside the @xmath35 nuclei .
the detailed description of the simulations is presented in ref .
@xcite .
-@xmath3 @xmath4 ( left ) and @xmath2-@xmath3 @xmath33 @xmath34 reaction ( right ) .
the acceptance is calculated for different widths of the presumed bound state and using the av18 potential model description of the nucleon momentum distribution inside @xmath1 .
[ reakcja_1],title="fig:",width=302,height=283 ] -@xmath3 @xmath4 ( left ) and @xmath2-@xmath3 @xmath33 @xmath34 reaction ( right ) .
the acceptance is calculated for different widths of the presumed bound state and using the av18 potential model description of the nucleon momentum distribution inside @xmath1 .
[ reakcja_1],title="fig:",width=302,height=283 ] the acceptance is almost a constant function of the excess energy and its average value is about 53% and 50% for and @xmath2-@xmath3 @xmath33 @xmath34 , respectively .
the high acceptance values allow high statistics measurements of these final states @xcite . during the experiment in 2010
, data were effectively taken for about 155 hours .
the average luminosity was estimated based on the trigger used for the elastic proton - proton scattering and is about l=8.15@xmath36 .
taking into account the fact that two reactions were measured , in total more than 40 times higher statistics were collected than in the experiment carried out in 2008 .
the search for @xmath1-@xmath0 bound states is performed with the wasa - at - cosy facility by taking advantage of its very high acceptance for the proton-@xmath26 and neutron-@xmath20 pairs as well as very good identification of helium nuclei .
in addition , wasa allows an exclusive measurement of the @xmath24 @xmath4 reaction .
+ at present the data analysis is in progress . in the optimistic case , the statistics could be sufficient to observe a signal from the @xmath0-mesic helium and in a pessimistic scenario the upper limit of the cross section for the @xmath1-@xmath0 bound state production will be decreased by a factor of about six .
the work is supported by the foundation for polish science - mpd program , co - financed by the european union within the european regional development fund , by the polish national science center through grants no .
0320/b / h03/2011/40 and 2011/01/b / st2/00431 , by the european commission under the 7th framework programme through the research infrastructures action of the capacities programme ( fp7-infrastructures-2008 - 1 , grant agreement n. 227431 ) and by the ffe grants of the research center jlich . + r.s .
bhalerao , l.c .
liu , q. haider , l.c .
liu , d. jido , h. nagahiro , s. hirenzaki , t. inoue , e. oset , s. d. bass , a. w. thomas , s. d. bass , a. w. thomas , _ acta phys .
_ b 41 ( 2010 ) q. haider , l. c. liu , c. wilkin , s. wycech , a. m. green and j. a. niskanen , a. m. green , h. machner , _ acta phys .
_ b 41 ( 2010 ) 2221 g. a. sokol , _ et al .
, arxiv : nucl - ex/0012010 _ ( 2000 ) a. gillitzer , _ acta .
slovaca _ 56 ( 2006 ) 269 j. berger _
et al . _ , b. mayer _ et al . _ , t. mersmann _ et al .
_ , j. smyrski _ et al .
_ , p. moskal , j. smyrski , _ acta phys .
_ b 41 ( 2010 ) 2281 j. smyrski _ et al .
, acta phys .
slovaca _ 56 ( 2006 )
213 w. krzemie _
et al . , int . j. mod
. phys . _ a 24 ( 2009 ) 576 j. smyrski _ et al .
_ , c. wilkin , m. pfeiffer _
et al . _ ,
b. krusche _
et al , _ _ acta phys .
_ b41 ( 2010 ) 2249 p. moskal _ et al .
phys . _ a 22 ( 2007 ) 305 w. krzemie , _ arxiv:1101.3103 _ ( 2011 ) g. a. sokol _
et al . , arxiv : nucl - ex/0106005 _ ( 2001 ) p. moskal , _
arxiv:0909.3979 _ ( 2009 ) w. krzemie , p. moskal , j. smyrski , _ acta phys .
_ 2 ( 2009 ) 141 w. krzemie , _ phd thesis , jagiellonian university _ ( 2011 ) m. skurzok , _ diploma thesis , jagiellonian university of cracow _ ( 2010 ) , berichte des fz - jlich , jl-4332 ( 2010 ) ,
_ arxiv:1009.5503 _ ( 2010 ) | the existence of @xmath0-mesic nuclei in which the @xmath0 meson is bound in a nucleus by means of the strong interaction was postulated already in 1986 , albeit not yet confirmed it by experiment .
the discovery of this new kind of an exotic nuclear matter would be very important as it might allow for a better understanding of the @xmath0 meson structure and its interaction with nucleons .
the search for @xmath0-mesic helium ( @xmath1-@xmath0 ) is carried out with high statistics and high acceptance with the wasa detector , installed at the cooler synchrotron cosy of the research center jlich .
the search is conducted via the measurement of the excitation function for selected decay channels of the @xmath1-@xmath0 system . in the experiment , performed in november 2010 , two reactions @xmath2-@xmath3 @xmath4 and @xmath2-@xmath3 @xmath5 were measured with a beam momentum ramped from 2.127gev / c to 2.422gev / c .
the report includes the description of the experimental method and status of the measurement . |
over a wide range of initial masses , stars appear to drive an outflow at some stage in their formation .
mounting evidence calls into question whether there is an upper limit on the mass of protostars that may drive collimated jets , despite uncertainty about whether the dominant physics of formation may change between low and high masses ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these jets will interact with and shape and be shaped by the circumstellar molecular envelope and larger natal cloud .
various jet properties and morphologies , such as asymmetries ( e.g. * ? ? ?
* ; * ? ? ?
* ) , wiggles and gaps ( e.g. * ? ? ?
* ) , the degree of collimation @xcite , and decreasing knot velocities have been explained by interaction with different kinds of environments .
interaction between the protostellar jet and ambient gas in the star forming cloud has been proposed as a mechanism to create molecular outflows .
@xcite offer a simple analytical model wherein molecular outflows are the product of environmental gas being entrained by a protostellar jet . while some molecular outflows are well - described by jet - entrainment ( see , e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , others show a conical morphology and broader range of velocities in the outflowing gas that have been interpreted as a wide - angle wind launched from the protostellar disc ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a full evolutionary scenario may involve a combination of these two , with different components dominating at different times ( e.g. * ? ? ?
* ; * ? ? ?
* ) . in either case , the outflow will clear material along the jet axis , creating a polar cavity as it travels through the protostellar envelope and into the surrounding molecular cloud .
clearing of the protostellar envelope may explain the apparent widening of molecular outflows associated with more evolved sources @xcite .
[ cols="<,<,<,<,<,<",options="header " , ] [ t : pm ] we compare new acs h@xmath0 images obtained in 2014 to the first epoch images obtained @xmath5 yrs earlier by @xcite and measure proper motions of jet features . in general , the measured velocities agree with those found by @xcite within the uncertainties ( see table 2 ) , although the better match of the instrument and filter used in this work permit higher precision than in our previous work . with the longer time baseline and better match of the observational setup , we can also measure proper motions of slower features and off - axis motions with greater fidelity .
the most remarkable motions we uncover with the latest epoch shows clear _ lateral _ expansion of the cocoon seen in h@xmath0 emission ( see figure [ fig : pm ] ) . in particular , we measure a transverse velocity for feature m12 of 64 km s@xmath3 and it appears to move at an angle @xmath6 away from the jet axis defined by the motion of features m8 , m5a , m10 .
similar lateral expansion in features o6
, o7 from hh 666 o traces the slow expansion of the h@xmath0 sheath surrounding the [ fe ii ] jet .
a slower velocity and an off - axis propagation direction are predicted for the expanding wings of a bow shock ( e.g. * ? ? ?
@xmath7{hh666_acs_diff_im.eps } \\ \begin{array}{cc } \includegraphics[angle=0,scale=0.335]{hh666_acs_pm_o.eps } & \includegraphics[angle=0,scale=0.35]{hh666_acs_pm_m.eps } \\ \includegraphics[angle=0,scale=0.335]{hh666_acs_pm_o_vectors.eps } & \includegraphics[angle=0,scale=0.35]{hh666_acs_pm_m_vectors.eps } \\ \end{array } \end{array}$ ]
@xmath8{hh666_unsm_color.eps } \\ \includegraphics[trim=50 mm 50 mm 50 mm 45mm , angle=90,scale=0.725]{hh666_big_cartoon.ps } \\ \end{array}$ ] given the striking disparity in the morphology and kinematics traced by h@xmath0 and [ fe ii ] emission in both images and spectra , we propose that these two different emission lines trace physically distinct components of the hh 666 outflow .
we illustrate the anatomy of a dense , externally irradiated protostellar outflow in figure [ fig : cartoon ] , and discuss the evidence for this picture below . in hh 666
o , the redshifted part of the outflow to the southeast of the protostar , h@xmath0 emission appears to form a cocoon that surrounds the [ fe ii ] jet core .
the large spatial offset of the h@xmath0 emission from the [ fe ii ] ( @xmath9 ) suggests that it traces an entirely different component of the outflow .
in fact , the thin arcuate morphology of the h@xmath0 and [ s ii ] emission along the face of the pillar resembles the molecular outflow lobes seen at longer ( mm ) wavelengths ( e.g. hh 211 , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
prompt entrainment of ambient material via bow shocks from episodic jets has been invoked to explain nested outflow shells ( e.g. * ? ? ?
jets in carina that are driven by protostars embedded in a dust pillar ( more than half the jets discovered by * ? ? ?
* ) will entrain material as they exit the cloud .
kinematics in the h@xmath0 spectrum also support the interpretation that h@xmath0 emission traces an externally irradiated outflow shell in hh 666 .
fast , ionized material at the head of the bow shock that creates this outflow shell corresponds to the highest velocities seen in the h@xmath0 spectrum .
velocities observed closer to the driving source are lower because they are dominated by emission from the slower wings of the bow shock .
the slowest velocities will be observed farthest from the apex of the bow shock , and therefore closest to the driving source ( see , e.g. figure 1 in * ? ? ?
* ) . in this picture
, the bow shock dominates entrainment of the ambient gas and will create the observed hubble - like velocity structure , as seen in the position - velocity diagrams of molecular outflows ( e.g. hh 300 , * ? ? ?
in contrast to the h@xmath0 emission , doppler velocities in the collimated [ fe ii ] jet are remarkably constant throughout hh 666 o , remaining close to @xmath10 km s@xmath3 , as expected for an unimpeded jet propagating away from its driving source .
we do not detect evidence for a hubble - flow in the [ fe ii ] emission from the jet .
if environmental interaction is dominated by the first impulse from the jet , portions of the jet traced by [ fe ii ] that travel in the wake of the initial bow shock will transfer little momentum to the environment . some optical emission from the jet body in the outer portions of hh 666
o , near the leading bow shock ( feature o3 in figure [ fig : pm ] , also see figure [ fig : hst_ims ] ) further supports the idea that there are two externally irradiated outflow components visible in hh 666 .
h@xmath0 , [ s ii ] , and [ fe ii ] emission begin to coincide just beyond the edge of the dust pillar , marking the first place where the newly liberated _ jet _ is laid bare to ionizing radiation .
optical emission from the _ jet _ beyond the pillar edge suggests a decrease in the intervening optical depth , and the increase of the [ fe ii ] ratio @xmath11 toward the terminus of hh 666 o supports this interpretation .
however , the ratio only approaches the intrinsic value beyond the end of the jet , suggesting that the observed reddening is primarily from the jet itself ( see figure [ fig : feii_ratio ] ) .
the blueshifted part of the outflow , hh 666 m , that extends to the northwest of the protostar reveals a similar picture to hh 666 o. close to the driving source ( within @xmath12 ) , the behavior of the h@xmath0 and [ fe ii ] emission in hh 666 m resembles that of hh 666 o. a narrow column of knotty [ fe ii ] emission bisects broad , arc - like h@xmath0 emission before the morphologies traced by the two lines converge at larger distances from the protostar .
new h@xmath0 images show that the head of the h@xmath0 arc moves along the jet axis while its wings expand laterally away from it .
features m8 , m5a , m10 ( see figure [ fig : pm ] ) move primarily along the jet axis with a transverse velocity of @xmath13 km s@xmath3 .
feature m12 travels at an @xmath6 angle away from the jet axis , indicating that this emission feature traces the expansion of the h@xmath0 arc away from the jet axis .
expanding h@xmath0 emission suggests that the arc may be the bow shock where hh 666 m breaks out of the front side of the parent cloud ( approximately at the location of feature m9 ) . in this picture , the ` dark patch ' to the northwest of the driving source is a cavity cleared by the jet as it plows out of the molecular cloud , leaving a deficit of material behind the bow shock . @xcite
also suggested that a cant nozzle @xcite may be responsible for refocusing the jet at this point .
spectra support the interpretation that the h@xmath0 arc traces the tip of the expanding outflow cavity .
we observe a hubble wedge in hh 666 m , with h@xmath0 emission tracing increasing velocities up until a sharp drop corresponding to features m5a , m10 , m12 ( see figure [ fig : spectra ] ) . inside feature m9 ,
closer to the driving source , there is a clear deficit of emission from the globule .
however , at the same place along the slit where globule emission decreases , we see blueshifted emission that extends smoothly toward feature m9 .
thus , it appears that h@xmath0 emission from the inner portion of hh 666 m also traces the expansion of the outflow shell .
a hubble - like velocity pattern in hh 666 m further suggests that the dark patch to the northwest of the driving source is an evacuated cavity in the dust pillar cleared by the jet .
if this is correct , hh 666 o should create a corresponding cavity to the southeast of the driving source as it propagates deeper into the cloud . to test this , we took an intensity tracing through the pillar along the axis of the jet to search for a decrease in the ambient h@xmath0 emission on either side of the driving source ( see figure [ fig : cavity_tracings ] ) . indeed , intensity tracings reveal dips in the h@xmath0 intensity along the pillar extending @xmath14 on either side of the driving source .
the emission trough is shallower to the southeast of the driving source , as expected for more foreground emission from the pillar itself .
@xmath15{hh666_halpha_tracing_area.eps } \\
\includegraphics[angle=90,scale=0.315]{hh666_cavity_tracing.eps } \\
\end{array}$ ] as in hh 666 o , doppler velocities measured in the [ fe ii ] spectra of hh 666 m remain steady , with velocities close to @xmath16 km s@xmath3 throughout .
fast , oppositely directed [ fe ii ] emission is what we expect from a spatially resolved steady _
jet_. this structure has been seen in other jets ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , including those driven by intermediate - mass protostars ( e.g. * ? ? ?
* ; * ? ? ?
hubble - like velocities have been observed in protostellar _ outflows _ , but usually in ir / mm tracers of molecular gas that has been accelerated by an underlying jet ( e.g. * ? ? ?
only two other jets show hubble - like velocities in h@xmath0 emission
hh 83 @xcite and hh 900 @xcite . beyond the h@xmath0 arc ( west of feature m10 , outside the globule ) , h@xmath0 and [ fe ii ] emission converges , thereafter tracing the same morphology .
the [ fe ii ] ratio remains relatively constant ( see figure [ fig : feii_ratio ] ) , and velocities traced by both lines remain relatively steady ( no strong shocks ) .
doppler velocities in [ fe ii ] spectra , however , are nearly 100 km s@xmath3 faster than those traced by h@xmath0 . low- and high - velocity components have been observed in jets from low - mass t tauri stars with velocity differences between the core of the jet and the slightly off - axis wings of as much as @xmath17 km s@xmath3 in some cases ( e.g. dg tau , * ? ? ?
* ; * ? ? ?
if the shell of material that is traced by h@xmath0 inside the globule is pushed back against the jet by feedback from massive stars in the region , then the two components may overlap spatially in images while the entrained material traced by h@xmath0 has decelerated as compared to the velocity of the [ fe ii ] jet . for embedded jets that entrain an outflow as they plow through a dusty pillar
, uv radiation from nearby massive stars will illuminate the shell of the outflow in addition to the jet .
we argue that h@xmath0 proper motions and doppler velocities trace the expansion and propagation of the externally irradiated outflow lobe entrained by the jet , but not the jet core . in this picture
, we expect larger proper motions from the jet itself ( @xmath18 mas yr@xmath3 for a velocity of 200 km s@xmath3 ) , although additional [ fe ii ] imaging at a later epoch is required to test this hypothesis .
our data show that h@xmath0 emission traces entrained outflowing gas with the same hubble - like velocity structure seen in molecular outflows associated with other hh jets ( e.g. * ? ? ?
however , the extreme uv radiation in the carina nebula illuminates entrained gas in hh 666 , leading to h@xmath0 emission from the outflow cocoon .
no h@xmath19 emission that is unambiguously associated with the jet has been observed in hh 666 @xcite , although weak h@xmath19 emission will be difficult to discern against the bright background of the photodissociation region ( pdr ) .
both morphology and kinematics reveal an h@xmath0 outflow lobe encasing the [ fe ii ] jet to the east and west of the hh 666 driving source .
proper motions reveal the dynamical age of these h@xmath0 outflow lobes to be @xmath20 yrs ( see table 2 ) .
this points to a sudden increase in the jet mass - loss rate @xmath4 yrs ago that provided a `` kick '' to the ambient material now seen as a hubble flow in h@xmath0 emission .
several more distant knots in hh 666 indicate that the jet had other outbursts like this in the past @xcite .
steady [ fe ii ] emission inside the h@xmath0 cocoons on either side of the jet indicates that the high - mass - loss rate episode has persisted , with a duration comparable to its age .
a burst duration of @xmath4 yrs is an order of magnitude larger than the estimated decay time of fu orionis outbursts ( e.g. * ? ? ?
this long duration is on the order of the time between ejections found in hh jets driven by low - mass stars ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and similar to the variability time - scale estimated for hh 46/47 @xcite .
while the burst duration is long compared to those inferred for low - mass protostars , the @xmath4 yr dynamical time of the inner jet is only a small fraction of the overall age of the hh 666 outflow ( @xmath21 yr , see * ? ? ?
this points to a strong , sustained increase in the accretion rate from a relatively evolved intermediate - mass protostar . a larger sample of intermediate - mass protostars is required to determine the frequency and duty cycle of such accretion bursts , and whether they remain stronger than their low - mass counterparts throughout their evolution .
[ fe ii ] emission from hh 666 traces a steady jet with higher densities and velocities than traced by h@xmath0 emission .
@xcite found that the mass - loss rate implied by the survival of fe@xmath22 in the jet is an order of magnitude higher than the mass - loss rate derived from the h@xmath0 emission measure @xcite .
new [ fe ii ] spectra presented here reveal velocities @xmath2 km s@xmath3 throughout the inner jet , similar to the fastest velocities in the h@xmath0 spectrum . applying the median tilt angle of the jet derived from the h@xmath0 velocities , @xmath23
( see table 2 ) , implies that the three - dimensional velocity of the [ fe ii ] jet is @xmath24 km s@xmath3 , almost twice the speed of the h@xmath0 jet . the higher mass - loss rate and velocities in the [ fe ii ] jet suggest that this component of the outflow has at least an order of magnitude more momentum .
however , the steady velocity of the [ fe ii ] jet traveling in the wake of the initial bow shock suggests little momentum transfer to the local environment .
this means that the high momentum of the [ fe ii ] jet will not be deposited locally in the molecular globule , but rather injected into the h ii region where its contribution to driving turbulence in the region will be dwarfed by feedback from massive stars .
in this paper , we propose a solution to the discrepancy observed in the morphology and kinematics of h@xmath0 and [ fe ii ] emission from hh 666 , an externally irradiated protostellar outflow in the carina nebula .
previous high - resolution imaging revealed different morphologies traced by the two lines .
we present new , near - ir spectra of the inner jet demonstrating that [ fe ii ] emission also traces different kinematics .
doppler velocities in the new [ fe ii ] spectra of the inner jet , features hh 666 m and o , remain remarkably consistent on either side of the jet , tracing velocities of @xmath2 km s@xmath3 throughout .
this stands in stark contrast to the velocity structure seen from the same portions of the jet in the longslit h@xmath0 spectrum from @xcite .
both hh 666 m and o show hubble - like velocities in the h@xmath0 spectrum with only the fastest doppler velocities in the hubble wedges on either side of the driving source approaching the velocities traced by [ fe ii ] .
h@xmath0 and [ fe ii ] emission also trace different morphologies , complementing the velocity discrepancies in spectra @xcite .
sheath - like h@xmath0 emission encompasses narrow [ fe ii ] emission on either side of the jet .
we argue that broad h@xmath0 emission traces the material swept up via prompt entrainment by a jet outburst that began @xmath4 yrs ago , with bright h@xmath0 arcs tracing the head of the bow shock that dominates entrainment .
constant velocity [ fe ii ] emission , inside the outflow sheath and hubble flow traced by h@xmath0 , indicates that the high - mass - loss rate episode from the jet is ongoing , tracing a strong accretion burst that has been sustained longer than the typical decay time of fu orionis outbursts . a new epoch of h@xmath0 imaging from _
hst_/acs provides a @xmath5 yr time baseline to measure the motion of bright h@xmath0 features .
proper motions trace the lateral expansion of the h@xmath0 sheath , consistent with an expanding shell around the jet .
we propose that the same underlying physics that drives the molecular outflows seen from more embedded intermediate - mass protostars creates the two - component outflow we observe in hh 666 . unlike jets with molecular outflows ,
hh 666 resides in a giant h ii region , so both the jet and the material it entrains are subject to the harsh uv background created by the many o - type stars in the carina nebula .
this leads to h@xmath0 emission from the walls of the outflow cavity and [ fe ii ] emission from the obscured jet .
outside the parent pillar where a cocoon no longer envelopes the jet , h@xmath0 and [ fe ii ] emission trace the same morphology .
thus , hh 666 provides a missing link that unifies the irradiated outflows driven by unobscured protostars with the molecular outflows typically observed from embedded intermediate - mass protostars .
we would like to thank rob simcoe for his assistance with reduction of the fire data .
support for this work was provided by nasa grant ar-12155 , go-13390 , and go-13391 from the space telescope science institute .
this work is based on observations made with the nasa / esa hubble space telescope , obtained from the data archive at the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
these _ hst _ observations are associated with programs go 10241 , 10475 , 13390 , and 13391 .
gemini observations are from the gs-2013a - q-12 science program . anderson j. , 2006 , `` the 2005 hst calibration workshop : hubble after the transition to two - gyro mode '' , proceedings of a workshop held at the space telescope science institute , baltimore , maryland , october 26 - 28 , 2005 .
eds koekemoer a.m. , goudfrooij p. , dressel l.l .
greenbelt , md , 11 anderson j. , king i.r . , 2006 ,
`` psfs , photometry , and astronomy for the acs / wfc '' , 2006 , 1 anderson j. , et al . , 2008a , aj , 135 , 2055 anderson j. , et al .
, 2008b , aj , 135,2114 anderson j. , van der marel r.p . , 2010 , apj , 710,1032 arce h.g . ,
goodman a.a . , 2001 , apj , 554 , 132 arce h.g . , goodman a.a . , 2001 , apjl , 551 , 171 arce h.g .
, goodman a.a .
, 2002 , apj , 575 , 928 arce h.g .
, sargent a.i . , 2005 ,
apj , 624 , 232 arce h.g . ,
sargent a.i . , 2006 ,
apj , 646 , 1070 arce h.g . , mardones d. , corder s.a . , garay g. , noriega - crespo a. , raga a.c . , 2013 , apj , 774 , 39 bacciotti f. , mundt r. , ray t.p . , eislffel j. , solf j. , camezind m. , 2000 , apj , 537 , 49 bally j. , reipurth b. , 2001 , apj , 546 , 299 bally j. , licht d. , smith n. , walawender j. , 2006 , aj , 131 , 473 beltrn m.t . , estalella r. , girart j.m .
, ho p.t.p .
, anglada g. , 2008 , a&a , 481 , 93 beuther h. , walsh a.j . , thorwirth s. , zhang q. , hunter t.r . ,
megeath s.t . ,
menten k.m . , 2008 , a&a , 481 , 169 cant j. , rodriguez l.f .
, 1980 , apj , 239 , 982 cant j. , rodriguez l.f . , barral j.f . , carral p. , 1981 ,
apj , 244 , 102 caratti o garatti a. , froebrich d. , eislffel j. , giannini t. , nisini b. , 2008 , a&a , 485 , 137 currie d.g .
et al . , 1996 ,
aj , 112 , 1115 de gouveia e.m . , birkinshaw m. , benz w. , 1996 , apjl , 460 , 111 dionatos o. , nisini b. , cabrit s. , kristensen l. , pineau des forts g. , 2010 , a&a , 521 , 7 duarte - cabral a. , bontemps s. , motte f. , hennemann m. , schneider n. , andr p. , 2013
, a&a , 558 , 125 ellerbroek l.e . , podio l. , kaper l. , sana h. , huppenkothen d. , de koter a. , monaco l. , 2013 , a&a , 551 , 5 ellerbroek l.e . , et al .
, 2014 , a&a , 563 , 87 forbrich j. , stanke th . , klein r. , henning th . , menten k.m . , schreyer k. , posselt b. , 2009 , a&a , 493 , 547 fuente a. et al . , 2009 , a&a , 507 , 1475 garay g. , brooks k.j . , mardones d. , norris r.p . , 2003 , apj , 587 , 739 garcia - lopez r. , nisini b. , giannini t. , eislffel j. , bacciotti f. , podio l. , 2008 , a&a , 487 , 1019 giannini , t. , et al . , 2015 , apj , 798 , 33 gueth f. , guilloteau s. , 1999 , a&a , 343 , 571 guzmn a.e . , garay g. , brooks k.j .
, voronkov m.a . , 2012 , apj , 753 , 51 hartigan p. , raymond j. , meaburn j. , 1990 , apj , 362 , 624 hartigan p. , morse j.a .
, raymond j. , 1994 , apj , 436 , 125 hartigan p. , morse j.a .
, reipurth b. , heathcote s. , bally j. , 2001 , apjl , 559 , 157 hartigan p. , et al . , 2011 , apj , 736 , 29 hartigan p. , reiter m. , smith n. , bally j. , 2015 , aj , 149 , 101 hartmann l. , kenyon , s.j . , 1996 , ara&a , 34 , 207 heathcote s. , morse j.a . , hartigan p. , reipurth b. , schwartz r.d .
, bally j. , stone j.m .
, 1996 , aj , 112 , 1141 klaassen p.d . , et al .
2013 , a&a , 555 , 73 lee c .- f . ,
stone j.m . ,
ostriker e.c . ,
mundy l.g .
, 2001 , apj , 557 , 429 lefloch b. , cernicharo j. , reipurth b. , pardo j.r . , neri r. , 2007 , apj , 658 , 498 liu c .- f .
, et al . 2012 , apj , 749 , 62 mcgroarty f. , ray t.p . , bally j. , 2004 , a&a , 415 , 189 melnikov s.y . , eislffel j. , bacciotti f. , woitas j. , ray t.p . , 2009 , a&a , 506 , 763 morse j.a . , kellogg j.r . , bally j. , davidson k. , balick b. , ebbets d. , 2001 , apjl , 548 , 207 mundt r. , ray t.p .
, raga a.c .
, 1991 , a&a , 252 , 740 nagar n.m . , vogel s.n . , stone j.m . , ostriker e.c . , 1997 ,
apj , 482 , 195 narayanan g. , walker c.k . , 1996 , apj , 466 , 844 noriega - crespo a. , et al . , 2004 ,
apjs , 154,352 ostriker e.c .
, lee c .- f . ,
stone j.m . ,
mundy l.g . , 2001 ,
apj , 557 , 443 povich m. et al . , 2011 ,
apjs , 194 , 14 pyo t .- s . , et al .
, 2003 , apj , 590 , 340 raga a.c
. , noriega - crespo a. , 1992 , rev .
astrofis . , 24 , 9 raga a.c . , cabrit s. , 1993 , a&a , 278 , 267 reipurth b. , 1989 , a&a , 220 , 249 reipurth b. , raga a.c . , heathcote s. , 1992 , apj , 392 , 145 reipurth b. , hartigan p. , heathcote s. , morse j.a .
, bally j. , 1997 , aj , 114 , 757 reipurth b. , bally j. , fesen r.a . , devine d. , 1998 , nature , 396 , 343 reipurth b. , yu k.c . , heathcote s. , bally j. , rodrguez l.f .
, 2000 , aj , 120 , 1449 reiter m. , smith , n. , 2013 , mnras , 433 , 2226 reiter m. , smith n. , 2014 , mnras , 445 , 3939 reiter m. , smith n. , kiminki m.m . ,
bally j. , anderson j. 2015 , mnras , accepted salyk c. , pontoppidan k. , corder s. , muoz d. , zhang k. , blake g.a . , 2014 , apj , 792 , 68 seplveda i. , anglada g. , estalella r. , lpez r. , girart j.m .
, yang j. , 2011 , a&a , 527 , 41 shang h. , allen a. , li z .- y . , liu c .- f . , chou m .- y . , anderson j. , 2006 , apj , 649 , 845 simcoe r.a .
et al . , 2013 , pasp , 125 , 270 smith n. , bally j. , brooks k.j . , 2004 ,
aj , 127 , 2793 smith n. , 2006a , mnras , 367 , 763 smith n. , 2006b , apj , 644 , 1151 smith n. , hartigan p. , 2006
, apj , 638 , 1045 smith n. , bally j. walborn n. , 2010 , mnras , 405 , 1153 sohn s.t . , anderson j. , van der marel r.p . , 2012 ,
apj , 753 , 7 takahashi s. , ho p.t.p . , 2012 , apj , 745 , 10 tappe a. , forbrich j. , martn s. , yuan y. , lada c.j . , 2012 ,
apj , 751 , 9 zapata l.a . ,
, brassfield e. , palau a. , patel n. , pineda j.e . , 2014 , mnras , 441 , 3696 zinnecker h. , mccaughrean m.j .
, rayner j.t . , 1998 , nature , 394 , 862 | hh 666 is an externally irradiated protostellar outflow in the carina nebula for which we present new near - ir [ fe ii ] spectra obtained with the fire spectrograph at magellan observatory .
earlier h@xmath0 and near - ir [ fe ii ] imaging revealed that the two emission lines trace substantially different morphologies in the inner @xmath1 of the outflow .
h@xmath0 traces a broad cocoon that surrounds the collimated [ fe ii ] jet that extends throughout the parent dust pillar .
new spectra show that this discrepancy extends to their kinematics .
near - ir [ fe ii ] emission traces steady , fast velocities of @xmath2 km s@xmath3 from the eastern and western limbs of the jet .
we compare this to a previously published h@xmath0 spectrum that reveals a hubble - flow velocity structure near the jet - driving source .
new , second - epoch _
h@xmath0 images reveal the _ lateral _ spreading of the h@xmath0 outflow lobe away from the jet axis .
proper motions also indicate a sudden increase in the mass - loss rate @xmath4 yr ago , while steady [ fe ii ] emission throughout the inner jet suggest that the burst is ongoing .
an accretion burst sustained for @xmath4 yr is an order of magnitude longer than expected for fu orionis outbursts , but represents only a small fraction of the total age of the hh 666 outflow .
altogether , available data suggests that [ fe ii ] traces the highly collimated protostellar _ jet _ while h@xmath0 traces the entrained and irradiated _
outflow_. hh 666 appears to be a missing link between bare jets seen in h ii regions and entrained molecular outflows seen from embedded protostars in more quiescent regions .
[ firstpage ] stars : formation jets outflows |
the recent massive work in the area of compressed sensing , surveyed in @xcite , rigorously demonstrated that one can algorithmically recover sparse ( and , more generally , compressible ) signals from incomplete observations .
the simplest model is a @xmath11-dimensional signal @xmath6 with a small number of nonzeros : @xmath12 such signals are called @xmath9-sparse .
we collect @xmath13 nonadaptive linear measurements of @xmath6 , given as @xmath14 where @xmath2 is some @xmath15 by @xmath11 measurement matrix .
we then wish to efficiently recover the signal @xmath6 from its measurements @xmath8 .
a necessary and sufficient condition for exact recovery is that the map @xmath2 be one - to - one on the set of @xmath9-sparse vectors .
cands and tao @xcite proved that under a stronger ( quantitative ) condition , the sparse recovery problem is equivalent to a convex program @xmath16 and therefore is computationally tractable .
this condition is that the map @xmath2 is an almost isometry on the set of @xmath7-sparse vectors : a measurement matrix @xmath2 satisfies the _ restricted isometry condition _ ( ric ) with parameters @xmath17 for @xmath18 if we have @xmath19 under the restricted isometry condition with parameters @xmath20 , the convex program exactly recovers an @xmath9-sparse signal @xmath6 from its measurements @xmath8 @xcite .
the restricted isometry condition can be viewed as an abstract form of the uniform uncertainty principle of harmonic analysis ( @xcite , see also @xcite and @xcite ) .
many natural ensembles of random matrices , such as partial fourier , bernoulli and gaussian , satisfy the restricted isometry condition with parameters @xmath21 , @xmath22 provided that @xmath23 see e.g. section 2 of @xcite and the references therein . therefore , a computationally tractable exact recovery of sparse signals is possible with the number of measurements @xmath15 roughly proportional to the sparsity level @xmath9 , which is usually much smaller than the dimension @xmath11 .
an important alternative to convex programming is greedy algorithms , which have roots in approximation theory .
a greedy algorithm computes the support of @xmath6 iteratively , at each step finding one or more new elements ( based on some `` greedy '' rule ) and subtracting their contribution from the measurement vector @xmath8 .
the greedy rules vary .
the simplest rule is to pick a coordinate of @xmath24 of the biggest magnitude ; this defines the well known greedy algorithm called orthogonal matching pursuit ( omp ) , known otherwise as orthogonal greedy algorithm ( oga ) @xcite .
greedy methods are usually fast and easy to implement , which makes them popular with practitioners .
for example , omp needs just @xmath9 iterations to find the support of an @xmath9-sparse signal @xmath6 , and each iteration amounts to solving one least - squares problem ; so its running time is always polynomial in @xmath9 , @xmath15 and @xmath11 .
in contrast , no known bounds are known on the running time of as a linear program . future work on customization of convex programming solvers for sparse recovery problems may change this picture , of course . for more discussion ,
see @xcite and @xcite .
a variant of omp was recently found in @xcite that has guarantees essentially as strong as those of convex programming methods .
this greedy algorithm is called regularized orthogonal matching pursuit ( romp ) ; we state it in section [ s : stable ] below . under the restricted isometry condition with parameters @xmath25
, romp exactly recovers an @xmath9-sparse signal @xmath6 from its measurements @xmath8 . summarizing ,
_ the uniform uncertainty principle is a guarantee for efficient sparse recovery ; one can provably use either convex programming methods or greedy algorithms ( romp ) . _
a more realistic scenario is where the measurements are inaccurate ( e.g. contaminated by noise ) and the signals are not exactly sparse . in most situations that arise in practice , one can not hope to know the measurement vector @xmath14 with arbitrary precision .
instead , it is perturbed by a small error vector : @xmath1 . here
the vector @xmath5 has unknown coordinates as well as unknown magnitude , and it needs not be sparse ( as all coordinates may be affected by the noise ) . for a recovery algorithm to be stable ,
it should be able to approximately recover the original signal @xmath6 from these perturbed measurements .
the stability of convex optimization algorithms for sparse recovery was studied in @xcite , @xcite , @xcite , @xcite .
assuming that one knows a bound on the magnitude of the error , @xmath26 , it was shown in @xcite that the solution @xmath27 of the convex program @xmath28 is a good approximation to the unknown signal : @xmath29 .
in contrast , the stability of greedy algorithms for sparse recovery has not been well understood .
numerical evidence @xcite suggests that omp should be less stable than the convex program , but no theoretical results have been known in either the positive or negative direction .
the present paper seeks to remedy this situation .
we prove that _ romp is as stable as the convex program .
_ this result essentially closes a gap between convex programming and greedy approaches to sparse recovery . regularized orthogonal matching pursuit ( romp ) [ t : stability ] assume a measurement matrix @xmath2 satisfies the restricted isometry condition with parameters @xmath30 for @xmath31 .
let @xmath6 be an @xmath9-sparse vector in @xmath32 .
suppose that the measurement vector @xmath33 becomes corrupted , so we consider @xmath1 where @xmath5 is some error vector .
then romp produces a good approximation to @xmath6 : @xmath34 note that in the noiseless situation ( @xmath35 ) the reconstruction is exact : @xmath36 .
this case of theorem [ t : stability ] was proved in @xcite .
our stability result extends naturally to the even more realistic scenario where the signals are only approximately sparse . here and henceforth ,
denote by @xmath37 the vector of the @xmath38 biggest coefficients in absolute value of @xmath39 .
[ t : stabsig ] assume a measurement matrix @xmath2 satisfies the restricted isometry condition with parameters @xmath40 for @xmath31 .
consider an arbitrary vector @xmath6 in @xmath32 .
suppose that the measurement vector @xmath33 becomes corrupted , so we consider @xmath1 where @xmath5 is some error vector .
then romp produces a good approximation to @xmath41 : @xmath42 * 1 .
* the term @xmath41 in the corollary can be replaced by @xmath43 for any @xmath44 .
this change will only affect the constant terms in the corollary .
* 2 . * by applying corollary [ t : stabsig ] to the largest @xmath45 coordinates of @xmath6 and using lemma [ l : ve ] below , we also have the error bound for the entire vector @xmath6 : @xmath46 * 3 .
* for the convex programming method , the stability bound was proved in @xcite , and even without the logarithmic factor .
we conjecture that this factor is also not needed in our results for romp .
* 4 . * unlike the convex program , romp succeeds with absolutely no prior knowledge about the error @xmath5 ; its magnitude can be arbitrary . in the terminology of @xcite , the convex programming approach needs to be `` noise - aware '' while romp needs not .
* one can use romp to approximately compute a @xmath45-sparse vector that is close to _ the best @xmath45-term approximation _
@xmath41 of an arbitrary signal @xmath6 . to this end ,
one just needs to retain the @xmath45 biggest coordinates of @xmath27 .
indeed , corollary [ c : napprox ] below shows that the best @xmath45-term approximations of the original and the reconstructed signals are close : @xmath47 * 6 . *
an important special case of corollary [ t : stabsig ] is for the class of compressible vectors , which is a common model in signal processing , see @xcite , @xcite .
suppose @xmath6 is a compressible vector in the sense that its coefficients obey a power law : for some @xmath48 , the @xmath49-th largest coefficient in magnitude of @xmath6 is bounded by @xmath50 .
then yields the following bound on the reconstructed signal : @xmath51 as observed in @xcite , this bound is optimal ( within the logarithmic factor ) ; no algorithm can perform fundamentally better . the rest of the paper
is organized as follows . in section [
s : proof ] , we prove our main result , theorem [ t : stability ] . in section [
s : consequences ] , we deduce the extension for approximately sparse signals , corollary [ t : stabsig ] , and a consequence for best @xmath9-term approximations , corollary [ c : napprox ] . in section [ s : implementation ] , we demonstrate some numerical experiments that illustrate the stability of romp .
we shall prove a stronger version of theorem [ t : stability ] , which states that _ at every iteration _ of romp , either at least @xmath52 of the newly selected coordinates are from the support of the signal @xmath6 , or the error bound already holds .
[ t : it ] assume @xmath2 satisfies the restricted isometry condition with parameters @xmath30 for @xmath31 .
let @xmath53 be an @xmath9-sparse vector with measurements @xmath1 . then at any iteration of romp , after the regularization step where @xmath54 is the current chosen index set
, we have @xmath55 and ( at least ) one of the following : 1 .
[ j support ] @xmath56 ; 2 .
[ error ] @xmath57 .
in other words , either at least @xmath52 of the coordinates in the newly selected set @xmath58 belong to the support of @xmath6 or the bound on the error already holds .
we show that the iteration invariant implies theorem [ t : stability ] by examining the three possible cases : * case 1 : ( ii ) occurs at some iteration .
* we first note that since @xmath59 is nondecreasing , if ( ii ) occurs at some iteration , then it holds for all subsequent iterations . to show that this would then imply theorem [ t : stability ]
, we observe that by the restricted isometry condition and since @xmath60 , @xmath61 then again by the restricted isometry condition and definition of @xmath27 , @xmath62 thus we have that @xmath63 thus ( ii ) of the iteration invariant would imply theorem [ t : stability ] .
* case 2 : ( i ) occurs at every iteration and @xmath58 is always non - empty . * in this case , by ( i ) and the fact that @xmath58 is always non - empty , the algorithm identifies at least one element of the support in every iteration .
thus if the algorithm runs @xmath9 iterations or until @xmath64 , it must be that @xmath65 , meaning that @xmath66 .
then by the argument above for case 1 , this implies theorem [ t : stability ] .
* case 3 : ( i ) occurs at each iteration and @xmath67 for some iteration .
* by the definition of @xmath58 , if @xmath67 then @xmath68 for that iteration . by definition of @xmath69
, this must mean that @xmath70 this combined with part 1 of proposition [ p : cons ] below ( and its proof , see @xcite ) applied with the set @xmath71 yields @xmath72 then combinining this with part 2 of the same proposition , we have @xmath73 since @xmath74 , this means that the error bound ( ii ) must hold , so by case 1 this implies theorem [ t : stability ] .
we now turn to the proof of the iteration invariant , theorem [ t : it ] .
we will use the following proposition from @xcite .
[ p : cons ] assume a measurement matrix @xmath2 satisfies the restricted isometry condition with parameters @xmath75
. then the following holds . 1 .
_ ( local approximation ) _ for every @xmath9-sparse vector @xmath0 and every set @xmath76 , @xmath77 , the observation vector @xmath78 satisfies @xmath79 2 .
_ ( spectral norm ) _ for any vector @xmath80 and every set @xmath76 , @xmath81 , we have @xmath82 3 . _
( almost orthogonality of columns ) _ consider two disjoint sets @xmath83 , @xmath84 .
let @xmath85 denote the orthogonal projections in @xmath86 onto @xmath87 and @xmath88 , respectively . then @xmath89 the proof of theorem [ t : it ] is by induction on the iteration of romp .
the induction claim is that for all previous iterations , the set of newly chosen indices is disjoint from the set of previously chosen indices @xmath54 , and either ( i ) or ( ii ) holds . clearly if ( ii ) held in a previous iteration , it would hold in all future iterations .
thus we may assume that ( ii ) has not yet held .
since ( i ) has held at each previous iteration , we must have @xmath90 let @xmath91 be the residual at the start of this iteration , and let @xmath58 , @xmath92 be the sets found by romp in this iteration . as in @xcite
, we consider the subspace @xmath93 and its complementary subspaces @xmath94 the restricted isometry condition in the form of part 3 of proposition [ p : cons ] ensures that @xmath95 and @xmath96 are almost orthogonal . thus @xmath96 is close to the orthogonal complement of @xmath95 in @xmath97 , @xmath98 the residual @xmath69 thus still has a simple description : [ residual ] here and thereafter , let @xmath99 denote the orthogonal projection in @xmath86 onto a linear subspace @xmath100 . then @xmath101 by definition of the residual in the algorithm , @xmath102 . to complete the proof
we need that @xmath103 .
this follows from the orthogonal decomposition @xmath104 and the fact that @xmath105 .
now we consider the signal we seek to identify at the current iteration , and its measurements : @xmath106 to guarantee a correct identification of @xmath107 , we first state two approximation lemmas that reflect in two different ways the fact that subspaces @xmath96 and @xmath108 are close to each other .
[ c : proj ] we have @xmath109 by definition of @xmath95 , we have @xmath110 .
therefore , by lemma [ residual ] , @xmath111 , and so @xmath112 now we use part 3 of proposition [ p : cons ] for the sets @xmath54 and @xmath113 whose union has cardinality at most @xmath114 by .
it follows that @xmath115 as desired .
[ l : uj ] consider the observation vectors @xmath116 and @xmath117 . then for any set @xmath118 with @xmath119 , @xmath120 since @xmath121
, we have by lemma [ c : proj ] and the restricted isometry condition that @xmath122 to complete the proof , it remains to apply part 2 of proposition [ p : cons ] , which yields @xmath123 .
we next show that the energy ( norm ) of @xmath124 when restricted to @xmath92 , and furthermore to @xmath58 , is not too small . by the regularization step of romp , since all selected coefficients have comparable magnitudes , we will conclude that not only a portion of energy but also of the _ support _ is selected correctly , or else the error bound must already be attained .
this will be the desired conclusion .
[ c : uj ] we have @xmath125 .
let @xmath126 = @xmath113 .
since @xmath127 , the maximality property of @xmath92 in the algorithm implies that @xmath128 by lemma [ l : uj ] , @xmath129 furthermore , since @xmath130 , by part 1 of proposition [ p : cons ] we have @xmath131 putting these three inequalities together , we conclude that @xmath132 this proves the lemma .
we next bound the norm of @xmath124 restricted to the smaller set @xmath58 , again using the general property of regularization . in our context , lemma 3.7 of @xcite applied to the vector @xmath133 yields @xmath134 along with lemma [ c : uj ] this directly implies : [ c : uj0 ] we have @xmath135
we now finish the proof of theorem [ t : it ] . the claim that @xmath55 follows by the same arguments as in @xcite .
the nontrivial part of the theorem is its last claim , that either ( i ) or ( ii ) holds .
suppose ( i ) in the theorem fails .
namely , suppose that @xmath136 , and thus @xmath137 set @xmath138 . by the comparability property of the coordinates in @xmath58 and since @xmath139 , there is a fraction of energy in @xmath140 : @xmath141 where the last inequality holds by lemma [ c : uj0 ] . on the other hand , we can approximate @xmath124 by @xmath142 as @xmath143 since @xmath144 , @xmath145 , and using lemma [ l : uj ] , we have @xmath146 furthermore , by definition of @xmath107 , we have @xmath147 .
so , by part 1 of proposition [ p : cons ] , @xmath148 using the last two inequalities and , we conclude that @xmath149 this is a contradiction to ( [ e : ubig ] ) so long as @xmath150 if this is true , then indeed ( i ) in the theorem must hold .
if it is not true , then by the choice of @xmath151 , this implies that @xmath152 this proves theorem [ t : it ] .
next we turn to the proof of corollary [ t : stabsig ] .
we first partition @xmath6 so that @xmath153 . then since @xmath2 satisfies the restricted isometry condition with parameters @xmath40 , by theorem [ t : stability ] and the triangle inequality , @xmath154 the following lemma as in @xcite relates the @xmath155-norm of a vector s tail to its @xmath156-norm .
an application of this lemma combined with will prove corollary [ t : stabsig ] .
[ l : ve ] let @xmath157 , and let @xmath158 be the vector of the @xmath38 largest coordinates in absolute value from @xmath159
. then @xmath160 let @xmath161 denote the @xmath162 largest entry of @xmath159 . if @xmath163 then @xmath164 so the claim holds .
thus we may assume this is not the case .
then we have @xmath165 simplifying gives the desired result . by lemma 29 of @xcite
, we have @xmath166 applying lemma [ l : ve ] to the vector @xmath167 we then have @xmath168 combined with , this proves the corollary .
we now show that by truncating the reconstructed vector , we obtain a @xmath45-sparse vector very close to the original signal .
[ c : napprox]assume a measurement matrix @xmath2 satisfies the restricted isometry condition with parameters @xmath40 for @xmath31 .
let @xmath6 be an arbitrary vector in @xmath32 , let @xmath169 be the measurement vector , and @xmath27 the reconstructed vector output by the romp algorithm . then @xmath170 where @xmath171 denotes the best @xmath38-sparse approximation to @xmath172 ( i.e. the vector consisting of the largest @xmath38 coordinates in absolute value ) .
let @xmath173 and @xmath174 , and let @xmath126 and @xmath175 denote the supports of @xmath176 and @xmath177 respectively . by corollary [ t : stabsig ] , it suffices to show that @xmath178 . applying the triangle inequality ,
we have @xmath179 we then have @xmath180 and @xmath181 since @xmath182 , we have @xmath183 . by the definition of @xmath175 , every coordinate of @xmath27 in @xmath175 is greater than or equal to every coordinate of @xmath27 in @xmath184 in absolute value .
thus we have , @xmath185 thus @xmath186 , and so @xmath187 this completes the proof .
* corollary [ c : napprox ] combined with corollary [ t : stabsig ] and implies that we can also estimate a bound on the whole signal @xmath6 : @xmath188
this section describes our experiments that illustrate the stability of romp .
we experimentally examine the recovery error using romp for both perturbed measurements and signals .
the empirical recovery error is actually much better than that given in the theorems .
first we describe the setup of our experiments . for many values of the ambient dimension @xmath11 , the number of measurements @xmath15 , and the sparsity @xmath9 , we reconstruct random signals using romp . for each set of values , we perform @xmath189 trials . initially , we generate an @xmath3 gaussian measurement matrix @xmath2 . for each trial , independent of the matrix , we generate an @xmath9-sparse signal @xmath6 by choosing @xmath9 components uniformly at random and setting them to one . in the case of perturbed signals , we add to the signal a @xmath11-dimensional error vector with gaussian entries . in the case of perturbed measurements ,
we add an @xmath15-dimensional error vector with gaussian entries to the measurement vector @xmath33 .
we then execute romp with the measurement vector @xmath14 or @xmath190 in the perturbed measurement case .
after romp terminates , we output the reconstructed vector @xmath27 obtained from the least squares calculation and calculate its distance from the original signal .
figure [ fig : meas2 ] depicts the recovery error @xmath191 when romp was run with perturbed measurements .
this plot was generated with @xmath192 for various levels of sparsity @xmath9 .
the horizontal axis represents the number of measurements @xmath15 , and the vertical axis represents the average normalized recovery error .
figure [ fig : meas2 ] confirms the results of theorem [ t : stability ] , while also suggesting the bound may be improved by removing the @xmath193 factor .
figure [ fig : sig4 ] depicts the normalized recovery error when the signal was perturbed by a gaussian vector .
the figure confirms the results of corollary [ t : stabsig ] while also suggesting again that the logarithmic factor in the corollary is unnecessary . | we demonstrate a simple greedy algorithm that can reliably recover a vector @xmath0 from incomplete and inaccurate measurements @xmath1 . here
@xmath2 is a @xmath3 measurement matrix with @xmath4 , and @xmath5 is an error vector .
our algorithm , regularized orthogonal matching pursuit ( romp ) , seeks to close the gap between two major approaches to sparse recovery .
it combines the speed and ease of implementation of the greedy methods with the strong guarantees of the convex programming methods . for any measurement matrix @xmath2 that satisfies a uniform uncertainty principle
, romp recovers a signal @xmath6 with @xmath7 nonzeros from its inaccurate measurements @xmath8 in at most @xmath9 iterations , where each iteration amounts to solving a least squares problem .
the noise level of the recovery is proportional to @xmath10 .
in particular , if the error term @xmath5 vanishes the reconstruction is exact .
this stability result extends naturally to the very accurate recovery of approximately sparse signals . |
the experimental and computational advances in our capacity of producing and trapping samples of cold ( below 1 k ) or ultracold ( below 1 mk ) molecules in the gas phase have made tremendous progress in recent years and have dramatically enhanced our detailed understanding of the processes at hand within their nanoscopic evaluation @xcite .
the ensuing cold and ultracold molecular species have many possible applications such as allowing for the accurate measurement of fundamental physical properties @xcite , the possible evaluations of small energy differences between enantiomers @xcite and the observation of temporal changes of the fine - structure constant @xcite .
furthermore , dipolar species have been suggested as qubits in quantum computers @xcite and polar gases in general , neutral and ionic , are expected to exhibit even more marked novel features due to their environment @xcite .
in addition , translationally cold molecular ions , and polar ions at that , can also become embedded inside coulomb crystals @xcite and provide ideal targets for a large variety of investigations like high precision spectroscopic measurements and state selected partners in ionic reactions @xcite .
earlier suggestions in this experimental area include the analyses of the ca@xmath0/na system by the storrs group @xcite , while the choice of a molecular partner as we shall be examining here is currently being considered by various experimental groups as a possible option @xcite .
the insertion of a molecular ion partner within a coulomb crystal @xcite , i.e. their phase transition to an ordered state at ultralow temperatures in the range of a few millikelvins , is one of the intriguing steps for the subsequent handling of such ions within a broad variety of molecular processes .
the present study therefore intends to provide , within a fully ab initio formulation of the problem , an accurate description of the interaction forces which drive some of the possible collisional processes at such low temperatures .
in particular , we wish to present the energy details and the spatial features of the various potential energy surfaces ( pes ) which become relevant in the direct process that involves mgh@xmath0(x@xmath1 ) in collision with rb(@xmath2s ) , both partners being current candidates in cold trap experiments @xcite . because of
the differences between the ionization potentials of the two partners we shall also show that the mgh(x@xmath3 ) molecule interacting with rb@xmath0(@xmath4s ) is indeed an important outgoing channel , which provides an interesting _ reactive _ outcome of the process and which remains open even at ultralow energies .
the section [ sec : compdet ] therefore discusses our computational details while section [ sec : result ] presents our results and analyses the novel features of our final pes s .
conclusions are given by the last section [ sec : concl ] .
in order to become familiar with the terms and the physical features of the chosen partners in the iontrap , it is useful to remind ourselves of the way in which their electronic features are put together from an ab initio viewpoint . a direct comparison between the ionization potentials ( ip ) of the mgh(x@xmath3 ) and rb(@xmath2s )
reveals that the ground state of the complex molecular ion rbmgh@xmath0 has its lowest energy correlation with the mgh(x@xmath3 ) + rb@xmath0(@xmath4s ) species .
consequently , the electronic state of the ion complex that correlates with mgh@xmath0(x@xmath1 ) + rb(@xmath2s ) , i.e. with fragments which do not undergo charge - exchange , is an excited state from the rbmgh@xmath0 electronic manifold .
if we start our analysis with the simpler case of a collinear geometry , an appropriate generalization of the wigner - witmer correlation rules for diatomic molecules ( see for example pagg .
281 - 284 of @xcite ) allows us to derive the electronic state manifold of the complex using only the knowledge of the spatial and spin symmetry properties of the electronic states of the combining fragments .
it is worth noting here that the use of the collinear geometry does not limit the generality of our results : it is always possible to univocally resolve the electronic states symmetry species for a particular spatial configuration of a molecule into those of a molecular geometry of lower symmetry by means of group theory .
furthermore , if we combine that qualitative information with the available spectroscopic data for the mgh and mgh@xmath0 , and the ip of rb and mgh , we would obtain the relative energies for the complex electronic states . in this particular case
, we have taken as origin of the energy axis the mgh(x@xmath3 ) + rb@xmath0(@xmath4s ) energy at infinite separation distance .
the relative energy of the mgh@xmath0(x@xmath1 ) + rb(@xmath2s ) asymptote , which is the lowest energy pathway to dissociate into the mgh@xmath0 molecular ion and a neutral rb atom , with respect to the origin of the energy axis is simply the electronic energy difference between the mgh(x@xmath3 ) and the rb(@xmath5 ) ips ( about 2.69 ev ) .
the asymptotic relative energy of any electronic state of the triatomic complex which correlates with an excited electronic state of one , or both of the fragments , is simply obtained by adding the electronic excitation energy to the energy of the appropriate asymptote .
for example , the mgh@xmath0(x@xmath1 )
+ rb(@xmath2p ) energy equals the energy of the mgh@xmath0(x@xmath1 ) + rb(@xmath2s ) plus the rb electronic @xmath2s-@xmath2p transition energy . in this way it is possible to determine the relative energy ordering of the triatomic complex electronic states , at least for an infinite separation between the fragments .
in fact , it should be clear that this approach while giving an exact idea of the possible electronic states , says nothing about their stability and even less about which nuclear conformation is the most stable .
nevertheless , this qualitative information provides valuable help in the choice of the more suitable _ ab initio _ method which should be used in the actual electronic structure calculation .
our figure [ fig : rbmgh_levels ] summarizes the results obtained from this type of analysis . at the very right side of the figure [ fig : rbmgh_levels ] we report the possible dissociation channels in the molecular ion mgh@xmath0 and neutral rb for the energy range we have considered relevant for our purposes in this study , while in the center the possible dissociation pathways into neutral mgh and rb@xmath0 are represented . finally , proceeding from the right to the left of figure [ fig : rbmgh_levels ] the wigner - witmer correlation rules bring the fragments electronic states into the complex ones .
a perusal of figure [ fig : rbmgh_levels ] suggests that a reasonable description of the rbmgh@xmath0 manifold of states could be obtained as a summation of three subsets of electronic states well separated in energy : in the first subset we put the system ground state , in the second one the a@xmath6 and the b@xmath3 states , while the third one collects a number of very high energy electronic states which will eventually correlate with the rubidium excited states and the mgh(b@xmath3 ) state .
it is also worth noting here that all the electronic states employed in the present scheme have the same spin symmetry : they are all doublets . with that information in mind , it is clear that an _ ab initio _ description of the rbmgh@xmath0 electronic structure should be able to treat at least the ground state and the first two excited states in a balanced manner .
in other words , in a case like the present one the use of a multireference ( mr ) electronic wavefunction is mandatory . we have therefore designed a mr wave function within the state averaged complete active space ( sa - cas ) approach : thus , the problem of the choice of a set of n - electron wave functions ( slater determinants or configuration state functions ) is recast into a partition of the molecular orbital space in * active orbitals , which can have different occupation numbers ( 0,1,2 ) and * inactive orbitals , which are doubly occupied in all the configurations .
the electrons which populate the active orbitals are then called the active electrons .
it is worth noting that , in our calculation the active electrons correspond exactly to the ones commonly defined in chemistry as the valence electrons .
in other words , we choose those molecular orbitals that are supposed to generate within the cas procedure all the n - electron configurations which are important for the description of the electronic states we are interested in . interacting with a rubidium atom .
see text for details.,scaledwidth=100.0% ] a prior knowledge of the electronic structure of the isolated fragments is therefore helpful in guiding the choice of the active orbital space within the complex : in figure [ fig : mgh_uamo ] a united - atom correlation diagram allows us to determine the symmetry and gives us an idea of the relative energies of the valence molecular orbitals of mgh and mgh@xmath0 by considering the corresponding structures of al and al@xmath0 .
following the aufbau principle , we therefore see that the first two low - lying electronic states , x@xmath3 and a@xmath7 , could be described by single electronic configurations : @xmath8 and @xmath9 respectively .
the case of the next higher energy state b@xmath3 is not so straightforward because more than one configuration could be important to describe it : @xmath8 , @xmath10 , @xmath11 , etc . in the case of the mgh@xmath0(x@xmath12 )
state the single configuration @xmath13 could be the one which manages to account for the main features of that electronic state .
we used that qualitative argument for the choice of a suitable orbitals active space for the calculation of the x@xmath3 , a@xmath6 and b@xmath3 mgh electronic states : @xmath14 , @xmath15 , @xmath16 , @xmath17 , @xmath18 and a @xmath19 anti - bond molecular orbital which has to be included in order to correctly describe the bond dissociation .
this active space produces the cas(3,6 ) ( 3 active electrons in 6 active orbitals ) multireference wavefunction that we used as reference in the multi reference configuration interaction with single and double excitations ( mrci - sd ) calculation of the mgh potential energy curves ( pecs ) reported in the upper panel of figure [ fig : mgh_pes ] .
a completely analogous orbital active space was used to calculate the pecs of mgh@xmath0 low lying electronic states ( reported in the lower panel of figure [ fig : mgh_pes ] ) , here the only difference is the number of active electrons : there are two of them .
it is interesting to note that the value of the coefficients of the configurations interaction ( ci ) expansion actually reflects the simple qualitative argument based on the aufbau principle and the molecular orbital scheme obtained from the united atom correlation diagram .
in fact , as we can see from table [ tab : cicoeff ] , the coefficient of the @xmath8 configuration is about 0.95 for the x@xmath3 state , while the a@xmath6 is chiefly described by the @xmath9 configuration ( ci coefficients of 0.96 ) . as expected , in the case of the b@xmath3 state more than one electronic configuration is important : the ci coefficients of the @xmath10 and @xmath20 configurations have comparable values : 0.77 and 0.55 respectively .
it is worth noting that the configurations in which the @xmath19 is occupied essentially do nt contribute to the ci expansion , at least at the interatomic distances around the equilibrium value .
0.60 @c||c|c||c||c conf .
& x@xmath3 & b@xmath3 & conf . & a@xmath6 + 2 + 00 00 & 0.9537 & -0.1410 & 2000 + 0 & 0.9619 + + 200 00 & 0.0683 & 0.7480 &
+ + 00 -0 & 0.1253 + 20 + 0 00 & -0.1951 & -0.5526 & 0200 + 0 & -0.1206 + + 000 02 & 0.0433 & -0.1626 & 0 + -0 + 0 & -0.0762 + + 000 20 & 0.0433 & -0.1626 & 0 + 0- + 0 & 0.0576 + 200 + 00 & 0.0204 & -0.0823 & & + + -+0 00 & 0.0792 & -0.0780 & & + 0 + 00 20 & -0.0328 & 0.0662 & & + 0 + 00 02 & -0.0328 & 0.0662 & & + + -0 + 00 & -0.0003 & 0.0606 & & + + + -0 00 & -0.0538 & 0.0146 & & + + 020 00 & -0.0089 & -0.0501 & & + on the rb atom side the situation is much simpler : asymptotically we are dealing with two species , rb@xmath0(@xmath4s ) and the rb(@xmath2s ) , both of which could be described by one electronic configuration : [ kr]@xmath21 and [ kr]@xmath22 . using that information one step further , we chose as active orbitals those complex molecular orbitals which correlate asymptotically with the mgh , mgh@xmath0 , rb , rb@xmath0 relevant configurations ; namely , mgh(@xmath14 , @xmath15 , @xmath16 , @xmath17 , @xmath18 ) , rb(@xmath23 ) .
this orbital space defines the cas(3,6 ) wavefunction that we finally used as a reference space in the mrci - sd calculations of the rbmgh@xmath0 pess .
@xmath24 . ] the geometry of the mgh - rb ion complex is described in jacobi ( r , r,@xmath25 ) coordinates where r is the distance between the atoms within the mgh molecule , r indicates the distance between the center of mass of mgh and the rb atom and the angle @xmath25 is taken between r and r : the @xmath26 orientation corresponds here to the mgh - rb collinear configuration with the rb atom approaching the h side of the neutral / ionic molecule : in the present calculations the distance r was kept fixed at the calculated equilibrium distance for the mgh@xmath0(x@xmath27 ( @xmath28 ) .
the equilibrium distance of the mgh@xmath0(x@xmath1 ) electronic states was extrapolated from the ground state pec reported in the lower panel of figure [ fig : mgh_pes ] .
the electronic structure calculations have been carried out on a discrete grid defined by varying the jacobi coordinates r and @xmath25 .
the r coordinate was allowed to vary between 3.00 and 16.00 .
a radial grid step size @xmath29 of 0.25 was used in the interval [ 3.00 , 7.50 ] , in the interval [ 7.50,10.0 ] @xmath30 , while in the interval [ 10.0,16.0] @xmath29 was 1.00 .
we defined an angular grid in the interval [ 0@xmath31,180@xmath31 ] with a step size @xmath33 .
the total number of grid points and , eventually , the number of computed ab initio energies was 450 .
the cas(3,6 ) wavefunction outlined in the previous section was used , for each grid point , in a state - averaged mcscf ( sa - mcscf ) calculation of the lowest 4 doublet states of the rbmgh@xmath0 ( the three @xmath34 and the one @xmath35 ) .
however , because the calculations were carried out in @xmath36 symmetry , the actual number of averaged states is indeed 5 ( four @xmath37 and one @xmath38 ) because the @xmath35 state splits into two components ( @xmath37 , @xmath38 ) which are fully degenerate only at collinear geometries .
the sa - mcscf therefore generates a common set of orbitals for the 5 states which in turn we have used for the subsequent mrci - sd calculations which we expect to correctly account for electron correlation effects : all the electrons , except the mg ( 1s@xmath2 ) , were treated as correlated in the present calculations .
it is worth noting that the highest @xmath3 state was included only to ensure correct convergence of the sa - mcscf wavefunction at all the considered geometries and does nt really have a simple physical meaning : from figure [ fig : rbmgh_levels ] one can see that the actual electronic states above the b@xmath3 are those which correlate with mgh@xmath0(x@xmath1)+rb(@xmath2p ) , i.e. with our present asymptotic state .
those electronic states , however , can not be described by the cas(3,6 ) wavefunction because the active orbitals space does not include molecular orbitals of the complex which correlate with the rb 5p atomic orbitals . however ,
this is not a real problem in the present work since we would like to gather information on the possible relaxation processes that specifically involve mgh@xmath0(x@xmath1 ) interacting with rb(@xmath2s ) . in other words
, we can disregard what is energetically above the b@xmath3 state of the complex and focus only on the first three lower lying electronic states of the latter . the gaussian basis set we selected was the cc - pvqz(spdf ) ( 16s,12p,3d,2f ) contracted as [ 6s,5p,3d,2f ] for the mg @xcite , the cc - pvqz(spd ) ( 6s,3p,2d ) contracted as [ 4s,3p,2d ] @xcite .
the rb atom was treated within the effective core potential ( ecp ) framework in order to reduce the number of electrons treated explicitly but to still account for the relativistic effects on the core electrons .
the ecp used was the ecp28mdf from the stuttgart group @xcite ; this ecp leaves the @xmath39 rb electrons explicitly described while the effect of the others is included within a pseudopotential formulation .
the gaussian basis set chosen for the rb atom was the companion basis set published along with the ecp28mdf parameters : the basis is ( 13s,10p,5d,3f ) contracted as [ 7s,6p,5d,3f ] @xcite .
all the calculations have been done by using the molpro 2000.6 suite of programs @xcite . in order to obtain the interaction energy from the calculated electronic energy
, we used as `` zero of the interaction energy '' the electronic energy of the system calculated at r=300 . at this distance
the angular variation of the ground state energy is about @xmath40 ( about 0.002 @xmath41 ) .
the remaining interaction energy at this distance could be estimated as not greater than @xmath42 which is the difference between the ground state energy calculated at r=300.0and the one calculated at r=275.0 .
0.85 @c|c basis & i.p .
( ev ) + cc - pvtz(sp)+ecp+cpp & 6.627 + cc - pvqz(sp)+ecp+cpp & 6.815 + full electron mg(spd ) h(sp ) cc - pvtz & 6.857 + full electron mg(spdf ) h(spd ) cc - pvqz & 6.862 + 0.85 @c|c|c||c method & @xmath2p ( @xmath43 ) & @xmath2d ( @xmath43 ) & i.p.(ev ) + mrci - sd & 12731.4 & 19605.2 & 4.147 + rs2 & 12849.4 & 20367.4 & 4.167 + rs3 & 12647.2 & 20187.0 & 4.132 + exp .
& 12697.75 & 19355.43 & 4.177 + because of the important role played by the ionization potential of the molecular partner in the present study , we report in table [ tab : ip ] the values obtained from our test calculations for the mgh molecule and in table [ tab : rbext ] the calculated ip along with the lower - lying energy levels of the rb atom . the molecular data
although not corrected for the zero point energy ( zpe ) effect , employ the mgh equilibrium geometry of 1.77 ( from ref .
@xcite ) and the mgh@xmath0 equilibrium bond length of about 1.64 ( from ref .
since our calculated values turned out to be very close to the experiments ( see before ) we see in the molecular case that the cc - pvqz basis set choice converges to a stabilized estimate of the ip value .
the atomic calculations also appear to get very close to experiments and surmise a difference between the two asymptotic charge - exchange systems of about @xmath44 .
to make pictorially clearer the relative shapes of the molecular potential curves which are relevant for the present study , we are presenting in the two panels of figure [ fig : mgh_pes ] the low - lying electronic states of mgh ( upper panel ) and the ones related to the molecular ion in the lower panel .
the data clearly show the small geometry change that takes place upon the charge exchange process when we start from the v=0 vibrational state of mgh@xmath0(x@xmath1 ) .
, title="fig : " ] , title="fig : " ]
if we were to adopt for simplicity a purely electrostatic model , which is increasingly valid as the r distance becomes greater , we would recognize that for the rbmgh@xmath0 complex we have two kinds of asymptotic behavior : the complex can dissociate as neutral mgh and rubidium ion or it can generate an ionic mgh molecule and a neutral rubidium atom .
it is obvious at this point that the long range behavior of the potential energy function in these two cases is likely to be quite different .
mgh is a polar molecule and therefore in the first case the long - range potential is dominated by the charge - dipole interaction which is proportional to @xmath45 . because the mgh dipole orientation , in our case this interaction is attractive for @xmath25 between 0@xmath31 and 90@xmath31 and repulsive for @xmath25 between 90@xmath31 and 180@xmath31 degrees . in the second option
the dominant long range term is due to the charge - induced dipole interaction which is proportional to @xmath46 and is always attractive since it depends on the spherical dipole polarizability of the neutral rb partner . ) .
see main text for discussion . ]
the results reported by figure [ fig : rbmgh_pes_cut0 ] further underline the characteristic behavior of the relevant surfaces : the lowest curve for the rb@xmath0 approach on the h atom of the neutral molecular partner clearly shows an attractive well of about @xmath47 , placed well below the incoming channel associated with the charge exchange process ( the b@xmath3 state ) and also below the more stable mgh@xmath48-rb@xmath0 complex of @xmath49 symmetry . in other words , the marked difference in energy between the ionization potential ( see table [ tab : ip ] and [ tab : rbext ] ) generates the dramatic energy gain that occurs when the interaction is included and which produces the neutral molecule in either an excited or its ground electronic state .
the configurations which describe the approach at the mg side are reported in figure [ fig : rbmgh_pes_cut180 ] . , but this time for the rb approach on the mg - side of the molecule ( @xmath50 ) . ]
the basic process remains the same in terms of providing the mgh(x@xmath3 ) as the lower energy molecular partner in the complex , while we see that , for the linear geometries , the possible splitting of the @xmath35 state associated with the nearest charge - exchange channel into its two components ( @xmath37 and @xmath38 ) strengthens the nonadiabatic coupling with the incoming channel since the latter also shows an @xmath37 symmetry component at those geometries . in other words ,
the formation of mgh@xmath0-rb complex at large distances ( around 6.0 ) is very strongly coupled with the quasi - degenerate charge - exchange process into mgh(a@xmath6)+rb@xmath0(@xmath4s ) , which can in turn radiatively decay into the @xmath37 complex with mgh(x@xmath3 ) as a final partner molecule .
thus , while the @xmath26 orientation indicates a more strongly bound ( mgh@xmath48-rb@xmath0 ) complex which could then radiatively decay to its stable ground state but which had to be reached by a nonadiabatic coupling with a sizable energy gap ( see figure [ fig : rbmgh_pes_cut0 ] ) , the approach on the other side of the molecule ( figure [ fig : rbmgh_pes_cut180 ] ) suggests a much smaller energy gap for the nonadiabatic event and the more likely radiative emission from a quasi - bound ( mgh@xmath48-rb@xmath0 ) complex .
the intermediate situation described by figure [ fig : rbmgh_pes_cut90 ] report the cuts associated with the @xmath51 orientation . the @xmath35 state is now split by the and [ fig : rbmgh_pes_cut180 ] but for the @xmath51 choice of the jacobi angle .
see main text for the discussion of the added symbols onto the computed curves . ]
interaction and therefore one of its components ( the @xmath37 state ) corresponds to a strongly bound ( mgh@xmath48+rb@xmath0 ) charge exchange complex close in energy to the incoming channel .
the latter , bound intermediate complex can now decay to a nearby lower state and form the most stable charge - exchange arrangement . here again , therefore we see that the analysis of the slopes and feature of the pes already indicates ( at least preliminarily ) that the formation of ultracold mgh could occur via charge - exchange complex into an intermediate , bound complex where the molecular partner is in its excited electronic state , the mgh(a@xmath6)(@xmath37 ) state , from which it can further decay by radiative transitions . from the above discussion , therefore , it is clear that the strength of the relative interactions , and the possible existence of complex bound states within any of the relevant pess for the title system , are crucial elements for selecting possible pathways to the decay , or excluding it , for the initial system containing the molecular cation .
we shall therefore present below the overall fitting of all the relevant interactions which we have discussed above along specific angular cuts and describe their overall feasibility for collisionally coupling radiative states into the final charge - exchange products .
the calculation of the potential energy surfaces have been carried out , as mentioned before , by considering the initial molecular ion at its equilibrium geometry so that only a 2-dimensional mapping of the pes s was generated using the previously discussed ab initio procedure .
each pes was then represented by using the two - center expansion already employed by us in our previous work @xcite and described there in greater detail . here ,
suffice it to say that we can represent each adiabatic ( born - oppenheimer ) pes as given by : @xmath52 where @xmath53 and @xmath54 stand for `` short - range '' and `` long - range '' respectively and the labels @xmath55 and @xmath56 refer to either atom in the molecular partner .
the two terms in equation [ eq : sr ] contain a set of parameters describing the relevant radial coefficients centered on either of the molecular atoms times specific legendre polynomials which are able to produce the angular dependence from that atom within the anisotropic interaction .
their actual forms have been given already in our previous work @xcite : here , it is obvious that the functional form of the long range part , @xmath58 , strictly depends on which fragment hosts the positive charge .
two cases can be distinguished : * the positive charge is carried by the diatom , * the positive charge is carried by the impinging atom . the two options lead to two possible long range behaviors : @xmath59 in the case of the mgh@xmath0 + rb asymptote , and
@xmath60 if the complex dissociates as mgh@xmath48 + rb@xmath0 .
this choice of the long range parts follows directly from the electrostatic model proposed and discussed in section [ sec : resulta ] . in both cases as described by the equations ( [ eq : lr1 ] ) and ( [ eq : lr2 ] ) , the @xmath61 term was included to take into account the dipole - induced dipole interaction effect .
all the employed long - range coefficients are not known and were obtained here from the numerical ( linear ) fitting of the long range part ( @xmath62 ) of the computed pess .
the computed values of the long range expansion coefficients are reported in table [ tab : lrcoeff ] .
0.85 @c||c|c|c & c@xmath63 & c@xmath64 & c@xmath65 + x@xmath3 & @xmath66 & @xmath67 & @xmath68 + a@xmath6 & @xmath69 & @xmath70 & @xmath71 + b@xmath3 & & @xmath72 & @xmath73 + potential energy surface for rbmgh@xmath0 ( linear symmetry term symbol : x@xmath3 ) .
the arrow marks the deep well region of the linear configuration . ]
the fitting parameters and the actual analytic representation of the present pess are available on request from the authors . a pictorial view of the lower electronic state is given by the 3d representation of figure [ fig : rbmgh_pes_e2 ] , where the deep attractive well in the collinear geometry with rb approaching the h - side of the molecule along charge - exchange surface , as presented by figure [ fig : rbmgh_pes_cut0 ] , is further visible and clearly shows the strong role played even at short range by the electrostatic forces .
the overall anisotropy of the ground - state complex configuration is seen to be rather strong , with a sizable barrier existing for the complex bending motion outside a narrow angular cone around @xmath26 . in the figure the arrow marks the minimum energy of that angular cone . in other words ,
the lower charge - exchange surface is dominated by strong attractive forces on the side of the molecule to which the dipole is pointing and is locking the incoming atom via the coupling term given asymptotically by equation ( [ eq : lr2 ] ) . symmetry .
lower panel : @xmath38 symmetry .
the two electronic states are degenerate for linear nuclear configuration and give rise to the a@xmath6 term symbol , title="fig : " ] symmetry .
lower panel : @xmath38 symmetry .
the two electronic states are degenerate for linear nuclear configuration and give rise to the a@xmath6 term symbol , title="fig : " ] the two panels of figure [ fig : rbmgh_pes_e2 ] further show the next higher electronic states : the lower panel reports the fitting of the raw points which describe the upper charge - exchange channel , i.e. the one which asymptotically yields the electronically excited state of the neutral molecular partner , the mgh(a@xmath6 ) configuration coupled with the ionic rb@xmath0(@xmath4s ) .
it is located at about @xmath74 above the lowest pes of figure [ fig : rbmgh_pes_e1 ] and we see that essentially it maintains its strong anisotropy with a linear complex formation of minimum energy as rb@xmath0 attaches to the h - atom of the neutral molecule .
the next higher pes is the one associated with the other component of the complex in the @xmath35 state i.e. it describes the charge - exchange compound of @xmath37 symmetry .
its nearly - spherical nature , therefore , is confirmed by the overall pes being very flat as the angle @xmath25 varies and without the marked barrier away from linearity shown by the more asymmetrical complex on the lower panel of figure [ fig : rbmgh_pes_e2 ] .
both components indicate very weak interaction on the mg - side approach of the impinging atom .
the overall anisotropy , however , changes very markedly as we look at the entrance channel side of the potential energy surface , i.e. at the electronic state of the complex which describes the @xmath3 state whereby the asymptotic partners are given by the mgh@xmath0 molecular ion in the x@xmath1 state and the neutral rb(@xmath2s ) atomic partner .
the change in anisotropy is shown pictorially by the data presented in figure [ fig : rbmgh_pes_e3 ] , where the spherical dipole polarizability of the atomic partner drives the asymptotic behavior of the whole pes .
the anisotropy of the surface presented in figure [ fig : rbmgh_pes_e3 ] is indeed showing a smaller attractive well located further out from the molecular center - of - mass and a more uniformly attractive region over the whole angular range .
the above findings therefore suggest a possible dynamical scheme that could be active during the initial interaction between the coulomb crystal arrangement of the molecular ions and the ultracold rb gas interacting with them . in other words ,
the uniformly attractive features of the entrance channel pes indicate long - range attraction between partners which is largely independent of their relative orientational approach and which could thus drive the system into its geometries of stronger interactions .
symmetry ( linear nuclear configuration term symbol : b@xmath3 ) ] a possible , pictorial path of the process is outlined over the data of figure [ fig : rbmgh_pes_cut90 ] , where we show superimposed over the potential curves : the entrance channel access ( dashed arrow on the upper right ) to the region of nonadiabatic coupling between the @xmath3 electronic state and the upper charge - exchange channel which produces mgh@xmath48+rb@xmath0 .
the likely dominance of collinear orientations in that channel , and the dramatic increase of the attractive well depth in both @xmath35 components , could drive the system to reach one of the many bound states supported by the @xmath37 components of the @xmath6 electronic configuration , as represented by the thick circular area .
those states could then undergo radiative stabilization into the vibrationally bound complexes of the electronic ground state ( this process is represented by the dotted downward arrow ) , thereby forming the neutral molecule : one can argue that the latter can then be replaced by the cold atomic ions within the coulomb crystal environment . to confirm the above mechanism obviously requires a more detailed analysis of its main ingredients , i.e. of the bound states supported by the [ mghrb]@xmath0 complex in both the relevant electronic states after charge - exchange transitions , the spatial features and the changing strength of the corresponding transition moments , together with the actual calculation of the relevant franck - condon integrals .
all of the above shall be the subject of an additional study , currently under way in our laboratory , while the present work chiefly intends to extract first and discuss the specific properties of accurate ab initio potential energy surfaces which could be employed to drive the ultracold dynamics .
the general characteristics of such interactions , to our knowledge , have never been obtained before .
in this work we have computed from ab initio methods a set of potential energy curves which are expected to be involved in the low - energy interaction of a molecular polar cation , the mgh@xmath0(@xmath1 ) , and the neutral rb(@xmath2s ) atomic partner .
the experimental setup which is considered relevant for the present study is the one provided , at ultralow temperatures , by a coulomb crystal arrangement of molecular ions interacting , and sympathetically cooled , by rb atoms in a cold trap environment @xcite .
the analysis of the accurate calculations which have been carried out here has provided the analytic fitting of 2d grids of raw ab initio points representing the relative strength and the orientational anisotropy of the overall interactions between the four possible asymptotic channel partners : mgh@xmath0 and rb and the three low - lying electronic states of mgh in interaction with rb@xmath0 .
the results indicate the existence of marked energy gain on charge exchange , as obviously expected from the very beginning , but also allow us to uncover the marked orientational effects of the dynamical coupling between partners along the entrance channels and the expectedly strong nonadiabatic coupling which would allow for transitions into intermediate charge - exchange states that include mgh@xmath48 as a partner within the complex .
the present calculations also indicate the possible radiative decay within the trap onto the ground state complex whereby the molecular ion within the coulomb crystal can be replaced by cold atomic ions : the latter partners are , in fact , formed within a fairly attractive potential supporting a large number of bound states that could then break up independently of their initial geometry of formation into the intermediate state . in conclusion , the present study constitutes the necessary , preliminary stage for any dynamical analysis of the collisional processes occurring in the cold environment of a coulomb crystal arrangement : it provides detailed information , which to our knowledge has not been available before , on the spatial extension and orientational features of four of the relevant interaction potentials which can be involved in understanding the nanoscopic dynamics of trap collisions with ultracold ions .
the outcomes of the calculations locate spatial regions where nonadiabatic couplings are more likely to play an important role , suggest possible intermediate complex formations during charge - exchange processes and indicate a possible radiative pathway to ionic replacement within the coulomb crystal environment .
although no experiments as yet exist on the present complex , the information which has been gathered by the present calculations should be a useful starting point for deciding on the likelihood of this molecular ion to be employed for sympathetic cooling with rb cold atomic condensates .
the financial support of the university of rome research committee , of the miur national research project prin 2006 and the computational aid from the caspur consortium are gratefully acknowledged .
we also thank professor s. schiller for suggesting this problem to us . | the electronic structures of the manifold of potential energy surfaces generated in the lower energy range by the interaction of the mgh@xmath0(x@xmath1 ) cationic molecule with rb(@xmath2s ) neutral atom are obtained over a broad range of jacobi coordinates from strongly correlated _ ab initio _ calculations which use a multireference ( mr ) wavefunction within a complete active space ( cas ) approach .
the relative features of the lowest five surfaces are analyzed in terms of possible collisional outcomes when employed to model the ultracold dynamics of ionic molecular partners . |
as a relativistic wave equation , the dirac equation provides naturally a description of an electron @xcite .
following dirac s discovery of the linear equation of the electron , there appears the fundamental idea of nonlinear description of an elementary spin-1/2 particle which makes it possible to take into account its self - interaction @xcite .
heisenberg put forward the idea to use a nonlinear dirac ( nld ) equation as a possible basis model for a unified field theory @xcite .
a key feature of the nld equation is that it allows solitary wave solutions or particle - like solutions the stable localized solutions with finite energy and charge @xcite .
that is , the particles appear as intense localized regions of field which can be recognized as the basic ingredient in the description of extended objects in quantum field theory @xcite .
different self - interactions give rise to different nld models mainly including the thirring model @xcite , the soler model @xcite , the gross - neveu model @xcite ( equivalent to the massless soler model ) , and the bag model @xcite ( _ _ i.e. _ _ the solitary waves with finite compact support ) , all of which attracted wide interest of physicists and mathematicians around the 1970s and 1980s , especially on looking for the solitary wave solutions and investigating the related physical and mathematical properties @xcite . for the nld equation in ( 1 + 1 ) dimensions ( _ _ i.e. _ _ one time dimension plus one space dimension ) ,
several analytical solitary wave solutions are derived in @xcite for the quadric nonlinearity , @xcite for fractional nonlinearity as well as @xcite for general nonlinearity by using explicitly the constraint resulting from energy - momentum conservation , and summarized by mathieu @xcite .
in contrast , there are few explicit solutions in ( 1 + 3 ) dimensions except for some particular cases shown in @xcite in spite of their existence claimed by mathematicians for various situations @xcite ( the readers are referred to an overview @xcite on this topic ) , and most understanding is based on numerical investigations , _ _
e.g. _ _ @xcite . beyond this , the study of the nld equation in ( 1 + 1 ) dimensions could be very helpful for that in ( 1 + 3 ) dimensions since the ( 1 + 1)-dimensional nld equation correspond to the asymptotic form of the equation in the physically interesting case of ( 1 + 3 ) dimensions as emphasized by kaus @xcite .
that is , some qualitative properties of the nld solitary waves could be similar in such two cases .
an interesting topic for the nld equation is the stability issue , which has been the central topic in works spread out over several decades in an effort that is still ongoing .
analytical studies of the nld solitary wave stability face serious obstacles @xcite , while results of computer simulations are contradictory @xcite .
the stability analysis of the nld solitary waves is still a very challenging mathematical problem to date @xcite .
recent efforts in this direction can be found in @xcite .
another rising mathematical interest related to the nld equation is the analysis of global well - posedness , _ _
e.g. _ _ see @xcite and references therein . in the case of that theoretical methods were not able to give the satisfactory results , numerical methods were used for obtaining the solitary wave solutions of the nld equation as well as for investigating the stability .
an important step in this direction was made by alvarez and carreras in 1981 @xcite , who simulated the interaction dynamics between the ( 1 + 1)-dimensional nld solitary waves of different initial charge for the soler model @xcite by using a second - order accurate crank - nicholson ( cn ) scheme @xcite .
they first saw there : charge and energy interchange except for some particular initial velocities of the solitary waves ; inelastic interaction in binary collisions ; and oscillating state production from binary collisions .
motivated by their work , shao and tang revisited this interaction dynamics problem in 2005 @xcite by employing a fourth - order accurate runge - kutta discontinuous galerkin ( rkdg ) method @xcite .
they not only recovered the phenomena mentioned by alvarez and carreras but also revealed several new ones , _ _
e.g. _ _ collapse in binary and ternary collisions of two - humped nld solitary waves @xcite ; a long - lived oscillating state formed with an approximate constant frequency in collisions of two standing waves @xcite ; full repulsion in binary and ternary collisions of out - of - phase waves @xcite .
their numerical results also inferred that the two - humped profile could undermine the stability during the scattering of the nld solitary waves .
note in passing that the two - humped profile was first pointed out by shao and tang @xcite and later gotten noticed by other researchers @xcite .
besides the often - used cn @xcite and rkdg methods @xcite , there exist many other numerical schemes for solving the ( 1 + 1)-dimensional nld equation : split - step spectral schemes @xcite , the linearized cn scheme @xcite , the semi - implicit scheme @xcite@xcite , legendre rational spectral methods @xcite , multi - symplectic runge - kutta methods @xcite , adaptive mesh methods @xcite _ etc_. the fourth - order accurate rkdg method @xcite is very appropriate for investigating the interaction dynamics of the nld solitary waves due to their ability to capture the discontinuous or strong gradients without producing spurious oscillations , and thus performs better than the second - order accurate cn scheme @xcite . however , the high cost due to the relatively more freedoms used in each cell and the stringent time step constraint reduce its practicality in more realistic simulations where realtime and quantitative results are required .
recently , there has been a remarkable upsurge of the interest in the nld models , as they emerge naturally as practical models in other physical systems , such as the gap solitons in nonlinear optics @xcite , bose - einstein condensates in honeycomb optical lattices @xcite , as well as matter influencing the evolution of the universe in cosmology @xcite . in view of such new trend ,
longtime stable , efficient , conservative and high - order accurate numerical methods for solving the nld equation are highly desirable .
finite difference methods , usually as the first try in practice , enable easy coding and debugging and thus are often used by physicists and engineers .
however , detailed discussion and careful comparison on finite difference solvers for the nld equation are not existed . to this end , the present work as the first step will extendedly propose the finite deference schemes for solving the nld equation with the scalar and vector self - interaction .
a general and precise comparison among them will be presented .
however , all of these finite difference methods are often of the second order accuracy and thus sustain fast error growth with respect to time .
to achieve relatively slow error growth , high - order accurate numerical methods are required . by exploiting the local conservation laws of the nld equation , we present exponential operator splitting ( os ) schemes which are time reversible and
can conserve the discrete charge .
one of the high - order accurate os schemes is afterwards adopted to investigate the interaction dynamics for the nld solitary waves under the general scalar and vector self - interaction .
it should be noted that the experiments carried out in the literatures are all limited to the collisions of the nld solitary waves under the quadric scalar self - interaction . here , the binary collisions of the nld solitary waves under the cubic scalar self - interaction or under the vector self - interaction or under the linear combination of the scalar and vector self - interactions are all studied for the first time .
the paper is organized as follows .
there is a brief review of the nld equation in section [ sec2:gnld ] and the solitary wave solutions are also derived there for the general scalar and vector self - interaction .
the numerical schemes are presented in section [ sec : diffsch ] and corresponding numerical analysis is given in section [ sec : discussion4fdm ] .
the numerical results are presented with discussion in section [ sec : discussion4fdm ] .
the paper is concluded in section [ sec : conclusion ] with a few remarks .
this section will introduce the ( 1 + 1)-dimensional nld equation with the scalar and vector self - interaction and derive its two solitary wave solutions . throughout the paper , units in which both the speed of light and the reduced planck constant are equal to one will be used .
our attention is restricted to the nld equation in @xmath0 dimensions which can be written in the covariant form @xmath1/\partial \wbar{{{\mbox{\boldmath \small $ \psi$}}}}=0 , \label{generalnld}\ ] ] where @xmath2 is the spinor with two complex components , @xmath3 denotes the adjoint spinor , @xmath4 is the complex conjugate transpose of @xmath2 , @xmath5 $ ] denotes the self - interaction lagrangian , the greek index @xmath6 runs from 0 to 1 , the einstein summation convection has been applied , @xmath7 is the imaginary unit , @xmath8 is the rest mass , @xmath9 stands for the covariant derivative , and @xmath10 , for @xmath11 , are the gamma matrices or the dirac matrices , chosen as those in @xcite , @xmath12 in fact , eq . is the equation of motion for the classical spinorial particle with the lagrangian being a sum of the dirac lagrangian and the self - interaction lagrangian , _ _
@xmath13=\wbar{{{\mbox{\boldmath \small $ \psi$ } } } } ( \mi { { \mbox{\boldmath \small $ \gamma$}}}^\mu\partial_\mu - m){{\mbox{\boldmath \small $ \psi$}}}+{l}_\text{i}[{\mbox{\boldmath \small $ \psi$ } } ] . \label{total_lag}\ ] ] there exist several different nld models in the literature , where two important models in ( 1 + 1 ) dimensions are the scalar self - interaction of soler @xcite @xmath14:=\wbar{{{\mbox{\boldmath \small $ \psi$}}}}{{\mbox{\boldmath \small $ \psi$}}},\ ] ] and the vector self - interaction of thirring @xcite @xmath15:= \wbar{{{\mbox{\boldmath \small $ \psi$}}}}{{\mbox{\boldmath \small $ \gamma$}}}^\mu{{\mbox{\boldmath \small $ \psi$}}}\wbar{{{\mbox{\boldmath \small $ \psi$}}}}{{\mbox{\boldmath \small $ \gamma$}}}_\mu{{\mbox{\boldmath \small $ \psi$}}},\ ] ] where @xmath16 with the minkowski metric @xmath17 on spacetime , which implies @xmath18 .
this paper will focus on the nld equation with a more general self - interaction @xcite @xmath19 = s \left(l_{\text{s}}[{{\mbox{\boldmath \small $ \psi$}}}]\right)^{k+1 } + v \left(l_{\text{v}}[{{\mbox{\boldmath \small $ \psi$}}}]\right)^{(k+1)/2},\ ] ] and extendedly propose and compare its numerical methods , where @xmath20 and @xmath21 are two real numbers and @xmath22 is a positive real number .
if the spinor @xmath2 is scaled by a constant factor as @xmath23 with @xmath24 , then the scaled self - interaction lagrangian will be @xmath25 $ ] which shows that the power exponent to @xmath26 is @xmath27 . in such sense , we call that the self - interaction lagrangian @xmath28 has the power exponent @xmath27 @xcite
. hereafter the _ quadric _ and _ cubic _ self - interaction will be referred to the case @xmath29 and the case @xmath30 , respectively .
the self - interaction implies the so - called homogeneity relation @xcite @xmath31}}{\partial{\wbar{{{\mbox{\boldmath \small $ \psi$}}}}}}}=(k+1){l}_\text{i}[{{\mbox{\boldmath \small $ \psi$ } } } ] .
\label{ppl = l}\ ] ] combining it with the definition of the lagrangian @xmath32 $ ] and gives @xmath33 = -k { l}_\text{i}[{{\mbox{\boldmath \small $ \psi$}}}].\ ] ] for the nld equation with , one may still verify the following local conservation laws for the current vector @xmath34 and the energy - momentum tensor @xmath35 : @xmath36 where @xmath37.\ ] ] for localized solitary waves @xmath38 , one may derive a direct corollary of , _ _
i.e. _ _ the following global conservation laws @xcite .
[ prop - cl ] assume that @xmath39 and @xmath40 hold uniformly for @xmath41 and @xmath42 .
the total energy @xmath43 , the total linear momentum @xmath44 , and the total charge @xmath45 , defined respectively by @xmath46 satisfy @xmath47 the properties and will be also exploited to find the solitary wave solutions of the ( 1 + 1)-dimensional nld equation with @xmath28 given in in the next subsection .
this subsection will derive the standing wave solutions of the ( 1 + 1)-dimensional nld equation with the self - interaction in the spirit of the technique used in @xcite .
the solution @xmath38 of the nld equation with @xmath28 in , having the form @xmath48 is wanted , where @xmath49 , and @xmath50 and @xmath51 are assumed to decay exponentially as @xmath52 or have the finite compact support .
for such solution , it is not difficult to verify that both the lagrangian @xmath32 $ ] and the energy - momentum tensor @xmath35 are independent of the time @xmath53 , because @xmath54 & \equiv \omega \wbar{{{\mbox{\boldmath \small $ \psi$}}}}{{\mbox{\boldmath \small $ \gamma$}}}^0{{\mbox{\boldmath \small $ \psi$}}}+\mi\wbar{{{\mbox{\boldmath \small $ \psi$}}}}{{\mbox{\boldmath \small $ \gamma$}}}^1{{\mbox{\boldmath \small $ \psi$}}}_x - m\wbar{{{\mbox{\boldmath \small $ \psi$}}}}{{\mbox{\boldmath \small $ \psi$}}}+{l}_\text{i}[{{\mbox{\boldmath \small $ \psi$ } } } ] , \\ t_{00 } & \equiv-\mi \wbar{{{\mbox{\boldmath \small $ \psi$}}}}{{\mbox{\boldmath \small $ \gamma$}}}^1{{\mbox{\boldmath \small $ \psi$}}}_x+m\wbar{{{\mbox{\boldmath \small $ \psi$}}}}{{\mbox{\boldmath \small $ \psi$}}}- { l}_\text{i}[{{\mbox{\boldmath
\small $ \psi$ } } } ] , \quad t_{01}\equiv \mi\wbar{{{\mbox{\boldmath \small $ \psi$}}}}{{\mbox{\boldmath \small $ \gamma$}}}^0{{\mbox{\boldmath \small $ \psi$}}}_x , \\
t_{10 } & \equiv-\omega \wbar{{{\mbox{\boldmath \small $ \psi$}}}}{{\mbox{\boldmath \small $ \gamma$}}}^1{{\mbox{\boldmath \small $ \psi$ } } } , \quad t_{11}\equiv -\mi \wbar{{{\mbox{\boldmath \small $ \psi$}}}}{{\mbox{\boldmath \small $ \gamma$}}}^1{{\mbox{\boldmath \small $ \psi$}}}_x+{l}[{{\mbox{\boldmath \small $ \psi$ } } } ] , \end{aligned}\ ] ] where @xmath55 . using the conservation laws further gives @xmath56 = 0 .
\label{t11}\ ] ] the first equation implies that @xmath57 is imaginary because @xmath58 thus , without loss of generality , we may assume that @xmath59 is real and @xmath60 is imaginary , and they are in the form @xmath61 where both @xmath62 and @xmath63 are pending real functions , and @xmath62 is assumed to satisfy the inequality @xmath64 . on the other hand , combining the first equation in with the second equation in yields @xmath65=0,\ ] ] which becomes for @xmath66 combining with leads to @xmath67 which reduces to for @xmath68 because @xmath69 for @xmath70 when @xmath49 , the solution of may be derived as follows : @xmath71 for initial data @xmath72 and @xmath73 . in fact , under the previous assumption , one may verify @xmath74 . if coefficients @xmath20 and @xmath21 in belong to the set @xmath75 for @xmath76 , or @xmath77 for @xmath78 , then from eq .
one has non - trivial @xmath62 for the localized solution as follows @xmath79 hereto , the standing wave solution of the nld equation with has been derived , and will be denoted as follows @xmath80 where @xmath63 and @xmath62 are given in eqs . and , respectively .
this solution represents a solitary wave with zero velocity and contains some special cases in the literature _
_ e.g. _ _ @xcite .
it has been pointed out in @xcite that the profile of the charge density @xmath81 for the standing wave under the scalar self - interaction ( _ _ i.e. _ _
@xmath82 and @xmath83 ) with @xmath29 can be either one - humped or two - humped , which is also recently confirmed for any @xmath73 by other researchers in @xcite .
they further pointed out there that the profile can only be one - humped for any @xmath73 in the case of @xmath84 and @xmath85 . for the linear combined self - interaction with @xmath82 and @xmath86
, we find that the profile can also be one - humped ( see figs .
[ fig.os-dg-rhoq ] and [ fig : k1-sv ] where the charge density is denoted by @xmath87 for convenience ) or two - humped ( see fig . [
fig : k1 - 2humped - sv ] ) .
this subsection will derive another solution of the ( 1 + 1)-dimensional nld equation with the self - interaction by using the lorentz covariance of the nld equation . consider a frame f with an observer o and coordinates
the observer o describes a particle by the wavefunction @xmath89 which obeys the nld equation with @xmath28 given in , _ _
i.e. _ _ @xmath90/\partial \wbar{{{\mbox{\boldmath \small $ \psi$}}}}\right)(x , t)=0.\ ] ] in another inertial frame f@xmath91 with an observer o@xmath91 and coordinates @xmath92 given by the lorentz transformation with a translation in the @xmath93-direction @xmath94 which is called `` boost '' in the @xmath93direction , where @xmath95 is any given position , @xmath96 is the relative velocity between frames in the @xmath93-direction , and @xmath97 is the lorentz factor . according to the relativity principle , the observer o@xmath91 describes the same particle by @xmath98 which should also satisfy @xmath99/\partial \wbar{{{\mbox{\boldmath \small $ \psi$}}}}'\right)(x',t')=0.\ ] ] using some algebraic manipulations , the `` transformation '' matrix @xmath100 may be found as follows @xmath101 which takes @xmath89 to @xmath98 under the lorentz transformation , _
_ i.e. _ _
@xmath102 where @xmath103 is the sign function which returns @xmath104 if @xmath105 , @xmath106 if @xmath107 , and @xmath108 if @xmath109 . applying the transformation with to the standing wave solution
gives another solution of the nld equation - , _
_ i.e. _ _ the moving wave solution @xmath110 this solution represents a nld solitary wave with velocity @xmath96 and will return to the standing wave if setting @xmath111 and @xmath112 .
this subsection will show that the nld equation with @xmath28 given in remains invariant under the time reversal operation @xmath113 where @xmath114 denotes the complex conjugate operation on @xmath89 , _ _ i.e. _ _
@xmath115 , the time - reversal operator @xmath116 satisfies @xmath117 due to the anticommutation relation @xmath118 , and @xmath119 is the unit matrix . from the relations
, it can be easily verified that @xmath120 so that the self - interaction lagrangian in satisfies @xmath121(x^\prime , t^\prime ) = l_\text{i}[{{\mbox{\boldmath \small $ \psi$}}}](x , t),\ ] ] under the time reversal transformation .
applying the time - reversal operator @xmath116 to the nld equation and using the definition as well as the relations and lead to an equation which has the same form as shown in .
that is , if a spinor @xmath89 satisfies the nld equation , then the transformed spinor @xmath122 by the time reversal operation will also satisfy the same nld equation , _ _
i.e. _ _ the nld equation is time reversible under the operation .
as we mentioned in section [ sec : intro ] , some numerical methods have been well developed for the nld equation with a scalar or vector self - interaction .
this section will extendedly present and compare several numerical methods for solving the ( 1 + 1)-dimensional nld equation with the general scalar and vector self - interaction .
their numerical analyses will be presented in section [ sec : discussion4fdm ] . for the sake of convenience , divide the bounded spatial domain @xmath123 into a uniform partition @xmath124 with a constant stepsize @xmath125 and give a grid in time @xmath126 with a time stepsize @xmath127 , and recast the nld equation into @xmath128 where @xmath129 all of which are real functions , and the dependence of @xmath89 on @xmath88 is implied .
use @xmath130 to denote approximation of @xmath131 and define the forward , backward and centered difference operators in space and time by : @xmath132 where @xmath133 is the identity operator , and @xmath134 and @xmath135 are the translation operators in time and space , respectively , defined by @xmath136 whose inverses exist and are defined by @xmath137 besides , several symbols are also introduced for arithmetic averaging operators : @xmath138 and for an extrapolation operator : @xmath139 * @xmath140 crank - nicolson schemes * the cn scheme and its linearized version have been studied in @xcite for the nld equation with the quadric scalar self - interaction . for the system
, the extension of the cn scheme in @xcite may be written as @xmath141 by approximating at point @xmath142 with compact central difference quotient in place of the partial derivative , where @xmath143 the above cn scheme ( named as cn hereafter ) is fully implicit and forms a nonlinear algebraic system . in practice ,
the linearization technique is often used to overcome difficulty in directly solving such nonlinear algebraic system .
two linearization techniques @xcite for numerical methods of the nonlinear schrdinger equation are borrowed here .
the first linearized cn scheme we consider is using wholly the extrapolation technique to the nonlinear self - interaction terms in @xmath144 which will be called by lcn1 . the second linearized cn scheme , denoted by lcn2 , is @xmath145 by partially applying the extrapolation technique to the nonlinear self - interaction terms .
it is worth noting that the above linearized cn schemes are not linearized version of the cn scheme .
the lcn2 scheme may conserve the charge and behaves better than the lcn1 scheme ( _ vide post _ ) .
for the ( 1 + 1)-dimensional nld equation with the quadric scalar self - interaction lagrangian , the cn scheme ( named by cn0 ) proposed in @xcite is @xmath146 and its linearized version ( called by lcn0 ) is given in @xcite as follows @xmath147 we will show in section [ sec : discussion4fdm ] that the cn , cn0 and lcn0 schemes conserve the charge and the cn scheme further conserves the energy .
* @xmath140 odd - even hopscotch scheme * the odd - even hopscotch scheme is a numerical integration technique for time - dependent partial differential equations , see @xcite .
its key point is that the forward euler - central difference scheme is used for the odd grid points while at the even points the backward euler - central difference scheme is recovered .
thus the odd - even hopscotch scheme may be explicitly implemented .
such scheme applied to the system becomes @xmath148 @xmath149 in the following we will call it by hs . * @xmath140 leapfrog scheme * the leapfrog scheme looks quite similar to the forward scheme , see _
_ e.g. _ _ ,
except it uses the values from the previous time - step instead of the current one . for the system , it is @xmath150 which is a three - level explicit scheme in time with a central difference in space and will be named by lf .
* @xmath140 semi - implicit scheme * another three - level scheme considered here for the system is @xmath151 which is obtained by approximating explicitly the nonlinear terms but implicitly the linear terms and will be called by si .
it is worth noting that such semi - implicit scheme has been studied for the nld equation with the quadric scalar self - interaction in @xcite .
this subsection goes into discussing exponential operator splitting scheme for the nld equation . for convenience ,
we rewrite the system as follows @xmath152 where the linear operator @xmath153 and both nonlinear operators @xmath154 and @xmath155 are defined by @xmath156 then the problem may be decomposed into three subproblems as follows @xmath157 due to the local conservation laws which are discussed in section [ sec : dis - nonlinear ] , solutions of the nonlinear subproblem or may be expressed as an exponential of the operator @xmath154 or @xmath155 acting on `` initial data '' .
thus we may introduce the exponential operator splitting scheme for the the nld equation or , imitating that for the linear partial differential equations , see _
_ e.g. _ _
@xcite and references therein . based on the exact or approximate solvers of those three subproblems , a more general @xmath158-stage @xmath159-th order exponential operator splitting method @xcite for the system can be cast into product of finitely many exponentials
as follows @xmath160 where @xmath161 denotes the time stepsize used within the @xmath162-th stage and satisfies @xmath163 and @xmath164 is any permutation of @xmath165 .
hereafter we call the operator splitting scheme by os(@xmath159 ) .
although one single product of finitely many exponentials exponentials is employed here , it should be pointed out that the linear combination of such finite products can also be used to construct exponential operator splitting schemes as shown in @xcite .
a simple example is the well - known second - order accurate operator splitting method of strang ( named by os(@xmath166 ) ) with @xmath167 another example is the fourth - order accurate operator splitting method @xcite with @xmath168 which is denoted by os(@xmath169 ) in the following .
another operator splitting scheme is studied in @xcite for the nld equation ( [ generalnld ] ) but only with the quadric scalar self - interaction lagrangian , and the second - order accurate strang method is applied there . for the system , it is based on the following operator decomposition @xmath170 with @xmath171 for the linear parabolic equation which is an irreversible system , a more general exponential operator splitting scheme and its accuracy as well as stability are discussed in @xcite , based on linear combinations of products of finitely many exponentials .
it is shown that for such irreversible system , negative weights or negative time stepsizes @xmath161 may lead to instability ; and the highest order of the stable exponential operator splitting approximation ( only with positive weights and positive sub - stepsizes in time ) is two . however , for time - reversible systems , such as the hamilton system , the schrdinger equations and the nld equation with @xmath28 given in , it is immaterial whether or not the weights or time stepsizes @xmath161 are positive @xcite , where a general framework was presented for understanding the structure of the exponential operator splitting schemes and both specific error terms and order conditions were analytically solved .
we are now solving the the linear subproblem . denote its `` initial data '' by @xmath172 at the @xmath162-th stage in .
if the spinor @xmath2 is periodic ( _ _ e.g. _ _
@xmath173-periodic ) with respect to @xmath93 , the fourier spectral method is employed to solve and gives @xmath174 here @xmath175 and @xmath176 denote the discrete fourier transform operator and its inverse , respectively , defined by @xmath177 where @xmath178 is the grid point number , and the matrix exponential in can be easily evaluated as follows @xmath179 with @xmath180 .
when the spinor @xmath2 is not periodic with respect to @xmath93 , the fifth - order accurate finite difference weno scheme will be used to solve the linear subproblem .
the readers are referred to @xcite for details . in this case , the linear subproblem can also be solved by using the characteristics method .
the nonlinear subproblems and are left to be solved now .
their `` initial data '' is still denoted by @xmath181 at the @xmath162-th stage in , and define @xmath182 for nonlinear subproblem , it is not difficulty to verify that @xmath183 so that @xmath184 using this local conservation law gives the solution at @xmath185 of with the `` initial data '' @xmath186 as follows @xmath187 for the nonlinear subproblem , one may still similarly derive the following local conservation laws @xmath188 by direct algebraic manipulations if using the fact that @xmath189 , @xmath190 and @xmath191 are all real . consequently , integrating from @xmath192 to @xmath193 gives its solution as follows @xmath194 where @xmath195 and @xmath196 .
it is because the local conservation laws and are fully exploited here that we can solve exactly the nonlinear subproblems and which imply the more higher accuracy of the os method than that of other methods . in summary
, we have * the cn and cn0 schemes are nonlinear and implicit , and could be solved by iterative algorithms such as picard iteration and newton method .
* the lcn0 , lcn1 , lcn2 and si schemes are linear and implicit . * the hs - , lf , and os(@xmath159 ) schemes are explicit .
before investigating the performance of the numerical methods proposed in section [ sec : diffsch ] , this section will go first into numerical analysis of them , including the accuracy in the sense of the truncation error , time reversibility and the conservation of the charge or energy .
if @xmath197 is periodic , then the cn , cn0 , lcn0 , lcn1 , lcn2 , hs , lf and si schemes are of order @xmath198 , and the os(@xmath159 ) scheme is of order @xmath199 for any arbitrary large @xmath200 .
the proof is very straightforward by using directly the taylor series expansion for the finite difference schemes and the fourier spectral analysis for the os(@xmath159 ) scheme , and thus is skipped here for saving space .
the cn , cn0 , hs , lf , si , and os(@xmath159 ) schemes are time reversible , but the lcn0 , lcn1 , lcn2 schemes are not .
we give the proof for the cn and lcn1 schemes as an example and the others can be proved in a similar way . according to the transformation
, the relation between the transformed finite difference solution and the original one should be @xmath201 with @xmath202 .
consequently , we have @xmath203 and then using the relations in yields @xmath204 for @xmath205 . applying the time - reversal operator @xmath116 to the cn scheme and using the commutation relation and eqs . and lead to @xmath206 which is exactly the cn scheme applied to @xmath207 .
that is , the cn scheme is invariant under the the time - reversal transformation , namely , it is time reversible .
the fact that the lcn1 scheme is not time reversible can be observed directly if noting @xmath208 \\ \neq & \mi \ell_t^\text{e } \left(f^\prime_\text{s}{{\mbox{\boldmath \small $ \gamma$}}}^0 { { \mbox{\boldmath \small $ \psi$}}}^\prime + f^\prime_\text{v } \wbar{{{\mbox{\boldmath \small $ \psi$}}}}^\prime{{\mbox{\boldmath \small $ \gamma$}}}_\mu { { \mbox{\boldmath \small $ \psi$}}}^\prime{{\mbox{\boldmath
\small $ \gamma$}}}^0{{\mbox{\boldmath \small $ \gamma$}}}^\mu { { \mbox{\boldmath \small $ \psi$}}}^\prime \right)_j^{n^\prime-1}.\end{aligned}\ ] ] next , we will discuss the conservation of the discrete energy , linear momentum and charge defined below for the numerical methods given in section [ sec : diffsch ] .
after performing the integration in the computational domain @xmath209 $ ] and then approximating the first derivative operator @xmath210 with the centered difference operator @xmath211 as well as the integral operator @xmath212 with the summation operator @xmath213 in eq . , we have the discrete energy , linear momentum and charge at the @xmath214-th time step @xmath215 where the inner product @xmath216 is defined as @xmath217 for two complex - valued vectors @xmath218 and @xmath219 , and @xmath178 is the grid point number . here
the values of @xmath2 at @xmath220 with @xmath221 are unknowns and those at @xmath95 and @xmath222 are determined by appropriate boundary conditions .
we first present the following lemma which can be verified through direct algebraic manipulations and will be used in discussing the conservation of the discrete charge , energy and linear momentum .
[ le:1 ] given that @xmath223 is a complex - valued vector mesh function with two components evaluated at the mesh @xmath224 ( @xmath225 , @xmath226 ) and a matrix @xmath227 , we have the following identities \(a ) @xmath228 ; \(b ) @xmath229 ; \(c ) @xmath230 ; \(d ) @xmath231 ; \(e ) @xmath232{l } -\delta_t^+\left ( h \sum_{j=1}^j ( \wbar{{{\mbox{\boldmath \small $ \phi$}}}}_j^n{{\mbox{\boldmath \small $ \gamma$}}}^1 \delta_x^0{{\mbox{\boldmath \small $ \phi$}}}_j^n ) \right ) \\
+ \frac12\big ( ( \ell_t^+\wbar{{{\mbox{\boldmath \small $ \phi$}}}}_{j+1}^n{{\mbox{\boldmath \small $ \gamma$}}}^1\delta_t^+{{\mbox{\boldmath \small $ \phi$}}}_j^n + \ell_t^+\wbar{{{\mbox{\boldmath \small $ \phi$}}}}_{j}^n{{\mbox{\boldmath \small $ \gamma$}}}^1\delta_t^+{{\mbox{\boldmath \small $ \phi$}}}_{j+1}^n ) \\
-(\ell_t^+\wbar{{{\mbox{\boldmath \small $ \phi$}}}}_{1}^n{{\mbox{\boldmath \small $ \gamma$}}}^1\delta_t^+{{\mbox{\boldmath \small $ \phi$}}}_0^n + \ell_t^+\wbar{{{\mbox{\boldmath \small $ \phi$}}}}_{0}^n{{\mbox{\boldmath \small $ \gamma$}}}^1\delta_t^+{{\mbox{\boldmath \small $ \phi$}}}_{1}^n ) \big ) .
\end{array}$ ] it can be checked that the following leibniz rules @xmath233 holds for any two spinors @xmath234 and @xmath235 , and then we have @xmath236 thus the identity @xmath237 holds .
because the operator @xmath238 is the adjoint operator of @xmath211 and @xmath239 is an hermite matrix , we get the identity @xmath240 directly by rearranging the summation .
the identity @xmath241 can be easily verified if using the fact @xmath242 .
the proof of @xmath243 ( resp .
@xmath244 ) is similar with that of @xmath240 ( resp .
@xmath241 ) .
[ thm : conservation ] the cn , cn0 , lcn0 , lcn2 , and os(@xmath159 ) schemes conserve the discrete charge , but only the cn scheme conserves the discrete energy . we begin with the discrete conservation law of charge for the cn scheme .
performing the inner product of @xmath245 and the cn scheme leads to @xmath246 and then the conservation law of the discrete charge can be easily verified by taking directly the real part as follows @xmath247 where lemma [ le:1 ] ( a ) is applied to the first term in eq .
, ( b ) to the second term and ( c ) to the third and fourth terms .
similarly , it can be verified that holds for the cn0 , lcn0 and lcn2 schemes .
performing the inner product of the cn scheme and @xmath248 , keeping the imaginary part and applying lemma [ le:1 ] give directly the conservation of the discrete energy @xmath249 for the fourier spectral method , using the fact that @xmath250 in is a unitary matrix yields @xmath251 and then @xmath252\right\|^2 \\ & = \left\|\exp\left ( -\mi \tau_i ( \kappa { { \mbox{\boldmath \small $ \gamma$}}}^0{{\mbox{\boldmath \small $ \gamma$}}}^1+m{{\mbox{\boldmath \small $ \gamma$}}}^0)\right ) \mathcal{f}\left({{\mbox{\boldmath \small $ \psi$}}}^{(0)}\right)\right\|^2 = \left\| \mathcal{f}\left({{\mbox{\boldmath \small $ \psi$}}}^{(0)}\right)\right\|^2 = \left\|{{\mbox{\boldmath \small $ \psi$}}}^{(0)}\right\|^2 , \ ] ] where parseval s identity is applied twice .
it can be readily verified that the matrix exponents in eqs . and are unitary , thus @xmath253 holds both for eqs . and , _
_ i.e. _ _
@xmath254 should be conserved for solutions of the nonlinear subproblems .
therefore , the os(@xmath159 ) scheme satisfies the conservation law of charge .
it will be verified by numerical results in section [ sec : numericalresults ] that the lcn1 , hs , lf and si schemes do not conserve the discrete charge or energy , and none of the numerical methods presented in section [ sec : diffsch ] conserves the discrete linear momentum .
this section will conduct numerical simulations to compare the performance of numerical schemes proposed in section [ sec : diffsch ] and then utilize the os(@xmath169 ) scheme to investigate the interaction dynamics for the nld solitary waves under the scalar and vector self - interaction . for those localized nld solitary waves , the periodic boundary condition for the os(@xmath159 ) scheme and the non - reflection boundary condition for other schemes
could be adopted at the boundaries of the computational domain if a relatively large computational domain has been taken in our numerical experiments .
all calculations are performed on a lenovo desktop computer with intel core i5 650 cpu and 4 gb ram using double precision in the 3.0.0 - 24-generic x86_64 linux operation system and the compiler is gcc 4.6.1 .
the computational domain @xmath255 will be taken as @xmath256 $ ] in examples [ eg.1]-[eg.scalar - vector ] and @xmath257 $ ] in example [ eg.scalar-2humped ] . and the particle mass @xmath8 in eq .
is chosen to be @xmath104 .
[ eg.1 ] the first example is devoted to comparing the numerical performance of all the numerical methods in section [ sec : diffsch ] in terms of the accuracy , the conservativeness , the efficiency and the error growth .
a one - humped solitary wave with the velocity @xmath258 is simulated here under the quadric scalar self - interaction ( _ _ i.e. _ _
@xmath83 and @xmath29 ) , traveling from right to left with the parameters in : @xmath259 , @xmath260 , and @xmath261 .
the @xmath262-rkdg method @xcite is also included here for comparison , which is assembled with a fourth - order accurate runge - kutta time discretization in time and the legendre polynomials of degree at most @xmath159 as local basis functions in the spatial galerkin approximation . .example [ eg.1 ] .
part i : numerical comparison of the accuracy , the conservativeness and the efficiency at @xmath263 with the time stepsize being set to @xmath264 .
the cpu time in seconds is recorded for the finest mesh . [ cols="<,>,^,^,^,^,^,^,^,^ " , ] [ eg.scalar ] the collision of two equal one - humped solitary waves under the scalar self - interaction , _ _ i.e. _ _ case b1 in table [ 3caselist ] , is studied in this example .
the interaction dynamics for the quadric case are shown in the left plot of fig .
[ fig : k1-s ] , where two equal waves with the initial amplitude of @xmath265 move close at a velocity of @xmath266 and overlap each other , then separate into a left moving wave and a right moving wave with the amplitude of @xmath267 and the velocity of @xmath268 .
similar phenomena are observed for the cubic case shown in the right plot of fig .
[ fig : k1-s ] except that ( 1 ) two waves overlap more stronger around @xmath269 now due to the stronger nonlinearity ; ( 2 ) after collision , the amplitude decreases to @xmath270 from the initial amplitude of @xmath271 while the velocity also decreases to @xmath272 . in both cases , the discrete charge ,
energy and linear momentum are approximately conserved in the interaction since the variation of them at @xmath273 is under @xmath274-@xmath275 .
[ eg.v ] the collision of two equal one - humped solitary waves under the vector self - interaction , _ _ i.e. _ _ case b2 in table [ 3caselist ] , is studied in this example .
to the best of our knowledge , it is the first time to study binary collision of the nld solitary waves under the vector self - interaction .
the interaction dynamics for the quadric case are shown in the left plot of fig .
[ fig : k1-v ] , where the waves keep the shape and the velocity after the collision .
a totally different phenomenon appears for the cubic vector self - interaction as displayed in the right plot of fig .
[ fig : k1-v ] .
the initial one - humped equal waves first merge into a single wave , then separate and overlap again . around @xmath263
, collapse happens and highly oscillatory waves are generated and moving outside with a big velocity near @xmath104 , meanwhile a one - humped wave with small amplitude is formed at the center . in both cases , the discrete charge ,
energy and linear momentum are approximately conserved in the interaction since the variation of them at @xmath276 is under @xmath277-@xmath278 .
note in passing that the collapse here is different from that shown in @xcite .
it was reported there that the strong negative energy and radiation appear when the collapse happens during the binary collision of two - humped waves .
[ eg.scalar-vector ] this example is devoted into investigating for the first time the collision of two equal nld solitary waves under the scalar and vector self - interaction , _
_ i.e. _ _ case b3 in table [ 3caselist ] .
the interaction dynamics for the quadric case are shown in the left plot of fig . [
fig : k1-sv ] , where two equal waves with the initial amplitude of @xmath279 move close at a velocity of @xmath266 and overlap each other , then separate into a left moving wave and a right moving wave with the amplitude of @xmath280 and the velocity of @xmath281 .
the collapse similar to that shown in right plot of fig .
[ fig : k1-v ] happens again for the cubic vector self - interaction , see the right plot of fig .
[ fig : k1-sv ] .
the initial one - humped equal waves first merge into a single wave at @xmath282 , then separate and overlap again . around @xmath263
, collapse happens and highly oscillatory waves are generated and moving outside with a big velocity near @xmath104 . in both cases , the discrete charge , energy and linear momentum are approximately conserved in the interaction since the variation of them at @xmath273 is under @xmath283-@xmath275 .
[ eg.scalar-2humped ] as reported before in @xcite , collapse happens in binary and ternary collisions of the nld solitary waves under the quadric scalar self - interaction if the two - humped waves are evolved . in this example
, we will show further that collapse could happen in binary collision of equal two - humped waves under the cubic scalar self - interaction and under the linear combination of scalar and vector self - interactions .
first , case b4 in table [ 3caselist ] is studied and the interaction dynamics are shown in fig .
[ fig : k1 - 2humped ] , which clearly shows that ( 1 ) collapse happens in both quadric and cubic cases but is more stronger in the latter ; ( 2 ) two initial waves at the same velocity are decomposed into groups with different velocities after the collision , but there is no such decomposition for the cubic case . in the left plot of fig .
[ fig : k1 - 2humped ] , the highly oscillating waves with small amplitude move outside at a big velocity of @xmath284 , while the one - humped waves with big amplitude follow them at a small velocity of @xmath285 . in both cases , the discrete charge , energy and linear momentum are approximately conserved in the interaction since the variation of them at @xmath273 is under @xmath286-@xmath287 .
second , binary collision of equal two - humped solitary waves under the scalar and vector self - interaction , _
_ i.e. _ _ case b5 in table [ 3caselist ] , is plotted in fig .
[ fig : k1 - 2humped - sv ] .
the phenomena are very similar to that shown in fig .
[ fig : k1 - 2humped ] , and the decomposition " phenomenon for the quadric case is more obvious than that shown in the left plot of fig .
[ fig : k1 - 2humped ] .
several numerical methods for solving the nld equation with the scalar and vector self - interaction have been presented and compared theoretically and numerically .
our results have revealed that among them , the os(@xmath169 ) scheme , one of the fourth - order accurate os methods , performs best in terms of the accuracy and the efficiency .
particularly , the os(@xmath169 ) scheme is usually more accurate than the @xmath288-rkdg method in the mesh of the same size , but the former needs much more less computational cost than the latter .
such superior performance of the os methods is credited to the full use of the local conservation laws of the nld equation such that the nonlinear subproblems resulted from them are exactly solved .
the interaction dynamics for the nld solitary waves under the quadric and cubic self - interaction have been investigated with the os(@xmath169 ) scheme .
we have found that such interaction dynamics depend on the exponent power of the self - interaction .
actually , it has been observed for the first time in our numerical experiments that , ( 1 ) collapse happens in collision of two equal one - humped nld solitary waves under the cubic vector self - interaction but such collapse does not appear for corresponding quadric case ; ( 2 ) two initial waves at the same velocity are decomposed into groups with different velocities after collapse in binary collision of two - humped nld solitary waves under the quadric scalar self - interaction or under the quadric scalar and vector self - interaction but such phenomenon does not show up for corresponding cubic case .
more efforts on the interaction dynamics for the nld solitary waves under more general self - interaction with the os(@xmath169 ) method are still going on .
sihong shao was partially supported by the national natural science foundation of china ( project no .
11101011 ) and the specialized research fund for the doctoral program of higher education ( project no .
20110001120112 ) .
huazhong tang was partially supported by the national natural science foundation of china ( project no .
10925101 ) .
the authors would also like to thank the referees for many useful suggestions . | this paper presents a review of the current state - of - the - art of numerical methods for nonlinear dirac ( nld ) equation .
several methods are extendedly proposed for the ( 1 + 1)-dimensional nld equation with the scalar and vector self - interaction and analyzed in the way of the accuracy and the time reversibility as well as the conservation of the discrete charge , energy and linear momentum .
those methods are the crank - nicolson ( cn ) schemes , the linearized cn schemes , the odd - even hopscotch scheme , the leapfrog scheme , a semi - implicit finite difference scheme , and the exponential operator splitting ( os ) schemes .
the nonlinear subproblems resulted from the os schemes are analytically solved by fully exploiting the local conservation laws of the nld equation .
the effectiveness of the various numerical methods , with special focus on the error growth and the computational cost , is illustrated on two numerical experiments , compared to two high - order accurate runge - kutta discontinuous galerkin methods .
theoretical and numerical comparisons show that the high - order accurate os schemes may compete well with other numerical schemes discussed here in terms of the accuracy and the efficiency .
a fourth - order accurate os scheme is further applied to investigating the interaction dynamics of the nld solitary waves under the scalar and vector self - interaction .
the results show that the interaction dynamics of two nld solitary waves depend on the exponent power of the self - interaction in the nld equation ; collapse happens after collision of two equal one - humped nld solitary waves under the cubic vector self - interaction in contrast to no collapse scattering for corresponding quadric case .
nonlinear dirac equation , solitary wave , interaction dynamics , finite difference method , operator splitting method |
in the physics of high-@xmath0 cuprates , optimizing their @xmath0 remains a fundamental yet still open problem .
empirically , important parameters that control @xmath0 have been identified for the cuprates , i.e. , chemical composition , structural parameters , the number of layers , etc , besides the doping concentration . for the structural parameters
specifically , several key quantities have been suggested : the bond length between copper and in - plane oxygen(@xmath7 , defined in fig.[fig1])@xcite , and the cu - apical oxygen distance ( @xmath8 ) @xcite .
now , the pressure effect is exceptionally valuable as an _ in situ _ way to probe the structure - dependence of @xmath0 . regarding this ,
two general observations have been made for the cuprates : ( i ) @xmath0 tends to be enhanced under hydrostatic pressure , while ( ii ) uniaxial pressures produce anisotropic responses of @xmath0 .
more precisely , ( i ) @xmath0 has been shown to monotonically increase for pressure @xmath9 gpa @xcite .
as for ( ii ) , an @xmath10-axis compression generally raises @xmath0 ( @xmath1 ) , while @xmath11-axis compression has an opposite effect ( @xmath2)@xcite . moreover , the magnitude of the pressure coefficient becomes smaller for materials having higher @xmath0 s , as summarized in fig.3 of ref.@xcite .
the purpose of the present study is to theoretically reveal the origin of these general trends , focusing on the single - layered cuprates for clarity , and to shed light on a possibility of further optimizing @xmath0 . conventionally , the theoretical model primarily used for the cuprates is a one - band hubbard model based on the @xmath4 orbital ( or sometimes cu-@xmath12o-@xmath13 orbital ) .
recently , we have shown@xcite that the @xmath5 orbital component strongly mixes into the states on the fermi surface in the relatively low-@xmath0 cuprates such as la@xmath14cuo@xmath15 ( la214)@xcite , where the hybridization works destructively against @xmath16-wave superconductivity .
while there have been some theoretical studies in the literature focusing on the role of the @xmath5 orbital@xcite , refs.@xcite conclude that larger the level offset @xmath17 between the @xmath4 and @xmath5 wannier orbitals , higher the @xmath0 , where @xmath17 is governed by the apical - oxygen height and the inter - layer distance .
one might then presume that the effects of uniaxial pressures can simply be captured in terms of the pressure - dependence of @xmath17 affected by the crystal field .
however , we reveal in the present work that the physics is not so simple .
we find that , while the variation of @xmath0 under pressure is indeed affected by @xmath17 , especially in the relatively low-@xmath0 cuprates , the large @xmath17 values in higher-@xmath0 cuprates such as hgba@xmath14cuo@xmath15 ( hg1201 ) make their relevance to the @xmath0 variation smaller .
we shall show that we have to turn our attention rather to the cu@xmath6 level , which is raised with pressure , resulting in a less rounded ( i.e. , better nested ) fermi surface .
this , along with the increase in the band width , is shown to cause a higher @xmath0 under pressure .
these results can be unified into a picture in which higher @xmath0 can be achieved by the `` distillation '' of the main ( i.e. , @xmath4 ) band , namely , the smaller the hybridization of other orbital components the better .
our theoretical procedure is as follows .
we first determine the lattice structure under uniaxial and hydrostatic pressures from a first - principles band calculation with the wien2k package@xcite . from the band structure ,
we construct the maximally - localized wannier orbitals@xcite to obtain the hopping integrals for a two - orbital tight - binding model that takes into account both the @xmath4 and the @xmath5 wannier orbitals explicitly@xcite . in this two - orbital model , we consider the onsite intra- and inter - orbital electron - electron repulsive interactions , which are given , in the standard notation , as @xmath18 where @xmath19 denote the sites while @xmath20 the two orbitals , the electron - electron interactions comprise the intraorbital repulsion @xmath21 , interorbital repulsion @xmath22 , and the hund s coupling @xmath23(= pair - hopping interaction @xmath24 ) . here
we take @xmath25 ev , @xmath26 ev , and @xmath23=0.3 ev@xcite .
these values conform to widely accepted , first - principles estimations for the cuprates that @xmath21 is 7 - 10@xmath27 ( with @xmath28 0.45 ev ) , while @xmath29 . here
we also observe the orbital su(2 ) requirement , @xmath30 .
to study the superconductivity in this multi - orbital hubbard model , we apply the fluctuation exchange approximation(flex)@xcite . in flex , we start with dyson equation to obtain the renormalized green s function , which is , in the multi - orbital case , a matrix in the orbital representation as @xmath31 , where @xmath32 and @xmath33 are orbital indices .
the bubble and ladder diagrams constructed from the renormalized green s function are then summed to obtain the spin and charge susceptibilities , @xmath34 @xmath35 where @xmath36 with wave vector @xmath37 and with matsubara frequency @xmath38 , and the irreducible susceptibility is @xmath39 with the interaction matrices @xmath40 @xmath41 with these susceptibilities , the fluctuation - mediated effective interactions are obtained , which are used to calculate the self - energy . then the renormalized green s functions are determined self - consistently from dyson equation .
the green s functions and the susceptibilities are used to obtain the spin - singlet pairing interaction in the form @xmath42 and this is used in the linearized eliashberg equation , @xmath43 the superconducting transition temperature , @xmath0 , corresponds to the temperature at which the maximum eigenvalue @xmath44 of the eliashberg equation reaches unity , so that @xmath44 at a fixed temperature can be used as a measure for @xmath0 .
@xmath0 of the hg cuprate is experimentally about three times higher than in la cuprate@xcite , so we calculate @xmath44 by putting @xmath45 ev for la and @xmath46 ev for hg for a clearer comparison .
as we shall see , the eigenvalues discussed in the present study are away from unity ( i.e. , the temperature is higher than @xmath0 ) due to the limitation in the number of matsubara frequencies and the @xmath47-point meshes .
therefore , for the la cuprate in particular , we restrict ourselves to qualitative argument for the @xmath0 variation under pressure .
for the hg cuprate , on the other hand , we can go down to lower temperatures @xmath48 where the eigenvalue approaches unity , and we have checked that the conclusions drawn from the @xmath46 calculation hold also for @xmath49 .
moreover , we estimate @xmath50 for hg with the @xmath49 results , as will be discussed in the final part of the paper .
we fix the total band filling ( number of electrons/ site ) at @xmath51 , for which the filling of the main band amounts to 0.85 ( 15 % hole doping ) .
we take a @xmath52 @xmath47-point meshes for the three - dimensional lattice with @xmath53 matsubara frequencies .
.structural and electronic parameters obtained from the first - principles(th ) and experiments(exp ) in ref@xcite . [ cols="^,^,^,^,^,^",options="header " , ] orbital level with respect to that for @xmath4 under uniaxial pressure ( top left inset ) .
bottom right : the shift of cu@xmath6 level under hydrostatic pressure and its effect on the fermi surface .
, width=302 ]
let us begin with the case of uniaxial pressure .
we first vary the lattice constants and calculate the total energy @xmath54 .
this is fit by the standard burch - marnaghan equation@xcite to determine the most stable structure with a unit cell volume @xmath55 , the @xmath10-axis lattice constant @xmath56 , and the @xmath11-axis @xmath57 . for simplicity
, we retain the tetragonal symmetry throughout , i.e. , @xmath58 ( so that the @xmath10-compression is actually biaxial ) .
we show in table i the lattice parameters , @xmath59 , @xmath60 , @xmath8 , @xmath61(la or ba height measured from cuo@xmath62 plane ) and @xmath63 , obtained for the la and hg cuprates .
the results are in good agreement with experimental values for the optimally doped compounds@xcite .
we then relax the structure perpendicular to the compression direction , namely , we allow the lattice constant in that direction to relax to obtain the value that gives the lowest energy . of the eliashberg equation
is plotted against @xmath64 ( a ) or @xmath65 ( b )
. triangles ( circles ) indicate the result for the la ( hg ) cuprates .
arrows depict the contributions ( see text ) to the @xmath44 variation from @xmath17 , @xmath66 , and @xmath67 , respectively , at @xmath68 .
lines are guide for the eye , with the dashed horizontal ones indicating the original values .
, width=226 ] figure [ fig2 ] plots the eigenvalue of the eliashberg equation @xmath44 against the lattice compression @xmath64 and @xmath65 for each compound .
the result shows that ( i ) in both compounds @xmath44 increases as @xmath64 is reduced , while decreases as @xmath65 is reduced , and ( ii ) the absolute value of the variations of @xmath44 is larger in la than in hg .
these features are in qualitative agreement with the experimental observations summarized in fig.3 of ref.@xcite , which shows @xmath1 and @xmath2 for both materials , with larger @xmath69 in la than in hg@xcite . to be more precise ,
while the compressibility in the @xmath10 direction is nearly the same between the two materials , that in the @xmath11 direction is about three times larger in hg than in la@xcite @xmath70 , but even if we take this into account , we find that @xmath71 is still larger for la than for hg in our calculation .
now we want to pinpoint the origin of this @xmath0 variation against uniaxial pressures . in both materials , @xmath72 increases as @xmath64
is reduced , while it decreases when @xmath65 is reduced .
this is natural , since the @xmath10 @xmath73 reduction pushes the in - plane ( out - of - plane ) ligands toward cu , resulting in a larger ( smaller ) crystal - field splitting and hence larger ( smaller ) @xmath17@xcite , as schematically depicted in fig.[fig1 ] .
one might then expect that this alone is the origin of the @xmath0 variation , since @xmath17 and @xmath0 are positively correlated@xcite . to see if this is indeed the case ,
we have considered a case where we increase @xmath17 alone to its value at @xmath74 or @xmath75 , and obtain @xmath44 with flex .
the results are indicated in fig.[fig2 ] with arrows labeled as `` @xmath17 '' . in la ,
the resulting @xmath44 is very close to those obtained for the actual compression , which implies that the main origin for @xmath44 variation under uniaxial pressure comes from @xmath17 .
by contrast , for hg , the @xmath17 contribution is too small to account for the actual @xmath44 variance ( see the blowups in fig.[fig2 ] ) .
the reason for this is that in hg , @xmath17 is @xmath76 times larger than in la ( table [ table1 ] ) , so that the effect of the @xmath5 orbital is tiny , while the contribution to the @xmath0 variation coming from other changes in the electronic structure become comparable with that from @xmath17 .
in particular , we focus on the change in the energy difference @xmath77 between cu@xmath6 and cu@xmath4 orbitals .
in fact , it has been shown that the cu@xmath6 orbital , which is implicitly included in the @xmath4 wannier orbital in the present scheme , affects the second @xmath78 and third @xmath79 neighbor hoppings@xcite .
note that the @xmath6 orbital can be integrated out ( implicitly included in the wannier orbitals ) prior to the many - body analysis , since the @xmath6 orbital sits in energy well away from the fermi level in contrast to the @xmath5 orbital ( fig.[fig1])@xcite .
smaller @xmath77 results in larger @xmath80 within the @xmath4 orbital sector , resulting in a more rounded fermi surface , which degrades @xmath16-wave superconductivity@xcite . to show how the roundness varies with @xmath77 , we consider a three - orbital model which explicitly includes the cu 4s wannier orbitals for the hg cuprate@xcite , and show in fig.[fig3 ] the fermi surface for various values of @xmath81 .
we stress here that , while larger @xmath17 and larger @xmath67 ( or smaller @xmath77 ) both give more rounded fermi surface , their effects on @xmath0 are opposite . under pressure , @xmath77 is enhanced , which in turn reduces @xmath67 . in fig.[fig2 ] , we show the effect of hypothetically reducing @xmath67 down to its values at @xmath74 or @xmath75 . while the effect of @xmath67 is much smaller than that of @xmath17 in la , the two effects are found to be comparable in hg .
hypothetical varied from 6ev ( nearly original value ) to 12ev .
, width=302 ] in addition to @xmath17 and @xmath67 , the band width @xmath66 ( the energy difference between @xmath82 and @xmath83 ) of the main band is also altered by pressure . in la
the change in @xmath44 due to the modification of @xmath66 is small compared to that arising from @xmath17 , but in hg the @xmath66 contribution is comparable with those from @xmath17 and @xmath67 , which in fact provides a full understanding of the net @xmath44 variation under uniaxial pressure .
namely , the @xmath10 @xmath73 reduction results in an increase ( decrease ) of the band width as expected , which enhances ( suppresses ) @xmath0 .
the increase of the band width results in a suppression of @xmath84 , hence the electron correlation effect .
this reduces the pairing interaction , while the self - energy correction due to the spin fluctuations is reduced at the same time .
the former has an effect of enhancing @xmath0 , while the latter suppresses superconductivity . in the case of hg compound
, the latter effect supersedes the former , resulting in an enhanced @xmath0 .
it should be noted that the contribution from @xmath67 , while relatively small for uniaxial compression , enhances @xmath0 for _ both _ of the @xmath10- and @xmath11- axis compressions in marked contrast with the contributions from @xmath17 and @xmath66 . this will become important in our analysis for hydrostatic pressures below .
having identified the ingredients that determine the @xmath0 variation against uniaxial pressures , let us now move on to _ hydrostatic _ compression . here
we optimize the lattice structure at a fixed unit cell volume @xmath85 by varying poisson s ratio , which we fit to the burch - marnaghan equation to obtain the most stable @xmath54 .
notably enough , for hydrostatic pressures @xmath44 in fig.[fig4 ] increases with the volume compression in _ both _ materials .
this result qualitatively agrees with experimental results@xcite . to understand its mechanism , we can ,
as done above for uniaxial pressures , decompose the pressure effect on @xmath44 into the contributions from @xmath17 , @xmath66 , and @xmath67 ( arrows in fig .
[ fig4](a)(b ) ) .
we can then realize that the variation of @xmath17 against hydrostatic pressure is not as straightforward as in uniaxial pressures .
namely , we can look at @xmath17 along with the aspect ratio " @xmath86 against the volume reduction in fig.[fig4](c)(d ) , where @xmath8 is the apical oxygen height and @xmath7 the in - plane cu - o distance . under hydrostatic pressure
, @xmath86 decreases in both materials because of the larger compressibility along the @xmath11 direction . one
might then expect that this would reduce the crystal field splitting and hence @xmath17 , but actually this is by no means always the case . in fact , @xmath17 increases with pressure for la , which is because the cu - o distance decreases , resulting in a larger crystal - field effect .
thus the @xmath0 enhancement in la mainly comes from the increase of @xmath17 .
the above argument for la does not directly apply to hg , since the original apical - oxygen height is larger , so that there is more room for the cuo octahedron to shrink along the @xmath11-axis than in la .
therefore , the @xmath86 reduction is larger , resulting in a nearly constant @xmath17 against the decrease of @xmath87 .
this further makes @xmath17 irrelevant to the @xmath0 variation in hg under hydrostatic pressures . as seen in fig.[fig4 ] with arrows , main contributions to the @xmath0 enhancement
come from @xmath66 and @xmath67 , with similar magnitudes . as shown in fig.[fig4](e)(f ) , the decrease of @xmath67 originally comes from an increase of the level offset @xmath77 introduced above .
the relatively large enhancement of the cu@xmath6 level under hydrostatic pressure can be understood from fig.[fig1 ] ( right ) , where all the ligands approaching cu push up the energy level of the extended and isotropic cu@xmath6 orbital to a larger extent than for the localized and anisotropic cu@xmath88 orbitals .
thus a message here is the _ hydrostatic and uniaxial pressures exert significantly different effects_. specifically , the importance of @xmath67 becomes prominent in hg in hydrostatic pressure because the @xmath67-contribution is positive for both of @xmath10- and @xmath11-axis compressions , while @xmath66-contribution has opposite effects as shown in fig.[fig2 ] .
as for the band - width effect , we have found here that hg exhibits an effect opposite to la for the present electron - electron interaction strength . to elaborate this ,
we have performed a flex calculation for various interaction strengths over @xmath89 , and found that increasing the band width always results in an enhanced @xmath44 in hg within the considered compression range , while in la a similar effect is obtained only for @xmath90 , with the effect reversing for smaller @xmath21 .
we have further noticed that this `` sign change '' in the band - width effect against @xmath21 is peculiar to the systems having smaller @xmath17 . at any rate ,
the band - width effect is much smaller than the effect of @xmath17 in la , so that the effect of pressure - dependence of @xmath21 does not affect the present conclusion .
let us finally comment on the relation between the @xmath44 variation for hydrostatic pressures and the @xmath0 enhancement in the actual pressure experiments . to see this ,
we have extended our calculation to lower temperatures for hg , where @xmath44 becomes closer to unity ( i.e. , @xmath91 approaches @xmath0 ) .
we find @xmath92 at @xmath45 ev for @xmath93 , and the same value of @xmath44 attained at @xmath94 ev for @xmath55 , so the temperature difference ( a rough estimate of @xmath95 ) amounts to @xmath96 k. since the compressibility is @xmath97 gpa@xmath98@xcite , this implies @xmath99 k / gpa , which has the same order of magnitude found experimentally@xcite .
to summarize , we have identified the parameters that govern the @xmath0 variation of the single - layered cuprates under pressure . for lower-@xmath0 materials with small @xmath17 as exemplified by la@xmath14cuo@xmath15 , @xmath0 is sensitive to @xmath17 , which is identified to be the main contribution . for higher-@xmath0 materials with large @xmath17
as exemplified by hgba@xmath14cuo@xmath15 , @xmath0 is rather insensitive to @xmath17 , and important contributions are revealed to come instead from the fermi surface roundness governed by the cu@xmath6 orbital as well as the variation of the band width @xmath66 .
these effects coming from the electronic structure in the multi - orbital systems can be unified into a single picture in which the orbital distillation of the main band results in a higher @xmath0 .
the present study can also shed light on a materials - science avenue for optimizing @xmath0 .
the strategy for enhancing @xmath0 , as conceived here , is : ( 1 ) keep the level offset between the @xmath4 and @xmath5 orbitals large ( ideally , larger than @xmath21 as shown in fig.[fig1 ] left ) , ( 2 ) expand the level offset between the cu@xmath6 and the cu@xmath100 as much as possible this makes the fermi surface more nested ( fig.[fig1 ] right ) , and ( 3 ) tune the band width @xmath66 to a moderate value . in this sense , it is important to keep the distance @xmath8 between apical oxygen and cu atom , and it is also important to decrease the in - plane cu - o bond length @xmath7 . in other words ,
the desired situation for optimizing @xmath0 should have an @xmath101 biaxial chemical pressure which reduces the length @xmath7 from those in existing compounds , with the value of @xmath8 kept high .
this may be coupled to the possibility of the level offset @xmath77 controlled independently of @xmath17 by tuning length @xmath7@xcite .
the numerical calculations were performed at the supercomputer center , issp , university of tokyo . this study has been supported by grants - in - aid for scientific research from jsps ( grants no .
23340095 , ra ; no . 23009446 , hs ;
21008306 , hu ; and no . 22340093 , kk and ha ) .
ra acknowledges financial support from jst - presto .
djs acknowledges support from the center for nanophase material science at oak ridge national laboratory .
jorgensen , d.g .
hinks , o.chmaissem , d.n .
argyiou , j.f .
mitchell , and b. dabrowski , in _ lecture notes in physics _ ,
* 475 * , p.1 ( 1996 ) .
a. bianconi , g. bianconi , s. caprara , d. di castro , h. oyanagi and n. l. saini , j. phys . : condens .
matter * 12 * , 10655 ( 2000 ) ; a. bianconi , s. agrestini , g. bianconi , d. di castro , and n. l. saini , j. alloys compd . *
317 - 318 * , 537 ( 2001 ) ; n. poccia , a. ricci and a. bianconi , adv .
matter phys .
* 2010 * , 261849 ( 2010 ) .
s. maekawa , j. inoue and t. tohyama , in _ the physics and chemistry of oxide superconductors _ , edited by y. iye and h. yasuoka ( springer , berlin , 1992 ) , vol . * 60 * , pp .
105 - 115 .
andersen , a.i liechtenstein , o. jepsen , and f. paulsen , j. phys .
solids * 56 * , 1573 ( 1995 ) .
feiner , j.h .
jefferson and r. raimondi , phys .
lett . * 76 * , 4939 ( 1996 ) .
e. pavarini , i. dasgupta , t. saha - dasgupta , o. jepsen , and o. k. andersen , phys .
lett . * 87 * , 047003 ( 2001 ) . c. weber , k. haule , and g. kotliar , phys .
b * 82 * , 125107(2010 ) . c. weber ,
yee , k. haule and g. kotliar , arxiv:1108.3028 .
t. takimoto , t. hotta and k. ueda , phys .
b * 69 * , 104504 ( 2004 ) .
h. sakakibara , h. usui , k. kuroki , r. arita , and h. aoki , phys . rev
. lett . * 105 * , 057003 ( 2010 ) .
h. sakakibara , h. usui , k. kuroki , r. arita and h. aoki , phys .
b * 85 * , 064501 ( 2012 ) . a .-
klehe , a. k. gangopadhyay , j. diederichs and j. s. schilling physica * 213c * , 266 ( 1993 ) ; * 223c * 121(1994 ) . l. gao , y. y. xue , f. chen , q. xiong , r. l. meng , d. ramirez , c. w. chu , j.h eggert , and h.k .
mao , phys .
b * 50 * , 4260 ( 1994 ) . f. hardy , n. j. hillier , c. meingast , d. colson , y. li , n. barisic , g. yu , x. zhao , m. greven , and j. s. schilling , phys . rev .
lett . * 105 * , 167002 ( 2010 ) .
f. gugenberger , c. meingast , g.roth , k. grube , v. breit , t. weber , h. wuhl , s. uchida , and y. nakamura , phys .
b * 49 * , 13137 ( 1994 ) . c. meingast , a. junod and e. walker , physica c * 272 * , 106 ( 1996 ) .
k. shiraishi , a. oshiyama , n. shima , t. nakayama and h. kamimura , solid state commun .
* 66 * , 629 - 632 ( 1988 ) .
h. kamimura and m. eto , j. phys .
jpn . * 59 * , 3053 ( 1990 ) ; m. eto and h. kamimura , j. phys .
* 60 * , 2311 ( 1991 ) .
a.j . freeman and j. yu , physica b * 150 * , 50 ( 1988 ) .
x. wang , h.t .
dang , and a. j.millis , phys .
b * 84 * , 014530(2011 ) .
l. hozoi , l. siurakshina , p. fulde and j. van den brink , sci . rep . * 1 * , 65 ( 2011 ) .
s. uebelacker and c. honerkamp , phys .
b * 85 * , 155122 ( 2012 ) .
m. mori , g. khaliullin , t. tohyama , and s. maekawa , phys .
101 , 247003 ( 2008 ) p. blaha , k. schwarz , g.k.h .
madsen , d. kvasnicka , and j. luitz , _ wien2k : an augmented plane wave _ + _ local orbitals program for calculating crystal properties _
( vienna university of technology , wien , 2001 ) .
n. marzari and d. vanderbilt , phys .
b * 56 * , 12847 ( 1997 ) ; i. souza , n. marzari and d. vanderbilt , phys .
b * 65 * , 035109 ( 2001 ) .
the wannier functions are generated by the code developed by a. a. mostofi , j. r. yates , n. marzari , i. souza , and d. vanderbilt , ( http://www.wannier.org/ ) .
j. kunes , r. arita , p. wissgott , a. toschi , h. ikeda , and k. held , comp .
. commun . * 181 * , 1888 ( 2010 ) .
bickers , d.j .
scalapino , and s.r .
white , phys .
lett . * 62 * , 961 ( 1989 ) .
t. dahm and l. tewordt , phys .
lett . * 74 * , 793 ( 1995 ) .
k. yada and h. kontani , j. phys .
* 74 * , 2161 ( 2005 ) .
h. eisaki , n. kaneko , d. l. feng , a. damascelli , p. k. mang , k. m. shen , z .- x .
shen , and m. greven , phys .
b * 69 * , 064512(2004 ) . f. birch , phys .
b * 71 * , 809(1947 ) .
j. d. jorgensen , h. -b .
schuttler , d. g. hinks , d. w. capone , k. zhang , and m. b. brodsky and d. j. scalapino , phys .
lett . * 58 * , 1024 ( 1987 ) .
wagner , p.g .
radaelli , d.g .
hinks , j.d .
jorgensen , j.f .
mitchell , b. dabrowski , g.s .
knapp , m.a .
beno , physica c * 210 * , 447 ( 1993 ) .
h. takahashi , h. shaked , b. a. hunter , p. g. radaelli , r. l. hitterman , d. g. hinks , and j. d. jorgensen , phys . rev .
b * 50 * , 3221 ( 1994 ) .
a. m. balagurov , d. v. sheptyakov , v. l. aksenov , e. v. antipov , s. n. putilin , p. g. radaelli and m. marezio , phys .
b * 59 * , 7209 ( 1999 ) . for a review , see d.j .
scalapino in _ handbook of high temperature superconductivity _ , chapter 13 ,
schrieffer and j.s .
brooks ( springer , new york , 2007 ) .
maier , m. jarrell , th .
pruschke , and j. keller , phys .
lett . * 85 * , 1524 ( 2000 ) .
p. r. c. kent , t. saha - dasgupta , o. jepsen , o. k. andersen , a. macridin , t. a. maier , m. jarrell , and t. c. schulthess , phys .
b * 78 * , 035132 ( 2008 ) . | the origin of uniaxial and hydrostatic pressure effects on @xmath0 in the single - layered cuprate superconductors is theoretically explored .
a two - orbital model , derived from first principles and analyzed with the fluctuation exchange approximation gives axial - dependent pressure coefficients , @xmath1 , @xmath2 , with a hydrostatic response @xmath3 for both la214 and hg1201 cuprates , in qualitative agreement with experiments .
physically , this is shown to come from a unified picture in which higher @xmath0 is achieved with an `` orbital distillation '' , namely , the less the @xmath4 main band is hybridized with the @xmath5 and @xmath6 orbitals higher the @xmath0 . some implications for obtaining higher @xmath0 materials
are discussed . |
the study of charged solutions to einstein s equations is certainly an interesting topic since the beginning of relativity , and the possibility for an astrophysical object to possess an appreciable amount of electric charge has been debated for a long time @xcite .
anisotropic charged solutions were studied in @xcite , where relevance of charge effects is highlighted , and , under self similarity assumption , in @xcite .
many numerical studies on the subject exist
see e.g. @xcite even outside the realm of relativistic fluids , such as charged scalar fields @xcite . in particular , charged collapsing object have attracted researchers attention , for different reasons .
on one side , if charge addition prevents the collapsing solution to become singular , reissner
nordstrm regular interiors could be modeled consequently @xcite . on the other side ,
if a singularity forms , it is extremely intriguing to investigate the charge effects to the causal behavior of the solution , in particular if the singularity is completely hidden inside a trapped region or a naked singularity exists , in the latter case providing counterexamples to cosmic censorship conjecture . in @xcite
the charged spherical dust cloud collapse was studied to find that shell focusing singularities are avoided in general , obtaining either shell crossing singularities or completely regular solutions , depending on mass to charge ratio . in the same paper
it was conjectured that shell crossing effects could be averted if spatial gradient of pressure remains nonzero . in the present paper
we determine analytically a class of charged spherical exact solutions undergoing gravitational collapse , and study their final state .
these solutions are in general anisotropic , but contain the charged dust model studied in @xcite as a limit case .
we find that shell focusing singularity formation is avoided for non central shells even for this wider class of solutions , and the solution either remains completely regular , or shell crossing singularity is developed .
the paper is organized as follows .
section [ sec : model ] is devoted to introducing the model in particular , in subsection [ sec : addq ] the procedure of charge addition to a spherically symmetric solution is sketched , while the class of ( uncharged ) background models is introduced in subsection [ sec : ns ] see also the appendix for the full list of conditions satisfied by these models .
the new ( charged ) solutions are derived in section [ sec : chargedsol ] and their relevant properties are studied in section [ sec : sing ] , whereas conclusions are left for the final section .
in this section we introduce relativistic charged models in general , and recall the class of model that we will perturb in next section using an electromagnetic charge .
we start from a spherical material described by the line element , written in comoving coordinates @xmath0 , @xmath1 carrying a non - vanishing electric charge density @xmath2 .
we introduce the skew - symmetrc maxwell tensor @xmath3 and , due to spherical symmetry and the absence of magnetic monopole @xcite , the only nonzero component is @xmath4 . from the maxwell s equations @xmath5}=0,\ ] ] being @xmath6 a timelike unit vector field , we find that @xmath7 where @xmath8 is arbitrary and is related to the charge density by the relation @xmath9 let us consider einstein equations @xmath10 for the line element , where @xmath11 is the sum of the contributions of the stress energy tensor induced by the material structure and the stress energy electromagnetic tensor @xmath12 that using is found to satisfy @xmath13 and therefore is diagonal when written as a @xmath14tensor .
the above procedure allows to add a charge to a given spherical model . in the following we are going to describe the spherical model that will be perturbed with an electromagnetic charge following the above scheme .
the model considered here has been found and fully described in @xcite , but for sake of completeness we briefly recall here the basic properties .
we consider a source of gravitational field given by an elastic material under isothermal conditions .
the stress energy tensor in comoving coordinates is diagonal , @xmath15 where @xmath16 is the energy density and @xmath17 and @xmath18 are the radial and the tangential pressure , respectively .
as described in @xcite , the additional information to close the system of einstein equations are supplied by a constitutive function @xmath19 where @xmath20 . note that @xmath21 is a function of the spatial coordinates and of the space
space part of the metric , not depending on the angular coordinates @xmath22 and @xmath23 for symmetry reasons .
the quantities in are related to @xmath21 by the identities @xmath24 where @xmath25 is the matter density and @xmath26 is an arbitrary function .
this framework suggests , as a more convenient way to describe the metric , the use of area
radius coordinates @xmath27 , in such a way that the line element becomes @xcite @xmath28 in @xcite it is described how , if we choose a @xmath29linear constitutive function , @xmath30 determined by the choice of two free functions @xmath31 and @xmath32 constrained at most to fulfil some reasonability conditions that we will review in a moment then einstein field equations for the metric takes the form @xmath33 where @xmath34 it is quite straightforward to verify that the arbitrary function @xmath35 is the misner sharp mass @xmath36 , whereas the @xmath37 is related to the comoving description of the matter through the relation @xmath38 . also observe that @xmath39 as one can find translating the line element into . using and ,
the components of @xmath40 as functions of @xmath41 are given by @xmath42 the above class contains , as particular cases , tolman
lemaitre dust models corresponding to @xmath43 and @xmath44 and vanishing radial pressure models taking @xmath43 and letting @xmath37 possibly depending on @xmath45 too . in @xcite
some particular cases separately studied in literature including the above cited are reviewed .
as said before , in order to describe reasonably a collapsing matter , some conditions must be placed on the arbitrary functions @xmath31 and @xmath32 . using the notation from (
* definition 2 ) , we suppose that they are chosen in such a way that is a _ collapsing area - radius separable ( ars ) spacetime _ see appendix [ sec : ars ] for a complete list of assumptions .
this implies , among the others , that w.e.c .
is satisfied and that @xmath35 and @xmath37 are positive and regular functions whose taylor development in a neighborhood of @xmath46 is given by @xmath47 note also that in view of the definition given in @xcite it must be @xmath48 and @xmath49 .
finally , in order to obtain a global model , a matching with an external space will be performed at @xmath50 , requiring that the fundamental forms of the two metrics at @xmath51 coincide ( israel darmois junction conditions ) . from
we observe that the radial pressure @xmath17 in general does not vanish along @xmath51 , which as well known is a necessary and sufficient condition to match the solution with a schwarzschild exterior . in this ore general case ,
a natural choice for the exterior metric is given by the generalized vaidya spacetime @xcite @xmath52 where @xmath53 is an arbitrary ( non negative ) function .
nordstrm metric is contained here as a particular case ( @xmath54 , with @xmath55 constants ) .
the immersion of @xmath51 in the two spacetimes can be parameterized respectively by @xmath56 and @xmath57 , and junction conditions are found to be @xmath58
let us consider the spherical model above recalled and let us add a charge density as described in section [ sec : addq ] .
coupling between elasticity and charge gives rise to a field theory where lagrangian density of the elastic material @xcite @xmath59 where @xmath60 denotes the matter density , @xmath61 sums up with the maxwell lagrangian density @xmath62 .
then using belinfante rosenfeld theorem the total stress energy tensor @xmath63 is now given by the sum @xmath64 , that in view of and is diagonal when expressed in comoving coordinates .
moreover , the perturbation given by depends on @xmath65 only .
this suggests the idea that the new model may belong to the class considered in section [ sec : ns ] . in particular , if we search for new functions @xmath66 , @xmath67 such that the new stress energy tensor obeys relations and for these new functions , it is easy to check that we are led to the following definition : @xmath68 in other words , the information about the charge effect can be completely encoded into the mass and the @xmath37 function in such a way that the solution obtained represent a _ fictitious _ material belonging to the same class , though with a _
new _ choice of the arbitrary functions .
note that this procedure is possibile since charge in is added to the _ off - shell _ relations and where @xmath69 is not explicitly depending on @xmath65 .
of course , the new degree of freedom introduced with the charge modifies the constitutive relation between the elements of the total stress energy tensor , and indeed now we have @xmath70 where @xmath71 is given by . at this stage , however , a remark is in place : although @xmath35 and @xmath37 satisfy all the properties sketched in section [ sec : ns ] and listed in the appendix especially regularity and in view of and one can not expect that the same properties are satisfied by @xmath66 and @xmath67 .
put it in another way , choosing the two arbitrary functions as in and above , the charged model obtained is _ not _ an ars collapsing spacetime in the sense of ( * ? ? ?
* definition 2 ) , and then results from there do not work here .
for instance , we can not exclude shell crossing singularity formation that on the contrary is likely to occur as we will see later .
nevertheless it is easy to see that the new model automatically satisfies the weak energy condition if the old does .
moreover , if we require regularity of the energy density in the center , recalling that @xmath72 at initial time , the condition @xmath73 as @xmath74 is needed , as it can be seen using together with the facts that @xmath75 and @xmath76 at initial time . assuming regularity of the charge function @xmath77 as done for the other arbitrary functions , this amounts to say that @xmath78 , with @xmath79 .
moreover , we have to assume that local isotropy at the center continues to hold once the charge is added . in comoving
coordinates , using , that would mean that @xmath80 supposing that the local isotropy holds for the uncharged model , this means that @xmath81 whenever we tend to the regular center .
recalling again the initial data @xmath72 , that implies that @xmath82 , and therefore the condition stated above on the behavior of the charge function must be strengthened as follows @xmath83
in view of the lack of regularity at the center see and one can not expect that these solutions collapse completely to a shell focusing singularity . to begin , an important preliminary result can be immediately stated .
[ thm : noncentral ] shell focusing singularity can not form from the collapse of non central shells of the charged elastic models described in section [ sec : chargedsol ] , unless a shell crossing singularity forms at an earlier comoving time . we will actually show that , for each @xmath84 , the function @xmath85 must vanish for some @xmath860,r_0[$ ] . since the equation @xmath87 controls shell crossing singularity formation , we will deduce that shell focusing non central singularities can not form before possibile shell crossing singularities . from , and
recalling from that @xmath32 is strictly positive , the zeroes of @xmath67 corresponds to zeroes of @xmath88 .
fixed @xmath89 , this quantity negatively diverges as @xmath90 , which means that the continuous function @xmath91 , that starts positive for @xmath92 , must vanish for some positive @xmath93 . in view of the above result
, non central shells either develop a shell crossing singularity or they stay regular , collapsing indefinitely to a stable non singular state or even bouncing and finally entering a period of expansion .
shell crossing formation corresponds to the zeroes of the function @xmath37 ( that in comoving coordinates reads as @xmath94 ) , while no singularity formation is possible if @xmath95 vanishes consider restated using @xmath66 and @xmath67 in place of @xmath35 , @xmath37 .
actually , points of the @xmath65 plane where @xmath95 is negative describe a region where the dynamics of the spacetime is not allowed ; taking initial data such that this quantity is positive , and studying the evolution of this quantity for a fixed shell @xmath89 , if the quantity vanishes for some @xmath96 and attains negative values for smaller values of @xmath45 , this means that the shell labeled @xmath97 can not collapse beyond the limit value @xmath96 , then indefinitely tends to it or arrives there in a finite amount of comoving time , bouncing back and entering an expanding phase . in view of the aforesaid , there exists a curve @xmath98 such that , for @xmath84 , @xmath99 is the greatest zero of the function @xmath85 between @xmath100 and @xmath97 ( recall that @xmath101 by the initial data choice ) .
recalling for @xmath67 , we have to determine the zeroes of @xmath102 and using and , it is found explicitly that @xmath103 moreover , to investigate the physics of the collapse , we need to study the apparent horizon , implicitly defined by solution of equation @xmath104 recalling equations and we find the existence of _ two _ different horizon curves satisfying , whose behavior depends on the charge function @xmath77 . in particular , we have the following situation : * when @xmath105 ( and @xmath106 ) : @xmath107 * when @xmath108 : @xmath109 as can be inferred from , when @xmath108 and @xmath110 , no horizon forms . in all other cases , we have two apparent horizons which bound the shell crossing curve @xmath111 one from below and the other from above . as said before , vanishing of the function @xmath112 could prevent shell crosing formation . from the above analysis , there is certainly a case when this happens , that is when no horizon formation takes place .
indeed , in that case it must be @xmath113 throughout the whole evolution , and then on the curve @xmath111 , the quantity under square root in is negative .
therefore , @xmath114 already vanished before @xmath111 which means that the solution remains regular for all times . to gain more information on the behavior of @xmath114 in this case ( @xmath115 ) , one can study the quantity @xmath116 along curves @xmath117 for @xmath1180,1]$ ] , finding that @xmath119 where @xmath120 with @xmath121 the denominator in is positive is the leading term of @xmath122 which is positive by definition , see subsection [ sec : ns ] therefore we have to study the sign of @xmath123 .
in particular we have @xmath124 ( as it is the initial data ) and @xmath125 ( since @xmath110 ) .
thus @xmath126 changes sign for some @xmath1270,1[$ ] , and then the zeroes of @xmath116 defines a curve @xmath128 which bounds @xmath111 from above as expected . where the quantity under square root in changes sign is defined , there exists a region the shaded area below that curve that can not be reached starting the collapse from @xmath129 . ]
the collapse behavior is sketched in figure [ fig:1 ] , where a graphic is plotted with @xmath130 on the vertical axis to get rid of the central coordinate singularity .
the collapse starts from @xmath129 and every shell collapses until it reaches @xmath131 where it can not proceed further since the quantity under square root in would be negative .
then , once the shell radius reaches the value @xmath131 , it starts bouncing back entering an expanding phase . since @xmath132 as @xmath133 ,
even the center remains regular for all times .
[ rem : ak ] it can be seen that the quantity @xmath123 is given by the leading order term of the function @xmath134 , evaluated for @xmath135 .
now let us examine the other cases .
when @xmath108 and @xmath106 we can perform a similar study as before , finding that @xmath136 , with @xmath126 given by above , and then its sign is related to the sign of @xmath123 again .
however , in this case both @xmath137 and @xmath138 are positive , so we have to consider separately the following two subcases : * if @xmath123 admits one root in the interval @xmath1390,1[$ ] , then as before it remains defined a curve @xmath128 which bounds @xmath111 from above , and then the situation is the same as that depicted in figure [ fig:1 ] . *
if @xmath140 in @xmath141 $ ] , then we have to evaluate @xmath116 along @xmath142 . for sake of simplicity
we consider here only the case @xmath143 ( the special case @xmath144 can be discussed with similar methods ) .
it can be found @xmath145 , where @xmath146 that is positive .
evaluating finally @xmath116 on curves @xmath147 , one gets again a behavior of the kind @xmath145 , where @xmath148 and therefore the zeroes of @xmath116 describe a curve @xmath149 whose behavior is given by @xmath150 + lies below the inner horizon , then a shell crossing singularity
denoted by the curve @xmath98 forms .
the shaded area represents the trapped region . ]
+ comparing the above equation with we conclude that @xmath151 for @xmath152 and then in this case a shell crossing singularity develops just after the formation of the outer horizon .
the situation is represented in figure [ fig:2 ] .
finally , case @xmath105 is yet left to be studied .
it is easy to check that zeroes of @xmath116 corresponds to solutions to the equation @xmath153 - 2 q q ' r_u y^2 \psi , _ r = q^2 \left [ { \psi , _ r}^2 - { q'}^2 y^2 \right].\ ] ] from this , using expansions , and and comparing the leading term of the left side member with the leading term of the right side member in the above equation leads to the following behavior for function @xmath154 such that @xmath155 : @xmath156 an immediate comparison with equation shows that @xmath151 for positive @xmath157 and then all shells as before undergo a shell crossing formation , and the situation is again represented in figure [ fig:2 ] .
summarizing , we have that the behavior of the collapse is driven by the parameter @xmath158 defined in in the following way : [ thm : fs ] under the assumptions made in section [ sec : chargedsol ] a shell crossing singularity forms up to the centre when : 1 .
the parameter @xmath158 defined in is greater than 1 , or 2 .
@xmath108 , @xmath159 and the function @xmath123 defined in is positive @xmath160 $ ] . in all other cases , provided
@xmath161 can be treated with similar methods as before , just taking into account higher order terms of the arbitrary functions .
] the solution remains regular , and no singularity forms up to the center .
the class of spherical models carrying an electromagnetic charge considered in this paper has been explicitly derived exploiting the description in terms of a state function @xmath19 see subsection [ sec : ns ] . since the addition of a charge modifies the stress energy tensor with a quantity depending on @xmath65 , it is possible to derive the state function for the new solution , seen as a fictitious material .
therefore , one can say that the process of solution charging is under a mathematical point of view internal to the class of models considered , i.e. those determined by the state function .
although the charged solution does not satisfy all the regularity properties of the original one , however it is possible to explicitly derive the metric and then to study its properties .
as we have seen , the electromagnetic charge deeply changes the final state of the solution .
it is a well established fact indeed that the uncharged solutions here considered always develop a shell focusing singularity that in some case is naked depending on taylor expansion of the arbitrary functions @xmath31 and @xmath32 , see @xcite for full details . instead , here the formation of a shell focusing singularity is impossible for non central shells , and otherwise either the solution bounces back and enters an expanding phase , or a shell crossing singularity takes place . as is well known shell crossing singularities are tipler
gravitationally weak @xcite and the spacetime may admit an extension beyond this kind of singularity , possibly developing a shell focusing singularity at a later comoving time .
this issue still remains an open problem in general , that has been partially solved only for particular cases , as tolman
bondi models @xcite , where metric can be extended in a natural but _ non _ unique and , yet bounded , _ non _ regular way beyond the shell crossing singularity .
the occurence of the two situations either regularity or shell crossing effect is determined ( theorem [ thm : fs ] ) by the taylor development of @xmath35 , @xmath37 and the charge function @xmath77 .
notice that , unlike the charged dust cloud case studied in @xcite , the final state is not determined exclusively by the charge to mass ratio at initial time @xmath162 , since now the mass depends on @xmath45 also and at initial time ( when @xmath72 ) we have @xmath163 , that in general does not allows to determine the coefficient @xmath164 needed in theorem [ thm : fs ] .
however the avoidance of shell focusing singularities up to spacetime extension as said before already found in @xcite remains true for this wider class of ( generally anisotropic ) solutions , giving a first answer ( in the negative ) to the conjecture raised in @xcite that the occurence of shell crossing singularities was determined by the vanishing of the pressure spatial gradient .
it emerges that electromagnetic charge completely changes the final state of the collapsing star in this anisitropic case .
the models introduced in @xcite and recalled in section [ sec : ns ] completely depend on the choice of two functions @xmath35 and @xmath37 .
an _ ars spacetime _ is described by the metric , where holds , and @xmath31 , @xmath32 are two regular functions .
since we consider initial data such that @xmath72 , the spacetime is defined in the set @xmath165 , and @xmath166 defines the junction hypersurface with the external solution see end of section [ sec : model ] above .
now we give a complete list of the properties that these functions must satisfy to obtain what in @xcite is termed as a _ collapsing _ ars spacetime . *
the arbitrary functions must be positive up to possible singularity formation , @xmath167 and @xmath168 , when @xmath169 ; * the weak energy condition ( w.e.c . )
must hold .
a sufficient condition is given by @xmath170 * regularity and isotropy of the metric at the center of symmetry must be imposed .
it can be seen that this is equivalent to the following conditions : @xmath171 * the initial density is required to be decreasing outwards , which means that @xmath172 * shell crossing singularity formation must be excluded .
this leads to the ( sufficient ) conditions @xmath173 @xmath174 * shell focusing singularities must form in a finite amount of comoving time .
the mathematical conditions to express this requirement are @xmath175 in particular , regularity of @xmath35 and @xmath37 , together with - and allows to write developments and .
moreover in @xcite to ensure the very last condition in , it is assumed that @xmath176 , that in view of means @xmath48 . | a class of spherical collapsing exact solutions with electromagnetic charge is derived .
this class of solutions in general anisotropic contains however as a particular case the charged dust model already known in literature . under some regularity assumptions that in the uncharged case give rise to naked singularities
, it is shown that the process of shell focusing singularities avoidance
already known for the dust collapse @xcite also takes place here , determing shell crossing effects or a completely regular solution .
[ section ] [ teo]proposition [ teo]lemma [ teo]corollary [ teo]definition [ teo]remark [ teo]example |
galaxy clusters in the early universe provide key clues to the relation between structure formation and galaxy evolution . in the local universe ,
galaxy clusters are located in the densest peaks of the dark matter distribution at the intersections of filaments ( e.g. , * ? ? ?
their galaxies have significantly different properties from field galaxies as evidenced by the morphology - density relation and the cluster red sequence ( e.g. , * ? ? ?
* ; * ? ? ?
in this way , cluster formation is closely linked to the large - scale structure and environmental effects on galaxy properties .
the direct observation of protoclusters , which are overdense regions of galaxies in the high - redshift universe , will provide us clues to how these relations are formed . in the local universe ,
clusters contain a large number of passive galaxies , but the fraction of star - forming galaxies in clusters is gradually increasing with redshift ( e.g. , * ? ? ?
* ) . around @xmath14
, some studies reported that galaxies in high - density environments form stars more actively than those in low - density environments ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . beyond @xmath15 , young and star - forming galaxies
appear to be a dominant galaxy population even in the most overdense regions .
these protoclusters are often identified by using lyman break galaxies ( lbgs ) or ly@xmath11 emitters ( laes ) , which enable us to trace large - scale structures at high redshift ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the highest - redshift protocluster discovered to date is at @xmath16 @xcite , and is composed of at least ten galaxies @xcite .
some candidates without spectroscopic confirmation beyond @xmath17 have also been found @xcite .
although the majority of protocluster members are young and star - forming galaxies , a red sequence , composed of bright and red galaxies , is found to appear in protoclusters around @xmath18 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these color differences between protocluster and field galaxies is the result of different galaxy properties , such as age , dust , or metallicity .
for example , stellar mass is a basic and readily observable property that can be used to determine details of the star - formation history ; protocluster galaxies appear to have higher stellar masses than their field counterparts at @xmath18 @xcite .
however , @xcite reported that star formation rate ( sfr ) is similar between protocluster and field galaxies at @xmath15 .
these results suggest that the differences in stellar mass at @xmath18 between protocluster and field galaxies may be attributed to differences in star - formation duration or the formation epoch . besides ordinary galaxies , very rare objects , such as ly@xmath11 blobs ,
submillimeter galaxies ( smgs ) , and active galactic nuclei ( agn ) , are also frequently discovered in high - density environments @xcite .
some contradictory results in the mass - metallicity relation have been revealed among protoclusters at the same redshift @xcite .
furthermore , @xcite have found a large amount of cold gas surrounding a @xmath19 protocluster , which may serve as the reservoir for significant future star formation .
as described above , some distinguishing features have been identified in each protocluster by comparing with field galaxies , but the still relatively small number of protoclusters known ( see the overview table in @xcite ) makes it difficult to determine which are common features .
in addition to the sample size , it should be noted that many previous studies used radio galaxies ( rgs ) or quasars ( qsos ) as probes of protoclusters ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , because the host galaxies of these objects are thought to be embedded in massive dark matter halos .
@xcite showed that rgs tend to reside in high - density environments ; on the other hand , protoclusters are also found in the regions without rgs or qsos ( e.g. , * ? ? ? * ; * ? ? ? * ) . despite the efficiency of searching for protoclusters around rg or qso fields ,
biased protoclusters might be selected because strong radiation from rgs and qsos may provide negative feedback and suppress nearby galaxy formation , especially for low - mass galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
furthermore , the mechanisms triggering agn activity may not be present in all overdense environments and , given that agn activity is a transient phenomenon with , in some cases , relatively short timescales ( e.g. , * ? ? ?
* ) , it is likely that many protoclusters are missed when using such objects as beacons .
in addition to rgs and qsos , smgs have been used as similar probes of overdensities at high redshift ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
. however , only cursory systematic studies of such objects have been performed ( e.g. , * ? ? ?
* ) finding that smgs have a complicated relationship with environment , and , indeed , simulations appear to underscore the complexity of this relationship @xcite .
therefore it is preferable , for any systematic study of such environments , to use a population that minimizes such biases and complexities .
this would allow us to address cluster formation and environmental effects on galaxy evolution based on systematic and less biased samples of protoclusters . here
, we present a systematic survey of protoclusters at @xmath20 based on wide - field imaging and follow - up spectroscopy .
this is a complementary approach to protocluster research compared with previous surveys targeting rg / qso fields using a relatively small field of view ( fov ) .
this survey was performed using the wide ( @xmath21 ) canada - france - hawaii telescope legacy survey ( cfhtls ) deep fields , whose depth and area are more adequate to measuring the overdensity of high - redshift galaxies to identify distant protoclusters than the cfhtls wide fields . although our protocluster candidates were detected by using a less biased method , our method still relies on the presence of ly@xmath11 emission , which will bring another potential bias , in confirming the protocluster members by spectroscopy .
section [ sec : phot ] describes the imaging data and the @xmath1 dropout galaxy sample used in this study . in section [ sec : over ] , we quantify overdensity based on the sky distribution of @xmath1 galaxies , and select the best protocluster candidates by comparing the most overdense regions with expectations from a cosmological simulation . the configuration and results of our follow - up spectroscopy are shown in section [ sec : spec ] . in section [ sec : discuss ] , we discuss the structure and properties of confirmed protoclusters and compare with those of field galaxies .
the conclusions are given in section [ sec : conc ] .
we assume the following cosmological parameters : @xmath22 , and magnitudes are given in the ab system .
we made use of publicly available data from the cfhtls ( t0007 : * ? ? ? * ) , which was obtained with megacam mounted at the prime focus of the cfht .
the deep fields of the cfhtls were used in this study , which consist of four independent fields of about @xmath23 area each ( @xmath21 area in total ) observed in the @xmath24- , @xmath25- , @xmath26- , @xmath27- , and @xmath28-bands .
the field center and limiting magnitudes of each field are summarized in table [ tab : photo ] . the seeing size ( fwhm ) and pixel scale of all the images are @xmath29 and @xmath30 , respectively .
although data at other wavelengths , such as near- or mid - infrared imaging , are available in a part of the cfhtls deep fields , the depth and coverage are significantly different from field to field . therefore , our protocluster search is conducted only based on the optical data to make an uniform survey .
ccccccccccccc d1 & 02:25:59 & @xmath3104:29:40 & 3063 & 28.12 & 28.32 & 27.77 & 27.30 & 26.39 & 17110 & 10416 & 2433 & 148 + d2 & 10:00:28 & @xmath3202:12:30 & 2902 & 28.07 & 28.19 & 27.70 & 27.30 & 26.45 & 14515 & 11160 & 2539 & 231 + d3 & 14:19:27 & @xmath3252:40:56 & 3161 & 28.14 & 28.38 & 27.91 & 27.48 & 26.43 & 21454 & 14896 & 2579 & 232 + d4 &
22:15:31 & @xmath3117:43:56 & 3035 & 27.96 & 28.19 & 27.67 & 27.17 & 26.26 & 10484 & 11288 & 1926 & 188 we created two multi - color catalogs with sextractor ( version 2.8.6 ; * ? ? ?
* ) , in which @xmath27- and @xmath28-band images were used as detection images respectively .
the detection images were first smoothed with gaussian function , then objects were detected by requiring a minimum of three adjacent pixels each above @xmath33 of the sky background rms noise .
then , the magnitudes and several other photometric parameters were measured in the other band images at exactly the same positions and with the same aperture of @xmath34 as in the detection - band images using the `` double image mode . ''
the galactic extinction was removed for each field based on the measurement of @xcite .
the individual catalogues were masked to remove the regions where the detection and photometric measurements of objects may have been seriously affected .
these regions are around bright stars , diffraction and bleed spikes from bright stars .
the regions near the frame edges , whose depth is systematically shallow , were also excluded .
as noted in @xcite , bright stars with @xmath35 produce a large halo , whose radius is @xmath36 .
the total masked regions were @xmath37 of the fov , and the effective areas of our analysis are shown in table [ tab : photo ] . finally , @xmath38330,00420,000 and 230,000270,000 objects were detected down to the @xmath39 limiting magnitudes , defined as the magnitude corresponding to three times the standard deviation in the sky flux measured in empty @xmath34 apertures , of the @xmath27- and @xmath28-bands in each field , respectively . to estimate the detection completeness of each @xmath27- and @xmath28-band image
, we used the iraf task mkobjects to create artificial objects on the original images .
artificial objects , which were given a gaussian profile with a fwhm the same as the seeing size , were randomly distributed on the real image outside of twice the fwhm of the real objects to avoid blending artificial objects with real objects .
we generated 50,000 artificial objects in the @xmath40 range , and extracted them using sextractor with the same parameter set .
this procedure was repeated ten times , and the detection completeness was 7050% at the @xmath39 limiting magnitudes of @xmath27- and @xmath28-bands in each field . at fainter than @xmath39 limiting magnitude , the detection completeness drops sharply to @xmath41 .
we confirmed that the results of our completeness tests are consistent with the values described in the cfhtls data release ( we find @xmath42 completeness at the same magnitude at which @xmath43 is expected ) .
it should be noted that detection completeness at bright magnitudes depends on blending with neighbor objects and we carefully masked out bright objects .
we selected @xmath44 galaxy candidates using the lyman break technique ( @xmath45- , @xmath46- , @xmath47- , and @xmath48-dropout galaxies ) .
the @xmath27-band detection catalog was used for the selection of @xmath45- , @xmath46- , and @xmath47-dropout galaxies , and @xmath48-dropout galaxies were selected from the @xmath28-band detection catalog , based on the following color selection criteria @xcite : @xmath49 where @xmath50 is a @xmath51 limiting magnitude . in redder bands than lyman break ( e.g. @xmath25- and @xmath26-bands for @xmath45-dropout galaxies ) , we only used objects detected with more than @xmath51 significance in order to accurately estimate their color ; on the other hand , @xmath51 limiting magnitude was used as color limit if objects were detected less than @xmath51 significance in bluer bands ( e.g. @xmath24-band for @xmath45-dropout galaxies ) .
the results do not significantly change even if we use 1 or @xmath39 significance as the limiting magnitude .
we estimated the redshift distribution resulting from these criteria by using the population synthesis model code galaxev @xcite and the absorption of intergalactic medium ( igm ) @xcite .
in galaxev , we simulated a large variety of galaxy spectral energy distributions ( seds ) using the padova 1994 simple stellar population model .
we assumed a @xcite initial mass function with lower and upper mass cutoffs @xmath52 and @xmath53 , two metallicities ( 0.2 and 0.02 @xmath54 ) , and two star formation histories ( sfhs ) of constant and instantaneous star formation .
we extracted model spectra with ages between @xmath55 and the age of the universe at that redshift and applied the reddening law of @xcite with @xmath56 between 0.00 and 1.50 .
the model magnitudes were estimated by convolving these simulated seds with the filter transmission curves .
we then added the photometric noise , which is typically @xmath57 and @xmath58 at the @xmath59 and @xmath39 limiting magnitudes , respectively . in this process , we assumed that the magnitude distribution of the simulated galaxies and the observed dropout galaxies were the same .
then , the redshift distributions of the @xmath45- , @xmath46- , @xmath47- , and @xmath48-dropout galaxies are estimated by applying the same color selection criteria of dropout galaxies to these simulated seds ( figure [ fig : select ] ) .
the expected redshift ranges of the dropout selection are @xmath60 , @xmath61 , @xmath62 , and @xmath63 for @xmath45- , @xmath46- , @xmath47- , and @xmath48-dropout galaxies , respectively .
it should be pointed out that these estimates rely on some assumptions ( e.g. , the model of igm absorption , sfh ) .
although we used the igm model of @xcite , other models have been proposed @xcite .
furthermore , @xcite found that igm absorption can vary significantly among different lines of sight , but there is a degeneracy between igm absorption and dust extinction .
however , as long as the properties of the galaxies in overdense regions and in the field are not too different , both populations will be affected in the same manner , implying that we can still search for relative overdensities as a tracer of protoclusters .
the numbers of selected dropout galaxies detected in each field are shown in table [ tab : photo ] .
note that the limiting magnitude of this study corresponds to about @xmath64 , @xmath65 , @xmath66 , and @xmath67 ( where @xmath67 is the characteristic magnitude of the schechter functions fitted the dropout galaxies at each redshift ) for @xmath45- , @xmath46- , @xmath47- , and @xmath48-dropout galaxies , respectively @xcite .
thus , the survey depth reaches at least the typical brightness of dropout galaxies even for @xmath48-dropout galaxies .
we evaluated the contamination rate for these color - selection criteria by comparing the dropout selection regions with the positions of contamination on two - color diagram ( figure [ fig : demo ] ) .
the major sources of the contamination are dwarf stars and old elliptical galaxies , the latter being possible to satisfy the color criteria due to the 4000 / balmer break .
to estimate the contamination rate of dwarf stars , we use the trilegal code @xcite , which can simulate number count and broad - band photometry of stars in any galaxy field .
since this model enables us to set up various structural parameters of thin disc , thick disc , halo , and bulge , we used an exponential disk model with default values of scale length and height , and a chabrier imf was applied .
the galactic latitudes were set to be the same as those of the observations ( @xmath68 ) .
then , photometry of the simulated dwarf stars was calculated for the cfht / megacam s filter set .
next , we simulated old galaxy seds using the galaxev , assuming two relatively high metallicities ( @xmath54 and @xmath69 ) , and extracted model spectra with ages of @xmath70 .
the redshift tracks of old galaxies are away from all dropout selection regions . and , although a few dwarf stars are located only within the @xmath47- and @xmath48-dropout selection regions , the main locus of dwarf stars lie far from these regions . actually , the contamination rate of dwarf stars in the @xmath47-dropout samples is expected to be @xmath71 depending on the galactic latitude , and the contamination rate in @xmath48-dropout samples is @xmath72 in the cfhtls deep fields . based on these simulations of dwarf stars and old galaxies , the dropout selection criteria used in this study
are confirmed to be able to separate high - redshift galaxies from contaminations .
we have estimated the local surface number density by counting the number of dropout galaxies within a fixed aperture in order to determine the overdensity significance quantitatively .
@xcite presented a useful definition of the characteristic radius of protoclusters based on cosmological simulations , which encloses 65% of the mass , based on a combination of @xmath73-body dark matter simulations and semi - analytical galaxy formation models . although their result was based on the three - dimensional distribution of protocluster galaxies , it is still useful guide for constructing a map of the projected local surface number density when searching for protoclusters . according to the characteristic radius of protoclusters having a descendant halo mass of @xmath74 at @xmath5 ,
the radius of @xmath75 was used for @xmath45- , @xmath46- , and @xmath47-dropout galaxies , which corresponds to @xmath76 , @xmath77 , and @xmath78 , respectively .
although the characteristic radius is expected to be larger for protoclusters having more massive descendant masses , these still show a clear overdensity even on @xmath75 scales ; thus , the aperture size of @xmath75 should be effective to find protoclusters with a descendant halo mass of @xmath79 .
however , we still have to consider the projection effects resulting from the large redshift uncertainty of our dropout selection .
we discuss the effectiveness of our protocluster search more quantitatively by using a theoretical model in the following subsection .
it should be noted that protoclusters have a nearly constant physical size at @xmath20 @xcite .
we used a slightly larger radius of @xmath80 ( @xmath81 ) for @xmath48-dropout galaxies in order to reduce the large poisson error resulting from a too small aperture .
the apertures were distributed over the cfhtls deep fields in a grid pattern at intervals of @xmath82 .
we measured the mean and the dispersion , @xmath83 , of the number of galaxies in an aperture over the field . the surface number density in masked regions
was assumed to be the same as the mean surface number density .
apertures in which more than @xmath84 area is masked are not used in the following analysis . using the mean and @xmath83 of the number of dropout galaxies in an aperture , surface number density contours of @xmath45- , @xmath46- , @xmath47- , and
@xmath48-dropout galaxies were calculated and are plotted in figures [ fig : cntr_udrop ] , [ fig : cntr_gdrop ] , [ fig : cntr_rdrop ] , and [ fig : cntr_idrop ] , respectively .
we note that this is the same procedure that was applied to the @xmath48-dropout galaxies in the subaru deep field ( sdf ) to draw their surface number density contour , which led to the discovery of the protocluster at @xmath16 from @xcite . in order to study the effects of incompleteness
, we performed the same overdensity measurement but for @xmath45- , @xmath46- , and @xmath47-dropout galaxies brighter than @xmath85 in @xmath27-band ( for @xmath48-dropout galaxies , it is hard to select a brighter subset due to the small sample size ) . from this analysis , we confirmed that the change of limiting magnitude does not have a significant effect on our protocluster selection .
although we can clearly see some overdense regions in the sky distribution maps presented in figures [ fig : cntr_udrop ] , [ fig : cntr_gdrop ] , [ fig : cntr_rdrop ] , and [ fig : cntr_idrop ] , it is not straightforward to find plausible protocluster candidates since the large redshift uncertainty of dropout technique , which is @xmath86 ( @xmath87 at @xmath88 ) , hampers the identification of clustering structure in three - dimensional space . on one hand , overdensities associated with real protoclusters could be weakened by fore / background galaxies ; on the other hand , chance alignments of the large - scale structure or superpositions of filaments could erroneously enhance the surface overdensity .
therefore , we will make use of predictions from simulations to understand the relation between surface overdensity significance and the probability of finding real protoclusters . to this end
, we used a set of light - cone models constructed by @xcite .
a brief outline of the light - cone models is presented below .
first , the assembly history of the dark matter halos was traced using an @xmath73-body simulation @xcite , in which the length of the simulation box was @xmath89 and the particle mass was @xmath90 .
the distributions of dark matter halos were stored at discrete epochs .
next , the processes of baryonic physics were added to dark matter halos at each epoch using a semi - analytic galaxy formation model @xcite . based on the intrinsic parameters of galaxies predicted by the semi - analytic model , such as stellar mass , star formation history , metallicity , and dust content ,
the photometric properties of simulated galaxies were estimated from the stellar population synthesis models developed by @xcite .
then , these simulated galaxies in boxes at different epochs were projected along the line - of - sight , and intergalactic medium ( igm ) absorption was applied @xcite in order to mimic a pencil - beam survey as described in @xcite .
finally , 24 light - cone models with @xmath91 fov were extracted using these procedures .
cccccccc d1ud01 & 02:24:35.4 & @xmath3104:19:58.9 & d1 & @xmath45-dropout & @xmath92 & 244 & yes + d2ud01 & 10:01:18.6 & @xmath3202:33:20.3 & d2 & @xmath45-dropout & @xmath93 & 182 & no + d3ud01 & 14:18:29.1 & @xmath3252:44:05.3 & d3 & @xmath45-dropout & @xmath94 & 300 & no + d3ud02 & 14:17:52.0 & @xmath3252:53:03.2 & d3 & @xmath45-dropout & @xmath95 & 268 & no + d4ud01 & 22:14:03.4 & @xmath3117:58:43.4 & d4 & @xmath45-dropout & @xmath95 & 157 & yes + d1gd01 & 02:25:36.3 & @xmath3104:15:57.4 & d1 & @xmath46-dropout & @xmath96 & 162 & yes + d1gd02 & 02:25:56.2 & @xmath3104:48:30.4 & d1 & @xmath46-dropout & @xmath92 & 153 & no + d3gd01 & 14:18:28.9 & @xmath3252:57:06.5 & d3 & @xmath46-dropout & @xmath97 & 214 & no + d3gd02 & 14:17:55.6 & @xmath3253:07:37.6 & d3 & @xmath46-dropout & @xmath98 & 201 & no + d4gd01 & 22:16:47.3 & @xmath3117:16:52.7 & d4 & @xmath46-dropout & @xmath99 & 153 & yes + d1rd01 & 02:24:45.3 & @xmath3104:55:56.5 & d1 & @xmath47-dropout & @xmath95 & 40 & yes + d2rd01 & 10:00:14.1 & @xmath3201:44:03.0 & d2 & @xmath47-dropout & @xmath100 & 48 & no + d2rd02 & 09:59:04.6 & @xmath3201:47:27.5 & d2 & @xmath47-dropout & @xmath98 & 64 & no + d3rd01 & 14:19:36.8 & @xmath3252:57:44.6 & d3 & @xmath47-dropout & @xmath98 & 39 & no + d4rd01 & 22:14:58.1 & @xmath3117:58:07.2 & d4 & @xmath47-dropout & @xmath92 & 31 & no + d4rd02 & 22:16:46.1 & @xmath3117:29:16.7 & d4 & @xmath47-dropout & @xmath101 & 31 & yes + d1id01 & 02:27:18.4 & @xmath3104:50:58.9 & d1 & @xmath48-dropout & @xmath102 & 10 & yes + d1id02 & 02:26:19.9 & @xmath3104:51:55.0 & d1 & @xmath48-dropout & @xmath96 & 9 & no + d3id01 & 14:19:14.2 & @xmath3252:55:15.7 & d3 & @xmath48-dropout & @xmath103 & 16 & yes + d3id02 & 14:20:09.3 & @xmath3252:28:17.1 & d3 & @xmath48-dropout & @xmath97 & 10 & no + d4id01 & 22:14:29.6 & @xmath3117:27:25.4 & d4 & @xmath48-dropout & @xmath101 & 7 & no the simulated @xmath45- , @xmath46- , @xmath47- , and @xmath48-dropout galaxy catalogs were made by matching the expected redshift distribution of each dropout galaxy sample ( figure [ fig : select ] ) .
we also applied the same limiting magnitude cut only for the detection band used for the observations to the simulated catalogs ( @xmath48-band for @xmath45- , @xmath46- , @xmath47-dropout galaxies , and @xmath104-band for @xmath48-dropout galaxies , respectively ) .
the average stellar mass of these simulated dropout galaxies is @xmath105 , and the 90% quantile range is @xmath106 .
this is consistent with that observed ( e.g. , * ? ? ?
* ) , implying that simulated dropout galaxies trace similar structures .
based on these simulated catalogs , we calculated local number density maps as in section [ sec : sky ] and selected overdense regions in the same way as in the cfhtls deep fields .
for each overdense region of dropout galaxies , we selected the strongest spike in the redshift distribution , and the dominant dark matter structure was defined by the most massive halo in that redshift spike .
the descendant halos at @xmath5 for each overdense region were then easily identified by tracing the halo merger tree of those halos .
as shown in figure [ fig : prob ] , although there is a large scatter , the significance of the overdensity at high redshift is clearly correlated with the descendant halo mass at @xmath5 , whose probability of no correlation is @xmath107 based on the spearman rank correlation test .
thus , it is possible to select reliable protocluster candidates through the surface overdensity . in this study , we set the criterion of protocluster candidates at @xmath108 overdensity significance in order to obtain a high purity of real protoclusters . based on this criterion , @xmath109 of these candidates of @xmath45-dropout ( @xmath48-dropout ) galaxies
are expected to be in real protoclusters . however , the completeness is very small ( @xmath110 and 10% for @xmath45- and @xmath48-dropout galaxies ) mainly due to the projection effect of the dropout technique .
it should be noted that the average descendant mass of protoclusters with @xmath108 overdensity significance is @xmath111 ( figure [ fig : prob ] ) .
a total of 21 candidates were identified from @xmath112 to @xmath13 ( five , five , six , and five candidates for the maps of @xmath45- , @xmath46- , @xmath47- , and @xmath48-dropout galaxies , respectively ) .
the coordinates and overdensity of the protocluster candidates are listed in table [ tab : target ] .
since these numbers of protocluster candidates are consistent with the model prediction , in which @xmath113 candidates per observed area ( @xmath21 ) are found in each redshift bin from @xmath112 to @xmath13 , most of the candidates are expected to be real protoclusters . to summarize , from the wide field imaging of the cfhtls deep fields ,
we have made a large protocluster sample at @xmath1 without the aid of any special probes , such as qsos , rgs , or smgs ; thus , this sample is not only large but also complementary with previous studies targeting qso , rg , or smgs fields .
despite our calibration of the selection of the protocluster candidates using light - cone projections described in the previous section , the overdense regions discovered could still be attributed to mere chance of alignments along the line - of - sight , given that the dropout technique samples a broad range of redshifts .
another possibility is that the overdense significance could be affected by the presence of highly clustered contaminating populations .
this possibility is negligible for the @xmath45- , @xmath46- , and @xmath47-dropout samples because of their high number density .
however , for the @xmath48-dropout galaxies , the average number density per aperture is only 1.6 due to the shallow @xmath28-band depth . since the contamination rate of @xmath48-dropout galaxies is @xmath110 ( section [ sec : select ] ) , the number of contaminants in an aperture is @xmath114 on average , which could result in an overdensity @xmath115 higher at worst .
therefore , further confirmation of clustering in redshift space is required to see whether our candidates are real or not . in addition to protocluster confirmation , spectroscopic observations
enables us to inquire into the internal structure or line properties of protoclusters .
it is necessary for revealing cluster formation to take various viewpoints .
ccccccccc 2012 may 13 & 14 & gmos & d3id01 & r600 & 4.5 & 7500 - 10000 & 330 & 1 & @xmath116 + 2012 oct .
21 & focas & d1id01 & vph900 & 5.7 & 7500 - 10100 & 220 & 1 & @xmath117 + 2014 aug .
24 & deimos & d1gd01 & 600zd & 3.5 & 5000 - 9300 & 120 & 1 & @xmath118 + & & d4gd01 & 600zd & 3.5 & 5000 - 9300 & 120 & 1 & @xmath118 + 2014 oct .
20 & 21 & focas & d1rd01 & vph650 & 5.5 & 6000 - 8300 & 280 & 1 & @xmath118 + & & d1gd01 & vph520 & 2.5 & 4900 - 6500 & 100 & 1 & @xmath117 + & & d1ud01 & vph520 & 2.5 & 4300 - 5900 & 60 & 4 & @xmath119 + & & d4gd01 & vph520 & 2.5 & 4900 - 6500 & 120 & 2 & @xmath118 + & & d4ud01 & vph520 & 2.5 & 4300 - 5900 & 60 & 3 & @xmath120 + 2014 oct .
24 & 25 & focas & d4rd02 & vph650 & 5.5 & 6000 - 8300 & 120 & 1 & @xmath120 before performing the follow - up spectroscopic observations , we first investigated how far protocluster members are typically spread from the center , again using the light - cone model . in the model ,
protocluster members are defined as galaxies whose descendants at @xmath5 reside in @xmath121 halos @xcite .
the center of a protocluster in three - dimensional space was estimated by using the median r.a .
, decl . , and redshift of all protocluster members .
the positional difference between the protocluster center defined as above and the peak of the surface overdensity observed , is typically less than @xmath122 and less than @xmath123 at worst .
then , we investigated the three - dimensional distribution of protocluster members in the overdense regions .
although each protocluster has a different structural morphology , such as filamentary or sheet - like , we simply estimated the probability of protocluster membership as a function of the distance to the center by calculating the ratio between protocluster members and non - members at a certain distance from the center .
we finally derive a probability map by taking the median stack of the probability maps of all the protocluster regions computed for each redshift .
figure [ fig:3d_model ] shows the probability map of protocluster members of @xmath45- , @xmath46- , @xmath47- , and @xmath48-dropout galaxies .
we found that galaxies lying within the volume of @xmath124 and @xmath125 at @xmath126 , will be protocluster members with a probability of @xmath127 .
based on this estimate , we defined the protocluster region as a sphere of @xmath128 radius .
it should be noted that we will evaluate protocluster existence based on significance of excess from homogeneous distribution rather than the absolute number of confirmed galaxies , because actual observations are incomplete and rely on ly@xmath11 emission to identify the redshifts of dropout galaxies .
the fraction of ly@xmath11 emitting galaxies among dropout galaxies has been investigated ( e.g. , * ? ? ?
* ; * ? ? ?
* ) ; however , it would yet to be explored for the fraction in overdense regions because the previous studies are mainly based on field galaxies .
@xcite found that a protocluster around a radio galaxy at @xmath129 exhibits a high overdensity of both laes and lbgs , while @xcite reported that there is no correlation between the distributions of laes and lbgs around a qso at @xmath130 .
hence , it is not yet fully understood what fraction of lbgs emits ly@xmath11 , especially in overdense regions .
although this may potentially bias protocluster identification , most protoclusters at @xmath18 tend to have high overdensities of both lbgs and laes ( e.g. , * ? ? ?
we carried out spectroscopic observations using subaru / focas @xcite , keckii / deimos @xcite , and gemini - n / gmos @xcite . in these observations , eight protocluster candidates from @xmath112 to @xmath13
were observed in total ( two at each redshift ) .
the target protocluster candidates and the configuration of spectroscopic observations are described in table [ tab : target ] and [ tab : obs_spec ] , respectively .
all these observations were conducted with multi - object spectroscopy ( mos ) mode .
the slits typically had a length of @xmath131 and a width of @xmath132 .
the used grisms were selected in order to have the highest efficiency at the wavelength of the redshifted ly@xmath11 line of targeted dropout galaxies and the spectral resolution of @xmath133})\,\mathrm{\aa}$ ] , where @xmath134 is the wavelength separation of the [ oii ] doublet ( @xmath135 ) in the rest - frame .
therefore , our spectroscopic observations were set up to have a resolution that is sufficient to resolve the [ oii ] emission into the doublet to check for contamination by foreground interlopers .
the wavelength coverage is also wide enough to cover the expected redshift range of the dropout galaxies . in the focas observations ,
the telescope was dithered along the slit to enable more accurate sky subtraction between exposures , and we used nod - and - shuffle mode , which allows increase in the accuracy of sky subtraction by real - time flipping to the sky position in the gmos observation .
although higher priority was given to brighter galaxies , we designed slit masks so as to allocate as many objects as possible .
furthermore , a slit of each mask was allocated for a bright star ( @xmath136 ) to monitor the time variations of seeing size or atmospheric transmission between exposures .
although the sky condition was good and stable during all observing nights , we removed only a few poor - quality frames by checking the bright star .
long slit exposures of one of the following spectroscopic standard stars hz44 , bd+28d4211 , and g191-b2b were taken each night with all configurations used in the night , and we corrected the difference of airmass between science targets and standard stars in the flux calibration .
the data taken by focas and gmos were reduced in a standard manner with iraf , and the pipeline spec2d was used for the reduction of the data taken by deimos .
the @xmath39 detection limits of emission line are typically @xmath137 , @xmath138 , @xmath139 , and @xmath140 for @xmath45- , @xmath46- , @xmath47- , and @xmath48-dropout galaxies , respectively , assuming the line width of @xmath141 .
all emission lines we detected are single emission lines , which are not likely to be h@xmath142 or [ ] emission lines because the wavelength coverage of our observation is wide enough to detect all these multiple lines simultaneously . only [ ]
@xmath143 emission , which is generally the strongest emission among them , might resemble a single emission line if the other lines are too faint to be detected .
however , according to the typical line flux ratio of [ ] @xmath143 and [ ] @xmath144 ( @xmath145 ) , even [ ] @xmath144 should be detected in our spectroscopic observations since the signal - to - noise ratio ( s / n ) of detected emission lines is @xmath146 on average ( and always @xmath147 ) .
therefore , we investigated the possibility that h@xmath11 and [ ] emission lines contaminate dropout galaxy samples based on both imaging and spectroscopic data .
only @xmath47- and @xmath48-dropout galaxies can be contaminated by h@xmath11 emitting galaxies according to its wavelength ( @xmath148 ) .
since higher - redshift dropout galaxies are selected using a redder color criterion to detect the strong lyman break , it is almost impossible to mimic this color by a balmer break object at lower redshift based on the expected color of passive galaxies as described in section [ sec : select ] .
even if balmer breaks of passive and old galaxies were strong enough to satisfy the dropout color criteria , these stellar population are unlikely to have h@xmath11 emission , as a diagnostic of the star - formation activity .
on the other hand , dusty starburst galaxies with strong balmer breaks could be contaminated because they would show an h@xmath11 emission line as well as a very red color , resulting from the combination of balmer break and dust reddening .
however , their h@xmath11 emissions can be discriminated from ly@xmath11 emissions , which appear right at the lyman break .
therefore , we consider the possibility of finding foreground h@xmath11 emission negligible . regarding [ ] doublet emission lines , it is possible to distinguish between ly@xmath11 and [ ] emission lines based on the line profile .
the spectral resolution of most of our spectroscopic observation was set high enough to resolve [ ] emission lines as doublets ( @xmath149 at @xmath150 ) , although it would be practically difficult to resolve these in most cases due to low s / n .
in this case , the [ ] emission line is typically skewed blueward , while the ly@xmath11 emission line from high redshift galaxies is skewed redward .
therefore , the skewness of the line profile allows us to distinguish between ly@xmath11 and [ ] emission lines ; however , it should be noted that [ oii ] emission is sometimes skewed redward when assuming an exotic physical properties of region ( e.g. , election density ) . to improve the way of distinguishing
, @xcite introduced `` weighted '' skewness , which makes use of line width as well as line profile .
ly@xmath11 emission usually has a larger line width than [ ] emission because ly@xmath11 emission typically emerges in an outflow or galactic wind .
therefore , we calculated the weighted skewness of all spectroscopically detected galaxies .
the asymmetric emission lines with @xmath151 are clear evidence of ly@xmath11 emission from high redshift galaxies , though it would be more difficult to distinguish them from nearby emission line galaxies at @xmath112 , where the igm attenuation is weaker than at higher redshifts . as shown in table [
tab : spec_all ] , most of the emission lines of this study have large @xmath152 .
however , 17% of all identified emission lines have @xmath153 ; although this would be caused by strong sky line residuals and low s / n data , we could not rule out the possibility of [ ] emission lines . in order to make the line profile measurement more accurate by reducing the effect of sky noise and low s / n data
, we made a composite spectrum of all 24 emission lines with @xmath153 by taking a median in the rest - frame , assuming they were ly@xmath11 , and normalized by the peak flux .
the weighted skewness of the composite spectrum was found to be @xmath154 , indicating that most of the emission lines even with @xmath153 in individual spectra are real ly@xmath11 emission lines from high - redshift galaxies .
in addition to the line profile , the possibility of [ ] emission can further be reduced by taking account of photometric data .
although [ ] emission is closer to balmer break than h@xmath11 emission , it will still be difficult to find a sharp break near the emission line except for peculiar galaxies such as dusty starburst galaxies . from these considerations ,
it is unlikely that h@xmath11 or [ ] emission lines contaminate our dropout samples , and ly@xmath11 is the most plausible interpretation to explain both photometric and spectral features . since the major contamination in the photometric selection is completely different from that in the spectroscopic observation , the combination of photometric and spectroscopic observations enables us to select a clean sample of high - redshift galaxies .
we can regard all single emission lines detected from our dropout sample as ly@xmath11 emission lines .
our protocluster confirmation completely depends on the detection of ly@xmath11 emission ; therefore , we might only select a part of the galaxy populations in these protoclusters .
we might miss protoclusters , if they were mainly composed of passive or dusty galaxies without ly@xmath11 emission .
this may lead to a possible selection bias in our protocluster search .
however , protoclusters at @xmath18 have been found to mainly contain star - forming galaxies .
it is worth noting that known protoclusters that include a large number of older or dustier galaxies , like the ssa22 ( @xmath155 ) and the spider - web ( @xmath156 ) protoclusters , also show a significant overdensity in laes as well @xcite .
these results suggest that the possible bias introduced by tracing protoclusters only by laes is not probably significant .
the number of dropout galaxies located in each protocluster candidate region is listed in table [ tab : target ] , and the number of spectroscopically observed galaxies is shown in table [ tab : cluster ] . from these numbers , about half , at least 35% , of dropout galaxies in protocluster candidate regions were observed by our follow - up spectroscopic observations .
we carefully discriminated real emission lines from sky lines or noise by examining both two - dimensional and one - dimensional spectra , and all emission lines identified in this study are shown in figure [ fig : spec_all ] .
we estimate the observed properties of the spectroscopically confirmed galaxies , such as uv absolute magnitude at 1300 in the rest - frame ( @xmath157 ) , ly@xmath11 luminosity ( @xmath158 ) , and rest - frame ly@xmath11 equivalent width ( @xmath159 ) , shown in table [ tab : spec_all ] . the redshifts were derived by the peak wavelength of the ly@xmath11 emission line , assuming the rest wavelength of ly@xmath11 to be 1215.6 .
these measurements could be overestimated if there was a galactic outflow .
when emission lines are located near strong sky lines , the position of the peak could be shifted .
these effects of sky lines and the wavelength resolution are taken into account when estimating the error .
observed line flux , @xmath160 , corresponds to the total amount of the flux within the line profile .
the slit loss was corrected based on the ratio of slit cccccccccc + 1 & 02:27:18.8 & -04:50:08.3 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 & @xmath166 & @xmath167 + 2 & 02:27:21.0 & -04:50:49.3 & @xmath168 & @xmath169 & @xmath170 & @xmath171 & @xmath172 & @xmath173 & @xmath174 + 3 & 02:27:19.0 & -04:53:48.0 & @xmath175 & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 & @xmath181 + + + 1 & 14:19:22.5 & + 52:57:22.5 & @xmath182 & @xmath183 & @xmath184 & @xmath185 & @xmath186 & @xmath187 & @xmath188 + 2 & 14:19:17.2 & + 52:56:14.4 & @xmath189 & @xmath190 & @xmath191 & @xmath192 & @xmath193 & @xmath194 & @xmath195 + + + 1 & 02:24:45.467 & -04:58:52.83 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 & @xmath201 & @xmath202 + 2 & 02:24:45.957 & -04:56:57.69 & @xmath203 & @xmath204 & @xmath205 & @xmath206 & @xmath207 & @xmath208 & @xmath209 + 3 & 02:24:42.586 & -04:58:36.00 & @xmath210 & @xmath211 & @xmath212 & @xmath213 & @xmath214 & @xmath215 & @xmath216 + 4 & 02:24:38.212 & -04:57:15.05 & @xmath217 & @xmath218 & @xmath219 & @xmath220 & @xmath221 & @xmath222 & @xmath223 + 5 & 02:24:45.964 & -04:54:34.80 & @xmath224 & @xmath225 & @xmath226 & @xmath227 & @xmath228 & @xmath229 & @xmath230 + 6 & 02:24:43.757 & -04:54:31.19 & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 + + + 1 & 22:16:46.722 & -17:28:02.00 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & @xmath243 & @xmath244 + 2 & 22:16:39.959 & -17:31:34.58 & @xmath245 & @xmath246 & @xmath247 & @xmath248 & @xmath249 & @xmath250 & @xmath251 + 3 & 22:16:45.765 & -17:29:19.89 & @xmath252 & @xmath253 & @xmath254 & @xmath255 & @xmath256 & @xmath257 & @xmath258 + + + 1 & 02:25:28.536 & -04:17:14.12 & @xmath259 & @xmath260 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & @xmath265 + 2 & 02:25:30.408 & -04:15:56.70 & @xmath266 & @xmath267 & @xmath268 & @xmath269 & @xmath270 & @xmath271 & @xmath272 + 3 & 02:25:32.014 & -04:17:03.56 & @xmath273 & @xmath274 & @xmath275 & @xmath276 & @xmath277 & @xmath278 & @xmath279 + 4 & 02:25:25.565 & -04:17:12.58 & @xmath280 & @xmath281 & @xmath282 & @xmath283 & @xmath284 & @xmath285 & @xmath286 + 5 & 02:26:12.550 & -04:18:41.23 & @xmath287 & @xmath288 & @xmath289 & @xmath290 & @xmath291 & @xmath292 & @xmath293 + 6 & 02:25:17.290 & -04:14:02.23 & @xmath294 & @xmath295 & @xmath296 & @xmath297 & @xmath298 & @xmath299 & @xmath300 + 7 & 02:25:51.739 & -04:14:37.26 & @xmath301 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & @xmath306 & @xmath307 + 8 & 02:25:39.708 & -04:14:20.73 & @xmath308 & @xmath309 & @xmath310 & @xmath311 & @xmath312 & @xmath313 & @xmath314 + 9 & 02:26:11.563 & -04:19:21.65 & @xmath315 & @xmath316 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 + 10 & 02:26:10.246 & -04:18:18.50 & @xmath322 & @xmath323 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 + 11 & 02:25:33.011 & -04:14:45.24 & @xmath329 & @xmath330 & @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath335 + 12 & 02:26:07.202 & -04:17:12.22 & @xmath336 & @xmath337 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & @xmath342 + 13 & 02:25:59.907 & -04:15:45.42 & @xmath343 & @xmath344 & @xmath345 & @xmath346 & @xmath347 & @xmath348 & @xmath349 + 14 & 02:25:57.460 & -04:18:10.27 & @xmath350 & @xmath351 & @xmath352 & @xmath353 & @xmath354 & @xmath355 & @xmath356 + 15 & 02:25:42.923 & -04:15:38.74 & @xmath357 & @xmath358 & @xmath359 & @xmath360 & @xmath361 & @xmath362 & @xmath363 + 16 & 02:25:44.405 & -04:14:11.81 & @xmath364 & @xmath365 & @xmath366 & @xmath367 & @xmath368 & @xmath369 & @xmath370 + 17 & 02:25:34.147 & -04:14:21.23 & @xmath371 & @xmath372 & @xmath373 & @xmath374 & @xmath375 & @xmath376 & @xmath377 + cccccccccc 18 & 02:25:56.529 & -04:17:27.85 & @xmath378 & @xmath379 & @xmath380 & @xmath381 & @xmath382 & @xmath383 & @xmath384 + 19 & 02:25:30.087 & -04:15:15.84 & @xmath385 & @xmath386 & @xmath387 & @xmath388 & @xmath389 & @xmath390 & @xmath391 + 20 & 02:25:49.845 & -04:14:53.42 & @xmath392 & @xmath393 & @xmath394 & @xmath395 & @xmath396 & @xmath397 & @xmath398 + 21 & 02:25:41.772 & -04:16:06.53 & @xmath399 & @xmath400 & @xmath401 & @xmath402 & @xmath403 & @xmath404 & @xmath405 + 22 & 02:25:56.593 & -04:15:15.20 & @xmath406 & @xmath407 & @xmath408 & @xmath409 & @xmath410 & @xmath411 & @xmath412 + 23 & 02:25:39.324 & -04:14:40.82 & @xmath413 & @xmath414 & @xmath415 & @xmath416 & @xmath417 & @xmath418 & @xmath419 + 24 & 02:26:11.555 & -04:17:39.51 & @xmath420 & @xmath421 & @xmath422 & @xmath423 & @xmath424 & @xmath425 & @xmath426 + 25 & 02:25:27.362 & -04:16:43.59 & @xmath427 & @xmath428 & @xmath429 & @xmath430 & @xmath431 & @xmath432 & @xmath433 + 26 & 02:25:25.816 & -04:16:38.94 & @xmath434 & @xmath435 & @xmath436 & @xmath437 & @xmath438 & @xmath439 & @xmath440 + 27 & 02:26:01.853 & -04:14:41.24 & @xmath441 & @xmath442 & @xmath443 & @xmath444 & @xmath445 & @xmath446 & @xmath447 + 28 & 02:25:31.239 & -04:15:49.81 & @xmath448 & @xmath449 & @xmath450 & @xmath451 & @xmath452 & @xmath453 & @xmath454 + 29 & 02:25:35.470 & -04:14:15.68 & @xmath455 & @xmath456 & @xmath457 & @xmath458 & @xmath459 & @xmath460 & @xmath461 + 30 & 02:25:34.433 & -04:15:05.46 & @xmath448 & @xmath462 & @xmath463 & @xmath464 & @xmath465 & @xmath466 & @xmath467 + 31 & 02:25:39.830 & -04:14:53.33 & @xmath468 & @xmath469 & @xmath470 & @xmath471 & @xmath472 & @xmath473 & @xmath474 + 32 & 02:25:11.525 & -04:16:20.17 & @xmath475 & @xmath476 & @xmath477 & @xmath478 & @xmath479 & @xmath480 & @xmath481 + 33 & 02:25:57.659 & -04:14:24.90 & @xmath482 & @xmath483 & @xmath484 & @xmath485 & @xmath486 & @xmath487 & @xmath488 + 34 & 02:25:21.473 & -04:16:01.50 & @xmath489 & @xmath490 & @xmath491 & @xmath492 & @xmath493 & @xmath494 & @xmath495 + 35 & 02:25:08.716 & -04:15:24.74 & @xmath496 & @xmath497 & @xmath498 & @xmath499 & @xmath500 & @xmath501 & @xmath502 + 36 & 02:25:28.097 & -04:14:54.46 & @xmath503 & @xmath504 & @xmath505 & @xmath506 & @xmath507 & @xmath508 & @xmath509 + + + 1 & 22:16:55.191 & -17:25:51.91 & @xmath510 & @xmath511 & @xmath512 & @xmath513 & @xmath514 & @xmath515 & @xmath516 + 2 & 22:17:00.190 & -17:25:06.37 & @xmath517 & @xmath518 & @xmath519 & @xmath520 & @xmath521 & @xmath522 & @xmath523 + 3 & 22:16:55.670 & -17:20:49.98 & @xmath524 & @xmath525 & @xmath526 & @xmath527 & @xmath528 & @xmath529 & @xmath530 + 4 & 22:16:49.872 & -17:21:53.02 & @xmath531 & @xmath532 & @xmath533 & @xmath534 & @xmath535 & @xmath536 & @xmath537 + 5 & 22:17:01.326 & -17:20:52.55 & @xmath538 & @xmath539 & @xmath540 & @xmath541 & @xmath542 & @xmath543 & @xmath544 + 6 & 22:16:54.811 & -17:28:37.91 & @xmath545 & @xmath546 & @xmath547 & @xmath548 & @xmath549 & @xmath550 & @xmath551 + 7 & 22:16:58.872 & -17:28:33.27 & @xmath552 & @xmath553 & @xmath554 & @xmath555 & @xmath556 & @xmath557 & @xmath558 + 8 & 22:17:07.296 & -17:28:45.15 & @xmath559 & @xmath560 & @xmath561 & @xmath562 & @xmath563 & @xmath564 & @xmath565 + 9 & 22:16:51.756 & -17:24:57.97 & @xmath566 & @xmath567 & @xmath568 & @xmath569 & @xmath570 & @xmath571 & @xmath572 + 10 & 22:16:42.993 & -17:15:53.36 & @xmath573 & @xmath574 & @xmath575 & @xmath576 & @xmath577 & @xmath578 & @xmath579 + 11 & 22:16:50.981 & -17:18:49.87 & @xmath580 & @xmath581 & @xmath582 & @xmath583 & @xmath584 & @xmath585 & @xmath586 + 12 & 22:16:53.509 & -17:19:06.60 & @xmath587 & @xmath588 & @xmath589 & @xmath590 & @xmath591 & @xmath592 & @xmath593 + 13 & 22:16:49.716 & -17:17:00.96 & @xmath594 & @xmath588 & @xmath595 & @xmath596 & @xmath597 & @xmath598 & @xmath599 + 14 & 22:16:53.576 & -17:19:07.20 & @xmath600 & @xmath588 & @xmath601 & @xmath602 & @xmath603 & @xmath604 & @xmath605 + 15 & 22:16:51.410 & -17:17:50.44 & @xmath606 & @xmath607 & @xmath608 & @xmath609 & @xmath610 & @xmath611 & @xmath612 + 16 & 22:16:54.326 & -17:18:34.98 & @xmath613 & @xmath614 & @xmath615 & @xmath616 & @xmath617 & @xmath618 & @xmath619 + 17 & 22:16:57.890 & -17:21:51.88 & @xmath620 & @xmath614 & @xmath621 & @xmath622 & @xmath623 & @xmath624 & @xmath625 + 18 & 22:16:51.591 & -17:18:12.00 & @xmath626 & @xmath627 & @xmath561 & @xmath628 & @xmath629 & @xmath630 & @xmath631 + cccccccccc 19 & 22:16:55.554 & -17:20:14.08 & @xmath632 & @xmath627 & @xmath633 & @xmath634 & @xmath635 & @xmath636 & @xmath637 + 20 & 22:16:48.909 & -17:15:31.09 & @xmath638 & @xmath639 & @xmath640 & @xmath641 & @xmath642 & @xmath643 & @xmath644 + 21 & 22:16:55.005 & -17:21:00.75 & @xmath645 & @xmath281 & @xmath646 & @xmath647 & @xmath648 & @xmath649 & @xmath650 + 22 & 22:16:46.962 & -17:21:06.42 & @xmath651 & @xmath652 & @xmath345 & @xmath653 & @xmath654 & @xmath655 & @xmath656 + 23 & 22:16:46.961 & -17:17:10.24 & @xmath657 & @xmath658 & @xmath659 & @xmath660 & @xmath661 & @xmath662 & @xmath663 + 24 & 22:16:42.903 & -17:17:35.09 & @xmath664 & @xmath665 & @xmath533 & @xmath666 & @xmath667 & @xmath668 & @xmath669 + 25 & 22:16:50.522 & -17:18:22.62 & @xmath670 & @xmath671 & @xmath672 & @xmath673 & @xmath674 & @xmath675 & @xmath676 + 26 & 22:16:49.533 & -17:16:44.13 & @xmath677 & @xmath678 & @xmath679 & @xmath680 & @xmath681 & @xmath682 & @xmath683 + 27 & 22:17:09.126 & -17:28:52.31 & @xmath336 & @xmath684 & @xmath685 & @xmath686 & @xmath687 & @xmath688 & @xmath689 + 28 & 22:16:56.467 & -17:17:20.11 & @xmath690 & @xmath691 & @xmath415 & @xmath692 & @xmath693 & @xmath694 & @xmath695 + 29 & 22:17:01.475 & -17:23:58.97 & @xmath696 & @xmath697 & @xmath698 & @xmath699 & @xmath700 & @xmath701 & @xmath702 + 30 & 22:17:04.114 & -17:29:22.86 & @xmath703 & @xmath704 & @xmath705 & @xmath706 & @xmath707 & @xmath708 & @xmath709 + 31 & 22:16:59.785 & -17:26:15.22 & @xmath710 & @xmath711 & @xmath712 & @xmath713 & @xmath396 & @xmath714 & @xmath715 + 32 & 22:17:00.167 & -17:27:32.72 & @xmath716 & @xmath717 & @xmath718 & @xmath719 & @xmath720 & @xmath721 & @xmath722 + 33 & 22:16:44.680 & -17:17:48.48 & @xmath651 & @xmath723 & @xmath724 & @xmath725 & @xmath726 & @xmath727 & @xmath728 + 34 & 22:16:49.846 & -17:17:16.49 & @xmath729 & @xmath730 & @xmath731 & @xmath732 & @xmath733 & @xmath734 & @xmath735 + 35 & 22:16:51.997 & -17:26:10.95 & @xmath385 & @xmath736 & @xmath737 & @xmath738 & @xmath739 & @xmath740 & @xmath741 + 36 & 22:16:53.458 & -17:20:03.45 & @xmath742 & @xmath743 & @xmath744 & @xmath745 & @xmath746 & @xmath747 & @xmath748 + 37 & 22:16:52.593 & -17:29:00.63 & @xmath749 & @xmath750 & @xmath751 & @xmath752 & @xmath753 & @xmath754 & @xmath755 + 38 & 22:16:59.778 & -17:22:16.93 & @xmath756 & @xmath757 & @xmath758 & @xmath759 & @xmath760 & @xmath761 & @xmath762 + 39 & 22:17:03.102 & -17:25:52.33 & @xmath763 & @xmath764 & @xmath765 & @xmath766 & @xmath767 & @xmath768 & @xmath769 + 40 & 22:16:48.708 & -17:15:41.17 & @xmath729 & @xmath770 & @xmath771 & @xmath772 & @xmath773 & @xmath774 & @xmath775 + 41 & 22:16:49.635 & -17:15:26.63 & @xmath776 & @xmath777 & @xmath778 & @xmath779 & @xmath780 & @xmath781 & @xmath782 + 42 & 22:16:56.050 & -17:24:57.12 & @xmath783 & @xmath784 & @xmath785 & @xmath786 & @xmath787 & @xmath788 & @xmath789 + + + 1 & 02:24:33.775 & -04:22:05.64 & @xmath790 & @xmath791 & @xmath792 & @xmath793 & @xmath794 & @xmath795 & @xmath796 + 2 & 02:24:24.047 & -04:19:30.14 & @xmath797 & @xmath798 & @xmath799 & @xmath800 & @xmath801 & @xmath802 & @xmath803 + 3 & 02:24:38.501 & -04:19:31.91 & @xmath804 & @xmath805 & @xmath806 & @xmath807 & @xmath808 & @xmath809 & @xmath810 + 4 & 02:24:32.251 & -04:20:05.64 & @xmath811 & @xmath812 & @xmath813 & @xmath814 & @xmath815 & @xmath816 & @xmath817 + 5 & 02:24:30.247 & -04:20:25.53 & @xmath818 & @xmath819 & @xmath820 & @xmath821 & @xmath822 & @xmath823 & @xmath824 + 6 & 02:24:35.414 & -04:20:32.25 & @xmath825 & @xmath826 & @xmath827 & @xmath828 & @xmath829 & @xmath830 & @xmath831 + 7 & 02:24:32.181 & -04:18:52.41 & @xmath832 & @xmath833 & @xmath834 & @xmath835 & @xmath836 & @xmath837 & @xmath838 + 8 & 02:24:26.931 & -04:18:09.40 & @xmath839 & @xmath840 & @xmath841 & @xmath842 & @xmath843 & @xmath844 & @xmath845 + 9 & 02:24:32.111 & -04:19:01.04 & @xmath846 & @xmath847 & @xmath848 & @xmath849 & @xmath850 & @xmath851 & @xmath852 + 10 & 02:24:32.361 & -04:18:33.93 & @xmath853 & @xmath854 & @xmath855 & @xmath856 & @xmath857 & @xmath858 & @xmath859 + 11 & 02:24:38.052 & -04:17:50.69 & @xmath860 & @xmath861 & @xmath862 & @xmath863 & @xmath864 & @xmath865 & @xmath866 + 12 & 02:24:36.424 & -04:20:40.01 & @xmath867 & @xmath868 & @xmath869 & @xmath870 & @xmath871 & @xmath872 & @xmath873 + 13 & 02:24:39.007 & -04:17:25.43 & @xmath874 & @xmath875 & @xmath876 & @xmath877 & @xmath878 & @xmath879 & @xmath880 + 14 & 02:24:35.609 & -04:19:31.99 & @xmath853 & @xmath881 & @xmath882 & @xmath883 & @xmath884 & @xmath885 & @xmath886 + cccccccccc 15 & 02:24:36.250 & -04:19:11.89 & @xmath887 & @xmath888 & @xmath889 & @xmath890 & @xmath891 & @xmath892 & @xmath893 + 16 & 02:24:36.988 & -04:18:09.47 & @xmath894 & @xmath895 & @xmath896 & @xmath897 & @xmath898 & @xmath899 & @xmath900 + 17 & 02:24:27.725 & -04:17:48.50 & @xmath901 & @xmath902 & @xmath903 & @xmath904 & @xmath905 & @xmath906 & @xmath907 + 18 & 02:24:35.157 & -04:17:00.64 & @xmath908 & @xmath909 & @xmath910 & @xmath911 & @xmath912 & @xmath913 & @xmath914 + 19 & 02:24:28.399 & -04:20:01.41 & @xmath915 & @xmath916 & @xmath917 & @xmath918 & @xmath919 & @xmath920 & @xmath921 + 20 & 02:24:38.367 & -04:17:16.11 & @xmath922 & @xmath923 & @xmath924 & @xmath925 & @xmath926 & @xmath927 & @xmath928 + 21 & 02:24:41.996 & -04:18:59.15 & @xmath929 & @xmath930 & @xmath931 & @xmath932 & @xmath933 & @xmath934 & @xmath935 + 22 & 02:24:37.488 & -04:19:20.22 & @xmath936 & @xmath937 & @xmath938 & @xmath939 & @xmath940 & @xmath941 & @xmath942 + 23 & 02:24:28.416 & -04:21:30.12 & @xmath943 & @xmath260 & @xmath944 & @xmath945 & @xmath946 & @xmath947 & @xmath948 + 24 & 02:24:36.548 & -04:18:26.31 & @xmath949 & @xmath950 & @xmath951 & @xmath952 & @xmath953 & @xmath954 & @xmath955 + 25 & 02:24:35.602 & -04:16:54.03 & @xmath956 & @xmath957 & @xmath958 & @xmath959 & @xmath960 & @xmath961 & @xmath962 + 26 & 02:24:44.626 & -04:19:35.65 & @xmath963 & @xmath964 & @xmath965 & @xmath966 & @xmath967 & @xmath968 & @xmath969 + 27 & 02:24:38.037 & -04:22:12.46 & @xmath970 & @xmath971 & @xmath972 & @xmath973 & @xmath974 & @xmath975 & @xmath976 + 28 & 02:24:29.531 & -04:21:43.03 & @xmath977 & @xmath978 & @xmath979 & @xmath980 & @xmath981 & @xmath982 & @xmath983 + 29 & 02:24:35.608 & -04:21:10.87 & @xmath984 & @xmath985 & @xmath986 & @xmath987 & @xmath988 & @xmath989 & @xmath990 + 30 & 02:24:24.653 & -04:19:31.71 & @xmath991 & @xmath992 & @xmath993 & @xmath994 & @xmath549 & @xmath995 & @xmath996 + + + 1 & 22:14:03.642 & -18:00:09.90 & @xmath908 & @xmath997 & @xmath998 & @xmath999 & @xmath1000 & @xmath1001 & @xmath1002 + 2 & 22:13:58.570 & -17:59:30.91 & @xmath1003 & @xmath1004 & @xmath1005 & @xmath1006 & @xmath1007 & @xmath1008 & @xmath1009 + 3 & 22:14:11.265 & -17:59:56.51 & @xmath1010 & @xmath1011 & @xmath1012 & @xmath1013 & @xmath1014 & @xmath1015 & @xmath1016 + 4 & 22:13:53.488 & -17:56:54.74 & @xmath949 & @xmath1017 & @xmath1018 & @xmath1019 & @xmath1020 & @xmath1021 & @xmath1022 + 5 & 22:13:51.171 & -17:57:18.84 & @xmath1023 & @xmath1024 & @xmath1025 & @xmath1026 & @xmath1027 & @xmath1028 & @xmath1029 + 6 & 22:13:53.773 & -17:57:40.48 & @xmath1030 & @xmath1031 & @xmath1032 & @xmath1033 & @xmath556 & @xmath1034 & @xmath1035 + 7 & 22:13:54.597 & -17:59:06.04 & @xmath1036 & @xmath1037 & @xmath1038 & @xmath1039 & @xmath1040 & @xmath1041 & @xmath1042 + 8 & 22:13:55.114 & -17:59:55.62 & @xmath1043 & @xmath1044 & @xmath1045 & @xmath1046 & @xmath1047 & @xmath1048 & @xmath1049 + 9 & 22:14:04.835 & -17:57:44.20 & @xmath1050 & @xmath1051 & @xmath1052 & @xmath1053 & @xmath1054 & @xmath1055 & @xmath1056 + 10 & 22:14:04.154 & -18:00:05.58 & @xmath1057 & @xmath1051 & @xmath931 & @xmath1058 & @xmath1059 & @xmath1060 & @xmath1061 + 11 & 22:14:03.430 & -17:59:22.71 & @xmath1062 & @xmath1063 & @xmath1064 & @xmath1065 & @xmath1066 & @xmath1067 & @xmath1068 + 12 & 22:14:09.396 & -17:57:58.56 & @xmath1069 & @xmath1070 & @xmath1071 & @xmath1072 & @xmath1073 & @xmath1074 & @xmath1075 + 13 & 22:14:16.330 & -17:57:22.22 & @xmath1076 & @xmath1077 & @xmath1078 & @xmath1079 & @xmath1080 & @xmath1081 & @xmath1082 + 14 & 22:14:09.371 & -17:58:15.61 & @xmath1083 & @xmath1077 & @xmath1084 & @xmath1085 & @xmath843 & @xmath1086 & @xmath1087 + 15 & 22:13:58.117 & -17:59:46.75 & @xmath1088 & @xmath1089 & @xmath1090 & @xmath1091 & @xmath1092 & @xmath1093 & @xmath1094 + 16 & 22:14:07.173 & -18:00:24.05 & @xmath1095 & @xmath1096 & @xmath1097 & @xmath1098 & @xmath1099 & @xmath1100 & @xmath1101 + ccccccc d1id01 & 8 & 3 & unclear & & & + d3id01 & 8 & 2 & possible & 2 & 5.75 & + d1rd01 & 15 & 6 & possible & 2 & 4.89 & + d4rd02 & 12 & 3 & unclear & & & + d1gd01 & 123 & 36 & no & & & + d4gd01 & 144 & 42 & yes & 11 & 3.67 & @xmath1102 + d1ud01 & 95 & 30 & yes & 5 & 3.13 & @xmath1103 + d4ud01 & 57 & 16 & yes & 5 & 3.24 & @xmath1104 width and seeing size for each observation , and its flux error was estimated from the combination of the line width and the noise level per 1 at wavelengths blueward of ly@xmath11 . since continuum flux was too faint to be detected in the observed spectra , @xmath157 was estimated from the broad - band magnitude ( @xmath25- , @xmath26- , @xmath27- , and @xmath28-bands for @xmath45- , @xmath46- , @xmath47- , and @xmath48-dropout galaxies ) .
the @xmath157 is defined as the absolute magnitude at 1300a in the rest - frame .
it is derived from the broad - band photometry after correcting the contribution of igm absorption and the ly@xmath11 emission to the broad - band photometry based on the spectroscopic . in this calculation
, we have assumed a uv slope , @xmath142 ( @xmath1105 ) to be -2 , which is consistent with the previous observations ( e.g. , bouwens et al .
2012 , apj , 754 , 83 ) .
the broad - band magnitude has a negligible dependence on uv slope , because the broad - band width is only @xmath1106 .
we have confirmed that @xmath157 chages only a few percent at maximum , when uv slope was varied from -3.0 to -1.0 .
in addition , @xmath159 was estimated by combining @xmath160 and @xmath157 .
the results found for each region are described in the following subsections .
we have observed eight @xmath48-dropout galaxies in the d1id01 region out of ten candidates .
almost all @xmath48-dropout galaxies in the d1id01 region were spectroscopically observed , as shown in figure [ fig : spec_obs_d1i ] .
three galaxies clearly have single emission lines , which can be identified as ly@xmath11 emission lines of @xmath13 galaxies .
their photometric and spectroscopic properties are summarized in table [ tab : spec_all ] .
two of three galaxies ( id=1 and 2 ) have close redshifts with difference of @xmath1107 , corresponding to the radial distance of @xmath1108 in physical scale . from our selection criteria
, we can expect @xmath1109 galaxy in a @xmath1110 bin if they were homogeneously distributed in redshift space .
the possibility to have two galaxies within @xmath1111 is 16% .
although their distribution is more concentrated than homogeneous , these two galaxies are unlikely to merge into a single halo by @xmath5 based on our analysis of the possible separations of protocluster galaxies at @xmath13 ( figure [ fig:3d_model ] ) . for now , we can not conclude that there is a real protocluster in the d1id01 region due to the small number of statistics . as for the d3id01 region , eight @xmath48-dropout galaxies
were observed out of sixteen candidates .
the completeness of our spectroscopic observation is smaller ( @xmath1112 ) than for the d1id01 region , which has less protocluster member candidates .
many faint @xmath48-dropout galaxies are still to be observed because we assigned the brighter @xmath48-dropout galaxies higher priorities .
ly@xmath11 emission lines were detected from two of the eight spectroscopic targets .
the sky distribution of the targets is shown in figure [ fig : spec_obs_d3i ] .
table [ tab : spec_all ] describes the properties of spectroscopically confirmed galaxies .
these two galaxies have almost the same redshift with the difference of @xmath1113 ( @xmath1114 in physical scale ) .
the possibility that two galaxies have this small redshift separation is only 1.2% , and these two galaxies can certainly be expected to be in the same halo at @xmath5 based on this small separation .
while we could not make a clear conclusion due to the small number of confirmed galaxies , the discovery of a close galaxy - pair at @xmath13 could imply the existence of a protocluster .
interestingly , @xcite have found that galaxies tend to be in close pairs in another protocluster at @xmath16 .
the relatively small number of confirmed candidate can be attributed to the observational limit since our spectroscopic samples are biased to brighter galaxies ( @xmath1115 ) .
we compared the d3id01 protocluster candidate with the @xmath16 protocluster in the sdf @xcite , which was also identified by the combination of dropout selection and follow - up spectroscopy targeting ly@xmath11 emission like this study . applying the same magnitude limit of @xmath1115 to the sdf protocluster ,
the number of remaining protocluster members was only two our of a total of ten .
furthermore , @xcite reported the discovery of two protoclusters at @xmath1116 .
these were discovered from a narrow - band survey , and six and four laes are included in each protocluster .
although lae selection is different from our dropout selection , it is useful to check the distribution of the uv continuum and the ly@xmath11 luminosity of protocluster galaxies . based on our observational limits of uv continuum and ly@xmath11 luminosity , only @xmath1117 laes would be identified for these protoclusters .
therefore , it is not unreasonable to expect only two confirmed member galaxies in this study even if there is a real protocluster . we have spectroscopically observed fifteen @xmath47-dropout galaxies in the d1rd01 region , and detected single emission lines from six galaxies .
the sky distribution of the observed galaxies is shown in figure [ fig : spec_obs_d1r ] . in the @xmath1118 overdense region
, there are @xmath1119 galaxies ; thus , only @xmath1120 @xmath47-dropout galaxies were observed by the follow - up spectroscopy .
two galaxies ( id=5 and 6 ) out of six are clustering both in spatial ( @xmath1121 ) and redshift space ( @xmath1122 ) at @xmath1123 , whose three - dimensional separation is @xmath1124 in physical scale . considering the observed volume ( @xmath1125 radius and @xmath1126 ) , it is unlikely ( @xmath1127 ) that the close pair is reproduced by uniform random distribution of six galaxies in three - dimensional space .
however , it is unclear whether this galaxy pair will grow into a cluster at @xmath5 due to the small number of confirmed galaxies ; at least , these two galaxies are expected to merge into a single halo .
since there are many spectroscopically unobserved galaxies , further follow - up observation will enable to clarify whether there is a protocluster or not .
in the d4rd02 region , the total integration time of follow - up spectroscopic observation was only two hours , which was half of that in the d1rd01 region .
thus , although twelve @xmath47-dropout galaxies were observed , ly@xmath11 emission lines were detected from only three galaxies .
the sky distribution of the observed galaxies is shown in figure [ fig : spec_obs_d4r ] .
these three galaxies are largely separated in redshift space .
since about 20 @xmath47-dropout galaxies remain to be spectroscopically observed , further follow - up observation will be necessary to make a conclusion . combining the deimos and focas follow - up observations , 123 @xmath46-dropout galaxies were observed , and the redshifts of 36 galaxies were determined by detecting ly@xmath11 emission lines . the sky distribution of the observed galaxies is shown in figure [ fig : spec_obs_d1 g ] .
figure [ fig : redshift_d1 g ] shows the redshift distribution of confirmed galaxies .
although galaxies seem to be clustering at @xmath1128 , these galaxies are spread over a large projected area , as shown in figure [ fig : spec_obs_d1 g ] . since the deimos has a wide fov ( @xmath1129 ) , which is larger than the area of the d1gd01 region , some @xmath46-dropout galaxies outside the overdense region
were also observed .
when we focus only on galaxies in the overdense region , shown by the red - line histogram in figure [ fig : redshift_d1 g ] , the peak around @xmath1130 becomes lower and only three galaxies are clustering within the expected redshift range of the protocluster candidate ( @xmath1131 ) .
given the total number of confirmed galaxies in the overdense region , the group of three galaxies can be reproduced even from a random homogeneous distribution with a probability of 21% .
based on this probability , it can not be dismissed that the observed redshift distribution is drawn from a uniform random distribution .
hence , we can not currently conclude that there is a protocluster in the d1gd01 region without further observations .
hereafter , we regard the d1gd01 region as not being a protocluster .
the high surface overdensity observed in this field could be attributed to a coincidental alignment of large - scale structure on a scale of @xmath1132 , which is too large to grow into a single halo by @xmath5 . combining the deimos and focas follow - up observations ,
144 @xmath46-dropout galaxies were spectroscopically observed in the d4gd01 protocluster region , and the redshifts of 42 galaxies were determined by detecting ly@xmath11 emission lines .
the sky distribution of the observed galaxies is shown in figure [ fig : spec_obs_d4 g ] .
the redshift distribution is shown in figure [ fig : redshift_d4 g ] .
there is a clear excess at @xmath8 , and , as contrasted with the d1gd01 region , eleven galaxies are clustered in a narrow redshift range of @xmath1133 , corresponding to @xmath1134 in physical scale . since it is almost impossible ( @xmath1135 ) to explain this clustering with a random homogeneous distribution , we concluded that there is a protocluster at @xmath8 , which includes eleven member galaxies ( id=10 - 20 ) .
it should also be noted that an agn was found in this region at @xmath1136 by our spectroscopy as shown in figure [ fig : spec_obs_d4 g ] .
the redshift was derived to be @xmath1137 based on its heii and ciii ] emission lines ( figure [ fig : spec_d4g_agn ] ) . according to this estimate , the redshift separation between the agn and the center of the protocluster is @xmath1138 , which corresponds to the radial distance of @xmath1139 .
as the redshift separation is too large , it is unlikely that this agn is a part of the protocluster members that will merge into a single halo by @xmath5 .
we have spectroscopically observed 95 @xmath45-dropout galaxies in the d1ud01 region , and 30 galaxies have single emission lines .
the sky distribution of the observed galaxies is shown in figure [ fig : spec_obs_d1u ] .
the redshift distribution is shown in figure [ fig : redshift_d1u ] .
there is a excess at @xmath1140 , including five galaxies within @xmath1141 .
the probability to reproduce this excess by drawing from a uniform random distribution of 30 galaxies is only 0.9% ; thus , the five galaxies were found to be significantly clustered though the absolute excess is only five .
the spatial and redshift separations among these five galaxies are small enough to merge into a single halo by @xmath5 compared with the model prediction ; therefore we confirmed a protocluster at @xmath1140 , which includes five member galaxies ( id=6 - 10 ) .
we have spectroscopically observed 57 @xmath45-dropout galaxies in the d4ud01 region , and 16 galaxies have single emission lines .
the sky distribution of the observed galaxies is shown in figure [ fig : spec_obs_d4u ] .
the redshift distribution is shown in figure [ fig : redshift_d4u ] .
there is a peak at @xmath1142 , consisting of five galaxies within @xmath1141 .
the probability to reproduce this excess by an uniform random distribution of 16 galaxies was found to be less than 0.1% .
these five galaxies are expected to merge into a single halo by @xmath5 compared with the model prediction .
therefore , we confirmed a protocluster at @xmath1142 , which includes five member galaxies ( id=7 - 11 ) . based on these follow - up spectroscopic observations , we were able to confirm three protoclusters in the d4gd01 , d1ud01 , and d4ud01 regions .
we could not confirm that the overdensity observed in the d1gd01 region is indeed a protocluster .
thus , at least , the success rate of our protocluster search is found to be 3/4 at @xmath0 , which is consistent with that of the model prediction ( @xmath1143 of @xmath2 overdense regions are expected to be real protoclusters ) . as for @xmath47- and @xmath48-dropout protocluster candidates , it is unclear whether they are real protoclusters or not because of the small number of spectroscopically confirmed galaxies , though the d3id01 and d1rd01 regions include close galaxy pairs , which could indicate the existence of protoclusters .
these results suggest that most of the other protocluster candidates will turn out to be genuine protoclusters once sufficient follow - up spectroscopic observations are performed . the summary of our protocluster confirmation is described in table [ tab : cluster ] .
these findings do not only increase the number of known protocluster at high redshift , but provide us with samples that are complementary to previous works in which protoclusters were mainly discovered in qso or rg regions .
although we discovered one agn in an overdense region by our spectroscopy , it was not associated with any protocluster . the precise relation between protoclusters and agns merits future investigation .
the radial velocity dispersions of the confirmed protoclusters were calculated by the redshifts of protocluster members assuming that the redshifts probe line - of - sight velocity . to calculate dispersion
, we used the biweight variance @xcite , which is an effective method even with a small sample .
since our follow - up spectroscopy is not complete , the effect of small number statistics should be taken into account in the uncertainty of the radial velocity dispersion .
the uncertainty of radial velocity dispersion was measured by the combination of the velocity error of our spectroscopic observations and the standard deviation of the measurements by bootstrap sampling the protocluster members .
in addition to the optical imaging and our spectroscopy , rich multi - wavelength data is available in part of the cfhtls deep fields .
although this enables us to make further analysis , such as sed fitting , to derive galaxy properties in more detail , these studies will be addressed in a future paper .
the d4gd01 protocluster , of which we confirmed eleven protocluster members , is the most extensively mapped protoclusters among the three confirmed systems described in section [ sec : spec ] .
thus , we focus on this protocluster in the following discussion of the internal structure of the protocluster and the properties of its galaxies .
we investigated the three - dimensional distribution of protocluster galaxies in the d4gd01 region , as shown in figure [ fig:3d ] . in this analysis ,
the distances are simply estimated from the measured redshift , including any possible peculiar velocity component .
however , we have checked that this does not significantly affect our estimates given the typical size of protoclusters . by @xmath5 , are still embedded in individual host halos at @xmath1144 , and their clustering is not probably strong enough to provoke a large peculiar velocity field .
therefore , we expect that their real three - dimensional distribution at some level can be determined by using their redshifts as proxies for their relative line of sight distances . according to the theoretical model of @xcite ,
the deviation of the difference between apparent and geometrical redshift is @xmath1145 , shown by black scale in figure [ fig:3d ] .
] this protocluster seems to have a region where galaxies are strongly concentrated , reminiscent of a cluster core . to discuss the internal structure quantitatively , we calculated the spatial separation of galaxies with respect to the @xmath73th nearest neighbor . in figure
[ fig : dnth ] , the distributions of the separation from the 1st to 6th nearest galaxies of individual protocluster galaxies are shown , and the red - line histograms indicate the expected distribution if eleven galaxies are randomly located inside the protocluster . the distribution of protocluster galaxies from the 3rd to 5th nearest are significantly different from the random distribution , whose significances are @xmath1146 based on the kolmogorov - smirnov ( ks ) test ; especially , the significance of 5th nearest is less than 0.01 .
in contrast , there are no significant differences in the distribution of 6th or higher nearest galaxies .
these results suggest that the protocluster has a subgroup consisting of six galaxies ; therefore , the subgroup can not be seen in the distribution of the 6th or higher nearest galaxies , which are consistent with the random distribution .
the six galaxies constituting the subgroup are defined by having shorter separation than @xmath80 from the 5th nearest galaxies ( id=11 - 16 ) .
the distributions of 1st and 2nd nearest galaxies could also be divided into two groups of shorter and longer separations though they are not so significant ( @xmath1147 ) due to focusing on too small scale .
these six galaxies are indicated by red points in figure [ fig:3d ] , and are located near the center of the protocluster .
there are several galaxies in the region surrounding the core , which could assemble into the core to form a rich cluster .
this is in clear contrast that we saw in the protocluster at @xmath13 , which was found to have several small subgroups , like galaxy pairs @xcite .
although we have only one protocluster at each redshift , if they are the progenitor and descendant of each other , the transformation of protocluster internal structure from @xmath13 to @xmath12 may be indicative of the virializing process over cosmic time , whereby protoclusters dynamically evolve into a more and more concentrated structure . at @xmath18 ,
the virializing process would not be completed yet though there are richer protoclusters where some massive and passive galaxies have already appeared .
the internal structures of some protoclusters have been more closely investigated using extensive spectroscopy , and some protoclusters were found to have significant substructure @xcite .
therefore , even at the same redshift , protoclusters could have a large variety of internal structure . in this study based on a small number of protoclusters , the difference of internal structure found in the @xmath13 and @xmath12 protoclusters
are assumed to be resulted from the evolutionary phase of the representative protoclusters .
however , larger samples at each redshift will be required to statistically match progenitors and descendants , allowing us to study cluster formation over cosmic time ( e.g. , * ? ? ?
we compared several galaxy properties between protocluster members and coeval field galaxies to investigate whether there are any differences due to their environment .
the d4gd01 protocluster consists of eleven members , while the number of field @xmath46-dropout galaxies available is 67 by combining the galaxies in the d1 and d4 fields .
the protocluster and field galaxies were identified by the same imaging and spectroscopic observations ; thus , some sample bias , if any , would affect both samples in the same way .
this should allow us to make a fair comparison between protocluster and field galaxies .
the average of @xmath158 , @xmath157 , and @xmath159 are described in table [ tab : ave_d4 g ] .
since the d4gd01 protocluster has a core - like substructure as mentioned in section [ sec : struc ] , the eleven members were divided into two groups of six galaxies in the core and five galaxies in the outskirt .
table [ tab : ave_d4 g ] also shows the average properties of these two subgroups . in figure
[ fig : m - ew ] , the properties of individual protocluster and field galaxies are plotted on the @xmath157 and @xmath159 diagram .
we found that protocluster galaxies have significantly smaller @xmath159 or @xmath158 luminosity than field galaxies ( ks p - value is @xmath1148 ) ; especially , fainter protocluster members in @xmath157 are strongly suppressed in their ly@xmath11 emissions compared with field galaxies .
no significant difference between the core and the outskirt in the @xmath159 and @xmath157 distribution was found .
cccc protocluster & @xmath1149 & @xmath1150 & @xmath1151 + field & @xmath1152 & @xmath1153 & @xmath1154 + p - value & 0.04 & 0.67 & 0.03 + core & @xmath1155 & @xmath1156 & @xmath1157 + outskirt & @xmath1158 & @xmath1159 & @xmath1160 previous studies have found little evidence for significant differences between the properties of galaxies inside and outside protoclusters , at least at @xmath1161 @xcite .
it is therefore interesting that we are finding a difference in the @xmath159 or @xmath158 distributions between field and protocluster in the d4gd01 system at @xmath8 .
this is perhaps an indication that the situation is changing around @xmath12 .
one simple mechanism that could reduce the ly@xmath11 @xmath159 would be dust , which traps ly@xmath11 photons .
if dust is a major reason for the small @xmath159 , @xmath157 is expected to be systematically smaller in the protocluster because uv flux can also be easily attenuated by dust as in the case with ly@xmath11 emission . however , we did not find any systematic difference in @xmath157 between the protocluster and field galaxies ( see table [ tab : ave_d4 g ] ) .
although dust attenuation does not seems to be the reason for this result , it is still possible that the amount of uv flux attenuated by dust is being compensated by higher sfr or other effects related to uv emission ; in that case , we would not find any difference in the distribution of apparent @xmath157 between protocluster and field galaxies even if there were to be a clear difference of dust attenuation .
therefore , in order to more directly estimate the dust attenuation , we compared the uv slope between protocluster and field galaxies .
the uv slope of each @xmath46-dropout galaxy was determined from the @xmath1162 color . however , since the @xmath27-band image was significantly deeper ( @xmath1163 ) than the @xmath28-band image , some @xmath46-dropout galaxies were not detected in the @xmath28-band . for the estimate of the uv slope ,
we therefore only used @xmath46-dropout galaxies which were detected in the @xmath28-band image with @xmath1164 significance ( @xmath1165 ) .
although this discordant depth between the @xmath27- and @xmath28-band images will lead to a bias in the estimate of uv slope , the protocluster and field galaxies would be equally biased as they were selected by the same criteria and from the same dataset .
the numbers of @xmath46-dropout galaxies used in the estimate was nine for the protocluster and 50 for the field .
note that there is no difference in the fraction of galaxies detected in @xmath28-band between protocluster ( 9/11 ) and field ( 50/67 ) .
the uv slope was calculated from @xmath1166 where @xmath1167 is the effective wavelength .
the average @xmath142 of the protocluster galaxies was @xmath1168 , and that of the field galaxies was @xmath1169 ; thus , it would be hard to explain the difference in @xmath159 simply by dust attenuation .
it should be noted that , the radiative process of ly@xmath11 photon to escape from a galaxy is very complicated affected by many other quantities , such as dust geometry , gas kinematics , and outflow ( e.g. , * ? ? ?
* ; * ? ? ?
neutral hydrogen gas within a protocluster is another possible reason for the small @xmath159 .
@xcite found a large amount of neutral hydrogen gas ( @xmath1170 ) in the intracluster space of a protocluster by examining spectra of background galaxies of a protocluster at @xmath19 that showed absorption at the same wavelength as the observed ly@xmath11 of the protocluster .
if the same is true in the d4gd01 protocluster , though our follow - up spectroscopy is not deep enough to check this even by stacking , a small @xmath159 could be explained as a result of resonant scattering by the intracluster neutral hydrogen gas .
while the uv photons can penetrate neutral hydrogen gas , ly@xmath11 emission is scattered and diffused , consistent with our results .
suppose that a nearly mature protocluster , such as the d4gd01 protocluster , had already accumulated significant cold intracluster gas at @xmath8 ; the intracluster gas would come either from the outside of the protocluster drawn in by the strong gravitational potential of the protocluster , or could be brought in with the evolved member galaxies themselves through in- and outflows .
although it is difficult to identify the cause , we have found that the average @xmath159 of @xmath8 protocluster members is significantly smaller than that of field galaxies .
however , it is still under debate how ly@xmath11 @xmath159 depends on environments .
actually , in contrast with our study , @xcite have found larger @xmath159 in the ssa22 protocluster at @xmath1171 .
@xcite reported a protocluster being composed of two subgroups at @xmath1140 , where the one subgroup has larger @xmath159 , but the other has smaller .
furthermore , there are some protoclusters which have no significant difference of ly@xmath11 @xmath159 from that of field galaxies @xcite .
since the number of known protoclusters is limited even around @xmath112 so far , it will be required to make a large sample enough to address a general feature .
this may have consequences for the measurement of the ly@xmath11 fraction @xcite , which is one of the ways to probe reionization at high redshifts .
a detailed study of galaxy properties will be made in the future by combining multi - wavelength data .
in this study , we have presented a protocluster survey from @xmath112 to @xmath13 in the cfhtls deep fields .
this survey was performed using wide - field imaging without using the preselection of common protocluster probes such as rgs and qsos .
protocluster candidates were identified by measuring the surface number density of dropout galaxies , and the follow - up spectroscopic observations identified three real protoclusters .
the major results and implications of this study are summarized below . 1 . we investigated the sky distribution of @xmath45- , @xmath46- , @xmath47- , and @xmath48-dropout galaxies in the wide - field imaging of the cfhtls deep fields , and quantified the local surface number density by counting galaxies within a fixed aperture .
we selected a total of 21 overdense regions with an overdensity significance greater than @xmath2 as protocluster candidates .
the number density of protocluster candidates was approximately one candidate per @xmath23 area for each redshift sample based on a @xmath1172 survey .
we investigated the relation between the overdensity at high redshifts and the descendant halo mass using light - cone models constructed from cosmological simulations .
we selected galaxy samples with the same redshift distribution as the observations , and the same overdensity measuring procedure was applied to this simulated sample of dropout galaxies .
a strong correlation between the overdensity at high redshifts and the descendant halo mass at @xmath5 was found , and @xmath1143 of the overdense regions with significance over @xmath2 are expected to grow to dark matter halos with @xmath1173 at @xmath5 . despite significant projection effects , the model predicts that protoclusters can be identified with high confidence by measuring the surface overdensity significance .
in addition , the model predicts that protocluster members are , on average , spread within a sphere of @xmath128 radius .
we carried out follow - up spectroscopic observations of eight protocluster candidates between @xmath112 and @xmath13 to confirm whether these were genuine protoclusters .
the redshifts of all the protocluster members were determined by detecting their ly@xmath11 emission lines , and no apparent contamination from low - redshift interlopers was found in our spectroscopic observation .
although the completeness of slit allocation to dropout galaxies in a protocluster candidate is about @xmath1174 , three of the eight protocluster candidates were confirmed to be genuine protoclusters by ascertaining that their member galaxies were clustering both in spatial and redshift directions ( @xmath1175 ) with @xmath1176 significance with more than five members spectroscopically identified .
spectroscopy revealed that chance alignment of dropout galaxies mimics an overdensity region for one candidate .
there are some signatures of clustering of several galaxies in the other four protocluster candidates at @xmath7 ; however , the numbers of spectroscopically confirmed galaxies are still too small to conclude that they are genuine protoclusters .
although there are still many protocluster candidates to follow - up spectroscopically , our method to search for protoclusters utilizing wide - field imaging is reliable and sufficiently effective to construct high - redshift protocluster samples based on the success rate of follow - up observations .
we investigated the internal structure of the d4gd01 protocluster at @xmath8 based on its eleven spectroscopically confirmed members .
the distribution of member galaxies exhibits a core - like structure : half of the members are concentrated in a central small region ( @xmath9 ) , and the others in the outskirts ( @xmath10 ) .
the result implies that this protocluster might be on the way to evolve into a virialized structure though further protocluster samples are required to confirm a general trend .
the d4gd01 protocluster galaxies have significantly smaller @xmath159 than coeval field galaxies , while there is no difference in @xmath157 .
we considered two physical mechanisms that may lead to this difference ; the first is dust in protocluster galaxies , and the second is intracluster neutral hydrogen gas .
although we were not able to draw definite conclusions based on current data , the uv slope was found not to favor an interpretation whereby the difference in @xmath159 is attributed only to dust .
our finding of a smaller @xmath159 implies that the properties of protocluster galaxies might be affected by the environment already at @xmath8 .
although we were successful in finding at least three new protoclusters using wide - field imaging and spectroscopy in a blank deep field , the sample size is still too small to elucidate a general picture of the structure formation and evolution of environmental effects .
however , this study is an important benchmark for finding large numbers of protoclusters and tracing the cluster formation history in upcoming deep , wide surveys using identical techniques . using the new instrument hyper suprimecam ( hsc ) on the subaru telescope
, we are performing an unprecedented wide and deep survey over the next four years .
the hsc strategic survey consists of three layers : the wide layer covers @xmath1177 with the @xmath48-band depth of @xmath1178 , the deep layer @xmath1179 with @xmath1180 , and the ultradeep layer @xmath1181 with @xmath1182 . from this study
, we estimate that the number of protoclusters , that the hsc survey will be able to find , will be @xmath1183 at @xmath7 and @xmath1184 at @xmath0 . from 2018
onwards , spectroscopic follow - up of tens of thousands of dropout galaxies selected from the hsc deep layer will be performed using the large multiplexing capability of the prime focus spectrograph , also on the subaru telescope .
this will allow us to understand the cluster formation process all the way from reionization to the present - day .
the cfhtls data used in this study are based on observations obtained with megaprime / megacam , a joint project of cfht and cea / irfu , at the canada - france - hawaii telescope ( cfht ) which is operated by the national research council ( nrc ) of canada , the institut national des science de lunivers of the centre national de la recherche scientifique ( cnrs ) of france , and the university of hawaii .
this study is based in part on data products produced at terapix available at the canadian astronomy data centre as part of the canada - france - hawaii telescope legacy survey , a collaborative project of nrc and cnrs . and
, this study based on data collected at the subaru , the w. m. keck , and the gemini north telescopes .
the subaru telescope is operated by the national astronomical observatory of japan .
the w. m. keck telescope is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration .
the w. m. keck observatory was made possible by the generous financial support of the w.m .
keck foundation .
the gemini north telescope is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministrio da cincia , tecnologia e inovao ( brazil ) and ministerio de ciencia , tecnologa e innovacin productiva ( argentina ) .
we are grateful to the subaru , the w. m. keck , and gemini observatory staff for their help with the observations , and wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community . the millennium simulation databases used in this study and the web application providing online access to them were constructed as part of the activities of the german astrophysical virtual observatory ( gavo ) .
we thank the anonymous referee for valuable comments and suggestions which improved the manuscript .
this research was supported by the japan society for the promotion of science through grant - in - aid for scientific research 12j01607 and 15h03645 .
adams , s. m. , martini , p. , croxall , k. v. , overzier , r. a. , & silverman , j. d. 2015 , , 448 , 1335 aravena , m. , bertoldi , f. , carilli , c. , et al .
2010 , , 708 , l36 barkana , r. , & loeb , a. 1999 , , 523 , 54 beers , t. c. , flynn , k. , & gebhardt , k. 1990 , , 100 , 32 bertin , e. , & arnouts , s. 1996 , , 117 , 393 bouwens , r. j. , illingworth , g. d. , franx , m. , & ford , h. 2007 , , 670 , 928 bruzual , g. , & charlot , s. 2003 , , 344 , 1000 butcher , h. , & oemler , a. , jr .
1984 , , 285 , 426 calzetti , d. , armus , l. , bohlin , r. c. , et al .
2000 , , 533 , 682 capak , p. l. , riechers , d. , scoville , n. z. , et al .
2011 , , 470 , 233 casey , c. m. , cooray , a. , capak , p. , et al .
2015 , , 808 , l33 chiang , y .- k . ,
overzier , r. , & gebhardt , k. 2013 , , 779 , 127 cucciati , o. , zamorani , g. , lemaux , b. c. , et al .
2014 , , 570 , a16 curtis - lake , e. , mclure , r. j. , pearce , h. j. , et al .
2012 , , 422 , 1425 de lapparent , v. , geller , m. j. , & huchra , j. p. 1986
, , 302 , l1 digby - north , j. a. , nandra , k. , laird , e. s. , et al .
2010 , , 407 , 846 dressler , a. 1980 , , 236 , 351 duval , f. , schaerer , d. , stlin , g. , & laursen , p. 2014
, , 562 , a52 faber , s. m. , phillips , a. c. , kibrick , r. i. , et al . 2003 , , 4841 , 1657 galametz , a. , stern , d. , pentericci , l. , et al .
2013 , , 559 , a2 girardi , l. , groenewegen , m. a. t. , hatziminaoglou , e. , & da costa , l. 2005 , , 436 , 895 guo , q. , white , s. , boylan - kolchin , m. , et al .
2011 , , 413 , 101 gwyn , s. d. j. 2012 , , 143 , 38 hatch , n. a. , kurk , j. d. , pentericci , l. , et al .
2011 , , 415 , 2993 hatch , n. a. , wylezalek , d. , kurk , j. d. , et al .
2014 , , 445 , 280 henriques , b. m. b. , white , s. d. m. , lemson , g. , et al .
2012 , , 421 , 2904 hook , i. m. , jrgensen , i. , allington - smith , j. r. , et al . 2004 , , 116 , 425 hopkins , p. f. 2012 , , 420 , l8 inoue , a. k. , shimizu , i. , iwata , i. , & tanaka , m. 2014 , , 442 , 1805 ishigaki , m. , ouchi , m. , & harikane , y. 2015 , arxiv:1509.01751 kacprzak , g. g. , yuan , t. , nanayakkara , t. , et al .
2015 , , 802 , l26 kashikawa , n. , aoki , k. , asai , r. , et al .
2002 , , 54 , 819 kashikawa , n. , shimasaku , k. , malkan , m. a. , et al .
2006 , , 648 , 7 kashikawa , n. , kitayama , t. , doi , m. , et al .
2007 , , 663 , 765 kodama , t. , tanaka , i. , kajisawa , m. , et al .
2007 , , 377 , 1717 kubo , m. , uchimoto , y. k. , yamada , t. , et al .
2013 , , 778 , 170 kuiper , e. , hatch , n. a. , rttgering , h. j. a. , et al . 2010 , , 405 , 969 kuiper , e. , hatch , n. a. , miley , g. k. , et al .
2011 , , 415 , 2245 kuiper , e. , venemans , b. p. , hatch , n. a. , miley , g. k. , rttgering , h. j. a. 2012 , , 425 , 801 kulas , k. r. , mclean , i. s. , shapley , a. e. , et al .
2013 , , 774 , 130 kurk , j. d. , pentericci , l. , rttgering , h. j. a. , & miley , g. k. 2004 , , 428 , 793 lehmer , b. d. , alexander , d. m. , geach , j. e. , et al .
2009 , , 691 , 687 lemaux , b. c. , cucciati , o. , tasca , l. a. m. , et al .
2014 , , 572 , a41 madau , p. 1995
, , 441 , 18 matsuda , y. , yamada , t. , hayashino , t. , et al .
2011 , , 410 , l13 mawatari , k. , yamada , t. , nakamura , y. , hayashino , t. , & matsuda , y. 2012 , , 759 , 133 meiksin , a. 2006 , , 365 , 807 miley , g. k. , overzier , r. a. , tsvetanov , z. i. , et al .
2004 , , 427 , 47 miller , t. b. , hayward , c. c. , chapman , s. c. , & behroozi , p. s. 2015 , , 452 , 878 muldrew , s. i. , hatch , n. a. , & cooke , e. a. 2015 , , 452 , 2528 ono , y. , ouchi , m. , mobasher , b. , et al .
2012 , , 744 , 83 ouchi , m. , shimasaku , k. , akiyama , m. , et al .
2005 , , 620 , l1 overzier , r. a. , bouwens , r. j. , cross , n. j. g. , et al .
2008 , , 673 , 143 overzier , r. a. , guo , q. , kauffmann , g. , et al . 2009a , , 394 , 577 overzier , r. a. , shu , x. , zheng , w. , et al .
2009b , , 704 , 548 overzier , r. , lemson , g. , angulo , r. e. , et al .
2013 , , 428 , 778 popesso , p. , rodighiero , g. , saintonge , a. , et al . 2011 , , 532 , a145 salpeter , e. e. 1955 , , 121 , 161 schlafly , e. f. , & finkbeiner , d. p. 2011
, , 737 , 103 shimakawa , r. , kodama , t. , tadaki , k .-
i . , et al .
2015 , , 448 , 666 smolcic , v. , miettinen , o. , tomicic , n. , et al . 2016 , arxiv:1604.01775 springel , v. , white , s. d. m. , jenkins , a. , et al . 2005 , , 435 , 629 stark , d. p. , ellis , r. s. , bunker , a. , et al .
2009 , , 697 , 1493 stark , d. p. , ellis , r. s. , & ouchi , m. 2011 , , 728 , l2 steidel , c. c. , adelberger , k. l. , dickinson , m. , et al .
1998 , , 492 , 428 steidel , c. c. , adelberger , k. l. , shapley , a. e. , et al .
2000 , , 532 , 170 steidel , c. c. , adelberger , k. l. , shapley , a. e. , et al .
2005 , , 626 , 44 tamura , y. , iono , d. , wilner , d. j. , et al .
2010 , , 724 , 1270 thomas , r. , le fvre , o. , cassata , v. l. b. p. , et al . 2014 , arxiv:1411.5692 toshikawa , j. , kashikawa , n. , ota , k. , et al . 2012 , , 750 , 137 toshikawa , j. , kashikawa , n. , overzier , r. , et al .
2014 , , 792 , 15 tran , k .- v .
h. , papovich , c. , saintonge , a. , et al .
2010 , , 719 , l126 trenti , m. , bradley , l. d. , stiavelli , m. , et al .
2012 , , 746 , 55 treu , t. , schmidt , k. b. , trenti , m. , bradley , l. d. , & stiavelli , m. 2013 , , 775 , l29 valentino , f. , daddi , e. , strazzullo , v. , et al .
2015 , , 801 , 132 venemans , b. p. , rttgering , h. j. a. , miley , g. k. , et al .
2007 , , 461 , 823 verhamme , a. , schaerer , d. , atek , h. , & tapken , c. 2008 , , 491 , 89 visvanathan , n. , & sandage , a. 1977 , , 216 , 214 van der burg , r. f. j. , hildebrandt , h. , & erben , t. 2010 , , 523 , a74 walter , f. , decarli , r. , carilli , c. , et al .
2012 , , 486 , 233 wylezalek , d. , galametz , a. , stern , d. , et al .
2013 , , 769 , 79 yamada , t. , nakamura , y. , matsuda , y. , et al .
2012 , , 143 , 79 zirm , a. w. , stanford , s. a. , postman , m. , et al .
2008 , , 680 , 224 | we present the discovery of three protoclusters at @xmath0 with spectroscopic confirmation in the canada - france - hawaii telescope ( cfht ) legacy survey deep fields . in these fields ,
we investigate the large - scale projected sky distribution of @xmath1 lyman break galaxies and identify 21 protocluster candidates from regions that are overdense at more than @xmath2 overdensity significance . based on cosmological simulations
, it is expected that more than @xmath3 of these candidates will evolve into a galaxy cluster of at least a halo mass of @xmath4 at @xmath5 .
we perform follow - up spectroscopy for eight of the candidates using subaru / focas , keckii / deimos , and gemini - n / gmos . in total
we target 462 dropout candidates and obtain 138 spectroscopic redshifts .
we confirm three real protoclusters at @xmath6 with more than five members spectroscopically identified , and find one to be an incidental overdense region by mere chance alignment .
the other four candidate regions at @xmath7 require more spectroscopic follow - up in order to be conclusive .
a @xmath8 protocluster , which has eleven spectroscopically confirmed members , shows a remarkable core - like structure composed of a central small region ( @xmath9 ) and an outskirts region ( @xmath10 ) .
the ly@xmath11 equivalent widths of members of the protocluster are significantly smaller than those of field galaxies at the same redshift while there is no difference in the uv luminosity distributions .
these results imply that some environmental effects start operating as early as at @xmath12 along with the growth of the protocluster structure .
this study provides an important benchmark for our analysis of protoclusters in the upcoming subaru / hsc imaging survey and its spectroscopic follow - up with the subaru / pfs that will detect thousands of protoclusters up to @xmath13 . |
liouville theory seems to be a universal building block that appears in various contexts such as noncritical string theory , two - dimensional gravity or d - brane physics .
it is also closely related to the @xmath0 or @xmath1 wznw models which are interesting as solvable models for string theory on noncompact curved backgrounds . from a more general point of view one
may regard liouville theory as a prototype for an interacting conformal field with noncompact target space .
it should therefore serve as a natural starting point for the development of techniques for the exact solution of such conformal field theories . in the case of liouville theory with periodic boundary conditions we now have a relatively complete characterization @xcite : knowledge of the spectrum of the theory and three point functions of primary fields allow one to consistently reconstruct arbitrary expectation values of local fields on the sphere or cylinder .
our understanding is less satisfactory in the case of liouville theory on two - dimensional domains with boundary such as the infinite strip , the upper half plane or the disk : one would again expect the theory to be fully characterized in terms of a finite set of structure functions together with the knowledge of the spectrum of the theory on the strip .
consistency of the reconstruction of the theory from these fundamental data requires them to satisfy consistency conditions very similar to those formulated by cardy and lewellen @xcite in the case of rational conformal field theories .
a part of these data has been determined and some of the basic consistency conditions have been verified @xcite@xcite@xcite .
what is missing are the determination of the three point function of boundary operators and the verification that these data satisfy the full set of conditions ensuring consistency of the reconstruction of the theory .
the aim of the present paper is to propose an explicit expression for the three point function of boundary operators as the solution to one of the most important consistency conditions expressing the associativity of the product of boundary operators .
the structure of this paper is as follows : the following section gathers those results on liouville theory that we will use in the present paper .
the third section then contains our proposal for the three point function of boundary operators .
it is based on the observation @xcite that an ansatz for that three point function in terms of the fusion coefficients naturally leads to a solution of the consistency condition that expresses the associativity of the product of boundary operators .
it remains to fix the remaining freedom by imposing certain normalization conditions .
some concluding remarks are made in section 4 , and the appendices contain some technical points used in the main text .
let us begin by recalling some results on liouville theory that will be relevant for the subsequent discussion , see @xcite for more details and references : lft on the sphere is semiclassically defined by the following action @xmath2 with the following boundary condition on the liouville field @xmath3 @xmath4 the parameter @xmath5 is related to planck s constant @xmath6 via @xmath7 , the scale parameter @xmath8 is often called the cosmological constant , and @xmath9 is the background charge @xmath10 it was first proposed in @xcite that liouville theory can be quantized as a conformal field theory with a space of states that decomposes as follows into irreducible unitary highest weight representations @xmath11 of the virasoro algebra : @xmath12 the highest weight @xmath13 of the representation @xmath14 was parametrized as @xmath15 .
the action of the virasoro algebra on @xmath16 is generated by the modes of the energy momentum tensor : @xmath17 the central charge of the virasoro algebra is then given in terms of @xmath5 via @xmath18 the local observables can be generated from the fields @xmath19 which semiclassically ( @xmath20 ) correspond to exponential functions @xmath21 of the liouville field .
the fields @xmath19 transform as primary fields under conformal transformations with conformal weight @xmath13 .
thanks to conformal symmetry , the fields @xmath19 are fully characterized by the three point functions @xmath22 an explicit formula for the three point function was proposed in @xcite @xmath23 where @xmath24 .
+ these pieces of information indeed amount to a full characterization of liouville theory on the sphere or cylinder : multipoint correlation functions can be factorized into three point functions by summing over intermediate states .
let us consider as prototypical example the four point function @xmath25 .
such four point functions may be represented by summing over intermediate states from the spectrum ( [ spec ] ) iff the variables @xmath26 are restricted to the range ) permits a meromorphic continuation to generic values of @xmath26 . ]
@xmath27 inserting a complete set of intermediate states between @xmath28 and @xmath29 would lead to an expression of the following form : @xmath30 @xmath31 is the s - channel conformal block which is completly determined by the conformal symmetry ( although no closed formula is known for it in general ) . @xmath32\ ! ( \eta ) \nonumber\end{aligned}\ ] ] where @xmath33 and @xmath34 , @xmath35 .
locality of the fields @xmath14 or associativity of the operator product expansion would lead to an alternative representation for @xmath25 as sum over _ t - channel _ conformal blocks @xmath36 : @xmath37 for the equivalence of the two representations ( [ s - channel ] ) and ( [ t - channel ] ) it is crucial that there exist @xcite invertible fusion transformations between s- and t - channel conformal blocks , defining the fusion coefficients : @xmath38 \,\mathcal{f}^{t}(\delta_{{\alpha}_i},\delta_{{\alpha}_{32}},z_i ) .
\label{transfost}\end{aligned}\ ] ] in @xcite , an explicit formula for the fusion coefficients was proposed in terms of the racah - wigner coefficients for an appropriate continuous series of representations of the quantum group @xmath39 with deformation parameter @xmath40 .
this formula was subsequently @xcite confirmed by direct calculation .
the resulting expression for the fusion coefficients is the following : @xmath41= \\ & \frac{{\gamma}(2q-\beta_1-\beta_2-\beta_{3}){\gamma}(\beta_2+\beta_{3}-\beta_1){\gamma}(q+\beta_2-\beta_1-\beta_{3}){\gamma}(q+\beta_3-\beta_2-\beta_1)}{{\gamma}(2q-\sigma_1-\beta_1-\sigma_{2}){\gamma}(\sigma_1+\sigma_{2}-\beta_1){\gamma}(q-\beta_1-\sigma_{2}+\sigma_1){\gamma}(q-\beta_1-\sigma_1+\sigma_{2 } ) } \\ & { \times}\frac{{\gamma}(q-\beta_{3}-\sigma_1+\sigma_3){\gamma}(\beta_{3}+\sigma_{1}+\sigma_{3}-q){\gamma}(\sigma_1+\sigma_3-\beta_{3}){\gamma}(\sigma_3+\beta_{3}-\sigma_1)}{{\gamma}(q-\beta_{2}-\sigma_2+\sigma_3){\gamma}(\beta_2+\sigma_2+\sigma_3-q){\gamma}(\sigma_2+\sigma_3-\beta_{2}){\gamma}(\sigma_{3}+\beta_2-\sigma_2 ) } \\ & { \times}\frac{{\gamma}(2q-2\sigma_{2}){\gamma}(2\sigma_{2})}{{\gamma}(q-2\beta_{3}){\gamma}(2\beta_{3}-q ) } \frac{1}{i}\int\limits_{-i\infty}^{i\infty}ds \;\ ; \frac{s_b(u_1+s)s_b(u_2+s)s_b(u_3+s)s_b(u_4+s ) } { s_b(v_1+s)s_b(v_2+s)s_b(v_3+s)s_b(q+s ) } , \end{aligned}\ ] ] where : @xmath42 an important identity satisfied by the fusion coefficients is the so - called pentagon equation , which follows from a similar identity satisfied by the racah - wigner coefficients mentioned previously @xcite .
@xmath43 one is also interested in understanding liouville theory on a simply connected domain @xmath44 with a nontrivial boundary @xmath45 .
for definiteness , we will only consider the conformally equivalent cases where @xmath44 is either the unit disk , the upper half plane or the infinite strip .
semiclassically , one may define the theory by means of the action @xmath46 where @xmath47 is the curvature of the boundary @xmath45 and @xmath48 is the so - called boundary cosmological constant . for the description of exact results in the quantum theory it was found to be useful @xcite to parametrize @xmath49 by means of a variable @xmath50 that is related to @xmath49 via @xmath51 requiring @xmath48 to be real one finds the two following regimes for the parameter @xmath52 : 1 . if @xmath53 , then @xmath52 is of the form @xmath54 2 . if @xmath55 , then @xmath52 is real . anticipating that all relevant objects will be found to possess meromorphic continuations w.r.t .
the boundary parameters @xmath50 , we shall discuss only the first regime explicitly in the following .
the hamiltonian interpretation of the theory @xcite is simplest in the case that @xmath44 is the infinite strip .
the associated hilbert space @xmath56 was found in @xcite to decompose as follows into irreducible representations of the virasoro algebra : @xmath57 the highest weight states generating the subrepresentation @xmath58 in @xmath59 will be denoted @xmath60 , where @xmath61 ( @xmath62 ) are the parameters of the boundary conditions associated to the left ( right ) boundaries of the strip .
it was proposed in @xcite@xcite that the states @xmath60 satisfy a reflection relation of the form @xmath63 ) .
the following formula was given in @xcite for the reflection coefficient @xmath64 : @xmath65 in addition to the fields @xmath66 localized in the interior of @xmath44 , one may now also consider operators @xmath67 that are localized at the boundary @xmath45 . the insertion point @xmath68 may separate segments of the boundary with different boundary conditions @xmath69 and @xmath70 . the boundary fields
@xmath67 are required to be primary fields with conformal weight @xmath71 .
they are therefore expected to create states @xmath60 and @xmath72 via @xmath73 to fully characterize lft on the upper half plane , one needs to determine some additional structure functions beside the bulk three point function @xmath74 : 1 .
bulk one point function @xcite : @xmath75 2 .
boundary two point function @xcite : @xmath76 let us remark that requiring the prefactor of the first delta - distribution on the right hand side of ( [ twopt ] ) to be unity partially fixes the normalization of boundary operators .
the appearance of the second term in ( [ twopt ] ) is a consequence of the reflection property ( [ reflprop ] ) .
bulk - boundary two point function@xcite : . ]
@xmath77 4 .
boundary three point function : @xmath78 taking advantage of the reflection property ( [ reflprop ] ) , we shall consider instead of @xmath79 the related quantity @xmath80 the present note will be devoted to the determination of this last structure function .
the basic consistency condition that the three - point function of boundary operators has to satisfy expresses the associativity of the product of boundary fields .
let us consider the 4 point function of boundary operators . inserting a complete set of intermediate states between the first two and the last two fields leads to an expansion into conformal blocks of the following form : , @xmath81 are close enough to q/2 . in this case , @xmath82 is of the form @xmath83 .
it turns out a posteriori that the general case can be treated by meromorphic continuation . ]
@xmath84 by using either cyclicity of correlation functions or associativity of the operator product expansion one would get a second expansion ( t - channel ) : @xmath85 using the fusion transformations ( [ transfost ] ) , the equivalence of the factorisation in the two channels can be rewritten : @xmath86 = c_{\beta_{4}|\beta_{32},\beta_{1}}^{\sigma_{4}\sigma_{2}\sigma_{1 } } c_{\beta_{32}|\beta_{3}\beta_{2}}^{\sigma_{4}\sigma_{3}\sigma_{2}}. \label{pentagone boundary}\ ] ] by means of the pentagon equation ( [ pentagone ] ) it easy to verify that the following ansatz @xmath87\ ] ] yields a solution to ( [ pentagone boundary ] ) , as was noticed in @xcite .
the coefficients @xmath88 appearing are unrestricted by ( [ pentagone boundary ] ) .
additional information is needed to determine them .
the boundary three point function @xmath79 should be meromorphic w.r.t .
the variables @xmath90 .
this far - reaching assumption can be motivated in various ways : one may e.g. use arguments like those reviewed in section 3 of @xcite concering the path integral for liouville theory . these arguments exhibit the analytic properties of correlation functions as a reflection of the asymptotic behavior of the liouville path integral measure in the region @xmath91 where the interaction terms vanish . ) . ]
such considerations lead in particular to the identification of the residues for the poles of @xmath79 with certain correlation functions in free field theory , which generalize the so - called screening - charge constructions of @xcite .
the resulting prescription for the calculation of these residues was formulated in @xcite .
most relevant for our purposes will be the observation that @xmath79 has a pole with residue 1 if @xmath92 : the relevant correlation functions in free field theory do not contain any screening charges .
on the other hand , it seems worth observing that the fusion coefficients themselves are meromorphic functions of all six variables they depend on , see ( * ? ? ? * lemma 21 ) .
this means that the function @xmath79 that is given by the expression ( [ ansatz ] ) will be meromorphic iff the function @xmath93 is meromophic w.r.t . @xmath94 . in the following
, we shall consider the special boundary field @xmath95 , which corresponds to a degenerate representation of the virasoro algebra . as pointed out in @xcite ,
it is in general not a trivial issue to decide when a boundary field that corresponds to a degenerate representation will satisfy the corresponding differential equations expressing null vector decoupling .
here , however , one may observe that one may create the boundary field @xmath95 by sending the bulk field @xmath96 to the boundary : it follows from the fact that @xmath96 satisfies a second order differential equation that the asymptotic behavior when @xmath96 approaches the boundary is described by a boundary field @xmath95 that satisfies a third order differential equation .
this last fact also implies that the operator product expansion of @xmath95 with a generic boundary operator can only contain three types of contributions : @xmath97 one may then consider the vacuum expectation values of the product of operators that is obtained by multiplying ( [ ope -b ] ) with the boundary fields @xmath98 , @xmath99 .
taking into account ( [ twopt ] ) , one is led to identify the structure functions @xmath100 ( @xmath101 with residues of the general three point function . as mentioned previously ,
the relevant residues can be represented as correlation functions in free field theory .
the structure function @xmath102 is nothing but a special case of the above - mentioned residue at @xmath92 , which is 1 .
this should be compared to what would follow from our ansatz ( [ ansatz ] ) .
let us note that it follows from appendix b(iii ) that the fusion coefficients indeed have a pole in the presently considered case .
the corresponding residue is most easily calculated by means of the recursion relations that follow from ( [ pentagone ] ) , see appendix b(iii ) for some details .
we find @xmath103 = \frac{\gamma(1+b^{2})}{\gamma(1 + 2b^{2 } ) } & \frac{\gamma(2b\sigma_1)\gamma(2b(q-\sigma_1))}{\gamma(b(q-\beta_2+\sigma_3-\sigma_1))\gamma(b(q-\beta_2+\sigma_1-\sigma_3))}{\times}\\ { \times } & \frac{\gamma(b(q-2\beta_2))\gamma(b(q-2\beta_2 + b))}{\gamma(b(\sigma_3+\sigma_1-\beta_2))\gamma(b(2q -\beta_2 -\sigma_3-\sigma_1))}. \end{aligned}\ ] ] our ansatz ( [ ansatz ] ) together with @xmath104 therefore implies the following first order difference equation for @xmath105 : @xmath106.\ ] ] this functional equation is solved by the following expression : @xmath107 in order to discuss the uniqueness of our solution ( [ gexp ] ) let us note that one may derive a second finite difference equation that is related to ( [ findiffg ] ) by substituting @xmath108 if one considers @xmath109 instead of @xmath110 .
taken together , these two functional equations allow one to conclude that our solution ( [ gexp ] ) is unique up to multiplication by a factor of the form @xmath111 , at least for irrational values of @xmath5 . to fix the remaining freedom it is useful to note that we now have two possible ways to calculate the structure function @xmath112 : on the one hand one may use our ansatz ( [ ansatz ] ) together with ( [ gexp ] ) and the following residue of the fusion coefficients : @xmath113 = \frac{\gamma(1+b^{2})}{\gamma(1 + 2b^{2})}. \\ & \cdot\frac{\gamma(2b\sigma_1)\gamma(2b(q-\sigma_1))\gamma(2b\beta_2 - 2bq)\gamma(2b\beta_2 - 1)}{\gamma(b(\beta_2+\sigma_3-\sigma_1))\gamma(b(\beta_2+\sigma_1-\sigma_3))\gamma(b(\sigma_3+\sigma_1+\beta_2 -q))\gamma(b(\beta_2 -\sigma_3-\sigma_1+q ) ) } \end{aligned}\ ] ] on the other hand , @xmath112 is one of the cases where a representation in terms of free field correlation functions is available@xcite : @xmath114 one finds a precisce coincidence of the expressions which one obtains by following these two ways if and only if the prefactor in the expression for @xmath115 is the one chosen in ( [ gexp ] ) . by collecting the pieces ,
one finally arrives at the following expression for the three point function of boundary operators : @xmath116 where the coefficients @xmath117 , @xmath118 and @xmath119 are defined as @xmath120 1 .
one recovers the expression for the boundary reflection amplitude ( [ reflcoeff ] ) from the boundary three point function the same way the bulk reflection amplitude was recovered from the bulk three point function in @xcite : using the fact that the fusion matrix depends on conformal weights only , and is thus invariant when @xmath121 , one finds : @xmath122 from the expression ( [ gexp ] ) for the function @xmath89 , one indeed finds formula ( [ reflcoeff ] ) for @xmath123 .
2 . one may explicitly check that the two - point function ( [ twopt ] ) is recovered by taking e.g. the limit @xmath124 if the three - point function : @xmath125 this is an easy consequence of the identity ( [ limfusion1 ] ) proven in appendix b(i ) 3 . with the help of symmetry properties of the fusion coefficients ( see appendix b(ii ) ) , it is possible to check that the boundary three point function is invariant w.r.t .
cyclic permutations .
we are finally going to sketch an argument in favor of the uniqueness of our expression for the boundary three point function : let us consider the associativity condition in the case that @xmath126 and that the boundary field @xmath127 is replaced by the degenerate field @xmath128 . due to ( [ ope -b ] )
, one finds that the associativity condition ( [ pentagone boundary ] ) gets replaced by @xmath129 \,c_{\beta_{4}|\beta_{3}\beta_{21}}^{\sigma_{4}\sigma_{3}\sigma_{1 } } = \ ; c_t({\beta}_{32};{\sigma}_4,{\sigma}_1)\ ; c_{\beta_{32}|\beta_{3}\beta_{2}}^{\sigma_{4}\sigma_{3}\sigma_{1 } } , \label{degasso}\ ] ] where @xmath130 takes the values @xmath131 , @xmath132 .
this can be read as a system of finite difference equations for the general boundary three point function @xmath133 . by specializing to the case that @xmath134 , one finds in particular a linear relation between the @xmath135 , @xmath99 .
replacing the degenerate field @xmath128 by @xmath136 leads to a similar second order finite difference equation which is related to the first by replacing @xmath137 in the coefficients , as well as replacing @xmath138 by @xmath139 .
it can be shown ( see appendix c ) that such self - dual systems of finite difference equations can for irrational @xmath5 only have at most two linearly independent solutions .
the relevant linear combination of these two solutions can be fixed e.g. by imposing the correct behavior w.r.t .
the reflection @xmath140 as given by ( [ reflprop ] ) . in this way
one arrives at the conclusion that the finite difference equations following from the associativity condition together with the reflection property ( [ reflprop ] ) suffice to uniquely determine the dependence of @xmath141 w.r.t . the variable @xmath142 .
but one may of course repeat that line of arguments by replacing any of the four operators in the four point function of boundary fields by the degenerate fields @xmath95 or @xmath143 , which would lead to finite difference equations that constrain the dependence of @xmath141 w.r.t .
@xmath144 and @xmath145 .
this leads to the conclusion that indeed the associativity condition in the presence of degenerate fields together with the reflection property ( [ reflprop ] ) uniquely determine the dependence of @xmath141 w.r.t .
all three variables @xmath90 .
the remaining freedom consists of multiplication with an arbitrary function of the boundary parameters @xmath146 .
this freedom is eliminated by requiring that the residue of the pole of @xmath141 at @xmath92 should indeed be unity , as discussed in subsection [ gcalc ] .
we now have determined the last of the structure functions that one needs to completely characterize liouville theory on domains with boundary .
what remains to be done is the verification that the expressions that have been put forward indeed satisfy the full set of cardy - lewellen type @xcite consistency conditions .
although some particularly important conditions have been verified ( an analog of the cardy condition @xcite , as well as the associativity condition studied in the present paper ) , it remains in particular to verify the conditions that link the boundary three point function with the bulk - boundary two - point function proposed in @xcite .
a beautiful characterization of the structure constants of certain classes of rational conformal field theories with boundaries has been given in @xcite , see also @xcite for closely related results .
it can be read as the statement that upon choosing a suitable normalization of the three point conformal blocks or chiral vertex operators , it becomes possible to recover all of the structure constants from the defining data of an associated modular tensor category .
validity of the cardy - lewellen conditions is automatic in this formalism .
what is not directly furnished by that formalism , though , is the explicit characterization for the necessary normalization of the three point conformal blocks
. it would certainly be nice to have at hand a similarly powerful formalism for non - rational conformal field theories such as liouville theory .
this should allow one in particular to carry out the missing proof that the structure functions satisfy the full set of cardy - lewellen type @xcite consistency conditions .
we will therefore try to verify whether our expression for the boundary three point function can be written in a form that one would expect to find in an extension of the formalism of @xcite to non - rational cft .
this turns out to be the case : let us write the three point function in terms of the b - racah - wigner coefficients : @xmath147 where @xcite @xmath148 it is easy to see that this can be rewritten as @xmath149 where the b - racah - wigner coefficients that appear on the right hand side have been modified w.r.t . to those considered in @xcite according to @xmath150 by using the the counterpart of ( [ crossid ] ) for the b - racah - wigner coefficients one may write ( [ c-6j ] ) as @xmath151 we consider ( [ c-6j ] ) as an encouraging hint that a verification of the cardy - lewellen conditions should be possible along similar lines as followed in @xcite for the case of rational cft .
this work was done while b.p .
was at laboratoire de physique thorique et mathmatique , universite montpellier ii , montpellier , france , as part of the tmr network with contract erbfmrx ct960012 .
b.p . now acknowledges support by the gif - project under contract nr i-645 - 130 - 14/1999 .
is supported by dfg sfb 288 `` differentialgeometrie und quantenphysik '' .
the function @xmath152 is a close relative of the double gamma function studied in @xcite .
it can be defined by means of the integral representation @xmath153 important properties of @xmath152 are @xmath154 the function @xmath155 may be defined in terms of @xmath152 as follows @xmath156 an integral that represents @xmath157 is @xmath158 the most important properties for our purposes are @xmath159 the @xmath160 may be defined in terms of @xmath161 as follows @xmath162 an integral representation convergent in the strip @xmath163 is @xmath164 properties : @xmath165
in this appendix we shall consider two important limiting cases of the fusion coefficients .
we are going to show : * if @xmath166 , @xmath167 then @xmath168 * introduce @xmath169 . then : @xmath170 to prove i ) , we will study the distribution on @xmath171 defined as @xmath172 it should be remarked that in sending @xmath173 some of the poles at @xmath174 and @xmath175 will cross the imaginary axis so that one has to deform the contour accordingly .
if one first considers @xmath176 for @xmath177 one finds by changing the integration variable to @xmath178 that the integral simplifies to a special value of the @xmath5-hypergeometric function : @xmath179 this particular value of the @xmath5-hypergeometric function vanishes , as follows from the identity @xcite @xmath180 and the fact that @xmath181 has a zero for @xmath182 .
one has thereby found that @xmath183 has support only for @xmath184 . in order to analyze the singular behavior near @xmath184 it will be useful to split off the residue contributions of the first poles that have crossed the real axis : @xmath185 where @xmath186 is defined by a contour that passes to the right of the poles at @xmath187 and @xmath188 .
one observes that @xmath186 is nonsingular at @xmath184 , and that @xmath189 near @xmath190 . the singular behavior near @xmath184 is therefore given by @xmath191 we have therefore found that @xmath192 .
our claim i ) is an easy consequence of this fact . in order to verify ii ) , one should observe that the prefactor of the integral in the expression for the fusion coefficients vanishes .
however , the contour of integration gets pinched between the poles from the factors @xmath193 and @xmath194 of the integrand in taking the limit . to isolate the singular contribution of the integral
, one may deform the contour @xmath195 into a contour that goes around the pole at @xmath196 in the right half plane plus a small circle around @xmath196 . due to the vanishing prefactor ,
only the residue contribution survives in the limit .
the rest is straightforward .
the fusion cofficients satisfy two types of symmetry relations : first , one may permute pairs of the variables @xmath197 : @xmath198 these identities follow from similar identities for the b - racah wigner symbols : @xmath199 which are easily derived from the definition of the b - racah wigner symbols given in @xcite taking into account the following properties of the b - clebsch - gordan coefficients : @xmath200_b}\right)^ * = & e^{-\pi i{\alpha}_2^*(q-{\alpha}_2^ * ) } \big[\begin{smallmatrix } q-{\alpha}_1^ * & { \alpha}_2^ * & q-{\alpha}_3^*\\ x_1-c_b & { x_2 } & x_3-c_b \end{smallmatrix}\big]\\ = & e^{+\pi i({\alpha}_3^*(q-{\alpha}_3^*)-{\alpha}_2^*(q-{\alpha}_2^*)-{\alpha}_1^*(q-{\alpha}_1^ * ) ) } { \big [ \,{}^{{\alpha}_3^*}_{x_3}\;{}^{{\alpha}_1^*}_{x_1}\;{}^{{\alpha}_2^*}_{x_2}\ , \big]_b } , \end{aligned}\ ] ] where we have used the notation @xmath201 .
second , there are identities that exchange the two `` internal indices '' with a pair of `` external indices '' @xmath202 the first of these identities is obtained from the pentagon ( [ pentagone ] ) by setting @xmath203 and taking the limit @xmath204 with the help of ( [ limfusion1 ] )
. the second can be obtained from the first by taking into account @xmath205 this identity in turn follows via standard moore - seiberg type arguments @xcite from the fundamental identity that assures crossing symmetry @xcite@xcite , together with ( @xmath206 , @xmath207 ) @xmath208 if one considers the special cases where one of @xmath197 , say @xmath209 equals @xmath210 and where a triple @xmath211 , @xmath212 which contains @xmath213 satisfies the fusion rules of @xcite , one will find that the right hand side of the fusion relation ( [ transfost ] ) reduces to a finite sum of terms selected by the fusion rules of @xcite .
the fusion coefficients that multiply the conformal blocks are residues of the general fusion coefficients , as can be seen by a generalization of our calculation leading to ( [ limfusion1 ] ) . in order to derive explicit expressions for these residues , it is useful to observe that the pentagon equation ( [ pentagone ] ) leads to recursion relations that determines the above - mentioned residues in terms of the following special case : @xmath214 , \nonumber\end{aligned}\ ] ] where @xmath215 .
the explicit expressions for these coefficients are : @xmath216 in subsection [ gcalc ] we need the following fusion coefficients : @xmath217 = { } f_{\sigma_{1},\beta_{2}\pm b } \left [ \begin{smallmatrix } - b & \beta_{2 } \\
\phantom{r}\sigma_{1 } & \sigma_{3 } \end{smallmatrix } \right].\ ] ] the pentagon identity ( [ pentagone ] ) then yields the formula that was used to calculate the expressions used in subsection [ gcalc ] : @xmath218\;=\ ; \sum_{t=\pm}\;\frac{f_{t+}\big(-{{\textstyle \frac{b}{2 } } } |{\beta},{\beta}+s\big ) } { f_{++}\big(-\frac{b}{2}|{\sigma}_2,{\sigma}_2\big)}\ ; f_{-t}\big({\beta}|{\sigma}_2-{{\textstyle \frac{b}{2 } } } , { \sigma}_1\big ) f_{+,s - t}\big({\beta}-{{\textstyle \frac{tb}{2 } } } |{\sigma}_2,{\sigma}_1\big ) . \ ] ]
let us indicate how one may obtain statements on the uniqueness of such equations : we will consider functions @xmath219 that are analytic in some domain @xmath220 that includes @xmath221 $ ] and satisfy @xmath222 where @xmath223 is the operator defined by @xmath224 , as well as the difference equation obtained by replacing @xmath137 .
we would like to show that there exist at most two linearly independent solutions .
assume having three solutions @xmath225 , @xmath226 , @xmath89 of which @xmath225 and @xmath226 are linearly independent .
one may consider @xmath227 the determinant vanishes due to the fact that each row is a linear combination of the two other by means of the difference equation ( [ diffeq ] ) .
but this implies that also the columns must be linearly dependent , in particular @xmath228 with coefficients @xmath229 , @xmath230 that might a priori depend on @xmath68 .
these coefficients are found as @xmath231 @xmath232 can be seen as a q - analogue of the wronskian relevant for second order it differential equations . by a direct calculation using ( [ diffeq ] )
one finds that @xmath233 in a similar way one obtains @xmath234 , @xmath235 .
it then follows for irrational @xmath5 that the @xmath236 , @xmath235 must be constant : from @xmath237 one finds periodicity of @xmath238 in the interval @xmath239 $ ] , so that @xmath236 can be represented by a fourier - series .
one may then use the equation @xmath234 to show vanishing of all fourier - coefficients but the one of the constant mode .
soc . a 196 ( 1901 )
265 - 388 r. behrend , p. pearce , v. petkova , j .- b .
zuber : _ nucl .
phys . _ * b570 * ( 2000 ) 525 - 589 j. cardy , _ nucl .
_ * b324 * ( 1989 ) 581 ; j. cardy , d. lewellen , _ phys .
_ * b259 * ( 1991 ) 274 ; d. lewellen , _ nucl .
phys . _ * b372 * ( 1992 ) 654 - 682 t.curtright , c.thorn .
`` _ conformally invariant quantization of liouville theory
_ phys.rev.lett.48 ( 1982 ) 1309 h.dorn and h .- j.otto.''_two and three point functions in liouville theory _ nucl.phys .
b429 ( 1994 ) , 375 vl.dotsenko , v.fateev .
b241 ( 1984 ) 333 g.felder , j.frhlich , j.fuchs and c.schweigert . `` _ conformal boundary conditions and three dimensional topological field theory _ '' , phys .
lett 84(2000 ) 1659 - 1662 ; preprint hep - th/9912239 g.moore , n.seiberg .
`` _ classical and quantum conformal field theory
phys 123 ( 1989 ) 177 - 254 b.l .
feigin , d.b .
fuchs : representations of the virasoro algebra , in : representations of lie groups and related topics , eds .
a.m. vershik , d.p .
zhelobenko ( gordon and breach , london 1990 ) v.fateev , a.zamolodchikov and al.zamolodchikov .
`` _ boundary liouville field theory .
i. boundary state and boundary two - point function _ '' , hep - th/0001012 . k.hosomichi . `` _ bulk - boundary propagator in liouville theory on a disc _ '' hep - th/0108093 + al.b.zamolodchikov , conference on liouville field theory , montpellier , january 1998 , unpublished .
b.ponsot and j.teschner .
`` _ louville bootstrap via harmonic analysis on a noncompact quantum group . _ '' hep - th/9911110 b.ponsot and j.teschner .
`` _ clebsch - gordan and racah - wigner coefficients for a continuous series of representations of @xmath39 _ '' , math - qa/0007097 , to appear in cmp .
i.runkel.''_boundary structure constants for the a series virasoro minimal models _ nucl .
b549 ( 1999 ) 563 - 578 + j.teschner , unpublished .
j.teschner `` _ liouville theory revisited _ '' , hep - th/0104158 j.teschner . `` _
remarks on liouville theory with boundary _ '' , hep - th/0009138 a.b.zamolodchikov and al.b.zamolodchikov .
_ structure constants and conformal bootstrap in liouville field theory _ nucl.phys .
b477 ( 1996 ) , 577 .
barnes : _ theory of the double gamma function _ , phil .
a196 * ( 1901 ) 265388 t. shintani : _ on a kronecker limit formula for real quadratic fields _ , j. fac .
tokyo sect .
1a math .
* 24 * ( 1977 ) 167199 | liouville field theory is considered on domains with conformally invariant boundary conditions .
we present an explicit expression for the three point function of boundary fields in terms of the fusion coefficients which determine the monodromy properties of the conformal blocks . |
the problem of the existence of periodic orbits for hamiltonian systems has a very far and rich history in the development of mathematics and physics .
in fact , some of the major contributions on this question include the celebrated poincar - birkhoff theorem on the existence of periodic orbits for area preserving twist maps on the annulus and the morse - lyusternik - schnirelman theory on the existence of closed geodesics on compact riemannian manifolds .
recently , a special class of symplectic invariants was introduced by h. hofer and e. zehnder @xcite as a link between rigidity phenomena of symplectic mappings and the existence of periodic orbits provided by classical variational methods .
more precisely , they introduced an axiomatic definition of symplectic capacities in the following way : consider the class of all symplectic manifolds @xmath33 of fixed dimension equal to 2n .
a symplectic capacity is a map @xmath34 that associates to @xmath2 a non negative number or @xmath35 and that satisfies the following axioms : * * monotonicity : * if there exists a symplectic embedding @xmath36 then @xmath37 * * conformality : * @xmath38 where @xmath39 is the ball of radius @xmath40 and @xmath41 is the cylinder over the @xmath42-plane .
a distinguished symplectic capacity was introduced by hofer and zehnder and is directly related to the question of the existence of periodic orbits on prescribed energy levels of hamiltonian flows : [ defchz ] given a symplectic manifold @xmath2 , define the hofer - zehnder capacity of @xmath18 by @xmath43 where @xmath44 is the set of _ admissible hamiltonians _ @xmath45 defined on @xmath18 , that is , * @xmath46 , where @xmath47 is the set of _ pre - admissible hamiltonians _ defined on @xmath18 , that is , @xmath45 satisfies the following properties : @xmath48 , there exist an open set @xmath49 such that @xmath50 and a compact set @xmath51 satisfying @xmath52 ; * every non - constant periodic orbit of @xmath53 has period greater than 1 .
intuitively , this capacity measure the sufficient oscillation for a pre - admissible hamiltonian to have fast periodic orbits , that is , of period less than one .
it is easy to see that it satisfies the monotonicity and conformality properties .
remarkably , by making use of classical variational methods , hofer and zehnder @xcite showed that it also satisfies the non - triviality axiom , giving , in particular , a dynamical new proof of gromov s nonsqueezing theorem @xcite .
moreover , this symplectic invariant has important consequences on the question of the existence of periodic orbits on prescribed energy levels of hamiltonian systems .
in fact , it is easy to prove that if a symplectic manifold @xmath2 has _ bounded hofer - zehnder capacity _ , that is , if @xmath54 for every open subset @xmath55 with compact closure , then given any hamiltonian @xmath56 with compact energy levels , there exists a dense subset @xmath57 such that for every @xmath58 the energy hypersurface @xmath59 has a periodic solution .
this result was improved by m. struwe in @xcite showing that if the energy hypersurfaces bound a compact submanifold then @xmath22 has total lebesgue measure .
this condition is not necessary as was proved recently in @xcite .
note that as an immediate consequence we have a proof of the weinstein conjecture @xcite for manifolds with bounded hofer - zehnder capacity , since we can construct in a neighborhood of a contact hypersurface @xmath60 a distinguished parametrized family of hypersurfaces with equivalent flows .
unfortunately , it is , in general , a difficult task to determine when a symplectic manifold has bounded hofer - zehnder capacity . for @xmath61 , where @xmath62 denotes the canonical symplectic form , it follows by hofer - zehnder theorem @xcite . for surfaces ,
it was proved by k. siburg @xcite , that the hofer - zehnder capacity of an open subset @xmath63 coincides with the area of @xmath63 .
however , the situation is dramatically different for higher dimensional manifolds .
actually , hofer and viterbo showed in @xcite that @xmath64 where @xmath2 is any closed symplectic manifold and @xmath65 is the two dimensional disk of radius @xmath66 , with @xmath40 satisfying the condition @xmath67 , where @xmath68 is the _ index of rationality _ of @xmath2 , that is , @xmath69}\om;\ \ [ u ] \in \pi_2(m ) \text { satisfies } \int_{[u]}\om > 0\bigg\}.\ ] ] if the above set is empty , we define @xmath70 . recall that a 2-form @xmath71 on @xmath18 is _ weakly exact _ if it is closed and @xmath70 .
this result was extended by g. lu @xcite to noncompact symplectic manifolds satisfying certain conditions at the infinity .
more precisely , he showed that @xmath72 provided that @xmath2 is geometrically bounded and @xmath73 .
a symplectic manifold @xmath2 is called _ geometrically bounded _ if there exists an almost complex structure @xmath74 on @xmath18 and a riemannian metric @xmath75 such that * @xmath74 is uniformly @xmath71-tame , that is , there exist positive constants @xmath76 and @xmath77 such that @xmath78 for all tangent vectors @xmath79 and @xmath80 to @xmath18 ; * the sectional curvature of @xmath75 is bounded from above and the injectivity radius is bounded away from zero .
the closed symplectic manifolds are clearly geometrically bounded ; a product of two geometrically bounded symplectic manifolds is also such a manifold .
moreover , it was proved in @xcite that the cotangent bundle of a compact manifold endowed with any twisted symplectic form is geometrically bounded ( a proof of this fact can also be found in @xcite ) . a _ twisted symplectic structure _ on a cotangent bundle @xmath81 is a symplectic form given by @xmath26 , where @xmath62 is the canonical symplectic form , @xmath82 is the canonical projection and @xmath15 is a closed 2-form on @xmath18 .
thus , using the monotonicity property , we can conclude that @xmath83 has bounded hofer - zehnder capacity , where @xmath6 is any compact manifold and @xmath62 is the canonical symplectic form ( because we can embed symplectically any bounded subset of @xmath84 into @xmath85 ) .
actually , by the same reason , we can prove that @xmath86 has bounded hofer - zehnder capacity for any geometrically bounded symplectic manifold @xmath18 .
it should be noted that it is necessary to consider product symplectic forms @xmath87 .
in fact , there is an example due to zehnder @xcite of a symplectic form on @xmath88 with unbounded hofer - zehnder capacity .
note that any symplectic form on @xmath89 is weakly exact .
other known example of cotangent bundle with bounded hofer - zehnder capacity is @xmath90 endowed with any twisted symplectic form @xmath91 , where @xmath15 is a closed 2-form on @xmath92 and @xmath93 is the canonical projection .
it was proved by j. jiang @xcite for @xmath94 and by v. ginzburg and e. kerman @xcite for the general case .
we will consider here a refinement of the hofer - zehnder capacity by considering periodic orbits whose homotopy class is contained in a given subgroup @xmath0 of @xmath95 .
more precisely , we define the hofer - zehnder semicapacity as follows ; it can be compared with the @xmath96-sensitive hofer - zehnder capacity defined by m. schwarz @xcite and lu @xcite and the relative symplectic capacity introduced by p. biran , l. polterovich and d. salamon @xcite : given a symplectic manifold @xmath2 and a subgroup @xmath3 define the _ hofer - zehnder g - semicapacity _ of @xmath18 ( or _ @xmath0-sensitive hofer - zehnder capacity _ ) by @xmath97 where @xmath98 is the set of _ @xmath0-admissible hamiltonians _ defined on @xmath18 , that is , * @xmath99 , that is , @xmath45 is pre - admissible ( see definition [ defchz ] ) ; * every nonconstant periodic orbit of @xmath53 whose homotopy class belongs to @xmath0 has period greater than 1 .
we call it a _ semicapacity _ because given a symplectic embedding @xmath100 such that @xmath101 it can not be expected , in general , that @xmath102 , where @xmath103 is the induced homomorphism on the fundamental group . however , we can state the following _ weak monotonicity property _ ( for a proof see section [ proofthma ] ) : given a symplectic embedding @xmath104 such that @xmath101 then @xmath105 for every subgroup @xmath106 .
in particular , if @xmath107 is injective we have that @xmath108 . note that obviously , @xmath109 for every subgroup @xmath3 . moreover , it can be show that if the hofer - zehnder @xmath0-semicapacity is bounded then there are periodic orbits with homotopy class in @xmath0 on almost all energy levels for every proper hamiltonian @xcite . a symplectic manifold @xmath2 has _ bounded hofer - zehnder @xmath0-semicapacity _ if for every open subset with compact closure @xmath110 , we have that @xmath111 we remark that the periodic orbits in @xmath112 given by the theorems of hofer - viterbo and lu are _
contractible_. thus , in terms of symplectic semicapacities , their theorems state the stronger result that @xmath113 for any geometrically bounded symplectic manifold @xmath2 and @xmath114 . by the weak monotonicity property
, we can conclude then that @xmath115 is bounded , where @xmath116 is the obvious subgroup generated by the circle in the cylinder @xmath117 .
in fact , the kernel of the induced homomorphism on the fundamental group of the symplectic embedding of @xmath118 into @xmath119 is given by @xmath120 .
the aim of the first part of this paper is to generalize this result in the following way : note that @xmath121 has a natural hamiltonian free circle action with hamiltonian given by the angle form @xmath122 viewed here as a function on @xmath123 .
we will consider the following : is the existence of a hamiltonian free circle action sufficient to ensure the boundedness of the hofer - zehnder semicapacity ( with respect to the subgroup of @xmath95 generated by the orbits of the action ) ?
before we state our first result we need a previous topological definition related to free circle actions : let @xmath124 be a free circle action on a connected open subset @xmath6 of a manifold @xmath18 .
we define the _ order of the action _ @xmath125 as the order of the cyclic subgroup of @xmath95 generated by the homotopy class of the orbits of @xmath125 . now , we are able to state our first theorem . in what follows
we will denote the oscillation of a hamiltonian @xmath45 by @xmath126 .
[ thma ] let @xmath2 be a geometrically bounded symplectic manifold and @xmath5 a connected open subset that admits a free hamiltonian circle action @xmath124 ( with period equal to @xmath127 ) given by the hamiltonian @xmath128 .
suppose that the order of this action @xmath129 ( considered as an action on @xmath18 ) satisfies @xmath130 .
given an open subset @xmath131 with compact closure , suppose that @xmath132 then , we have that @xmath133 where @xmath134 is the induced homomorphism on the fundamental group and @xmath135 is the subgroup generated by the orbits of @xmath125 .
* note that the previous theorem can not be stated directly for @xmath6 because we need to assume that @xmath18 is geometrically bounded and it not clear that an open subset of a geometrically bounded symplectic manifold is geometrically bounded . * the circle action above needs to be hamiltonian .
in fact , the example of herman - zehnder @xcite has unbounded hofer - zehnder capacity and it is easy to see that it admits a free symplectic circle action with infinite order .
before we explain the main ideas involved in the proof of theorem [ thma ] , let us state some corollaries and applications .
[ coreulerclass ]
let @xmath2 be a geometrically bounded symplectic manifold .
suppose that @xmath18 admits a free hamiltonian circle action generated by @xmath136 whose euler class @xmath137 \in h^2(m / s^1,\r)$ ] satisfies @xmath137|_{\pi_2(m / s^1 ) } = 0 $ ] .
then , for any open subset @xmath110 with compact closure , we have that @xmath138 provided that @xmath139 .
actually , it can be proved that given a free circle action on a manifold @xmath18 such that the euler class @xmath137 \in h^2(m / s^1,\r)$ ] satisfies @xmath137|_{\pi_2(m / s^1 ) } = 0 $ ] then the order of the action is infinite ( see section [ eulerclass ] ) .
now , consider the finite subgroup @xmath140 .
note that , since the circle action is hamiltonian , the induced action of @xmath141 on @xmath18 is symplectic . since the action is free , the quotient manifold @xmath142 has a induced symplectic structure @xmath143 defined uniquely by the property that the pullback of @xmath143 by the quotient projection is @xmath71 .
moreover , @xmath142 has an induced free hamiltonian circle action ( with period @xmath144 ) with order greater than or equal to @xmath145 .
consequently , we have the following immediate corollary : [ cor1 ] let @xmath2 be a geometrically bounded symplectic manifold .
suppose that @xmath18 admits a free hamiltonian circle action generated by @xmath136 and let @xmath55 be an open subset with compact closure .
if @xmath146 and @xmath147 , then @xmath148 where @xmath149 is the quotient projection , @xmath150 is the inclusion , @xmath71 is the induced symplectic form on @xmath142 and @xmath125 is the induced circle action .
there are a lot of non - trivial examples of symplectic manifolds admitting such free hamiltonian circle actions .
the first one that we will consider is the following : let @xmath18 be a compact manifold admitting a free circle action . then the lift of this action to the cotangent bundle of @xmath18 is hamiltonian with respect to the canonical symplectic form . in fact , it is hamiltonian with respect to any twisted symplectic form @xmath151 , where @xmath15 is a closed two - form on @xmath18 given by the pullback of a closed two - form @xmath152 on @xmath153 .
thus , we have the following corollaries : [ cor2 ] let @xmath18 be a compact manifold admitting a free circle action @xmath125 whose euler class @xmath137 \in h^2(m / s^1,\r)$ ] satisfies @xmath137|_{\pi_2(m / s^1 ) } = 0 $ ] and @xmath15 a weakly exact two - form on @xmath18 given by the pullback of a two - form @xmath152 on @xmath153 .
then the cotangent bundle @xmath154 has bounded hofer - zehnder @xmath155-semicapacity .
let @xmath18 be a compact manifold such that @xmath156 and @xmath157 the total space of a circle bundle @xmath158 .
then @xmath159 has bounded hofer - zehnder @xmath155-semicapacity .
[ cor3 ] let @xmath18 be a compact manifold admitting a free circle action @xmath125 .
then @xmath160 has bounded hofer - zehnder @xmath155-semicapacity for any @xmath161 and any weakly exact 2-form @xmath162 , where @xmath163 is the quotient projection .
it is easy to construct specific non - trivial examples of such circle bundles .
for instance , the lens spaces @xmath164 @xmath165 over @xmath166 , the heisenberg manifold over the 2-torus and , more generally , any 2-step nilmanifold @xcite .
in fact , it can be proved that the group of equivalence classes of circle bundles over a manifold @xmath18 is isomorphic to @xmath167 @xcite .
thus , any non - trivial element of @xmath167 for a manifold @xmath18 such that @xmath156 corresponds to a non - trivial circle bundle with infinite order . the heisenberg manifold , for example , corresponds to the cohomology class of the area form on @xmath89 . in the section [ proofthmb ] ,
we show some applications of the theorem [ thma ] where the circle action is not a lifted action . we will consider circle actions given by the magnetic flow associated to monopoles on surfaces .
more precisely , let @xmath168 be a closed riemannian manifold , @xmath15 a closed 2-form on @xmath18 and @xmath25 endowed with the twisted symplectic form @xmath91 , where @xmath62 is the pullback of the canonical symplectic form via the riemannian metric .
when @xmath18 is a hyperbolic surface and @xmath15 is the area form , this manifold has remarkable properties and is related to mcduff s example of a symplectic manifold with disconnected contact - type boundary @xcite ( see section [ proofthmb ] , where we give a simple construction of such a manifold ) .
the hamiltonian flow @xmath169 given by the kinetic energy @xmath170 with respect to @xmath26 is called the magnetic flow generated by the riemannian metric @xmath75 and the magnetic field @xmath15 .
it can be show that , when @xmath18 is a surface of constant negative curvature and @xmath15 is the area form , the magnetic flow defines a free circle action with infinite order on certain subsets of @xmath25 . using this fact and the expanding completion of convex manifolds developed by eliashberg and gromov @xcite
, we can prove the following theorem : [ thmb ] let @xmath18 be a surface of genus @xmath171 endowed with the hyperbolic metric @xmath172 and the khler form @xmath173 .
let @xmath174 be the usual hamiltonian given by the kinetic energy @xmath175 and @xmath176 a positive real number .
consider the open subset @xmath177 .
then , @xmath178 as long as @xmath179 , where @xmath31 is the subgroup generated by the fibers of the unitary bundle . in particular , the periodic orbits have contractible projection on @xmath18
. on the other hand , we have that @xmath180 for every @xmath181 . by moser s theorem
, we have the following corollary : let @xmath18 be a surface of genus @xmath182 and @xmath15 a symplectic form on @xmath18 . then there exists a riemannian metric @xmath75 on @xmath18 such that @xmath183 as long as @xmath184 , where @xmath185 and @xmath186 is the subgroup generated by the fibers of the unitary bundle as above .
when @xmath18 is the 2-torus and @xmath15 is a symplectic form , it is well know that @xmath187 for every open subset @xmath188 with compact closure .
in fact , it is easy to prove that @xmath189 is symplectomorphic to @xmath190 , where @xmath191 is the canonical symplectic form on @xmath192 @xcite .
this result can also be obtained by the theorem [ thma ] .
now , we will use theorem [ thma ] to show that for closed khler manifolds @xmath193 the possible obstruction to the boundedness of the hofer - zehnder capacity is a very thin subset of @xmath18 which , in many situations , can be explicitly described . to prove it , we will need the following nice decomposition result of p. biran @xcite which enable us to represent a khler manifold as a disjoint union of two basic components whose symplectic nature is very standard : [ bir ] let @xmath9 be a closed khler manifold with @xmath11 \in h^2(m,\z)$ ] and @xmath194 a complex hypersurface whose homology class @xmath195 \in h_{2n-2}(m)$ ] is the poincar dual to @xmath13 $ ] for some @xmath196 .
then , there exists an isotropic cw - complex @xmath197 whose complement - the open dense subset @xmath198 - is symplectomorphic to a standard symplectic disc bundle @xmath199 modeled on the normal bundle @xmath200 of @xmath12 in @xmath18 and whose fibers have area @xmath201 .
the symplectic form @xmath202 is given by @xmath203 where @xmath204 is the bundle projection , @xmath40 is the radial coordinate using a hermitian metric @xmath205 and @xmath206 is a connection form on @xmath19 such that @xmath207 .
the form @xmath208 is uniquely characterized by the requirements that its restriction to the zero section @xmath12 equals @xmath209 , the fibers of @xmath204 are symplectic and have area @xmath201 and @xmath202 is invariant under the obvious circle action along the fibers .
it is called standard because the symplectic type of @xmath210 depends only on the symplectic type of @xmath211 and the topological type of the normal bundle @xmath200 @xcite .
let us recall that the pair @xmath23 is called _ subcritical _
@xcite if @xmath24 is a subcritical stein manifold , that is , if there exists a plurisubharmonic morse function @xmath125 on @xmath24 such that @xmath212 for every critical point @xmath213 of @xmath125 .
it is equivalent to the condition that the dimension of @xmath214 ( that is , the maximal dimension of the cells of @xmath214 ) is strictly less than @xmath145 .
[ thmc ] let @xmath9 be a closed khler manifold ( @xmath10 ) with @xmath11 \in h^2(m,\z)$ ] and @xmath194 a complex hypersurface whose homology class @xmath195 \in h_{2n-2}(m)$ ] is the poincar dual to @xmath13 $ ] for some @xmath196 .
then there exists an open dense subset @xmath19 of @xmath18 symplectomorphic to a standard symplectic disc bundle over @xmath22 whose complement is an isotropic cw - complex @xmath215 such that if @xmath11|_{\pi_2(m)}=0 $ ] then @xmath216 where @xmath217 is the subgroup generated by the orbits of the obvious @xmath218-action on @xmath20 . if @xmath11|_{\pi_2(m ) } \neq 0 $ ] and
@xmath17 , we have that @xmath219 moreover , if @xmath23 is subcritical and @xmath11|_{\pi_2(m)}=0 $ ] , then @xmath220 where @xmath221 . * in @xcite it is proved that the hypothesis that @xmath10 and that either @xmath11|_{\pi_2(m ) } = 0 $ ] or @xmath17 are not necessary to ensure that @xmath222 is finite . *
the result above can not be stated directly for symplectic disk bundles and concluded as a property of symplectic disk bundles plus biran s decomposition result because symplectic disk bundles are not geometrically bounded .
actually , a fundamental ingredient in the proof of theorem [ thmc ] is that , by the lefschetz theorem , the circle action on @xmath20 has the same order considered both as an action on @xmath20 itself and as an action on @xmath24 ( this construction is not necessary in @xcite since we do not need there the topological assumption on the order of the action ) .
* under the hypothesis that either @xmath223 or @xmath11|_{\pi_2(m)}=0 $ ] , it was proved by p. biran @xcite that the _ gromov capacity _ of @xmath19 satisfies @xmath224 .
these assumptions are not necessary as was proved recently by g. lu @xcite . * the result above for _ subcritical _ manifolds
was proved by c. viterbo in @xcite .
* when @xmath23 is subcritical we do not consider the case @xmath17 because , as was proved in @xcite , every subcritical polarization has degree 1 .
we remark that the cw - complex @xmath214 above is given by the union of the stable manifolds of the gradient flow of a plurisubharmonic function @xmath125 defined on @xmath24 such that it can be explicitly computed in many examples @xcite .
the next theorem shows that we can get a similar result in the neighborhood of a symplectic hypersurface of donaldson type @xcite : [ thmd ] let @xmath9 be a closed symplectic manifold ( @xmath10 ) with @xmath11 \in h^2(m,\z)$ ] and @xmath194 a symplectic hypersurface whose homology class @xmath195 \in h_{2n-2}(m)$ ] is the poincar dual to @xmath13 $ ] for some @xmath225 .
then there exists @xmath226 such that if @xmath227 there exists an open neighborhood @xmath228 of @xmath22 such that if @xmath11|_{\pi_2(m)}=0 $ ] then @xmath229 where @xmath230 , as above , is the subgroup generated by the orbits of the obvious @xmath218-action on @xmath231 . if @xmath11|_{\pi_2(m ) } \neq 0 $ ] , we have that @xmath232 it is important to remark that an extension of biran s result for symplectic hypersurfaces considered above ( which is generally expected to remains valid ) will enable us to extend theorem [ thmc ] for general symplectic manifolds . the essential idea in the proof of theorem [ thma ] is to relate the hofer - zehnder capacity of a symplectic manifold endowed with a hamiltonian circle action with the hofer - zehnder capacity of its reduced symplectic manifold ( in the sense of marsden - weinstein ) with respect to this action .
more precisely , we consider the diagonal hamiltonian circle action on @xmath233 whose reduced symplectic manifold @xmath234 is given by @xmath2 . then , given a pre - admissible hamiltonian @xmath235 we construct a @xmath218-invariant pre - admissible hamiltonian @xmath236 on @xmath157 whose reduced dynamics is given by a reparametrization of the hamiltonian vector field of @xmath45 on @xmath18 .
thus , we can apply the results of hofer - viterbo and lu to get a periodic orbit for @xmath237 and so , by reduction , a periodic orbit for @xmath53 .
the essential step is to show the non - triviality of the projected closed orbit and it is here that the hypothesis on the order of the action plays an essential role .
the idea is to use the condition on the homotopy of the periodic orbit together with an upper bound on the period of the orbit to ensure that the orbit can not be tangent to the trajectories of the diagonal action . in the second part of this paper
, we will consider the problem of the existence of periodic orbits on prescribed energy levels for a special class of hamiltonian dynamical systems given by the magnetic flows .
this problem was first considered by v. arnold @xcite and s. novikov @xcite and for a discussion of the results in this area we refer to @xcite .
it was proved by l. polterovich @xcite that for every nontrivial weakly exact magnetic field on a manifold whose euler characteristic vanishes , there exist non - trivial contractible closed orbits of the magnetic flow in a sequence of arbitrarily small energy levels .
the proof uses the geometry of the hofer s metric in the group of hamiltonian diffeomorphisms and the fundamental fact that the displacement energy of the zero section of @xmath25 , with respect to the twisted symplectic form given by a non - vanishing magnetic field , is equal to zero .
recently , e. kerman @xcite gives the same result for magnetic fields given by symplectic forms .
in @xcite we proved the existence of contractible closed orbits for almost all low energy levels , provided that the magnetic field is also symplectic .
we will prove here an extension of polterovich s theorem for any manifold , without the assumption on the euler characteristic .
[ thme ] let @xmath18 be any closed riemannian manifold and @xmath15 a non - trivial weakly exact magnetic field .
then there exists a sequence of arbitrarily small energy levels containing non - trivial contractible periodic orbits .
the main idea in the proof is similar to that of theorem [ thma ] .
in fact , we consider a lift of the magnetic flow to @xmath84 and use the topological condition on the periodic orbit to ensure the non - triviality of the projected periodic orbit by symplectic reduction .
the details are given in the section [ proofthme ] .
i am very grateful to gabriel paternain for many helpful comments and suggestions and to the centro de investigacin en matemtica , guanajuato , mexico for the hospitality while part of this work was in progress .
i am also grateful to kai cieliebak , dusa mcduff and felix schlenk for very useful comments .
before we begin the proof of theorem [ thma ] we need some auxiliary results .
[ hzin2 ] let @xmath2 be a geometrically bounded symplectic manifold and @xmath238 endowed with the symplectic form @xmath239 , where @xmath240 is the canonical symplectic form on @xmath117 .
let @xmath241 be an open subset with compact closure and suppose that @xmath242 , where @xmath243 is the projection onto the second factor . then given a hamiltonian @xmath244
, there exists a nonconstant periodic orbit @xmath245 of @xmath53 with period @xmath246 moreover , the homotopy class @xmath247 $ ] of @xmath248 belongs to the subgroup @xmath249 .
since @xmath250 is compact , there exists a positive constant @xmath251 such that @xmath252 \subset t^*s^1 $ ] .
let @xmath62 be the canonical symplectic form on @xmath192 and consider the symplectomorphism @xmath253 \to a : = \{(x , y ) \in \r^2 ; a \leq x^2+y^2 \leq 5a \}$ ] given by @xmath254 note that @xmath255 , where @xmath256 ) \simeq \pi_1(t^*p ) \to \pi_1(t^*m \times \r^2 ) \simeq \pi_1(t^*m)$ ] is the homomorphism induced on the fundamental group by the transformation given by the identity and @xmath169 in the first and second factors respectively .
thus , we can suppose , without loss of generality , that @xmath261 is a connected two - dimensional compact submanifold with boundary , such that there exists an open disk of radius @xmath262 with @xmath263 distinct points ( @xmath264 ) @xmath265 and an orientation preserving diffeomorphism @xmath266 such that @xmath267 from a theorem of dacorogna and moser ( see lemma 2.2 in @xcite ) this @xmath268 can be required to be symplectic .
thus , consider the hamiltonian @xmath269 given by @xmath270 which is obviously @xmath271 , since @xmath272 , where @xmath273 is a compact subset such that @xmath274 .
then it is sufficient to prove that @xmath288 . to do it ,
suppose , by contradiction , that there exists @xmath289 such that @xmath290 has a nonconstant periodic orbit @xmath245 such that @xmath291 \in \psi_*g$ ] and the period of @xmath245 is less than 1 .
then the image of @xmath245 by @xmath292 ( note that the image of @xmath245 is contained in @xmath293 , since @xmath294 outside @xmath293 ) is a periodic orbit of @xmath295 of the same period and @xmath296 \in \psi_*^{-1}(\psi_*g),\ ] ] a contradiction .
[ [ construction - of - a - reparametrized - diagonal - circle - action - rho - on - n - times - ts1-and - its - symplectic - reduction . ] ] construction of a reparametrized diagonal circle action @xmath297 on @xmath298 and its symplectic reduction .
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ let @xmath299 , where @xmath300 is endowed with the symplectic form @xmath301 , where @xmath240 is the canonical symplectic form on @xmath117 .
note that @xmath117 has a natural hamiltonian circle action given by the hamiltonian @xmath302 .
let @xmath303 be such that @xmath304 and @xmath305 .
consider the diagonal action @xmath306 on @xmath157 given by the hamiltonian @xmath307 and let @xmath308 be the diffeomorphism given by @xmath309 where @xmath124 is the flow generated by @xmath310 .
the factor @xmath145 is not important in this section and it will be used only in the next section in the construction of a @xmath297-invariant pre - admissible hamiltonian from @xmath45 . to show that @xmath315 , it is more convenient to write @xmath268 as @xmath316 where @xmath317 is the projection onto the circle and @xmath318 is the unit vector field tangent to the fibers of @xmath117 such that @xmath319 .
consequently , @xmath320 where @xmath321 is the hamiltonian vector field generated by @xmath310 and @xmath322 is the pullback of the angle form @xmath323 on @xmath218 to @xmath157 .
thus , @xmath324 but note that @xmath325 and @xmath326 since the vectors are colinear . on the other hand , @xmath327 and @xmath328 because the vectors are orthogonal with respect to the product decomposition @xmath329 and by the definition of @xmath311 .
we have then that , @xmath330 where the last equality follows from the fact that @xmath331 .
now , note that @xmath332 because @xmath333 and @xmath334 . on the other hand
, we have that @xmath335 because @xmath336 and @xmath337 ( note that @xmath322 here is the pullback of the angle form @xmath323 on @xmath218 to @xmath157 ) .
consequently , @xmath338 as desired .
we claim that the quotient projection @xmath342 is given by @xmath343 , where @xmath344 is the projection onto the first factor . in effect , note that , by the previous lemma , @xmath268 sends the orbits of @xmath345 to the orbits of @xmath346 . on the other hand , @xmath96 is the quotient projection at @xmath347 with respect to the trivial circle bundle given by the orbits of @xmath348 .
[ [ construction - of - a - rho - invariant - pre - admissible - hamiltonian - hh - on - n - times - ts1-from - h . ] ] construction of a @xmath297-invariant pre - admissible hamiltonian @xmath236 on @xmath298 from @xmath45 .
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ in view of the last lemma , consider the map @xmath351 defined by @xmath352 such that @xmath353 is the quotient projection . now , let @xmath45 be a pre - admissible hamiltonian on @xmath63
. we will construct from @xmath45 a @xmath297-invariant hamiltonian defined on @xmath157 .
firstly , fix a sufficiently small constant @xmath354 satisfying @xmath355 define the new hamiltonian by @xmath356 where @xmath357 is a @xmath271 function such that @xmath358 , @xmath359 , @xmath360 and @xmath361 for every @xmath362 $ ] ( see figure [ graph1 ] ) .
thus , @xmath363 is the lift of @xmath364 by the quotient projection @xmath365 .
it is easy to see that @xmath368 where @xmath369 is given by ( see figure [ graph2 ] ) @xmath370 in effect , it is clear that @xmath371 .
moreover , @xmath372 where @xmath373 is the subset of @xmath157 given by @xmath374 and @xmath49 is an open set such that @xmath50 ( that exists because @xmath375 ) since @xmath376 . note that the interior of @xmath373 is not empty .
[ [ existence - of - a - non - trivial - periodic - orbit - of - x_h - given - by - the - reduction - of - a - periodic - orbit - of - x_hh . ] ] existence of a non - trivial periodic orbit of @xmath53 given by the reduction of a periodic orbit of @xmath237 .
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ let us now prove that there exists a periodic orbit @xmath248 of @xmath53 of period less than @xmath380 and homotopy class @xmath247 \in g_\vr$ ] .
note firstly that @xmath381 in effect , the image of @xmath382 under the projection @xmath383 is the graph of @xmath384 restricted to @xmath63 and the projection of @xmath385 is the graph of @xmath386 : note that @xmath387 has the hamiltonian vector field with respect to @xmath71 given by @xmath388 .
thus , the reduced dynamics of @xmath237 at @xmath339 is a reparametrization of the dynamics of @xmath53 . since @xmath389 , the nonconstant periodic orbits of @xmath390 have period greater than or equal to those of @xmath53 .
consequently , by the lemma [ mwreduction ] , it is sufficient to show that the periodic orbit of @xmath236 given by the proposition [ hzin2 ] is not tangent to a fiber given by the action of @xmath125 ( such that it is projected onto a singularity of @xmath53 ) .
suppose that there exists a periodic orbit @xmath391 \to \u$ ] of @xmath237 tangent to an orbit of @xmath297 and whose homotopy class belongs to @xmath279 .
we will show that the period @xmath392 of @xmath248 is strictly greater than @xmath393 ( see proposition [ hzin2 ] ) .
now , let @xmath395 be an almost complex structure on @xmath157 compatible with @xmath311 and @xmath396 be the corresponding riemannian metric .
we have that along @xmath248 , @xmath397 on the other hand , the period of the orbits of @xmath345 is equal to @xmath398 because the action on @xmath18 is free .
consequently , since @xmath247 \in \pi_1(s^1)$ ] , the projection @xmath399 must wrap around itself at least @xmath129 times .
hence , @xmath245 also must wrap around itself at least @xmath129 times and so the period of @xmath248 is given by @xmath400 but , by the definition of @xmath145 , @xmath401 actually , note that the inequality @xmath402 is equivalent to @xmath403 that is , @xmath404 which is exactly the condition on @xmath145 and @xmath405 .
thus , it is sufficient to prove that @xmath406 which implies @xmath407 to conclude this inequality , note that @xmath408 where in the second equation we used the fact that @xmath395 defines an isometry . on the other hand
, we have that for every @xmath409 , @xmath410 but note that , since @xmath248 is tangent to a fiber , it is projected onto a singularity of @xmath53 , that is , @xmath411 for every @xmath412 .
consequently , we have that @xmath413 as desired .
finally , note that the periodic orbit @xmath276 of @xmath237 given by the proposition [ hzin2 ] satisfies @xmath414 \in \pi_1(s^1)$ ] .
consequently , the corresponding reduced periodic orbit @xmath248 satisfies @xmath247 \in g_\vr$ ] since @xmath415 .
let @xmath18 be a surface of genus @xmath182 endowed with a riemannian metric @xmath75 of constant negative curvature .
let @xmath25 be the tangent bundle of @xmath18 with the symplectic form @xmath62 given by the pullback of the canonical symplectic form of @xmath81 by the bundle isomorphism induced by @xmath75 .
consider on @xmath18 the almost complex structure @xmath74 such that @xmath443 is the khler form .
as defined in the introduction , the magnetic flow associated to @xmath75 and @xmath15 is the hamiltonian flow given by the kinetic energy @xmath170 with respect to the twisted symplectic form @xmath444 .
it is easy to see that a curve @xmath445 defines an orbit of this flow if and only if @xmath446 . in other words ,
the trajectories are the curves with constant geodesic curvature .
consequently , the dynamics of this magnetic flow can be explicitly described : on the energy levels greater than one half , the magnetic flow is topologically equivalent to the geodesic flow ( which has periodic orbits in every nontrivial homotopy class and does not have projected contractible periodic orbits , because there is no conjugate points ) ; on the sphere bundle , it coincides with the horocycle flow and hence is minimal , that is , all the orbits are dense ; and on the energy levels below than one half , all the orbits are closed and homotopic to the fibers of the sphere bundle . in particular , it defines ( after a suitable reparametrization ) a hamiltonian circle action with infinite order on @xmath447 ( since @xmath156 ) such that we can apply theorem [ thma ] .
the problem is the condition on @xmath448 to be geometrically bounded . to circumvent this problem
we need the following proposition : to fix the notation , let @xmath456 , where @xmath457 is the curvature map , be the pullback of the canonical symplectic form of @xmath81 by the riemannian metric and @xmath458 . define @xmath459 and @xmath460 .
consider the liouville form @xmath461 and the 1-form @xmath462 note that @xmath463 and @xmath464 , where the last equality follows from the fact that @xmath465 defines a connection form on @xmath466 and that @xmath18 is a hyperbolic surface .
let us compute the lie derivative of @xmath467 with respect to @xmath468 .
initially , we have that , @xmath469 such that @xmath470 but , we have that @xmath471 consequently , @xmath472 thus , @xmath473 and so @xmath474 on the other hand , the lie derivative of @xmath452 with respect to @xmath475 is given by @xmath476 since @xmath477 consequently , we have that @xmath478 it was proved in @xcite that the tangent bundle of a compact manifold endowed with any twisted symplectic form is geometrically bounded .
thus , consider the almost complex structure @xmath74 on @xmath25 such that @xmath481 defines a riemannian metric with bounded sectional curvature and positive injectivity radius . by the previous lemma , given any @xmath482 sufficiently small , the subset @xmath483 has contact type boundary , that is , its boundary is convex in the sense of eliashberg - gromov @xcite .
thus , we can consider the set @xmath484 ( that we can identify with @xmath448 ) and a symplectic form @xmath450 on @xmath485 such that @xmath450 coincides with @xmath452 on @xmath486 and the vector field @xmath487 is an extension of @xmath453 to @xmath488 ( that is , @xmath489 coincides with @xmath490 on @xmath491 ) such that @xmath492 .
note that @xmath450 is exact , since @xmath452 on @xmath486 is exact and @xmath450 restricted to @xmath493 is obviously exact .
consider now the almost complex structure @xmath494 on @xmath485 given by @xmath495 on @xmath486 and by the pushforward @xmath496 of @xmath74 by the flow @xmath125 of @xmath487 on @xmath497 . by the construction
, @xmath494 is invariant by @xmath125 on @xmath498 .
define @xmath499 . by our previous discussion
, we have that @xmath500 defines a geometrically bounded riemannian metric on @xmath486 .
we need to prove that it also defines a geometrically bounded riemannian metric on @xmath493 .
but , we have that @xmath501 , that is , @xmath502 consequently , since @xmath494 is invariant by @xmath125 on @xmath497 , we obtain that @xmath503 thus , the sectional curvature @xmath504 of @xmath500 on @xmath505 is given by @xmath506 and the injectivity radius @xmath507 is greater than or equal to @xmath508 but , @xmath500 has bounded sectional curvature and positive injectivity radius on @xmath486 , proving the proposition . by our previous discussion ,
the magnetic flow restricted to an energy level @xmath181 is topologically conjugated to the geodesic flow such that it has no projected contractible periodic orbits .
consequently , we have that @xmath509 for every @xmath181 . on the other hand
, we have that for energies @xmath510 the ( reparametrized ) magnetic flow defines a free hamiltonian circle action ( of period 1 ) on @xmath27 of infinite order on @xmath448 . by the previous proposition , given @xmath482 sufficiently small
, there exists a geometrically bounded symplectic form @xmath450 on @xmath511 that coincides with @xmath452 on @xmath512 .
thus , by theorem [ thma ] we have that for every @xmath354 there exists a finite constant @xmath513 such that , for every @xmath482 sufficiently small , @xmath514 but , @xmath515 before we give the proof of theorem [ thmc ] , we need some previous lemmas
. the first one can be derived by the arguments in the proof of proposition [ tame ] and the fact @xmath24 has a convex boundary .
however , for the reader s convenience we present an independent proof using the structure of stein manifolds : [ tame1 ] given a neighborhood @xmath63 of @xmath12 , there exists an exact geometrically bounded symplectic form @xmath297 on @xmath24 such that @xmath297 coincides with @xmath15 on the complement of @xmath63 .
note that @xmath12 can be obtained as the zero set of a holomorphic section @xmath516 of a complex line bundle defined by the divisor @xmath12 .
it can be show that the function @xmath517 given by @xmath518 is plurisubharmonic @xcite .
the next lemma proved in @xcite shows that we can construct a new plurisubharmonic function on @xmath24 that coincides with @xmath125 on @xmath519 and whose gradient vector field is complete .
since @xmath18 is closed , we have that @xmath533 defines a geometrically bounded riemannian metric on @xmath534 , that is , the sectional curvature of @xmath533 restricted to @xmath534 is bounded from above and the injectivity radius is positive .
we need to prove that it also defines a geometrically bounded riemannian metric on all @xmath24 .
given @xmath535 , there exists @xmath536 such that @xmath537 for some @xmath538 , where @xmath539 is the gradient flow of @xmath525 with respect to @xmath533 .
but , we have that @xmath540 , that is , @xmath541 consequently , since the complex structure @xmath74 is invariant by @xmath539 , we obtain that @xmath542 thus , the sectional curvature @xmath504 of @xmath533 on @xmath505 is given by @xmath506 and the injectivity radius @xmath507 is greater than or equal to @xmath508 but , @xmath533 has bounded sectional curvature and positive injectivity radius on @xmath534 , proving that @xmath543 is geometrically bounded .
[ nullset ] suppose that @xmath18 is subcritical and let @xmath545 be an open subset with compact closure .
let @xmath297 be a geometrically bounded symplectic form on @xmath24 given by the lemma [ tame1 ] .
then there exists a symplectomorphism isotopic to the identity @xmath546 such that @xmath547 .
remember that the isotropic cw - complex @xmath214 is given by the union of the stable manifolds of the gradient vector field @xmath548 of a certain plurisubharmonic function @xmath125 with a complete gradient flow such that @xmath549 @xcite .
since @xmath18 is subcritical , we have that @xmath550 and so there exists a hamiltonian isotopy @xmath551 compactly supported in an arbitrarily small neighborhood of @xmath214 such that @xmath552 .
since @xmath214 is compact , there exists a neighborhood @xmath228 of @xmath214 such that @xmath553 .
fix @xmath392 so large that @xmath554 . since @xmath555 moves @xmath228 away from itself
, we have that @xmath556 because @xmath557 leaves @xmath558 invariant . since @xmath548 is conformally symplectic , we have that @xmath559 is the desired symplectomorphism
. by theorem [ bir ] , there exists an isotropic cw - complex @xmath215 ( given by the union of the stale manifolds of the gradient vector field associated to a plurisubharmonic function defined on @xmath560 ) whose complement @xmath561 is symplectomorphic to a standard symplectic disc bundle @xmath199 over @xmath22 modeled on the normal bundle @xmath562 of @xmath22 in @xmath18 , where @xmath563 moreover , this symplectomorphism @xmath564 sends @xmath22 to the zero section of @xmath565 .
note that @xmath566 has an obvious hamiltonian free circle action ( of period @xmath127 ) generated by the hamiltonian @xmath567 such that @xmath568 also generates a hamiltonian free circle action on @xmath569 . by the lefschetz theorem @xcite
, we have that the inclusion @xmath570 induces an injective map on the fundamental group such that the order of the circle action on @xmath571 is equal to the order of the circle action @xmath572 generated by @xmath573 on @xmath574 . since the chern class of @xmath565
is given by @xmath575 $ ] , we have that if @xmath11_{\pi_2(m)}=0 $ ] then , by the proposition [ propeulerclass ] and the lefschetz theorem , the order of the action generated by @xmath310 is infinite . on the other hand ,
the order of the action is always grater than or equal to @xmath28 . in effect , note that the @xmath218-bundle @xmath576 is equivalent to the quotient by @xmath577 of a @xmath218-bundle @xmath578 with euler class @xmath579 $ ] such that the subgroup of @xmath580 generated by the orbits of the action has order greater than or equal to @xmath28 .
let @xmath450 be the exact geometrically bounded symplectic form on @xmath24 given by the lemma [ tame1 ] that coincides with @xmath15 on the complement @xmath581 of a tubular neighborhood of radius @xmath582 of @xmath12 .
note that @xmath583 also has a hamiltonian free circle action that coincides with @xmath584 on @xmath581 .
now , let @xmath585 be an open subset with compact closure and @xmath482 sufficiently small such that @xmath586 .
then , by the theorem [ thma],we have that if @xmath11_{\pi_2(m)}=0 $ ] then , @xmath587 where @xmath588 is the subgroup generated by the homotopy class of the orbits of @xmath584 . on the other hand ,
if @xmath11_{\pi_2(m ) } \neq 0 $ ] and @xmath17 we have the inequality @xmath589 finally , suppose that @xmath23 is subcritical and let @xmath590 be an open subset with compact closure and @xmath482 sufficiently small such that @xmath591 .
then , by the lemma [ nullset ] , there exists a symplectomorphism isotopic to the identity @xmath592 such that @xmath593 .
take @xmath482 sufficiently small such that @xmath594 .
thus , by the weak monotonicity property , we conclude that @xmath595 in fact , by the symplectic neighborhood theorem , there exists a neighborhood @xmath228 of @xmath12 and a symplectomorphism @xmath596 between @xmath597 and a neighborhood @xmath395 of the zero section of the canonical disc bundle @xmath199 , where @xmath563 moreover , @xmath598 sends @xmath22 to the zero section of @xmath565 .
now , notice that @xmath453 point towards the interior of the compact domains bounded by the hypersurfaces @xmath601 , that is , the subset @xmath602 has a convex boundary in the sense of eliashberg - gromov @xcite , where @xmath603 and @xmath604 is the radial coordinate of @xmath605 .
thus , we can consider the set @xmath606 and a symplectic form @xmath450 on @xmath485 such that @xmath450 coincides with @xmath202 on @xmath607 and the vector field @xmath608 is a complete extension of @xmath453 to @xmath609 ( that is , @xmath610 coincides with @xmath490 on @xmath611 ) such that @xmath612 .
note that @xmath614 has an obvious hamiltonian free circle action ( of period @xmath127 ) generated by the hamiltonian @xmath615 such that @xmath568 also generates a hamiltonian free circle action on @xmath616 . by the lefschetz theorem for symplectic hypersurfaces of donaldson type @xcite
, there exists @xmath617 such that if @xmath227 then the inclusion @xmath618 induces an injective map on the fundamental group and so the order of the circle action on @xmath619 is equal to the order of the circle action @xmath572 generated by @xmath573 on @xmath620 .
let @xmath621 . since the chern class of @xmath565
is given by @xmath575 $ ] , we have that , as in the proof of theorem [ thmc ] , if @xmath11_{\pi_2(m)}=0 $ ] then the order of the action generated by @xmath310 is infinite . on the other hand ,
the order of the action is always grater than or equal to @xmath28 .
let @xmath450 be the geometrically bounded symplectic form on @xmath24 given by the lemma [ tame2 ] that coincides with @xmath15 on the complement @xmath581 of a tubular neighborhood of radius @xmath582 of @xmath12 .
now , let @xmath626 be a relative compact open subset and @xmath482 sufficiently small such that @xmath627 .
then , by the theorem [ thma ] , we have that if @xmath11_{\pi_2(m)}=0 $ ] then , @xmath587 where @xmath628 is the subgroup generated by the homotopy class of the orbits of @xmath584 .
analogously , if @xmath11_{\pi_2(m ) } \neq 0 $ ] and @xmath629 we have the inequality @xmath589 a diffeomorphism of a symplectic manifold @xmath2 is called hamiltonian if it is given by the time one map of a time - dependent hamiltonian vector field .
let @xmath630 be the group of hamiltonian diffeomorphisms of @xmath2 .
it was discovered by hofer @xmath631 that this group carries a natural biinvariant finsler metric with a non - degenerate distance function .
this distance is defined as follows .
the lie algebra @xmath632 of @xmath630 consists of all smooth functions on @xmath18 satisfying a normalization condition which ensures that , when @xmath18 is open , @xmath633 iff @xmath598 has compact support , and , when @xmath18 is closed , @xmath633 iff @xmath598 has zero mean with respect to the volume form induced by @xmath71 .
the adjoint action of @xmath630 on @xmath632 is the standard action of diffeomorphisms on functions .
consider the @xmath634-norm in @xmath632 , @xmath635 this norm is invariant under the adjoint action , and thus defines a biinvariant finsler metric on @xmath630 .
now , let @xmath636 , @xmath637 $ ] , a path of hamiltonian diffeomorphisms , that is , a curve @xmath638 \to ham(m,\om)$ ] .
let @xmath639 be its normalized hamiltonian function , that is , @xmath640 for all @xmath641 .
we define then the length of this curve by @xmath642 the hofer s metric is then defined by @xmath643 where the infimum is taken over all paths @xmath248 which join @xmath169 and @xmath268 . for
a subset @xmath63 of @xmath18 define the displacement energy @xmath644 as the measure of the distance between the identity map and the set of @xmath645 which displaces @xmath63 from itself , in the sense that @xmath646 . using the geometry of the hofer s metric and that , for a manifold whose euler characteristic vanishes
, the displacement energy of the zero section of @xmath81 , with the twisted symplectic form given by a non - vanishing magnetic field , is equal to zero , polterovich @xcite proved the following theorem : [ polterovich @xcite ] let @xmath18 be a closed manifold whose euler characteristic vanishes and @xmath15 a non - vanishing weakly exact magnetic field .
then there exist contractible magnetic closed orbits on a sequence of arbitrarily small energy levels .
the idea of the proof is the following @xcite : take a smooth function @xmath647 , @xmath648 , which equals @xmath649 on @xmath650 $ ] , vanishes on @xmath651 and is strictly increasing on @xmath652 .
thus , the closed orbits of the ( normalized ) hamiltonian @xmath653 corresponds to reparametrized magnetic periodic orbits whose energy is less than @xmath654 .
the key point is that the minimum set of @xmath598 coincides with the zero section of @xmath81 and so its displacement energy vanishes @xcite .
it implies that the asymptotic non - minimality of @xmath598 is strictly less than 1 and thus that @xmath598 ( as a vector in the lie algebra @xmath632 ) does not generate a minimal geodesic in @xmath630 @xcite .
then , using a result of lalonde and mcduff @xcite , we conclude that the hamiltonian @xmath598 has a non - constant contractible closed orbit , since @xmath15 is weakly exact .
it should be noted that the contractibility of the periodic orbits given by lalonde and mcduff follows by the fact that the proof of theorem 5.4 of @xcite relies on the hofer - zehnder capacity - area inequality ( theorem 1.17 of @xcite ) proved by floer , hofer and viterbo @xcite , where the periodic orbits are contractible ( see @xcite ) .
consider the trivial @xmath218-bundle @xmath655 , the flat connection @xmath656 and the product metric @xmath657 .
let @xmath658 be the pullback of the magnetic field @xmath15 to @xmath157 .
the lifted circle action to @xmath659 is hamiltonian with respect to the twisted symplectic form @xmath660 , where @xmath661 is the canonical projection .
in fact , let @xmath662 the vector field on @xmath659 that generates the circle action .
then , @xmath663 where @xmath656 above is viewed as a function in @xmath659 . moreover , the form @xmath656 is invariant by the magnetic flow of @xmath659 , because taking a magnetic orbit @xmath248 in @xmath157 , we have that @xmath664 where @xmath665 denotes the lorentz force of the magnetic field @xmath658 ( defined as the bundle map @xmath666 uniquely characterized by the property that @xmath667 ) and in the third equation we used the fact that @xmath453 is parallel . thus , the magnetic flow on @xmath659 is equivariant with respect to the hamiltonian lifted action of @xmath218 to @xmath659 . on the other hand ,
it is easy to see that the marsden - weinstein reduced symplectic manifold with respect to the circle action on @xmath668 is given by @xmath669 .
moreover , the pullback of the hamiltonian metric @xmath670 in @xmath25 by @xmath671 is given by @xmath672 .
consequently , the reduced dynamics of the magnetic flow @xmath673 of @xmath157 on @xmath674 is given by the magnetic flow @xmath675 on @xmath18 .
now , suppose that @xmath15 is weakly exact .
since @xmath157 is the product , the fibers of @xmath157 are non - contractible and , in fact , the subgroup generated by the fibers in @xmath676 is isomorphic to @xmath677 .
since the euler characteristic of @xmath157 vanishes and @xmath678 is weakly exact ( because @xmath679 ) , we can apply polterovich s theorem to conclude that @xmath673 has contractible closed orbits @xmath680 of arbitrarily small energy .
since the subgroup in @xmath676 generated by the fibers of @xmath157 is cyclic infinite , these orbits can not be given by the fibers .
this ensures that they are projected on non - constant contractible periodic orbits of @xmath675 . finally , note that the energy of these projected orbits is equal to @xmath681 where @xmath682 , because @xmath683 as @xmath145 grows to infinite .
y. eliashberg , m. gromov , _ convex symplectic manifolds _ several complex variables and complex geometry , part 2 ( santa cruz , ca , 1989 ) , 135 - 162 , proc . sympos .
pure math .
, 52 , part 2 , amer .
soc . , providence , ri , 1991 .
m. herman , _
exemples de flots hamiltoniens do nt aucune perturbation en topologie @xmath271 na dorbites periodiques sur un ouvert de surfaces denergies _ , comptes rendus de lacadmie des sciences de paris * 412 * ( 1991 ) , 989994 .
initially , note that @xmath11|_{\pi_2(m)}=0 $ ] iff @xmath420 = 0 $ ] , where @xmath421 is the universal cover . in effect , it is clear that if @xmath422 is exact , then @xmath11|_{\pi_2(m)}=0 $ ] . to prove the converse , note that , by the hurewicz s theorem , we have that @xmath423
thus , the condition that @xmath15 is weakly exact is equivalent that the integral of @xmath420 $ ] over any 2-cycle is equal to zero .
but , by de rham s theorem , it is equivalent to @xmath422 be exact . to prove the claim , note that , since the fiber is path connected , @xmath427 is simply connected
thus , it is sufficient to prove that @xmath427 covers @xmath18 .
now , observe that @xmath429 where @xmath430 is the group of deck transformations on @xmath425 . on the other hand
, @xmath424 commutes with the action of @xmath430 , since it is a lifted action from @xmath157 .
consequently , it induces an action of a discrete group @xmath431 on @xmath427 such that @xmath432 proving the claim . | we introduce the hofer - zehnder @xmath0-semicapacity @xmath1 of a symplectic manifold @xmath2 ( or @xmath0-sensitive hofer - zehnder capacity ) with respect to a subgroup @xmath3 ( @xmath4 ) and prove that given a geometrically bounded symplectic manifold @xmath2 and an open subset @xmath5 admitting a hamiltonian free circle action with order greater than two then @xmath6 has bounded hofer - zehnder @xmath0-semicapacity , where @xmath7 is the subgroup generated by the orbits of the action , provided that the index of rationality of @xmath2 is sufficiently great ( for instance , if @xmath8|_{\pi_2(m)}=0 $ ] ) .
we give several applications of this result . using p. biran s decomposition theorem ,
we prove the following : let @xmath9 be a closed khler manifold ( @xmath10 ) with @xmath11 \in h^2(m,\z)$ ] and @xmath12 a complex hypersurface representing the poincar dual of @xmath13 $ ] , for some @xmath14 .
suppose either that @xmath15 vanishes on @xmath16 or that @xmath17 .
then there exists a decomposition of @xmath18 into an open dense subset @xmath19 such that @xmath20 has finite hofer - zehnder @xmath0-semicapacity and an isotropic cw - complex , where @xmath21 is the subgroup generated by the obvious circle action on the normal bundle of @xmath22 .
moreover , we prove that if @xmath23 is subcritical then @xmath24 has finite hofer - zehnder @xmath0-semicapacity .
we also show that given a hyperbolic surface @xmath18 and @xmath25 endowed with the twisted symplectic form @xmath26 , where @xmath15 is the khler form on @xmath18 , then the hofer - zehnder @xmath0-semicapacity of the domain @xmath27 bounded by the hypersurface of kinetic energy @xmath28 minus the zero section @xmath29 is finite if @xmath30 , where @xmath31 is the subgroup generated by the fibers of @xmath32 . finally , we will consider the problem of the existence of periodic orbits on prescribed energy levels for magnetic flows .
we prove that given any weakly exact magnetic field @xmath15 on any compact riemannian manifold @xmath18 then there exists a sequence of contractible periodic orbits of energy arbitrarily small , extending a previous result of l. polterovich . |
accretion disks are believed to be the energy source of various activities in active galactic nuclei , x - ray binaries , protostars and so on . using the rxte satellite ,
detailed timing analysis of x - ray binaries has been carried out ( e.g. , @xcite , @xcite , @xcite ) .
black hole candidates ( bhcs ) show transitions between a low / hard state ( lhs ) dominated by the hard power - law component and a high / soft state ( hss ) characterized by the soft black body component . during the transition , bhcs
stay in hard intermediate state ( hims ) or in a soft intermediate state ( sims ) @xcite .
the light curves during lhs of bhcs are subject to violent x - ray fluctuations and sometimes accompany quasi - periodic oscillations ( qpo ) in the fourier power spectral density ( psd ) .
the lhs is associated with steady outflows ( e.g. , @xcite ) . in this state ,
mass accretes to the black hole as an optically thin , advection - dominated accretion flow ( e.g. , @xcite ; @xcite , ) .
the energy spectrum of the hims is softer than that in the lhs .
low frequency ( @xmath18 ) qpos are observed in the luminous hard state and in the intermediate state @xcite .
high frequency qpos ( @xmath19 ) are sometimes observed in these states .
the frequency of low - frequency qpos in the intermediate state moves to higher frequency and the low - frequency qpos disappear when the state changes to hss ( see @xcite ) .
these observations indicate that qpos are associated with the cooling of the disk . by carrying out global three - dimensional ( 3d ) magneto - hydrodynamic ( mhd ) simulations including radiative cooling
, @xcite showed that when the accretion rate exceeds the limit for the onset of the cooling instability , the radiatively inefficient , optically thin disk transitions into a magnetically supported , cool , intermediate state .
@xcite constructed a steady model of such disks and showed that their luminosity can exceed @xmath20 , where @xmath21 is the eddington luminosity .
when the transition to the cool disk takes place in the outer region , cool gas will be supplied to the inner region .
low frequency qpos are sometimes attributed to the oscillation at the interface between the hot inner disk and the cool outer disk ( e.g. , @xcite ) .
the excitation mechanism of such an oscillation , however , was not clear .
another puzzle is the coexistence of low - frequency and high - frequency qpos .
@xcite proposed that high - frequency qpos are generated by the resonance between radial and vertical oscillations of accretion disks .
s. kato ( 2001a , b ) pointed out that non - axisymmetric g - mode oscillations can be trapped in a thin , relativistic disk and these oscillations are excited by the corotation resonance .
non - linear couplings of disk oscillations and disk warp were examined by @xcite .
@xcite reproduced high - frequency qpos by 3d mhd simulations of optically thin , geometrically thick accretion flows .
they showed that the qpos appear around the epicyclic frequency @xmath22 at @xmath23 and @xmath24 , where @xmath25 is the kepler frequency and showed that the amplitude of qpos are damped after a while .
the @xmath26-noise - like fluctuations observed in bhcs have been reproduced by global 3d mhd simulations ( e.g. , @xcite , @xcite ) .
@xcite showed that the psd changes its slope around @xmath27 hz when they assumed a @xmath28 black hole .
@xcite pointed out that magnetic reconnection in the innermost region of the disk can be the origin of intermittent x - ray flares known as x - ray shots ( @xcite ) , which produce flat psd at low - frequency ( @xmath29 hz ) . in this paper , we report the results of global 3d mhd simulations which produced low - frequency qpos and discuss their excitation mechanisms . in section [ method ] , we describe basic equations and initial conditions .
the results of simulations are given in section [ result ] . in section [ form ] , we concentrate on the oscillation excited in the inner torus . section [ discuss ] is devoted to discussion and conclusion .
we solved the following resistive mhd equations in a cylindrical coordinate system @xmath30 ; @xmath31 @xmath32 = - \nabla p - \rho \nabla \phi + \frac{\mbox{\boldmath $ j$ } \times \mbox{\boldmath $ b$}}{c } ~ , \label{eqn : b2}\ ] ] @xmath33 @xmath34 where @xmath35 , @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 , and @xmath43 are the density , pressure , gravitational potential , velocity , magnetic field , current density , resistivity , temperature and specific entropy , respectively .
the specific entropy is expressed as @xmath44 , where @xmath45 is the specific heat capacity and @xmath46 is the specific heat ratio .
we included the joule heating term but neglected the radiative cooling term in the energy equation .
we assume the anomalous resistivity @xmath47 ^ 2 $ ] @xcite , where @xmath48 is the electron ion drift speed and @xmath49 is the threshold above which anomalous resistivity sets in .
general relativistic effects are simulated using the pseudo - newtonian potential @xmath50 @xcite , where @xmath51 is the gravitational constant , @xmath52 is the mass of the black hole , @xmath53 , and @xmath3 is the schwarzschild radius .
we neglect the self - gravity of the disk .
we solved the resistive mhd equations using a modified lax
wendroff scheme @xcite with an artificial viscosity @xcite .
the units of length and velocity are the schwarzschild radius @xmath3 and the light speed @xmath54 , respectively .
the unit time is @xmath55 .
the unit temperature is given by @xmath56 , where @xmath57 is the proton mass and @xmath58 is the boltzmann constant .
the number of grids is @xmath59 .
the grid size is @xmath60 for @xmath61 , and @xmath62 . for model
lt , we set the grid interval as follows ; @xmath63 , @xmath64 , where @xmath65 and @xmath66 .
the outer boundaries at @xmath67 and at @xmath68 are free boundaries where waves can be transmitted . for model
ht , we set @xmath69 and @xmath70 . therefore , the outer boundaries are located at @xmath71 and @xmath72 , respectively .
the grid size in the azimuthal direction is @xmath73 .
we included the full circle @xmath74 in the simulation region , and applied periodic boundary conditions in the azimuthal direction .
an absorbing boundary condition is imposed at @xmath75 by introducing a damping factor , @xmath76 the physical quantities @xmath77 inside @xmath78 are re - evaluated by @xmath79 which means that the deviation from initial values @xmath80 is artificially reduced with damping rate @xmath81 .
waves propagating inside @xmath82 are absorbed in the transition region ( @xmath83 ) .
the initial state of our simulation is an equilibrium torus threaded by a weak toroidal magnetic field . at the initial state , the torus is assumed to have a specific angular momentum , @xmath84 .
the magnetic field distribution is determined according to @xcite . by using the polytropic relation @xmath85 at the initial state and by assuming @xmath86 where @xmath87 is the initial plasma @xmath88 at the initial pressure maximum of the torus @xmath89 , and @xmath90 is the azimuthal magnetic field .
we integrated the equation of motion into a potential form , @xmath91 where @xmath92 is the sound speed , @xmath93 is the alfvn speed , and @xmath94 . at @xmath95 , the rotation speed of the torus
@xmath96 equals the keplerian velocity . by using equation ( [ eqn : poten ] ) ,
we obtain the density distribution as @xmath97 } } { k[\gamma/(\gamma-1 ) ] [ 1+\beta_{\rm b}^{-1}\varpi^{2(\gamma-1 ) } /\varpi_{\rm b}^{2(\gamma-1 ) } ] } \right\}^{1/(\gamma -1 ) } ~ , \ ] ] where @xmath98 is the density at @xmath89 . outside the torus
, we assumed a hot , isothermal ( @xmath99 ) spherical halo .
the density distribution of the halo is given by @xmath100 $ ] , where @xmath101 is the gravitational potential at @xmath102 .
in this paper , we report the results of simulations for two models .
model ht assumes a hot outer torus with the sound speed @xmath103 at @xmath104 . in model
ht , the torus has a constant specific angular momentum ( @xmath105 ) .
model lt is a cool disk model in which @xmath106 at @xmath107 and @xmath108 . in both models
, we adopted @xmath109 , @xmath110 , @xmath111 , @xmath112 , @xmath113 , and @xmath114 .
since we do not include radiative cooling , @xmath98 is arbitrary .
we adopt @xmath115 .
( 120mm,90mm)f1.eps figure [ fig : f1]a shows the time evolution of mass accretion rate for model ht measured at @xmath116 .
figure [ fig : f1]b shows the time evolution of mass accretion rate at @xmath116 ( black ) and at @xmath117 ( gray ) .
the mass accretion rate @xmath15 at @xmath118 is computed by @xmath119 the mass accretion rate at @xmath120 is measured by integrating the accretion rate in the equatorial region @xmath121 .
the unit of the mass accretion rate is @xmath122
. mass accretion rate at @xmath118 becomes quasi - steady after about 10 rotational periods at the initial density maximum .
the increase in mass accretion rate saturates when @xmath123 .
the mass accretion takes place due to the efficient angular momentum transport by maxwell stress in an mri - driven turbulent field . the equatorial mass accretion rate at @xmath124 shows time variation with a timescale of @xmath125 .
figure [ fig : f1]c shows the time evolution of mass outflow rate measured at @xmath126 computed by @xmath127 the mass outflow rate correlates with the mass accretion rate with a time lag of about @xmath128 .
( 120mm,90mm)f2.eps figure [ fig : f2]a shows the correlation between the mass accretion rate at @xmath118 and the mass outflow rate measured at @xmath129 for model ht . the correlation is computed by @xmath130 d t~.\ ] ] here , @xmath131 and @xmath132 are average during the interval @xmath133 . in figure 2 , we adopted @xmath134 and @xmath135 .
the correlation function has peaks at @xmath136 and @xmath137 . the mass outflow rate at @xmath138 correlates with the mass accretion rate at @xmath118 with delay of @xmath139 .
this timescale is comparable to the time scale of the propagation time of the outflow whose average speed is @xmath140 .
figure [ fig : f2]b shows the self correlation function of the mass outflow rate at @xmath129 in the time interval @xmath141 .
in addition to the peak at @xmath142 , positive peaks appear around @xmath143 and @xmath144 .
this indicates that mass outflow rate oscillates quasi - periodically with period @xmath145 .
( 120mm,90mm)f3.eps figure [ fig : f3 ] shows the isosurface of the density and the vertical velocity .
blue surfaces and green surfaces depict the density isosurface ( @xmath146 ) and the isosurface of vertical velocity ( @xmath147 ) , respectively .
winds emerge intermittently from the inner region of the accretion disk .
toroidal magnetic fields are dominant in the wind .
the intermittent ejection is driven by expansion of magnetic loops anchored to the accretion disk @xcite .
( 120mm,120mm)f4.eps figure [ fig : f4]a shows the time evolution of magnetic energy integrated in @xmath148 , and @xmath149 normalized by the initial gas pressure at @xmath150 .
figure [ fig : f4]b shows the time evolution of @xmath151 averaged in @xmath152 , and @xmath149
. figures [ fig : f4]c and [ fig : f4]d show the time evolution of @xmath153 and the ratio of the gas pressure to magnetic pressure @xmath154 , respectively .
although magnetic energy decreases , the plasma @xmath88 is nearly constant ( @xmath155 ) because gas pressure also decreases .
the fluctuations of the magnetic energy and radial magnetic field correlate with the mass accretion rate .
( 120mm,90mm)f5.eps figure [ fig : f5]a shows the time evolution of the mass accretion rate for model lt measured at @xmath156 .
figure [ fig : f5]b shows the mass accretion rate at @xmath157 ( black ) and the equatorial mass accretion rate at @xmath158 ( gray ) which enlarges the time range @xmath159 . figure [ fig : f5]c shows the time evolution of the mass outflow rate measured at @xmath126 . in figure
[ fig : f5]a , @xmath160 is an order of magnitude smaller than that for model ht .
this result indicates that the angular momentum transport rate strongly depends on the temperature of the gas supplied from the outer region .
( 120mm,90mm)f6.eps figure [ fig : f6]a shows the correlation between the mass accretion rate at @xmath156 and the mass outflow rate at @xmath161 in the interval @xmath162 for model lt .
the mass outflow rate correlates with the mass accretion rate with time delay of @xmath163 .
figure [ fig : f6]b shows the self correlation of the mass outflow rate at @xmath161 in the interval @xmath164 .
positive peaks appear at @xmath165 , and @xmath166 .
it indicates that the mass outflow rate oscillates with period @xmath167 .
( 120mm,90mm)f7.eps figure [ fig : f7 ] shows the time evolution of ( a ) magnetic energy , ( b ) @xmath168 , ( c ) angular momentum transport rate @xmath169 , and ( d ) plasma @xmath88 for model lt .
the magnetic energy is averaged in the region @xmath170 , @xmath171 , and @xmath172 . magnetic energy gradually decreases when @xmath173 .
figure [ fig : f7]c shows that @xmath7 in model lt . since magnetic energy decreases , plasma @xmath88 increases and stays around @xmath174 .
physical quantities shown in figure [ fig : f7 ] show short time scale oscillations and longer time scale ( @xmath125 ) time variations .
the latter is due to the time variation of the mass accretion rate from the outer region .
( 120mm,120mm)f8.ps figure [ fig : f8 ] shows the snap shots of the density distribution .
figures [ fig : f8]a and [ fig : f8]b show the density distribution in @xmath175 plane for model ht and lt , respectively .
the density is averaged in the azimuthal direction .
figures [ fig : f8]c and [ fig : f8]d show the density distribution in @xmath176 plane averaged in @xmath149 . in model lt
, an inner torus is created around @xmath177 .
the inner torus is deformed into a crescent - like shape .
the inner torus is formed because angular momentum transport becomes inefficient .
figures [ fig : f7]a and [ fig : f7]b indicate that maxwell stress decreases due to the decrease in magnetic energy .
magnetic energy decreases partly because magnetic flux is swallowed into the black hole with accreting gas and partly because magnetic energy dissipates by magnetic reconnection .
the deformation of the inner torus into a crescent shape takes place due to the growth of the papaloizou - pringle instability ( @xcite , @xcite )
. when the disk is hot , since the inner torus is not formed , disk gas accretes to the black hole through dense , spiral channels ( figure [ fig : f8]c ) .
this result is consistent with the result obtained by simulations of the hot disk reported by @xcite .
( 120mm,90mm)f9.eps figure [ fig : f9 ] shows the time - averaged radial distribution of physical quantities .
the black curves show the results for model ht averaged in @xmath178 .
the gray curve shows the result for model lt averaged in @xmath179 .
figure [ fig : f9]a displays the specific angular momentum distribution .
the dotted curve shows the keplerian specific angular momentum using pseudo - newtonian potential .
since angular momentum continuously decreases in model ht , no inner torus is formed in this model . in model
lt ( dashed curve ) , since @xmath169 is small , nearly constant angular momentum , inner torus is formed in @xmath180 .
figure [ fig : f9]b shows the radial distribution of @xmath181 . inside the inner torus , the angular momentum transport rate becomes very small .
solid and dashed curves in figure [ fig : f9]c show the sound speed and alfvn speed , respectively .
figure [ fig : f9]d shows the radial velocity .
the accretion proceeds subsonically in @xmath182 . in model
lt , accretion speed becomes very low in the inner torus .
( 120mm,120mm)f10.eps figure [ fig : f10 ] shows the distribution of azimuthally averaged temperature ( [ fig : f10]a , [ fig : f10]b ) , and vertical velocity ( [ fig : f10]c , [ fig : f10]d ) . figures [ fig : f10]a and [ fig : f10]b show that accreting matter is significantly heated up .
figures [ fig : f10]c and [ fig : f10]d show that mass outflow emerges from the disk with an average speed of @xmath183 .
the outflow is anti - symmetric to the equatorial plane .
the outflow is more powerful in model ht . by comparing figures [ fig : f8]a and [ fig : f10]a , we can distinguish three regions ; equatorial disk , hot funnel near the rotation axis , and outflows between the disk and the funnel .
accretion proceeds in the equatorial disk .
a fraction of the accreting matter is ejected from the inner region of the disk .
( 120mm,120mm)f11.eps figure [ fig : f11 ] shows the distribution of the azimuthal magnetic field ( color ) and magnetic fields depicted from the poloidal components of magnetic fields averaged in the azimuthal direction ( curves ) .
the azimuthal magnetic field is antisymmetric to the equatorial plane at this stage ( @xmath184 for model
ht and @xmath185 for model lt ) .
magnetic fields are turbulent inside the disk but show more coherent structures in the interface between the disk and the halo where mass outflow takes place .
magnetic field lines are stretched along this interface .
a large - scale poloidal magnetic field is created in the funnel near the rotation axis .
( 120mm,120mm)f12.eps figure [ fig : f12 ] shows the snapshots of the density distribution on the @xmath186 plane for model lt .
density is averaged in the same region as that in figures [ fig : f8]c and [ fig : f8]d .
the inner torus deforms its shape from a circle into a crescent , and from a crescent into a circle , repeatedly .
( 120mm,120mm)f13.eps the top panel of figure [ fig : f13 ] plots the time evolution of the ratio of the joule heating rate to the magnetic energy averaged in @xmath187 , @xmath149 , and @xmath188 .
the ratio increases when magnetic energy is released ( when @xmath189 decreases or @xmath190 increases ) .
the ratio changes quasi - periodically with a time scale of @xmath191 .
the bottom panel of figure [ fig : f13 ] shows the time evolution of the amplitude of non - axisymmetric @xmath192 mode ( @xmath193 is the mode number in the azimuthal direction ) of the density .
the amplitude of the azimuthal mode is computed by fourier decomposing the density contrast @xmath194 the running average of the amplitude is plotted using the amplitude during @xmath195 for each point .
the gray curve depicts the same curve as that in figure [ fig : f13 ] .
the amplitude of @xmath192 mode anti - correlates with @xmath196 .
this indicates that the magnetic energy is released when the @xmath192 mode disappears .
the amplitude of the @xmath192 mode also correlates with @xmath197 ( figure [ fig : f7]b ) and @xmath169 ( figure [ fig : f7]c ) .
they all show long timescale ( @xmath198 ) variation and short timescale ( @xmath199 ) oscillations .
the peaks of the short timescale oscillations in the fourier amplitude of the @xmath192 mode in figure [ fig : f13 ] coincide with those in figure [ fig : f7 ] .
( 120mm,120mm)f14.eps figure [ fig : f14 ] shows the mass accretion rate measured at @xmath200 ( black ) , magnetic energy ( gray ) , and the fourier amplitude of the @xmath192 mode for the density ( dashed ) for model lt . when the magnetic energy increases inside the inner torus , the mass accretion rate increases because the angular momentum transport rate increases as magnetic energy ( and magnetic stress ) is accumulated in the torus . on the other hand , when the magnetic energy is released , the mass accretion rate decreases .
the amplitude of the @xmath192 mode ( dashed curve in figure [ fig : f14 ] ) correlates with the magnetic energy ( gray ) .
the magnetic fields are amplified when the amplitude of the @xmath192 mode increases . on the other hand ,
the amplitude of the @xmath192 mode decreases when the magnetic energy is released .
when the maxwell stress decreases due to the decrease in magnetic energy , the papaloizou & pringle instability grows again inside the torus .
therefore , the torus deforms itself into a crescent shape .
the interval between the magnetic energy releases in the inner torus is @xmath191 .
( 120mm,90mm)f15.eps figure [ fig : f15]a shows the correlation between the amplitude of the @xmath192 mode and the joule heating rate normalized by the magnetic energy for model ht . the correlation function has positive peaks at @xmath201 and @xmath202 .
the positive correlation at @xmath203 indicates that magnetic energy is released when @xmath192 spiral channel ( see figure [ fig : f8]c ) develops .
this result is consistent with that reported in @xcite , where we showed that magnetic reconnection takes place in the spiral channel .
other peaks in the correlation function indicate that @xmath192 mode develops quasi - periodically with interval @xmath204 .
figure [ fig : f15]b shows the correlation between the amplitude of the @xmath192 mode and the joule heating rate for model lt .
in contrast to that in model ht , no positive peak appears at @xmath203 .
instead , negative peak appears at @xmath205 and positive peaks appear at @xmath206 .
it indicates that @xmath192 mode anti - correlates with the joule heating rate and that the growth of the @xmath192 mode precedes the release of the magnetic energy by @xmath207 .
( 120mm,90mm)f16.eps figure [ fig : f16 ] shows the spatial distribution of the fourier amplitude @xmath208 of time variabilities in mass accretion rate for model ht in @xmath209 , and for model lt in @xmath210 . in model ht ,
various peaks appear at various radius . on the other hand , in model lt ,
low frequency qpos around @xmath27hz appear in @xmath211 , where the inner torus is formed .
( 120mm,90mm)f17.eps figure [ fig : f17 ] shows the power spectral density ( psd ) of the time variation of the mass accretion rate averaged in @xmath212 and @xmath213 .
black and gray curves show psd ( @xmath208 where @xmath214 is the fourier power ) for model lt and model ht , respectively .
we adapted the data in the time range @xmath209 in model ht and @xmath210 in model lt .
the psd for model lt has a broad low - frequency peak around @xmath10 .
this low - frequency peak corresponds to the oscillation involving amplification and release of magnetic energy in the inner torus .
the psd for model ht is flat in @xmath215 , which means that @xmath216 and changes its slope around @xmath217 .
since oscillations are excited at various radius in model ht , psd shows flat spectrum .
the psd in model lt has a slope steeper than that in model ht in @xmath218 . in model lt in figure
[ fig : f17 ] , a small peak appears at @xmath219 . based on the simulation using 32 azimuthal grid points , @xcite pointed out that a high - frequency qpo appears when the disk shows low - frequency ( @xmath220 ) sawtooth - like oscillations of magnetic energy .
here we confirmed that low - frequency qpos and high - frequency qpos coexist by a simulation including twice as many grid points in the azimuthal direction .
in this paper , we studied the dependence of the structure and time variation of the black hole accretion flows on the gas temperature supplied from the outer region .
when hot gas ( @xmath221 ) is supplied , the angular momentum is transported efficiently all the way to the black hole .
the average ratio of the maxwell stress to gas pressure , @xmath222 , consistent with previous simulations by @xcite and @xcite .
such disks correspond to the radiatively inefficient , optically thin disk in the low / hard state of black hole candidates .
our simulation clearly showed the appearance of outflows with a maximum speed of @xmath223 from such hot disks .
the mass outflow rate @xmath224 is comparable to the mass accretion rate to the black hole .
the magnetic fields are turbulent inside the disk but they show large - scale coherent structures near the rotation axis .
plasma flows out in the interface between the funnel and the accretion disk .
the hot accretion flow shows time variations whose psd is flat in @xmath225 when the mass of the central black hole is @xmath226 .
when cool plasma is supplied , an inner torus is formed around @xmath227 .
such a torus is formed when the mri - generated turbulent magnetic field is dissipated inside the disk .
since maxwell stress decreases due to the decrease of magnetic energy , @xmath169 decreases to @xmath228 .
thus a nearly constant - angular - momentum torus is formed in the innermost region ( @xmath229 ) of the accretion flow .
we found that the inner torus deforms itself into a crescent shape .
such deformation takes place due to the growth of the non - axisymmetric instability in geometrically thick tori @xcite .
such a non - axisymmetric structure enhances the growth of mri . as magnetic energy increases , the angular momentum transport rate increases .
thus , the accretion rate increases .
when the magnetic energy accumulated in the disk is released , the disk comes back to the weakly magnetized , axisymmetric torus .
new cycle begins as magnetic energy is amplified by mri .
the period of the cycle is about @xmath230 in model lt .
it creates a low - frequency peak around @xmath231 in psd of the mass accretion rate .
low - frequency qpos sometimes observed in low / hard state and hard intermediate state ( hims ) of black hole candidates can be reproduced by such magnetic cycles .
we also showed that a high - frequency qpo is excited when low - frequency qpo appears .
let us discuss why the magnetic cycle is excited when low - temperature gas is supplied .
when hot gas accretes , since the scale height of the gas is large , large eddies are formed in the magnetically turbulent disk and create a large scale coherent magnetic field which transports angular momentum efficiently . on the other hand ,
when low - temperature gas is supplied , since the scale height and eddy size become smaller , angular momentum is transported only in the local region .
thus the angular momentum is transported more efficiently in high temperature disks .
low temperature plasmas can be supplied when cooling instability takes place in the outer disk .
@xcite carried out a global 3d mhd simulation of black hole accretion flows by including radiative cooling .
they showed that when the density of the outer disk is sufficiently high , cooling instability takes place .
the outer disk shrinks vertically , and forms a magnetically supported , cool , optically thin disk .
when such cool plasma accretes , the magnetic cycle may be excited in the inner region .
the amplitude of the @xmath192 non - axisymmetric mode correlates positively with the joule heating rate in model ht but anti - correlates in model lt . in model
ht , since the angular momentum transport rate is large enough , mass smoothly accretes along the spiral channel , in which the magnetic fields are stretched and form a current sheet .
magnetic energy is released in such current sheets .
therefore , the amplitude of the @xmath192 mode correlates with the joule heating rate . on the other hand , in model lt ,
non - axisymmetric @xmath192 pattern disappears when the magnetic energy is released .
@xcite pointed out that during the transition from the lhs to hss , gx339@xmath2324 shows a sub - transition from hims dominated by power - law x - ray radiation to soft intermediated state ( sims ) dominated by radiation from an optically thick disk .
they also showed that the low - frequency qpos appear when the x - ray spectrum stays in hims and sims and sometimes in lhs .
the high - frequency qpos appear in the hims and sims .
when the high - frequency qpo is observed , a low - frequency qpo is always observed in the x - ray spectrum .
this tendency is consistent with our simulation results .
a number of theoretical models have been proposed for high - frequency qpos .
@xcite proposed that high - frequency qpos are formed by the resonance of radial and vertical oscillations .
@xcite studied the excitation of high - frequency qpos by resonance with the disk warp . in our simulation
, a high - frequency peak appears around @xmath233 when the low - frequency oscillation is prominent . in our simulation , however , the low - frequency oscillation was weakened due to the heating of the inner torus .
we expect that when the extra heating is extracted by radiative cooling , the magnetic cycle may continue and excite high - frequency qpos .
we would like to report the results of numerical simulations including radiative cooling in subsequent papers . in this work ,
we treated the relativistic effect by using the pseudo - newtonian potential .
the accuracy of this approximation is worst near the black hole s horizon especially for spinning black holes .
when we include the effects of black hole spin , the inner torus will be formed closer to the black hole .
thus , higher frequency qpos will appear due to the oscillation of the inner torus .
we should note that disk luminosity depends on the relativistic beaming and light bending . in subsequent papers
, we would like to report psds obtained by relativistic ray - tracing of numerical results .
we are grateful to m.a .
abramowicz , s. kato , w. kulniak , and m. burusa for discussion .
numerical computations were carried out on vpp5000 at center for computational astrophysics , cfca of naoj ( p.i .
this work is supported in part by japan society for the promotion of science ( jsps ) research fellowships for young scientists ( mm : 18 - 1907 ) , and in part by grants - in - aid for scientific research of the ministry of education , culture , sports , science , and technology ( rm : 17030003 ) . | we present the results of global three dimensional magneto - hydrodynamic simulations of black hole accretion flows .
we focus on the dependence of numerical results on the gas temperature @xmath0 supplied from the outer region .
general relativistic effects are taken into account using the pseudo - newtonian potential .
we ignore the radiative cooling of the accreting gas .
the initial state is a torus whose density maximum is at @xmath1 or @xmath2 from the gravitating center , where @xmath3 is the schwarzschild radius .
the torus is initially threaded by a weak azimuthal magnetic field .
we found that mass accretion rate and the mass outflow rate strongly depend on the temperature of the initial torus .
the ratio of the average maxwell stress generated by the magneto - rotational instability ( mri ) to gas pressure , @xmath4 is @xmath5 in the hot torus ( @xmath6 at @xmath2 ) and @xmath7 in the cool torus ( @xmath8 at @xmath1 ) . in the cool model ,
a constant angular momentum inner torus is formed around @xmath9 .
this inner torus deforms itself from a circle to a crescent quasi - periodically . during this deformation , the mass accretion rate , the magnetic energy and the maxwell stress increase . as the magnetic energy is released , the inner torus returns to a circular shape and starts the next cycle .
power spectral density ( psd ) of the time variation of the mass accretion rate in the cool model has a low frequency peak around @xmath10 when we assumed a 10@xmath11 black hole .
the psd of the hot model is flat in @xmath12 .
the slope of the psd in the cool model is steeper than that in the hot model in @xmath13 .
the mass outflow rate in the low temperature model also shows quasi - periodic oscillation .
intermittent outflows are created in both models .
the outflow speed is @xmath14 .
the mass outflow rate is comparable to the mass accretion rate @xmath15 at @xmath16 in the high temperature model and about @xmath17 in the low temperature model . |
the yang baxter equation originally appeared in statistical physics and in quantum integrable systems , and its study has then led to the discovery of quantum groups , which have found numerous applications in mathematics and mathematical physics ( for literature on the yang
baxter equation and related subjects , see , _
e.g. _ , @xcite and references therein ) . the fusion procedure for the yang baxter equation allows the construction of new solutions of the yang baxter equation starting from a given fundamental solution .
we refer to the new solutions as `` fused '' solutions of the yang baxter equation .
the fusion procedure has first been introduced in @xcite .
several aspects of the fusion procedure for the yang baxter equation and of its applications to theoretical physics can be found for example in @xcite . in this paper
we start by reviewing the fusion procedure for an arbitrary fundamental solution of the yang baxter equation , together with the identification and the study of distinguished invariant subspaces for the fused solutions .
we provide complete proofs as the details are scattered in the literature and as , moreover , often only some particular solutions , depending on the physical model under consideration , were considered .
let @xmath0 be a finite - dimensional complex vector space .
starting from an arbitrary solution @xmath1 of the yang
baxter equation on @xmath2 ( see section [ sec - prel ] for more precision ) , a fused solution is built given two sets of complex numbers @xmath3 and @xmath4 and is denoted by @xmath5 .
the element @xmath5 is an endomorphism of the vector space @xmath6 .
linear subspaces @xmath7 and @xmath8 are constructed , with the property that the fused solution @xmath5 preserves the subspace @xmath9 of @xmath10 .
thus , by restriction , the operator @xmath5 induces a solution of the yang
baxter equation acting on the space @xmath9 .
a crucial feature of the invariant subspaces is that , for any @xmath3 , the subspace @xmath11 is defined as the image of an endomorphism @xmath12 of @xmath13 , and moreover the endomorphism @xmath12 is expressed in terms of the original solution @xmath1 .
the second part of the paper is devoted first to the review of the identification of interesting spaces among the spaces @xmath11 when @xmath1 is the so - called yang solution .
then we generalize the obtained results to the case where @xmath1 is a generalization of the yang solution for a @xmath14-graded vector space @xmath0 ( _ i.e. _ a super " vector space ) , and to the case where @xmath1 is the standard deformation of the yang solution , both in the usual and @xmath14-graded situations . for the yang solution ,
the main result is the identification of subspaces @xmath11 which are isomorphic to finite - dimensional irreducible representations of the general linear lie algebra @xmath15 , where @xmath16 is the dimension of @xmath0 .
it results in a family of fused solutions of the yang
baxter equation acting on finite - dimensional irreducible representations of @xmath15 .
let @xmath17 be a partition of length less than or equal to @xmath16 .
a subspace @xmath11 isomorphic to the irreducible representation of highest weight @xmath17 is obtained when the set of numbers @xmath18 is a string of classical contents associated to a standard young tableau of shape @xmath17 .
we show moreover that , by replacing the set @xmath18 by a string of contents associated to another standard young tableau of the same shape , we obtain an equivalent fused solution up to a change of basis .
as it is suggested in @xcite , the study of the subspaces @xmath11 when @xmath1 is the yang solution is made using the classical schur weyl duality ( see , _ e.g. _ , @xcite ) together with the so - called `` fusion formula '' for the symmetric group . the fusion formula for the symmetric group @xmath19 originates in @xcite and has then been developed in @xcite , see also @xcite .
the fusion formula expresses each element of a complete set of primitive idempotents of the group ring @xmath20 as a certain evaluation of a rational function in several variables with coefficients in @xmath20 ( the rational function is the same for all the idempotents ) .
as the name suggests , the fusion formula is intimately related to the fusion procedure .
indeed the rational function is built from elementary `` universal '' solutions of the yang
baxter equation , called baxterized elements of @xmath20 , and it turns out that its expression reproduces the expression for the operator @xmath12 , mentioned above , in terms of @xmath1 .
when @xmath1 is the generalization of the yang solution for a @xmath14-graded vector space @xmath21 , we still can use the fusion formula for the symmetric group @xmath19 .
however , in this situation , another action of @xmath19 on @xmath13 is relevant .
this action gives rise to a super " analogue of the classical schur
weyl duality @xcite , which holds between the symmetric group and the general linear lie superalgebra @xmath22 , where @xmath23 and @xmath24 . with these tools , generalizing the preceding results concerning the yang solution , we obtain a family of fused solutions of the yang baxter equation acting on finite - dimensional irreducible representations of @xmath22 . for the standard deformation of the yang solution , we use the fusion formula for the hecke algebra @xcite together with the quantum analogue of the schur
weyl duality @xcite .
this duality holds between the hecke algebra and the quantum group @xmath25 associated to @xmath15 .
then we apply the fusion procedure to obtain a family of fused solutions of the yang
baxter equation acting on finite - dimensional irreducible representations of @xmath25 .
the parameters @xmath18 are now related to quantum contents of standard young tableaux . here again , replacing a standard young tableau by another standard young tableau of the same shape leads to equivalent fused solutions up to a change of basis .
the standard deformation of the yang solution also admits a generalization for a @xmath14-graded vector space @xmath0 and we obtain in this case , via the fusion procedure , solutions of the yang baxter equation acting on finite - dimensional irreducible representations of a deformation @xmath26 of @xmath27 .
in this situation , we use the analogue of the schur - weyl duality between the hecke algebra and @xmath26 proved in @xcite . for an arbitrary fundamental solution @xmath1
, we study the possible equivalence of the fused solutions under a permutation of the parameters @xmath3 . by `` equivalence '' of two fused solutions ,
we mean that the two corresponding invariant subspaces have the same dimension , and moreover the restrictions of the fused solutions to these subspaces coincide up to a change of basis .
assuming an unitary property for @xmath1 , we obtain that particular permutations ( that we call `` admissible '' for @xmath18 and @xmath1 , see subsection [ subsec - dep ] ) of the parameters @xmath3 lead to equivalent fused solutions . for the four solutions @xmath1 considered above ,
the fact that the fused solutions depend on the standard young tableaux only through their shapes is a consequence of this general property .
we notice that , even for solutions @xmath1 considered in this paper , the study of ( proper ) invariant subspaces for the fused solutions is , though quite satisfactory , not fully complete .
namely , it is still an open problem to determine all the sets of parameters @xmath18 such that the operator @xmath12 is non - invertible ( and thus leads to a proper invariant subspace ) and to study the obtained fused solutions .
even the classification of the sets of parameters @xmath18 such that the operator @xmath12 is proportional to a projector is not completely understood .
one motivation for a detailed study of the fusion procedure for an arbitrary solution of the yang
baxter equation came from recent results generalizing the fusion formula of the symmetric group and of the hecke algebra to several other structures : the brauer algebras @xcite , the birman
murakami algebras @xcite , the complex reflection groups of type @xmath28 @xcite , the ariki koike algebras @xcite and the wreath products of finite groups by the symmetric group @xcite .
these fusion formulas involve as well a rational function ( with coefficients in the algebras under consideration ) which is built from elementary universal " solutions of the yang
baxter equation , the analogues for these structures of the baxterized elements of @xmath20 .
however , in these fusion formulas , baxterized solutions of the reflection equation also appear in the rational functions .
it is therefore expected , as it is already mentioned in @xcite , that these fusion formulas admit an interpretation in the framework of the fusion procedure for the reflection equation .
we postpone to another article an analogous study of the fusion procedure for the reflection equation and of invariant subspaces of the fused solutions of the reflection equation .
the paper is organized as follows . in section [ sec - prel ] ,
we give necessary definitions and notations . in section [ sec - fus ] , starting from an arbitrary solution @xmath1 of the yang
baxter equation acting on @xmath2 , we describe the construction of the endomorphisms @xmath5 and prove that the obtained operators @xmath5 are solutions of the yang
baxter equation as well . in section [ sec - inv ]
, we define endomorphisms @xmath12 and show that their images provide invariant subspaces for the fused solutions .
we also give an alternative formula for @xmath12 and prove the equivalence of the fused solutions under admissible permutation of the parameters @xmath18 . both sections [ sec - fus ] and [ sec - inv ] deal with an arbitrary solution @xmath1 of the yang
baxter equation ( only the unitary property is assumed in subsection [ subsec - dep ] ) . in section [ sec - sn ]
, we apply the whole procedure to the situation where @xmath1 is the yang solution , and identify invariant subspaces isomorphic to finite - dimensional irreducible representations of the general linear lie algebra .
the main results are summarized in corollary [ coro - fin - sn ] .
we also provide explicit examples of fused solutions and discuss open problems concerning non - standard " evaluations of the fusion function . in sections [ sec - sn - s ] and [ sec - hn ] ,
we generalize the results of the preceding section to the situation where @xmath1 is a generalization of the yang solution for a super vector space and where @xmath1 is the standard deformation of the yang solution or of its super " generalization .
[ [ acknowledgements . ] ] * acknowledgements .
* + + + + + + + + + + + + + + + + + + + the author would like to thank oleg ogievetsky for many fruitful discussions and for his valuable comments on this work .
[ [ section ] ] * 1 . *
+ + + + let @xmath29 and @xmath30 be two integers such that @xmath31 .
we will use the following notation for the product of non - commuting quantity @xmath32 depending on @xmath33 : @xmath34 similarly , for non - commuting quantities depending on two indices @xmath33 and @xmath35 , the arrow
@xmath36 " indicates that the factors are ordered lexicographically , while the arrow
@xmath37 " indicates that the factors are ordered in the reverse lexicographic order .
for example , we have : @xmath38 [ [ section-1 ] ] * 2 . * + + + + the product of two ( and any number of ) permutations is written with the standard notation for the composition of functions , namely , if @xmath39 are two permutations then @xmath40 means that @xmath41 is applied first and then @xmath42 ; for example , @xmath43 in the standard cyclic notation .
[ [ section-2 ] ] * 3 . *
+ + + + for a vector space @xmath0 over @xmath44 , we denote by @xmath45 the set of endomorphisms of @xmath0 .
we fix a finite - dimensional vector space @xmath0 over @xmath44 and let @xmath46 denote the identity endomorphism of @xmath0 and @xmath47 denote the permutation endomorphism of @xmath2 ( that is , @xmath48 for @xmath49 ) .
we will denote by @xmath50 the identity operator on @xmath51 , for @xmath52 .
we will use the standard notation for operators in @xmath53 ; namely , if @xmath54 then @xmath55 , for @xmath56 , will denote the operator in @xmath53 acting as @xmath57 on the @xmath33-th copy and trivially anywhere else .
for example , for @xmath58 , we have @xmath59 similarly , if @xmath60 is an operator in @xmath61 then @xmath62 will denote the operator in @xmath53 acting as @xmath60 in the @xmath33-th and @xmath35-th copies and trivially anywhere else .
for example , if @xmath58 : @xmath63 we also have for example @xmath64 .
note that we slightly abuse by not indicating the integer @xmath65 in the notation @xmath55 , @xmath62 , @xmath66 , etc .
this should not lead to any confusion as it will be clear on which space the operators act .
we recall some obvious commutation relations that we will often use throughout the text without mentioning : @xmath67 t_ir_{j , k}&=&r_{j , k}t_i & \text{if $ i\notin\{j , k\}$\,,}\\[0.2em ] r_{i , j}r_{k , l}&=&r_{k , l}r_{i , j } & \text{if $ \{i , j\}\cap\{k , l\}=\emptyset$\,,}\\[0.2em ] { \textrm{p}}_{i , j}t_k&=&t_{s_{i , j}(k)}{\textrm{p}}_{i , j } & \text{for all $ i , j , k=1,\dots , n$ with $ i\neq j$,}\\[0.2em ] { \textrm{p}}_{i , j}r_{k , l}&=&r_{s_{i , j}(k),s_{i , j}(l)}{\textrm{p}}_{i , j}\ \ \ & \text{for all $ i , j , k , l=1,\dots , n$ with $ i\neq j$ and $ k\neq l$ , } \end{array}\ ] ] where @xmath68 denotes the transposition of @xmath33 and @xmath35 .
[ [ yangbaxter - equation . ] ] * 4 .
baxter equation .
* + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath60 be a function of one variable @xmath69 , taking values in @xmath61 .
the variable @xmath70 is called the _
spectral parameter_. the function @xmath60 is a solution , on @xmath2 , of the yang
baxter equation if the following functional relation is satisfied : @xmath71 where both sides take values in @xmath72 and @xmath73 means @xmath74 , etc .
( we will always use this standard notation ) . by abuse of speaking
, we will sometimes say that the operator @xmath1 itself is a solution of the yang
baxter equation .
more generally , let @xmath75 be an indexing set , @xmath76 be a collection of finite - dimensional vector spaces over @xmath44 , and let @xmath77 be a family of functions of @xmath69 such that @xmath78 takes values in @xmath79 , for any @xmath80 .
we say that the functions @xmath78 form a family of solutions of the yang
baxter equation if the following functional relations are satisfied : @xmath81 where both sides take values in @xmath82 ( in ( [ yb - add - fam ] ) , @xmath83 stands for @xmath84 , and similarly for @xmath85 and @xmath86 ) . equation ( [ yb - add1 ] ) is the yang
baxter equation with additive " spectral parameters .
we will sometimes use the version of the equation with multiplicative " spectral parameters , namely @xmath87 for a function @xmath60 of the variable @xmath88 taking values in @xmath61 .
[ [ partitions - and - young - tableaux . ] ] * 5 . partitions and young tableaux . *
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath89 be a partition of a positive integer @xmath65 , that is , @xmath90 is a family of integers such that @xmath91 and @xmath92 .
we say that @xmath17 is a partition _ of size _ @xmath65 and _ of length _ @xmath93 and set @xmath94 and @xmath95 .
the young diagram of @xmath17 is the set of elements @xmath96 such that @xmath97 and @xmath98 .
a pair @xmath96 is usually called a _ node_. the young diagram of @xmath17 will be represented in the plan by a left - justified array of @xmath93 rows such that the @xmath35-th row contains @xmath99 nodes for all @xmath100 ( a node will be pictured by an empty box ) .
we number the rows from top to bottom .
we identify partitions with their young diagrams and say that @xmath101 is a node of @xmath17 , or @xmath102 , if @xmath101 is a node of the diagram of @xmath17 . for a node @xmath103 , we set @xmath104 . the number @xmath105 is called the _ classical content _ of @xmath106 . for any complex number @xmath107 , we set @xmath108 and call @xmath109 the _ @xmath107-quantum content _ of @xmath106 , or simply the _ quantum content _ of @xmath106 when @xmath107 is fixed .
the hook of a node @xmath110 is the set of nodes of @xmath17 consisting of the node @xmath106 and the nodes which lie either under @xmath106 in the same column or to the right of @xmath106 in the same row ; the hook length @xmath111 of @xmath106 is the cardinality of the hook of @xmath106 .
we define , for any partition @xmath17 , @xmath112 for any non - zero complex number @xmath107 , we also define for a partition @xmath17 : @xmath113}\ , \ ] ] where @xmath114:=\frac{q^n - q^{-n}}{q - q^{-1}}$ ] for @xmath115 . a young tableau @xmath116 of shape @xmath17 is a bijection between the set @xmath117 and the set of nodes of @xmath17 . in other words , a young tableau of shape @xmath17 is obtained by placing without repetition the numbers @xmath118 into the nodes of @xmath17 .
we use the notation @xmath119 to denote the shape of @xmath120 . a young tableau @xmath120 of shape @xmath17 is _ standard _
if its entries increase along any row and down any column of @xmath17 , that is , the number corresponding to the node @xmath101 is greater than the number corresponding to the node @xmath121 when @xmath122 and @xmath123 . for a young tableau @xmath120 ,
we denote respectively by @xmath124 and @xmath125 the classical content and the @xmath107-quantum content of the node containing the number @xmath33 , for all @xmath126 .
we fix @xmath0 a finite - dimensional vector space over @xmath44 and let @xmath60 be a function of @xmath69 , taking values in @xmath61 , which satisfies the yang baxter equation : @xmath127 where both sides operate on @xmath128 ( in sections [ sec - fus ] and [ sec - inv ] , we work with the additive convention for the spectral parameters , see paragraph * 4 * of section [ sec - prel ] ; equivalently , we could have chosen the multiplicative version , as indicated in remark [ rem - mult ] below ) .
[ ex - yang ] one of the simplest example of a solution of equation ( [ yb - add ] ) is the yang solution : @xmath129 the fact that the function given by ( [ yang ] ) satisfies the yang baxter equation ( [ yb - add ] ) can be checked by a direct calculation.@xmath130 let @xmath65 and @xmath131 be positive integers .
we consider the space @xmath6 and label the copies of @xmath0 by @xmath132 ( from left to right ) such that , for example , the operator @xmath133 with @xmath134 and @xmath135 stands for the operator @xmath136 with the notation explained in section [ sec - prel ] ( we prefer the underlined notation to make a clear distinction between the indices corresponding to @xmath13 and the indices corresponding to @xmath137 ) .
let @xmath138 be an @xmath65-tuple of complex parameters and @xmath139 be an @xmath131-tuple of complex parameters .
we define a function @xmath140 of one variable @xmath69 taking values in @xmath141 by : @xmath142 moving all the operators with @xmath65 as a first index to the left and repeating the process for @xmath143 , we find the following alternative form for @xmath5 : @xmath144 we prove in the following theorem that the set of functions @xmath145 , where @xmath146 , @xmath147 and @xmath148 , forms a family of solutions of the yang baxter equation
. we will call elements of this family fused solutions " of the yang baxter equation , and operators @xmath5 fused operators " .
[ thm - fus - r ] let @xmath65 , @xmath131 and @xmath149 be positive integers and let @xmath138 , @xmath139 and @xmath150 be , respectively , an @xmath65-tuple , an @xmath131-tuple and an @xmath149-tuple of complex parameters .
we have the functional equation @xmath151 where both sides takes values in @xmath152 and the copies of @xmath0 are labelled by @xmath118 , @xmath153 , @xmath154 ( from left to right ) . in equation ( [ fus - yb - add ] )
, @xmath5 stands for the operator @xmath155 and similarly for @xmath156 and @xmath157 .
note first that the equality @xmath158 , valid for any @xmath159 , @xmath160 and @xmath161 , ensures that , in the product of operators in the left hand side of ( [ fus - yb - add ] ) , we can always apply the yang
baxter equation ( [ yb - add ] ) as soon as the indices are properly arranged .
in other words , we will never have to worry about the spectral parameters .
therefore , for saving place during the proof , we will drop them out of the notation , namely we set , until the end of the proof , @xmath162 for any @xmath159 , @xmath160 and @xmath161 .
we prove the formula ( [ fus - yb - add ] ) by induction on @xmath65 .
if @xmath163 then the left hand side of ( [ fus - yb - add ] ) is @xmath164 we have , for @xmath165 , @xmath166 = \!\!\ ! & r_{1,{\underline{1}}}r_{1,{\underline{2}}}\dots r_{1,{\underline{n'-1}}}\cdot r_{{\underline{n'}},{\underline{\underline{i}}}}r_{1,{\underline{\underline{i}}}}r_{1,{\underline{n'}}}\cdot r_{{\underline{n'-1}},{\underline{\underline{i}}}}\dots r_{{\underline{2}},{\underline{\underline{i}}}}r_{{\underline{1}},{\underline{\underline{i}}}}\,\quad\text{(as $ r_{1,{\underline{n'}}}r_{1,{\underline{\underline{i}}}}r_{{\underline{n'}},{\underline{\underline{i}}}}=r_{{\underline{n'}},{\underline{\underline{i}}}}r_{1,{\underline{\underline{i}}}}r_{1,{\underline{n'}}}\,$)}\\[0.4em ] = \!\!\ ! & r_{{\underline{n'}},{\underline{\underline{i}}}}\cdot r_{1,{\underline{1}}}r_{1,{\underline{2}}}\dots r_{1,{\underline{n'-1}}}\cdot r_{1,{\underline{\underline{i}}}}\cdot r_{{\underline{n'-1}},{\underline{\underline{i}}}}\dots r_{{\underline{2}},{\underline{\underline{i}}}}r_{{\underline{1}},{\underline{\underline{i}}}}\cdot r_{1,{\underline{n'}}}\ . \end{array}\ ] ] repeating this calculation with @xmath131 replaced by @xmath167 and so on , we find that
@xmath168 so we finally obtain that the left hand side of ( [ fus - yb - add ] ) for @xmath163 is equal to @xmath169 which coincides with the right hand side of ( [ fus - yb - add ] ) for @xmath163 .
now let @xmath170 and set @xmath171 . using ( [ def - fus - r2 ] ) and
commutation relations , we reorganize the left hand side of ( [ fus - yb - add ] ) and write it as @xmath172 we use the induction hypothesis to transform this expression into @xmath173 then we use the induction basis ( with the space labelled here by @xmath65 playing the same role as the space labelled by @xmath174 in the calculation for the induction basis ) to move @xmath157 to the left and we obtain @xmath175 this is equal , using again commutation relations , to @xmath176 , that is , to the right hand side of ( [ fus - yb - add ] ) .
[ rem - graph ] to the operator @xmath1 is associated the following figure : ( 300,300)(0,-300 ) ( 0,0 ) ( 1 , -1)300 ( 300,0 ) ( -1 , -1)300 ( 155,-125 ) roughly speaking , this represents the interaction of two particules at the intersection of the two lines , and the interaction is governed , in some sense , by the operator @xmath1 ( see , _
e.g. _ , @xcite for the physical meaning of the graphical formulation of the yang
baxter equation ) .
then , the yang baxter equation ( [ yb - add ] ) is formulated graphically as ( 1200,550)(0,-500 ) ( 0,20 ) ( 150,20 ) ( 500,20 ) ( 0,0 ) ( 1 , -1)500 ( 150,0 ) ( 0 , -1)500 ( 500,0 ) ( -1 , -1)500 ( 600,-250 ) ( 137,-120 ) ( 225,-210 ) ( 175,-320 ) ( 700,20 ) ( 1050,20 ) ( 1200,20 ) ( 700,0 ) ( 1 , -1)500 ( 1050,0 ) ( 0 , -1)500 ( 1200,0 ) ( -1 , -1)500 ( 1075,-120 ) ( 925,-210 ) ( 1040,-320 ) the numbers above the lines are the indices of the copy of @xmath0 in @xmath177 . by convention , we read the figure from top to bottom and we write the corresponding operators from left to right ; for example , the left hand side of the picture above corresponds to @xmath178 . within this graphical interpretation ,
the fused solutions , given by formula ( [ def - fus - r ] ) , correspond to interactions of multiplets of particules , namely , @xmath65 particules interacting with @xmath131 particules . as an example , for @xmath179 , the corresponding interaction is depicted by : ( 650,530)(0,-500 ) ( 0,20 ) ( 150,20 ) ( 500,20 ) ( 650,20 ) ( 0,0 ) ( 1 , -1)500 ( 150,0 ) ( 1 , -1)500 ( 500,0 ) ( -1 , -1)500 ( 650,0 ) ( -1 , -1)500 ( 230,-215 ) ( 380,-215 ) ( 308,-140 ) ( 308,-290 ) where @xmath180 for @xmath181 , and @xmath182 are the parameters of the fused solutions .
we note that there is no ambiguity in the ordering of the factors , because @xmath183 and @xmath184 commute .
the expression obtained from the above picture is equal to the right hand side of the defining formula ( [ def - fus - r ] ) for @xmath179 . @xmath130
[ ex - fus ] let @xmath185 and @xmath186 .
we consider the space @xmath187 and the four copies of @xmath0 are labelled ( from left to right ) by @xmath188 . by definition
, we have @xmath189 @xmath190 we illustrate on this example the calculation made in the proof of theorem [ thm - fus - r ] ( we use the same abbreviated notation as in the proof ) : @xmath191 = & r_{2,{\underline{1}}}\cdot r_{2,{\underline{\underline{1}}}}\,r_{1,{\underline{1}}}\cdot r_{1,{\underline{\underline{1}}}}\,r_{{\underline{1}},{\underline{\underline{1 } } } } & \text{(as $ r_{1,{\underline{1}}}\,r_{2,{\underline{\underline{1}}}}=r_{2,{\underline{\underline{1}}}}\,r_{1,{\underline{1}}}$)}\\[0.4em ] = & r_{2,{\underline{1}}}\,r_{2,{\underline{\underline{1}}}}\cdot r_{{\underline{1}},{\underline{\underline{1}}}}\,r_{1,{\underline{\underline{1}}}}\,r_{1,{\underline{1}}}\ \
\quad & \text{(as $ r_{1,{\underline{1}}}\,r_{1,{\underline{\underline{1}}}}\,r_{{\underline{1}},{\underline{\underline{1}}}}=r_{{\underline{1}},{\underline{\underline{1}}}}\,r_{1,{\underline{\underline{1}}}}\,r_{1,{\underline{1}}}$)}\\[0.4em ] = & r_{{\underline{1}},{\underline{\underline{1}}}}\,r_{2,{\underline{\underline{1}}}}\,r_{2,{\underline{1}}}\cdot r_{1,{\underline{\underline{1}}}}\,r_{1,{\underline{1 } } } & \text{(as $ r_{2,{\underline{1}}}\,r_{2,{\underline{\underline{1}}}}\,r_{{\underline{1}},{\underline{\underline{1}}}}=r_{{\underline{1}},{\underline{\underline{1}}}}\,r_{2,{\underline{\underline{1}}}}\,r_{2,{\underline{1}}}$)}\\[0.4em ] = & r_{{\underline{1}},{\underline{\underline{1}}}}\,r_{2,{\underline{\underline{1}}}}\cdot r_{1,{\underline{\underline{1}}}}\,r_{2,{\underline{1}}}\cdot r_{1,{\underline{1 } } } & \text{(as $ r_{2,{\underline{1}}}\,r_{1,{\underline{\underline{1}}}}=r_{1,{\underline{\underline{1}}}}\,r_{2,{\underline{1}}}$)}\\[0.6em ] = & r_{{\underline{{\textbf{c}}}},{\underline{\underline{{\textbf{c}}}}}}(v)r_{{\textbf{c}},{\underline{\underline{{\textbf{c}}}}}}(u+v)r_{{\textbf{c}},{\underline{{\textbf{c}}}}}(u)\ . & \end{array}\ ] ] this calculation ( and the `` higher '' ones as well ) is maybe more clear within the graphical interpretation explained in remark [ rem - graph ] .
the calculation above corresponds to the following graphical calculation : ( 2400,550)(0,-500 ) ( 0,40 ) ( 100,40 ) ( 300,40 ) ( 700,40 ) ( 0,0 ) ( 1 , -1)500 ( 100,0 ) ( 1 , -1)500 ( 300,0 ) ( 0 , -1)500 ( 700,0 ) ( -1 , -1)500 ( 750,-250 ) ( 850,40 ) ( 950,40 ) ( 1225,40 ) ( 1550,40 ) ( 850,0 ) ( 1 , -1)500 ( 950,0 ) ( 1 , -1)500 ( 1225,0 ) ( 0 , -1)500 ( 1550,0 ) ( -1 , -1)500 ( 1600,-250 ) ( 1700,40 ) ( 1800,40 ) ( 2150,40 ) ( 2400,40 ) ( 1700,0 ) ( 1 , -1)500 ( 1800,0 ) ( 1 , -1)500 ( 2150,0 ) ( 0 , -1)500 ( 2400,0 ) ( -1 , -1)500 where we omit the spectral parameters at the intersections as they are fixed by the indices of the interacting lines . moving the line indexed by @xmath192 in the picture ( using the pictorial version of the yang
baxter equation of remark [ rem - graph ] ) corresponds to moving the operator @xmath193 from right to left in the calculation written above . here also , we note that there is no ambiguity in the ordering of the factors , due to commutation relations .
@xmath130 [ ex1 ] in this example , we let @xmath1 be the yang solution ( [ yang ] ) .
we set @xmath185 , @xmath194 , @xmath195 and @xmath196 .
we have : @xmath197 we consider the case @xmath198 and let @xmath199 be a basis of @xmath0 .
we define @xmath200 , @xmath201 , @xmath202 and @xmath203}:=e_1\otimes e_2-e_2\otimes e_1 $ ] .
we calculate the matrix of the endomorphism ( [ ex - fus1 ] ) in the following basis of @xmath204 : @xmath205}\otimes e_1,\,e_{[12]}\otimes e_2\}\ .\ ] ] we obtain ( points indicate coefficients equal to @xmath206 ) : @xmath207 we remark that , for the particular choice of @xmath208 made here , the fused operator @xmath5 leaves invariant the subspace @xmath209 of @xmath204 , where @xmath210 is the symmetric square of @xmath0 .
this phenomenon can be explained by the relation : @xmath211 together with the fact that the space @xmath210 is the image of the endomorphism @xmath212 of @xmath213 ( since @xmath214 ) .
@xmath130 [ ex2 ] we still take for @xmath1 the yang solution ( [ yang ] ) and we set again @xmath185 , @xmath194 and @xmath195 .
but now we set @xmath215 .
we then have @xmath216 we consider again the case @xmath198 and calculate the matrix of the endomorphism ( [ ex - fus2 ] ) in the same basis of @xmath204 as in example [ ex1 ] .
we obtain ( points indicate coefficients equal to @xmath206 ) : @xmath217 we remark now that , for the particular choice of @xmath208 made here , the fused operator @xmath5 leaves invariant the subspace @xmath218 of @xmath204 , where @xmath219 is the alternating square of @xmath0 . similarly to the preceding example , this phenomenon can be explained by the relation : @xmath220 together with the fact that the space @xmath219 is the image of the endomorphism @xmath221 of @xmath213 ( since @xmath222).@xmath130
in both examples [ ex1 ] and [ ex2 ] , we have noted that , for some choices of the parameters @xmath18 and @xmath223 , the fused operator @xmath5 leaves invariant a subspace of @xmath6 ; moreover this subspace can be seen as the image of an endomorphism of @xmath6 written in terms of the original solution @xmath1 of the yang baxter equation .
the goal of this section is to show that examples [ ex1 ] and [ ex2 ] are actually ( simplest ) examples of a general phenomenon .
let @xmath65 be a positive integer such that @xmath224 and @xmath138 be an @xmath65-tuple of complex parameters . recall that @xmath1 is an arbitrary solution of the yang
baxter equation ( [ yb - add ] ) on @xmath2 .
we define the following endomorphism of @xmath13 : @xmath225 it will be useful to consider that @xmath226 are complex variables and to see @xmath227 as a function of @xmath18 with values in @xmath53 .
depending on the function @xmath60 , the function @xmath227 can be singular for some values of @xmath18 ( see , for example , ( [ yang ] ) ) .
we will need a preliminary lemma concerning the operator @xmath12 .
[ lem - f ] we have @xmath228 as in the proof of the theorem [ thm - fus - r ] , we do not need to pay attention to the spectral parameters since @xmath229 .
thus , until the end of the proof , we use the abbreviated notation @xmath230 for any @xmath231 .
we prove the formula ( [ f2 ] ) by induction on @xmath65 . for @xmath185
there is nothing to prove , so we let @xmath232 and write @xmath233 where we have moved to the right the operators @xmath234 using commutation relations .
now we use the induction hypothesis and we have @xmath235 & = \displaystyle\bigl(\prod_{2\leq i < j\leq n-1}^{\leftarrow}r_{i , j}\bigr)\cdot r_{1,n-1}\dots r_{1,2}\cdot r_{1,n}r_{2,n}\dots r_{n-1,n}\ . \end{array}\ ] ] note that @xmath236 & = r_{2,n}\cdot r_{1,n-1}\dots r_{1,3}r_{1,n}r_{3,n}\dots r_{n-1,n}\cdot r_{1,2}\ , , \end{array}\ ] ] and thus , repeating a similar calculation the necessary number of times , we arrive at @xmath237 we use again the induction hypothesis and commutation relations to conclude as follows : @xmath238 & = \displaystyle\bigl(\prod_{2\leq i < j\leq n}^{\rightarrow}r_{i , j}\bigr)r_{1,n}r_{1,n-1}\dots r_{1,2}\\[1.3em ] & = \displaystyle\bigl(\prod_{2\leq i < j\leq n}^{\leftarrow}r_{i , j}\bigr)r_{1,n}r_{1,n-1}\dots r_{1,2}= \displaystyle\prod_{1\leq i < j\leq n}^{\leftarrow}r_{i , j}\ .
\end{array}\ ] ] for an @xmath65-tuple of complex numbers @xmath18 such that the function @xmath227 is non - singular at @xmath18 , let @xmath7 denote the image of the endomorphism @xmath12 : @xmath239 let @xmath240 be positive integers and @xmath138 , respectively @xmath139 , be an @xmath65-tuple , respectively an @xmath131-tuple , of complex numbers such that @xmath11 and @xmath241 are defined . [ thm - inv - sub ] the fused operator @xmath5 preserves the subspace @xmath9 of @xmath6 .
let @xmath242 be a complex variable .
we first prove the following lemma : [ lem - inv - sub ] 1 .
we consider the space @xmath243 with the copies of @xmath0 labelled by @xmath244 ( from left to right ) .
the endomorphism @xmath245 preserves the subspace @xmath246 .
we consider the space @xmath247 with the copies of @xmath0 labelled by @xmath248 ( from left to right ) .
the endomorphism @xmath249 preserves the subspace @xmath250 . 1 .
set @xmath251 .
the image of the endomorphism @xmath252 of @xmath243 is @xmath253 .
we have , using lemma [ lem - f ] , @xmath254 = & \displaystyle\prod_{0\leq i < j\leq n'}^{\rightarrow}r_{{\underline{i}},{\underline{j}}}(c_{{\underline{i}}}-c_{{\underline{j}}})= \displaystyle\prod_{0\leq i
< j\leq n'}^{\leftarrow}r_{{\underline{i}},{\underline{j}}}(c_{{\underline{i}}}-c_{{\underline{j}}})\\[1.2em ] = & \displaystyle\bigl(\prod_{1\leq i < j\leq n'}^{\leftarrow}r_{{\underline{i}},{\underline{j}}}(c_{{\underline{i}}}-c_{{\underline{j}}})\bigr)\cdot r_{{\underline{0}},{\underline{n'}}}(v - c_{{\underline{n'}}})\dots r_{{\underline{0}},{\underline{2}}}(v - c_{{\underline{2}}})r_{{\underline{0}},{\underline{1}}}(v - c_{{\underline{1}}})\\[1.4em ] = & \bigl(\textrm{id}\otimes f({\underline{{\textbf{c}}}})\bigr)\cdot r_{{\underline{0}},{\underline{n'}}}(v - c_{{\underline{n'}}})\dots r_{{\underline{0}},{\underline{2}}}(v - c_{{\underline{2}}})r_{{\underline{0}},{\underline{1}}}(v - c_{{\underline{1}}})\ . \end{array}\ ] ]
thus the operator @xmath245 restricted on @xmath253 has its image contained in @xmath253 .
now set @xmath255 .
the image of the endomorphism @xmath256 of @xmath247 is @xmath257 .
we have @xmath258 = & r_{n , n+1}(c_n - v)\dots r_{2,n+1}(c_2-v)r_{1,n+1}(c_1-v)\cdot\displaystyle\bigl(\prod_{1\leq i <
j\leq n}^{\leftarrow}r_{i , j}(c_{i}-c_{j})\bigr)\\[-0.5em ] = & \displaystyle\bigl(\prod_{1\leq i < j\leq n+1}^{\leftarrow}r_{i , j}(c_{i}-c_{j})\bigr)\\[1.2em ] = & \displaystyle\bigl(\prod_{1\leq i < j\leq n+1}^{\rightarrow}r_{i , j}(c_{i}-c_{j})\bigr)\\[1.2em ] = & \displaystyle\bigl(\prod_{1\leq i < j\leq n}^{\rightarrow}r_{i , j}(c_{i}-c_{j})\bigr)\cdot r_{1,n+1}(c_1-v)r_{2,n+1}(c_2-v)\dots r_{n , n+1}(c_n - v ) \\[1.4em ] = & \bigl ( f({\textbf{c}})\otimes\textrm{id}\bigr)\cdot r_{1,n+1}(c_1-v)r_{2,n+1}(c_2-v)\dots r_{n , n+1}(c_n - v)\ .
\end{array}\ ] ] we used lemma [ lem - f ] in the first and third equalities , and commutation relations in the second and fourth equalities .
we conclude that the image of the restriction on @xmath257 of the operator @xmath249 is contained in @xmath257 .
we return to the proof of theorem [ thm - inv - sub ] .
recall that the operator @xmath5 is given by formula ( [ def - fus - r ] ) , namely @xmath259 for any @xmath260 , applying lemma [ lem - inv - sub](ii ) with @xmath261 and the label @xmath262 replaced by @xmath263 , we obtain that the operator @xmath264 preserves the subspace @xmath265 , and so in turn that the operator @xmath5 preserves @xmath265 .
now recall the alternative formula ( [ def - fus - r2 ] ) for the operator @xmath5 : @xmath266 for any @xmath159 , applying lemma [ lem - inv - sub](i ) with @xmath267 and the label @xmath268 replaced by @xmath33 , we obtain that the operator @xmath269 preserves the subspace @xmath270 , and so in turn that the operator @xmath5 preserves @xmath270 .
we conclude that the operator @xmath5 preserves the subspace @xmath9 .
[ rem - graph - inv ] for @xmath179 , part of theorem [ thm - inv - sub ] is illustrated as follows , using the pictorial version of the yang
baxter equation explained in remark [ rem - graph ] , ( 1500,530)(0,-500 ) ( 0,20 ) ( 150,20 ) ( 500,20 ) ( 650,20 ) ( 0,0 ) ( 1 , -1)425 ( 425,-425 ) ( 1 , 0)150 ( 150,0 ) ( 1 , -1)350 ( 500,-350 ) ( 0 , -1)150 ( 500,0 ) ( -1 , -1)500 ( 650,0 ) ( -1 , -1)500 ( 750,-250 ) ( 900,-75 ) ( 1000,20 ) ( 1350,20 ) ( 1500,20 ) ( 1000,-150 ) ( 1 , -1)350 ( 1000,0 ) ( 0 , -1)150 ( 1075,-75 ) ( 1 , -1)425 ( 1075,-75 ) ( -1 , 0)150 ( 1350,0 ) ( -1 , -1)500 ( 1500,0 ) ( -1 , -1)500 where we omit the spectral parameters .
this graphical equality corresponds to the fact that the operator @xmath5 preserves the subspace @xmath265 ( a similar picture can be drawn to illustrate the other half of theorem [ thm - inv - sub ] ) .
@xmath130 the content of theorem [ thm - inv - sub ] is empty if the endomorphisms @xmath271 and @xmath272 are both invertible ( as in this situation @xmath273 ) .
this is the goal of sections [ sec - sn][sec - hn ] to explain that , for standard examples of @xmath1 , there are some values of @xmath18 such that the images of the operators @xmath12 are interesting proper subspaces of @xmath13.@xmath130 we will give an alternative formula for @xmath12 which will be useful later . to do this ,
we define a function @xmath274 with values in @xmath61 by : @xmath275 where we recall that @xmath47 is the permutation operator on @xmath2 . for any @xmath42 in the symmetric group @xmath19 on @xmath65 letters ,
we define @xmath276 by : @xmath277 note that if @xmath42 is written as a a product of transposition , say @xmath278 , then @xmath279 .
let @xmath280 be the longest element of the symmetric group @xmath19 .
we recall the following property of the element @xmath280 : @xmath281 where @xmath282 is the longest element of @xmath283 , seen as an element of @xmath19 acting only on the letters @xmath284 .
note that @xmath280 is an involution .
[ lem - f - alt ] we have @xmath285 we prove formula ( [ f4 ] ) by induction on @xmath65 .
for @xmath185 there is nothing to prove .
let @xmath232 and write @xmath286 the induction hypothesis allows to replace @xmath287 by @xmath288 so it remains to prove that @xmath289 \hspace{4.5cm}={\widehat{r}}_{n-1,n}(c_1-c_{n})\dots{\widehat{r}}_{2,3}(c_{n-2}-c_{n}){\widehat{r}}_{1,2}(c_{n-1}-c_{n}){\textrm{p}}_{w_n}\ , .
\end{array}\ ] ] using ( [ wn ] ) , the left hand side of ( [ eq - lem1 ] ) is equal to @xmath290 = & { \widehat{r}}_{n-1,n}(c_1-c_n){\textrm{p}}_{n-1,n}\dots\dots { \widehat{r}}_{2,n}(c_{n-2}-c_n){\textrm{p}}_{2,n}\,{\widehat{r}}_{1,n}(c_{n-1}-c_n){\textrm{p}}_{1,n}\cdot{\textrm{p}}_{w_{n-1 } } \end{array}\ ] ] moving all the permutation operators to the right , we obtain that the left hand side of ( [ eq - lem1 ] ) is equal to @xmath291 it is easy to see that @xmath292 and thus , with ( [ wn ] ) , this concludes the proof of ( [ eq - lem1 ] ) and in turn of the lemma .
we define the following function of @xmath18 with values in @xmath53 : @xmath293 assume that the space @xmath11 is defined ( that is , @xmath227 is non - singular at @xmath18 ) .
lemma [ lem - f - alt ] implies that @xmath294 and thus has the following corollary .
[ cor - hf ] the subspace @xmath7 coincides with the image of the endomorphism @xmath295 .
let @xmath65 , @xmath131 be positive integers such that @xmath240 , let @xmath138 be an @xmath65-tuple of complex numbers and @xmath139 be an @xmath131-tuple of complex numbers . for any permutation @xmath42 in the symmetric group @xmath19
, we define @xmath297 .
we denote by @xmath298 the transposition @xmath299 , for @xmath300 ; then we have , for @xmath300 , @xmath301 .
we recall that the copies of @xmath0 in @xmath6 are labelled by @xmath132 ( from left to right ) .
[ lem - dep - r ] 1 . for @xmath300 , we have , on @xmath6 , @xmath302 2 . for @xmath300
, we have , on @xmath13 , @xmath303 \(i ) we have , for any @xmath260 , @xmath304 we use in the second equality commutation relations , together with the yang
baxter relation on the copies labelled by @xmath305 , @xmath306 and @xmath263 . due to formula ( [ def - fus - r ] ) , this proves the item ( i ) .
\(ii ) formula ( [ dep - f ] ) is equivalent to @xmath307 we prove formula ( [ dep - f2 ] ) by induction on @xmath65 . if @xmath185 , formula ( [ dep - f2 ] ) is trivial .
let @xmath232 .
we deal first with the case @xmath308 .
we have , using commutation relations , @xmath309 the induction hypothesis gives @xmath310 and besides we have @xmath311 = & r_{2,n}(c_2-c_n)r_{1,n}(c_1-c_n)r_{3,n}(c_3-c_n)\dots r_{n-1,n}(c_{n-1}-c_n)\cdot r_{1,2}(c_1-c_2)\ , \end{array}\ ] ] where we used the yang
baxter relation on the copies labelled by @xmath174 , @xmath312 and @xmath65 , and commutation relations .
formula ( [ dep - f2 ] ) for @xmath308 follows .
then we assume that @xmath313 , and we write @xmath314 similarly to above , we first have , using commutation relations and the yang
baxter equation on the copies labeled by @xmath174 , @xmath305 and @xmath306 , that @xmath315 = & r_{1,2}(c_1-c_2)\dots r_{1,k+1}(c_1-c_{k+1})r_{1,k}(c_1-c_{k})\dots r_{1,n}(c_{1}-c_n)\cdot r_{k+1,k}(c_{k+1}-c_k)\ , \end{array}\ ] ] and then we use the induction hypothesis , namely ( since @xmath313 ) @xmath316 this yields the desired result .
for the remaining of this section , we assume that @xmath317 where @xmath318 . in the physical literature
, this condition is called the `` unitarity '' condition .
we note that @xmath319 .
[ def - adm ] for @xmath320 , we say that the transposition @xmath321 is _ admissible _ for @xmath1 and for the set of parameters
@xmath3 if @xmath322 . let @xmath320 such that the transposition @xmath298 is _ admissible _ for @xmath1 and for the set of parameters @xmath3 .
we define an endomorphism @xmath323 of @xmath13 by @xmath324 the admissibility of @xmath298 implies that the operator @xmath323 is invertible with , from ( [ inv - r ] ) , @xmath325 in this situation , the assertions of lemma [ lem - dep - r ] can be rewritten , with @xmath326 , @xmath327 we sum up the consequences of these two formulas in the following proposition .
we assume that @xmath11 ( and thus @xmath328 as well ) are defined .
[ prop - inv ] let @xmath320 such that the transposition @xmath298 is _ admissible _ for @xmath1 and for the set of parameters @xmath3 .
let @xmath329 such that @xmath330 is a basis of the subspace @xmath11 . 1 .
then @xmath331 is a basis of @xmath328 .
in particular , @xmath332 and @xmath328 have the same dimension .
moreover , for any basis @xmath333 of @xmath241 , the matrix of the endomorphism @xmath334 in the basis @xmath335 coincides with the matrix of the endomorphism @xmath336 in the basis @xmath337 ( we use the notation @xmath338 and similarly for @xmath337 ) . 1 .
as the operator @xmath323 is invertible , the linear independence of the vectors in @xmath339 is immediate .
moreover , due to the second relation in ( [ a - r - f ] ) , we have that the dimensions of @xmath332 and @xmath328 coincide and also that the vectors in @xmath339 belong to @xmath328 .
this proves the item ( i ) .
2 . the item ( ii ) is a direct consequence of the first formula in ( [ a - r - f ] ) .
-0.4 cm .2 cm [ rem - mult ] in sections [ sec - fus ] and [ sec - inv ] , we worked with a solution of the yang baxter equation with additive " spectral parameters ( see paragraph * 4 * in section [ sec - prel ] ) .
the whole construction in sections [ sec - fus ] and [ sec - inv ] has an equivalent multiplicative " version , where one replaces addition for the spectral parameters by multiplication . as examples , the fused solution ( [ def - fus - r ] ) is given by @xmath340 the invariant subspace for this fused solution is the image in @xmath13 of the operator @xmath341 and the unitarity condition ( [ inv - r ] ) becomes @xmath342 , with @xmath343 . all proofs and calculations made in sections [ sec - fus ] and [ sec - inv ] are completely similar for the multiplicative version . in the sequel we will use both versions . @xmath130
starting from an arbitrary solution @xmath1 of the yang baxter equation on @xmath2 , we obtained in the preceding sections , via the fusion procedure , a set @xmath345 of functions of @xmath69 , where @xmath146 , @xmath147 and @xmath148 , which forms a family of solutions of the yang baxter equation , and where @xmath346 .
moreover , for generic @xmath18 and @xmath223 , we identified a subspace @xmath347 which is preserved by @xmath5 .
therefore , by restriction , the operators @xmath5 induce a family of solutions of the yang
baxter equation acting on the spaces @xmath9 . in this section
, we will further study the subspaces @xmath11 , when we start from the yang solution , ( [ yang2 ] ) below , of the yang baxter equation on @xmath2 .
this will lead to a family of fused solutions acting on irreducible representations of the general linear lie algebra @xmath15 .
.2 cm we fix in this section : @xmath348 let @xmath170 and recall that @xmath227 is the following function of @xmath349 with values in @xmath53 : @xmath350 moreover , recall that the subspace we are interested in , denoted by @xmath11 , is the image of the operator @xmath12 , for points @xmath3 where the function @xmath227 is non - singular . for the solution ( [ yang2 ] ) ,
the expression for the operator @xmath12 reads : @xmath351 set @xmath352 and denote by @xmath15 the lie algebra of endomorphism of @xmath0 and by @xmath344 its universal enveloping algebra .
the lie algebra @xmath15 acts on the tensor product @xmath13 and we denote by @xmath353 this representation , given by : @xmath354 the representation @xmath353 extends to a representation , which we still denote by @xmath353 , of @xmath344 .
the symmetric group @xmath19 acts on @xmath13 by permuting the copies and we denote this representation of @xmath19 by @xmath355 . explicitly , the representation @xmath355 is given by : @xmath356 where @xmath357 is defined by ( [ p - pi ] ) .
the representation @xmath355 extends to a representation , which we still denote by @xmath355 , of the group algebra @xmath20 .
the irreducible complex representations of @xmath20 are parametrized by the partitions of @xmath65 . for any partition @xmath17 of @xmath65
, we denote by @xmath358 the corresponding irreducible @xmath20-module ( with the convention that @xmath359 is the trivial representation ) . for any partition @xmath17 such that @xmath360 ,
we denote by @xmath361 the irreducible highest - weight @xmath344-module of highest weight @xmath17 . the schur
weyl duality between the symmetric group @xmath19 and the algebra @xmath344 can be expressed by the following assertions ( see _ e.g. _ @xcite ) .
[ thm - sw - sn ] 1 .
the subalgebra @xmath362 of @xmath363 is the centraliser of @xmath364 .
the subalgebra @xmath364 of @xmath363 is the centraliser of @xmath362 .
3 . as an @xmath365-module defined by @xmath353 and @xmath355
, the space @xmath13 decomposes as : @xmath366 where @xmath17 runs over the set of partitions such that @xmath367 and @xmath360 . according to formulas ( [ f - sn ] ) and ( [ sw - sn ] ) , we are interested in the following rational function in @xmath368 with values in the group algebra @xmath20 : @xmath369 we refer to the rational function @xmath370 as the `` fusion function '' of the symmetric group .
it is easy to check that , for distinct @xmath371 , @xmath372 the functions @xmath373 with values in @xmath20 are called _
baxterized elements_. we say that they are universal " solutions of the yang baxter equation ( associated to the symmetric group ) as they provide solutions of the yang baxter equation in the representations of @xmath19 .
let @xmath17 be a partition of @xmath65 and let @xmath116 be a standard young tableau of shape @xmath17 . for brevity , set @xmath374 for @xmath56 .
the following result traces back to the work of jucys @xcite , see also @xcite ( we use a formulation as in @xcite ) .
[ thm - idem - sn ] the element obtained by the following consecutive evaluations @xmath375 is a primitive idempotent of @xmath20 which generates a minimal left ideal isomorphic , as an @xmath20-module , to the irreducible module @xmath358 .
recall that @xmath376 is the non - zero complex number defined in ( [ f - lam ] ) .
[ rem - sing ] for some standard young tableau @xmath116 , it may happen that @xmath377 for @xmath378 ( for example , if the standard young tableau @xmath116 has the shape @xmath379 ) .
in this situation , theorem [ thm - idem - sn ] asserts in particular that the rational function @xmath380 is non - singular for the evaluations as in ( [ idem - sn ] ) , even if the factor @xmath381 is singular at @xmath382 and @xmath383.@xmath130 let @xmath116 be a standard young tableau such that @xmath385 with @xmath367 and @xmath360 .
we let @xmath386 where @xmath3 is seen here as an @xmath65-tuple of variables . now , in view of theorem [ thm - idem - sn ] , we define @xmath387 due to formulas ( [ f - sn ] ) and ( [ sw - sn ] ) , we have that @xmath388 the assertions of theorem [ thm - idem - sn ] imply that @xmath227 is non - singular for the consecutive evaluations of the variables as in ( [ def - ft ] ) .
we denote by @xmath389 the image of the operator @xmath390 , which is well - defined and which coincides , due to ( [ f - e ] ) , with the image in @xmath13 of the operator @xmath391 .
moreover , the element @xmath392 acts on the irreducible @xmath20-module @xmath358 as a projector on a one - dimensional subspace , and annihilates any irreducible @xmath20-module @xmath393 with @xmath394 . the preceding discussion , which follows from theorem [ thm - idem - sn ] together with the schur weyl duality , namely formula ( [ sw - sn - for ] ) in theorem [ thm - sw - sn ] , leads to the identification of the space @xmath389 .
[ thm - conc - sn ] the subspace @xmath389 of @xmath395 is an irreducible @xmath344-module isomorphic to @xmath396 .
let @xmath116 , @xmath397 be two standard young tableaux such that @xmath385 and @xmath398 , with @xmath367 , @xmath399 and @xmath400 .
we set @xmath401 , @xmath402 and we define @xmath403 the restriction of the operator @xmath404 to the subspace @xmath405 of @xmath6 .
let @xmath320 and denote by @xmath406 the young tableau of shape @xmath17 obtained from @xmath116 by exchanging the nodes with numbers @xmath306 and @xmath305 .
we assume that @xmath406 is also standard .
this is equivalent to say that @xmath407 .
moreover , the yang solution ( [ yang2 ] ) satisfies the following unitarity condition : @xmath408 therefore , according to definition [ def - adm ] , the condition for the tableau @xmath406 to be standard is equivalent to the condition for @xmath321 to be an admissible transposition for theyang solution @xmath1 and the set of parameters @xmath409 .
theorem [ thm - conc - sn ] implies that @xmath405 and @xmath410 are both isomorphic to @xmath384 and , moreover , proposition [ prop - inv ] applied here asserts that the endomorphisms @xmath411 and @xmath412 coincide up to a change of basis . besides , it is well - known that , for two standard young tableaux @xmath116 and @xmath413 of the same shape @xmath17 , there is a sequence of admissible transpositions @xmath414 transforming @xmath116 into @xmath413 . finally , we sum up the results obtained in this section for the yang solution .
the item ( i ) below follows from theorems [ thm - fus - r ] , [ thm - inv - sub ] and [ thm - conc - sn ] .
the item ( ii ) is a consequence of the above discussion together with the proposition [ prop - inv ] .
[ coro - fin - sn ] 1 .
the set of functions @xmath415 , where @xmath116 , @xmath397 are standard young tableaux such that @xmath416 , forms a family of solutions of the yang
baxter equation , where @xmath411 is an endomorphism of a space isomorphic to @xmath417 .
2 . for four standard young tableaux
@xmath116 , @xmath413 , @xmath397 and @xmath418 as above such that @xmath419 and @xmath420 , the endomorphisms @xmath411 and @xmath421 coincide up to a change of basis .
.2 cm [ ex - fus - sn ] let @xmath422 .
the expression for the operator @xmath423 is @xmath424 the results of this section , applied to this example , give that the operator @xmath423 preserves the subspace @xmath425 , where @xmath426 is the space of the irreducible representation of @xmath344 corresponding to the partition @xmath427 .
in fact the subspace @xmath428 coincides with the symmetric square of @xmath0 , since it is the image of the operator @xmath214 .
let @xmath429 and fix a basis @xmath199 of @xmath0 .
we set @xmath430 , @xmath431 and @xmath432 .
we give the matrix of the endomorphism @xmath411 of @xmath433 written in the basis @xmath434 ordered lexicographically ( that is , @xmath435 , @xmath436 , @xmath437 , @xmath438 , ... ):
@xmath439 where the points indicate the coefficients equal to @xmath206 .
@xmath130 [ ex - inv - sn ] let @xmath440 \fbox{\scriptsize{2}}\end{array}$ ] , @xmath441 \fbox{\scriptsize{3}}\end{array}$ ] and @xmath442 .
the expression for the operator @xmath423 is : @xmath443 here @xmath444 , whereas the subspace @xmath389 is the image in @xmath177 of the operator @xmath445 let @xmath429 and fix a basis @xmath199 of @xmath0 .
one can directly verify that @xmath446 is a basis of the subspace @xmath389 where @xmath447 ( we indicate that @xmath448 and @xmath449 ) . we give the matrix of the endomorphism @xmath411 in the basis @xmath450 of @xmath405
( points replace coefficients equal to @xmath206 ) : @xmath451 on the other hand , the expression for the operator @xmath452 is : @xmath453 here , the subspace @xmath454 is the image in @xmath177 of the operator @xmath455 as @xmath456 and @xmath457 , proposition [ prop - inv ] ( or item ( ii ) in corollary [ coro - fin - sn ] ) asserts that the matrix of the endomorphism @xmath458 in a certain basis of @xmath459 coincides with ( [ ex - sn-21a ] )
. moreover , proposition [ prop - inv ] provides explicitly this basis .
indeed , define , according to formula ( [ def - ak ] ) , the following operator on @xmath177 : @xmath460 then the basis of @xmath459 in which the matrix of the endomorphism @xmath458 coincides with ( [ ex - sn-21a ] ) is @xmath461 . in this simple example ,
the assertions of proposition [ prop - inv ] can be directly checked since , using that @xmath462 we easily verify that @xmath463 and @xmath464 , and in turn that @xmath465 is a basis of @xmath454 .
the calculation of the matrix of the endomorphism @xmath458 in the basis @xmath461 is straightforward and leads , as expected , to the matrix ( [ ex - sn-21a ] ) .
@xmath130 in the framework of the fusion procedure for the yang solution , we are interested in any set of complex parameters @xmath3 such that the fusion function @xmath370 , defined by @xmath466 is non - singular at @xmath18 , and moreover such that the resulting element of @xmath20 is non - invertible .
indeed , such a set of parameters @xmath18 will lead , via ( [ sw - sn ] ) , to a non - invertible endomorphism @xmath12 , and in turn to a proper subspace @xmath7 such that @xmath265 is invariant for the fused operators of the form @xmath5 . with the help of theorem [ thm - idem - sn ]
, we obtained in the preceding subsection such a set of parameters for any standard young tableau ( actually , in these situations , the evaluations of the fusion function are proportional to primitive idempotents of @xmath20 ) .
one may easily see that there are other sets of parameters @xmath3 such that the fusion function is non - singular and gives a non - invertible element of @xmath20 .
the classification of such sets of parameters @xmath18 and the study of the associated subspaces @xmath11 and the associated fused solutions is an open question .
we note also that the less general problem of finding all sets of parameters @xmath3 such that @xmath370 is non - singular at @xmath18 and gives an element proportional to an idempotent is not solved
. we will give in this subsection some examples for small @xmath65 .
[ [ case - n2 . ] ] * case n=2 . * + + + + + + + + + + + in this situation , the fusion function , with values in @xmath467 , is given by @xmath468 it is immediate to see that the only pairs of complex numbers @xmath469 such that the evaluation @xmath470 is non - singular and is a non - invertible element of @xmath467 are the pairs such that @xmath471 ( this can be seen from ( [ inv - r - sn ] ) ) . as the rational function @xmath472 only depends on @xmath473 , there is no loss of generality to consider that @xmath474 . as a conclusion , for @xmath185 ,
the only pairs @xmath469 leading to a proper subspace @xmath475 are the pairs associated to the standard young tableaux @xmath476 and @xmath477 \fbox{\scriptsize{2}}\end{array}$ ] .
[ [ case - n3 . ] ] * case n=3 .
* + + + + + + + + + + + the fusion function with values in @xmath478 is given by @xmath479 as above , we can fix the freedom of translating all variables @xmath480 by the same constant with the assumption that @xmath481 is evaluated to @xmath206 .
[ prop - rem ] the evaluation of the function @xmath482 at @xmath483 and @xmath484 is proportional to an idempotent of @xmath478 if and only if : @xmath485 to prove the proposition , we solve by a direct analysis the system of 6 equations in @xmath486 , obtained by equating to 0 the coefficient in front of each element of @xmath487 in the expression @xmath488 ( we skip the details ) .
the first four evaluations found in the preceding proposition correspond to standard tableaux , namely , with the same notation as in ( [ et ] ) , @xmath489 \fbox{\tiny{3}}\end{array}}\!\!\!=\!\frac{1}{3}\phi(0,1,-1)\,,\ \ \ , e_{\!\!\!\begin{array}{l}\fbox{\tiny{1}}\fbox{\tiny{3}}\\[-0.3em ] \fbox{\tiny{2}}\end{array}}\!\!\!=\!\frac{1}{3}\phi(0,-1,1)\,,\ \ \ , e_{\!\!\!\begin{array}{l}\fbox{\tiny{1}}\\[-0.3em ] \fbox{\tiny{2}}\\[-0.3em ] \fbox{\tiny{3}}\end{array}}\!\!\!=\!\frac{1}{6}\phi(0,-1,-2)\,.\ ] ] for brevity , let @xmath490 and @xmath491 \fbox{\tiny{2}}\\[-0.3em ] \fbox{\tiny{3}}\end{array}}\!\!$ ] . for the two `` non - standard '' evaluations of the fusion function obtained in the above proposition , one can verify that the following elements @xmath492 are idempotents of @xmath478 such that @xmath493 where @xmath494 is the unit element of @xmath478 .
thus , @xmath495 is a complete system of pairwise orthogonal primitive idempotents of @xmath478 , and @xmath496 ( respectively , @xmath497 ) generates a minimal left ideal isomorphic , as an @xmath478-module , to the irreducible module corresponding to the partition @xmath498 . by the same arguments as in the preceding subsection , these two non - standard evaluations ( namely , with @xmath499 or @xmath500 ) lead to fused solutions of the yang baxter equation acting on spaces @xmath501 , where @xmath502 is any partition , alternative to the ones obtained in corollary [ coro - fin - sn ] .
it is an open question to determine if these `` non - standard '' fused solutions are related in some sense to the standard ones . in general ( for any @xmath65 ) , we think that `` non - standard '' fused solutions deserve a better study .
it is interesting to note that the two non - standard evaluations found in proposition [ prop - rem ] correspond actually to classical contents of non - standard young tableaux , namely the young tableaux @xmath503 and @xmath504 \fbox{\scriptsize{3}}\\[-0.2em]\fbox{\scriptsize{2}}\end{array}\!$ ] .
further , translating the three variables by @xmath174 or @xmath505 , which do not affect the fusion function , the non - standard evaluations also correspond to the non - standard young tableaux @xmath506 \fbox{\scriptsize{1}}\end{array}$ ] and @xmath507 \fbox{\scriptsize{3}}\end{array}$ ] ( we note that these young tableaux are maybe more natural in view of the coefficients @xmath508 in ( [ ns - idem ] ) ) . finally , we indicate that , for @xmath509 , there are some evaluations of the fusion function leading to an element of @xmath510 proportional to an idempotent and which do not correspond to any young tableaux , standard or not ( such an evaluation is , for example , @xmath511 , using a notation as in ( [ fus - s3 ] ) extended for @xmath509 ) .
@xmath130 as we have already explained above , not only the evaluations of the fusion function providing elements proportional to idempotents are of interest for the fusion procedure ; the evaluations of the fusion function giving non - invertible elements of @xmath20 furnish proper invariant subspaces as well . for @xmath58 , with the notation ( [ fus - s3 ] ) and @xmath481 evaluated to 0 , obvious examples of such evaluations ( which include the ones studied previously ) are @xmath512 further , one can verify that the two elements obtained by the following consecutive evaluations @xmath513 are non - singular and non - invertible ( we also mention that the evaluations described in this remark exhaust the evaluations of @xmath482 providing non - invertible elements of @xmath478).@xmath130
we explain how , with the help of a super " analogue of the schur
weyl duality @xcite , the fusion formula for the symmetric group can also be used in the context of the fusion procedure for a generalization of the yang solution .
we obtain a family of fused solutions acting on irreducible representations of the general linear lie superalgebra @xmath22 .
[ [ generalization - of - the - yang - solution - for - a - mathbbz-2mathbbz - graded - vector - space . ] ] * generalization of the yang solution for a @xmath514-graded vector space . * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we fix a @xmath514-decomposition of the vector space @xmath0 as @xmath515 and we set @xmath23 and @xmath516 . a non - zero vector @xmath517 is called homogeneous if @xmath518 or @xmath519 . for a homogeneous vector @xmath520 , @xmath521 , we set @xmath522 and let @xmath523 be the endomorphism of @xmath0 defined by @xmath524 for any homogeneous vector @xmath517 .
we define an operator @xmath525 by : @xmath526 and we let @xmath60 be the following function of @xmath69 with values in @xmath61 : @xmath527 where @xmath47 is the permutation operator on @xmath2 .
it is straightforward to check that the function @xmath60 is a solution of the yang
baxter equation on @xmath2 : @xmath528 moreover , this solution satisfies the same unitarity condition ( [ inv - r - sn ] ) as the yang solution , namely @xmath529 [ [ schurweyl - duality - in - the - mathbbz-2mathbbz - graded - setting . ] ] * schur weyl duality in the @xmath514-graded setting .
* + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath170 .
we recall that the symmetric group @xmath19 on @xmath65 letters is isomorphic to the group generated by elements @xmath530 subject to the defining relations : @xmath531 s_is_{i+1}s_i = s_{i+1}s_is_{i+1 } & & \text{for $ i=1,\dots , n-2$\,,}\\[0.3em ] s_is_j = s_js_i & & \text{for $ i , j=1,\dots , n-1 $ such that $ |i - j|>1$\ , , } \end{array}\ ] ] the isomorphism being given by @xmath532 , @xmath533 .
we define an operator @xmath534 by @xmath535 , that is , we have : @xmath536 the following map from the set of generators @xmath537 to the set @xmath53 , @xmath538 extends to an algebra homomorphism from @xmath20 to @xmath53 ( one easily checks that the relations ( [ sn ] ) are satisfied by the images of the generators ) .
we denote by @xmath539 this representation of @xmath20 .
let @xmath22 be the lie superalgebra of endomorphisms of @xmath0 , defined by the following @xmath14-grading : a non - zero element @xmath540 is homogeneous of degree 0 if @xmath541 , @xmath521 , and is homogeneous of degree 1 if @xmath542 and @xmath543 .
we let @xmath544 denote the degree of an homogeneous element @xmath540 .
let @xmath27 be the universal enveloping algebra of the lie superalgebra @xmath22 .
the map , with values in @xmath53 , defined on homogeneous element @xmath540 by @xmath545 g\otimes \textrm{id } \otimes\dots\otimes \textrm{id}+{\widetilde{\textrm{i}}}_v \otimes g\otimes \textrm{id}\dots\otimes \textrm{id}+\cdots\dots+{\widetilde{\textrm{i}}}_v \otimes\dots\otimes { \widetilde{\textrm{i}}}_v\otimes g\ , & \text{if $ |g|=1$\ , , } \end{array}\right.\ ] ] extends to a representation of the algebra @xmath27 on @xmath13 .
we denote by @xmath546 this representation .
we recall the generalization of the schur
weyl duality , which holds between the symmetric group @xmath19 and the algebra @xmath27 @xcite .
[ thm - sw - sn - s ] 1 .
the subalgebra @xmath547 of @xmath363 is the centraliser of @xmath548 .
the subalgebra @xmath548 of @xmath363 is the centraliser of @xmath547 .
3 . as an @xmath549-module defined by @xmath546 and @xmath539
, the space @xmath13 decomposes as : @xmath550 where @xmath90 runs over the set of partitions such that @xmath367 and @xmath551 if @xmath552 , and where the modules @xmath553 are irreducible @xmath27-modules .
[ [ fused - solutions - on - m_lambdaumathfrakgl_nmotimes - m_lambdaumathfrakgl_nm . ] ] * fused solutions on @xmath554 . * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + as a consequence of formulas ( [ yang - s ] ) and ( [ rho - sn - s ] ) , we have : @xmath555 therefore , defining the following rational function in the complex variables @xmath368 with values in @xmath20 : @xmath556 we have , recalling the definition ( [ hf ] ) , @xmath557 moreover , the following lemma relates the function @xmath558 with the fusion function @xmath370 , defined by ( [ phi - sym ] ) , of the symmetric group .
[ phi - tphi ] we have ( where @xmath280 is the longest element of @xmath19 ) : @xmath559 recall that the relation ( [ ijk ] ) holds for the functions @xmath373 with values in @xmath20 .
as moreover @xmath560 for @xmath561 , the lemma is a consequence of lemma [ lem - f - alt ] together with the fact that @xmath562 ( one applies lemma [ lem - f - alt ] with @xmath563 replaced by @xmath373 and @xmath357 replaced by @xmath42 , for @xmath561 ) .
thus , as @xmath280 is invertible , the fusion formula for the symmetric group in theorem [ thm - idem - sn ] can be used , as in subsection [ subsec - conc - sn ] , to analyze the image of @xmath564 in @xmath13 when @xmath368 are evaluated to classical contents of standard young tableaux as in ( [ idem - sn ] ) .
using corollary [ cor - hf ] , formula ( [ trho - hf ] ) and the schur weyl duality stated in theorem [ thm - sw - sn - s ] , we reproduce the same reasoning as in subsection [ subsec - conc - sn ] , to obtain the generalization of the corollary [ coro - fin - sn ] for the solution ( [ yang - s ] ) .
namely , let @xmath116 , @xmath397 are two standard young tableaux such that @xmath565 and @xmath566 , with @xmath567 , @xmath568 and @xmath569 if @xmath552 .
we obtain , by restriction of the fused operators to their invariant subspaces , a function @xmath570 with the following properties .
[ coro - fin - sn - s ] 1 .
the set of functions @xmath415 , where @xmath116 , @xmath397 are standard young tableaux as above , forms a family of solutions of the yang baxter equation , and
@xmath411 is an endomorphism of a space isomorphic to @xmath571 .
2 . for four standard young tableaux
@xmath116 , @xmath413 , @xmath397 and @xmath418 as above such that @xmath419 and @xmath420 , the endomorphisms @xmath411 and @xmath421 coincide up to a change of basis .
assume that @xmath572 .
let @xmath185 and define @xmath573 note that , contrary to the situation @xmath574 , the function @xmath575 is not a solution of the yang baxter equation .
this does not contradict the relation ( [ ijk ] ) because if @xmath232 and @xmath576 , we have @xmath577 .
actually , we have , if @xmath576 , @xmath578 it follows from ( [ rho - sn - s ] ) and ( [ ijk ] ) that the function @xmath579 satisfies , instead of ( [ yb - add - s ] ) , a braided " yang
baxter equation : @xmath580 with the braiding defined by the operator @xmath581 , namely with @xmath582 and @xmath583 . @xmath130
in this section , we will consider another example of solution of the yang
baxter equation , which is the standard deformation of the yang solution ( [ yang2 ] ) considered in section [ sec - sn ] .
we present the generalization of the construction in section [ sec - sn ] , which leads here to a family of fused solutions of the yang baxter equation acting on irreducible representations of the quantum group @xmath25 ( the standard deformation of @xmath344 ) .
we fix a basis @xmath584 of the vector space @xmath0 where @xmath352 .
let @xmath107 be a non - zero complex number and define an endomorphism @xmath274 of @xmath2 by , for @xmath585 , @xmath586 it is well - known , and it can be directly checked , that @xmath274 satisfies the quadratic relation @xmath587 and verifies as well , on the space @xmath128 , @xmath588 relations ( [ quadr - r ] ) and ( [ braid - r ] ) imply by a direct calculation that the function of @xmath88 , with values in @xmath61 , given by : @xmath589 satisfies the equation ( operators act on @xmath128 ) @xmath590 as a direct consequence , the function @xmath60 given by : @xmath591 is a solution , on @xmath2 , of the yang baxter equation with multiplicative spectral parameters : @xmath592 [ rem - def ] the solution @xmath593 is a deformation of the yang solution ( [ yang2 ] ) in the following sense . consider @xmath107 as a variable in @xmath594 and set @xmath595 . since @xmath596 , we obtain @xmath597 where we have used that @xmath598 .
thus @xmath599 -0.6cm@xmath130 [ ex - r - hn ] in this example , let @xmath429 .
we write the matrix of the endomorphism @xmath593 in the basis @xmath601 of @xmath2 ( points indicate coefficients equal to @xmath206 ) : @xmath602 -0.6cm@xmath130 the hecke algebra ( of type a ) is the associative algebra @xmath604 over @xmath44 generated by @xmath605 with the defining relations : @xmath606 \sigma_i\sigma_{i+1}\sigma_i=\sigma_{i+1}\sigma_i\sigma_{i+1 } & & \text{for $ i=1,\dots , n-2$\,,}\\[0.3em ] \sigma_i\sigma_j=\sigma_j\sigma_i & & \text{for $ i , j=1,\dots , n-1 $ such that $ |i - j|>1$\ , . } \end{array}\ ] ] the following map from the set of generators of @xmath604 to @xmath53 : @xmath607 extends to an algebra homomorphism .
this follows from relations ( [ quadr - r ] ) and ( [ braid - r ] ) , together with the obvious commutation relation @xmath608 if @xmath576 .
we denote by @xmath355 this representation of @xmath604 on the space @xmath13 .
the standard deformation of the universal enveloping algebra of @xmath15 is the associative algebra @xmath25 over @xmath44 defined by generators and relations as follows .
the generators are @xmath609 , @xmath610 , and @xmath611 , @xmath612 , and the defining relations are : @xmath613 k_ie_jk_i^{-1}=q^{a_{i , j}}e_j,\quad k_if_jk_i^{-1}=q^{-a_{i , j}}f_j & & \text{for $ i=1,\dots , n$ and $ j=1,\dots , n-1$\,,}\\[0.5em ] e_if_j - f_je_i=\displaystyle\delta_{i , j}\frac{k_ik_{i+1}^{-1}-k_i^{-1}k_{i+1}}{q - q^{-1 } } & & \text{for $ i , j=1,\dots , n-1$\ , , } \end{array}\ ] ] where @xmath614 is the kronecker delta , @xmath615 , @xmath616 , and @xmath617 otherwise , together with @xmath618 f_jf_i^2-(q+q^{-1})f_if_jf_i+f_i^2f_{j}=0 & & \text{for $ i , j=1,\dots , n-1 $ such that $ |i - j|=1$\,,}\\[0.5em ] e_ie_j = e_je_i,\quad f_if_j = f_jf_i & & \text{for $ i , j=1,\dots , n-1 $ such that $ |i - j|>1$\ , . } \end{array}\ ] ] the algebra @xmath25 admits a hopf algebra structure such that , in particular , the coproduct is the algebra homomorphism @xmath619 from @xmath25 to @xmath620 defined on the generators by : @xmath621 \delta(e_j)=e_j\otimes 1+k_jk_{j+1}^{-1}\otimes e_j,\quad \delta(f_j)=f_j\otimes
k_j^{-1}\!k_{j+1}+1\otimes f_j & & \text{for $ j=1,\dots , n-1$\ , . }
\end{array}\ ] ] the vector representation @xmath622 of the algebra @xmath25 on @xmath0 is given , on the basis @xmath623 , by : @xmath624 \eta(e_j)\bigl(e_k\bigr)=\delta_{j+1,k}e_{k-1},\quad \eta(f_j)\bigl(e_k\bigr)=\delta_{j-1,k}e_{k+1 } & & \text{for $ k=1,\dots , n$ and $ j=1,\dots , n-1$\ , . }
\end{array}\ ] ] via the coproduct @xmath619 , the representation @xmath622 induces a representation @xmath353 of the algebra @xmath25 on the space @xmath13 .
more precisely , let @xmath625 and define inductively @xmath626 for @xmath627 , where @xmath628 is the identity homomorphism of @xmath25 .
thus , for all @xmath629 , @xmath630 is an algebra homomorphism from @xmath25 to @xmath631 .
explicitly , we have : @xmath632 \delta^{(k)}(e_j)=\displaystyle\sum_{p=0,\dots , k-1}(k_jk_{j+1}^{-1})^{\otimes p}\otimes e_j\otimes 1^{\otimes k-1-p } & & \text{for $ j=1,\dots , n-1$\,,}\\[1.4em ] \delta^{(k)}(f_j)=\displaystyle\sum_{p=0,\dots , k-1}1^{\otimes p}\otimes f_j\otimes ( k_j^{-1}\!k_{j+1})^{\otimes k-1-p } & & \text{for $ j=1,\dots , n-1$\ , . }
\end{array}\ ] ] then the representation @xmath353 of the algebra @xmath25 on the space @xmath13 is given by : @xmath633 to state the analogue of the schur
weyl duality ( theorem [ thm - sw - sn ] ) between the hecke algebra @xmath604 and the quantum algebra @xmath25 , we recall that the irreducible representations of @xmath604 are parametrized by the partitions of @xmath65 . for any partition @xmath17 of @xmath65 , we denote by @xmath634 the corresponding irreducible @xmath604-module ( with the convention that @xmath635 is the one - dimensional @xmath604-module on which the generators @xmath605 act by multiplication by @xmath107 ) . for any partition @xmath17 such that @xmath360 ,
we denote by @xmath636 the irreducible highest weight @xmath25-module of highest weight @xmath17 ( where the eigenvalues of the generators @xmath637 are of the form @xmath638 , @xmath639 ) . 1 .
the subalgebra @xmath640 of @xmath363 is the centraliser of @xmath641 .
the subalgebra @xmath641 of @xmath363 is the centraliser of @xmath640 .
3 . as an @xmath642-module defined by @xmath353 and @xmath355
, the space @xmath13 decomposes as : @xmath643 where @xmath17 runs over the set of partitions such that @xmath367 and @xmath360 .
let @xmath170 and recall the definition ( [ hf ] ) of the function @xmath644 of @xmath349 ( with the multiplicative version for the spectral parameters , see remark [ rem - mult ] ) with values in @xmath53 : @xmath645 due to corollary [ cor - hf ] , when the function @xmath644 is non - singular at a particular value @xmath138 , the subspace @xmath7 is the image of the operator @xmath295 . following formulas ( [ hf2 ] ) , ( [ sw - hn ] ) and ( [ hra - hn ] )
, we define the following rational function in variables @xmath646 with values in the algebra @xmath604 : @xmath647 where @xmath648 for all @xmath533 .
the elements @xmath649 with values in @xmath604 are the _
baxterized elements _ of the hecke algebra .
we also set @xmath650 the main result concerning the function @xmath651 is the following ( we follow @xcite ) .
let @xmath17 be a partition of @xmath65 and let @xmath116 be a standard young tableau of shape @xmath17 . for brevity , set @xmath652 for @xmath56 .
[ thm - idem - hn ] the element obtained by the following consecutive evaluations @xmath653 is a primitive idempotent of @xmath604 which generates a minimal left ideal isomorphic , as an @xmath604-module , to the irreducible module @xmath634 .
[ rem - cl ] in the same spirit as in remark [ rem - def ] , we verify that @xmath655 where we recall that @xmath656 with the isomorphism given by @xmath657 , @xmath533 . thus ,
setting @xmath658 , for @xmath659 , we obtain : @xmath660 using lemma [ phi - tphi ] , we conclude that @xmath661 where @xmath370 is the fusion function of the symmetric group defined by ( [ phi - sym ] ) . in this sense ,
theorem [ thm - idem - sn ] is the classical limit ( @xmath662 ) of theorem [ thm - idem - hn ] .
@xmath130 according to formulas ( [ hf2 ] ) and ( [ phi - hn ] ) , we have that @xmath666 the assertions of theorem [ thm - idem - hn ] imply that @xmath644 is non - singular for the consecutive evaluations of the variables as in ( [ def - hft ] ) .
thus the image @xmath667 of the operator @xmath668 is well - defined and coincides , since @xmath669 is invertible , with the image in @xmath13 of the operator @xmath670 .
let @xmath320 and denote by @xmath406 the young tableau of shape @xmath17 obtained from @xmath116 by exchanging the nodes with numbers @xmath306 and @xmath305 .
we assume that @xmath406 is also standard , which is equivalent to the fact that @xmath673 .
the solution ( [ ra - hn ] ) satisfies the following unitarity condition , implied by ( [ quadr - r ] ) , @xmath674 therefore , according to definition [ def - adm ] , the condition for the tableau @xmath406 to be standard is equivalent to the condition for @xmath321 to be an admissible transposition for @xmath593 and the set of parameters @xmath409 . as in subsection [ subsec - conc - sn ] , corollary [ coro - fin - sn ] , we sum up the results obtained for the standard deformation of the yang solution . for two standard young tableaux @xmath116 , @xmath397 such that @xmath385 and @xmath398 , with @xmath367 , @xmath399 and @xmath400 , we set @xmath675 , @xmath676 and we define @xmath677 to be the restriction of the operator @xmath404 to the subspace @xmath678 : @xmath679 1 .
the set of functions @xmath415 , where @xmath116 , @xmath397 are standard young tableaux such that @xmath416 , forms a family of solutions of the yang
baxter equation , where @xmath680 is an endomorphism of a space isomorphic to @xmath681 .
2 . for four standard young tableaux
@xmath116 , @xmath413 , @xmath397 and @xmath418 as above such that @xmath682 and @xmath683 , the endomorphisms @xmath680 and @xmath684 coincide up to a change of basis .
[ ex - fus - hn ] this example is the ( deformed ) analogue of example [ ex - fus - sn ] .
namely , we consider the fused operator @xmath685 , with @xmath179 , @xmath686 and @xmath687 , obtained from the solution ( [ ra - hn ] ) .
the expression for @xmath685 is @xmath688 the results of this section , applied to this example , give that the operator @xmath685 preserves the subspace @xmath689 , where @xmath690 is the space of the irreducible representation of @xmath25 corresponding to the partition @xmath427 .
the subspace @xmath691 is the image of the operator @xmath692 and is a ( quantum ) analogue of the symmetric square of @xmath0 .
we give the matrix of the endomorphism @xmath685 on the space @xmath695 written in the basis @xmath696 ordered lexicographically ( the points indicate the coefficients equal to @xmath206 ) : @xmath697_{_{\alpha}}}{[-1]_{_{\alpha}}}}{{[0]_{_{\alpha}}}{[1]_{_{\alpha}}}}\!\!\!&\cdot&\cdot&\cdot&\cdot&\cdot&\cdot&\cdot&\cdot \\
\cdot&\cdot&1&\cdot&\frac{-q [ 2]^2\alpha}{{[1]_{_{\alpha}}}}&\cdot&\!\!\!\frac{[2]\alpha^2}{{[0]_{_{\alpha}}}{[1]_{_{\alpha}}}}\!\!\!&\cdot&\cdot \\ \cdot&\frac{-[2]{[-1]_{_{\alpha}}}}{{[0]_{_{\alpha}}}{[1]_{_{\alpha}}}}&\cdot&\frac{{[-1]_{_{\alpha}}}}{{[1]_{_{\alpha}}}}&\cdot&\cdot&\cdot&\cdot & \cdot\\ \cdot&\cdot&\frac{-q^{-1}}{{[1]_{_{\alpha}}}}&\cdot&\frac{p(\alpha)}{{[0]_{_{\alpha}}}{[1]_{_{\alpha}}}}&\cdot&\frac{-q^{-1}}{{[1]_{_{\alpha}}}}&\cdot & \cdot\\ \cdot&\cdot&\cdot&\cdot&\cdot&\frac{{[-1]_{_{\alpha}}}}{{[1]_{_{\alpha}}}}&\cdot&\frac{-[2]{[-1]_{_{\alpha}}}\alpha}{{[0]_{_{\alpha}}}{[1]_{_{\alpha}}}}\!\!\!&\cdot \\ \cdot&\cdot&\!\!\!\frac{[2]}{{[0]_{_{\alpha}}}{[1]_{_{\alpha}}}}\!\!\!&\cdot&\frac{-q [ 2]^2}{{[1]_{_{\alpha}}}}&\cdot&1&\cdot&\cdot \\ \cdot&\cdot&\cdot&\cdot&\cdot&\frac{-[2]{[-1]_{_{\alpha}}}}{{[0]_{_{\alpha}}}{[1]_{_{\alpha}}}}&\cdot&\frac{{[-1]_{_{\alpha}}}}{{[1]_{_{\alpha } } } } & \cdot\\ \cdot&\cdot&\cdot&\cdot&\cdot&\cdot&\cdot&\cdot & \frac{{[-2]_{_{\alpha}}}{[-1]_{_{\alpha}}}}{{[0]_{_{\alpha}}}{[1]_{_{\alpha}}}}\\ \end{array}\right)\,,\ ] ] where we have set ( @xmath698 ) @xmath699:=\displaystyle\frac{q^n - q^{-n}}{q - q^{-1}}\,,\ \ \ { [ n]_{_{\alpha}}}:=\displaystyle\frac{q^n\alpha - q^{-n}}{q - q^{-1}}\ \ \ \ \text{and}\ \ \ \ p(\alpha):=\displaystyle\frac{q^{-1}\alpha^2+(q^3 - 2q-2q^{-1}+q^{-3})\alpha+q}{(q - q^{-1})^2}\,.\ ] ] taking @xmath595 and performing the `` classical '' limit ( @xmath662 ) on the coefficients of the above matrix , we obtain the matrix ( [ mat - sn ] ) displayed in example [ ex - fus - sn ] ( as it should be in view of remark [ rem - def ] ) . to perform the classical limit
, we use that : @xmath699\bigr\vert_{q=1}=n\,,\ \ \ \ [ n]_{q^{2u}}\bigr\vert_{q=1}=u+n\ \ \ \ \text{and}\ \ \ \
p(q^{2u})\bigr\vert_{q=1}=u^2-u+2\ ; \ ] ] this can be obtained , for example , by setting @xmath700 , expanding in powers of @xmath701 ( up to the order 2 ) and letting @xmath701 tend to 0 .
we remark the following facts appearing in the above deformed version of the matrix ( [ mat - sn ] ) : * while the factors @xmath312 in ( [ mat - sn ] ) are deformed in @xmath702 $ ] , the factors @xmath703 become @xmath702 ^ 2 $ ] ( and not @xmath704 $ ] ) .
* the factors @xmath705 are deformed into @xmath706_{_{\alpha}}}$ ] , while the factor @xmath707 is deformed into @xmath708 . * in addition to the `` deformation rules '' described in the preceding items , powers of @xmath107 and of @xmath709 , which do not affect the classical limit , appear in some coefficients of the matrix .
@xmath130 as in subsection [ subsec - rem ] , other evaluations of the fusion function ( [ phi - hn ] ) leading to non - invertible elements of the hecke algebra @xmath604 are of interest in the framework of the fusion procedure for the solution ( [ ra - hn ] ) .
we only indicate here the analogues of the `` non - standard '' idempotents found in proposition [ prop - rem ] ( see ( [ ns - idem ] ) ) : @xmath710 which are pairwise orthogonal primitive idempotents of @xmath711 generating minimal left ideals isomorphic to the irreducible module @xmath712 .
the construction of fused solutions acting on irreducible representations of @xmath27 in section [ sec - sn - s ] also admits , for any @xmath713 , a generalization to the quantum " setting .
it relies on the schur
weyl duality between the hecke algebra @xmath604 and a quantum analogue of @xmath27 proved in @xcite .
we only indicate the relevant generalization ( for any @xmath713 ) of the solution @xmath593 , see ( [ ra - hn ] ) , and the corresponding action of @xmath604 on @xmath13 . [ [ generalization - of - the - solution - ralpha - for - a - mathbbz-2mathbbz - graded - vector - space . ] ] * generalization of the solution @xmath593 for a @xmath514-graded vector space . * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we fix a basis @xmath714 of the vector space @xmath0 , such that @xmath715 is a basis of the subspace @xmath716 and @xmath717 is a basis of the subspace @xmath718 .
let @xmath719 be defined by , for @xmath720 , @xmath721 ( -1)^{|e_i||e_j|}\,e_j\otimes e_i & \text{if $ i < j$,}\\[0.4em ] ( -1)^{|e_i||e_j|}\,e_j\otimes e_i + ( q - q^{-1})\,e_i\otimes e_j & \text{if $ i > j$. } \end{array}\right.\ ] ] note that , for @xmath574 , this operator coincides with the operator defined by ( [ hr - hn ] ) , and that , if we take @xmath722 , we obtain the operator @xmath581 defined by ( [ def - tp ] ) . as in subsection [ subsec - def - yang ] ,
one directly checks that : @xmath723 { \widehat{r}}_{1,2}{\widehat{r}}_{2,3}{\widehat{r}}_{1,2}={\widehat{r}}_{2,3}{\widehat{r}}_{1,2}{\widehat{r}}_{2,3}\ \ & \text{on $ v\otimes v\otimes v$\ , , } \end{array}\ ] ] and thus that the following function of @xmath88 with values in @xmath61 : @xmath724 is a solution , on @xmath2 , of the yang baxter equation with multiplicative spectral parameters : @xmath592 for this action of @xmath604 , an analogue of the jimbo schur weyl duality , theorem [ thm - sw - hn ] , is proved in @xcite , and involves a quantum deformation of @xmath27 instead of @xmath25 ( we refer to @xcite for the precise definition of the deformation of @xmath27 ) .
alternatively , the result in @xcite can be seen as a quantum analogue of theorem [ thm - sw - sn - s ] , with @xmath27 replaced by its quantum deformation and the symmetric group @xmath19 replaced by the hecke algebra @xmath604 .
we note that the direct sum in ( [ sw - sn - for - s ] ) remains , in the quantum setting , over the same set of partitions .
now , the exact same reasoning as in subsection [ subsec - fin - hn ] can be reproduced and leads to the construction of fused solutions acting on spaces isomorphic to irreducible representations of the quantum deformation of @xmath27 .
namely , we obtain a @xmath14-graded analogue of corollary [ coro - fin - hn ] , or in other words , a quantum analogue of corollary [ coro - fin - sn - s ] .
we omit the details as they are repetitions of the preceding sections .
m. nazarov , _ yangians and capelli identities _ , in : `` kirillov s seminar on representation theory '' ( g. i. olshanski , ed . ) amer .
181 , amer .
soc . , providence , ri , ( 1998 ) , 139163 .
arxiv : q - alg/9601027 | we first review the fusion procedure for an arbitrary solution of the yang baxter equation and the study of distinguished invariant subspaces for the fused solutions .
then we apply these general results to four particular solutions : the yang solution , its standard deformation and their generalizations for super vector spaces . for the yang solution , respectively , its super " generalization , we explain how , using the fusion formula for the symmetric group together with the ( super ) schur weyl duality , the fusion procedure allows to construct a family of fused solutions of the yang
baxter equation acting on irreducible representations of the general linear lie algebra , respectively , of the general linear lie superalgebra . for the deformations of the two previous solutions , we use the fusion formula for the hecke algebra together with the ( super ) quantum schur
weyl duality to obtain fused solutions acting on irreducible representations of the quantum groups associated to the general linear lie ( super)algebras . |
the hot gas that is filling clusters of galaxies delineates the gravitational potential of the cluster dark matter . this enables a derivation of the cosmologically important cluster masses from the x - ray data ( e.g. , bahcall & sarazin 1977 ; sarazin 1988 ) .
clusters form via mergers that stir the gas ( e.g. , roettiger , loken , & burns 1997 and other simulations ) , after which the cluster should be left alone for a few gyr to reach hydrostatic equilibrium for the mass measurement to work .
clusters with regular x - ray morphology and the x - ray brightness strongly peaked on a cd galaxy should have formed sufficiently long ago and are therefore believed to be the safest candidates for x - ray mass measurements . for a number of clusters , however , the x - ray - derived masses are systematically lower than the independent gravitational lensing measurements ( e.g. , miralda - escud & babul 1995 and later works ) .
both lensing and x - ray measurements can suffer from systematic effects such as line of sight substructure and inadequate x - ray modeling ( e.g. , bartelmann & steinmetz 1996 ; allen , ettori , & fabian 2001 ) .
a1795 ( @xmath2 ) is one of the most relaxed clusters in x - rays ( e.g. , buote & tsai 1996 ; briel & henry 1996 ) .
it has had sufficient time to develop a cool core ( e.g. , edge , stewart , & fabian 1992 ) , although its precise nature is debated ( tamura et al .
zooming in on the center with , fabian et al .
( 2001 ) discovered a 40 filament of cool gas extending from the cd galaxy , coincident with the optical h@xmath3 filament ( cowie et al .
a total mass profile was derived from data by ettori et al .
( 2000 ) .
below we analyze the same data set and show that even in the centers of such relaxed clusters as a1795 , full hydrostatic equilibrium may not have been achieved , resulting in considerable x - ray mass underestimates . a similar finding
was earlier reported for the otherwise relaxed cluster rxj1720 + 26 ( mazzotta et al .
we measure the bulk velocity and , for the first time , acceleration of the gas near the cluster center by analyzing a mild brightness edge , similar to those detected by in several other clusters ( markevitch et al . 2000 ; vikhlinin , markevitch & murray 2001 ; mazzotta et al .
we use @xmath4 and @xmath5 ; if not specified , confidence intervals are 68% for one - parameter .
( 0,14.3)(18.5,24 ) ( 0.3,24.0)=8.5 cm ( 9.8,24.0)=8.5 cm ( 8.0,23.3)a ( 17.5,23.3)b ( -0.1,15.8 ) = 3.5 mm fig .
1.(_a _ ) acis image in the 0.75 kev band .
pixels are 2 , point sources are excluded .
a mild brightness edge is seen @xmath6 kpc south of the center .
( _ b _ ) the same image smoothed by a gaussian with @xmath7 ( brightness increases from gray to red to green ) .
the cool filament ( fabian et al . 2001 ) is seen in the center .
dashed ellipse centered on the cd galaxy approximately delineates the edge .
dashed lines show two 60 sectors used for extracting radial profiles .
we have combined two archival acis - s observations of a1795 performed in december 1999 and march 2000 , limiting the analysis to the s3 chip .
the standard data cleaning resulted in the total useful exposure of 36 ks .
the background was modeled as in , e.g. , markevitch et al .
( 2000 ) . the cluster 0.75 kev image is shown in fig . 1 .
in the center , there is the gas filament discovered by fabian et al .
( 2001 ) . south of the center , outside the filament region
, one can see a subtle edge feature that is the subject of this work . the radial brightness profile , extracted in the southern 60 sector centered on the cd galaxy ( see fig .
1_b _ ) is shown in fig .
2_a_. as other profiles below , it is derived in narrow elliptical segments parallel to the edge ( the ellipse in fig .
1_b _ ) , and plotted as a function of the average emission - weighted radius in each segment . in the narrow sectors of interest , these segments are reasonably close to circular , so below we assume spherical symmetry for simplicity of our qualitative analysis .
the brightness profile has the characteristic shape of a projected spherical gas density discontinuity . in the radial range
@xmath8 ( within a factor of 2 of the edge radius ) , the profile is fit very well by a gas density model consisting of two power laws ( @xmath9 and @xmath10 ) and a jump by a factor of @xmath11 ( see fig .
2_c _ ) . the observed temperature differences ( described below ) have a negligible effect on the density model .
closer to the center lies the filament region , so we do not include smaller radii in the analysis and show them only for illustration . at larger radii
, the density profile is known to steepen ( vikhlinin , forman , & jones 1998 ) but this is unimportant for our fits . for spectral modeling , we averaged the position - dependent telescope effective area , the detector nonuniformity and spectral response over a given region with the cluster brightness as weight .
response matrices released on 2001 august 8 were used . the temperature profile in the southern sector ,
derived using galactic absorption and the kaastra ( 1992 ) plasma model , is shown in fig .
2_b_. for the bins inside the edge , we have subtracted the projected contribution from the outer , hotter gas , assuming a 6.7 kev temperature ( the average for the 3 outer bins ) and the best - fit density model .
the resulting profile is consistent with two constants inside and outside the edge ; thus , any more complex deprojection is not warranted .
the outside temperature is in agreement with the 67 kev average observed at larger radii ( e.g. , markevitch et al .
1998 ; arnaud et al.2001 ) .
since our analysis includes regions with short cooling time , we checked if any multicomponent gas could bias our measurements .
we first tried to fit our spectra with a simple two - temperature or standard cooling flow model , which did not improve the fits . on the other hand ,
the spectrum of the central region ( @xmath12 ) exhibits lines of ionized si and s at 1.9 kev and 2.4 kev ( rest frame ) that can be fit with a heavily - absorbed ( @xmath13 ) , cool ( @xmath14 kev ) thermal component , in addition to the main 34 kev gas . the origin of this emission is beyond the scope of this paper , but we have looked for a similar contaminant in the regions of our interest . for this ,
only the second observation taken at the acis temperature of @xmath15c was used , for which the current response matrices have adequate accuracy ( the @xmath16c matrices are sufficient to fit the continuum shape but not the lines ) . in the combined 3 spatial bins outside the edge ( solid crosses in fig .
2_b _ ) , such a component is not required .
for the inner 3 bins , the fit is marginally improved ( @xmath17 for 3 additional d.o.f . ) ; however , the best - fit temperature of the main gas does not change and the fit does not allow a significant filling factor for these cool clouds .
also , the cool component gives only 2% of the flux in the 0.75 kev band and can not bias our density profile for the main gas .
thus , the derived quantities for the diffuse gas in our regions of interest should not be affected .
( 0,-0.5)(18.5,15.3 ) ( 0.2,15.9)=9.5 cm ( 0.2,10.5)=9.5 cm ( 9.4,16.18)=9.5 cm ( 9.4,10.5)=9.5 cm ( 8.1,14.0)a ( 8.1 , 2.8)b ( 17.3,14.0)c ( 17.3,10.5)d ( 17.3 , 6.8)e ( -0.1,1.4 ) = 3.5 mm fig .
2.profiles from the southern 60 sector containing the edge .
errors are 68% .
vertical dotted lines show the edge position .
( _ a _ ) surface brightness in the 0.75 kev band .
the line is a fit in the @xmath18 range ; the corresponding density model is shown in panel ( _ c _ ) .
( _ b _ ) temperature profile with approximate deprojection ( see text ) .
solid crosses show values used for derivation of the pressure and total mass across the edge .
( _ d _ ) pressure profile .
( _ e _ ) total mass within a given radius ( confidence bands are shown in gray ) . for comparison , a fit in the sector opposite to the edge ( see figs . 1_b _ and 3 ) is shown by dashed line
; they must show the same mass if the gas is in hydrostatic equilibrium .
the temperature profile reveals that the edge is a `` cold front '' , or a contact discontinuity between two moving gases , similar to ( although not quite as prominent as ) those discovered by in several other clusters . from the edge geometry
, the denser gas should be moving away from the center ( or the outer gas flowing onto the center ) . multiplying the gas density by its temperature
, we obtain a profile of thermal pressure across the edge ( fig .
it is continuous ; if we fit the inner 3 and outer 3 temperature values ( solid crosses in fig .
2_b _ ) by power law functions of radius , the pressure jump is @xmath19 ( a 90% error including the scatter of the power law slopes ) .
this leaves little or no room for a relative gas motion , since the ram pressure from such a motion would cause the inner thermal pressure to be higher compared to that on the outside ( beyond the small stagnation region ) , as is indeed seen in a3667 ( vikhlinin et al .
2001 ) the above upper limit on the pressure jump corresponds to a 90% upper limit on the mach number of 0.5 ( see , e.g. , fig .
6 in vikhlinin et al . 2001 ) the limit is probably even stronger , because by fitting the temperature profiles with power laws in narrow radial intervals , we may allow too much unphysical freedom . if instead one assumes constant temperatures inside and outside , the mach number should be very close to 0 .
thus the inner and outer gases appear very nearly at rest and in pressure equilibrium .
therefore , one expects them to be in hydrostatic equilibrium in the cluster gravitational potential . under this assumption and
that of spherical symmetry , we can derive the cluster total mass profile in this radial range using our power - law fits to the temperature and density profiles in the immediate vicinity of the edge .
this mass profile is shown in fig .
2_e _ ( the error bands include uncertainties of the local temperature gradients ; uncertainties on the density gradients are negligible in comparison ) .
the profile reveals an unphysical discontinuity by a factor of 2 , with a formal significance of 99% . for comparison
, we analyzed the northern sector of the cluster opposite to the edge .
a brightness profile there shows no features and is fit well by the customary @xmath20-model ( fig .
the temperature is nearly constant , but for exactness , we have modeled its slow radial rise by a smooth function .
its emission - weighted projection is overlaid on the data in fig .
3_b_. the total mass profile derived using these fits is shown by dashed line in fig.2_e_. outside the edge radius , the masses derived from the two opposite sectors agree , as they must .
this indicates that the gas immediately outside the edge is indeed near hydrostatic equilibrium , while the gas inside the edge is not .
similar mass discontinuities at the cold fronts were observed in a3667 ( vikhlinin et al .
2001 ) and rxj1720 + 26 ( mazzotta et al .
however , in the former , the cold front is a boundary of a subcluster flying with the sonic velocity where one does not expect hydrostatic equilibrium .
the latter cluster is apparently relaxed on large scales so the situation is similar to a1795 , but there , too , a high velocity of the front could not be excluded .
mazzotta et al .
( 2001a ) proposed that the core of rxj1720 + 26 is moving relative to the rest of the cluster as a result of the peculiar history of that cluster s collapse .
however , such an explanation for a1795 ( as well as several other generally relaxed clusters with similar features , m. markevitch et al . in preparation ) may require too much of a coincidence .
we propose instead that the cool gas just inside the edge is `` sloshing '' north - south in the central potential well ( which may be static at present ) and is now observed near the point of maximum displacement , where it has zero velocity but nonzero centripetal acceleration .
such a gas would not be in hydrostatic equilibrium with the potential instead , its distribution would reflect the reduced gravity force in the accelerating reference frame , resulting in the apparent mass underestimate .
assuming that the gas outside the edge is hydrostatic , the acceleration of the moving gas can be estimated from the mass jump @xmath21 as @xmath22 cm s@xmath23 or @xmath24 , where @xmath25 is the radius of the edge .
it is interesting to estimate the gravitational potential energy of this gas . approximately , @xmath26 erg g@xmath27 . for a 4 kev gas , this is about half of its thermal energy .
as the gas accelerates , this energy becomes available for dissipation .
thus , such sloshing can produce significant heating in the cluster center that may offset some radiative cooling .
unlike a3667 ( the site of the most prominent cold front ) where the cool gas is retained by a merging subcluster and travels together with its distinct dark matter halo ( a. vikhlinin et al . in preparation ) , here the moving gas is probably not attached to any underlying moving dark matter body .
indeed , simulations of cluster mergers ( e.g. , roettiger et al .
1997 ) predict that late into a merger , the gas decouples from dark matter and sloshes in the changing gravitational potential for a significant time before coming to a hydrostatic equilibrium .
= 3.5 mm fig .
3.profiles from the northern 60 sector opposite to the edge .
( _ a _ ) surface brightness .
the line shows a fit by a @xmath20-model .
( _ b _ ) projected temperature profile .
the line shows a best - fit model used to estimate the total mass ( fig .
errors are 68% .
however , a recent major merger in a1795 is unlikely because of the exceptional regularity of its x - ray image .
also , the gas temperature decreases toward the center , indicating that either the gas has had time to cool ( @xmath28 yr at the edge radius ) or the cool gas from subcluster infall has had time to settle down , both processes requiring the absence of significant disturbance .
perhaps a localized gas motion in the core might be caused by a disturbance of the central gravitational potential by an infalling small subcluster .
such a disturbance would supply kinetic energy to the low - entropy gas accumulated over time in the cluster center , which is more responsive to small potential perturbations than the hotter outer gas .
alternatively , in clusters where the central agn produces large bubbles of hot plasma ( e.g. , churazov et al .
2000 ; mcnamara et al.2000 ) , such activity could supply kinetic energy to the surrounding cool gas .
there are no obvious large bubbles in a1795 .
we note that the central cool filament , extending southward from the cd galaxy ( fig .
1_b _ ) , is aligned with the apparent direction of the most recent gas movement .
this suggests that the cd galaxy in a1795 does not move with the gas but probably stays with the gravitational potential ( to within its observed small 150 velocity relative to the cluster average ; oegerle & hill 1994 ) .
the filament is probably a cooling wake arising as the cd galaxy is moving relative to the gas , as proposed by fabian et al.(2001 ) , only their relative motion is not due to the cd galaxy s oscillations in the potential but rather to the sloshing of the gas .
the filament length , @xmath1 kpc , should then give the amplitude of the gas oscillations in the central potential well .
the observed mass discontinuity may have significant effect on x - ray measurements of the total mass at the cluster centers .
it is at these radii where strong lensing often implies 23 times higher values ( e.g. , miralda - escud & babul 1995 and later works ) , although more accurate x - ray modeling was shown to reduce the discrepancy ( e.g. , allen et al.2001 ) .
given the filament , the edge and the overall image geometry , it appears likely that most of the gas at small radii in a1795 participates in the gas bulk flow and is therefore nonhydrostatic , even though we can see the effect on the mass profile only in the southern sector where sloshing produces a brightness edge .
data on a number of other relaxed clusters reveal similar fronts in the cluster central regions ( m. markevitch 2001 , in preparation ) so such nonequilibrium may be common .
note that the volume inside the edge radius contains only @xmath29% of the gas within the virial radius , so our finding can not be generalized to the whole cluster . in summary
, we have studied a mild gas density edge near the center of the otherwise relaxed cluster a1795 .
the gas on two sides of the edge appears at rest ; however , the inner gas has significant centripetal acceleration , indicating that the gas is sloshing in the cluster gravitational potential .
if one assumes hydrostatic equilibrium , one obtains an underestimate of the total mass near the center . | data on a1795 reveal a mild edge - shaped discontinuity in the gas density and temperature in the southern sector of the cluster at @xmath0 kpc .
the gas inside the edge is 1.31.5 times denser and cooler than outside , while the pressure is continuous , indicating that this is a `` cold front '' , the surface of contact between two moving gases .
the continuity of the pressure indicates that the current relative velocity of the gases is near zero , making the edge appear to be in hydrostatic equilibrium .
however , a total mass profile derived from the data in this sector under the equilibrium assumption , exhibits an unphysical jump by a factor of 2 , with the mass inside the edge being lower .
we propose that the cooler gas is `` sloshing '' in the cluster gravitational potential well and is now near the point of maximum displacement , where it has zero velocity but nonzero centripetal acceleration .
the distribution of this non - hydrostatic gas should reflect the reduced gravity force in the accelerating reference frame , resulting in the apparent mass discontinuity .
assuming that the gas outside the edge is hydrostatic , the acceleration of the moving gas can be estimated from the mass jump , .
the gravitational potential energy of this gas that is available for dissipation is about half of its current thermal energy .
the length of the cool filament extending from the cd galaxy ( fabian et al . ) may give the amplitude of the gas sloshing , @xmath1 kpc . such gas bulk motion might be caused by a disturbance of the central gravitational potential by past subcluster infall . |
since its induction @xcite , distance correlation has had many applications in , e.g. , life science @xcite and variable selection @xcite , and has been analyzed @xcite , extended @xcite in various aspects . if distance correlation were implemented straightforwardly from its definition , its computational complexity can be as high as a constant times @xmath0 for a sample size @xmath2 .
this fact has been cited for numerous times in the literature as a disadvantage of adopting the distance correlation . in this paper , we demonstrate that an o(@xmath3 ) algorithm for a version of the distance correlation exits .
the main idea behind the proposed algorithm is to use an idea rooted in the the avl tree structure @xcite .
the same idea has been utilized to develop fast algorithm for computing the kendall s @xmath4 rank correlation coefficient @xcite @xcite .
we extend it to make it suitable for our purpose .
the derivation of the fast algorithm also involves significant reformulation from the original version of the distance correlation .
details are presented in this paper . in simulations ,
not only we demonstrate the effectiveness of the fast algorithm , but also we testify that the advantage of using distance correlation ( in comparison with other existing methods ) become more evident when the sample sizes increase .
these experiments become feasible due to the availability of the proposed fast algorithm . in one experiment ( see details in section [ sec : screening ] ) , we increased the sample size by @xmath5 fold from a previously published simulation study .
the rest of this paper is organized as follows .
section [ sec : reviewdistcorr ] reviews the distance covariance / correlation and its relevant properties . in section [ sec : unbiased ] , we consider a reformulation of the distance covariance , such that the new estimator is both unbiased and a u - statistic . in section [ sec : algorithm ] , an algorithm with the average complexity of @xmath6 was presented .
extensive simulations are presented in section [ sec : simulations ] to demonstrate the additional capability we obtained due to the proposed fast algorithm .
finally , some concluding remarks are made in section [ sec : conclude ] .
detailed description of the algorithm is relegated to the appendix , along with most of the technical proofs .
distance covariance and distance correlation was introduced in 2005 by one of the co - authors of this paper , g. j. szkely , in several lectures to address the deficiency of pearson s correlation , namely that the pearson s correlation can be zero for dependent variables . in the following , we start with a definition of the distance covariance .
the population distance covariance between random vectors @xmath7 and @xmath8 with finite first moments is the nonnegative number @xmath9 defined by a weighted @xmath10 norm measuring the distance between the joint characteristic function ( c.f . )
@xmath11 of @xmath7 and @xmath8 , and the product @xmath12 of the marginal c.f.s of @xmath7 and @xmath8 .
if @xmath7 and @xmath8 take values in @xmath13 and @xmath14 , respectively , @xmath15 is @xmath16 where @xmath17 .
the integral exists provided that @xmath7 and @xmath8 have finite first moments .
this immediately shows that distance covariance is zero if and only if the underlying random variables are independent .
the beauty of kernel @xmath17 is that the corresponding sample statistic has the following surprisingly simple form .
denote the pairwise distances of the @xmath7 observations by @xmath18 and the pairwise distances of the @xmath8 observations by @xmath19 for @xmath20 and denote the corresponding double centered distance matrices by @xmath21 , and @xmath22 where @xmath23 @xmath24 it is clear that the row sums and column sums of these double centered matrices are 0 .
the squared sample distance covariance is the following simple formula @xmath25 the corresponding squared sample variance is @xmath26 and we can define the sample distance correlation as the standardized sample covariance : @xmath27 for more details see and a discussion paper @xcite .
it is clear that @xmath28 is rigid motion invariant and scale invariant . for recent applications of distance correlation ,
see e.g. , * li2012 and .
the population version of distance covariance and distance correlation can be defined without characteristic functions , see .
this definition is as follows .
let @xmath29 and @xmath30 be random variables with finite expectations .
the random distance functions are @xmath31 and @xmath32 . here
the primed random variable @xmath33 denotes an independent and identically distributed ( i.i.d . )
copy of the variable @xmath7 , and similarly @xmath34 are i.i.d .
introduce the real - valued function @xmath35 = e|x - x| = \int |x - x'| df_x(x'),\ ] ] where @xmath36 is the cumulative distribution function ( cdf ) of @xmath7 , and @xmath37 which is a real - valued random variable .
for simplicity we write @xmath38 and @xmath39 .
next we define the counterpart of centered distance matrices .
the centered distance function is @xmath40.\ ] ] for random variables we have @xmath41,\end{aligned}\ ] ] where @xmath42 = e[m(x , f_x ) ] = \int \int |x - x'| \
, df_x(x ' ) \ , df_x(x).\ ] ] similarly define the centered distance function @xmath43 and the random variable @xmath44 . now for @xmath45 i.i.d . , and @xmath34 i.i.d . ,
such that @xmath7 and @xmath8 have finite expectations , the population distance covariance @xmath46 is defined by @xmath47.\ ] ] we have that @xmath48 is always nonnegative , and equates zero if and only if @xmath7 and @xmath8 are independent .
it is clear by inspection that without further efforts the implementation of the sample distance covariance and the corresponding sample distance correlation requires o(@xmath0 ) steps . in this paper
we show that for real - valued random variables @xmath7 and @xmath8 , we do not need more than o(@xmath1 ) steps .
in this section , a reformulation is given in section [ sec : reformulation ] .
we then show in section [ sec : u - stat ] that the newly formed unbiased estimator is a u - statistic .
we will work with the unbiased version of the squared sample distance covariance , which is published in .
the definition is as follows .
let @xmath49 be a symmetric , real valued @xmath50 matrix with zero diagonal , @xmath51 .
define the @xmath52-centered matrix @xmath53 as follows : the @xmath54-th entry of @xmath53 is @xmath55 here the
@xmath52-centered " is so named because as shown below , the corresponding inner product ( which will be specified in ) defines an unbiased estimator of the squared distance covariance .
let @xmath56 denote a sample of observations from the joint distribution @xmath57 of random vectors @xmath7 and @xmath8 .
let @xmath49 be the euclidean distance matrix of the sample @xmath58 from the distribution of @xmath7 , and @xmath59 be the euclidean distance matrix of the sample @xmath60 from the distribution of @xmath8 .
then if @xmath61 , for @xmath62 , the following @xmath63 is an unbiased estimator of squared population distance covariance @xmath15 .
the proof of the above proposition is in the appendix of .
let @xmath64 denote the inner product defined in .
the following notations will be used . define the column and row sums as follows : @xmath65
we will need the following lemma .
[ lem : omega1 ] if @xmath64 is the inner product defined in then we have @xmath66 for the proof see the appendix .
formula will be used to prove that ( i ) the estimator in is a u - statistic and thus we can apply the relevant limit theorems to study its asymptotic behavior ; ( ii ) the estimator in can be computed in @xmath67 steps .
suppose @xmath68 is a sample . for positive integer @xmath69 ,
let @xmath70 denote all the distinct @xmath69-subsets of @xmath71 .
for a set @xmath72 , we define notation @xmath73 .
let @xmath74 be a symmetric real - valued or complex - valued kernel function of @xmath69 variables . for each @xmath75 ,
the associated u - statistic of order @xmath69 , @xmath76 , is equal to the average over ordered samples of size @xmath69 of the sample values @xmath77 . in other words ,
@xmath78 for u - statistics , we can verify the following lemma .
[ lem : check01 ] for @xmath79 , we denote @xmath80 where @xmath81 is defined in after removing the element @xmath82 .
then we must have @xmath83 in , each term @xmath77 is counted @xmath84 times on both sides .
hence the equality holds .
in fact , using arithmetic deduction , one can prove that the converse of the above is also true .
in other words , the _ jackknife invariance _ is a necessary and sufficient condition for being u - statistics . for a very similar ( equivalent ) approach
[ lem : check02 ] if there exists a positive integer @xmath85 , such that for any @xmath86 , function @xmath87 satisfies and , then there must be a kernel function @xmath88 of order @xmath69 , such that @xmath87 can be written in a form as in ; i.e. , @xmath87 is a u - statistic .
a proof of the above can be found in the appendix .
the two lemmas above show that the recursive relation is a necessary and sufficient condition for a u - statistic . for later use ,
we explicitly restate the result below .
[ th : check01 ] let @xmath89 be a statistic of a sample @xmath90 .
let @xmath91 , @xmath92 , be a statistic of a reduced sample @xmath93 ; i.e. , @xmath91 is the statistic after removing the observation @xmath82 .
the necessary and sufficient condition for @xmath89 to be a u - statistic of order @xmath69 is @xmath94 holds for all @xmath95 .
the above can be extended to a two - sample problem , in which a sample is @xmath96 for @xmath97 . by replacing @xmath82 with @xmath98 ,
all the previous arguments still hold . .
combine lemma [ lem : check01 ] and lemma [ lem : check02 ] , and simplify , we have .
let @xmath64 denote the inner product that is defined in .
note @xmath64 is based on the entire sample ( i.e. , @xmath99 . for @xmath100 ,
let @xmath101 denote the corresponding statistic after knocking out pair @xmath98 from the entire sample .
the following lemma establish counterpart for @xmath102 , where @xmath103 .
[ lem : omega2 ] for @xmath104 , let @xmath105 , @xmath106 , @xmath107 , and @xmath108 denote the corresponding sums after entry @xmath109 is removed from the sample . if @xmath102 is the inner product that is defined in after knocking off the @xmath110-th entry @xmath109 , we have @xmath111 we will not provide the proof for lemma [ lem : omega2 ] , because it will be identical with the proof of lemma [ lem : omega1 ] . [
th : u - stat ] estimator @xmath64the inner product that is defined in is a u - statistic .
the kernel function of the corresponding u - statistic is the inner product that was defined in with @xmath112 .
see a proof in the appendix .
now we know that is a u - statistic and it is easy to see that is in fact a u - statistic with a degenerate kernel under the null hypothesis of independence of @xmath7 and @xmath8 , thus we can see from corollary 4.4.2 of that if the second moments of @xmath7 and @xmath8 are finite then under the null hypothesis , the limit distribution of @xmath113 has the form @xmath114 , where @xmath115 , and @xmath116 are i.i.d .
standard normal random variables . under the alternative hypothesis
we have that @xmath117 , thus we can easily construct a consistent test of independence . for a technically much more difficult approach ,
see where a similar result was derived for a related v - statistic using deep results on complex - valued gaussian processes .
we now argue that when @xmath7 and @xmath8 are univariate , there is an o@xmath118 algorithm to implement .
we start with several intermediate results , which are presented as lemmas below .
[ lem : aidot ] denote @xmath119 for @xmath100 , we also denote @xmath120 we have @xmath121 a proof is relegated to the appendix . due to symmetry , the following is the counterpart fact for @xmath8 .
we state it without a proof .
[ lem : bidot ] denote @xmath122 for @xmath100 , we denote @xmath123 we have @xmath124 using formulas and , the following two equations can be easily established .
we state them without a proof .
[ cor : abdots ] we have @xmath125 and @xmath126 the following lemma will be used .
[ lem : aijbij ] we define a sign function , for @xmath127 , @xmath128 for any sequence @xmath129 , for @xmath100 , we define @xmath130 the following is true : @xmath131.\ ] ] we have @xmath132.\end{aligned}\ ] ] per the definition of @xmath133 , one can verify that the above equates to .
[ lem : gamma ] for any sequence @xmath129 , there is an o@xmath134 algorithm to compute for all @xmath135 ( @xmath136 ) , where @xmath137 .
again , we relegate the proof to the appendix .
the main idea of the proposed algorithm is a modification as well as an extension of the idea that was used in and , which developed a fast algorithm for computing the kendall s @xmath4 rank correlation coefficient .
the principle of the avl tree structure @xcite was adopted .
despite they are in a similar spirit , the algorithmic details are different .
we now present the main result in the following theorem .
[ th : fastalgo ] the unbiased estimator of the squared population distance covariance ( that was defined in ) can be computed by an o(@xmath1 ) algorithm . in lemma
[ lem : omega1 ] , the unbiased statistic has been rewritten as in . for the first term on the right hand side of , per lemmas [ lem : aijbij ] and
[ lem : gamma ] , there is an o(@xmath1 ) algorithm to compute it . for the second term on the right hand side of , note that quantities @xmath138 , and @xmath139 that were defined in lemmas [ lem : aidot ] and [ lem : bidot ] , respectively , are partial sums , which can be computed for all @xmath140 s with o(@xmath1 ) algorithms .
the @xmath141 factor is inserted , because one may need to sort @xmath82 s or @xmath142 s in order to compute for @xmath138 , and @xmath139 .
then by and , all @xmath143 and @xmath144 can be computed at order o(@xmath1 ) .
consequently , the second term on the right hand side of can be computed by using an o(@xmath1 ) algorithm . for the third term on the right hand side of , using and in corollary [ cor : abdots ] , we can easily see that it can be computed via an o(@xmath1 ) algorithm .
from all the above , the theorem is established .
for readers convenience , we present a detailed algorithm description in appendix , where algorithm [ alg : partialsum2d ] realizes the idea that is described in the proof of lemma [ lem : gamma ] ; algorithm [ alg : sub01 ] is a subroutine that will be called in algorithm [ alg : partialsum2d ] ; and the algorithm [ alg : fadcor ] is the algorithm that can compute for the distance covariance at o(@xmath1 ) .
in section [ sec : implement ] , we describe a matlab and c based implementation of the newly proposed fast algorithm .
this fast algorithm enables us to run some simulations with sample sizes that were impossible to experiment with before its appearance .
we report some numerical experiments in section [ sec : effectiveness ] .
distance correlation has been found helpful in feature screening . in section [ sec : screening ] , we redo experiments on this regard , increasing the sample size from @xmath145 to @xmath146 . it is observed that the advantage of using the distance correlation is more evident when the sample size becomes larger .
the fast algorithm was implemented in matlab , with a key step ( of dyadic updating ) being implemented in c. it was then compared against the direct ( i.e. , slow ) implementation .
table [ tab : fadcor01 ] presents the average running time for the two different implementations in matlab with @xmath147 replications at each sample size .
the sample size goes from @xmath148 ( @xmath149 ) to @xmath150 ( @xmath151 ) . in all these cases , the two methods ended with identical solutions ; this validates our fast algorithm .
note a comparison in matlab is not desirable for our fast algorithm .
the direct method calls some matlab functions , which achieve the speed of a low - level language implementation , while the implementation of the fast method is not . in theory
, the fast algorithm will compare more favorably if both methods are implemented in a low - level language , such as in c or c++ .
.running times ( in seconds ) for the direct method and the fast method for computing the distance correlations .
the values in the parentheses are sample standard errors . at each sample size ,
@xmath147 repetitions were run.[tab : fadcor01 ] [ cols=">,>,>",options="header " , ] the fast method allows us to study how the sample distance correlation converge to the population counterpart as a function of the sample size .
[ fig : converge01 ] shows the convergence of the sample distance correlation and pearson s correlation .
it is worth noting that in cases ( 3)-(8 ) , the pearson s correlation quickly converges to zero , while the sample distance correlation clearly stays away from zero .
this experiments shows that a previous observation in fig .
[ fig : showcases02 ] should occur with large probability . covering interval of both sample pearson s correlation ( solid line , with low and upper sample quartiles marked by ` @xmath152 ' ) and sample distance correlation ( dotted lines , with both quartiles marked by ` @xmath153 ' ) .
the horizontal axis equates the @xmath154sample size@xmath155 .
the vertical axis corresponds to the values of correlations . in cases ( 3 ) through ( 8) , the two correlations clearly converge to different constants , when the pearson s correlation always seems to converge to zero . ] in * li2012 , distance correlation has been proposed to facilitate feature screening in ultrahigh - dimensional data analysis .
the proposed sure independence screening procedure based on the distance correlation ( dc - sis ) has been proven to be effective in their simulation study . due to the use of the direct method , they restricted their sample size to @xmath145 .
we redo the simulations as in * li2012 , however increases the sample size to @xmath146 , i.e. , @xmath5 times of the originally attempted .
it is observed that the use of distance correlation becomes more advantageous when the sample size increases
. the screening algorithm , which was initially advocated by , works as follows . for each covariate @xmath156 , a ` marginal utility ' function was computed .
such a marginal utility function can be the pearson s correlation , the distance correlation that was discussed in this paper , or other dependence measure such as the one in * jasa2011sir that was also used in the simulation studies of * li2012 .
the ` screening ' is based on the magnitude of the values of these marginal utility function . sometimes , forward ,
backward , or a hybrid stepwise approach is proposed . in this paper , we refrain from further discussion in this potential research direction .
our simulation setup follows the one in .
note that an alternative approach named _ sure independent ranking and screening _ ( sirs ) @xcite was compared against . for a sample , @xmath157 , of two random variable @xmath7 and @xmath8 , the sirs dependence measure ( i.e. , the marginal utility function )
is defined as @xmath158 ^ 2,\ ] ] where @xmath159 is an indicator function . the formulation in the above definition seemingly hint an o@xmath160 algorithm .
the following theorem will show that it can be computed via an o@xmath134 algorithm .
the proof and the algorithmic details are relegated to the appendix .
[ th : sirs ] for a sample , @xmath157 , of a bivariate random vector @xmath161 , the sirs measure @xcite in can be computed via an algorithm whose average complexity is o@xmath134 . for completeness , we state our simulation setup below .
we generate @xmath162 from normal distribution with zero mean and covariance matrix @xmath163 , and the error term @xmath164 from the standard normal distribution @xmath165 .
two covariance matrices are considered to assess the performance of the dc - sis and to compare with existing methods : ( 1 ) @xmath166 and ( 2 ) @xmath167 .
note that a covariance matrix with entries @xmath168 enjoys a known cholesky decomposition : @xmath169 , where @xmath170 if @xmath171 , and @xmath172 , @xmath173 , for @xmath174 and @xmath175 , @xmath176 . in our simulations , we take advantage of this known decomposition .
the dimension @xmath177 varies from @xmath178 to @xmath179 .
each experiment was repeated @xmath180 times , and the performance is evaluated through the following three criteria : 1 .
@xmath181 : the minimum model size to include all active predictors .
we report the @xmath182 , and @xmath183 quantiles of @xmath181 out of @xmath180 replications .
@xmath184 : the proportion that an individual active predictor is selected for a given model size @xmath185 in the @xmath180 replications .
@xmath186 : the proportion that all active predictors are selected for a given model size @xmath185 in the @xmath180 replications .
the @xmath181 is used to measure the model complexity of the resulting model of an underlying screening procedure .
the closer to the minimum model size the @xmath181 is , the better the screening procedure is .
the sure screening property ensures that @xmath184 and @xmath186 are both close to one when the estimated model size @xmath185 is sufficiently large .
different from , the @xmath185 is chosen to be @xmath187 $ ] , @xmath188 , and @xmath189 throughout our simulations to empirically examine the effect of the cutoff , where @xmath190 $ ] denotes the integer part of @xmath191 .
an innovative stopping rule is introduced in for dc - sis .
we did not implement it here , because the new stopping rule requires a multivariate version of the distance correlation , which is not covered by this paper .
the example is designed to compare the finite sample performance of the dc - sis with the sis @xcite and the sirs @xcite . in this example , we generate the response from the following four models : * @xmath192 , * @xmath193 , * @xmath194 , * @xmath195 , where @xmath196 is an indicator function .
the regression functions @xmath197 in models ( 1.a)-(1.d ) are all nonlinear in @xmath198 .
in addition , models ( 1.b ) and ( 1.c ) contain an interaction term @xmath199 , and
model ( 1.d ) is heteroscedastic . following fan and lv ( 2008 ) ,
we choose @xmath200 for @xmath201 , and @xmath202 , where @xmath203 , @xmath204bernoulli@xmath205 and @xmath206 .
we set @xmath207 in this example to be consistent with the experiments in : challenging the feature screening procedures under consideration . for each independence screening procedure
, we compute the associated marginal utility between each predictor @xmath208 and the response @xmath8 .
that is , we regard @xmath209 as the predictor vector in this example .
tables [ tab : ex1a ] , [ tab : ex1b ] , and [ tab : ex1b-2 ] present the simulation results for @xmath210 , and @xmath186 .
the performances of the dc - sis , sis , and sirs are quite similar in model ( 1.a ) , indicating that the sis has a robust performance if the working linear model does not deviate far from the underlying true model .
the dc - sis outperforms the sis and the sirs significantly in models ( 1.b)-(1.d ) .
both the sis and the sirs have little chance to identify the important predictors @xmath211 and @xmath212 in models ( 1.b ) and ( 1.c ) , and @xmath213 in model ( 1.d ) . comparing tab.s [ tab : ex1a ] and [ tab : ex1b ] with the counterparts in
, one can clearly see that the advantage of using the distance correlation becomes more evident , observing smaller sample quantiles of @xmath181 for dc - sis , and larger coverage probabilities in @xmath184 and @xmath186 .
distance correlation has been found useful in many applications @xcite .
a direct implementation of the distance correlation led to an @xmath214 algorithm with sample size @xmath2 . we propose a fast algorithm .
its computational complexity is @xmath6 on average .
armed with this fast algorithm , we carry out some numerical experiments with sample sizes that have not been attempted before .
we found that in many cases , the advantage of adopting the distance correlation becomes even more evident .
the proposed fast algorithm certainly makes the distance correlation more applicable in situations where statistical dependence needs to be evaluated .
algorithm [ alg : fadcor ] is the algorithm that can compute for the distance covariance at o(@xmath1 ) .
algorithm [ alg : partialsum2d ] realizes the idea that is described in the proof of lemma [ lem : gamma ] .
algorithm [ alg : sub01 ] is a subroutine that will be called in algorithm [ alg : partialsum2d ] .
[ alg : fadcor ] algorithm : fast computing for distance covariance ( fadcor ) * inputs : * observations @xmath90 , and @xmath215 . +
* outputs : * the distance covariance that was defined in . 1 . sort @xmath90 , and @xmath215 .
let @xmath216 and @xmath217 denote the order indices ; i.e. , if for @xmath218 , @xmath219 , then @xmath82 is the @xmath110th smallest observations among @xmath90 . + similarly if for @xmath218 , @xmath220 , then @xmath142 is the @xmath110th smallest observations among @xmath215 .
2 . let @xmath221 , and @xmath222 denote the order statistics .
+ denote the partial sums : @xmath223 they can be computed using the following recursive relation : @xmath224 , @xmath225 3 .
compute @xmath226 , @xmath227 , @xmath228 , and @xmath139 that are defined in lemma [ lem : aidot ] and [ lem : bidot ] , using the following formula : for @xmath229 , we have @xmath230 4 .
compute @xmath231 and @xmath232 per their definitions in lemma [ lem : aidot ] and [ lem : bidot ] .
5 . using and ,
compute @xmath233 .
[ alg:1.part2 ] 6 . using and ,
compute @xmath234 and @xmath235 .
[ alg:1.part3 ] 7 .
_ partialsum2d _ to compute for @xmath236 , @xmath237 , @xmath238 , and @xmath239 . 8 . using to compute @xmath240 .
[ alg:1.part1 ] 9 .
finally , apply the results of steps [ alg:1.part2 ] .
, [ alg:1.part3 ] . , and [ alg:1.part1 ] .
[ alg : partialsum2d ] algorithm : fast algorithm for a @xmath241-d partial sum sequence ( partialsum2d ) * inputs : * observations @xmath90 , @xmath215 , and @xmath242 . + * outputs : * quantity @xmath243 that is defined in lemma [ lem : aijbij ] . 1 .
compute for the order statistics @xmath244 for @xmath90 . then rearrange triplets @xmath245 s such that we have @xmath246 .
each triplet @xmath245 ( @xmath100 ) stay unchanged .
2 . let @xmath247 denote the order statistics for @xmath215 , and assume that @xmath248 , are the order indices ; i.e. , if @xmath220 , then @xmath142 is the @xmath110-th smallest among @xmath215 . without loss of generality
, we may assume that @xmath249 .
evidently aforementioned function @xmath250 is invertible .
let @xmath251 denote its inverse .
define the partial sum sequence : for @xmath100 , @xmath252 the following recursive relation enables an o(@xmath2 ) algorithm to compute for all @xmath253 s , @xmath254 4 . for @xmath100 ,
define @xmath255 again the above partial sums can be computed in o(@xmath2 ) steps .
compute @xmath256 6 . call
subroutine _ dyadupdate _ to compute for @xmath257 for all @xmath258
by , we have that @xmath259 [ alg : sub01 ] subroutine : a dyadic updating scheme ( dyadupdate ) * inputs : * sequence @xmath215 and @xmath242 , where @xmath215 is a permutation of @xmath260 .
+ * outputs : * quantities @xmath261 , @xmath92
. 1 . recall that we have assumed @xmath262 .
if @xmath2 is not dyadic , we simply choose the smallest @xmath263 such that @xmath264 . recall that for @xmath265 , @xmath266 , we define a close interval @xmath267.\ ] ] 2 .
assign @xmath268 , and @xmath269 .
3 . for @xmath270 , we do the following . 1 .
fall all @xmath271 s , such that @xmath272 .
then for these @xmath271 s , do update @xmath273 2 .
find nonnegative integers @xmath274 such that @xmath275 let @xmath276 .
for @xmath277 , compute @xmath278 3 .
one can verify the following equalities : @xmath280 the following will be used in our simplification too .
we have @xmath281 similarly , we have @xmath282 in the following , we simplify the statistic in .
we have @xmath283.\end{aligned}\ ] ] furthermore , we have @xmath284 \\ & & + \frac{1}{n(n-2)^2 ( n-3 ) } \sum_{i \neq j } ( a_{i\cdot } + a_{\cdot j})(b_{i\cdot } + b_{\cdot j})\\ & & -\frac{a_{\cdot \cdot } b_{\cdot \cdot}}{(n-1)(n-2)^2(n-3 ) } \\ & \stackrel{\mbox{\eqref{eq : check04},\eqref{eq : check06}}}{= } & \frac{1}{n(n-3 ) } \sum_{i \neq j } a_{ij}b_{ij } -\frac{4}{n(n-2)(n-3 ) } \sum_{i=1}^n a_{i\cdot}b_{i\cdot } \\ & & + \frac{1}{n(n-2)^2 ( n-3 ) } \sum_{i \neq j } ( a_{i\cdot } + a_{\cdot j})(b_{i\cdot } + b_{\cdot j } ) -\frac{a_{\cdot \cdot } b_{\cdot \cdot}}{(n-1)(n-2)^2(n-3)}.\end{aligned}\ ] ] now bringing in and , we have @xmath285 \\ & = & \frac{1}{n(n-3 ) } \sum_{i
\neq j } a_{ij}b_{ij } -\frac{2}{n(n-2)(n-3 ) } \sum_{i=1}^n a_{i\cdot}b_{i\cdot } + \frac{a_{\cdot \cdot } b_{\cdot \cdot}}{n(n-1)(n-2)(n-3)},\end{aligned}\ ] ] which is .
we use arithmetic induction .
suppose @xmath286 , becomes @xmath287 by defining @xmath288 , we can verify that @xmath88 is a kernel function with @xmath69 variables .
consequently , @xmath289 can be written as .
now suppose for any @xmath290 , @xmath87 has the form as in , with the function @xmath88 that was defined above .
applying with @xmath291 , we can show that @xmath292 still has the form as in , with the same function @xmath88 that was defined above .
we omit further details .
it is evident to verify that the followings are true : for @xmath293 , @xmath294 we then have @xmath295 for the right hand side of , we have the following : @xmath296}{(n-1)(n-3)(n-4 ) } \\ & & + \frac{(n-4)a_{\cdot \cdot } b_{\cdot \cdot } + 4\sum^n_{k=1 } a_{k\cdot } b_{k \cdot}}{(n-1)(n-2)(n-3)(n-4 ) } \\ & = & \frac { \sum_{i \neq j } a_{ij}b_{ij}}{n-3 } - \frac{2}{(n-2)(n-3 ) } \sum^n_{i=1 } a_{i\cdot } b_{i \cdot } + \frac{a_{\cdot \cdot } b_{\cdot \cdot}}{(n-1)(n-2)(n-3)}.\end{aligned}\ ] ] compare with , we can verify that the above equates to @xmath297 , which ( per theorem [ th : check01 ] ) indicates that @xmath64 is a u - statistic .
the kernel function of the corresponding u - statistic is the inner product that was defined in with @xmath112 . without loss of generality ( wlog ) , we assume that @xmath302 .
we have @xmath303 note that we can verify the following equations : @xmath304 where @xmath305 .
we can rewrite @xmath135 as follows : @xmath306 we will argue that the three summations on the right hand side can be implemented by o@xmath134 algorithms . first , term @xmath307 is a formula for partial sums .
it is known that an o(@xmath2 ) algorithm exists , by utilizing the relation : @xmath308 second , after sorting @xmath309 s at an increasing order , sums @xmath310 is transferred into a partial sums sequence .
hence it can be implemented via an o(@xmath2 ) algorithm .
if quicksort @xcite ( * ? ? ?
* section 5.2.2 : sorting by exchanging ( pages 113 - 122 ) ) is adopted , the sorting of @xmath309 s can be done via an o(@xmath1 ) algorithm .
we will argue that sums @xmath311 can be computed in an o(@xmath1 ) algorithm .
wlog , we assume that @xmath312 , is a permutation of the set @xmath313 .
wlog , we assume that @xmath2 is dyadic ; i.e. , @xmath262 , where @xmath314 or @xmath263 is a nonnegative integer . for @xmath265 , @xmath266
, we define an close interval @xmath267.\ ] ] we then define the following function @xmath315 where @xmath229 , @xmath265 , and @xmath266 .
we argue that computing the values of @xmath316 for all @xmath317 , can be done in o(@xmath1 ) .
first of all , it is evident that for all @xmath318 , we have @xmath319 suppose for all @xmath320 , @xmath321 s have been computed for all @xmath322 and @xmath110 . for each @xmath323 , there is only one @xmath324 , such that @xmath325 $ ] . by the definition of @xmath326
, we have @xmath327 the above dynamic programming style updating scheme needs to be run for @xmath2 times ( i.e. , for all @xmath328 ) , however each stage requires no more than @xmath329 updates .
overall , the computing for all @xmath316 takes no more than o(@xmath1 ) . for a fixed @xmath140 , @xmath100 ,
we now consider how to compute for @xmath330 .
if @xmath331 , obviously we have @xmath332 . for @xmath333 , there must be a unique sequence of positive integers @xmath334 , such that @xmath335 since @xmath336 , we must have @xmath337 .
we then define @xmath338 as follows @xmath339 one can then verify the following : for @xmath340 , @xmath341 since @xmath337 , the above takes no more than o(@xmath141 ) numerical operations .
consequently , computing @xmath330 for all @xmath258 , can be done in o(@xmath1 ) .
( we realized that the above approach utilized the avl tree structure @xcite . ) from all the above , we established the result .
we have the following sequence of equations : @xmath342 ^ 2 & = & \sum_{j=1}^n \left[\sum_{i=1}^n x_i
\mathbf{1}(y_i < y_j ) \right ] \cdot \left[\sum_{k=1}^n x_k \mathbf{1}(y_k < y_j ) \right ] \\ & = & \sum_{j=1}^n \sum_{i=1}^n \sum_{k=1}^n x_i \cdot x_k \cdot \mathbf{1}(y_i < y_j \mbox { and } y_k < y_j)\\ & = & \sum_{i=1}^n x_i \left [ \sum_{k : y_i \le y_k } x_k \sum_{j=1}^n \mathbf{1}(y_k
< y_j ) + \sum_{k : y_i > y_k } x_k \sum_{j=1}^n \mathbf{1}(y_i < y_j ) \right ] .\end{aligned}\ ] ] the last expression implies the following steps to compute for sirs@xmath161 . since @xmath348 s , @xmath349 s , and @xmath350 s are partial sums , it is easy to verify that each of the above steps can be done within o@xmath134 operations on average , hence the entire algorithm takes o@xmath134 operations on average .
adelson - velskii , g. , and landis , e. m. 1962 , `` an algorithm for the organization of information , '' _ proceedings of the ussr academy of sciences _ , 146 , 263266 .
( russian ) english translation by myron j. ricci in soviet math .
doklady , 3:1259 - 1263 , 1962 .
kong , j. , klein , b. e. k. , klein , r. , and wahba , g. 2012 , `` using distance correlation and ss - anova to acess associations of familial relationnships , lifestyle factors , diseases , and mortality , '' _ proceedings of the national academy of sceineces _ , 109(50 ) , 2035220357 .
koroljuk , v. s. , and borovskich , y. v. 1994 , _ theory of u - statistics _ , vol .
273 of _ mathematics and its applications _ , dordrecht : kluwer academic publishers group .
translated by p. v. malyshev , d.v .
malyshev from the 1989 russian original edition . | distance covariance and distance correlation have been widely adopted in measuring dependence of a pair of random variables or random vectors .
if the computation of distance covariance and distance correlation is implemented directly accordingly to its definition then its computational complexity is o(@xmath0 ) which is a disadvantage compared to other faster methods . in this paper
we show that the computation of distance covariance and distance correlation of real valued random variables can be implemented by an o(@xmath1 ) algorithm and this is comparable to other computationally efficient algorithms .
the new formula we derive for an unbiased estimator for squared distance covariance turns out to be a u - statistic .
this fact implies some nice asymptotic properties that were derived before via more complex methods .
we apply the fast computing algorithm to some synthetic data .
our work will make distance correlation applicable to a much wider class of applications .
* author s footnote : * dr .
xiaoming huo is a professor in the school of industrial and systems engineering at the georgia institute of technology .
he has been serving as a rotator at the national science foundation since august 2013 .
mailing address : 765 ferst dr , atlanta , ga 30332 ( email : [email protected] ) .
dr .
gbor j. szkely is a program officer at the national science foundation .
mailing address : 4201 wilson blvd , arlington , va 22203 ( email : [email protected] ) .
keywords : distance correlation , fast algorithm , statistical dependence |
one of basic problems in knot theory is to distinguish knots in @xmath6 from each other using knot invariants .
there are several knot invariants each being powerful enough to determine if a given knot in @xmath6 is the unknot , such as the knot floer homology @xcite , the a - polynomial @xcite @xcite , and the khovanov homology @xcite . in other words ,
these invariants are each an unknot - detector .
it is also known that the knot floer homology can detect the trefoil knot and the figure 8 knot @xcite . in this paper
we first consider the problem of detecting the set of torus knots @xmath5 in @xmath6 using knot invariants .
to reach this goal either the knot floer homology or the @xmath1-polynomial alone is not enough ; for instances the torus knot @xmath7 has the same knot floer homology as the @xmath8-cable over @xmath9 @xcite , and the torus knot @xmath10 has the same @xmath1-polynomial as the torus knot @xmath11 .
however when the two invariants are combined together , the job can be done .
we have [ main1]if a knot @xmath12 in @xmath6 has the same knot floer homology and the same @xmath1-polynomial as a torus knot @xmath5 , then @xmath13 .
we then go further to find out a one - parameter family of mutually distinct hyperbolic knots @xmath14 in @xmath6 , where @xmath15 is integer valued , each of which is detected by the combination of its @xmath1-polynomial and its knot floer homology .
a knot diagram for @xmath14 is illustrated in figure [ fig ( l,-1 ) ] .
note that @xmath16 is the @xmath17-pretzel knot .
also note that @xmath14 is a subfamily of the hyperbolic knots @xmath18 ( with some forbidden values on integers @xmath19 ) given in @xcite each of which admits a half - integral toroidal surgery ( the slope formula is given in @xcite , and recalled in section [ topological properties of j ] of this paper ) and by @xcite these hyperbolic knots @xmath18 are the only hyperbolic knots in @xmath6 which admit non - integral toroidal surgeries .
[ main2]the family of knots @xmath20 are mutually distinct hyperbolic knots in @xmath6 .
let @xmath21 be any fixed @xmath14 , @xmath15 . if a knot @xmath12 in @xmath6 has the same knot floer homology and the same @xmath1-polynomial as @xmath21 , then @xmath22 .
the key input from the knot floer homology in proving theorems [ main1 ] and [ main2 ] is that @xmath12 is fibred @xcite ( since @xmath5 or @xmath21 is ) and has the same alexander polynomial as @xmath5 or @xmath21 @xcite . with these conditions ( conclusions ) in place , we can then identify the knot using the a - polynomial and the alexander polynomial ( in particular the genus of the knot ) .
the proof of theorems [ main1 ] will be given in section [ proof1 ] after we prepare some general properties on @xmath1-polynomials in section [ apoly ] , where we also derive a cabling formula for @xmath1-polynomials of cabled knots in @xmath6 ( theorem [ cabling formula ] ) and in particular we give explicitly the a - polynomials of iterated torus knots ( corollary [ a - poly of iterated torus ] ) . the argument for theorem [ main2 ] is more involved than that for theorem [ main1 ] , for which we need to make some more preparations ( besides those made in section [ apoly ] ) in the next three sections . in section [ topological properties of j ]
we collect some topological properties about the family of knots @xmath18 , in particular we give a complete genus formula for @xmath18 and show that @xmath14 is a class of small knots in @xmath6 . in section [ a - poly of j ] we collect some info about the @xmath1-polynomials of the knots @xmath18 without knowing the explicit formulas of the @xmath1-polynomials , and with such info we are able to show that if a hyperbolic knot @xmath12 has the same @xmath1-polynomial as a given @xmath23 , then @xmath12 has the same half - integral toroidal surgery slope as @xmath21 and @xmath12 is one of @xmath24 with @xmath25 being divisible by @xmath26 .
we then in section [ distinguish j from k(l , m,0,p ) ] identify each @xmath23 among @xmath24 with @xmath27 using the genus formula and the half - integral toroidal slope formula for @xmath24 .
results obtained in these three sections , together with some results from section [ apoly ] , are applied in section [ proof2 ] to complete the proof of theorem [ main2 ] .
note that the @xmath1-polynomial @xmath28 ( for a knot @xmath12 in @xmath6 ) used in this paper is a slightly modified version of the original @xmath1-polynomial given in @xcite .
the only difference is that in the current version , the @xmath1-polynomial of the unknot is @xmath29 and @xmath30 may possibly occur as a factor in @xmath28 for certain knots @xmath12 contributed by some component of the character variety of the knot exterior containing characters of irreducible representations ; while in the original version @xmath30 is a factor of the @xmath1-polynomial for every knot contributed by the unique component of the character variety of the knot exterior consisting of characters of reducible representations ( see section [ apoly ] for details ) .
the current version contains a bit more information than the original one . * acknowledgements.*the first author was partially supported by nsf grant numbers dms-1103976 , dms-1252992 , and an alfred p. sloan research fellowship .
first we need to recall some background material on @xmath1-polynomials and set up some notations . for a finitely
generated group @xmath31 , @xmath32 denotes the set of representations ( i.e. group homomorphisms ) from @xmath31 to @xmath33 . for each representation @xmath34 ,
its character @xmath35 is the complex valued function @xmath36 defined by @xmath37 for @xmath38 .
let @xmath39 be the set of characters of representations in @xmath32 and @xmath40 the map sending @xmath41 to @xmath35 .
then both @xmath32 and @xmath39 are complex affine algebraic sets such that @xmath42 is a regular map ( see @xcite for details ) . for an element @xmath38 , the function @xmath43 is defined by @xmath44 for each @xmath45 .
each @xmath46 is a regular function on @xmath39 .
obviously @xmath45 is a zero point of @xmath46 if and only if either @xmath47 or @xmath48 is a parabolic element .
it is also evident that @xmath46 is invariant when @xmath49 is replaced by a conjugate of @xmath49 or the inverse of @xmath49 .
note that if @xmath50 is a group homomorphism between two finitely generated groups , it naturally induces a regular map @xmath51 by @xmath52 and a regular map @xmath53 by @xmath54 .
note that if @xmath55 is an irreducible subvariety of @xmath56 , then the zariski closure of @xmath57 in @xmath39 is also irreducible .
if in addition the homomorphism @xmath58 is surjective , each of the regular maps @xmath59 and @xmath60 is an embedding , in which case we may simply consider @xmath61 and @xmath56 as subsets of @xmath32 and @xmath39 respectively , and write @xmath62 and @xmath63 . for a compact manifold @xmath64
, we use @xmath65 and @xmath66 to denote @xmath67 and @xmath68 respectively .
the @xmath1-polynomial was introduced in @xcite .
we slightly modify its original definition for a knot @xmath12 in @xmath6 as follows .
let @xmath69 be the exterior of @xmath12 in @xmath6 and let @xmath70 be the standard meridian - longitude basis for @xmath71 .
let @xmath72 be the regular map induced by the inclusion induced homomorphism @xmath73 , and let @xmath74 be the set of diagonal representations of @xmath71 , i.e. @xmath75 then @xmath74 is a subvariety of @xmath76 and @xmath77 is a degree @xmath78 , surjective , regular map .
we may identify @xmath74 with @xmath79 through the eigenvalue map @xmath80 which sends @xmath81 to @xmath82 if @xmath83 and @xmath84 .
for every knot in @xmath6 , there is a unique component in @xmath85 consisting of characters of reducible representations , which we call the trivial component of @xmath85 ( the trivial component is of dimensional one ) .
now let @xmath86 be the set of nontrivial components of @xmath85 each of which has a @xmath29-dimensional image in @xmath87 under the map @xmath88 .
the set @xmath86 is possibly empty , and in fact with the current knowledge it is known that @xmath86 is empty if and only if @xmath12 is the unknot .
so when @xmath12 is a nontrivial knot , @xmath89 is @xmath29-dimensional in @xmath74 and in turn @xmath90 is @xmath29-dimensional in @xmath91 .
let @xmath92 be the zariski closure of @xmath90 in @xmath93 .
then @xmath92 is a plane curve in @xmath93 defined over @xmath94 .
let @xmath95 be the defining polynomial of @xmath92 normalized so that @xmath96 $ ] with no repeated factors and with @xmath29 as the greatest common divisor of its coefficients
. then @xmath28 is uniquely defined up to sign . for the unknot
we define its @xmath1-polynomial to be @xmath29 .
as remarked in the introduction section , @xmath30 might occur as a factor of @xmath28 for certain knots . also by @xcite @xcite ,
@xmath97 if and only if @xmath12 is the unknot ( in fact for every nontrivial knot @xmath12 , @xmath28 contains a nontrivial factor which is not @xmath30 ) .
we note that from the constructional definition of the @xmath1-polynomial we see that each component @xmath55 of @xmath86 contributes a factor @xmath98 in @xmath28 , i.e. @xmath98 is the defining polynomial of the plane curve @xmath99 which is the zariski closure of @xmath100 in @xmath93 , and moreover @xmath98 is balanced , i.e. if @xmath101 is a generic zero point of @xmath98 then @xmath102 is also a zero point of @xmath98 .
also note that @xmath98 is not necessarily irreducible over @xmath103 but contains at most two irreducible factors over @xmath103 .
we shall call such @xmath98 a balanced - irreducible factor of @xmath28 .
obviously @xmath28 is a product of balanced - irreducible factors and the product decomposition is unique up to the ordering of the factors .
we now define a couple of functions which will be convenient to use in expressing the @xmath1-polynomials for torus knots and later on for cabled knots and iterated torus knots .
let @xmath104 be a pair of relative prime integers with @xmath105 .
define @xmath106 $ ] to be the polynomial determined by the pair @xmath107 as follows : @xmath108 and define @xmath109 $ ] to be the polynomial determined by the pair @xmath107 as follows : @xmath110 note that the ring @xmath111 $ ] is a unique factorization domain .
the following lemma can be easily checked .
[ lem : irrfac ] among the polynomials in ( [ polynomial of ( p , q ) ] ) and ( [ g polynomial of ( p , q ) ] ) , the first two in ( [ polynomial of ( p , q ) ] ) and the two in ( [ g polynomial of ( p , q ) ] ) are irreducible over @xmath103 , and the last two in ( [ polynomial of ( p , q ) ] ) can be factorized as the product of two irreducible polynomials over @xmath103 : @xmath112 the set of nontrivial torus knots @xmath5 are naturally indexed by pairs @xmath107 satisfying @xmath113 , @xmath114 . note that @xmath115 is the mirror image of @xmath5 .
the @xmath1-polynomial of a torus knot @xmath5 is given by ( e.g. ( * ? ? ?
* example 4.1 ) ) : @xmath116 in particular the @xmath1-polynomial distinguishes @xmath5 from @xmath117 . for the exterior @xmath69 of a nontrivial knot in @xmath6
, @xmath118 can be considered as a subgroup of @xmath119 which is well defined up to conjugation . in particular , the function @xmath120 on @xmath85 is well defined for each class @xmath121 . as
@xmath120 is also invariant under the change of the orientation of @xmath122 , @xmath120 is also well defined when @xmath122 is a slope in @xmath123 .
later on for convenience we will often not make a distinction among a primitive class of @xmath124 , the corresponding element of @xmath71 and the corresponding slope in @xmath123 , so long as it is well defined .
it is known ( e.g. @xcite ) that any irreducible curve @xmath55 in @xmath85 belongs to one of the following three mutually exclusive types : ( a ) for each slope @xmath122 in @xmath123 , the function @xmath120 is non - constant on @xmath55 ; ( b ) there is a unique slope @xmath125 in @xmath123 such that the function @xmath126 is constant on @xmath55 ; ( c ) for each slope @xmath122 in @xmath123 , the function @xmath120 is constant on @xmath55 .
obviously a curve of type ( a ) or ( b ) has one dimensional image in @xmath87 under the map @xmath127 .
note that the trivial component of @xmath85 is of type ( b ) .
hence a curve of type ( a ) is contained in @xmath86 and so is a curve of type ( b ) if it is not the trivial component of @xmath85 .
an irreducible curve in case ( a ) is named a _ norm curve_. indeed as the name indicates , a norm curve in @xmath85 can be used to define a norm , known as culler - shalen norm , on the real @xmath78-dimensional plane @xmath128 satisfying certain properties .
such curve exists when @xmath69 is hyperbolic , namely any component of @xmath85 which contains the character of a discrete faithful representation of @xmath119 is a norm curve .
for an irreducible curve @xmath55 in @xmath85 , let @xmath129 be the smooth projective completion of @xmath55 and let @xmath130 be the birational isomorphism .
the map @xmath58 is onto and is defined at all but finitely many points of @xmath129 .
the points of @xmath129 where @xmath58 is not defined are called ideal points and all other points of @xmath129 are called regular points .
the map @xmath58 induces an isomorphism from the function field of @xmath55 to that of @xmath129 .
in particular every regular function @xmath46 on @xmath55 corresponds uniquely to its extension @xmath131 on @xmath129 which is a rational function .
if @xmath131 is not a constant function on @xmath129 , its degree , denoted @xmath132 , is equal to the number of zeros of @xmath131 in @xmath129 counted with multiplicity , i.e. @xmath133 where @xmath134 is the zero degree of @xmath131 at point @xmath135
. we shall identify @xmath128 with the real @xmath136-plane so that @xmath137 are integer lattice points with @xmath138 being the meridian class and @xmath139 the longitude class .
so each slope @xmath140 corresponds to the pair of primitive elements @xmath141 .
[ properties of a norm curve ] let @xmath55 be a norm curve of @xmath85
. then the associated culler - shalen norm @xmath142 on @xmath143 has the following properties : let @xmath144 and let @xmath145 be the disk in @xmath143 with radius @xmath146 with respect to the norm @xmath142 , then \(1 ) for each nontrivial element @xmath147 , @xmath148 and thus @xmath149 .
( 2 ) the disk @xmath145 is a convex finite sided polygon symmetric to the origin whose interior does not contain any non - zero element of @xmath124 and whose boundary contains at least one but at most four nonzero classes of @xmath124 up to sign .
( 3 ) if @xmath150 is a vertex of @xmath145 , then there is a boundary slope @xmath140 of @xmath123 such that @xmath151 lie in the line passing through @xmath150 and @xmath152 .
( that is , @xmath153 is a boundary slope of @xmath123 for any vertex @xmath150 of @xmath145 ) .
( 4 ) if a primitive class @xmath147 is not a boundary class and @xmath154 has no noncyclic representations , then @xmath155 lies in @xmath156 ( i.e. @xmath157 ) and is not a vertex of @xmath145 .
( 5 ) if the meridian class @xmath158 is not a boundary class , then for any non - integral class @xmath155 if it is not a vertex of @xmath145 then it does not lie in @xmath156 and thus @xmath159 .
theorem [ properties of a norm curve ] is originated from ( * ? ? ?
* chatper 1 ) although it was assumed there that the curve @xmath55 contains the character of a discrete faithful representation of @xmath69 .
the version given here is contained in @xcite . recall that if @xmath160 $ ] is a two variable polynomial in @xmath161 and @xmath162 with complex coefficients , the newton polygon @xmath163 of @xmath98 is defined to be the convex hull in the real @xmath136-plane of the set of points @xmath164 the following theorem is proved in @xcite .
[ dual polygons ] let @xmath55 be a norm curve of @xmath85 and let @xmath98 be the balanced - irreducible factor of @xmath28 contributed by @xmath55 . then the norm polygon @xmath145 determined by @xmath55 is dual to the newton polygon @xmath163 of @xmath98 in the following way : the set of slopes of vertices of @xmath145 is equal to the set of slopes of edges of @xmath163 . in fact
@xmath145 and @xmath163 mutually determine each other up a positive integer multiple .
we remark that although in @xcite there was some additional condition imposed on @xmath55 and the version of the @xmath1-polynomial defined in @xcite is mildly different from the one as given here , the above theorem remains valid with identical reasoning as given in @xcite .
we only need to describe the exact relation between @xmath145 and @xmath163 as follows , to see how they determine each other up to an integer multiple .
let @xmath165 be the zariski closure of the restriction @xmath166 of @xmath55 in @xmath87 and let @xmath167 be the degree of the map @xmath168 . as explained in @xcite ( originated from @xcite ) , the newton polygon @xmath163 determines a width function @xmath169 on the set of slopes given by @xmath170 if @xmath171 is the number of lines in the @xmath136-plane of the slope @xmath172 which contain points of both @xmath173 and @xmath163 .
the width function in turn defines a norm @xmath174 on the @xmath136-plane @xmath143 such that @xmath175 for each primitive class @xmath176 .
finally @xmath177 [ non - hyperbolic ]
if every balanced - irreducible factor of @xmath28 over @xmath103 has two monomials , then @xmath12 is not a hyperbolic knot .
the condition of the corollary means that the newton polygon of each balanced - irreducible factor of @xmath28 consists of a single edge . on the other hand for a hyperbolic knot
, its character variety contains a norm curve component @xmath55 which contributes a balanced - irreducible factor @xmath98 to the @xmath1-polynomial such that the newton polygon of @xmath98 has at least two edges of different slopes . an irreducible curve in @xmath86 of type ( b ) ( such curve exists only for certain knots )
is named a _ semi - norm curve _ as suggested by the following theorem which is contained in @xcite .
[ properties of a semi - norm curve ] suppose that @xmath178 is an irreducible curve of type ( b ) with @xmath125 being the unique slope such that @xmath126 is constant on @xmath55 .
then a semi - norm @xmath142 can be defined on @xmath143 , with the following properties : ( 1 ) for each slope @xmath179 , @xmath148 .
( 2 ) for the unique slope @xmath125 associated to @xmath55 , @xmath180 and @xmath125 is a boundary slope of @xmath69 .
( 3 ) if @xmath122 is a primitive class and is not a boundary class and @xmath154 has no non - cyclic representation , then @xmath181 . (
4 ) let @xmath182 .
then for any slope @xmath122 , @xmath183 .
note that for each torus knot @xmath5 , every nontrivial component in its character variety is a semi - norm curve with @xmath184 as the associated slope .
if @xmath12 is a small knot , i.e. if its exterior @xmath69 does not contain any closed essential surface , then every nontrivial component of @xmath85 is either a norm curve or a semi - norm curve .
we now proceed to get some properties on @xmath1-polynomials of satellite knots in @xmath6 .
recall that a knot @xmath12 in @xmath6 is a satellite knot if there is a pair of knots @xmath185 and @xmath186 in @xmath6 , called a companion knot and a pattern knot respectively , associated to @xmath12 , such that @xmath185 is nontrivial , @xmath186 is contained in a trivial solid torus @xmath187 in @xmath6 but is not contained in a @xmath188-ball of @xmath187 and is not isotopic to the core circle of @xmath187 , and there is a homeomorphism @xmath189 from @xmath187 to a regular neighborhood @xmath190 of @xmath185 in @xmath6 which maps a longitude of @xmath187 ( which bounds a disk in @xmath6 ) to a longitude of @xmath190 ( which bounds a seifert surface for @xmath185 ) and maps a meridian of @xmath187 to a meridian of @xmath190 , and finally @xmath191 .
we sometimes write a satellite knot as @xmath192 to include the above defining information ( @xmath12 still depends on how @xmath186 is embedded in @xmath187 ) .
[ p - factor]let @xmath193 be a satellite knot in @xmath6
. then @xmath194 in @xmath195 $ ] .
the lemma is obviously true when @xmath186 is the unknot in @xmath6 .
so we may assume that @xmath186 is a nontrivial knot in @xmath6 .
let @xmath69 , @xmath196 , @xmath197 be the exteriors of @xmath12 , @xmath185 and @xmath186 in @xmath6 respectively .
there is a degree one map @xmath198 such that @xmath199 and @xmath200 is a homeomorphism .
the map @xmath201 is given by a standard construction as follows . from the definition of the pattern knot given above
, we see that if @xmath64 is the exterior of @xmath186 in the trivial solid torus @xmath187 , then @xmath197 is obtained by dehn filling @xmath64 along @xmath202 with a solid torus @xmath203 such that the meridian slope of @xmath203 is identified with the longitude slope of @xmath187 .
also if we let @xmath204 , then @xmath205 . now the degree one map @xmath206 is defined to be : on @xmath207 it is the homeomorphism @xmath208 and on @xmath196 it maps a regular neighborhood of a seifert surface in @xmath196 to a regular neighborhood of a meridian disk of @xmath203 in @xmath203 and maps the rest of @xmath196 onto the rest of @xmath203 ( which is a @xmath188-ball ) .
the degree one map @xmath201 induces a surjective homomorphism @xmath209 from @xmath119 to @xmath210 such that @xmath211 is an isomorphism , mapping the meridian to the meridian and the longitude to the longitude . in turn
@xmath209 induces an embedding @xmath212 of @xmath213 into @xmath85 in such a way that the restriction of @xmath214 on @xmath87 with respect to the standard meridian - longitude basis @xmath215 of @xmath123 is the same as the restriction of @xmath213 on @xmath216 with respect to the standard meridian - longitude basis @xmath217 of @xmath218 , i.e. for each @xmath219 , we have @xmath220 , @xmath221 and @xmath222 .
the conclusion of the lemma now follows easily from the constructional definition of the @xmath1-polynomial . for polynomials
@xmath223 $ ] and @xmath224 $ ] both with nonzero degree in @xmath225 , let @xmath226 denote the resultant of @xmath227 and @xmath228 eliminating the variable @xmath225
. in general @xmath229 may have repeated factors even when both @xmath227 and @xmath228 are irreducible over @xmath103 . for a polynomial @xmath230 $ ] , let @xmath231\ ] ]
denote the polynomial obtained from @xmath232 by deleting all its repeated factors .
[ non - zero winding]let @xmath233 be a satellite knot such that the winding number @xmath169 of @xmath186 in the solid torus @xmath187 is non - zero .
then every balanced - irreducible factor @xmath234 of the @xmath1-polynomial @xmath235 of @xmath185 extends to a balanced factor @xmath236 of the @xmath1-polynomial @xmath28 of @xmath12 .
more precisely , ( 1 ) if the @xmath225-degree of @xmath234 is non - zero , then @xmath237 $ ] .
in particular if @xmath238 or @xmath239 for some nonnegative integer @xmath240 and @xmath241 ( such factor is irreducible and balanced ) , then @xmath242 or @xmath243 respectively ; ( 2 ) if the @xmath225-degree of @xmath234 is zero , i.e. @xmath244 is a function of @xmath245 only , then @xmath246 .
let @xmath69 , @xmath196 , @xmath207 , @xmath247 , @xmath248 be defined as in the proof of lemma [ p - factor ] .
we have @xmath205 .
note that @xmath249\oplus\z[\l_{\sc}]$ ] , and @xmath250=w[\l_\sc]$ ] , @xmath251=w[\m_\sk]$ ] . given a balanced - irreducible factor @xmath234 of @xmath235 ,
let @xmath55 be a component of @xmath252 which gives rise the factor @xmath234 .
for each element @xmath253 , the restriction of @xmath41 on @xmath254 can be extended to an abelian representation of @xmath255 , and thus @xmath41 can be extended to a representation of @xmath119 , which we still denote by @xmath41 , such that @xmath256 it follows that @xmath55 extends to a component or components of @xmath86 whose restriction on @xmath123 is or are one dimensional and thus all together gives rise a balanced factor @xmath236 of @xmath28 .
moreover the variables @xmath257 of @xmath236 and the variables @xmath245 and @xmath225 of @xmath234 are related by @xmath258 therefore when the @xmath225-degree of @xmath234 is positive , @xmath236 can be obtained by taking the resultant of @xmath259 and @xmath260 , eliminating the variable @xmath225 , and then deleting possible repeated factors . in particular if @xmath238 or @xmath239 for some nonnegative integer @xmath240 and @xmath261 , then the resultant of @xmath262 or @xmath263 with @xmath260 , eliminating the variable @xmath225 , is @xmath264 or @xmath243 respectively ( which is irreducible over @xmath103 and is banlanced ) . also if the degree of @xmath265 in @xmath225 is zero , then obviously @xmath246 . in proposition
[ non - zero winding ] , the @xmath162-degree of @xmath236 is at most equal to the @xmath225-degree of @xmath266 ( and generically they are equal ) .
this follows directly from the definition of the resultant ( cf .
* iv,8 ) ) .
next we are going to consider cabled knots .
let @xmath107 be a pair of relatively prime integers with @xmath267 , and @xmath12 be the @xmath107-cabled knot over a nontrivial knot @xmath185 .
that is , @xmath12 is a satellite knot with @xmath185 as a companion knot and with @xmath5 as a pattern knot which lies in the defining solid torus @xmath187 as a standard @xmath107-cable with winding number @xmath268 .
as the @xmath269-cable over a knot is equal to the @xmath107-cable over the same knot , we may always assume @xmath105 .
the following theorem gives a cabling formula for the @xmath1-polynomial of a cabled knot @xmath12 over a nontrivial knot @xmath185 , in terms of the @xmath1-polynomial @xmath270 of @xmath185 .
[ cabling formula ] let @xmath12 be the @xmath107-cabled knot over a nontrivial knot @xmath185 , with @xmath105
. then @xmath271\ ] ] if the @xmath225-degree of @xmath270 is nonzero and @xmath272 if the @xmath225-degree of @xmath270 is zero .
for a polynomial @xmath273 $ ] , define @xmath274 = \left\{\begin{array}{ll } red[res_{\overline y}(f(x^q,\overline y ) , \overline{y}^q - y ) ] , & \mbox{if the degree of $ f(\overline x ,
\overline y)$ in $ \overline y$ is nonzero},\\ f(x^q ) , & \mbox{if the degree of $ f(\overline x , \overline y)$ in $ \overline y$ is zero } \end{array}\right.\ ] ] then proposition [ non - zero winding ] was saying that @xmath275\ ] ] and theorem [ cabling formula ] is saying that @xmath276.\ ] ] note that @xmath277=red[\prod ext^q[f_\sc(\overline x,\overline y)]]$ ] where the product runs over all balanced - irreducible factors @xmath266 of @xmath270 and thus by proposition [ non - zero winding ] , @xmath278 $ ] is a balanced factor of @xmath28 .
so we only need to show [ f is a factor of a_k ] @xmath279 is a balanced factor of @xmath28 .
( note that each irreducible factor of @xmath279 is balanced ) .
[ no other factor ] besides @xmath279 and @xmath278 $ ] , @xmath28 has no other balanced factors . to prove the above two claims ,
let @xmath69 , @xmath196 , @xmath207 , @xmath247 , @xmath248 be defined as in lemma [ p - factor ] with respect to @xmath280 where @xmath281 is embedded in @xmath187 as a standard @xmath107-cable .
we have @xmath205 and @xmath282 . for a convenience to the present argument , we give a direct description of @xmath207 as follows .
we may consider @xmath283 as @xmath284 , where @xmath92 is a disk of radius @xmath78 , such that @xmath285 has slope zero for each point @xmath286 .
let @xmath287 be the concentric sub - disk in @xmath92 with radius @xmath29 .
then @xmath288 is a solid torus in @xmath190 sharing the same core circle @xmath185 .
we may assume that the knot @xmath12 is embedded in the boundary of @xmath289 as a standard @xmath107-curve , where @xmath290 has the meridian - longitude coordinates consistent with that of @xmath291 ( i.e. for a point @xmath292 , @xmath285 is a longitude of @xmath290 ) .
then @xmath207 is the exterior of @xmath12 in @xmath293 .
note that @xmath207 is a seifert fibred space whose base orbifold is an annulus with a single cone point of order @xmath294 , a seifert fiber of @xmath207 in @xmath295 has slope @xmath296 and a seifert fiber of @xmath207 in @xmath123 has slope @xmath184 .
let @xmath297 be a seifert fibre of @xmath207 lying in @xmath298 and @xmath299 be a seifert fibre of @xmath207 lying in @xmath300 .
up to conjugation , we may consider @xmath297 and @xmath299 as elements of @xmath255 . also note that @xmath297 is conjugate to @xmath299 in @xmath255 .
it is well known that each of @xmath297 and @xmath299 lies in the center of @xmath255 which is independent of conjugation .
it follows that if @xmath301 is an irreducible representation then @xmath302 , for some fixed @xmath303 , where @xmath304 is the identity matrix .
hence if @xmath55 is an irreducible subvariety of @xmath305 which contains the character of an irreducible representation , then for every @xmath253 , we have @xmath302 , which is due to the fact that the characters of irreducible representations are dense in @xmath55
. now we are ready to prove claim [ f is a factor of a_k ] . if @xmath306 , then by lemma [ p - factor ] and the formula ( [ apoly of t(p , q ) ] ) , @xmath279 is a factor of @xmath28 .
so we may assume that @xmath307 .
under this assumption , one can see that the fundamental group of @xmath207 has the following presentation : @xmath308 such that @xmath309 where @xmath122 is a based simple loop free homotopic in @xmath207 to the center circle of @xmath293 and @xmath310 is a based simple loop free homotopic in @xmath207 to @xmath311 in @xmath295 . to see these assertions , note that @xmath207 contains the essential annulus @xmath312 with @xmath313 of the slope @xmath184 ( the cabling annulus , consisting of seifert fibers of @xmath207 ) and @xmath314 decompose @xmath207 into two pieces @xmath315 and @xmath316 , such that @xmath315 is a solid torus ( which is @xmath317 ) and @xmath316 is topologically @xmath295 times an interval .
the above presentation for @xmath255 is obtained by applying the van kampen theorem associated to the splitting of @xmath318 along @xmath314 .
we should note that as @xmath307 , a longitude in @xmath290 intersects @xmath12 geometrically exactly once .
it is this curve pushed into @xmath315 which yields the element @xmath122 and pushed into @xmath316 which yields @xmath310 . also because @xmath307 , @xmath319 form a basis for @xmath320 .
therefore by van kampen , @xmath255 is generated by @xmath321 with relations @xmath322 and @xmath323 , which yields presentation ( [ presentation of for fund gp of y ] ) after canceling the element @xmath297 . with a suitable choice of orientation for @xmath310
i.e. replacing @xmath310 by its inverse if necessary , we also have @xmath324 .
see figure [ cross - section ] for an illustration when @xmath325 . of @xmath207 and the elements @xmath122 and @xmath310 of @xmath255 when @xmath325 .
] by ( * ? ?
* lemma 7.2 ) , @xmath326 , which denotes the manifold obtained by dehn surgery on @xmath12 in the solid torus @xmath293 with the slope @xmath184 , is homeomorphic to @xmath327 , and @xmath328 , which denotes the manifold obtained by dehn filling of @xmath69 with the slope @xmath184 , is homeomorphic to @xmath329 ( where the meaning of the notation @xmath330 should be obvious ) . as @xmath307 , @xmath330 is a homology sphere . by @xcite
, @xmath331 contains at least one irreducible representation @xmath332 .
note that the restriction of @xmath332 on @xmath254 is not contained in @xmath333 .
that is , we have : @xmath334 and @xmath335 ( we may consider @xmath310 as the longitude @xmath311 of @xmath254 ) . if @xmath336 , we can extend @xmath332 to a curve of representations of @xmath119 with one dimensional characters .
in fact for every @xmath337 , we may define @xmath338 as follows .
on @xmath339 , let @xmath340 , so in particular @xmath341 , define @xmath342 where @xmath343 is a fixed order @xmath294 matrix in @xmath33 .
it s routine to check that @xmath344 is well defined , the trace of @xmath345 varies as @xmath1 runs over @xmath33 , and @xmath346 .
hence we get a curve in @xmath85 whose restriction on @xmath123 is one - dimensional , and moreover this curve generates @xmath347 as a factor of @xmath28 if @xmath348 or generates @xmath349 as a factor of @xmath28 if @xmath350 . we now show that @xmath351 is a factor of @xmath28 if @xmath348 or @xmath352 is a factor of @xmath28 if @xmath350 , for all @xmath105 . for the representation @xmath332 given in the last paragraph , defined @xmath353 by @xmath354 where @xmath355 is the onto homomorphism @xmath356 . as @xmath357 is a generator of @xmath358 , @xmath359 .
similarly as @xmath311 is trivial in @xmath360 , @xmath361 .
hence @xmath362 .
we can now extend @xmath363 to @xmath364 over @xmath69 similarly as for @xmath332 to @xmath344 , only this time we choose @xmath343 as a fixed order @xmath365 matrix in @xmath33 , so that the trace of @xmath366 varies as @xmath1 runs over @xmath33 , and @xmath367 .
the existence of the factor of @xmath28 that we set to prove now follows .
this completes the proof of claim [ f is a factor of a_k ] .
we lastly prove claim [ no other factor ] . given a balanced - irreducible factor @xmath98 of @xmath28 , let @xmath55 be an irreducible component of @xmath368 over @xmath103 which produces @xmath98 .
let @xmath369 and @xmath370 be the zariski closure of the restriction of @xmath55 on @xmath207 and @xmath196 respectively .
note that each of @xmath369 and @xmath370 is irreducible .
also @xmath369 is at least one dimensional since its restriction on @xmath123 is one dimensional .
if @xmath369 does not contain irreducible characters , then the restriction of @xmath369 on @xmath295 is also one dimensional .
so the restriction of @xmath370 on @xmath295 is one dimensional and @xmath55 is an abelian extension of @xmath370 ( since the character of a reducible representation is also the character of an abelian representation ) .
it follows that @xmath370 can not be the trivial component in @xmath371 for otherwise @xmath55 woulde be the trivial component of @xmath85 .
hence @xmath372 and @xmath55 is its abelian extension , which means that @xmath98 is a factor of @xmath277 $ ] .
hence we may assume that @xmath373 contains irreducible characters .
then , as noted before , we have @xmath374 for each @xmath41 with @xmath375 . if @xmath376 , then @xmath377 or @xmath378 is the factor contributed by @xmath55 to @xmath28 corresponding to @xmath379 is positive or negative respectively .
that is , @xmath98 is a factor of @xmath279 .
hence we may assume that @xmath380 .
it follows that @xmath369 is a positive dimensional component of @xmath326 .
therefore @xmath336 ( because @xmath381 ) and @xmath55 contributes the factor @xmath382 or @xmath383 to @xmath28 corresponding to @xmath379 is positive or negative respectively .
that is , @xmath384 is a factor of @xmath279 .
this proves claim [ no other factor ] and also completes the proof of the theorem .
perhaps we should note that in the above proof consistent choice of base points can be made for all the relevant manifolds such as @xmath69 , @xmath196 , @xmath207 , @xmath314 , @xmath315 , @xmath316 , @xmath295 , @xmath123 , @xmath326 , @xmath385 , @xmath328 so that their fundamental groups are all well defined as relevant subgroups or quotient groups .
in fact we can choose a simply connected region in @xmath386 ( as shown in figure [ cross - section ] ) so that its intersection with each relevant manifold listed above is a simply connected region which is served as the ` fat base point ' of that manifold .
let @xmath185 be the figure @xmath387 knot .
its @xmath1-polynomial is @xmath388 ( ( * ? ? ?
* appendix ) ) .
if @xmath12 is the @xmath389-cable over @xmath185 , @xmath390 , then @xmath391\\ & = ( 1+x^{2p}y)[x^{16 } + ( -1 + 2 x^4 + 3 x^8 - 2 x^{12 } - 6 x^{16 } - 2 x^{20 } + 3 x^{24 } + 2 x^{28 } - x^{32 } ) y + x^{16 } y^2].\end{array}\ ] ] if @xmath12 is the @xmath392-cable over @xmath185 , @xmath390 , then @xmath393\\ & = ( -1+x^{6p}y^2)[x^{36 } + ( -1 + 3 x^6 + 3 x^{12}- 8 x^{18 } - 12 x^{24 } + 6 x^{30}+ 20 x^{36 } + 6 x^{42 } - 12 x^{48 } \\&\;\;\ ; - 8 x^{54}+ 3 x^{60}+ 3 x^{66 } - x^{72 } ) y+ x^{36 } y^2].\end{array}\ ] ] finally we would like to give an explicit formula for the @xmath1-polynomials of iterated torus knots .
let @xmath394\ ] ] be an @xmath240-th iterated torus knot , i.e. @xmath395 is a nontrivial torus knot and when @xmath396 , for each @xmath397 , @xmath398 , @xmath399 $ ] is a satellite knot with @xmath400 $ ] as a companion knot and with @xmath401 , @xmath402 , as a pattern knot lying in the trivial solid torus @xmath187 as a @xmath403-cable with winding number @xmath404 , where @xmath405 may be less than @xmath404 and @xmath406 is also allowed .
[ a - poly of iterated torus]let @xmath407 $ ] be an iterated torus knot . if for each @xmath408 , @xmath404 is odd , then @xmath409 and if for some @xmath408 , @xmath404 is even , we let @xmath410 be the smallest such integer , then @xmath411 \(1 ) in the @xmath1-polynomial given in the corollary , each @xmath412 is a nontrivial polynomial even when @xmath406 .
\(2 ) the polynomial expression for @xmath413 given in the corollary has no repeated factors .
\(3 ) the boundary slopes detected by @xmath414 are precisely the following @xmath240 integer slopes : @xmath415 the proof goes by induction on @xmath240 applying theorem [ cabling formula ] .
when @xmath416 , the proposition holds obviously .
suppose for @xmath417 the proposition holds .
note that @xmath12 has @xmath418 $ ] as a companion knot and @xmath419 as the corresponding pattern knot ( which maybe a trivial knot in @xmath6 , which occurs exactly when @xmath420 ) . by induction ,
the @xmath1-polynomial @xmath421 of @xmath185 is of the corresponding form as described by the corollary .
now applying theorem [ cabling formula ] one more time to the pair @xmath422 we see that the corollary holds .
we omit the routine details . for instance
the a - polynomial of the @xmath423-cable over the @xmath107-torus knot is @xmath424
suppose that @xmath12 is a knot in @xmath6 with the same knot floer homology and the same a - polynomial as a given torus knot @xmath5 .
our goal is to show that @xmath13 . by ( [ apoly of t(p , q ) ] ) and lemma [ non - hyperbolic ] , @xmath12 is not a hyperbolic knot .
so @xmath12 is either a torus knot or a satellite knot .
[ torus knot case]suppose that @xmath425 is a torus knot whose @xmath1-polynomial divides that of @xmath5 and whose alexander polynomial divides that of @xmath5 , then @xmath426 .
since @xmath427 , from the formula ( [ apoly of t(p , q ) ] ) , we have @xmath428 . from the condition
@xmath429 we have @xmath430 now if @xmath431 , then either @xmath432 or @xmath433 . if @xmath432 , then from ( [ pqrs ] ) we must have @xmath434 . by our convention for parameterizing torus knots , @xmath113 .
thus we must also have @xmath435 .
but @xmath379 and @xmath294 are relatively prime , we have @xmath436 , which contradicts the early conclusion @xmath428 .
if @xmath433 , again by our convention @xmath437 , we see that @xmath438 does not divide @xmath439 , and so ( [ pqrs ] ) can not hold .
this contradiction completes the proof of the lemma . by lemma [ torus knot case ]
, we see that if @xmath12 is a torus knot , then @xmath13 .
we are going to show that it is impossible for @xmath12 to be a satellite knot , which will consist of the rest of the proof of theorem [ main1 ] .
suppose that @xmath12 is a satellite knot .
we need to derive a contradiction from this assumption .
let @xmath185 and @xmath186 be a pair of associated companion knot and pattern knot to @xmath12 , and let @xmath169 be the winding number of @xmath186 in its defining solid torus @xmath187 ( recall the definition of a satellite knot given in section [ apoly ] ) . as @xmath5 is a fibred knot , @xmath12 is also fibred .
according to ( * ? ? ?
* corollary 4.15 and proposition 8.23 ) , each of @xmath185 and @xmath186 is a fibred knot in @xmath6 , @xmath440 , and the alexander polynomials of these knots satisfy the equality @xmath441 we may choose @xmath185 such that @xmath185 is itself not a satellite knot , and thus is either a hyperbolic knot or a torus knot .
the companion knot @xmath185 can not be a hyperbolic knot .
suppose that @xmath185 is hyperbolic .
then @xmath235 contains a balanced - irreducible factor @xmath234 whose newton polygon detects at least two distinct boundary slopes of @xmath185 .
as @xmath440 , by proposition [ non - zero winding ] , @xmath234 extends to a balanced factor @xmath236 of @xmath442 .
moreover from the relation ( [ x and x bar ] ) we see that the newton polygon of @xmath236 detects at least two distinct boundary slopes of @xmath12 .
but clearly the newton polygon of @xmath442 only detects one boundary slope .
we arrive at a contradiction .
so @xmath443 is a torus knot .
[ nontrivial pattern ] the pattern knot @xmath186 of @xmath12 can not be the unknot .
suppose otherwise that @xmath186 is the unknot .
then as noted in @xcite the winding number @xmath169 of @xmath186 in its defining solid torus @xmath187 is larger than @xmath29 .
equality ( [ alexander poly ] ) becomes @xmath444 for some integer @xmath445 . on the other hand
it is easy to check that the degrees of the leading term and the second term of @xmath446 differ by @xmath29 and thus @xmath447 can not be of the form @xmath448 , @xmath445 .
this contradiction completes the proof .
if @xmath449 , then by proposition [ non - zero winding ] , @xmath450 divides @xmath442 , and by ( [ alexander poly ] ) @xmath451 divides @xmath447 .
hence by lemma [ torus knot case ] we have @xmath452 . by lemma [ nontrivial pattern ]
, @xmath186 is a non - trivial knot .
hence from formula ( [ alexander poly ] ) , we see that the genus of @xmath453 is less than that of @xmath12 .
but the genus of @xmath12 is equal to that of @xmath5 .
we derive a contradiction .
hence @xmath445 . by lemma [ p - factor ]
, @xmath454 divides @xmath28 .
now if @xmath186 is itself a satellite knot with its own companion knot @xmath455 and pattern knot @xmath456 .
then again each of @xmath455 and @xmath456 is a fibred knot and the winding number @xmath457 of @xmath456 with respect to @xmath455 is larger than zero . arguing as above
, we see that @xmath455 may be assumed to be a torus knot and that @xmath458 .
also we have @xmath459 from which we see that @xmath456 can not be the trivial knot just as in the proof of lemma [ nontrivial pattern ] .
so after a finitely many such steps ( the process must terminate by @xcite ) , we end up with a pattern knot @xmath460 for @xmath461 such that @xmath460 is nontrivial but is no longer a satellite knot .
thus @xmath460 is either a hyperbolic knot or a torus knot . by lemma [ non - hyperbolic ] and lemma [ p - factor ]
, @xmath460 can not be hyperbolic .
hence @xmath460 is a torus knot . again
because @xmath462 divides @xmath442 and @xmath463 divides @xmath447 , we have @xmath464 by lemma [ torus knot case ] .
but once again we would have @xmath465 .
this gives a final contradiction .
in @xcite a family of hyperbolic knots @xmath18 in @xmath6 ( where at least one of @xmath379 and @xmath240 has to be zero ) was constructed such that each knot in the family admits one ( and only one ) half - integral toroidal surgery . to be hyperbolic ,
the following restrictions on the values for @xmath19 are imposed : @xmath466 from now on we assume that any given @xmath18 is hyperbolic , i.e. @xmath19 satisfy the above restrictions . the half - integral toroidal slope @xmath467 of @xmath18 was explicitly computed in ( * ? ? ?
* proposition 5.3 ) as given below : @xmath468 it turns out that @xmath18 are the only hyperbolic knots in @xmath6 which admit non - integral toroidal surgeries .
[ gl2](@xcite ) if a hyperbolic knot @xmath12 in @xmath6 admits a non - integral toroidal surgery , then @xmath12 is one of the knots @xmath18 .
[ prop : dupl ] the knots @xmath18 have the following properties : ( a ) @xmath469 is the mirror image of @xmath470 .
( b ) @xmath24 is the mirror image of @xmath471 .
( c ) @xmath472 .
( d ) @xmath473 .
this follows from ( * ? ? ?
* proposition 1.4 ) .
explicit closed braid presentations for the knots @xmath18 are given in @xcite .
figure [ fig : braid ] shows the braids whose closure are @xmath18 .
the left two pictures are a reproduction of ( * ? ? ?
12 ) , but the right two pictures are different from ( * ? ? ?
13 ) . here
an arc with label @xmath474 means @xmath474 parallel strands , and a box with label @xmath42 means @xmath42 positive full - twists when @xmath475 and @xmath476 negative full - twists when @xmath477 .
we only give the picture for the case @xmath478 , since the case @xmath479 can be treated by applying proposition [ prop : dupl ] .
( 375,167 ) ( 0,0 ) ( 19,0 ) 90 @xmath480 ( 42,3)@xmath481 ( 38,135)@xmath240 ( 60,0 ) 90 @xmath480 ( 25,25 ) 45 @xmath482 ( 2,34)@xmath483 ( 6,64)@xmath481 ( 47,58 ) 78 @xmath484 ( 15,-10)@xmath485 ( 120,0 ) 90 @xmath486 ( 143,3)@xmath481 ( 161,0 ) 90 @xmath486 ( 102,30)@xmath481 ( 122,18 ) 50 @xmath487 ( 123,62 ) 50 @xmath482 ( 141,135)@xmath240 ( 117,32)@xmath483 ( 159,30)@xmath483 ( 116,-10)@xmath488 ( 220,1 ) ( 249,135)@xmath379 ( 226,3 ) 90 @xmath489 ( 250,3 ) 90 @xmath482 ( 282,93 ) 87 @xmath482 ( 273,3 ) 90 @xmath489 ( 243,84 ) 60 @xmath490 ( 260,89 ) 0 @xmath481 ( 256,69 ) 0 @xmath491 ( 270,117)@xmath483 ( 247,117)@xmath483 ( 220,-10)@xmath492 ( 337,135)@xmath379 ( 367,24)@xmath483 ( 311,10)@xmath491 ( 331,10)@xmath481 ( 348,2 ) 90 @xmath486 ( 369,2 ) 90 @xmath486 ( 312,84 ) 45 @xmath482 ( 307,58 ) 50 @xmath487 ( 314,-10)@xmath493 [ prop : genus ] suppose that @xmath478 , let @xmath494 then the genus of @xmath469 is @xmath495 and the genus of @xmath24 is @xmath496 in @xcite , it is noted that @xmath18 is the closure of a positive or negative braid .
hence it is fibered , and the genus can be computed by the formula @xmath497 where @xmath185 is the crossing number in the positive or negative braid , and @xmath293 is the braid index .
when @xmath498 , the braid is a negative braid , @xmath499 , @xmath500 so @xmath501 when @xmath502 , we can cancel all the negative crossings in the braid to get a positive braid of index @xmath499 .
we have @xmath503 so @xmath504 the computation for other cases are similar .
using proposition [ prop : genus ] and proposition [ prop : dupl ] ( a)(b ) , we can compute the genus of @xmath18 when @xmath479 .
for example , when @xmath505 , the genus of @xmath24 is @xmath506 the @xmath507-surgery on the knot @xmath508 was explicitly given in @xcite which is the double branched cover of @xmath6 with the branched set in @xmath6 being a link shown in figure [ tangle1 ] . from the tangle decomposition of the branched link one can see that @xmath509 is a graph manifold obtained by gluing two seifert fibred spaces , each over a disk with two cone points , together along their torus boundaries . for our purpose , we need to give a more detailed description of the graph manifold @xmath509 as follows . let @xmath510 denote a two string tangle , i.e. @xmath343 is a @xmath188-ball and @xmath42 is a pair of disjoint properly embedded arcs in @xmath343 .
here we may assume that @xmath343 is the unit @xmath188-ball in the @xmath511-space @xmath512 ( with the @xmath136-plane horizontal ) and that the four endpoints of @xmath42 lies in the lines @xmath513 , @xmath514 and @xmath515 .
let @xmath92 be the unit disk in @xmath343 which is the intersection of @xmath343 with the @xmath516-plane .
then the four endpoints of @xmath42 divides @xmath517 into four arcs , naturally named the east , west , north , south arcs .
the _ denominator closure _ of @xmath510 is the link in @xmath6 obtained by capping off @xmath42 with the east and west arcs , and the _ numerator closure _ of @xmath510 is the link in @xmath6 obtained by capping off @xmath42 with the north and south arcs .
let @xmath518 , @xmath519 , be the two tangles shown in figure [ tangle1 ] and let @xmath520 be the double branched cover of @xmath518 .
the denominator closure of @xmath521 is the twisted knot of type @xmath522 ( which is the trivial knot when @xmath523 , the trefoil knot when @xmath350 ) and therefore the double branched cover of @xmath6 over the link is the lens space of order @xmath524 .
the numerator closure of @xmath521 gives a composite link in @xmath6 and in fact the composition of two nontrivial rational links corresponding to the rational numbers @xmath525 and @xmath526 .
for the half - integral toroidal surgery ] the double branched cover of the east arc ( as well as the west arc ) is a simple closed essential curve in @xmath527 , which we denote by @xmath528 , with which dehn filling of @xmath529 is a lens space of order @xmath524 .
the double branched cover of the north arc ( as well as the south arc ) is a seifert fiber of @xmath529 , which we denote by @xmath530 , with which dehn filling of @xmath529 is a connected sum of two nontrivial lens spaces of orders @xmath531 and @xmath532 .
similarly the numerator closure of @xmath533 is the twisted knot of type @xmath534 and the denominator closure of @xmath533 is a composite link of two nontrivial rational links corresponding to the rational numbers @xmath535 and @xmath536 .
so the double branched cover of the north arc of @xmath533 is a simple closed essential curve in @xmath537 , denoted @xmath538 , with which dehn filling of @xmath539 is a lens space of order @xmath540 , and the double branched cover of the west arc is a seifert fiber of @xmath539 , denoted by @xmath541 , with which dehn filling is a connected sum of two nontrivial lens spaces of orders @xmath78 and @xmath542 .
finally @xmath509 is obtained by gluing @xmath529 and @xmath539 along their boundary tori such that @xmath543 is identified with @xmath541 and @xmath530 with @xmath538 . note that when @xmath544 or @xmath29 , @xmath529 is the exterior of a torus knot in @xmath6 , and when @xmath545 or @xmath29 , @xmath539 is the exterior of a @xmath546 torus knot in @xmath6 .
[ only one incomp surface ] let @xmath508 , @xmath507 the unique half - integral toroidal slope of @xmath3 and @xmath547 the exterior of @xmath3 . up to isotopy , there is a unique closed orientable incompressible surface in @xmath509 , which is a torus . as discussed above , @xmath509 is a graph manifold with the torus decomposition @xmath548 .
we just need to show that any connected closed orientable incompressible surface @xmath549 in @xmath509 is isotopic to @xmath527 .
suppose otherwise that @xmath549 is not isotopic to @xmath527 .
as each @xmath520 does not contain closed essential surfaces , @xmath549 must intersect @xmath527 .
we may assume that @xmath550 is incompressible and boundary incompressible , for each @xmath519 .
as @xmath520 is seifert fibred , @xmath551 is either horizontal ( i.e. consisting of seifert fibres or vertical ( transverse to seifert fibres ) in @xmath520 , up to isotopy .
so the boundary slope of @xmath551 is either @xmath552 ( when @xmath551 is horizontal ) or is the rational longitude of @xmath520 ( when @xmath551 is vertical ) .
but @xmath552 is identified with @xmath553 , for @xmath519 , @xmath554 .
so the boundary slope of @xmath551 must be the rational longitude of @xmath520 for each @xmath397 .
it follows that @xmath555 is infinite , yielding a contradiction .
[ meridian not bdy slope ] for each @xmath508 , its meridian slope is not a boundary slope .
if the meridian slope of @xmath3 is a boundary slope , then the knot exterior @xmath547 contains a connected closed orientable incompressible surface @xmath549 of genus
larger than one such that @xmath549 remains incompressible in any non - integral surgery , by ( * ? ? ?
* theorem 2.0.3 ) .
but by lemma [ only one incomp surface ] , @xmath509 does not contain any closed orientable incompressible surface of genus larger than one .
this contradiction completes the proof . and
@xmath556 in @xmath557
lastly in this section we are going to show that @xmath14 is a class of small knots .
let @xmath558 be a standard genus two handlebody in @xmath6 and let @xmath559 and @xmath556 be oriented loops in the surface @xmath560 as shown in figure [ genus two surface ] .
as observed in ( * ? ?
* figure 8) , a regular neighborhood of @xmath561 in @xmath562 is a genus one seifert surface of a trefoil knot . under the basis @xmath563,[g_2]$ ] , the monodromy of the trefoil is represented by the matrix @xmath564 from ( * ? ? ?
* figure 11 ) , the knot @xmath14 has a knot diagram as shown in figure [ fig ( l,-1 ) ] ( a ) , which in turn can be embedded in the seifert surface with the homology class @xmath565+(l_*+1)[g_2]$ ] as shown in figure [ fig ( l,-1 ) ] ( b ) . for any knot @xmath3 lying in the seifert surface , an algorithm to classify all closed essential surfaces in the knot exterior of @xmath3 is given in ( * ? ? ?
* theorem 10.1 ( 3 ) ) . for simplicity
, we only state the part that we need : ] suppose that @xmath3 is a simple closed curve on the genus one seifert surface of a trefoil knot in the homology class @xmath566+a_2[g_2]$ ] .
the slope @xmath567 has continued fraction expansion @xmath568=b_1-\frac1{\displaystyle b_2-\frac1{\displaystyle\ddots-\frac1{b_k}}},\ ] ] where the coefficients alternate signs , @xmath569 when @xmath570 , and @xmath571 . if @xmath572 and @xmath573 , then every closed essential surface in the complement of @xmath574 corresponds to a solution of the following equation : @xmath575 where @xmath304 and @xmath3 are subsets of @xmath576 each not containing consecutive integers and @xmath577 .
[ lem : small ] each @xmath14 is a small knot .
namely , the exterior of @xmath14 contains no closed essential surfaces . in this case @xmath578 , @xmath579 $ ] .
clearly , equation ( [ eq : esssurf ] ) has no solution .
so @xmath14 is small .
to obtain explicit expressions for the @xmath1-polynomials of the knots @xmath580 could be a very tough task .
but we can obtain some useful info about the @xmath1-polynomials of these knots without computing them explicitly .
[ lem : abelrep ] suppose a graph manifold @xmath64 is obtained by gluing two torus knot exteriors @xmath581 together , such that the meridian of @xmath520 is glued to the seifert fiber of @xmath582 , @xmath519 , where @xmath583 .
then @xmath584 has no non - cyclic @xmath585 representations . to prove this lemma
, we will use the following well - known fact whose proof is elementary .
[ lem : sl2comm ] suppose that @xmath586 are two commuting matrices , @xmath587 .
( i ) if there exits @xmath588 such that @xmath589 then @xmath590 ( ii ) if there exits @xmath588 such that @xmath591 then @xmath592 as @xmath64 has cyclic homology group , it is equivalent to show that every @xmath33-representation of @xmath584 is abelian .
we choose a base point of @xmath64 on the common boundary torus @xmath593 of @xmath581 , then @xmath594 are naturally the subgroups of @xmath584 .
suppose that @xmath595 is a representation .
let @xmath596 represent the seifert fiber of @xmath520 , @xmath519 .
we first consider the case that one of @xmath597 , say @xmath598 , is in @xmath599 which is the center of @xmath585 .
since @xmath600 represents the meridian of @xmath539 , @xmath600 normally generates @xmath601 , hence @xmath602 is contained in @xmath603 . since @xmath604
, we see that @xmath605 is also in @xmath603 , and thus for the same reason as just given @xmath606 is in @xmath603 .
hence @xmath607 is contained in @xmath599 .
now suppose that neither of @xmath597 is in @xmath603 , then there exists @xmath588 such that @xmath608 is in the form of either ( [ eq : diag ] ) or ( [ eq : para ] ) .
since @xmath600 is in the center of @xmath609 , by lemma [ lem : sl2comm ] , @xmath610 ( including @xmath611 ) is contained in an abelian subgroup @xmath1 of @xmath585 ( @xmath1 is either the set of diagonal matrices or the set of upper triangular trace @xmath612 matrices ) . since @xmath613 is in the center of @xmath601 , again by lemma [ lem : sl2comm ] , @xmath614
is also contained in @xmath1 .
hence @xmath615 , being generated by @xmath616 , is an abelian group .
[ prop : abelrep ] the fundamental group of the half - integral toroidal surgery on @xmath617 has no non - cyclic @xmath585 representations and has no non - cyclic @xmath618 representations . by the discussion preceding lemma [ only one incomp surface ] , the half - integral toroidal surgery , @xmath509 , is a graph manifold satisfying the conditions of lemma [ lem : abelrep ] .
hence @xmath509 has no non - cyclic @xmath33 representations .
the manifold can not have non - cyclic @xmath618 representations either since every @xmath618 representation of @xmath509 lifts to a @xmath33-representation because the manifold has odd cyclic first homology .
[ r is a vertex]let @xmath547 be the knot exterior of any given hyperbolic knot @xmath617 .
let @xmath55 be any norm curve in @xmath619 and @xmath145 the norm polygon determined by @xmath55 .
then the half - integral toroidal slope @xmath620 of @xmath3 is associated to a vertex of @xmath145 as described in theorem [ properties of a norm curve ] ( 3 ) , i.e. @xmath621 is the slope of a vertex of @xmath145 in the @xmath136-plane @xmath622 .
suppose otherwise that @xmath507 is not associated to a vertex of @xmath145 .
as the meridian slope of @xmath547 is not a boundary slope by lemma [ meridian not bdy slope ] , it follows from theorem [ properties of a norm curve ] ( 4 ) and ( 5 ) that @xmath623 is contained in @xmath624 but @xmath507 is not , which means that @xmath625 for some point @xmath135 . as @xmath509 has no noncyclic representations by lemma [ prop : abelrep ] , the point @xmath626 can not be a regular point of @xmath129 ( by ( * ? ? ? * proposition 1.5.2 ) or ) .
so @xmath626 is an ideal point of @xmath129 . as @xmath507 is not a slope associated to any vertex of @xmath145 ,
@xmath627 is finite for every class @xmath122 in @xmath628 .
now we may apply ( * ? ? ?
* proposition 4.12 ) to see that @xmath547 contains a closed essential surface @xmath549 such that if @xmath549 compresses in @xmath509 and @xmath629 , then @xmath630 . by lemma
[ only one incomp surface ] , @xmath549 must compress in @xmath509 and of course @xmath549 compresses in @xmath631 .
but @xmath632 . we arrive at a contradiction .
[ j has only 3 roots]let @xmath633 be the @xmath1-polynomial of any given hyperbolic knot @xmath617 .
let @xmath620 be the half - integral toroidal slope of @xmath3 . if @xmath634 is a solution of the system @xmath635 then @xmath636 .
suppose otherwise that @xmath637 .
then by the constructional definition of the @xmath1-polynomial , there is a component @xmath529 in @xmath638 which contributes a factor @xmath98 in @xmath633 such that @xmath639 is a solution of @xmath640 let @xmath165 be the zariski closure of @xmath641 in @xmath642 .
we knew @xmath165 is an irreducible curve .
we may find an irreducible curve @xmath55 in @xmath529 such that @xmath165 is also the zariski closure of @xmath166 .
now it also follows from the constructional definition of the @xmath1-polynomial that there is a convergent sequence of regular points @xmath643 such that @xmath644 and @xmath645 , i.e. if @xmath626 is the limit point of @xmath646 in @xmath129 , then @xmath647 and @xmath648 .
note that @xmath649 is not constant on @xmath55 .
for otherwise @xmath55 would be a semi - norm curve with @xmath507 as the associated slope and would have a contradiction with theorem [ properties of a semi - norm curve ] ( 3 ) ( since by lemma [ meridian not bdy slope ] @xmath623 is not a boundary slope ) .
so we have @xmath650 .
again @xmath626 can not be a regular point of @xmath129 due to lemma [ prop : abelrep ] .
so @xmath626 is an ideal point of @xmath129 such that @xmath651 and @xmath652 is finite .
we get a contradiction with ( * ? ? ?
* prposition 4.12 ) as in the proof of lemma [ r is a vertex ] . [ nonhyp surgery ] let @xmath12 be a hyperbolic knot in @xmath6 .
for a given slope @xmath296 , if every solution @xmath634 of the system of equations @xmath653 has @xmath654 , then @xmath655 is not a hyperbolic manifold . some of the ideas for the proof come from @xcite .
suppose otherwise that @xmath655 is a hyperbolic @xmath188-manifold .
then @xmath656 has a discrete faithful representation @xmath657 into @xmath618 . by thurston ( ( * ? ? ?
* proposition 3.1.1 ) ) , this representation can be lifted to a @xmath33-representation @xmath658 .
it follows from the mostow rigidity that the character @xmath659 of @xmath658 is an isolated point in @xmath660 .
note that @xmath658 can be considered as an element in @xmath661 and @xmath659 can be considered as an element in @xmath85 since @xmath662 embeds in @xmath661 and @xmath660 embeds in @xmath85 .
of course we have @xmath663 but @xmath664 .
let @xmath55 be a component of @xmath85 which contains @xmath659 .
by thurston ( ( * ? ?
* proposition 3.2.1 ) ) , @xmath55 is positive dimensional . [
claim : non - constant ] the function @xmath665 is non - constant on @xmath55 .
suppose otherwise .
then @xmath665 is constantly zero on @xmath55 since @xmath666 .
so for every @xmath253 , @xmath667 is either @xmath304 or @xmath668 or is a parabolic element .
let @xmath529 be an irreducible curve in @xmath55 which contains the point @xmath659 .
for a generic point @xmath669 , @xmath667 can not be a parabolic element since otherwise @xmath670 is either @xmath304 or @xmath668 or parabolic for all @xmath669 and this happens in particular at the point @xmath659 , yielding a contradiction . for a generic point @xmath669 , @xmath667 can not be @xmath668 either for otherwise by continuity , @xmath671 for every point @xmath669 , and this happens in particular at the point @xmath659 , yielding another contradiction .
so for a generic point @xmath669 , @xmath672 and again by continuity , @xmath672 for every point @xmath669 .
so @xmath529 factors though the @xmath296-surgery on @xmath12 and becomes a subvariety of @xmath660 .
but this contradicts the fact that @xmath659 is an isolated point of @xmath660 .
the claim is thus proved .
it also follows from the proof of claim [ claim : non - constant ] that @xmath55 is one dimensional . for
otherwise there would be a curve @xmath529 in @xmath55 such that @xmath673 and @xmath665 is constantly zero on @xmath529 , which is impossible by the proof of claim [ claim : non - constant ] .
it follows from claim [ claim : non - constant ] that the restriction of @xmath55 in @xmath87 is one dimensional and thus @xmath674 and contributes a factor @xmath98 to @xmath28 .
we may assume , up to conjugation of @xmath658 , that @xmath675 note that @xmath676 ( as @xmath664 and can not be a parabolic element of @xmath33 ) . by the construction of @xmath28
, @xmath639 is a solution of the system @xmath677 and thus is a solution of @xmath678 we get a contradiction with the assumption of the lemma .
[ same r slope ] if @xmath2 is a hyperbolic knot whose @xmath1-polynomial divides the @xmath1-polynomial of @xmath617 , then @xmath12 has the same half - integral toroidal slope as @xmath3 and thus @xmath12 is one of @xmath18 .
since @xmath12 is hyperbolic , @xmath85 contains a norm curve component @xmath679 which contributes a balanced - irreducible factor @xmath98 to @xmath28 such that the newton polygon of @xmath98 is dual to the norm polygon determined by @xmath679 by theorem [ dual polygons ] . by the assumption that @xmath28 divides @xmath633
, @xmath98 is also a factor of @xmath633 .
thus there is a curve @xmath55 in a component of @xmath638 which contributes @xmath98 and @xmath55 must be a norm curve whose norm polygon @xmath145 is dual to the newton polygon of @xmath98 . by lemma
[ r is a vertex ] , the half - integral toroidal slope @xmath620 of @xmath3 is associated to a vertex of @xmath145 and thus @xmath620 is also associated to an edge of the newton polygon of @xmath98 .
hence @xmath507 is also a boundary slope of @xmath12 .
again by the assumption that @xmath28 divides @xmath633 , together with lemma [ j has only 3 roots ] , we see that if @xmath634 is a solution of the system @xmath680 then @xmath636 . now applying lemma [ nonhyp surgery ]
, we see that @xmath681 is not a hyperbolic manifold . applying ( * ? ? ?
* theorem 2.0.3 ) and @xcite we see that @xmath681 must be a haken manifold and thus must be a toroidal manifold ( as it has finite first homology ) .
finally @xmath12 is one of the knots @xmath18 by theorem [ gl2 ] .
[ lmn - irr - rep ] the half - integral toroidal @xmath507-surgery on @xmath682 , @xmath683 , is a manifold with an irreducible @xmath33 representation @xmath658 whose image contains no parabolic elements .
we knew from section [ topological properties of j ] that @xmath548 , where @xmath529 is a @xmath150-torus knot exterior and @xmath539 is seifert fibred with base orbifold @xmath684 for some odd integer @xmath685 , such that the meridian slope @xmath543 of @xmath529 is identified with the seifert fibre slope @xmath541 of @xmath539 and the seifert fiber slope @xmath530 of @xmath529 is identified with a lens space filling slope @xmath686 of @xmath539 ( and the lens space has order @xmath540 ) .
perhaps it is easier to construct a @xmath618 representation @xmath687 of @xmath688 with the required properties .
as @xmath509 has zero @xmath689-homology , every @xmath618 representation of @xmath688 lifts to a @xmath33 representation . the representation @xmath687 will send @xmath601 to a cyclic group of order @xmath540 which is possible by factoring through @xmath690 .
so @xmath691 . we claim @xmath692 is not the identity element . for otherwise @xmath687 factors through the group @xmath693 which is the trivial group as @xmath694 is odd .
hence the order of @xmath692 is an odd number @xmath695 which is a factor of @xmath696 . on the @xmath529 side ,
we need to have @xmath697 and @xmath698 of order @xmath294 .
so @xmath687 factors through the triangle group @xmath699 such representation exists and can be required to be irreducible . also as at least one of @xmath700 and @xmath701 is odd and @xmath294 is odd
, we may require the image of @xmath687 containing no parabolic elements .
in fact the triangle group is either a spherical or a hyperbolic triangle group and so we may simply choose @xmath687 to be a discrete faithful representation of the triangle group into @xmath702 ( when the triangle group is spherical ) or into @xmath703 ( when the triangle group is hyperbolic ) , and thus the image of @xmath687 has no parabolic elements .
[ lmp - irr - rep]for any given @xmath704 , with @xmath705 , and with @xmath25 not divisible by @xmath524 , the half - integral toroidal surgery on @xmath3 is a manifold with an irreducible @xmath33 representation @xmath658 whose image contains no parabolic elements .
the proof is similar to that of lemma [ lmn - irr - rep ] . we knew @xmath548 , where @xmath539 is a @xmath546-torus knot exterior and @xmath529 is seifert fibred with base orbifold @xmath706 , such that the meridian slope @xmath686 of @xmath539 is identified with the seifert fibre slope @xmath530 of @xmath529 and the seifert fiber slope @xmath541 of @xmath539
is identified with a lens space filling slope @xmath543 of @xmath529 ( and the lens space has order @xmath524 ) . as in lemma
[ lmn - irr - rep ] , we jus need to construct a @xmath618 representation @xmath687 of @xmath688 which is irreducible and whose image contains no parabolic elements .
the representation @xmath687 will send @xmath609 to a cyclic group of order @xmath524 which is possible by factoring through @xmath707 .
so @xmath708 . we claim @xmath709 is not the identity element . for otherwise @xmath687 factors through the group @xmath710 which is a cyclic group of order less than @xmath524 since @xmath25 is not divisible by @xmath26 by our assumption .
hence the order of @xmath692 is an odd number @xmath695 which is a factor of @xmath26 . on the @xmath539 side ,
we need to have @xmath711 and @xmath712 of order @xmath294 .
so @xmath687 factors through the triangle group @xmath713 such representation exists and can be required to be irreducible .
also as both @xmath700 and @xmath294 are odd , we may require the image of @xmath687 containing no parabolic elements .
[ ( l , m , n ) is different from ( l,m ) ] the @xmath1-polynomial of any @xmath714 , @xmath683 , does not divide the @xmath1-polynomial of any @xmath715 .
suppose otherwise that @xmath716 . then by proposition [ same r slope ] , @xmath3 and @xmath717 have the same half - integral toroidal slope @xmath620 , @xmath718 odd .
let @xmath658 be an irreducible representation of @xmath509 provided by lemma [ lmn - irr - rep ] . then @xmath719 but @xmath664 . we know that @xmath659 is contained in a positive dimensional component @xmath529 of @xmath619 .
let @xmath55 be an irreducible curve in @xmath529 containing @xmath659 .
@xmath720 is not constant on @xmath55 .
otherwise @xmath720 is constantly equal to @xmath721 on @xmath55 .
if @xmath722 is not a constant on @xmath55 , then @xmath55 would be a semi - norm with @xmath507 as its associated boundary slope , which is impossible by theorem [ properties of a semi - norm curve ] as @xmath723 is not a boundary slope and @xmath724 .
so @xmath722 is a constant not equal to @xmath721 on @xmath55 since @xmath664 and is not parabolic .
so for any point @xmath253 , @xmath725 can not be a parabolic element ( for otherwise @xmath726 is also parabolic and thus @xmath727 ) .
so @xmath728 for any @xmath253 .
we now get a contradiction with ( * ? ? ? * proposition 4.10 ) , which proves the claim .
so @xmath720 is not constant on @xmath55 which means that the component @xmath729 belongs to @xmath638 and thus contributes a factor in @xmath633 .
moreover the point @xmath659 contributes a root @xmath634 to the system @xmath730 such that @xmath731 .
as @xmath716 , @xmath634 is also a solution of the system @xmath732 which contradicts lemma [ j has only 3 roots ] . with a similar proof replacing lemma [ lmn - irr - rep ] by lemma [ lmp - irr - rep ] ,
we have [ ( l , m , p ) is different from ( l,m)]the @xmath1-polynomial of any @xmath24 , @xmath705 , @xmath25 not divisible by @xmath26 , does not divide the @xmath1-polynomial of any @xmath733 .
[ non - small seifert surgery ] let @xmath12 be a hyperbolic knot in @xmath6 . for a given slope @xmath296 , @xmath379 odd , if every solution @xmath634 of the system of equations @xmath653 has @xmath654 , then @xmath655 can not be a seifert fibred space whose base orbifold is a @xmath78-sphere with exactly three cone points .
the proof is similar to that of lemma [ nonhyp surgery ] .
suppose otherwise that @xmath655 is a a seifert fibred space whose base orbifold is a @xmath78-sphere with exactly three cone points . as @xmath379 is odd , the base orbifold is either spherical or hyperbolic .
hence @xmath656 has an irreducible @xmath618 representation @xmath657 which factors through the orbifold fundamental group of the base orbifold such that the image of @xmath687 contains no parabolic elements . as @xmath379 is odd , the @xmath618 representation @xmath657 lifts to a @xmath33 representation @xmath658 .
it is well known that the character @xmath659 of @xmath658 is an isolated point in @xmath660 .
considered as a point in @xmath85 , @xmath659 is contained in a positive dimensional component @xmath55 of @xmath85 .
arguing exactly as in claim [ claim : non - constant ] we have @xmath665 is non - constant on @xmath55 .
hence @xmath55 is contained @xmath86 and contributes a factor @xmath98 to @xmath28 .
exactly as in the proof of lemma [ nonhyp surgery ] , the point @xmath659 provides a solution @xmath639 to the system @xmath678 such that @xmath734 , giving a contradiction with the assumption of the lemma .
the goal of this section is to prove the following proposition : [ prop : l-100 ] suppose that two knots @xmath14 and @xmath24 have the same genus @xmath735 and the same half - integral toroidal slope @xmath507 , where @xmath15 and @xmath736 , then @xmath737 .
let @xmath738 for @xmath24 , by ( [ eq : rslope ] ) , @xmath739 when @xmath505 , using ( [ eq : genus0p ] ) , we get @xmath740 where @xmath741 consider the family of knots @xmath14 , where @xmath15 . in this case , @xmath742 are given by @xmath743 so @xmath744 [ lem : p<0 ] suppose that @xmath745 and @xmath736 , then the knot @xmath24 has different @xmath746 from the knot @xmath14 , where @xmath15 .
otherwise , assume @xmath24 has the same @xmath746 as @xmath14 . using ( [ eq : d ] ) , we see that @xmath747using ( [ eq : s0p ] ) and ( [ eq : lmcondition ] ) , we have @xmath748using ( [ eq : sd ] ) and the previous two inequalities , we get @xmath749 so @xmath750 , hence @xmath751 . if @xmath350 or @xmath752 , @xmath753 . by ( [ eq : s0p ] ) ,
@xmath754 , which contradicts ( [ eq : ds-100 ] ) . if @xmath755 , then @xmath756 . by ( [ eq : s0p ] ) ,
it follows from ( [ eq : sd ] ) that @xmath758 is a quadratic residue modulo @xmath759 , which is not true .
if @xmath760 , then @xmath761 and @xmath762 .
it follows from ( [ eq : sd ] ) that @xmath78 is a quadratic residue modulo @xmath758 , which is not true . for the family of knots @xmath763
, the following proposition expresses @xmath764 in terms of @xmath474 and @xmath718 .
[ prop : lmformula ] for the family of knots @xmath763 , the pair @xmath765 is determined by @xmath474 and @xmath718 by the following formulas : if @xmath766 , then @xmath767 if @xmath768 , then @xmath769 using ( [ eq : d ] ) , we can express @xmath770 as a linear function of @xmath718 and @xmath25 .
substitute such expression of @xmath770 into ( [ eq : s0p ] ) , we get @xmath771 if @xmath766 , by ( [ eq : lquad ] ) , @xmath25 is a root of the quadratic polynomial @xmath772 whose two roots are @xmath773 by ( [ eq : d ] ) , @xmath774 and @xmath775 , so @xmath776 using ( [ eq : d ] ) again , we can compute @xmath410 as in the statement . if @xmath768 , @xmath25 is a root of the quadratic polynomial @xmath777 whose two roots are @xmath778 by ( [ eq : d ] ) , @xmath774 and @xmath779 , so @xmath780 using ( [ eq : d ] ) again , we can compute @xmath410 as in the statement .
[ lem : lm>0 ] suppose that two knots @xmath763 and @xmath781 have the same @xmath735 and @xmath507 .
suppose further that @xmath782 then the quadruple @xmath783 is either @xmath784 or @xmath785 for @xmath786 .
using proposition [ prop : lmformula ] and ( [ eq : lmcondition ] ) , we get @xmath787 and @xmath788 hence @xmath789 which implies @xmath790 from ( [ eq : d ] ) and ( [ eq : lmcondition ] ) , we can conclude that @xmath791 by ( [ eq : l_ge2 ] ) , @xmath792 is a perfect square which has the same parity as @xmath793 , hence @xmath794 for some @xmath795 .
then @xmath796 using ( [ eq:4s ] ) , we also get @xmath797 by ( [ eq : lle-2 ] ) , @xmath798 is a perfect square with the same parity as @xmath799 , so @xmath800 for some @xmath801
. then @xmath802 compare ( [ eq:2c ] ) and ( [ eq:2c ] ) , we get @xmath803 moreover , since both @xmath804 and @xmath805 are integers , @xmath806 is an integer , so @xmath807 if @xmath808 , then @xmath809 . by ( [ eq : dcc ] ) ,
@xmath810 , a contradiction to ( [ eq : d_ge8 ] ) .
if @xmath811 , using ( [ eq : dcc ] ) we get @xmath812 .
hence @xmath813 , @xmath814 .
if @xmath815 , using ( [ eq : dcc ] ) we get @xmath816 . using ( [ eq : d_ge8 ] ) , the only possible case is @xmath817 , and @xmath818 .
when @xmath745 , this result follows from lemma [ lem : p<0 ] . when @xmath390 , from the formula ( [ eq : s0p ] ) it is easy to see @xmath819 for @xmath24 , but @xmath820 for @xmath14
. now we consider the case @xmath544 . by proposition [ prop : lmformula ] , for any given @xmath746 , there are at most three knots @xmath821 , @xmath822 , having this @xmath746 .
there is at most one pair @xmath823 in each of the three cases : * @xmath824 * @xmath825 * @xmath826 the pair @xmath827 is in the third case . by proposition [ prop : dupl ] ( c ) , @xmath828 , and
@xmath829 is in the second case .
suppose that there is also a pair @xmath765 in the first case with the same @xmath746 , then lemma [ lem : lm>0 ] implies that @xmath830 and @xmath831 . by proposition [ prop : dupl ] ( d ) ,
the three knots @xmath832 are the same .
recall that the unique half - integral toroidal slope of @xmath14 is @xmath833 , where @xmath834 .
hence the family of knots @xmath835 are mutually distinct .
suppose @xmath12 is hyperbolic . then by lemma [ same r slope ] , @xmath12 has the same half - integral toroidal slope @xmath507 as @xmath21 and @xmath12 is one of @xmath18 .
applying lemma [ ( l , m , n ) is different from ( l,m ) ] , we have @xmath545 or @xmath29 . since the half - integral toroidal slope of @xmath836 is positive while the half - integral toroidal slope of @xmath23 is negative , we see that @xmath240 must be zero . similarly applying lemma [ ( l , m , p ) is different from ( l,m ) ]
, we have @xmath379 is non - positive , and @xmath26 divides @xmath25 .
that is , we have @xmath837 with @xmath27 .
now by proposition [ prop : l-100 ] , we have @xmath22 .
suppose otherwise that @xmath838 contains a factor of the form @xmath839 or @xmath840 . as this factor is irreducible and balanced , it is contributed by a curve component @xmath55 in @xmath842 .
moreover @xmath55 is a semi - norm curve with @xmath843 as the unique associated boundary slope .
we claim that @xmath843 is either @xmath844 or @xmath845 . from theorem [ properties of a semi - norm curve ] ( 3 ) and lemma [ meridian not bdy slope ]
, we see that the meridian slope @xmath623 has the minimal semi - norm @xmath146 .
to prove the claim , we just need to show , by theorem [ properties of a semi - norm curve ] ( 4 ) , that for the half - integral toroidal slope @xmath507 of @xmath21 , we have @xmath846 , which is equivalent to show that @xmath847 for every @xmath135 . as @xmath848 has no non - cyclic representation by lemma [ prop : abelrep ] , at every regular point @xmath849 we have @xmath850 . if at an ideal point @xmath626 of @xmath129 , @xmath851 , then @xmath627 is finite for every class @xmath122 in @xmath852 ( because both @xmath853 and @xmath854 are finite at @xmath626 ) . we can now derive a contradiction with ( * ? ? ?
* proposition 4.12 ) just as in the proof of lemma [ r is a vertex ] , and the claim is thus proved . on the other hand it is shown in ( * ? ? ? * proposition 5.4 ) that each of @xmath855 and @xmath856 is a small seifert fibred space . as @xmath21 is a small knot ,
each of @xmath844 and @xmath845 can not be a boundary slope by ( * ? ? ?
* theorem 2.0.3 ) and thus can not be the slope @xmath843 .
we arrive at a contradiction .
since @xmath21 is a fibered knot by @xcite , @xmath12 is also fibered .
hence if @xmath857 is any pair of companion knot and pattern knot associated to @xmath12 , then each of @xmath185 and @xmath186 is fibred and the winding number @xmath169 of @xmath186 with respect to @xmath185 is larger than zero . as is true for any satellite knot
, @xmath12 has a companion knot @xmath185 which is either a torus knot or a hyperbolic knot .
so we just need to rule out the possibility that @xmath185 be a torus knot .
if @xmath185 is a @xmath107-torus knot , then by formula ( [ apoly of t(p , q ) ] ) , @xmath421 contains a factor of the form @xmath858 or @xmath859 for some integer @xmath860 and @xmath861 . as the winding number @xmath169 of the pattern knot @xmath186 with respect to @xmath185 is non - zero , by proposition [ non - zero winding ]
, @xmath28 contains a factor of the form @xmath862 or @xmath863 . but this contradicts lemma [ a_j has no seminorm factor ] . [ surgery on c ] for any hyperbolic knot @xmath185 in @xmath6 , any surgery with a slope @xmath864 , @xmath865 , satisfying @xmath866 and @xmath867 odd , will produce either a hyperbolic manifold or a seifert fibred space whose base orbifold is @xmath868 with exactly three cone points . as @xmath866
, @xmath869 is irreducible @xcite and atoroidal @xcite .
thus @xmath869 is either a hyperbolic manifold or an atoroidal seifert fibred space . in latter case ,
the seifert fibered space has non - cyclic fundamental group @xcite . as @xmath379 is odd
, the base orbifold of the seifert fibered space can not be non - orientable .
thus the base orbifold is a @xmath78-sphere with exactly three cone points .
suppose otherwise .
let @xmath872 be the rational number @xmath873 in its reduced form , i.e. @xmath874 and @xmath875 .
then @xmath876 and @xmath877 is odd .
so by lemma [ surgery on c ] , the surgery on @xmath185 with the slope @xmath872 will yield either a hyperbolic manifold or a seifert fibred space whose base orbifold is a @xmath78-sphere with exactly three cone points .
applying either lemma [ nonhyp surgery ] or lemma [ non - small seifert surgery ] , we see that the @xmath1-polynomial @xmath878 of @xmath185 has a zero point @xmath879 such that @xmath880 and @xmath881 . now from proposition [ non - zero winding ] and its proof
, we see that @xmath878 can be extended to a factor @xmath232 of @xmath28 and the variables of @xmath878 and @xmath232 satisfying the relations : @xmath882 , @xmath883 .
in particular for some @xmath639 we have @xmath884 , @xmath885 , and @xmath634 is a zero point of @xmath232 .
obviously @xmath886 . from @xmath887
, we have @xmath888 , i.e. @xmath889 .
if @xmath897 or @xmath896 , it implies that there is a curve component @xmath55 in @xmath898 such that @xmath129 has a point at which @xmath899 but @xmath900 .
this is impossible as @xmath901 has the minimal zero degree at every point of @xmath129 ( because @xmath21 is a small knot , @xmath893 is a cyclic surgery slope but is not a boundary slope ) .
the proof is similar to that of lemma [ w divides 2s-1 ] , only replacing lemma [ j has only 3 roots ] by lemma [ s-1 has only 2 roots ] .
first note that @xmath893 is an odd number ( as @xmath834 is even ) .
so if @xmath893 is not divisible by @xmath871 , then the reduced form @xmath872 of the rational number @xmath902 has denominator @xmath876 . now arguing as in the proof of lemma [ w divides 2s-1 ] starting from the @xmath872-surgery on @xmath185
, we see that the system @xmath894 has a solution @xmath639 with @xmath903 .
this gives a contradiction with lemma [ s-1 has only 2 roots ] .
it follows from corollary [ w is one ] and proposition [ non - zero winding ] that @xmath421 is a factor of @xmath890 .
so proposition [ same r slope ] says that @xmath507 is also a toroidal slope of @xmath185 and @xmath185 is one of @xmath18 . now arguing as in the case when @xmath12 is hyperbolic , we see that @xmath185 is one of @xmath24 with @xmath379 non - positive and with @xmath27 .
if @xmath908 , then by ( * ? ? ?
* theorem 2.1 ( d ) ) the @xmath909-surgery on @xmath904 is a seifert fibred manifold whose base orbifold is a @xmath78-sphere with exactly three cone points , except when @xmath185 is one of @xmath910 , @xmath911 , @xmath912 , @xmath913 , @xmath914 .
as we knew that @xmath379 is non - positive and @xmath26 divides @xmath25 , these exceptional cases can be excluded except @xmath911 or @xmath915 .
so we just need to deal with the case when the @xmath909-surgery on @xmath185 is a seifert fibred manifold whose base orbifold is a @xmath78-sphere with exactly three cone points .
note that @xmath893 is odd .
hence by lemma [ non - small seifert surgery ] , the system @xmath916 has a solution @xmath639 with @xmath903 .
as @xmath421 is a factor of @xmath838 , the point @xmath639 is also a solution of the system @xmath894 which yields a contradiction with lemma [ s-1 has only 2 roots ] .
recall that @xmath185 is one of @xmath918 , @xmath379 non - positive , @xmath919 , @xmath915 , and has the same @xmath507 slope as @xmath21 and thus has the same @xmath474 slope as @xmath21 .
from ( [ eq : rslope ] ) and ( [ eq : genus0p ] ) we have that for @xmath918 , @xmath379 non - positive , its @xmath474 slope and genus @xmath735 are given by @xmath920 for @xmath911 its @xmath474 and @xmath735 values are @xmath921 and for @xmath915 , @xmath922 in each case , one can check directly that @xmath923 holds , being aware of some forbidden values on @xmath924 given by ( [ eq : lmcondition ] ) . as noted in the proof of theorem [ main1 ] ,
when the winding number @xmath449 , the pattern knot @xmath186 is a nontrivial knot .
we also have , by lemma [ p - factor ] , that @xmath454 is a factor of @xmath890 . combining this fact with ( [ apoly of t(p , q ) ] ) and lemma [ a_j has no seminorm factor ] we know that @xmath186 can not be a torus knot .
so @xmath186 is either a hyperbolic knot or a satellite knot .
if @xmath186 is a hyperbolic knot , then arguing as in the proofs of lemmas [ c and j have the same r ] , [ s-1 is cyclic for c ] and [ less than 4 g ] , we have that @xmath186 has the same half - integral toroidal slope as @xmath21 , that @xmath186 is one of @xmath925 with @xmath926 non - positive or @xmath911 or @xmath927 , and that @xmath928 . now from @xmath929 we have that the genus of the given satellite knot @xmath12 ( which is equal to that of @xmath21 ) is equal to the sum of the genus of @xmath185 and the genus of @xmath186 .
so the genus of one of @xmath185 and @xmath186 , say @xmath185 ( same argument for @xmath186 ) , is less than or equal to the half of the genus of @xmath21 , i.e. @xmath930 so @xmath931 but @xmath834 and @xmath932 , which do not fit in the inequality ( [ s inequality ] ) .
this contradiction shows that @xmath186 can not be hyperbolic .
so @xmath186 is a satellite knot .
let @xmath933 be a pair companion knot and pattern knot for @xmath186 .
once again as @xmath186 is fibred , each of @xmath455 and @xmath456 is fibred , and the winding number @xmath457 of @xmath456 with respect to @xmath455 is larger than zero . making use of the fact that @xmath934
, one can show , similarly as for the pair @xmath857 , that @xmath455 can be assumed to be hyperbolic , that @xmath935 , that @xmath455 has the same @xmath507 and @xmath474 values as @xmath21 , that @xmath936 for some non - positive @xmath937 or @xmath911 or @xmath938 , and that @xmath939 .
now from the equality @xmath940 we see that one of @xmath941 and @xmath942 is less than or equal to @xmath943 .
this leads to a contradiction just as in the preceding paragraph .
so @xmath186 can not be a satellite knot , and this final contradiction completes the proof of theorem [ main2 ] . | we say that a given knot @xmath0 is detected by its knot floer homology and @xmath1-polynomial if whenever a knot @xmath2 has the same knot floer homology and the same @xmath1-polynomial as @xmath3 , then @xmath4 . in this paper
we show that every torus knot @xmath5 is detected by its knot floer homology and @xmath1-polynomial .
we also give a one - parameter family of infinitely many hyperbolic knots in @xmath6 each of which is detected by its knot floer homology and @xmath1-polynomial .
in addition we give a cabling formula for the a - polynomials of cabled knots in @xmath6 , which is of independent interest .
in particular we give explicitly the a - polynomials of iterated torus knots .
v c u _ sl(2,c ) ps .
psl(2,c ) _ |
the extragalactic magnetic field ( egmf ) is of great interest to the overall understanding of astrophysical magnetic fields and related processes .
it could act as a seeding field for magnetic fields in galaxies and clusters @xcite , and its origin may be related to inflation or other periods in the early history of the universe @xcite .
faraday rotation measurements @xcite and analysis of cobe data anisotropy @xcite have put an upper bound on the egmf strength at @xmath5 gauss . on the other hand , gamma - ray - initiated electromagnetic cascades deflected by the egmf in intergalactic space have a characteristic angular spread @xcite and time delay @xcite , both of which provide a probe for lower egmf strengths @xcite .
present and next generation gamma - ray telescopes have the possibility to measure the egmf strength by observing the angular and temporal distributions of cascade photons from extragalactic gamma - ray sources such as blazars @xcite . with current very - high - energy ( vhe , @xmath0 gev ) and high - energy ( he , 100 mev @xmath6 gev ) gamma - ray data on vhe - selected blazars , a lower limit on the egmf strength
can be placed by requiring that the cascade flux of vhe emission not exceed the measured flux or upper limit in the he band @xcite .
analytic cascade models assuming a simple relationship between the cascade flux and egmf strength have put the lower limit at @xmath7 to @xmath8 gauss when the source livetime is unlimited @xcite , similar to the results of monte carlo simulations @xcite .
if the cascade time delay is limited to the @xmath9 years of simultaneous he and vhe observations , the lower limit becomes @xmath10 to @xmath11 gauss according to the simple cascade models @xcite , or @xmath11 to @xmath12 gauss according to the simulations @xcite . in this letter
, we present a new semi - analytic model of the electromagnetic cascade .
in contrast to previous analytic cascade models @xcite , our cascade model considers the full track of the primary photon without the assumption of interacting exclusively at the mean free path .
this simultaneously accounts for both angular and temporal constraints in a natural way .
in addition , we model the radiation backgrounds and source emission in greater detail . as a complementary approach to full monte carlo simulations @xcite
, our model serves as a tool for clarifying the cascade picture , rapidly searching through the parameter space , and interpreting simulation results .
we also develop a systematic framework for applying our cascade model s predictions to derive lower limits on the egmf strength at specific confidence levels .
this framework is applicable to the results of monte carlo simulations as well .
consider a blazar emitting gamma rays at distance @xmath13 .
gamma rays emitted at an angle @xmath14 relative to the line of sight may produce an @xmath15 pair via absorption on the photon background @xcite at a distance @xmath16 from the source , not necessarily equal to the mean free path . after being deflected by the egmf through the angle @xmath17 , the pairs could scatter background photons to gamma - ray energies , redirecting them toward the observer .
these secondary gamma rays would arrive at an incidence angle @xmath18 ( see fig .
[ fig : geometry ] ) . in this picture , the influence of the egmf enters solely through @xmath17 .
the angles @xmath18 and @xmath14 are uniquely specified in terms of @xmath17 , @xmath13 , and @xmath16 , provided @xmath19 .
because the energy density of cmb photons far exceeds that of the extragalactic background light ( ebl ) , we assume that inverse compton scattering proceeds in the thomson regime via @xmath15 interactions with the cmb exclusively .
an electron with lorentz factor @xmath20 will on average scatter secondary photons to energy @xmath21 , where @xmath22 mev is the average cmb photon energy @xcite .
the energy loss rate of the electron is @xmath23 where @xmath24 is the speed of light , @xmath25 is the cmb photon density , and @xmath26 is the thomson cross section . for an egmf @xmath27 perpendicular to the electron momentum @xmath28 ,
the deflection rate in terms of the larmor radius @xmath29 is @xmath30 thus , when an electron s lorentz factor has changed from @xmath31 to @xmath20 , the electron will have been deflected by @xmath32 however , if the angle @xmath33 between @xmath27 and @xmath28 is other than @xmath34 , eq . [ eq : defangle ] should be generalized to @xmath35 we use the full cmb black - body spectrum to determine the observed cascade spectrum .
the number of secondary photons between energies @xmath36 and @xmath37 produced by an electron that changes lorentz factor from @xmath38 to @xmath20 is the number density of cmb photons between @xmath39 and @xmath40 times the thomson cross section and the distance travelled by the electron : @xmath41 where @xmath42 is the planck constant , @xmath43 the boltzmann constant , and @xmath44 k the cmb temperature . in terms of the mean free path
@xmath45 inferred from the optical depth @xmath46 of the ebl , the probability for a photon of energy @xmath47 to be absorbed between @xmath16 and @xmath48 is @xmath49 . approximating both particles in the resulting pair to have initial energy @xmath50
, we calculate the differential flux of observed secondary photons by integrating over @xmath16 , @xmath47 , and @xmath20 , and averaging over @xmath33 : @xmath51 here , @xmath52 is the probability distribution of @xmath33 , equal to @xmath53 for an egmf uniformly distributed in direction , and @xmath54 is the intrinsic flux of the source , with @xmath55 from fig .
[ fig : geometry ] .
the @xmath56 factor from the @xmath57 being deflected into the surface of a cone with opening angle @xmath17 cancels the @xmath58 enhancement from a similar effect at the source .
we take the integral over @xmath33 from @xmath59 to @xmath34 .
the physical lower bound for the integration over the primary energy @xmath47 is @xmath60 .
nearly all of the photons above 200 tev will be absorbed within 1 mpc of the source , and the @xmath15 pairs will be isotropized by the strong field of the surrounding galaxy , resulting in negligible cascade contribution .
we therefore adopt an upper limit of @xmath61 tev , suggesting an upper limit of @xmath62 on the @xmath20 integration . as a practical matter
, we enforce a lower limit of @xmath63 on the @xmath20 integration , motivated by the cmb density becoming negligible at energies above 3 mev and our disinterest in the cascade spectrum below 100 mev .
observational effects enter eq .
[ eq : model ] through limits on the @xmath16 integration . as seen in fig .
[ fig : geometry ] , we can express @xmath18 as @xmath64 so that a cut on @xmath18 translates directly into a cut on @xmath16 .
similarly , the time delay @xmath65 of cascade photons may be written as @xmath66 exchanging a limit on the source livetime @xmath65 for another constraint on @xmath16 .
we adopt the intersection of the @xmath16 cuts from eqs .
[ eq : angle ] and [ eq : time ] in evaluating the cascade flux via eq .
[ eq : model ] .
we now briefly examine several assumptions made in the construction of this cascade model .
( i ) exact energy distributions of pair production products are approximated as each having half the energy of the primary photon , as the cascade spectrum only weakly depends on the spectral distribution of pairs @xcite .
( ii ) the thomson limit assumption , appropriate when @xmath67 @xmath68 @xmath69 tev , certainly produces spectra valid for the range of existing tev gamma - ray instruments . (
iii ) we assume the egmf to be coherent over the cooling length of the electrons , at most a few mpc for the most energetic electrons , making this assumption valid for coherence lengths @xmath70 1 mpc .
( iv ) only secondary cascade photons are considered .
for observed photons above 100 mev , the primary energy must be @xmath71 gev .
the requirement for third - generation photons is thus @xmath72 tev . while a primary photon above this energy does produce higher - generation cascade photons , the power going into the higher generations is small for conventional blazar emission models , leading to negligible contribution from higher - order cascades @xcite .
( v ) cosmological effects enter the cascade model solely in the calculation of @xmath73 .
cosmic expansion , energy redshift , ebl evolution , and other cosmological effects are ignored , limiting the application of the cascade model to nearby sources ( @xmath74 ) , considering the @xmath75 radiation density evolution . ( vi )
we assume axial symmetry in the intrinsic emission @xmath54 , following the current understanding of blazar emission , that is , boosted isotropic emission from hot blobs with the jet direction pointing toward the earth @xcite .
( vii ) the blazar intrinsic emission should be steady over the time interval @xmath65 . to get the predicted cascade flux from eq .
[ eq : model ] , we require a source emission model @xmath54 . motivated by models of blazars as relativistically beamed sources @xcite we model the intrinsic emission as boosted isotropic emission , @xmath76 that is , a power law of index @xmath77 boosted by the lorentz factor @xmath78 with an exponential cutoff energy @xmath79 . the second term in eq .
[ eq : boost ] models the blazar counter jet , which we find does not significantly impact our result . to get a conservative estimate for the cascade flux
, we employ the ebl model from @xcite which is relatively transparent for vhe gamma rays . having no prior assumptions on the egmf structure , we take @xmath80 to calculate the cascade flux in eq . [
eq : model ] .
we demonstrate the cascade model by investigating rgb j0710 + 591 , a high - frequency - peaked bl lacertae ( hbl ) object located at a redshift of @xmath81 , for which simultaneous he and vhe data are available with no variability observed @xcite . in the vhe regime , we take the spectrum reported by @xcite , while to get the he spectrum we analyze publicly available _ fermi _ large area telescope ( lat ) data taken from a @xmath82-year period between 2008 august and 2011 january , using the _ fermi _ science tools v9r18p6 and the p6_v3_diffuse instrument response functions ( irfs ) , with models gll_iem_v02 for the galactic diffuse and isotropic_iem_v02 for the isotropic background .
accounting for nearby point sources , we perform an unbinned likelihood analysis to get the spectrum between 100 mev and 300 gev , finding a best - fit index of @xmath83 .
next , we bin the data into five energy bins , demanding that each bin have a test statistic greater than 10 and a maximum relative uncertainty on the flux of @xmath84 , as suggested by @xcite .
the combined spectra appear in fig .
[ fig : spectrumexample ] , demonstrating that the _ fermi _ spectral points are consistent with the 68% confidence band from the unbinned likelihood analysis . to compare with the measured spectra
, we compute the total flux as the sum of the cascade flux calculated by eq .
[ eq : model ] and the direct flux @xmath85 , and then normalize it to observational data .
for example , the results for @xmath86 , @xmath87 tev , @xmath88 , and a range of different egmf strengths appear in fig . [
fig : spectrumexample](a ) for a blazar with unlimited livetime , and in fig .
[ fig : spectrumexample](b ) for a blazar active for only 3 years . in both cases
we require @xmath18 to be smaller than the 68% containment radius of the point spread function ( psf ) @xcite in the _ fermi _ range , or smaller than the @xmath89 cut @xcite in the vhe range . setting the simple requirement that the flux be lower than the measured spectrum , we observe from fig .
[ fig : spectrumexample ] that the lower limit on the egmf strength should be between @xmath7 and @xmath8 gauss for the infinite - time case , and between @xmath11 and @xmath12 gauss for the 3-year case , consistent with results shown by @xcite for rgb j0710 + 591 . to be more rigorous , we fit the predicted flux to the data and use the resulting @xmath90 values to constrain the egmf strength .
[ fig : fitexample](a ) shows a sample fit for @xmath86 , @xmath87 tev , @xmath88 , and @xmath91 gauss . because we can not exclude the existence of additional components contributing to the blazar emission ( e.g. , *
* ) and modifying the intrinsic spectrum of eq .
[ eq : boost ] , we also fit a broken power law with an index @xmath92 below 80 gev , as shown for instance in fig .
[ fig : fitexample](b ) .
we find that the additional free parameter does not greatly affect our constraints on the egmf . versus egmf strength @xmath93 for various livetimes for rgb j0710 + 591 .
lower limits on @xmath93 are indicated at 90% and 95% confidence levels for the 1-year and unlimited cases but omitted at intermediate times for clarity . ] for each egmf strength @xmath93 , we find the best - fit model of the combined cascade and intrinsic emission for a wide range of cutoff energies @xmath94 tev , 100 tev ) , allowing @xmath77 and @xmath92 to vary between 2.5 and the physically motivated constraint of 1.5 ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and fixing @xmath95 to a typical value of 10 .
we plot the @xmath90 value of the best - fit model as a function of @xmath93 in fig . [
fig : time_cuts ] for blazar livetimes from 1 year to @xmath96 years , and for the unlimited case . at low @xmath93 ,
the @xmath90 values converge because the cascade arrives promptly and at small angles . as @xmath93 increases ,
the arrival angles and times of the cascade begin to spread out , diminishing the observed emission and providing a better fit to the observed data . the convergence of the curves in fig .
[ fig : time_cuts ] to the infinite - livetime curve is easily understood .
combining eq .
[ eq : angle ] with eq .
[ eq : time ] and assuming small angles , we can translate a cut on the instrument angle @xmath18 into a time cut @xmath97 : @xmath98 for example , source photons at 1 tev , which have a mean free path of @xmath99 mpc , will produce cascade photons of energy @xmath100 gev , for which an angular cut of @xmath101 is appropriate for the _ fermi _ lat . at the distance of rgb j0710 + 591 ( @xmath102 mpc ) , this translates into a time cut of @xmath103 years .
this is close the the livetime of @xmath104 years at which the @xmath90 curves in fig .
[ fig : time_cuts ] begin to converge to the unlimited - livetime case , becoming nearly indistinguishable at @xmath105 years . if the blazar livetime is smaller , the livetime cut outweighs the angle cut , and the position of the @xmath90 curve depends on the blazar livetime
. for longer livetimes , the angle cut becomes more constraining and the curves converge to the unlimited - livetime case . agree with the curves for small times but begin to diverge for livetimes between @xmath106 and @xmath107 years , as expected . at longer times , the instrument psf limits the possible delay time of the photons , causing the constraint to level off . ]
we constrain the egmf strength for a given livetime by finding the point at which @xmath90 exceeds its minimum value by @xmath108 for each curve in fig .
[ fig : time_cuts ] .
two sample confidence levels ( 90% and 95% ) , corresponding to @xmath108 values of 2.72 and 3.84 ( see e.g. , * ? ? ?
* ) are indicated in the figure .
the limits derived from these confidence levels are shown in fig .
[ fig : confidence_limits ] as a function of the blazar livetime . for livetimes below @xmath109 years
, the limit on the egmf strength scales with the blazar livetime @xmath65 as @xmath110 , as expected from combining eq .
[ eq : deflection ] and eq .
[ eq : timecut ] .
this relation breaks down when the time limit enforced by eq .
[ eq : timecut ] is smaller than the blazar livetime .
our limit at 95% confidence is @xmath111 gauss if the blazar s livetime is infinite and @xmath112 gauss if the livetime is 3 years , in agreement with the results of @xcite . in all cases ,
we reject the hypothesis of zero egmf at greater than @xmath113 confidence .
the cascade model we present builds upon previous models to give a more complete semi - analytic treatment , accounting for the photon trajectories , background spectra , and intrinsic emission in sufficient detail to produce realistic cascade predictions for different blazar livetimes . for example , the spectra shown in fig .
[ fig : spectrumexample ] converge to the zero field case at high energies and become suppressed by the field approximately as @xmath114 at low energies ( @xmath115 gev ) , in agreement with the spectra produced by a full monte carlo simulation as presented in @xcite .
previous cascade models @xcite overestimate the suppression of the cascade flux in the egmf compared to our result . combined with a systematic framework for interpreting the simultaneous fit of he and vhe spectra to predictions ,
our model derives robust lower limits on the egmf strength .
the application of the model to rgb j0710 + 591 data yields lower limits at 95% confidence that agree with the results of @xcite for the same source , further indicating that the physical picture of the cascade presented by the model is sufficiently complete to produce accurate results within the limits discussed in sec .
[ sec : model ] .
the scaling of the lower limits with source livetime , depicted in fig .
[ fig : confidence_limits ] , demonstrates that our analysis framework is also physically meaningful .
the limits are conservative because we choose a low ebl model , as well as for a few other reasons that we now discuss .
we assume that the egmf is configured as coherent domains of the same field strength and random field orientation , neglecting domain crossing by the @xmath57 pairs . in the energy range of interest , this restricts the domain size @xmath116 to be @xmath117 mpc .
if @xmath118 mpc instead , domain crossing will result in smaller deflection angles for pairs @xcite and a higher bound on the field strength than we report .
in testing various models of the intrinsic emission , we only consider cutoff energies below 100 tev and spectral indices softer than 1.5 . while it is possible to achieve intrinsic emission that is even harder by invoking some unconventional mechanism ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the he cascade flux could only be higher in that case , leading to a more stringent lower limit .
the spectral index could also be further constrained using multi - wavelength information ( e.g. , * ? ? ?
* ) , which depends on detailed blazar modeling and hence is beyond the scope of this letter .
the p6_v3_diffuse irfs used in the cascade model integration may underestimate the psf at large energies @xcite , but fig .
[ fig : spectrumexample ] shows that the most constraining part of the he spectrum for rgb j0710 + 591 is the low - energy region where this effect is small , and , as discussed in sec .
[ sec : app ] , it can only affect the cases with livetimes larger than @xmath106 years .
furthermore , we expect the irfs we use to underestimate the cascade flux , so our lower limit is still conservative . in the future
, there will be more realistic irfs for the _ fermi _ lat , as well as more blazars with simultaneous he - vhe baseline data available , and the lower limit could be further improved using the framework we present .
the computations used in this work were performed on the joint fermilab - kicp supercomputing cluster , supported by grants from fermilab , kavli institute for cosmological physics , and the university of chicago . | very - high - energy ( vhe , @xmath0 gev ) gamma rays emitted by extragalactic sources , such as blazars , initiate electromagnetic cascades in the intergalactic medium .
the cascade photons arrive at the earth with angular and temporal distributions correlated with the extragalactic magnetic field ( egmf ) .
we have developed a new semi - analytical model of the cascade properties which is more accurate than previous analytic approaches and faster than full monte carlo simulations . within its range of applicability , our model can quickly generate cascade spectra for a variety of source emission models , egmf strengths , and assumptions about the source livetime . in this letter
, we describe the properties of the model and demonstrate its utility by exploring the gamma - ray emission from the blazar rgb j0710 + 591 . in particular , we predict , under various scenarios , the vhe and high - energy ( he , 100 mev @xmath1 e @xmath1 300 gev ) fluxes detectable with the veritas and _ fermi _ large area telescope ( lat ) observatories . we then develop a systematic framework for comparing the predictions to published results , obtaining constraints on the egmf strength . at a confidence level of 95% ,
we find the lower limit on the egmf strength to be @xmath2 gauss if no limit is placed on the livetime of the source or @xmath3 gauss if the source livetime is limited to the past @xmath4 3 years during which _ fermi _ observations have taken place . |
the high - energy scattering of a dipole off a nucleus / hadron in the _ mean field approximation _ is decribed by the bk - equation @xcite .
the main results following from the bk - equation are the so - called geometric scaling behaviour of the scattering amplitude @xcite and the , roughly , powerlike energy dependence of the saturation scale @xcite which are supported by the hera data @xcite . over the last few years
, we have had real breakthroughs in our understanding of high - energy scattering near the unitarity limit .
namely , we have understood how to deal with small-@xmath0 dynamics at high energy _ beyond the mean field approximation _ ,
i.e. , beyond the bk - equation . in this work , after briefly introducing the known dynamics in the mean - field case , we discuss the main steps of the recent theoretical progress as follows : we start with a discussion of the first step beyond the mean field approximation , which was done in ref . @xcite by enforcing the bfkl evolution in the scattering process to satisfy natural requirements as unitarity limits and lorentz invariance .
the consequence was a correction to the saturation scale and the breaking of the geometric scaling at high energies .
then , we explain the relation between high - energy qcd and statistical physics found in ref .
@xcite which has clarified the physical picture of , and the way to deal with , the dynamics beyond the bk - equation .
we explain that _ gluon number fluctuations _ from one scattering event to another and the _ discreteness _ of gluon numbers , both ignored in the bk evolution and also in the balitsky - jimwlk equations @xcite , lead to the breaking of the geometric scaling and to the correction to the saturation scale , respectively . in a next step
we show the new evolution equations , the so - called pomeron loop equations @xcite , which include a new element in the evolution , the pomeron loop .
finally , we discuss the possibility of phenomenological implications @xcite of the recent theoretical advances .
( for further studies on the recent theoretical advances ( not discussed here ) see also @xcite . )
consider the high - energy scattering of a dipole of transverse size @xmath1 off a target ( hadron , nucleus ) at rapidity @xmath2 .
the @xmath3-dependence of the @xmath4-matrix in the mean field approximation is given by the bk - equation ( @xmath3-dependence is suppressed for simplicity ) @xmath5 \ .
\label{eq : bk_eq}\ ] ] this equation can be interpreted as follows ; if increasing the rapidity of the dipole by @xmath6 while keeping the rapidity of the target fixed , the probability for the dipole to emit a gluon increases . in the large-@xmath7
limit the initial quark - antiquark state plus the emitted gluon can be viewed as two dipoles - one of the dipoles consists of the inital quark and the aniquark part of the gluon whike the other dipoles is given by the quark part of the gluon and the inital antiquark .
the probability for the spliting of the inital dipole @xmath8 into two doughter dipoles with transverse sizes @xmath9 and @xmath10 is given by the weight in eq.([eq : bk_eq ] ) , @xmath11 , where @xmath12 is the transverse size of the emitted gluon and @xmath13 $ ] @xcite .
on the right - hand side of eq.([eq : bk_eq ] ) , the first three terms ( first one is virtual ) describe the scattering of a single dipole with the target whereas the last term gives the simultanous scattering of the two doughter dipoles with the target , as shown in fig .
[ fig : bk_eq ] . without the last term
, the bk - equation reduces to a linear equation , the bfkl equation , which gives the growth of @xmath14 with rapidity , while the nonlinaer term , @xmath15 , tames the growth of @xmath16 such that the unitarity limit , @xmath17 , is satisfied .
one of the main results following from the bk - equation is the _ geometric scaling _
behaviour of the @xmath4-matrix @xcite in a large kinematical window @xmath18 where @xmath19 is the so - called _ saturation momentum _
defined such that @xmath20 be of @xmath21 .
( [ eq : gs ] ) means that the @xmath4-matrix scales with a single quantity @xmath22 rather than depending on @xmath23 and @xmath3 separatelly .
this behaviour implies a similar scaling for the dis cross section , @xmath24 , which is supported by the hera data @xcite .
another important result that can be extracted from the bk - equation is the rapidity dependence of the saturation momentum ( leading-@xmath3 contribution)@xcite , @xmath25 \ , \label{eq : qrmf}\ ] ] where @xmath26 is the bfkl kernel and @xmath27 .
the shape of the @xmath4-matrix resulting from the bk - equation is preserved in the transition region from weak ( @xmath28 ) to strong ( @xmath29 ) scattering with increasing @xmath3 , showing a `` travelling wave '' behaviour as sketched in fig.[fig : bk_eq1 ] , on the left hand side . with increasing @xmath3 , the saturation region at @xmath30 where @xmath29 however widens up , including smaller and smaller dipoles , due to the growth of the saturation momentum .
as we will see in the next sections , the situation changes a lot once gluon number fluctuations are taken into account .
let s start with an elementary dipole of size @xmath31 at rapidity @xmath32 and evolve it using the bfkl evolution up to @xmath33 .
the number density of dipoles of size @xmath34 at @xmath3 in this dipole , @xmath35 , obeys a completeness relation @xmath36 where on the right hand side the rapidity evolution is separated in two successive steps , @xmath37 . with @xmath38 eq.([eq : nlf ] ) can be approximately rewritten in terms of the @xmath4-matrix as @xmath39 where @xmath40 . in ref . @xcite it was realized that the above completeness relations , or , equivalently , the lorentz invariance , is satisfied by the bk evolution only by violating unitarity limits .
this can be illustrated as follows : suppose that @xmath34 is close to the saturation line , @xmath41 , so that the left hand side of eq.([eq : tlf ] ) is large . on the right hand side of eq.([eq : tlf ] ) it turns out that @xmath42 is typically very small in the region of @xmath43 which dominates the integral .
this means that @xmath44 must be typically very large and must violate unitarity , @xmath45 , in order to satisfy ( [ eq : tlf ] ) . the simple procedure used in ref .
@xcite to solve the above problem was to limit the region of the @xmath43-integration in eq.([eq : tlf ] ) by a boundary @xmath46 so that @xmath44 would never violate unitarity , or @xmath42 would always be larger than @xmath47 .
the main consequence of this procedure , i.e. , bk evolution plus boundary correcting it in the weak scattering region , is the following scaling behaviour of the @xmath4-matrix near the unitarity limit @xmath48 and the following energy dependence of the saturation momentum @xmath49 \label{eq : qrbmf}\ ] ] with @xmath50 eq.([eq : trbmf ] ) shows the breaking of the geometric scaling , which was the hallmark of the bk - equation shown in eq.([eq : gs ] ) , and eq.([eq : qrbmf ] ) shows the correction to the saturation momentum ( cf .
eq.([eq : qrmf ] ) ) , both emerging as a consequence of the evolution beyond the mean field approximation .
the high energy evolution can be viewed also in another way which is inspired by dynamics of reaction - diffusion processes in statistical physics @xcite . to show it ,
let s consider an elementary target dipole of size @xmath31 which evolves from @xmath51 up to @xmath33 and is then probed by an elementary dipole of size @xmath23 , giving the amplitude @xmath52 .
it has become clear that the evolution of the target dipole is _ stochastic _ leading to random dipole number realizations inside the target dipole at @xmath3 , corresponding to different events in an experiment . the physical amplitude , @xmath52 ,
is then given by averaging over all possible dipole number realizations / events , @xmath53 , where @xmath54 is the amplitude for dipole @xmath23 scattering off a particular realization of the evolved target dipole at @xmath3 .
an illustration is shown in fig.[fig : bk_eq1 ] , the right - hand side plot , where the @xmath4-matrix for different events ( thin lines ) and the average over all events ( thick lines ) , @xmath55 , are shown at two different rapidities , respectivelly .
the mean field description breaks down at low target dipole occupancy due to the _ discreteness and the fluctuations of dipole numbers_. because of discreteness the dipole occupancy can not be less than one for any dipole size . taking this fact into account by using the bk equation with a cutoff when @xmath4 becomes of order @xmath56 @xcite , or the occupancy of order one ( see eq.([eq : tn ] ) ) , leads exactly to the same correction for the saturation momentum as given in eq.([eq : qrmf ] ) .
the latter cutoff is essentially the same as , and gives a natural explanation of , the boundary used in ref.@xcite and briefly explained in the previous section .
the dipole number fluctuations in the low dipole occupancy region result in fluctuations of the saturation momentum from event to event , with the strength @xmath57 extracted from numerical simulations of statistical models .
the averaging over all events with random saturation momenta , in order to get the physical amplitude , causes the breaking of the geometric scaling and replaces it by a new scaling law , the so - called _ diffusive scaling _ , in which case the scattering amplitude is a function of another variable , @xmath58 this equation differs from eq.([eq : trbmf ] ) since eq.([eq : trbmf ] ) misses dipole number fluctuations .
note that because of the geometric scaling violation , the result in eq.([eq : tqs ] ) changes the shape as the rapidity increases , as illustrated in fig .
[ fig : bk_eq1 ] ( right - hand side ) by the decreasing slope of the thick line with growing rapidity , in contrast to the solution to the bk - equation in eq.([eq : gs ] ) .
the statistical physics / high - density qcd correspondence suggests the following picture for the wavefunction of a highly evolved hadron which is probed by a dipole of transverse size @xmath23 : as the hadron is boosted to high rapidities the density of gluons inside the hadron grows .
also the fluctuation in gluon numbers , which is characterized by the dispersion in eq .
( [ eq : disp ] ) , grows with rising rapidity .
however , as long as @xmath59 , which means @xmath60 , the effects of fluctuations can be neglected and the evolution of the hadron is described to a good approximation by the bk - equation . thus , for @xmath61 , as shown in fig.[fig : picture ] , to the left of the saturation line , @xmath62 , is the `` saturation region '' with the `` large - size '' ( small momentum ) gluons at a large density , of order @xmath63 or the @xmath64 , while the shadowed region is the transition region from high to low gluon density , or the front of the at @xmath4-matrix ( geometric scaling regime ) . at higher rapidities , @xmath65 , where the fluctuations become important , the geometric scaling regime is replaced by the diffusive scaling given in eq .
( [ eq : tqs ] ) .
it was always clear that the bk equation does not include fluctuations .
however , it took some time to realize that also the balitsky - jimwlk equations do miss them .
it turned out ( see first reference in @xcite ) that the balitsky - jimwlk equations do include bfkl evolution , `` pomeron mergings '' but not also `` pomeron splittings '' , which are represented by the three graphs in fig .
[ fig : bjimwlk ] for two dipoles scattering off a target , respectivelly . after this insight , the so - called pomeron loop equations @xcite
have been constructed to account for `` pomeron splittings '' or dipole number fluctuations .
the pomeron loop equations can be expressed in a hamiltonian language , in which case one extends the jimwlk - equation @xcite , or in terms of scattering amplitudes @xcite , in which case the balitsky equations are extended . in order to be close to the bk - equation discussed in sec .
[ sec : mfa ] , we show the pomeron loops using the scattering amplitude . in the large-@xmath7 limit
, they can be written either as a stochastic equation of langevin - type , @xmath66 \nonumber \\ & + & \frac{\alpha_s}{2 \pi } \sqrt{\frac{\alpha_s n_c}{2\pi^2 } } \int\limits_{{\bf u , v , z } } { \cal a}({\bf x},{\bf y}|{\bf u},{\bf z } ) \frac{|{\bf u}-{\bf v}|}{({\bf u}-{\bf z})^2 } \sqrt{\nabla^2_{\bf u}\nabla^2_{\bf v}\,t_{\bf uv}}\ \nu({\bf u},{\bf v},{\bf z};y ) \label{eq : le}\end{aligned}\ ] ] or , equivalently , as a hierarchy of coupled equations of averaged amplitudes , where for simplicity we show only the first two of them , which read @xmath67 \nonumber \\
\frac{\partial \langle t_{\bf xz } t_{\bf zy}\rangle}{\partial y}&= & \frac{\alpha_sn_c}{2\pi^2 } \int\limits_{{\bf t } } \
m_{{\bf xzt}}\,\left [ -\langle t_{\bf xz}t_{\bf zy}\rangle + \langle t_{\bf xt}t_{\bf zy}\rangle + \langle t_{\bf tz}t_{\bf zy}\rangle - \langle t_{\bf xt}\,t_{\bf tz}t_{\bf zy}\rangle \right ] \nonumber \\ & + & \frac{\alpha_sn_c}{2\pi^2 } \int\limits_{\bf t } \ m_{\bf zyt}\,\left [ -\langle t_{\bf xz}t_{\bf zy}\rangle + \langle t_{\bf xz}t_{\bf zt}\rangle + \langle t_{\bf xz}t_{\bf ty}\rangle - \langle t_{\bf xz}\,t_{\bf zt}t_{\bf ty}\rangle \right ] \nonumber \\ & + & \left(\frac{\alpha_s}{2 \pi}\right)^2 \frac{\alpha_s n_c}{2\pi^2}\ \int\limits_{{\bf u},{\bf v } } { \cal r}({\bf x},{\bf z},{\bf z},{\bf y}|{\bf
u},{\bf v } ) \
\langle t_{\bf uv } \rangle \label{eq : hie}\end{aligned}\ ] ] where the noise is non - diagonal ( non - gaussian ) in the first two arguments @xmath68 the triple pomeron kernel @xcite reads @xmath69\ ] ] and @xmath70 is the amplitude for dipole - dipole scattering in the two - gluon exchange approximation and for large-@xmath7 , with @xmath71 \ .\ ] ] in above equations the integrations are always over transverse sizes , @xmath72 .
last term in eq.([eq : le ] ) , containing the non - gaussian noise @xmath73 , is new as compared with the bk - equation and accounts for fluctuations in the dipole numbers or the stochastic nature of the evolution in small-@xmath0 physics .
the hierarchy in eq.([eq : hie ] ) reduces to the bk - equation only in the mean field approximation , i.e. , when @xmath74 .
the hierarchy in eq.([eq : hie ] ) , as compared with the balitsky - jimwlk hierarchy , involves in addition to linear bfkl evolution ( fig.[fig : bjimwlk](a ) ) and pomeron mergings ( fig.[fig : bjimwlk](b ) ) , also pomeron splittings ( fig.[fig : bjimwlk](c ) ) , and therefore , in the course of the evolution , also _
pomeron loops_. the three pieces of evolution are represented by the linear terms , nonlinear terms and the last term on the right - hand side of the second equation in eq.([eq : hie ] ) , respectivelly , which describes the scattering of two dipoles off a target .
it is nt yet clear at which energy fluctuation / pomeron loop effects start becoming important .
the results shown in the previous sections , eq.([eq : qrbmf ] ) and eq.([eq : tqs ] ) , are valid at asymptotic energies .
a solution to the evolution equations , which is not yet available because of their complexity , would have helped to better understand the subasymptotics . using the statistical physics / high density qcd correspondence , phenomenological consequences of fluctuations in the fixed coupling case
have been studied , for example for dis and diffractive cross sections @xcite , forward gluon production in hadron - hadron collisions @xcite and for the nuclear modification factor @xmath75 @xcite , in case fluctuations become important in the range of lhc energies .
recently , in the fixed coupling case , it has been shown that dipole - proton scattering amplitudes which include fluctuation effects seem to describe better the hera data .
also the parameters turn out reasonable : the diffusion coefficient @xmath76 ( @xmath77 ) is in agreement with numerical simulations of approximations to pomeron loop equations @xcite and the saturation exponent @xmath78 ( @xmath79 ) is decreased as expected theoretically . on the other hand , allowing the coupling to run , however , within a toy model @xcite which is supposed to mimic the qcd evolution equations with pomeron loops , it has been argued that gluon number fluctuations / pomeron loops can be neglected in the range of hera and lhc energies .
see also refs .
@xcite for more studies on running coupling plus gluon number fluctuations . | it has been recently understood how to deal with high - energy scattering beyond the mean field approximation .
we review some of the main steps of this theoretical progress , like the role of lorentz invariance and unitarity requirements , the importance of discreteness and fluctuations of gluon numbers ( pomeron loops ) , the high - energy qcd / statistical physics correspondence and the consequences for the saturation scale , the scattering amplitude and other , also measurable , quantities . |
@xcite presented high signal - to - noise ratio _ spitzer
_ infrared spectrograph ( irs ) observations of 18 virgo early - type galaxies ( etgs hereafter ) , selected from those that define the color - magnitude relation of the cluster , with the aim of detecting the silicate emission of their dusty , mass - losing evolved stars .
they found that 13 etgs , i.e.76% of the sample , show a pronounced broad silicate feature that is spatially extended and likely of stellar origin , in agreement with model predictions of old populations @xcite .
the absence of any other feature in their spectra suggested that these galaxies could be considered as the prototype of passively evolving etgs .
in contrast , the irs spectrum of the other four etgs presented various levels of activity . among these ,
ngc 4486 ( m 87 ) , the well - known giant elliptical located at the cluster centre , stands out .
its mid - infrared ( mir hereafter ) spectrum , besides narrow atomic emission lines , shows an excess of emission at longer wavelengths with respect to passive etgs .
the silicate emission at 10@xmath0 m is also evident in the spectrum , and since m 87 hosts a strong central radio source ( virgo a ) and a jet , visible from radio to x - ray wavelengths , it could be an indication of possible torus emission , as already found in other agns ( siebenmorgen et al .
2005 , shi et al .
2006 , perlman et al .
2007 ) .
the importance of the torus in the context of the agn energetics makes the debate about the origin of the nuclear mir emission in m 87 very intriguing .
early , high spatial resolution , ground - based mir observations of m 87 @xcite have shown that the emission from the nucleus is consistent with synchrotron emission .
both investigations have excluded the possibility that m 87 hosts a quasar - like nucleus similar to that at the center of cen a. however , neither could exclude the possibility of a low - luminosity agn .
indeed @xcite detected a faint , extended , circum - nuclear component that could originate in a torus contributing @xmath37% of the nuclear luminosity .
alternatively , they suggested that this extended component could be stellar emission from red giant stars .
more recently , @xcite reanalyzed this issue by means of high resolution ground based imaging and _ spitzer _ spectroscopy .
they confirmed the consistency with synchrotron emission for the mir fluxes of the knot a / b complex in the jet .
however , they measured a spectral index @xmath40.41 , much lower than before and different from that found in the nearby structures ( e.g. * ? ? ? * ) .
they still noticed deviations from the pure synchrotron emission and speculated that , at least in the shorter wavelength range , this could be due either to contribution from m and k stars or contamination from the low frequency tail of the emission from knot hst-1 , not resolved by _
spitzer _ but also not apparent in the 11.7@xmath0 m _ subaru _ image of the nucleus .
we revisit here the nature of the mir emission of m 87 exploiting our high signal - to - noise _ spitzer _ spectra of passively evolving etgs .
taking advantage of these data we have constructed very accurate mir templates of the stellar populations in these galaxies ( valdes et al . 2009 , in prep . ) .
_ these templates will be used to disentangle the nuclear non - stellar emission of m 87 by a proper subtraction process . _ in section [ data ] we summarize our _ spitzer irs _ observations and data extraction as well as all supplementary data needed to perform the subtraction . in section [ subtraction ]
we describe the subtraction of the spectral energy distribution ( sed ) of a passive etg from the irs spectrum of m 87 while , in section [ residual ] we discuss the properties of the residual component .
our conclusions are presented in section [ conclusions ] .
in this section we describe how we have obtained the nir / mir seds of m 87 , and of three other elliptical galaxies , ngc 4365 , ngc 4382 and ngc 4636 .
these were selected to be `` passive '' , insofar as they show no evidence of either recent past star formation or agn activity .
they will be used as templates for the stellar component in our attempt to isolate the non stellar emission of the nucleus of m 87 .
standard staring mode short ( sl1 and sl2 ) and long ( ll2 ) low resolution irs spectral observations of m 87 and other selected virgo cluster etgs were obtained during the first _ spitzer _ general observer cycle on 2005 june 1 as part of program i d 3419 ( pi a. bressan ) .
the spectra were extracted in a fixed aperture ( 3.6``@xmath518 '' for sl and 10.2``@xmath510.4 '' for ll ) , using a purpose built program that allows the reconstruction of the intrinsic galaxy profile @xcite .
position angles of the sl ( 3.6 `` wide ) and ll ( 10.2 '' wide ) irs slits in the case of m 87 , are shown superimposed to the mips 24@xmath0 m image , in figure [ fig1 ] . at these wavelengths
the emission of ngc 4486 appears quite complex even at the low resolution of mips 24@xmath0 m .
perlman et al . (
2001 ) have provided a thorough description of the mir spatial emission of the central regions of m87 .
their 10.8 @xmath0 m image shows the `` nucleus '' and the d , f and a , b and c knots , whose fluxes are reported in their table 1 . to check which of these external components can affect our spectra
we show the 2-d spectral images of of the ll2 and sl1 segments in the bottom and top panels of figure [ fig1 ] , respectively . in the ll2
slit we can easily recognize both the spectra of the nucleus and that of the nearest bright a , b and c knots , whereas the sl1 slit is clearly not affected by these three components .
thus , the only extra - nuclear component that may contribute to our sl spectrum is knot d. however this feature is located 3 `` from the nucleus , has low level flux ( 7% of the nucleus , perlman et al .
2001 ) and thus is likely to provide only a negligible contribution to the sl spectrum .
the other components , located at greater distances , are well outside the sl slit . on the other hand
our ll slit contains the strong knots a , b and c that are clearly visible in the spectral images . in this case
the extraction is limited to a width of 10.4 '' in order to avoid these strong knots that are not present in sl . as we will see later
our results are independent of the inclusion or exclusion of the ll segment from our analysis . the estimate of the flux _ emitted _ by an extended source within a fixed slit aperture involves the deconvolution of the received flux with the psf of the instrument .
this correction is important to obtain the shape of the intrinsic spectral energy distribution ( sed ) of the galaxy , because from the slit loss correction we have estimated that for a point source the losses due to the psf amount to @xmath320% at 5@xmath0 m and to @xmath340% at 15@xmath0 m .
conversely , a uniform source will suffer no net losses . in order to recover the intrinsic sed we have convolved a surface brightness profile model with the instrumental psf , and we have simulated the corresponding linear profile along the slits , taking into account the relative position angles of the slits and the galaxy .
the adopted intrinsic model profile is a wavelength dependent two dimensional modified king law ( elson at al .
1987 ) : @xmath6^{-\gamma/2 } \label{elson}\ ] ] where @xmath7 and @xmath8 are the coordinates along the major and minor axis of the galaxies , @xmath9 is the axial ratio taken from the literature .
@xmath10 , @xmath11 and @xmath12 are free parameters that are functions of the wavelength and are obtained by fitting the observations with the simulated profile . in order to get an accurate determination of the parameters of the profiles
several wavelength bins have been co - added .
* this procedure allows us ( i ) to recognize whether a particular feature is resolved or not , ( ii ) to reconstruct the intrinsic profile of the source and ( iii ) to account for the wavelength dependent psf losses , in a more general way than for point like sources .
* examples of the fitting procedure for ngc 4486 at two selected wavelengths are shown in fig.[fig2 ] .
the dashed lines show the reconstructed one dimensional intrinsic spatial emission profiles ( in electrons / s ) sampled along the slit , while the solid lines represent the corresponding one dimensional psf convolved models that are fitted to the data .
the error bars in the figure show 1-@xmath13 uncertainties that were computed taking into account the poissonian noise from the source and the background , the readout noise and dark current noise , as described in the spitzer observer manual .
the spatial emission parameters obtained for all the galaxies used in this paper are shown in table 1 . at short irs wavelengths the spatial emission of ngc 4486 appears complex .
it is a superposition of an unresolved central component and an extended component . indeed , in fitting only a single component to the data we find an unresolved central source with @xmath11=0.1 ``
nonetheless , there is an evident excess over this unresolved component beyond @xmath14 , that does affect our rectangular extraction , 18 '' long .
the excess remains visible up to @xmath314@xmath0 m beyond which the profile is clearly that of the psf ( @xmath11=0.1 " ) , and the excess in the wings is much less important .
from this spatial analysis we conclude that ngc 4486 is progressively more dominated by un unresolved nuclear component as the wavelength increases .
however , the presence of an underlying extended emission , evident at 5 @xmath0 m , can be detected even at 9@xmath0 m .
since we have shown that the external jet does not contribute to the sl segment , we argue that this emission is of stellar origin .
lrrrrr [ tab0 ] & & 5.4@xmath0m&&14.0@xmath0 m + galaxy & @xmath11&@xmath12 & @xmath11&@xmath12 + ngc 4486 & 0.1&0.5&0.1&0.7 + ngc 4365 & 3.5&2.0&4.0&2.2 + ngc 4382 & 1.3&1.4&1.6&1.3 + ngc 4636 & 3.7&0.8&3.5&0.9 + irac fluxes were measured from archival `` post basic calibrated data '' images using a customized program written in idl . before flux measurements ,
the background in each image was determined as the modal sky value of the off - source portion of each image .
fluxes were then measured in circular apertures centred on each galaxy .
the error on each flux measurement was calculated as the quadrature sum of the calibration error ( taken to be 10% ) and the standard deviation of the background value .
the calibration error is the dominant term . in order to place m 87 irac and irs data on the same flux scale , we measured irac 5.8 @xmath0 m fluxes in the circular aperture containing the same flux sampled by irs at the same wavelengths this aperture has a radius of 4 and has been selected as the reference aperture for any band .
consequently , all broad band fluxes in all other galaxies were measured within the same aperture and their irs spectra were normalized to their flux in irac channel 3 . in the case of m 87 , we have also measured the emission at 24@xmath0 m from the mips image , which may serve as a useful check for our irs extraction procedure .
since the irs extraction procedure indicates that toward these wavelengths m87 is dominated by the nuclear component , we have taken into account the effects of the psf to obtain the mips 24@xmath0 m flux . to this end
we have divided the flux measured within the nuclear @xmath15 radius aperture with the fractional energy of the mips 24@xmath0 m psf image , encircled by the same radius .
nir images of the sample galaxies in the j , h and k@xmath16 bands were downloaded from the all - sky data release from the two micron all sky survey . using these data , in addition to
our irs spectra , we have obtained the nir / mir sed of m87 and of the other three etgs that will be used as templates for the subtraction of the stellar component .
in order to disentangle the nuclear and the extended component we resort to the comparison of the high signal - to - noise spectral energy distribution ( sed ) of m 87 with that of the passive ellipticals we have also observed with spitzer . here we describe the process of subtraction of the stellar component . before subtraction ,
all the template galaxy seds were normalized to the flux of m 87 in the 2mass ks band .
this band was selected because it is the nearest band to the irs spectrum with a negligible non - stellar component .
the nir / mir sed of m 87 and of the template etgs , after this normalization , are shown in figure [ fig3 ] .
we notice that all seds are very similar in the nir bands .
this similarity , that extends to optical wavelengths , indicates a uniformity of the underlying stellar populations .
the seds of passive etgs are very similar also in the mir , especially where the integrated spectrum is dominated by the stellar photospheres , i.e. below 8@xmath0 m . at longer wavelengths
some differences appear , likely caused by small differences in the average properties of circumstellar envelopes of mass loosing agb stars @xcite . *
this uniformity is strengthened by the recent comparison of the surface brightness distribution in the ks and irs peak - up 16@xmath0 m bands , of bright etgs in virgo and coma clusters @xcite*. ngc 4636 shows also narrow emission lines but its stellar continuum is that of a passive elliptical .
in contrast , m 87 shows a clear mir excess which reaches about one order of magnitude at @xmath1720@xmath0 m ( see also perlman et al .
an estimate of the quality of our irs extraction process is provided by the independent measure of the 24@xmath0 m mips flux , plotted in the same figure
. in summary , inspection of figure [ fig3 ] shows that m87 stands out in the mir spectral window . in this region , we hypothesize that the stellar populations of m 87 and the other passive etgs are similar , and we subtract the spectra of the passive etgs from that of m 87 .
lcccccc [ tab2 ] galaxy&@xmath18&@xmath19&s / f@xmath20&s@xmath21&s@xmath22 + template & & & mjy & mjy + ngc 4365&[email protected]&[email protected]&0.06&8.9&20.0 + ngc 4382&[email protected]&[email protected]&0.05&7.7&21.3 + ngc 4636&[email protected]&[email protected]&0.06&8.1&22.0 +
to highlight the nature of the mir excess we simply subtract the seds of the passive etgs from that of m 87 . figure [ fig4 ] shows the results obtained using the sed of ngc 4382 , but similar results are obtained adopting the other galaxies as templates for the stellar emission . the residual , without emission lines that have been masked out ,
is well - fitted with a power - law and does not show any evidence of other particular features reminiscent of additional emission components .
this is even more surprising considering the prominent broad features visible in the stellar template and given the relative contribution of the stellar and non - stellar components .
table 2 summarizes the power - law exponents ( s@xmath2 ) , the fractional contribution of the power - law in the ks band , and its absolute flux at 3.6 and 11.7@xmath0 m , for the three templates adopted .
errors in the exponents are only formal errors of the fits and do not include errors from the normalizations involved in our procedure . a realistic uncertainty is provided by the range of slopes obtained with different galaxy templates . * our residual flux
( table 2 ) is consistent with the high resolution , ground based nuclear values by @xcite ( s@xmath24= [email protected] mjy at 11.67@xmath0 m , fwhm=@xmath26 , figure [ fig4 ] red dot ) and @xcite ( s@xmath27 mjy at 10.8@xmath0 m , fwhm@xmath28 , figure [ fig4 ] red triangle ) .
* in the same figure , we have also plotted the observed mips 24 @xmath0 m flux , before ( open square ) and after ( filled square ) a 15% correction due to the contribution of stars , estimated from the extrapolation of the passive elliptical template .
the mips 24 @xmath0 m flux is also a time independent measure and falls on the same residual power - law the power - law has a slope @xmath18 between 0.77 - 0.82 , depending on the adopted stellar population template .
this value is significantly larger than that found by @xcite ( @xmath30.41 ) .
they fitted different irs observations ( pid82 , rieke ) , within a lower extraction width and using the spectral range 7.5 to 15@xmath0 m , specifically selected to avoid the shorter wavelength region where the contamination from the stars could be significant .
this large discrepancy could be real , for example due to a variability of the nuclear component , or could simply be due to the different method adopted for the analysis . to understand its origin we have re - analysed the irs data set used by @xcite , with our extraction procedure .
the resulting spectrum is plotted in figure [ fig4 ] ( crosses ) .
we notice that our flux values are larger than those published by @xcite , but this is due to the 2.5 times larger extraction width that we have adopted .
since , apart from this , the two irs datasets are very similar , we argue that the different values of the slope arise from the different approach adopted .
a crude fit to our m 87 sed , in the 7.5 to 15@xmath0 m region , gives a slope @xmath18 @xmath30.17 .
but in our case the strength of the underlying stellar population is not negligible and changes from 70% at 5 @xmath0 m to 25% at 15 @xmath0 m . after the subtraction ,
the slope of the residual is @xmath18 @xmath3 0.8 . within the smaller aperture used by @xcite
the contribution of the stellar population is certainly less severe than in our case , but not negligible , in particular around 7.5 @xmath0 m .
the lack of this correction is most likely the cause of the different slope obtained . finally , we notice that the slope we derived is more consistent with the previous @xcite analysis .
our slope is also similar to the value found for jet knots a / b and c / g by shi et al .
2007 , and to the slope found at radio wavelengths when the entire , extended radio source is measured .
exploiting our high signal - to - noise _ spitzer _ irs spectra and broad band data from the literature , we have built template seds of the passive stellar populations in etgs . after a passive stellar component is properly subtracted from our m 87 irs spectrum , the residual , almost independent of the adopted template , is a power - law , and lacks any signature of the 10@xmath0 m feature .
this suggests that the 10@xmath0 m bump observed in the irs spectrum of m 87 can be entirely explained in terms of the underlying stellar population , typical of all other passively evolving virgo ellipticals ( bressan et al .
the residual intensity matches fairly well the high spatial resolution ground based _ subaru _ nuclear measurements at 10.8@xmath0 m and 11.67@xmath0 m , by @xcite and @xcite , respectively . since the contribution of stars at these wavelengths is not negligible , this indicates that our procedure , even if applied to a large aperture ( r@xmath34 corresponding to @xmath17300 pc at 16 mpc ) is accurate enough to reveal the emission from the @xmath291 core .
the above findings confirm , in a new independent way , that the inner ( @xmath291 " ) core of m 87 is dominated by synchrotron emission , leaving little room for a torus component @xcite . since the signatures of mir silicate emission have been detected in many agn observed with _
@xcite , we suggest that , in some circumstances , the underlying stellar population may be an important component to be taken into consideration . this work is based on observations made with the _ spitzer _ space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
we thank the referee for his comments that helped to improve the final version of this paper . l.s .
, a. b. and g.l .
g. thank inaoe for warm hospitality .
a.m. acknowledges support from the italian scientists and scholars of north america foundation .
we acknowledge contracts asi - inaf i/016/07/0 and conacyt 54511 .
bower , r. g. , lucey , j. r. , & ellis , r. s. 1992 , , 254 , 589 bressan , a. , granato , g.l . , & silva , l. 1998 , aa , 332 , 135 bressan , a. , panuzzo , p. , buson , l. , clemens , m. , granato , g. l. , rampazzo , r. , silva , l. , valdes , j. r. , vega , o. , & danese l. 2006 , apj , 639 , l55 clemens , m. s. , bressan , a. , panuzzo , p. , rampazzo , r. , silva , l. , buson , l. , & granato , g. l. 2009 , , 392 , 982 elson , r. a. w. , fall , s. m. , & freeman , k. c. 1987 , apj , 323 , 54 higdon , s.j.u . ,
devost , d. , higdon , j. l. , brandl , b. r. , houck , j. r. , hall , p. , barry , d. , charmandaris , v. , smith , j. d. t. , sloan , g. c. , & green , j. 2004 , pasp , 116 , 975 perlman , e. s. , sparks , w. b. , radomski , j. , packham , c. , fisher , r. s. , pia , r. , & biretta , j. a. 2001 , , 561 , l51 perlman , e. s. , mason , r. e. , packham , c. , levenson , n. a. , elitzur , m. , schaefer , j.j . , imanishi , m. , sparks , w.b . , & radomski , j. 2007 , , 663 , 808 shi , y. , rieke , g. h. , hines , d. c. , gorjian , v. , werner , m. w. , cleary , k. , low , f. j. , smith , p. s. , & bouwman , j. 2006 , , 653 , 127 shi , y. , rieke , g. h. , hines , d. c. , gordon , k. d. , & egami , e. 2007 , , 655 , 781 siebenmorgen , r. , haas , m. , krgel , e. , & schulz , b. 2005 , , 436 , l5 whysong , d. , & antonucci , r. 2004 , , 602 , 116 | we study in detail the mid - infrared _ spitzer_-irs spectrum of m 87 in the range 5 to 20 @xmath0 m .
thanks to the high sensitivity of our _ spitzer_-irs spectra we can disentangle the stellar and nuclear components of this active galaxy . to this end we have properly subtracted from the m 87 spectrum , the contribution of the underlying stellar continuum , derived from passive virgo galaxies in our sample . the residual is a clear power - law , without any additional thermal component , with a zero point consistent with that obtained by high spatial resolution , ground based observations . the residual is independent of the adopted passive template .
this indicates that the 10@xmath0 m silicate emission shown in spectra of m 87 can be entirely accounted for by the underlying old stellar population , leaving little room for a possible torus contribution .
the mir power - law has a slope @xmath1 0.77 - 0.82 ( s@xmath2 ) , consistent with optically thin synchrotron emission .
# 1 # 2 _ mem . soc . astron .
it . _ * # 1 * , # 2 # 1 # 2 _ the messenger _ # 1 , # 2 # 1 # 2 astron .
nach .
# 1 , # 2 # 1 # 2 a&a # 1 , # 2 # 1 # 2 a&a # 1 , l#2 # 1 # 2 a&ar # 1 , # 2 # 1 # 2 a&as # 1 , # 2 # 1 # 2 aj # 1 , # 2 # 1 # 2 ara&a # 1,#2 # 1 # 2 apj # 1 , # 2 # 1 # 2 apj # 1 , l#2 # 1 # 2 apjs # 1 , # 2 # 1 # 2 ap&ss # 1 , # 2 # 1 # 2 adv .
space res .
# 1 , # 2 # 1 # 2 bull . astron
. inst .
czechosl .
# 1 , # 2 # 1 # 2 j. quant .
spectrosc .
radiat .
transfer # 1 , # 2 # 1 # 2 mnras # 1 , # 2 # 1 # 2 mem .
r. astr .
soc .
# 1 , # 2 # 1 # 2 phys .
lett .
rev .
# 1 , # 2 # 1 # 2 publ .
astron .
soc .
japan # 1 , # 2 # 1 # 2 publ .
astr .
soc .
pacific # 1 , # 2 # 1 # 2 nat # 1 , # 2 # 1 # 2 acta astron . # 1 , # 2 |
in @xcite , christopher hardin and alan taylor present a remarkable consequence of the axiom of choice .
consider an arbitrary set @xmath0 and a function @xmath1 .
we think of the domain of @xmath2 as representing time , and for each @xmath3 , we think of @xmath4 as the state of some system at time @xmath5 .
the question hardin and taylor address is whether or not it is possible to predict the state of the system from its past behavior . in other words
, can we predict @xmath4 from @xmath6 ? if there is no restriction on the function @xmath2
, then it seems that @xmath6 gives us no information about @xmath4 , and therefore it is hard to see how such a prediction could be possible .
nevertheless , using the axiom of choice , hardin and taylor produce a prediction strategy with the property that for every function @xmath2 , the prediction is correct for all but countably many values of @xmath5
. it may be helpful to imagine that at each time @xmath5 there is an agent who is making the prediction of @xmath4 .
each agent can see the past but not the present or future , so the agent at @xmath5 knows @xmath6 but no other values of @xmath2 .
hardin and taylor make an extensive study of the predictions that can be made by agents in this and similar situations in their book @xcite .
one idea considered by hardin and taylor is the possibility that the agents can see the past but do nt know what time it is in other words , the agent at time @xmath5 knows the value of @xmath7 for every @xmath8 , but does nt know the value of @xmath5 .
they define a prediction strategy to be _ anonymous _ if it can be used by agents whose knowledge is restricted in this way . and
they show that even with this restriction on the knowledge of the agents , there is a prediction strategy that guarantees that for every function @xmath2 , all but countably many agents are correct . in this paper
we consider several variations on the notion of anonymity .
what if the agents not only do nt know what time it is , they also do nt know the rate at which time passes ?
what if each agent knows the order of past events , but nothing about the length of time between events ?
we will see that in some cases the agents are still able to perform well , with only countably many making incorrect predictions , while in other cases they do very badly , with only countably many making correct predictions .
to state our results formally , we will need some definitions .
let @xmath9 denote the set of all functions from @xmath10 to @xmath0 .
each such function represents a possible time evolution of the system under study . following hardin and taylor , we call such functions _
scenarios_. given a scenario @xmath2 , the predictions made by all of the agents could be thought of as determining another scenario , which we will denote @xmath11 .
thus , @xmath12 , and for any time @xmath3 , @xmath13 denotes the prediction of @xmath4 made by the agent at time @xmath5 under the scenario @xmath2 .
since the agent at @xmath5 can see only the past , scenarios @xmath2 and @xmath14 will be indistinguishable to him if @xmath15 , and therefore his predictions in these scenarios should be the same .
we therefore make the following definition . a function @xmath16 is called a _ predictor
_ if it has the property that for all functions @xmath17 and all @xmath3 , if @xmath15 then @xmath18 .
we say that the agent at @xmath5 _ guesses correctly _ if his prediction of the value of @xmath4 is correct that is , if @xmath19 . otherwise , agent @xmath5 _ guesses incorrectly_. we can now state hardin and taylor s main result from @xcite .
[ thm : ht ] there is a predictor @xmath20 such that for every scenario @xmath2 , only countably many agents guess incorrectly ; in other words , @xmath21 is countable .
the details of the proof can be found in @xcite , but the idea is simple . using the axiom of choice , we can let @xmath22 be a well ordering of @xmath9 .
now for any @xmath23 and any @xmath3 , let @xmath24_x = \{g \in { } ^\mathbb{r}s : f \upharpoonright ( -\infty , x ) = g \upharpoonright ( -\infty , x)\}$ ] , and let @xmath25 be the @xmath22-least element of @xmath24_x$ ] .
in other words , @xmath25 is the least scenario that is consistent with the information available to the agent at time @xmath5
. let @xmath26 .
hardin and taylor show that @xmath21 is well ordered ( by the usual ordering on @xmath10 ) , and any well ordered subset of @xmath10 is countable .
( to see why a well ordered set @xmath27 must be countable , define a one - to - one function @xmath28 by letting @xmath29 be a rational number larger than @xmath5 but smaller than the next element of @xmath30 , if there is one . ) thus , with this predictor , only countably many agents guess incorrectly .
next we present hardin and taylor s definition of anonymity .
suppose that @xmath2 is a scenario , and define another scenario @xmath14 by the equation @xmath31 , for some constant @xmath32 .
in other words , @xmath33 , where @xmath34 is the function @xmath35 . we will call @xmath34 a _ shift function_. we could think of @xmath14 as the scenario that represents precisely the same time evolution of the system as @xmath2 , but shifted in time by @xmath32 units . now consider agents at times @xmath5 and @xmath36 who can see the past , but who do nt know where they are located in time . for any @xmath37
, we have @xmath38 this means that the past from the point of view of the agent at @xmath5 in scenario @xmath14 looks exactly the same as the past from the point of view of the agent at @xmath36 in scenario @xmath2 .
we would therefore expect these agents to make the same predictions .
in other words , we would expect @xmath39 , or equivalently @xmath40 .
this motivates hardin and taylor s definition of anonymity .
a predictor @xmath20 is _ anonymous _ if for every scenario @xmath2 and every real number @xmath32 , @xmath41 anonymity of a predictor can be thought of as a sort of invariance under shifts .
this suggests a natural generalization of the notion of anonymity ( see @xcite ) .
let @xmath42 be a family of functions from @xmath10 to @xmath10 .
a predictor @xmath20 is
_ @xmath42-anonymous _
if for every scenario @xmath2 and every @xmath43 , @xmath44 thus , anonymity is just @xmath45-anonymity , where @xmath45 is the family of shift functions .
but there are other families of functions that we might use instead : @xmath46 if @xmath2 is a scenario and @xmath47 , then @xmath48 is a scenario that is the same as @xmath2 , except that events have been shifted in time , and also the rate at which events happen has been changed .
thus , a @xmath49-anonymous predictor is one that can be used by agents who not only do nt know what time it is , they also do nt know how fast time passes . similarly , @xmath50- and @xmath51-anonymous predictors are predictors that can be used by agents who are insensitive to more extreme distortions of time .
hardin and taylor ask in @xcite how successful a @xmath51-anonymous predictor can be ( see question 7.8.3 on p. 82 ) .
note that @xmath52 , and therefore the requirement of @xmath53-anonymity becomes stronger as @xmath54 increases . we will show that there is a @xmath49-anonymous predictor that always performs well ( see theorem [ thm : main1 ] ) , but there is a scenario in which all @xmath50-anonymous predictors perform badly ( see theorem [ thm : main2 ] ) .
we need a few definitions and lemmas before we can prove our main result about @xmath49-anonymous predictors .
recall that a scenario @xmath2 is said to be _ periodic _ if there is a number @xmath55 such that for all @xmath5 , @xmath56 ; @xmath32 is called a _ period _ of @xmath2 .
equivalently , @xmath2 is periodic with period @xmath32 if @xmath57 .
we will need to generalize this concept to functions in @xmath49 .
we will say that a scenario @xmath2 is _ affine - invariant _ if there is some @xmath47 such that @xmath58 is not the identity function and @xmath59 .
since our agents can see only the past , it will also be convenient to have names for scenarios that are periodic or affine - invariant up to some point in time .
suppose that @xmath5 is a real number and @xmath2 is a scenario .
if there is some @xmath55 such that @xmath60 , then we say that @xmath2 is _ past - periodic on @xmath61 with period @xmath32_. similarly , if there is some nonidentity function @xmath47 such that @xmath62 , then we say that @xmath2 is _ past - affine - invariant on @xmath61_. clearly , if a scenario is periodic or affine - invariant , then it is also past - periodic or past - affine - invariant . in the other direction ,
if a scenario is past - periodic or past - affine - invariant , then it can be modified to make it periodic or affine - invariant , as our next two lemmas show .
[ lem : perext ] suppose the scenario @xmath2 is past - periodic on @xmath61 .
then there is a periodic scenario @xmath14 with the property that @xmath15 .
choose @xmath55 so that @xmath2 is past - periodic on @xmath61 with period @xmath32 , and define an equivalence relation @xmath63 on @xmath10 by @xmath64 let @xmath65_b$ ] denote the equivalence class of @xmath66 under @xmath63 . since @xmath2 is past - periodic on @xmath61 with period @xmath32 , it is constant on @xmath65_b \cap ( -\infty , x)$ ] , for every real number @xmath66 . to define @xmath14 ,
we simply let the value of @xmath14 on any equivalence class @xmath65_b$ ] be the constant value of @xmath2 on @xmath65_b \cap ( -\infty , x)$ ] .
it is easy to see that @xmath14 agrees with @xmath2 on @xmath61 and @xmath14 is periodic .
[ lem : afext ] suppose the scenario @xmath2 is past - affine - invariant on @xmath61 .
then there is a scenario @xmath14 that is affine - invariant and has the property that @xmath15 .
since @xmath2 is past - affine - invariant on @xmath61 , there is some nonidentity function @xmath47 such that @xmath62 . for any positive integer @xmath67 , let @xmath68 denote the @xmath67-fold composition of @xmath58 with itself
similarly , let @xmath69 denote the @xmath67-fold composition of @xmath70 with itself , and let @xmath71 be the identity function .
now if we define @xmath72 to mean that there is some integer @xmath67 such that @xmath73 , then @xmath74 is an equivalence relation on @xmath10 .
the rest of the proof is similar to the proof of lemma [ lem : perext ] : @xmath2 is constant on the intersection of each equivalence class with @xmath61 , and we can define @xmath14 to agree with @xmath2 on @xmath61 and to be constant on all equivalence classes .
we will also need to know that the periods in periodicity and past - periodicity of a function always match .
[ lem : permatch ] suppose @xmath2 is past - periodic on @xmath61 with period @xmath32 .
if @xmath2 is also periodic , then it has period @xmath32 .
suppose @xmath2 is periodic with period @xmath75 . to see that @xmath32 is also a period of @xmath2 , consider any real number @xmath66 .
choose an integer @xmath67 such that @xmath76 is less than @xmath5 .
then @xmath77 we are now ready to prove that there is a @xmath49-anonymous predictor that performs very well .
[ thm : main1 ] there is a @xmath49-anonymous predictor @xmath20 such that for every scenario @xmath2 , @xmath21 is countable . in other words ,
only countably many agents guess incorrectly .
the predictor we use is similar to the one described in @xcite . using the axiom of choice ,
fix a well ordering @xmath22 of @xmath78 that lists periodic scenarios first , then other affine - invariant scenarios , and then the rest of the scenarios .
suppose @xmath2 is an arbitrary scenario and @xmath5 is an arbitrary agent .
let @xmath24_x = \ { g \in { } ^{\mathbb{r } } s : f \upharpoonright ( -\infty , x ) = ( g \circ t ) \upharpoonright ( -\infty , x ) \text { for some $ t \in t_2$}\}$ ] .
let @xmath25 denote the @xmath22-least element of @xmath24_x$ ] .
to define @xmath13 , choose any @xmath47 for which @xmath79 , and let @xmath80 we first need to show that @xmath20 is well defined that is , that the value of @xmath13 does nt depend on the choice of @xmath58 .
suppose that for some scenario @xmath2 and some agent @xmath5 , there are two distinct functions @xmath81 such that @xmath82 in order to prove that @xmath13 is well defined , we need to show that @xmath83 . in other words , letting @xmath84 and @xmath85 , we must show that @xmath86 . of course
, if @xmath87 then this clearly holds , so we may assume that @xmath88 .
let @xmath89 .
then @xmath90 , and it follows from that @xmath91 and @xmath92 in particular , reveals that @xmath25 is past - affine - invariant on @xmath93 , so lemma [ lem : afext ] tells us that there is an affine - invariant function @xmath14 that agrees with @xmath25 on @xmath93 .
therefore @xmath94 agrees with @xmath2 on @xmath61 , so @xmath14 is an element of @xmath24_x$ ] .
since @xmath25 is the @xmath22-least element of @xmath24_x$ ] , it follows that @xmath95 .
since @xmath22 lists affine - invariant scenarios first , and @xmath14 is affine - invariant , we deduce that @xmath96 is affine - invariant as well .
so we may fix an @xmath97 in @xmath49 with the property that @xmath97 is not the identity function and @xmath98 an immediate consequence of this is that @xmath99 . by replacing @xmath97 with @xmath100 if necessary , we may assume that @xmath101 . to prove that @xmath86 , we now consider two cases .
case 1 : @xmath97 and @xmath102 commute .
if @xmath103 , then @xmath104 since @xmath105 , this means that the linear function @xmath97 has two distinct fixed points a contradiction , since @xmath97 is not the identity function .
hence @xmath106 , and since we have @xmath101 , we conclude that @xmath107 .
therefore @xmath108 as required
. case 2 : @xmath97 and @xmath102 do not commute .
consider any @xmath109 .
in what follows , we abbreviate expressions like @xmath110 to @xmath111 to prevent parentheses from stacking up .
note that @xmath109 implies that @xmath112 , and therefore @xmath113 .
thus , @xmath114 we let the reader verify that @xmath115 is a shift function , and since @xmath97 and @xmath102 do nt commute , it is not the identity function .
therefore we can fix @xmath55 so that @xmath116 . equation now tells us that @xmath117 ; in other words , @xmath25 is past - periodic on @xmath118 with period @xmath32 . by lemma [ lem : perext ] we know that there is a periodic scenario @xmath14 that agrees with @xmath25 on @xmath118 .
fix such a @xmath14 .
then @xmath119 agrees with @xmath2 on @xmath61 , and therefore @xmath14 is an element of @xmath24_x$ ] .
hence , @xmath95 .
since @xmath14 is periodic , and @xmath22 lists periodic scenarios first , @xmath25 must also be periodic . by lemma [ lem : permatch ]
, @xmath25 has period @xmath32 .
putting this all together , we have shown that @xmath120 replacing @xmath32 with @xmath121 if necessary , we may assume that @xmath122 , so that @xmath123 for all @xmath66 in @xmath10 .
therefore for every @xmath124 , @xmath125 , so @xmath126 it is easy to verify that @xmath127 is also a nonidentity shift function , so we can choose some @xmath128 such that @xmath129 . thus , equation says that @xmath130 .
in other words , @xmath25 is past - periodic on @xmath131 with period @xmath75 .
as before , we can use lemma [ lem : permatch ] to deduce that @xmath132 to complete our proof that @xmath20 is well defined , we can now compute @xmath133 we must now show that @xmath20 is @xmath49-anonymous . given a scenario @xmath2 and an element @xmath58 of @xmath49 , we need to show that @xmath134 . fix an arbitrary agent @xmath5 .
we first claim that @xmath135_x = [ f]_{t(x)}$ ] . to see why ,
suppose @xmath14 is an element of @xmath135_x$ ] .
fix @xmath136 so that @xmath137
. then @xmath138 .
since @xmath139 , it follows that @xmath14 is in @xmath24_{t(x)}$ ] .
the other inclusion has an almost identical proof .
the fact that @xmath135_x = [ f]_{t(x)}$ ] implies that the two sets have the same @xmath22-least element : @xmath140 since @xmath141 is in @xmath24_{t(x)}$ ] , we can choose some @xmath142 such that @xmath143 , and therefore @xmath144 applying the definition of @xmath20 ( equation ) , we have @xmath145 since @xmath5 was arbitrary , it follows that @xmath146 .
finally , we must prove that with the predictor @xmath20 , only countably many agents guess incorrectly .
let @xmath30 be the set of agents who guess incorrectly ; that is , @xmath147 .
suppose that @xmath5 and @xmath66 are both in @xmath30 and @xmath148 .
we claim that @xmath149 .
to prove this claim , we first show that @xmath24_y \subseteq [ f]_x$ ] .
consider an arbitrary @xmath14 in @xmath24_y$ ] .
fix a @xmath58 in @xmath49 so that @xmath2 agrees with @xmath150 on @xmath151 . since @xmath152 , it follows trivially that @xmath2 agrees with @xmath150 on @xmath61 ; hence , @xmath14 is in @xmath24_x$ ] . since @xmath24_y \subseteq [ f]_x$ ] , the @xmath22-least element of @xmath24_y$ ] greater than or equal to the @xmath22-least element of @xmath24_x$ ] . in other words , @xmath153 .
we now claim that @xmath154 .
suppose , toward contradiction , that @xmath155 .
choose a @xmath58 in @xmath49 such that @xmath2 agrees with @xmath156 on @xmath151 .
then by the definition of @xmath20 , @xmath157 where the last equal sign follows from the fact that @xmath2 and @xmath158 agree on @xmath151 and @xmath159 . however , this is a contradiction , since we assumed that @xmath5 guesses incorrectly .
so @xmath154 .
combining @xmath153 and @xmath154 , we have @xmath160 . now , suppose @xmath30 has an infinite descending chain @xmath161 .
by the claim we have just proven , this gives us an infinite descending chain @xmath162 in @xmath163 , contradicting the fact that @xmath22 is a well ordering .
thus @xmath30 has no infinite descending chain , so it is well ordered , and therefore countable .
in this section , we construct a scenario in which agents using _ any _ @xmath50-anonymous predictor perform poorly . broadly speaking
, this scenario ensures that every agent sees the same thing looking into the past , up to distortion by elements in @xmath50 .
the condition of @xmath50-anonymity then implies that every agent guesses the same thing .
the scenario takes each of its values only countably many times , so only countably many agents guess correctly .
we spend the rest of the section working out the details of this argument .
central to the construction of this scenario will be functions that are _ smooth _ ( that is , infinitely differentiable at each point in their domain ) , but fail to be analytic at certain points . for example the function @xmath164 is smooth everywhere , but fails to be analytic at 0 . note in particular that for every positive integer @xmath165 , @xmath166 , where @xmath167 denotes the @xmath165th derivative of @xmath168 .
we can use @xmath168 to define a _ smooth transition function _
\to [ 0,1]$ ] as follows : @xmath170 the graph of @xmath97 transitions smoothly from being very flat at the point @xmath171 to being very flat at the point @xmath172 ( see figure [ fig : s ] ) .
more precisely , we have @xmath173 , @xmath174 , @xmath97 is increasing and smooth on @xmath175 $ ] , and for every positive integer @xmath165 , @xmath176 . (
the subscripts @xmath177 and @xmath178 here indicate that these are one - sided derivatives . ) by shifting and rescaling @xmath97 , we can define similar smooth transition functions connecting other points .
let @xmath179 and @xmath180 be two points in @xmath181 , and suppose that @xmath182 and @xmath183 ; we say that @xmath184 is _ above and to the right of _
we define the function @xmath186 \to [ q_1 , q_2]$ ] by the equation @xmath187 the graph of @xmath188 passes through the points @xmath185 and @xmath184 , and @xmath188 is increasing and smooth on @xmath189 $ ] .
when we use the notation @xmath188 , we implicitly assume that @xmath184 is above and to the right of @xmath185 .
notice that since @xmath188 is increasing , it is invertible
. also , by the chain rule , we have @xmath190 and using the chain rule @xmath165 times , we have @xmath191 in particular , at the endpoints of the domain we have the one - sided derivatives @xmath192 .
thus , @xmath188 is very flat at @xmath185 and @xmath184 . before continuing ,
we note one reason that these smooth transitions functions will be useful to us .
suppose that @xmath185 , @xmath184 , and @xmath193 are three points in the plane , with @xmath193 above and to the right of @xmath184 , which is above and to the right of @xmath185 .
if we concatenate @xmath188 and @xmath194 ( in the obvious way ) , we obtain a smooth increasing function passing through all three of the points @xmath185 , @xmath184 , and @xmath193 . repeating this reasoning
, we can create a smooth increasing function passing through a sequence of points going up and to the right in @xmath181 .
we now use our smooth transition functions to define an equivalence relation on @xmath10 .
we will say that a point @xmath195 in @xmath181 is a _ rational pair _ if both @xmath196 and @xmath197 are rational numbers .
since the set of rational pairs is countable , the set @xmath198 is countable as well . for real numbers
@xmath5 and @xmath66 , we define @xmath199 to mean that there is some @xmath200 such that either @xmath201 or @xmath202 .
we will say that a finite sequence of numbers @xmath203 is an _
@xmath204-path of length @xmath67 _ from @xmath5 to @xmath66 if @xmath205 now we let @xmath206 mean that there is an @xmath204-path from @xmath5 to @xmath66 . for example , if @xmath199 , then @xmath207 is an @xmath204-path of length 2 from @xmath5 to @xmath66 , so @xmath206 .
for any real number @xmath5 , the one - term sequence @xmath208 is an @xmath204-path of length 1 from @xmath5 to @xmath5 , and therefore @xmath209 ; in other words , @xmath210 is reflexive .
we have already seen that @xmath210 is reflexive .
clearly @xmath212 is symmetric .
it follows that the reverse of an @xmath204-path is another @xmath204-path , and therefore @xmath210 is symmetric .
transitivity of @xmath210 can be proven by concatenating @xmath204-paths . to see that the equivalence classes are countable , observe that since @xmath204 is countable , for any @xmath3 there are only countably many @xmath66 such that @xmath199 .
it follows that there are only countably many @xmath204-paths starting at @xmath5 . for each @xmath217
, use the mean value theorem to fix a @xmath218 so that @xmath219 . rearranging this equation and taking the limit of both sides yields
@xmath220 the left side of this equation is the definition of @xmath221 , and the right side equals @xmath222 by hypothesis , so @xmath223 .
but now the same reasoning can be applied to @xmath224 , yielding @xmath225 .
since we can continue to repeat this reasoning , the lemma follows by induction .
[ thm : main2 ] there is a set @xmath0 and a scenario @xmath226 such that for every @xmath50-anonymous predictor @xmath20 , @xmath227 is countable . in other words ,
only countably many agents guess correctly .
let @xmath228_{\sim_f^ * } : x \in \mathbb{r } \}$ ] , and let @xmath1 be defined by the equation @xmath229_{\sim_f^*}.\ ] ] thus , the state of the scenario @xmath2 at time @xmath5 is @xmath211_{\sim_f^*}$ ] .
suppose @xmath20 is a @xmath50-anonymous predictor .
we claim that using this predictor , all agents make the same guess in the scenario @xmath2 .
assuming this claim , we can let @xmath211_{\sim_f^ * } \in s$ ] be the common guess made by all agents .
then for every @xmath230 , @xmath231_{\sim_f^ * } = [ y]_{\sim_f^*}\\ & \leftrightarrow y \in [ x]_{\sim_f^*}.\end{aligned}\ ] ] in other words , the countably many agents in the equivalence class @xmath211_{\sim_f^*}$ ] guess correctly , and everyone else guesses incorrectly .
thus , proving this claim will suffice to prove the theorem . to construct @xmath58 ,
let @xmath240 be an increasing sequence of rational numbers converging to @xmath232 with @xmath241 .
let @xmath242 to be an increasing sequence of rationals converging to @xmath233 with the property that for every @xmath243 , @xmath244 notice that for every @xmath243 we have @xmath245 and @xmath246 and therefore @xmath247 for each negative integer @xmath54 ,
let @xmath248 and @xmath249 .
for every integer @xmath54 , let @xmath250 , and for @xmath251 let @xmath252 .
we claim that the following function @xmath58 has the properties that we want : @xmath253 $ , } \\
s_{b_i b_{i+1}}(x ) , & \text{if $ x \in [ w+i , w+i+1]$ , where $ i \ge 0$. } \\
\end{cases}\ ] ] it is not hard to see that @xmath58 is an increasing bijection from @xmath10 to @xmath10 ; therefore it is continuous on @xmath10 . by our earlier comment about concatenating smooth transition functions
, @xmath58 is smooth everywhere except perhaps at @xmath232 .
it is also clear that for every positive integer @xmath165 , @xmath254 .
if we can show that @xmath255 as well , then it will follow that @xmath58 is smooth at @xmath232 , and therefore @xmath236 . to do this , by lemma [ lem : smooth ] , it will suffice to show that @xmath256 . fix a positive integer @xmath165 , and suppose @xmath257 . according to ,
@xmath258 where @xmath97 is the function defined in .
since @xmath97 is smooth , each of its derivatives is continuous . using the extreme value theorem
, we let @xmath259 be the maximum value of @xmath260 for @xmath261 $ ] . then for each @xmath5 in @xmath262
$ ] , we have @xmath263 using the inequalities @xmath264 and @xmath257 , and then applying @xmath265 , we have @xmath266 combining and , we conclude that if @xmath257 and @xmath267 $ ] , then @xmath268 now , to show that @xmath256 , suppose @xmath269 .
fix @xmath270 large enough that @xmath271 .
now suppose @xmath272 . then @xmath273 $ ] for some @xmath274 . by ,
@xmath275 as required .
therefore , @xmath58 is an element of @xmath50 . since the graph of @xmath58 passes through @xmath276 , @xmath237 .
all that remains is to show that @xmath2 agrees with @xmath48 on @xmath213 .
suppose @xmath217 .
then @xmath277 $ ] for some integer @xmath67 , and @xmath278 .
note that @xmath279 , since @xmath280 . therefore @xmath281_{\sim_f^ * } & & \text{(by the definition of $ f$)}\\ & = [ x]_{\sim_f^ * } & & \text{(since $ s_{a_n a_{n+1}}(x ) \sim_f^ * x$)}\\ & = f(x ) & & \text{(by the definition of $ f$).}\end{aligned}\ ] ] there is plenty to discover about @xmath42-anonymous predictors when @xmath282 .
in particular ,
what if @xmath42 consists of increasing polynomials or , alternatively , increasing analytic functions ?
what if @xmath42 consists of increasing computable functions ? at what point do agents start to perform poorly ?
9 hardin , c. s. and taylor , a. d. , a peculiar connection between the axiom of choice and predicting the future , _ amer .
math . monthly _ * 115 * ( 2008 ) , pp .
hardin , c. s. and taylor , a. d. , _ the mathematics of coordinated inference : a study of generalized hat problems _ , springer international publishing , cham , switzerland , 2013 . | consider an arbitrary set @xmath0 and an arbitrary function @xmath1 .
we think of the domain of @xmath2 as representing time , and for each @xmath3 , we think of @xmath4 as the state of some system at time @xmath5 .
imagine that , at each time @xmath5 , there is an agent who can see @xmath6 and is trying to guess @xmath4in other words , the agent is trying to guess the present state of the system from its past history . in a 2008 paper , christopher hardin and alan taylor
use the axiom of choice to construct a strategy that the agents can use to guarantee that , for every function @xmath2 , all but countably many of them will guess correctly . in a 2013 monograph
they introduce the idea of anonymous guessing strategies , in which the agents can see the past but do nt know where they are located in time . in this paper
we consider a number of variations on anonymity .
for instance ,
what if , in addition to not knowing where they are located in time , agents also do not know the rate at which time is progressing ?
what if they have no sense of how much time elapses between any two events ? we show that in some cases agents can still guess successfully , while in others they perform very poorly . |
a simple atomic system in its liquid state is a prototypical strongly coupled system with no small expansion parameters available to simplify calculations . under conditions where classical mechanics is applicable
there are a number of theoretical and computational methods available to address such strong coupling conditions .
these include integral equation and density functional methods , and numerical techniques such as molecular dynamics ( md ) and monte carlo ( mc ) simulation . for quantum systems
there are complementary many - body methods to address the new features of diffraction and exchange or degeneracy effects
. conditions of both quantum degeneracy and strong coupling pose challenging problems with few choices for accurate computation of properties sensitive to both features ( e.g. , diffusion and path integral monte carlo ( pimc ) simulation ) .
one approach to describe such quantum systems is to extend the classical methods in a phenomenological way .
for example , md has been used with effective potentials modified to include short range diffraction effects @xcite . a related approach , wave packet md , uses a similar variant of classical mechanics to define trajectories guiding wave - packets .
an interesting new approach is to use classical integral equations for pair correlations modified to include an effective potential and an effective temperature dharma00 .
quantum effects are imbedded in these effective properties , while strong correlations are generated by the classical form of the integral equation . a surprisingly wide class of quantum systems
have been described in this way with considerable success @xcite .
the objective here is to formalize the exploitation of classical methods by defining a classical statistical mechanics whose equilibrium properties are the same as those for an underlying quantum system .
the exact definition is given in the next section , providing the basis for the introduction of practical approximations .
one simple realization of electron gas as a classical system is described in the following sections .
the emphasis here is on the definition of the mapping and its approximate evaluation .
applications will be described elsewhere .
consider for simplicity a one component system at equilibrium represented in the grand canonical ensemble .
the grand potential @xmath0 is proportional to the pressure and is given by@xmath1here @xmath2 denotes a trace over properly symmetrized @xmath3 particle states .
the thermodynamics for the system is specified as functions of the inverse temperature @xmath4 and functionals of the local chemical potential @xmath5 ( an external potential has been included for generality ) and pair potential @xmath6 .
the hamiltonian @xmath7 contains the kinetic energy @xmath8 and sum over pair potentials : @xmath9 .
a corresponding classical system is considered with hamiltonian @xmath10 in the same volume @xmath11 at equilibrium described by the classical grand canonical ensemble , with inverse temperature @xmath12 and local chemical potential @xmath13 .
the hamiltonian has the same form except the potential energy functions @xmath14 and @xmath15 are different .
the classical grand potential is defined in terms of these quantities by@xmath16here , @xmath17 is the thermal de broglie wavelength associated with the classical temperature .
the integration for the partition function is taken over the @xmath3 particle configuration space .
the classical system has undefined ingredients : the effective inverse temperature , @xmath18 , the local chemical potential , @xmath19 , and the pair potential for interaction among the particles , @xmath14 .
a correspondence between the classical and quantum systems is defined by expressing these quantities as functions or functionals of @xmath4 , @xmath20 , and @xmath21 .
this is accomplished by requiring the numerical equivalence of two independent thermodynamic properties and one structural property .
the first two are chosen to be the equivalence of the grand potential and its first functional derivative with respect to the local chemical potential .
@xmath22an equivalent form for these conditions can be given in terms of the pressure and density@xmath23these two relations provide two independent relations between @xmath18 , @xmath24 and the physical variables @xmath4 and @xmath25 .
it remains to have a structural equivalence to relate the pair potential @xmath14 to @xmath26 , which are two particle functions .
this is accomplished by equating the functional derivatives of the grand potentials with respect to these pair potentials@xmath27an equivalent form for this definition is the equivalence of pair correlation functions@xmath28 in this way the three equations of ( [ 2.5 ] ) and ( [ 2.7 ] ) determine , formally , the classical parameters @xmath29 @xmath30 and @xmath31 as functions of @xmath32 and functionals of @xmath25 , and @xmath33 : @xmath34 .
this completes the definition of the thermodynamics of a classical system that is representative of that for a given quantum system .
it is also possible to deduce from this a classical density functional theory that is representative of the quantum density functional theory .
this will be discussed elsewhere .
is finite at @xmath35 . , title="fig:",width=302,height=207 ] is finite at @xmath35
. , title="fig:",width=302,height=207 ] exact representations of the classical parameters as functions / functionals of the quantum properties follow from classical density functional theory which has the form@xmath36here @xmath37 and @xmath38 denote formal functionals of the local density and pair correlation function . but according to the definitions ( 2.5 ) and ( [ 2.7 ] ) these can be replaced by their corresponding quantum properties to give the desired result @xmath39 this provides the formal expressions for the classical local chemical potential and classical pair potential in terms of properties of the quantum system , suitable for the introduction of approximations . for example a good approximation to @xmath40 for coulomb systems is the hypernetted chain ( hnc ) form @xmath41 , where @xmath42 is the direct correlation function , defined in terms of @xmath43 by the ornstein - zernicke equation .
it remains to find an equivalent expression for @xmath18 .
this is obtained from the first definition of ( [ 2.5 ] ) and the classical virial equation for the pressure , which has the form @xmath44 \rightarrow \mathcal{k}% \left [ \beta _ { c}\mu _ { c},\beta _ { c}\phi _ { c},g\right ] .
\label{3.4}\]]in the last step @xmath45 has been replaced by @xmath46 according to ( [ 2.7 ] ) . since @xmath47 and @xmath48 are known from ( [ 3.2 ] ) , @xmath49 is given entirely in terms of properties of the quantum system .
an interesting first illustration of this correspondence of classical and quantum system is the uniform ideal fermi gas .
the quantum description of this non - interacting system with only exchange effects is of course simple to analyze .
however , the corresponding classical system is a true many - body problem since those same effects translate into a pair potential among all particles @xcite .
the full many - body problem must be solved to find that effective potential . here
that problem is addressed in terms of the approximate hnc integral equation .
only the results will be given . in figure
( * 1a * ) the classical pair potential is shown as a function of @xmath50 where @xmath51 is the thermal de broglie wavelength defined in terms of the quantum @xmath4 .
a family of curves is given for different values of the degeneracy parameter @xmath52 for the quantum system .
of interest is the development of long range @xmath53 behavior as the temperature is lowered ( large z ) . for this reason the mapping to a classical system requires a uniform neutralizing background to cancel this divergent behavior .
the classical temperature is determined from the virial equation , with @xmath54 defined as above from the hnc integral equation .
this is shown in figure ( * 1b * ) . as expected for @xmath55 while @xmath56 approaches a finite value for @xmath35 .
it is interesting to note that while the pressures of the classical and quantum systems are the same by definition , the internal energies are the same as well .
however , this is true only if the internal energy is defined thermodynamically and not mechanically .
showing the regularization of coulomb potential at the origin and comparison with the kelbg potential.,width=340,height=264 ] [ sec5]a practical classical representation for the uniform interacting electron gas can be obtained in a similar way .
the classical potential is divided into its ideal gas contribution @xmath57 plus a contribution from the coulomb interactions @xmath58 : @xmath59 .
an approximation for the latter is obtained by requiring the correct rpa form at weak coupling . in the classical context
this means that the direct correlation function is given by the classical effective potential .
this is obtained from the ornstein - zernicke equation with the classical @xmath60 replaced by the quantum @xmath54 , according to the definition ( [ 2.7 ] ) , evaluated in the rpa . the final result for the effective classical potential is@xmath61 .
\label{5.1}\]]here @xmath62 is the ideal gas static structure factor and @xmath63 is the rpa structure factor .
it is possible to show that the familiar kelbg potential @xcite is recovered in the weak coupling , low density limit .
figure ( * 2 * ) illustrates the quantum corrections to the bare coulomb form for @xmath64 and @xmath65 ( where @xmath66 is the usual ion sphere radius relative to the bohr radius , and @xmath67 is the fermi temperature ) .
this research has been supported by nsf / doe partnership in basic plasma science and engineering award de - fg02 - 07er54946 nsf / doe , and by us doe grant de - sc0002139 . | a quantum system at equilibrium is represented by a corresponding classical system , chosen to reproduce the thermodynamic and structural properties .
the objective is to develop a means for exploiting strong coupling classical methods ( e.g. , md , integral equations , dft ) to describe quantum systems .
the classical system has an effective temperature , local chemical potential , and pair interaction that are defined by requiring equivalence of the grand potential and its functional derivatives with respect to the external and pair potentials for the classical and quantum systems .
practical inversion of this mapping for the classical properties is effected via the hypernetted chain approximation , leading to representations as functionals of the quantum pair correlation function . as an illustration ,
the parameters of the classical system are determined approximately such that ideal gas and weak coupling rpa limits are preserved . |
the interiors of gas giant planets provide ground truth for planet formation theories and the properties of materials under high pressure and temperature . accordingly
, many studies aimed at deriving the interior states of giant planets in our solar system have been undertaken in the past half century ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the study of giant planets in our solar system has been recently augmented by the growing body of mass and radius measurements for transiting extrasolar planets .
these measurements have enabled the first studies of the heavy element components of gas giants orbiting other stars , as has been done for the super - neptune hats-7b @xcite and the hot saturn hd 149026b @xcite , and in the statistical characterization of heavy - element enrichment in extrasolar gas giant planets ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . nonetheless , characterizing the interior structure of exoplanets@xmath9 in particular , determining the presence of a heavy element core@xmath9remains challenging , since mass and radius measurements alone can not in general uniquely constrain the interior density profile nor the chemical makeup of a planet . in particular
, determining whether heavy elements are concentrated in the core or distributed uniformly within the envelope is especially difficult for jupiter - sized planets since the large , predominantly light - element envelope masks the signal of the radial distribution of heavy elements .
however , the orbital configuration in a subset of multi - planet systems is such that the dynamical evolution of the system depends on the love number ( @xmath10 ) of its innermost planet @xcite .
the love number ( @xmath10 ) quantifies the elastic deformation response of a planet to external forces and thus encodes information about its interior structure , including clues about its core mass @xcite . utilizing the secular theory of @xcite
, @xcite showed that , in a system of two planets orbiting a central body , @xmath10 of the inner planet can be determined if ( i ) the mass of the inner planet is much smaller than the mass of the central body , ( ii ) the semi - major axis of the inner planet is much less than the semi - major axis of the outer planet , ( iii ) the eccentricity of the inner planet is much less than the eccentricity of the outer planet , ( iv ) the planet is transiting , and ( v ) the planet is sufficiently close to its host star , such that the tidal precession is significant compared to the precession induced by relativistic effects .
the hat - p-13 system is the first and only currently known system to fulfill these criteria .
the hat - p-13 system consists of three bodies in orbit around a central star with a mass of @xmath11 and radius @xmath12 @xcite .
hat - p-13b is a low - eccentricity transiting planet with mass @xmath13 , radius @xmath14 , and an orbital period of 2.9 days @xcite .
hat - p-13c is a radial velocity companion with a minimum mass @xmath15 , an orbital period of 446 days , and an eccentricity of 0.66 @xcite .
this system also exhibits a long - term radial velocity trend indicative of a third companion located between @xmath16 au with a minimum mass of @xmath17 @xcite .
however , @xcite demonstrated that the existence of this third companion does not disrupt the secular dynamics that allows the eccentricity of hat - p-13b ( @xmath0 ) to be related to its love number ( @xmath1 ) . using existing constraints on the orbital eccentricity of hat - p-13b from radial velocity measurements ,
@xcite were able to place an upper bound on the core mass ( @xmath2 ) of 120 @xmath6 ( @xmath18 ) . in this study
we present new observational measurements of secondary eclipses of hat - p-13b ( i.e. when hat - p-13b passes behind its host star ) obtained using @xmath3 , which we use to place stronger constraints on the eccentricity of hat - p-13b .
we combine these new secondary eclipse times with the most recent transit and radial velocity measurements of the system @xcite in order to derive an improved constraint on @xmath1 and @xmath2 .
the paper is structured as follows .
first , we describe our data acquisition , post - processing , and analysis ( section 2 ) .
we then present the results of the secondary eclipse measurements and corresponding determination of the eccentricity , @xmath10 , core mass , and atmospheric properties of hat - p-13b ( section 3 ) .
finally , we discuss the implications of our findings in section 4 .
two observations of hat - p-13 were taken using the infrared array camera ( irac ) on board the @xmath3 ( @xmath19 ; * ? ? ?
* ) , one using the 3.6 @xmath8 m band on ut 2010 may 09 and the other using the 4.5 @xmath8 m band on ut 2010 june 08 , 11 orbits later ( pi j. harrington , program i d 60003 ) .
each dataset comprises 68,608 sub - array images taken with 0.4 s integration times over 8.7 hours of observation .
we extract the utc - based barycentric julian date ( bjd@xmath20 ) , subtract the sky background , and remove transient hot pixels from each of the images as described in @xcite and @xcite . to calculate the flux from the hat - p-13 system in each image
, we first estimate the position of the star on the array using the flux - weighted centroid method @xcite with radii ranging between 2.0 - 5.0 pixels in 0.5 pixel increments .
we then calculate the corresponding stellar flux using a circular aperture with either a fixed or time - varying radius .
we consider fixed radii ranging between 2.0 - 5.0 pixels in 0.5 pixel increments , and calculate the time - varying aperture using the square root of the noise pixel parameter as described in @xcite .
this parameter is proportional to the full width at half max ( fwhm ) of the star s point spread function , and is calculated for each image using a circular aperture with radii ranging between 2.0 - 5.0 pixels in 0.5 pixel increments .
we then either multiply the square root of the noise pixel parameter by a constant scaling value of [ 0.6 , 0.7 , 0.8 , 0.85 , 0.9 , 0.95 , 1.00 , 1.05 , 1.10 , 1.15 , or 1.20 ] pixels or add a constant offset of [ -0.9 , -0.8 , -0.7 , -0.6 , -0.5 , -0.4 , -0.3 , -0.2 , -0.1 , 0.0 , 0.1 , 0.2 , 0.3 , 0.4 , or 0.5 ] pixels in order to determine the aperture radius for each image .
we next create a time series for each photometric aperture where we trim the first 90 minutes ( 11,904 images ) of each time series in order to remove the well - known ramp that occurs at the start of each new telescope pointing ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we replace non - numerical ( nan ) flux values with the median flux value of each time series and replace values that deviate by more than three standard deviations from the local mean , determined from the nearest 100 points , with the local mean .
we compare this approach to one in which we instead trim outliers from our light curves and find that our best - fit eclipse depths and times change by less than 0.2 sigma in both channels .
0.2% of the measurements were outliers or nan in each channel .
we then normalize each time series to one by dividing by the median value .
the photometric time series in both channels is dominated by an instrumental effect related to irac s well - known intrapixel sensitivity variations , combined with the pointing oscillation of the @xmath19 .
we correct for this effect using pixel level decorrelation ( pld ) , as described by @xcite .
this method models the variation in flux intensity in each image due to this instrumental effect by tracking the change in intensity over time within a small box of pixels centered on the flux - weighted centroid .
we use a total of nine pixels arranged in a 3x3 box centered on the position of the stellar centroid .
we remove images from the time series where one of these nine pixels deviates from its mean flux by more than 3@xmath21 ( 0.3% of the data at 3.6 @xmath8 m and 0.1% of the data at 4.5 @xmath8 m ) .
most of these deviations correlate with large pointing excursions in the photometric time series .
we identify two pointing excursions in the 3.6 @xmath8 m data , one of 0.7 pixels for 10 s and one of 0.5 pixels for 20 s , and one of 0.9 pixels for 10 s the 4.5 @xmath8 m data .
we divide the flux in each individual pixel by the summed flux across all nine pixels , weighting each pixel by its contribution to the flux and thereby isolating the instrument noise from astrophysical signals ( see * ? ? ?
* ) , and repeat this operation for each image in our photometric time series .
we also incorporate a constant and a linear term in time to model baseline instrument noise . unlike @xcite
, we do not include a quadratic term because we found that the linear fit has an equivalent rms ( root mean square residual ) to the quadratic fit and so adding the quadratic parameter is not justified .
in addition , the quadratic term was correlated with the eclipse depth in our model fits .
we fit a combined instrumental noise and eclipse @xcite model to the light curve for each combination of photometric apertures listed in sec .
[ subsec : aquisition ] using the ` leastsq ' routine in scipy v0.14.0 with python 2.7.6 and examine the residuals from the best - fit solution in order to determine the optimal aperture set for each bandpass . as discussed in @xcite and @xcite , we first bin the photometric light curves and time series for individual pixels by a factor of 512 ( @xmath224 minute intervals ) before fitting the model , then apply the resulting best - fit model coefficients to the unbinned light curve .
this allows us to identify solutions that minimize noise on longer time scales , which are most important for determining the best - fit eclipse parameters , in exchange for a moderately higher scatter in the unbinned residuals .
we allow the center of eclipse time , eclipse depth , pixel weights , constant , and linear terms to vary as free parameters in our fits .
we excluded from consideration any apertures with an unbinned rms more than 1.1 times that of the aperture with the lowest rms in each band , focusing instead on the subset of apertures with low scatter .
we then compared the relative amounts of time - correlated or `` red '' noise in the remaining apertures by calculating the standard deviation of the residuals as a function of bin size . for light curves with minimal red noise
, we would expect the standard deviation of the residuals to vary by the @xmath23 gaussian scaling relation @xcite where @xmath24 is the number of points in each bin and @xmath25 is the number of bins .
we evaluate the actual amount of red noise in the time series for each aperture by calculating the least - squares difference between the observed and theoretical noise scaling ( fig .
[ fig : rootn ] ) and select the aperture that minimizes this quantity in each bandpass .
we next find the optimal bin size to use to fit the lightcurve in each channel via the same least - squares approach with which we find the optimal aperture . after determining the optimal bin size in each bandpass , we repeat our aperture optimization at the new bin size .
we iterate on searching for the optimal aperture and bin size until we converge on the optimal pairing of aperture and bin size for each bandpass . after optimizing our choice of bin size and aperture
, we found that the 4.5 @xmath8 m light curve displayed a residual ramp - like signal despite our decision to trim the first 90 minutes of data .
we therefore experimented with fits where we trimmed up to 3 hours of data from the start of the light curve ( i.e. , up to the beginning of the eclipse ) .
we found that the best - fit eclipse times were correlated with the amount of data trimmed from the start of the light curve over the full range of trim durations considered , indicating that the ramp extended to the start of the eclipse .
we then considered an alternative approach in which we returned to our original 90 minute trim duration and deliberately used larger than optimal bin sizes in our fits , effectively forcing the models to identify solutions with less structure on long time scales .
we found that fits with bin sizes larger than 100 points ( 40 s ) effectively removed the ramp from the light curve , avoiding the need to increase the trim interval to values larger than 90 minutes .
these fits resulted in best - fit secondary eclipse times approximately 2 minutes ( 0.6 @xmath21 ) earlier than our original fits with a smaller bin size .
we tested for a residual ramp by repeating the large bin size fits with trim intervals ranging from 30 minutes up to 3 hours , and found no evidence for a correlation between the trim interval and the best - fit eclipse time .
we then repeated our optimization for bin size considering bin sizes between 128 - 2048 points in powers of two .
we found that our best - fit eclipse depths and times varied by less than 0.4@xmath21 across this range , and were in good agreement with the best - fit values for the 3-hour trim interval using the smaller bin size .
we also considered fits using a smaller bin size where we included an exponential function of time to account for the observed ramp , but found that this exponential function was a poor match for the shape of the observed ramp .
we speculate that a sum of several exponentials might provide a better fit ( e.g. , * ? ? ?
* ) , but felt that the added free parameters were not justified given the success of using larger bin sizes .
we also find that enforcing larger bin sizes in the 4.5 @xmath8 m channel leads to better agreement of the secondary eclipse timing between the two channels .
we also tried decorrelating instrumental noise in our data using pixel - mapping ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this non - parametric technique constructs an empirical map of the pixel response across the chip by comparing the measured flux from each image to those of other images with similar stellar positions .
we model the pixel sensitivity at each point in our time series using a gaussian spatial weighting function over the 50 nearest neighbors in stellar centroid @xmath26 and @xmath27 position and noise pixel parameter space .
the inclusion of the noise pixel parameter in the weighting ensures that the pixel map incorporates systematics unrelated to changes in the star s position that affect the shape of the stellar point - spread function .
the number of neighbors was chosen to be large enough to adequately map the pixel response across the range of star positions in each eclipse data set while maintaining a reasonably low computational overhead @xcite .
@xcite found that pld is generally more effective in removing time - correlated ( i.e. , red ) noise than other decorrelation methods as long as the range of star positions across the data set remains below @xmath220.2 pixels .
the range of star centroid positions in our eclipse data sets lies below this threshold , and therefore , we expect pld to perform optimally .
we also directly compare the performance of pld for cases where we fit to either the unbinned or optimally binned photometry , as well as to the fit acquired from photometry using the @xcite pixel mapping technique described in the previous paragraph .
we find that that the optimally binned pld have lower levels of correlated noise than the other methods ( fig .
[ fig : rootn ] ) .
in addition , binned pld gives center of eclipse phases in the two bandpasses that are most consistent with each other ( at the 1.3@xmath21 level ) ; the unbinned pld and pixel mapping techniques produced center of eclipse phases consistent at the 2.6@xmath21 and 5.0@xmath21 levels , respectively .
we therefore select the pld technique applied to the binned dataset for our final analysis . for the fits described in the rest of this paper we use the following optimal aperture set and bin size . for the 3.6 @xmath8 m channel
we select a bin size of 21 points ( @xmath228 sec ) , a 3.0 pixel radius aperture to find the centroid , a 2.0 pixel aperture to find the noise pixel parameter , and add 0.3 pixels to the square root of the noise pixel parameter to obtain the aperture within which we sum the flux .
for the 4.5 @xmath8 m channel we select a bin size of 128 points ( @xmath2250 sec ) , a 4.5 pixel radius aperture to find the centroid , a 4.0 pixel aperture to find the noise pixel parameter , and add 0.3 pixels to the square root of the noise pixel parameter to obtain the aperture within which we sum the flux .
we determine the uncertainties on our model parameters using the markov - chain monte carlo ( mcmc ) code emcee v2.1.0
@xcite on python 2.7.6 .
we allow the center of eclipse time , eclipse depth , pixel weights , constant , and linear terms to vary as free parameters in our fits .
we set the uncertainties on individual points in each light curve equal to the standard deviation of the residuals after subtracting the best - fit solution in each bandpass .
we run the mcmc with 250 walkers for 20,000 steps ; the first 5000 steps from each walker were ` burn - in ' steps and removed from the chain . for the observations in the 4.5 @xmath8 m band we found that the 1@xmath21 uncertainties on the rms overlap with the errors theoretically expected in the absence of correlated noise on the timescale of the eclipse ingress and egress ( 30 min ; fig .
[ fig : rootn ] ) and therefore report the uncertainties in measurements from the 4.5 @xmath8 m band directly from the mcmc analysis .
however , for the observations in the 3.6 @xmath8 m band , the calculated rms consistently deviate above the expected improvement with increased binning for timescales longer than 1 min .
we therefore choose a conservative approach and multiply the uncertainties in the center of eclipse time derived from the mcmc in the 3.6 @xmath8 m band by a factor of 1.3 , the factor by which the rms lies above the theoretical improvement at the 30 min timescale @xcite . since the timescale of the eclipse is approximately half of the length of the dataset , we are unable to accurately estimate the red noise on that timescale and so adopt the same factor of 1.3 scaling for the eclipse depth uncertainty in this band .
we next calculate an updated value for the eccentricity of hat - p-13b using the approach described in @xcite .
we fit the available radial velocity observations for this planet from @xcite simultaneously with the best - fit transit ephemeris from @xcite and measured secondary eclipse times from this study .
we first allow the apsides of each planet ( @xmath28 and @xmath29 ) to vary independently and then repeat the fits imposing a prior that the posterior distribution of @xmath28 matches the posterior distribution of @xmath29 that was calculated from the fit in which @xmath28 and @xmath29 were allowed to vary independently .
we use the latter version of the fits in our final analysis , and discuss the rationale for this assumption in sec .
[ subsec : seculardynamics ] and [ subsec : apsalignment ] .
we use the mesa code @xcite , a 1-dimensional thermal evolution model , for interior modeling . in the pressure - temperature space
relevant to hat - p-13b , mesa uses the scvh tables @xcite for the equation of state .
we adopt a solar composition envelope and evolve an array of interior models of hat - p-13b with varying core masses and energy dissipation rates .
specifically , we consider core masses of 0.1 - 80 @xmath6 and dissipation rates equal to 0.05% , 0.10% , or 0.50% of the insolation .
the thermal dissipation range we adopt here encapsulates both ( i ) the energy deposition typically quoted for hot jupiters residing on circular orbits ( e.g. ohmic dissipation , kinetic deposition ) and ( ii ) an additional component of energy arising due to the sustained tidal dissipation ( e.g. , * ? ? ?
* ; * ? ? ?
we calculate the insolation ( @xmath30 ) using an equilibrium temperature of 1725 k @xcite .
we assume that the total mass of hat - p-13b is the best fit value reported by @xcite , 0.906 @xmath31 , and acknowledge that a more recent value ( 0.899@xmath31 , * ? ? ?
* ) is available but that the mass - radius relationship for giant planets is famously independent of mass and so our choice of the @xcite mass makes a negligible difference in our analysis .
we also note that the errors on the mass are negligible compared to the uncertainties inherent in the equation of state ( see * ? ? ?
we assume a bond albedo of zero and a core density of 10 @xmath32 ; varying the core density by 2 @xmath32 has a negligible affect on the radial density profile obtained by mesa .
we let the mesa models evolve for 3.0 gyr , based on the best fit age of 3.5 gyr reported by @xcite .
however , the radial density structure reaches a quasi - steady solution after @xmath221 gyr , so the results are insensitive to the assumed system age . for each pairing of core mass and dissipation rate
we calculate @xmath1 based on the density profile , using the equations of @xcite : @xmath33 + @xmath34 is the radius of the planet and @xmath35 is a dimensionless quantity that is obtained by integrating the ordinary differential equation radially in @xmath36 outward from @xmath37 : @xmath38 + in the above expression , @xmath39 is the density obtained from the density distribution @xmath40 output from mesa , and @xmath41 is the mean density interior to @xmath42 .
note that if the core density is constant then @xmath43 , where @xmath44 is the core radius , ( i.e. @xmath10 is 3/2 for a body of constant density , e.g. , * ? ? ?
we use a linear spline to interpolate the coarse grid of @xmath1 and @xmath45 values , corresponding to various core mass and dissipation input pairings evolved in mesa , along both the core mass axis and the dissipation axis , and extend the grid from @xmath46 to @xmath47 with a linear extrapolation .
once we determine the model values of @xmath1 and @xmath45 for each pair of core mass and dissipation , we evaluate the probability of each core mass and dissipation pairing , given the probability distributions of the measured values of @xmath1 and @xmath45 for the hat - p-13 system .
while the probability distribution for @xmath45 is measured from observation , the probability distribution of @xmath1 must be calculated .
we describe this calculation below .
.hat - p-13 system properties [ cols= " < , < " , ] + @xmath48 @xcite , @xmath49 @xcite the octupole - order secular theory of @xcite , augmented with a description of a tidally - facilitated apsidal advance @xcite , can be used to describe the non - keplerian components of motion in the hat - p-13 system and provides a method by which the relationship between @xmath0 and @xmath1 can be obtained @xcite . in the hat - p-13 system
, tidal dissipation quickly drains energy and acts to circularize the orbit of hat - p-13b .
however , the presence of the distant and highly eccentric hat - p-13c acts to prevent complete circularization of the orbit of hat - p-13b . instead , the system tends towards a nearly elliptic equilibrium point , which acts as an attractor in phase space .
as long as the orbits of hat - p-13b and hat - p-13c are coplanar , this minimization is achieved through aligning the apsides .
apsidal alignment is typically reached within @xmath223 circularization timescales @xcite . however ,
once orbital equilibrium is achieved , both orbits decay slowly and the orbital configuration remains quasi - stable for the rest of the lifetime of the system . in order to maintain alignment of the apsides ,
the apsidal precession of both hat - p-13b and hat - p-13c must be equal .
that is : @xmath50 + the secular apsidal precession of hat - p-13c , @xmath51 , dominates all other contributions to the total apsidal precession of hat - p-13c .
the terms that dominate the apsidal precession of hat - p-13b are the secular precession , @xmath52 , the tidal precession , @xmath53 , and general relativistic precession , @xmath54 .
the minor effects due to rotational precession , @xmath55 , are also included but we neglect the negligible contribution to the apsidal precession from the stellar rotational bulge ( e.g. , * ? ? ?
the equations of apsidal precession are comprehensively discussed in @xcite and given here for convenience : @xmath56 \label{eq : varpi_c}\end{gathered}\ ] ] @xmath57 \label{eq : varpi_bsec}\end{gathered}\ ] ] @xmath58 @xmath59 @xmath60 in the preceding equations , @xmath61 is the newtonian gravitational constant and @xmath62 is the speed of light .
the subscripts ` b ' , ` c ' , and ` * ' denote properties of hat - p-13b , hat - p-13c , and the star , respectively .
@xmath63 is the semimajor axis , @xmath64 is the eccentricity , @xmath24 is the mean motion , @xmath34 is the radius , and @xmath25 is the mass . under the assumption that the apsides are aligned , the @xmath65 terms in eqs .
[ eq : varpi_c ] and [ eq : varpi_bsec ] are zero . since all of the system properties that appear in the equations of apsidal precession have been measured , with the exception of @xmath1 , eq .
[ eq : apsides ] can be rearranged to solve for the love number of hat - p-13b purely in terms of known quantities .
note that it is not necessary to measure the apsidal precession rate of either hat - p-13b nor hat - p-13c , it is sufficient to know only that they are equal .
we construct the posterior probability distribution for @xmath1 from mcmc chains comprising 10@xmath66 normally distributed values for each of the measured hat - p-13 system properties ( table [ tab : properties ] ) using eqs .
[ eq : apsides]-[eq : varpi_gr ] .
we then multiply the probability distributions for @xmath1 and @xmath45 obtained from mesa and map that distribution into a 2-dimensional probability distribution of core mass and heat dissipation .
finally , we obtain the 1-dimensional probability distribution of the core mass of hat - p-13b by marginalizing the 2-dimensional distribution over dissipation , assuming a uniform prior on dissipation between 0.05%-0.5% @xmath30 .
we determine the dayside temperature of hat - p-13b from the measured secondary eclipse depths in each bandpass . to do so , we first calculate the stellar flux by integrating a phoenix stellar flux model @xcite for each bandpass weighted by the subarray average spectral response curve .
we utilize a phoenix model with an effective temperature of @xmath67 = 5700 k , a surface gravity of @xmath68 = 4.0 , and a modestly enhanced metallicity of @xmath69 $ ] = 0.5 . for comparison ,
hat - p-13 has a measured @xmath67 = @xmath70 k , @xmath69 $ ] = @xmath71 @xcite , and @xmath68 = @xmath72 @xcite .
we calculate the flux of the planet as a fraction of the total system flux based on the depth of the secondary eclipse .
we then find the temperature that gives a blackbody curve that , when integrated over its respective bandpass , matches the planetary flux .
we calculate the errors on the temperature by constructing the posterior distribution for the temperature in each wavelength using mcmc chains of length 2.5 @xmath73 10@xmath74 , based on the measured eclipse depths and @xmath75 .
the effective dayside temperature was calculated by taking the error - weighted mean of the best - fit temperatures in each bandpass .
we find that the hat - p-13b secondary eclipses are centered at 2455326.70818 @xmath76 0.00406 and 2455355.87672 @xmath76 0.00226 bjd@xmath20 in the 3.6 @xmath8 m and 4.5 @xmath8 m bands , respectively .
these times are 24.2 @xmath76 5.8 minutes and 15.5 @xmath76 3.3 minutes earlier ( orbital phase 0.49424 @xmath76 0.00139 and 0.49633 @xmath76 0.00079 ) , respectively , than the predicted time based on a circular orbit ( fig .
[ fig : eclipses ] ) , where we have accounted for the 41 s light travel time delay @xcite and the uncertainty in the @xcite ephemeris ( 9.7 and 11 seconds for the 3.6 @xmath8 m and 4.5 @xmath8 m observations , respectively ) .
the eclipse depths for the 3.6 @xmath8 m and 4.5 @xmath8 m channel are 0.0662 @xmath76 0.0113% and 0.1426 @xmath76 0.0130% , respectively ( fig .
[ fig : eclipses ] ) .
these secondary eclipse times are consistent at the 1.3@xmath21 level .
we therefore take the error - weighted mean and find that the observed center of secondary eclipse time occurs 17.6 @xmath76 2.9 minutes earlier ( orbital phase 0.49582 @xmath76 0.00069 ) than the predicted value for a circular orbit .
assuming apsidal alignment , the eccentricities of the orbits for the two innermost planets in this system are @xmath77 and @xmath78 .
we use these eccentricities to calculate a love number for the innermost planet ( @xmath79 of @xmath80 , where values of @xmath81 are inconsistent with the mesa interior models ( i.e. , would require a negative core mass ) .
when we combine this constraint on @xmath1 with the measured planet radius ( @xmath45 ) we find that the core mass of hat - p-13b is less than 25 @xmath6 ( less than 9% @xmath82 ; 68% confidence interval ) , with a most likely core mass of 11 @xmath6 ( 4% @xmath82 ; fig .
[ fig : coremass ] ) .
the constraint from @xmath1 strongly favors smaller core masses , while the constraint from @xmath45 modestly favors larger core masses , up to @xmath83 ( fig .
[ fig : coremass ] ) .
we find best - fit brightness temperatures of 1680 @xmath76 119 k at 3.6 @xmath8 m and 2265 @xmath76 150 k at 4.5 @xmath8 m and compare our measured eclipse depths in each bandpass to atmosphere models from @xcite and @xcite ( fig .
[ fig : atmosphere ] ) .
both models assume a solar composition , plane - parallel atmosphere with molecular abundances set to the local thermal equilibrium values .
the @xcite models assume even heat distribution across the dayside and vary the amount of energy incident at the top of the dayside atmosphere in order to approximate the effects of redistribution to the night side . in these models
the zero redistribution case is labeled as ` 2@xmath84 ' and the full redistribution case is labeled as ` 4@xmath84 ' .
we also consider versions of the model with and without an equilibrium abundance of tio ; when present , this molecule absorbs at high altitudes and produces a temperature inversion in the dayside atmosphere .
the @xcite models account for the presence or absence of a dayside temperature inversion by introducing a gray absorber at low pressures where the opacity @xmath85 can be adjusted as a free parameter .
atmospheric circulation is included as a heat sink between 0.01 - 0.1 bars , where the parameter @xmath86 defines the fractional amount of energy redistributed to the night side and ranges from 0 - 50% ( from no redistribution to the nightside to complete redistribution across both hemispheres ) .
the @xcite model satisfactorily reproduces the observed eclipse depths in both bandpasses when including a dayside temperature inversion due to absorption from tio and relatively efficient circulation between the day and night sides .
although none of the @xcite models are able to match the observed 3.6 micron eclipse depth within the 3@xmath21 uncertainty , we obtain the closest match with models that include an absorber ( @xmath85 = 0.1 ) and relatively efficient circulation ( @xmath86 = 40% ) .
correlations between the apsidal orientation ( @xmath87 ) and eccentricity ( @xmath64 ) introduce errors on the determination of eccentricity of hat - p-13b ( @xmath0 ) . since @xmath0 is relatively small
, we obtain a correspondingly poor constraint on @xmath28 of @xmath88 degrees in fits where we allow @xmath28 to vary independently of @xmath29 .
however , since @xmath89 is large , we are able to measure @xmath29 with an uncertainty of less than a degree ( @xmath90@xmath91 ) . the measured apsidal angles for planets b and c
are thus consistent with apsidal alignment , although the relatively large uncertainties on @xmath28 preclude a definitive determination .
when we allow @xmath28 to vary freely in our fits we find that @xmath92 .
this eccentricity is nonzero at the 3.1@xmath21 level , providing independent confirmation that the orbit of hat - p-13b has not yet been circularized and therefore that the secular orbital coupling mechanism discussed by @xcite and @xcite is applicable to this system .
note that the uncertainty in @xmath0 is more than five times greater than in the case when we assume apsidal alignment .
if the planets are coplanar , their apsides will align in much less than the age of the hat - p-13 system @xcite .
@xcite showed that an initial mutual inclination between the orbits of hat - p-13b and hat - p-13c would evolve to a limit cycle in @xmath0 and apsidal orientation , rather than to a fixed @xmath0 and apsidal alignment .
that study explored the effects of the inclination angle between the orbits of hat - p-13b and hat - p-13c ( @xmath93 ) on @xmath0 and found that if the orbits are nearly coplanar ( @xmath94 ) then the limit cycle in @xmath0 will have a width of less than @xmath95 and the width of the limit cycle of the angle between the apsides is @xmath96 ( calculated from eq .
15 , 16 , and 17 of * ? ? ?
thus , the @xmath0 measured at a particular epoch of the hat - p-13 system is insensitive to this limit cycle if @xmath93 is low .
we propose that @xmath93 is indeed likely to be small , based on both observational constraints and theoretical arguments .
first , the exploration by @xcite found that a configuration of either ( i ) prograde , near - coplanar orbits or ( ii ) @xmath97 is strongly favored .
second , @xcite measured the rossiter - mclaughlin effect @xcite during a transit of hat - p-13b and found that the spin axis of the star and the angular momentum vector of hat - p-13b s orbit are well aligned on the sky ( @xmath98 ) .
this is significant because hat - p-13b orbits far enough from the star that the the orbital precession rate is dominated by torque from hat - p-13c rather than the @xmath99 quadrupole moment of the star @xcite .
if @xmath93 were large , as in case ( ii ) of @xcite , nodal precession of hat - p-13b s orbit around hat - p-13c s orbital axis would ensue , manifesting as cyclic variations in the angle between stellar equator and the orbital plane of hat - p-13b ( @xmath100 ) .
therefore , it is unlikely that a small value for @xmath100 would be measured at a randomly selected epoch unless @xmath93 is small @xcite .
however , the initial orbital configuration of the system is unknown and the sky - projected angle ( @xmath101 ) , rather than the true @xmath100 , is measured , so it is not possible to definitively determine @xmath93 from the rossiter - mclaughlin measurement alone .
we therefore argue that @xmath93 must be small , without attempting to place a definitive upper limit on @xmath93 .
a direct measurement of @xmath93 may be forthcoming by studying transit timing variations ( ttvs ) in the orbit of hat - p-13b , since mutual inclination can induce a detectable ttv signature @xcite .
@xcite found that there is no compelling evidence for large ttvs in the orbit of hat - p-13b , although ttvs of less than 100 s are possible @xcite .
@xcite explored theoretical ttvs for hat - p-13b and found that hat - p-13c should induce ttvs on the order of tens of seconds , and that a precise determination of ttvs would make it possible to discriminate between the two allowed scenarios ( @xmath93 near 0@xmath102 or 130 - 135@xmath102 ) found by @xcite .
astrometry of hat - p-13 could also be used to probe @xmath93 .
we calculate an expected astrometric signal from hat - p-13b of either ( i ) 61 @xmath8as , if the orbit of hat - p-13c is effectively edge - on as seen from earth or ( ii ) 86 @xmath8as if it is inclined at 135@xmath102 as seen from the earth .
astrometry from the gaia mission should be accurate to @xmath2210 @xmath8as @xcite and thus will be sensitive enough to discriminate between these two scenarios .
although a direct measurement of the apsidal precession of the system ( i.e. , @xmath103 ) would allow a direct check of the secular perturbation theory that allows us to calculate @xmath1 , we calculate that the precession rate for this planet is on the order of @xmath104 deg / year and is therefore beyond the reach of current radial velocity observations
. however , the presence of a third companion @xcite in the system may complicate the determination of @xmath93 using any of these methods .
the initial characterization of @xmath2 by @xcite was limited by the relatively large uncertainty in the published eccentricity for the innermost planet .
based on radial velocity data alone they concluded that @xmath2 must be less than 120 @xmath6 at the 1@xmath21 level , and argued that core masses greater than 40 @xmath6 were disfavored based on the required effective tidal dissipation ( @xmath105 ) .
more recently , @xcite used updated measurements of the hat - p-13 system from @xcite to find an allowed range of @xmath1 based on the 1@xmath21 error on @xmath0 by using the polynomial relating @xmath0 and @xmath1 given in @xcite .
they then used that @xmath1 range to place constraints on the interior structure of hat - p-13b using the values of @xmath82 and @xmath45 from @xcite and complex interior models .
their analysis indicated that @xmath2 is less than @xmath106 .
however , caution must be exercised when using the polynomial equation of @xcite , since the shape of the curve strongly depends on all of the measured system parameters ( fig .
[ fig : k2_vs_e ] ) .
in addition , the polynomial does not include uncertainties in the @xmath0-@xmath1 relationship due to observational measurement uncertainties .
our analysis offers an improved estimate of @xmath2 ( less than @xmath107 with 68% confidence ) by taking into account both the change in the dependence of @xmath1 on @xmath0 due to updated measurements of @xmath82 , @xmath108 , @xmath109 , @xmath45 , @xmath110 , @xmath111 , and @xmath89 and the effect of the uncertainties in those measured values on the @xmath0-@xmath1 relationship and @xmath2 determination , which had been neglected in previous studies .
when combined with new radial velocity measurements from @xcite , the secondary eclipse measurements of hat - p-13 provide strong constraints on @xmath0 and our assumption of apsidal alignment further reduces uncertainty on this parameter .
our method also allows us to explore the full probability distribution for @xmath2 instead of only placing an upper bound on its value .
there are several caveats worth mentioning in regard to our estimated core mass .
we note that @xmath10 is only the lowest harmonic describing the internal yielding of a body to external forces , and is thus an inherently degenerate quantity ( as noted for specific models of hat - p-13b by * ? ? ?
the effects of metallicity on atmospheric opacity may also affect the thermal evolution and thus the radial structure of the planet ( as noted for brown dwarfs by * ? ? ?
* ) but are neglected here .
we adopt a solar composition envelope for definitiveness and expect that increasing the metallicity will have only a small effect on our predicted core mass based on the extensive exploration of this effect on interior models performed by @xcite .
we also note that an inhomogeneous heavy element distribution may lead to an overestimation of @xmath2 @xcite .
thus , our estimate is specific to a model with a refractory element core and a solar composition envelope .
imperfect knowledge of the equations of state of materials at high pressure and temperature also introduces additional uncertainties ( e.g. * ? ? ?
* ) that are not accounted for in this study .
in addition , strong constraints on the internal heat dissipation are not available , although we can determine how the uncertainty in the internal dissipation impacts our conclusions for @xmath2 by re - calculating the @xmath2 probability distribution assuming either extremely high or extremely low dissipation rates .
we find that the main effect of the dissipation rate is to shift the peak of the probability distribution for @xmath2 lower for higher values of dissipation , while maintaining a comparable distribution shape . when we specify dissipation as 0.05% @xmath30 , the probability distribution peaks at @xmath2 = 22 @xmath6 . for a dissipation of 0.50% @xmath30 ,
the probability distribution peaks at @xmath2 = 3 @xmath6 .
we therefore conclude that uncertainties in the internal heat dissipation introduce modest , but not overwhelming , uncertainties in the estimate of @xmath2 ( i.e. , lack of knowledge of the heat dissipation yields uncertainties that are within the 1@xmath21 errors from the observational uncertainties ) .
@xcite compare the irradiation temperatures ( @xmath112 ) of a large sample of hot jupiters to their measured dayside brightness temperatures ( @xmath113 ) from secondary eclipse observations , and find that hotter planets appear to have relatively inefficient day - night circulation . for hat - p-13b @xmath114 k , yielding a predicted @xmath115 k ( from fig
. 2 of * ? ? ?
* ) , which is 2@xmath21 above the effective dayside temperature we measure ( @xmath116 k ) .
the @xmath117 that we obtain for hat - p-13b ( @xmath118 ) indicates relatively efficient redistribution of energy to the night side for the case of zero bond albedo ( see fig .
7 of * ? ? ?
* ) , in good agreement with our findings in sec .
[ sec : temperature ] .
the @xmath119 of hat - p-13b also fits the trend of decreasing @xmath119 with lower planetary mass found by @xcite ( their fig .
the circulation model of @xcite , which depends on the equilibrium temperature of the planet , also predicts moderately efficient energy redistribution such that the nightside flux from hat - p-13b should be 0.55 - 0.75 that of its dayside flux , depending on the drag timescale .
our analysis indicates that @xmath2 is comparable to the core masses of jupiter ( @xmath120 , * ? ? ?
* ) and saturn ( @xmath121-@xmath122 , * ? ? ?
* ) in our own solar system .
core accretion models for gas giant planet formation suggest that minimum core masses of approximately 10 @xmath6 are needed in order to form jovian planets , although this limit depends on both the composition of the core and the properties of the gas disk near the planet s formation location ( e.g * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
although our observation is consistent with core accretion theory @xcite , our 1@xmath21 confidence interval extends down to zero core mass and therefore does not preclude alternative formation models such as disk instability ( e.g. , * ? ? ? * ) , nor does it provide a definitive test of post - formation core erosion ( e.g. * ? ? ?
* ; * ? ? ?
work has been undertaken to probe the heavy - element fractions of gas giant planets across a broad range of planets , from the hot super - neptune hats-7b @xcite and hot saturn hd 149026b ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) to super - jupiters ( e.g. gj436b and hat - p-2b ;
the constraints on the heavy - element component of these planets are often accompanied by statements about their inferred core mass , with the caveat that there are degeneracies between models with heavy element cores and models with heavy elements distributed throughout the envelope ( e.g. , * ? ? ?
avenues for partially breaking the degeneracies between thermal evolutionary models with heavy - elements distributed throughout the planet and models with heavy - element cores are available for extremely metal - rich planets , such as hats-7b and hd 149026b .
however , in general , measurements of mass and radii can only be used to constrain the overall fraction of the planetary mass composed of heavy elements .
the inference of a radial distribution of refractory elements , and therefore assertions related to the mass of a solid core , require additional information ( e.g. knowledge of @xmath10 ) . in this regard
hat - p-13b is unique because it is the only member of the extrasolar planetary census for which this additional information exists .
our constraint on the core mass of hat - p-13b is consistent with the determination of heavy - element enrichment , with the accompanying inference of the presence of cores in hot jupiters by @xcite and @xcite .
@xcite invoke the presence of heavy element cores to explain the small radii of the metal - rich 0.60 @xmath31 hat - p-3b and 0.62 @xmath31 xo-2b , and @xcite investigated a sample of 14 hot jupiters and found that a subset of those planets had smaller radii than allowed by models without either a solid core or metal - rich envelope .
we stress , though , that the independent measurement of the degree of central mass concentration , such as done in this paper , is necessary to determine the radial distribution of heavy elements for jovian - mass planets .
finally , we also compare the results of our study to empirical scaling relations from @xcite , which are based on mass and radius measurements from a sample of 15 planets with moderate irradiation levels ( incident flux @xmath123 erg s@xmath124 ) around stars with metallicities ranging from [ fe / h]@xmath125 = -0.030 to + 0.390 .
that study found a positive correlation between the bulk metallicity of a planet and that of its host star and a negative correlation between a planet s mass and its metallicity .
it also provided an empirical relationship relating the heavy element complement of giant planets ( @xmath126 ) to their host star : @xmath127_{*}$ ] . applying this relation to hat - p-13b , which orbits a relatively metal - rich star ( [ fe / h]@xmath128 ; @xcite ) , we find an estimated heavy element mass of @xmath129 , i.e. , 84% of the total mass of hat - p-13b , a much higher percentage than we determine for the core mass of hat - p-13b and also a higher percentage than is found for most of the planets considered by @xcite .
this may indicate that the empirical relation can not be extrapolated to planets around stars with metallicities higher than those of the stars they studied , or that there are additional parameters , such as formation location , that can affect the final core masses for these planets .
other systems analogous to the hat - p-13 system , i.e. , systems that allow us to measure the @xmath10 of the inner planet , will be useful for exploring the distribution of core masses over a larger sample of giant planets .
in order to exploit the models utilized in this study we require that such a planet ( i ) be transiting , ( ii ) have a circularization timescale less than one third of the age of the system , ( iii ) have an equilibrium eccentricity large enough to be measured with high precision ( eq .
36 of * ? ? ?
* ) , and ( iv ) have a @xmath53 comparable to or larger than @xmath54 ( eq
. 12 of * ? ? ?
radial velocity observations of the kepler-424 @xcite , wasp-41 @xcite , hat - p-44 , hat - p-45 , and hat - p-46 @xcite systems indicate that they may have architectures that would make them amenable to this kind of study .
we note that many of the hot jupiters detected by ongoing transit surveys have relatively sparse radial velocity observations , making it difficult to determine whether or not they have a suitable outer companion .
@xcite find that approximately half of all hot jupiters have massive long - period companions , suggesting that there is a high probability that future radial velocity campaigns will discover additional systems analogous to hat - p-13b .
although the current observations of hat - p-13 provide an improved estimate of the innermost planet s orbital eccentricity , the uncertainty in this parameter is still the single largest contribution to the uncertainty in the love number .
we therefore conclude that this system could benefit from additional secondary eclipse measurements .
one of our model limitations is the lack of constraint on the metallicity of hat - p-13b s envelope ( see * ? ? ?
. therefore further atmospheric studies are critical to refine our understanding of hat - p-13b s structure and composition .
atmospheric circulation models for tidally locked planets suggest that high metallicity atmospheres may have less efficient atmospheric circulation than their lower - metallicity counterparts @xcite , which does not appear to be the case for hat - p-13b based on the atmospheric models we perform .
since hat - p-13 is currently one of the most metal - rich stars known to host a hot jupiter , it is intriguing that neither hat - p-13b s core mass nor its atmosphere suggest significant heavy element enrichment .
the hat - p-13 system will likely provide invaluable leverage when exploring the relationship between host star and planetary metallicity .
in addition , full - orbit phase curve observations with spitzer would also allow us to break degeneracies between the planet s dayside albedo and the efficiency of its atmospheric circulation ( e.g. , * ? ? ? * ) .
the possibility of independently constraining both the core mass and the atmospheric properties of hat - p-13b makes this planet an ideal target for future observations .
in this study we present observations of two secondary eclipses of hat - p-13b centered at 2455326.70818 @xmath76 0.00406 and 2455355.87672 @xmath76 0.00226 bjd@xmath20 .
this corresponds to an error - weighted mean eclipse time that is 17.6 @xmath76 2.9 minutes minutes earlier ( at orbital phase 0.49582 @xmath76 0.00069 ) than the predicted time for a circular orbit , indicating that this planet has a non - zero orbital eccentricity .
we fit the measured eclipse times simultaneously with the available radial velocity data in order to derive an eccentricity of @xmath4 for this planet , under the assumption that the orbits of hat - p-13b and hat - p-13c are coplanar . using this eccentricity
, we calculate a corresponding constraint on the planet s love number ( @xmath130 ) .
we then use this @xmath130 and the measured radius of hat - p-13b as constraints on interior structure models , which allow us to directly estimate the mass of the planet s core .
moderate mutual inclinations ( up to @xmath131 between the orbits of hat - p-13b and hat - p-13c ) do not significantly alter the constraint from @xmath0 on the determination of the core mass . we calculate that the core mass of hat - p-13b is less than 25 @xmath6 ( 9% of the planet s mass ; 68% confidence interval ) , with a most likely core mass of 11 @xmath6 ( 4% of the planet s mass ) .
we also use the secondary eclipse depths to find that the dayside temperature is @xmath116 k. comparing these depths and the dayside temperature to models , we find that it is likely that hat - p-13b has a strong atmospheric absorber and efficient dayside energy redistribution . obtaining the love number of hat - p-13b is crucial to determining its core mass because the presence of a modest core in a jupiter - mass planet is typically masked by its overlying envelope . the unique opportunity to independently constrain the core mass and atmospheric properties of this hot jupiter with a modestly sized core
makes the hat - p-13 system an important case study for dynamical constraints on the core masses of gas giant planets .
we thank henry ngo and joseph harrington for insightful discussions .
this work is based on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology , under contract with nasa .
this material is based upon work supported by the national science foundation graduate research fellowship under grant no .
any opinion , findings , and conclusions or recommendations expressed in this material are those of the authors(s ) and do not necessarily reflect the views of the national science foundation .
this work was based in part on observations obtained at the w. m. keck observatory using time granted by the university of hawaii , the university of california , and the california institute of technology .
we thank the observers who contributed to the measurements reported here and acknowledge the efforts of the keck observatory staff .
we extend special thanks to those of hawaiian ancestry on whose sacred mountain of mauna kea we are privileged to be guests . | hat - p-13b is a jupiter - mass transiting exoplanet that has settled onto a stable , short - period , and mildly eccentric orbit as a consequence of the action of tidal dissipation and perturbations from a second , highly eccentric , outer companion . due to the special orbital configuration of the hat - p-13 system ,
the magnitude of hat - p-13b s eccentricity ( @xmath0 ) is in part dictated by its love number ( @xmath1 ) , which is in turn a proxy for the degree of central mass concentration in its interior .
thus , the measurement of @xmath0 constrains @xmath1 and allows us to place otherwise elusive constraints on the mass of hat - p-13b s core ( @xmath2 ) . in this study
we derive new constraints on the value of @xmath0 by observing two secondary eclipses of hat - p-13b with the infrared array camera on board the @xmath3 .
we fit the measured secondary eclipse times simultaneously with radial velocity measurements and find that @xmath4 .
we then use octupole - order secular perturbation theory to find the corresponding @xmath5 . applying structural evolution models
, we then find , with 68% confidence , that @xmath2 is less than 25 earth masses ( @xmath6 ) .
the most likely value of @xmath7 , which is similar to the core mass theoretically required for runaway gas accretion .
this is the tightest constraint to date on the core mass of a hot jupiter .
additionally , we find that the measured secondary eclipse depths , which are in the 3.6 @xmath8 m and 4.5 @xmath8 m bands , best match atmospheric model predictions with a dayside temperature inversion and relatively efficient day - night circulation . |
a number of possibilities to obtain negative coupling between two superconductors in a josephson junction have been proposed.@xcite the negative sign of the josephson current corresponds to the coupling energy minimum at the phase difference @xmath1 between the superconducting electrodes . in superconducting rings containing an odd number of such @xmath1 shifts , the flow of spontaneous supercurrent may lead to anomalous flux quantization .
this has been confirmed in experiments with high-@xmath2 superconductors , where the negative coupling is due to @xmath3-wave symmetry of the superconducting order parameter.@xcite also , reversing the direction of the supercurrent appears to be observed in out - of - equilibrium mesoscopic superconducting structures.@xcite recently , a metastable @xmath1 state has been observed in @xmath4he weak links.@xcite the possibility of @xmath1 coupling through a barrier containing paramagnetic impurities has been proposed first by bulaevskii et al.,@xcite and in superconductor / ferromagnet / superconductor ( s / f / s ) metallic weak links by buzdin at al.@xcite subsequently , the progress in fabrication of s / f multilayers has motivated considerable experimental@xcite and theoretical@xcite investigations of @xmath1 coupling .
theory of proximity coupled s / f superlattices in the dirty limit,@xcite and of atomic - scale s / f superlattices with tunneling between s and f layers,@xcite is well established . a clear evidence of @xmath1 coupling , which has been sought in several s / f systems from the oscillations of the critical temperature , predicted in refs . and
, now appears to be established unambiguously.@xcite recently , another evidence of @xmath1 state is found in the nonmonotonic variation of the critical current in s / f / s junctions.@xcite the modified andreev reflection at s / f interfaces has been explored theoretically , both for @xmath5-wave@xcite and @xmath3-wave superconductors,@xcite and experimentally.@xcite the josephson current in @xmath5-wave superconductor / ferromagnetic insulator / superconductor ( s / fi / s ) junctions has been calculated in several papers,@xcite whereas more recent works deal with @xmath3-wave s / f / s and f / i / s junctions,@xcite and with the general case of unconventional superconductor junctions with ferromagnetic barrier.@xcite at low temperature the andreev reflection is dominant , and the ferromagnetic barrier may induce a characteristic anharmonic current - phase relation with new nodes in @xmath6.@xcite the similar type of anharmonicity may appear in @xmath3-wave pinhole junctions,@xcite or in s / i / s junctions,@xcite as found experimentally in the grain - boundary josephson junctions in high-@xmath7 superconductors.@xcite in this paper we present new consequences of the modified andreev reflection in s / f / s junctions .
we show that coexisting stable and metastable @xmath0 and @xmath1 states appear in the vicinity of a critical strength of the ferromagnetic barrier influence , sec .
this coexistence leads to new modes of the magnetic flux penetration in superconducting loops interrupted by a single ferromagnetic junction ( rf squid ) , or two junctions ( dc squid ) , sec .
iii and iv , respectively .
section v contains a brief summary and discussion .
we consider an s / f / s junction with a ferromagnetic metal barrier of thickness @xmath3 and constant exchange energy @xmath8 .
both superconducting and ferromagnetic metals are assumed clean , with same dispersion relations and with same fermi energies .
the pair potential variation is taken as a step function : @xmath9 in the barrier , and @xmath10 in superconducting electrodes , where @xmath11 is the phase difference at the junction .
the josephson current density is calculated within the quasiclassical theory of superconductivity,@xcite @xmath12 where @xmath13 is the fermi velocity , @xmath14 denotes the angular averaging over the fermi surface , and @xmath15 is the density of states at the fermi level .
the quasiclassical normal green s functions @xmath16 , where @xmath17 , and @xmath18 , are calculated from the eilenberger equations in the previous paper.@xcite for @xmath5-wave pairing in superconducting electrodes and transparent interfaces between two metals , the normal green s function in the barrier is @xmath19 where @xmath20 for @xmath17 respectively , and @xmath21 is the parameter measuring the ferromagnetic barrier influence , @xmath22 being the angle between the direction of @xmath23 and the axis perpendicular to the barrier . for @xmath24 , the green s function , eq .
( 2.2 ) , has the well known form for s / n / s junction.@xcite for spherical fermi surface , the josephson current through the barrier of area @xmath25 , @xmath26 , is @xmath27\over u^3 } \,{\rm d}u , \label{i}\ ] ] where @xmath28 , and the normal resistance is given by @xmath29 .
the temperature dependence of @xmath30 is approximated by @xmath31 where @xmath7 is the critical temperature of superconducting electrodes .
the current - phase relation @xmath6 is calculated numerically as a function of @xmath11 , @xmath32 and of parameters measuring the influence of the ferromagnetic barrier , @xmath33 and @xmath3 .
the results , illustrated in figs . 1 and 2 for fixed thickness of the barrier , @xmath34 , where @xmath35 is the bcs coherence length at zero temperature , remain practically unchanged for any barrier thickness @xmath36 at fixed @xmath33 . the ground state phase difference @xmath37 is determined from the minimum of the s / f / s junction energy@xcite @xmath38 where @xmath39 is the flux quantum , fig .
3 . at high temperature , @xmath40 ,
the current - phase relation is quite close to @xmath41 where @xmath42 signs correspond to @xmath0 and @xmath1 junction , respectively .
the amplitude of oscillations in @xmath6 decreases when @xmath33 increases from zero to the critical value @xmath43 . for @xmath44 , the supercurrent flow in the opposite direction ( fig .
1(a ) ) , and @xmath45 starts to increase .
this is the signature of the @xmath1 junction with ground state phase difference @xmath46 , fig .
3(a ) . at low temperature , for clean and transparent normal metal barrier , the andreev reflection is dominant , and deviations from the sinusoidal @xmath6 law are pronounced.@xcite in the s / f / s case , due to the influence of the exchange interaction in the barrier , the current - phase relation is even more anharmonic .
positive for small @xmath33 , where @xmath0 state is stable , the sign of @xmath6 for @xmath47 is negative for large @xmath33 , where @xmath1 state is stable , and multinode oscillations in @xmath6 appear in the vicinity of @xmath48 , fig .
in the latter case , there are stationary states @xmath49 , @xmath50 both for @xmath51 ( @xmath52 ) and @xmath53 .
one of these states is stable , and the other is metastable .
for example , at @xmath54 for @xmath55 the metastable state is @xmath53 , @xmath51 being the stable state , and _
vice versa _ for @xmath56 , fig .
3(b ) . zero and
@xmath1 states are equally stable at @xmath57 .
while @xmath48 is practically temperature and barrier thickness independent , the interval of the coexistence of @xmath0 and @xmath1 states , @xmath58 , rapidly decreases with increasing temperature . for @xmath59 , @xmath60 and
the coexistence interval is @xmath61 at @xmath62 , while @xmath63 and the coexistence interval is @xmath64 at @xmath65 .
the appearence of the @xmath1 junction for @xmath66 is also manifested in the nonmonotonous dependence of the critical current @xmath67 on @xmath33 at given @xmath32 , fig .
2 . for @xmath68 , @xmath69 again , and near @xmath70 metastable states would appear as well .
however , large @xmath33 would induce the decoupling of s electrodes.@xcite for fixed @xmath8 , such a nonmonotonic variation of the critical current as a function of the barrier thickness @xmath3 has been observed in nb / al / gd / al / nb junctions.@xcite the same effect is also predicted for the atomic - scale s / f superlattices.@xcite recently measured @xmath6 in @xmath4he weak links@xcite are of the same form as the curves in fig .
1(b ) for @xmath71 , corresponding to the metastable @xmath1 state . in @xmath3-wave s / f / s junctions , stable and metastable states may also coexist due to the exchange interaction .
however , in this case the ground state phase differences can vary from @xmath0 to @xmath1 , depending not only on @xmath33 and @xmath32 , but also on the crystal orientation of the superconducting electrodes.@xcite
when s / f / s junction is a part of a superconducting loop ( rf squid ) , the ground state is determined by minimizing the squid free energy @xmath72 , \label{wl}\ ] ] where @xmath73 represents the normalized external magnetic flux through the loop , and @xmath74 is the inductance.@xcite therefore , the total magnetic flux through the loop @xmath75 is related to the external flux @xmath76 by @xmath77 for @xmath78 , and for @xmath33 sufficiently large ( @xmath1 junction ) , there is a flow of spontaneous supercurrent,@xcite with nonmonotonic dependence on the normalized inductance @xmath79 note that when @xmath80 , where @xmath81 , @xmath82 from eqs .
( [ i ] ) and ( [ gsl1 ] ) @xmath83 is calculated numerically .
the results are illustrated in figs .
4 and 5 for several values of @xmath84 and @xmath33 , for @xmath85 . at high temperature , @xmath86 ,
for @xmath87 and @xmath88 the double - well squid energy gives stable and metastable solutions of eq .
( [ gsl1 ] ) , equally stable at @xmath89 . corresponding flux jump between @xmath90 and @xmath91 is indicated by vertical line linking the stable solutions , fig .
in experiments , the hysteretic loops enclosing these lines are usually seen . for @xmath1 junctions , @xmath92 , fig .
4(b ) , there is only one stable solution with @xmath93 , and flux jumps at @xmath94 and @xmath95 , without hysteresis . for @xmath96 ,
the effect of supercurrents in the loop is negligible , and @xmath97 for any @xmath33 . at low temperature ,
similar results are obtained only in the limiting cases of normal metal and strong ferromagnet barriers . for @xmath24 and @xmath98 , there is the usual integer flux quantization , fig .
5(a ) . for @xmath99 and @xmath98
, there is the half - integer flux quantization , fig .
5(d ) , @xmath83 being nonhysteretic in contrast to the previous case .
corresponding free energies are illustrated in fig .
new features appear for @xmath33 close to @xmath48 , related to the presence of stable and metastable minima of the triple - well squid energy , figs .
6(b ) and 6(c ) .
for relatively small inductance , @xmath100 , the squid has two degenerate configurations : the nearly zero and one - half fluxoid state , when @xmath101 , and the nearly one - half and one fluxoid states , when @xmath102 , as shown in fig . 7 .
for example , at @xmath103 , for @xmath104 and @xmath100 , new hysteretic jumps of @xmath105 are found , from @xmath106 to @xmath93 in the interval @xmath107 , and from @xmath108 to @xmath109 in the interval @xmath110 , figs .
5(b ) and 5(c ) .
this anomalous periodicity of @xmath83 resemble to the usual one , fig .
5(a ) , scaled by @xmath111 .
for a loop containing two s / f / s junctions ( dc squid ) the total phase difference is @xmath112 , where @xmath113 and @xmath114 are the phase differences at the first and the second junction , respectively . in the ground state ,
the magnetic flux through the loop @xmath83 is calculated from two coupled equations@xcite @xmath115 and @xmath116 where @xmath117 is the external current , @xmath118 and @xmath119 . as expected ,
@xmath83 dependence for dc squid in the absence of the external current , @xmath120 , is similar to that for rf squid . at high temperature , @xmath86 , and for two @xmath0 junctions , @xmath121 , @xmath122 , or two @xmath1 junctions , @xmath123 , @xmath124 , @xmath83 dependence is qualitatively the same as in a hysteretic single @xmath0 junction loop , fig .
for one @xmath0 junction and one @xmath1 junction , the two - junction loop behaves as a nonhysteretic single @xmath1 junction loop , fig .
4(b ) . at low temperature , outside the coexistence interval of @xmath0 and @xmath1 states ,
for two @xmath0 junctions , @xmath125 , @xmath126 , or for two @xmath1 junctions , @xmath127 , @xmath128 , the two - junction loop behaves as a hysteretic single @xmath0 junction loop . for one @xmath0 junction , and one @xmath1 junction , @xmath125 , @xmath129 or _
vice versa _ , the two - junction loop behaves as a nonhysteretic single @xmath1 junction loop .
for example , at @xmath103 , @xmath83 curves for @xmath130 and @xmath131 would be practically the same as in fig .
5(a ) , and for @xmath132 and @xmath133 as in fig .
5(d ) , if @xmath134 , the first junction normalized inductance @xmath135 corresponding to @xmath84 of a single - junction loop . when @xmath33 , for at least one of the junctions , lies within the coexistence interval , @xmath136 new jumps between the stable solutions appear for sufficiently small inductance , similarly to the single - junction loop .
for example , at @xmath103 and for @xmath134 , @xmath83 curves for @xmath132 and @xmath137 are the same as in fig .
5(b ) , and for @xmath132 and @xmath138 , as in fig .
5(c ) , see fig .
magnetic flux penetration in the case when metastable states exist in both junctions is illustrated for @xmath139 and @xmath137 in fig .
9(a ) . for nonzero external current , up to its maximum value @xmath140
, @xmath141 curves are similar to those for @xmath142 , the flux through the loop becoming closer to the external flux . however ,
when @xmath0 and @xmath1 states coexist , for high external current flux jumps appear for any inductance , as illustrated for @xmath143 in fig .
in general , @xmath144 curves have characteristic dips at the same positions as the jumps in @xmath141 curves .
three typical cases at low temperature are illustrated in fig .
10 for one period of @xmath144 . for two @xmath0 junction loop
, there is the usual @xmath145 dependence,@xcite with a single dip at @xmath89 , fig .
for one @xmath0 and one @xmath1 junction , there are two dips at the ends of the period , @xmath94 and @xmath95 , fig
new characteristic behavior is obtained when there is the coexistence of @xmath0 and @xmath1 states in one of the junctions at least : @xmath145 curve has three maxima and two central dips at the flux jumps positions in @xmath146 , fig .
the ground state phase difference @xmath37 in @xmath5-wave s / f / s junctions is always @xmath0 or @xmath1 . in the vicinity of the critical strength of the ferromagnetic barrier influence @xmath147 , where @xmath0 and @xmath1 states are equally stable
, the characteristic multinode anharmonicity of @xmath6 implies the coexistence of stable @xmath0 and metastable @xmath1 state for @xmath148 , or _
vice versa _ for @xmath149 . for @xmath150 and for @xmath151
only stable states exist , @xmath152 , and @xmath46 , respectively .
the interval of @xmath33 , where the metastable states appear , rapidly decreases with increasing temperature , whereas @xmath48 is practically unchanged .
stable and metastable states can also coexist in @xmath3-wave s / f / s junctions,@xcite in contrast to @xmath3-wave non - magnetic junctions where the similar anharmonicity of the current - phase relation does not lead to the metastable states.@xcite the peculiar low temperature behavior of clean and transparent s / f / s junctions is due to the modification of the andreev reflection by the exchange interaction in the ferromagnet .
the effects of the barrier interfacial nontransparency , and of the scattering on impurities , neglected in this paper , suppress the andreev reflection . as a consequence , anharmonicity of the current - phase relation becomes less pronounced , the value of @xmath48 changes , and the metastability region narrows.@xcite however ,
when these effects are relatively weak , they do not change the above results qualitatively .
we also expect that the suppression of the pair potential near the interfaces , not included in the present non - selfconsistent calculation , would not destroy the metastable states neither.@xcite the coexistence of @xmath0 and @xmath1 states in the s / f / s - junction leads to the coexistence of integer and half - integer fluxoid configurations in squids , corresponding to the minima of the triple - well potential energy , and generating two flux jumps per one external flux quantum .
this is the case when @xmath33 is close to @xmath48 in rf squid , and at least in one of the junctions in dc squid .
another feature is the multipeak dependence of the dc squid critical current on the external magnetic flux . in conclusion , experimental investigation of the predicted new manifestation of the modified andreev reflection
is desirable . beside possible device applications ,
e.g. quantum computing,@xcite the coexistence of integer and half - integer flux quantum configuration in s / f / s junction squids could be also used for the experimental study of the quantum superposition of macroscopically distinct states.@xcite | new modes of the magnetic flux penetration in squids with ferromagnetic junctions are predicted theoretically .
characteristic multinode anharmonicity of the current - phase relation implies the coexistence of stable and metastable @xmath0 and @xmath1 states for a large interval of strength of the ferromagnetic barrier influence at low temperature . as a consequence
, the coexistence of integer and half - integer fluxoid configurations appears in both rf and dc squids , and generates two flux jumps per one external flux quantum . in dc squids , new flux jumps are manifested as two dips in the critical current dependence on the external magnetic flux . |
considerable effort has been made on the characterization of non - local properties of quantum states and operations .
@xcite have computed locally invariant polynomial functions of density matrix elements .
makhlin @xcite has recently analyzed non - local properties of two - qubit gates and presented local invariants for an operation @xmath1 .
makhlin also studied some basic properties of perfect entanglers , which are defined as the unitary operations that can generate maximal entangled states from some initially separable states .
also shown were entangling properties of gates generated by several different hamiltonian operators .
all these results are crucial for physical implementations of quantum computation schemes . determining the entangling capabilities of operations generated by a given physical system is another intriguing and complementary issue .
zanardi @xcite has explored the entangling power of quantum evolutions .
the most extensive recent effort to characterize entangling operations is due to cirac and coworkers @xcite .
kraus and cirac @xcite focused on finding the best separable two - qubit input states such that some given unitary transformation can create maximal entanglement .
@xcite developed the interaction cost for a non - local operation as the optimal time to generate it from a given hamiltonian .
the same group , hammerer _ et al . _
@xcite then extended these considerations to characterize non - local gates .
these works are closely related to time optimal control as addressed recently by khaneja _
@xcite , who studied systems described by a hamiltonian that contains both a non - local internal or drift term , and a local control term .
all these studies assume that any single qubit operation can be achieved almost instantaneously .
this is a good approximation for the situation when the control terms in the hamiltonian can be made large compared to the internal couplings .
universality and controllability are issues of crucial importance in physical implementations of quantum information processing @xcite .
a series of important results have been obtained since questions of universality were first addressed by deutsch in his seminal papers on quantum computing @xcite .
deutsch @xcite proved that any unitary operation can be constructed from generalized toffoli gate operating on three qubits .
divincenzo @xcite proved universality for two - qubit gates by reconstructing three - qubit operations using these gates and a local not gate .
similarly , barenco @xcite and sleator and weinfurter @xcite identified the controlled unitary operation as a universal two - qubit gate .
barenco @xcite showed the universality of the cnot gate supplemented with any single qubit unitaries , and pointed out advantages of cnot in the context of quantum information processing .
lloyd @xcite showed that almost any quantum gate for two or more qubits is universal .
@xcite proved that almost any two - qubit gate is universal by showing that the set of non - universal operations in @xmath2 is of lower dimension than the @xmath2 group .
universal properties of quantum gates acting on an @xmath3 dimensional hilbert space have been studied by brylinski @xcite .
_ @xcite have pointed out that universal quantum computation can be achieved by any entangling gate supplemented with local operations .
_ @xcite recently demonstrated this by extending the results of brylinski , giving a constructive proof that any two - qubit entangling gate can generate cnot if arbitrary single qubit operations are also available .
universal sets of quantum gates for @xmath4-qubit systems have been explored by vlasov @xcite in connection with clifford algebras .
general results on efficient simulation of any unitary operation in @xmath5 by a discrete set of gates are embodied in the solovay - kitaev theorem @xcite and in recent work due to harrow _ et al .
the solovay - kitaev theorem implies the equivalence of different designs of universal quantum computers based on suitable discrete sets of single - qubit and two - qubit operations in a quantum circuit .
an example is the standard universal set of gates including controlled - not and three discrete single - qubit gates , namely hadamard , phase and @xmath6 gates @xcite .
other universal sets have also been proposed . according to the solovay - kitaev theorem
, every such design can represent a circuit that is formulated using the standard set of gates .
consequently , all quantum computation constructions including algorithms , error - correction , and fault - tolerance can be efficiently simulated by physical systems that can provide a suitable set of operations , and do not necessarily need to be implemented by the standard gates .
this moves the focus from study of gates to study of the hamiltonians whose time evolution gives rise to the gates . in this context , burkard _ et al . _
@xcite studied the quantum computation potential of the isotropic exchange hamiltonian .
this interaction can generate @xmath7 gate directly .
however , cnot can not be obtained directly from the exchange interaction .
_ showed that it can be generated via a circuit of two @xmath7 gates and a single - qubit phase rotation .
more recently , whaley and co - workers have shown that the two - particle exchange interaction is universal when physical qubits are encoded into logical qubits , allowing a universal gate set to be constructed from this interaction alone @xcite .
this has given rise to the notion of `` encoded universality '' , in which a convenient physical interaction is made universal by encoding into a subspace @xcite .
isotropic , anisotropic , and generalized forms of the exchange interaction have recently been shown to possess considerable power for efficient construction of universal gate sets , allowing explicit universal gate constructions that require only a small number of physical operations @xcite . in this paper , we analyze non - local two - qubit operations from a geometric perspective and show that considerable insight can be achieved with this approach .
we are concerned with three main questions here .
first , achieving a geometric representation of two - qubit gates .
second , characterizing or quantifying all operations that can generate maximal entanglement .
third , exact simulation of any arbitrary two - qubit gates from a given two - body physical interaction together with single qubit gates .
the fundamental mathematical techniques we employ are cartan decomposition and weyl group in the lie group representation theory .
the application of these theories to the lie algebra @xmath0 provides us with a natural and intuitive geometric approach to investigate the properties of non - local two - qubit operations .
this geometric approach reveals the nature of the problems intrinsically and allows a general formulation of solutions to the three issues of interest here . because it is the non - local properties that generate entanglement in quantum systems , we first study the invariants and geometric representation of non - local two - qubit operations . a pair of two - qubit operations are called locally equivalent if they differ only by local operations .
we apply the cartan decomposition theorem to @xmath0 , the lie algebra of the special unitary group @xmath8 .
we find that the geometric structure of non - local gates is a 3-torus . on the other hand , the cartan decomposition of @xmath0 derived from the complexification of @xmath9
yields an easy way to derive the invariants for local transformations .
these invariants can be used to determine whether two gates are locally equivalent .
moreover , we establish the relation of these local invariants to the coordinates of the 3-torus .
this provides with a way to compute the corresponding points on the 3-torus for a given gate .
it turns out that a single non - local gate may correspond to finitely many different points on the 3-torus .
if we represent these points in a cube with side length @xmath10 , there is obvious symmetry between these points .
we then use the weyl group theory to reduce this symmetry . we know that in this case the weyl group is generated by a set of reflections in @xmath11 .
it is these reflections that create the kaleidoscopic symmetry of points that correspond to the same non - local gate in the cube .
we can explicitly compute these reflections , and thereby show that the local equivalence classes of two - qubit gates are in one - to - one correspondence with the points in a tetrahedron except on the base .
this provides a complete geometric representation of non - local two - qubit operations .
the second objective of this paper is to explore the properties of perfect entanglers , that is , the quantum gates that can generate maximally entangled states from some initially separable states .
we start with criteria to determine whether a given two - qubit gate is a perfect entangler .
a condition for such a gate has been stated in @xcite .
we provide here a proof of this condition and show that the condition can be employed within our geometric analysis to determine which fraction of all non - local two - qubit gates are perfect entanglers .
we show that the entangling property of a quantum gate is only determined by its geometric representation on the 3-torus . using the result that every point on the tetrahedron corresponds to a local equivalence class
, we then show that the set of all perfect entanglers is a polyhedron with seven faces and possessing a volume equal to exactly half that of the tetrahedron .
this implies that amongst all the non - local two - qubit operations , exactly half of them are capable of generating maximal entanglement . finally , we explore universality and controllability aspects of non - local properties of given physical interactions and the potential of such specified hamiltonians to generate perfect entanglers .
our motivation is related to that of encoded universality , namely , determining the potential for universal quantum computation and simulation of a given physical hamiltonian . however , whereas encoded universality sought to construct encodings to achieve universality of quantum logic , here we focus on the simulation of any arbitrary two - qubit gate .
achievement of this , together with our second result above , allows generation of maximal entanglement as well as providing universality . to realize this , we consider here the conventional scenario of a hamiltonian acting on a physical set of qubits such that any arbitrary single qubit operation and certain specific two - qubit operations may be turned on for selected time durations in series .
generally speaking , two - qubit interactions include both local and non - local terms .
the non - local terms can give rise to not only well - known entangling gates such as cnot , but also to many other classes of gates that may or may not lie in the perfect entangling sector .
we therefore seek a systematic way to construct quantum circuits from a given physical hamiltonian that can simulate _ any _ arbitrary two - qubit gate exactly , without restriction by limitations of speed due to time - scales of control interactions ( e.g. as in nmr ) or to fundamental considerations @xcite . as in the study of encoded universality
we start with the gates that can be directly generated by a given hamiltonian .
generally , these gates form a one dimensional subset on the 3-torus . to construct an exact simulation of any arbitrary two - qubit gate , we make use of the quantum circuit model .
we explicitly construct a quantum circuit that contains three non - local gates generated by a given two - body interaction hamiltonian for corresponding finite time durations , together with at most four local gates .
we prove that such a quantum circuit can simulate _ any _ arbitrary two - qubit operation exactly and is therefore universal .
in particular , it can therefore _ efficiently _ provide maximal entanglement from any arbitrary hamiltonian of this form .
such efficient construction from any given hamiltonian is extremely useful for design and experimental implementation of quantum information processing schemes .
in this section , we briefly review some basic facts about cartan decomposition and the weyl group within lie group representation theory @xcite , and then apply these results to @xmath0 , the lie algebra of the special unitary group @xmath8 .
applications of cartan decomposition to quantum system control can also be found in @xcite .
we concentrate on @xmath8 when studying two - qubit gates .
it is well - known that an arbitrary two - qubit gate @xmath12 can be decomposed as the product of a gate @xmath13 and a global phase shift @xmath14 , where @xmath15 .
because the global phase has no significance in quantum mechanics , we can thereby reduce the study of the group @xmath2 of two - qubit quantum evolution operators to @xmath8 .
extensions of results from the group @xmath8 back to @xmath2 are made when appropriate .
we heuristically introduce a partition of the set of two - qubit operations represented by the group @xmath8 .
this set splits into two subsets , one of local gates @xmath16 and the other of non - local gates @xmath17 .
the latter splits further into a set of perfect entanglers , _
i.e. _ , those that can generate maximally entangled states , an example of which is cnot , and the complementary set of those non - local gates that are not perfect entanglers . this schematic partition is illustrated in fig .
[ fig : partition ] .
a rigorous definition of perfect entanglers is presented in sec . [
sec : perfect ] .
[ ] [ ] @xmath18 [ ] [ ] @xmath19 [ ] [ ] @xmath20 [ ] [ ] @xmath21 [ ] [ ] @xmath22 [ ] [ ] @xmath17 [ ] [ ] @xmath16 [ ] [ ] @xmath23 , title="fig : " ] our first goal is to establish fundamentals for a geometric picture of non - local unitary operations with emphasis on their generators , which are represented by hamiltonian operators in physical context .
we start with a summary of some basic definitions @xcite .
consider a lie group @xmath24 and its corresponding lie algebra @xmath25 .
the adjoint representation @xmath26 is a map from the lie algebra @xmath25 to @xmath25 which is the differential of the conjugation map @xmath27 from the lie group @xmath24 to @xmath24 given by @xmath28 . for matrix lie algebras , @xmath29 , where @xmath30 , @xmath31 are both represented as matrices of compatible dimensions .
the differential of the adjoint representation is denoted by @xmath32 , and @xmath33 is a map from the lie algebra @xmath25 to @xmath25 given by the lie bracket with @xmath34 , that is , @xmath35 $ ] .
we now define an inner product on @xmath25 by the killing form @xmath36 .
let @xmath37 be a basis for @xmath25 .
the numbers @xmath38 such that @xmath39=\sum_{i=1}^n c_{jk}^i x_i\ ] ] are the _ structure constants _ of the lie algebra @xmath25 with respect to the basis , where @xmath40 , @xmath41 run from @xmath42 to @xmath4 . since @xmath43&=[\sum_{i=1}^n c_{j1}^i x_i , \dots , \sum_{i=1}^n c_{jn}^i x_i]\\ & = [ x_1 , \dots , x_n]\left [ \begin{matrix } c_{j1}^1&\cdots & c_{jn}^1\\ \vdots&&\vdots \\
c_{j1}^n&\cdots & c_{jn}^n \end{matrix } \right ] , \endaligned\end{aligned}\ ] ] the matrix representation of @xmath44 with respect to the basis is @xmath45.\ ] ] thus , the trace of @xmath46 , which is @xmath47 , is @xmath48 , which is also the @xmath49-th entry of the matrix of the quadratic form @xmath50 .
the lie algebra @xmath25 is semisimple if and only if the killing form is nondegenerate , _
i.e. _ , the determinant of its matrix is nonzero .
let @xmath51 be a compact subgroup of @xmath24 , and @xmath52 the lie algebra of @xmath51 .
assume that @xmath25 admits a direct sum decomposition @xmath53 , such that @xmath54 with respect to the metric induced by the inner product .
let @xmath25 be a semisimple lie algebra and let the decomposition @xmath55 , @xmath54 satisfy the commutation relations @xmath56\subset \k,\quad [ \p , \k]\subset\p,\quad [ \p , \p]\subset \k.\ ] ] this decomposition is called a cartan decomposition of @xmath25 , and the pair @xmath57 is called an orthogonal symmetric lie algebra pair .
a maximal abelian subalgebra @xmath58 contained in @xmath59 is called a _ cartan subalgebra _ of the pair @xmath60 . if @xmath61 is another cartan subalgebra of @xmath60 , then there exists an element @xmath62 such that @xmath63 .
moreover , we have @xmath64 .
[ prop : cartan ] given a semisimple lie algebra @xmath25 and its cartan decomposition @xmath53 , let @xmath58 be a cartan subalgebra of the pair @xmath65 , then @xmath66 . for @xmath67 ,
let @xmath68 be an eigenvector of @xmath33 and @xmath69 the corresponding eigenvalue , _
i.e. _ , @xmath70=\alpha ( x)w.\ ] ] the linear function @xmath71 is called a _ root _ of @xmath25 with respect to @xmath58 .
let @xmath72 denote the set of nonzero roots , and @xmath73 denote the set of roots in @xmath72 which do not vanish identically on @xmath58 .
note that if @xmath74 , it is also true that @xmath75 .
let @xmath76 and @xmath77 denote the centralizer and normalizer of @xmath58 in @xmath51 , respectively .
in other words , @xmath78 the quotient group @xmath79 is called the _ weyl group _ of the pair @xmath80 .
it is denoted by @xmath81 .
one can prove that @xmath81 is a finite group .
each @xmath82 defines a hyperplane @xmath83 in the vector space @xmath58 .
these hyperplanes divide the space @xmath58 into finitely many connected components , called the _
weyl chambers_. for each @xmath82 , let @xmath84 denote the reflection with respect to the hyperplane @xmath83 in @xmath58 .
[ prop : weyl ] the weyl group is generated by the reflections @xmath84 , @xmath85 . this proposition is proved in corollary 2.13 , ch .
vii in @xcite .
now we apply the above results to @xmath0 , the lie algebra of the special unitary group @xmath8 . the lie algebra @xmath86 has a direct sum decomposition @xmath53 , where @xmath87 here @xmath88 , @xmath89 , and @xmath90 are the pauli matrices , and @xmath91 . if we use @xmath92 to denote the matrices in eq .
( [ eq:29 ] ) , where @xmath40 runs from left to right in eq .
( [ eq:29 ] ) , we can derive the lie brackets of @xmath92 and @xmath93 .
these are summarized in the following table : @xmath94&x_1&x_2&x_3&x_4&x_5&x_6&x_7&x_8&x_9&x_{10}&x_{11 } & x_{12}&x_{13}&x_{14}&x_{15}\\ \hline x_1&0&-x_3&x_2&0&0&0&0&0&0&-x_{13}&-x_{14}&-x_{15}&x_{10}&x_{11}&x_{12}\\ x_2&x_3&0&-x_1&0&0&0&x_{13}&x_{14}&x_{15}&0&0&0&-x_7&-x_8&-x_9\\ x_3&-x_2&x_1&0&0&0&0&-x_{10}&-x_{11}&-x_{12}&x_7&x_8&x_9&0&0&0\\ x_4&0&0&0&0&-x_6&x_5&0&-x_9&x_8&0&-x_{12}&x_{11}&0&-x_{15}&x_{14}\\ x_5&0&0&0&x_6&0&-x_4&x_9&0&-x_7&x_{12}&0&-x_{10}&x_{15}&0&-x_{13}\\ x_6&0&0&0&-x_5&x_4&0&-x_8&x_7&0&-x_{11}&x_{10}&0&-x_{14}&x_{13}&0\\ \hline x_7&0&-x_{13}&x_{10}&0&-x_9&x_8&0&-x_6&x_5&-x_3&0&0&x_2&0&0\\ x_8&0&-x_{14}&x_{11}&x_9&0&-x_7&x_6&0&-x_4&0&-x_3&0&0&x_2&0\\ x_9&0&-x_{15}&x_{12}&-x_8&x_7&0&-x_5&x_4&0&0&0&-x_3&0&0&x_2\\ x_{10}&x_{13}&0&-x_7&0&-x_{12}&x_{11}&x_3&0&0&0&-x_6&x_5&-x_1&0&0\\ x_{11}&x_{14}&0&-x_8&x_{12}&0&-x_{10}&0&x_3&0&x_6&0&-x_4&0&-x_1&0\\ x_{12}&x_{15}&0&-x_9&-x_{11}&x_{10}&0&0&0&x_3&-x_5&x_4&0&0&0&-x_1\\ x_{13}&-x_{10}&x_7&0&0&-x_{15}&x_{14}&-x_2&0&0&x_1&0&0&0&-x_6&x_5\\ x_{14}&-x_{11}&x_8&0&x_{15}&0&-x_{13}&0&-x_2&0&0&x_1&0&x_6&0&-x_4\\ x_{15}&-x_{12}&x_9&0&-x_{14}&x_{13}&0&0&0&-x_2&0&0&x_1&-x_5&x_4&0 \end{array}\end{aligned}\ ] ] now the structure constants @xmath95 can be found from the above table ( see eq .
( [ eq:44 ] ) ) so that we can evaluate @xmath96 it is easy to verify that @xmath97 , and thus the killing form of @xmath0 is @xmath98 . since @xmath99 and @xmath100 , from the lie bracket computation table above , it is clear that @xmath101\subset \k,\quad [ \p ,
\k]\subset \p,\quad [ \p , \p]\subset \k.\ ] ] therefore the decomposition @xmath102 is a cartan decomposition of @xmath0 . note that the abelian subalgebra @xmath103 is contained in @xmath59 and is a maximal abelian subalgebra , _
i.e. _ , we can not find any other abelian subalgebra of @xmath59 that contains @xmath58 .
hence it is a cartan subalgebra of the pair @xmath60 .
further , since the set of all the local gates @xmath51 is a connected lie subgroup @xmath16 of @xmath8 , and there is one - to - one correspondence between connected lie subgroups of a lie group and subalgebras of its lie algebra @xcite , it is clear that @xmath52 in eq .
( [ eq:29 ] ) is just the lie subalgebra corresponding to @xmath51 . from proposition [ prop : cartan ]
, any @xmath104 can be decomposed as @xmath105 where @xmath106 , @xmath107 , and @xmath108 , @xmath109 , @xmath110 .
another more intuitive cartan decomposition of @xmath0 can be obtained via the complexification of @xmath9 .
consider @xmath111 , the real special linear group , and @xmath112 , the special orthogonal group .
the lie algebra @xmath9 is the set of @xmath113 real matrices of trace zero , and @xmath114 the set of @xmath113 real skew symmetric matrices .
then @xmath9 can be decomposed as @xmath115 , where @xmath59 is the set of @xmath113 real symmetric matrices .
this is nothing but the decomposition of a matrix into symmetric and skew symmetric parts , and it is indeed a cartan decomposition of @xmath9 . consider the following subset of the complexification of @xmath9 : @xmath116 it can be verified that @xmath117 is exactly @xmath0 , and thus @xmath118 is an orthogonal symmetric lie algebra pair .
the isomorphism carrying @xmath52 in eq .
( [ eq:29 ] ) into @xmath114 is just the transformation from the standard computational basis of states to the bell basis in @xcite .
this procedure is of crucial importance in computing the invariants for two - qubit gates under local transformations .
see section [ sec : non - local ] for more details .
now let us compute the weyl group @xmath81 .
let @xmath119 .
identify @xmath58 with @xmath11 , then @xmath120 $ ] .
the roots of @xmath25 with respect to @xmath58 are eigenvalues of the matrix of @xmath33 : @xmath121 for @xmath122 , the plane @xmath83 in @xmath58 is the set @xmath123 , where @xmath124^t$ ] .
the reflection of @xmath120 $ ] with respect to the plane @xmath83 is @xmath125.\ ] ] similarly , we can compute all the reflections @xmath84 as follows : @xmath126,\quad s_{i(c_2+c_3)}(x)=[c_1 , -c_3 , -c_2],\\ s_{i(c_2-c_1)}(x)=[c_2 , c_1 , c_3],\quad s_{i(c_1+c_2)}(x)=[-c_2,-c_1 , c_3],\\ s_{i(c_1-c_3)}(x)=[c_3 , c_2 , c_1],\quad s_{i(c_1+c_3)}(x)=[-c_3 , c_2 , -c_1 ] .
\endgathered\end{aligned}\ ] ] from proposition [ prop : weyl ] , the weyl group @xmath81 is generated by @xmath84 given in eq .
( [ eq:37 ] ) .
therefore , the reflections @xmath84 are equivalent to either permutations of the elements of @xmath127 $ ] , or permutations with sign flips of two elements .
we now study non - local two - qubit operations within the group theoretical framework of the previous section .
the cartan decomposition of @xmath0 provides us with a good starting point to explore the invariants under local gate operations .
it also reveals that the geometric structure of the local equivalence classes is none other than a 3-torus . every point on this 3-torus corresponds to a local equivalence class of two - qubit gates .
different points may also correspond to the same equivalence class . to reduce this symmetry ,
we apply the weyl group theory .
we show that the local equivalence classes of two - qubit gates are in one - to - one correspondence with the points in a tetrahedron , except on the base where there are two equivalent areas .
this tetrahedral representation of non - local operations plays a central role in our subsequent discussion of perfect entanglers and the design of universal quantum circuits .
two unitary transformations @xmath128 , @xmath13 are called _ locally equivalent _ if they differ only by local operations : @xmath129 , where @xmath106 , @xmath130 are local gates .
this clearly defines an equivalence relation on the lie group @xmath8 .
we denote the equivalence class of a unitary transformation @xmath128 as @xmath131 $ ] . from the cartan decomposition of @xmath0 in section [ sec : cartansu4 ] , any two - qubit gate @xmath132 can be written in the following form @xmath133 where @xmath106 , @xmath107 .
because the two - qubit gate @xmath128 is periodic in @xmath134 and the minimum positive period is @xmath10 , the geometric structure of @xmath127 $ ] is a 3-torus , @xmath135 . in @xcite ,
local invariants were given for two - qubit gates .
here we will connect these invariants of makhlin to the coordinates @xmath127 $ ] on the 3-torus .
we first consider the case of the two - qubit gates in @xmath8 , and then extend the results to the general case of @xmath2 .
consider the transformation from the standard basis of states @xmath136 , @xmath137 , @xmath138 , @xmath139 to the bell basis @xmath140 , @xmath141 , @xmath142 , @xmath143 . in this basis , the two - qubit gate @xmath128 in eq .
( [ eq:12 ] ) can be written as @xmath144 where @xmath145 recalling that @xmath146 is a basis for @xmath52 , it is not hard to verify that @xmath147 forms a basis for @xmath114 , the lie algebra of the special orthogonal group @xmath148 .
hence @xmath149 can be written as @xmath150 where @xmath151 eq . ( [ eq:57 ] ) can also be obtained from the cartan decomposition of @xmath0 derived from the complexification of @xmath9 , as discussed in section [ sec : cartansu4 ] .
an abelian subalgebra @xmath58 is generated by @xmath152 , and the transformation to the bell basis takes these operators to @xmath153 .
therefore , we have @xmath154 , where @xmath155
let @xmath156 where @xmath157 is defined by eq .
( [ eq:60 ] ) .
the complete set of local invariants of a two - qubit gate @xmath104 is given by the spectrum of the matrix @xmath158 @xcite , and hence by the eigenvalues of @xmath159 : @xmath160 since @xmath158 is unitary and @xmath161 , the characteristic polynomial of @xmath158 is then @xmath162 therefore the spectrum of @xmath158 is completely determined by only the two quantities @xmath163 and @xmath164 . for a two - qubit gate @xmath128 given in eq .
( [ eq:12 ] ) , its local invariants can be derived from eq .
( [ eq:62 ] ) as : @xmath165 now let us consider the local invariants for the general case of @xmath2 @xcite .
an arbitrary two - qubit gate @xmath166 can be decomposed as the product of a gate @xmath13 and a global phase shift @xmath14 , where @xmath167 .
it follows that @xmath168 , where @xmath169 and @xmath170 it is clear that the global phase factor just rotates the eigenvalues of @xmath171 along the unit circle in the complex plane , while keeping their relative phase invariant .
therefore , it does not affect the entangling properties and we can consequently divide by @xmath172 .
the local invariants of a two - qubit gate @xmath128 are thus given by @xmath173 where the numerical factors are incorporated into the denominators to provide convenient normalization .
if @xmath128 is now written in the following form : @xmath174 we can compute its local invariants as : @xmath175 because the local invariants @xmath176 and @xmath177 characterize the non - local properties of unitary operations , we can use these two invariants to check whether a pair of two - qubit gates are locally equivalent .
the invariants @xmath176 and @xmath177 are evaluated by taking the matrix representation of a gate in the bell basis and then using eqs .
( [ eq:88 ] ) and ( [ eq:32 ] ) .
for example , cnot and controlled - z ( referred as c(z ) ) possess identical values of the local invariants , given by @xmath178 and @xmath179 .
therefore , they belong to the same local equivalence class .
we refer to this class as [ cnot ] . on the other hand ,
the local invariants for @xmath7 are @xmath180 and @xmath181 .
hence this gate belongs to a different local equivalence class that we refer to as @xmath182 $ ] .
note that from eq .
( [ eq:32 ] ) , since the local invariants are functions of eigenvalues of the matrix @xmath158 , the local equivalence class can alternatively be defined simply via the set of eigenvalues of the matrix @xmath158 .
( [ eq:11 ] ) reveals the relation between the local invariants @xmath176 and @xmath177 and the coordinates @xmath127 $ ] of the 3-torus structure of non - local two - qubit gates . from this relation ,
given a set of coordinates @xmath127 $ ] , we can easily compute the local invariants for a local equivalence class .
vice versa , from a given pair of values of the local invariants @xmath176 and @xmath177 , we can also find the points on the 3-torus that correspond to a given two - qubit operation .
in general , we expect to find multiple points on the 3-torus for a given pair @xmath176 and @xmath177 .
we now show how this multiple - valued nature can be removed by using the weyl group to construct a geometric representation that allows the symmetry to be reduced . to visualize the geometric structure of the two - qubit gates , we first consider a cube with side length @xmath10 in the vector space @xmath58 .
this provides an equivalent representation of the points on the 3-torus , since @xmath183 .
clearly , every point in this cube corresponds a local equivalence class .
however , different points in the cube may belong to the same local equivalence class .
for example , both the points @xmath184 $ ] and @xmath185 $ ] correspond to the gate @xmath7 .
we use the theory of the weyl group to reduce this symmetry in the cube . from the lie group representation theory , the orbits of local gates @xmath51 acting on @xmath186 are in one - to - one correspondence with the orbits of the weyl group @xmath81 on @xmath58 @xcite . from proposition
[ prop : weyl ] , the weyl group @xmath81 is generated by the reflections @xmath84 as given in eq .
( [ eq:37 ] ) .
note that in eq .
( [ eq:37 ] ) , the reflections @xmath84 are either permutations or permutations with sign flips of two entries in @xmath127 $ ] .
therefore , if @xmath127 $ ] is an element in a local equivalence class @xmath131 $ ] , then @xmath187 $ ] , @xmath188 $ ] , @xmath189 $ ] , and @xmath190 $ ] are also in @xmath131 $ ] , where @xmath191 is a permutation of @xmath192 . with the meaning clear from the context of the discussion , in the remainder of this paper we shall use the triplet @xmath127 $ ] to denote either the corresponding local equivalence class of a two - qubit gate , or simply to refer to a specific point on the 3-torus or cube . [ cols="^,^,^ " , ]
recall that the flow generated by @xmath193 on the 3-torus is @xmath194 $ ] . by choosing some appropriate @xmath195 from the weyl group @xmath81 , we can steer the flow generated by the hamiltonian .
for example , if we want to change the flow from @xmath194 $ ] into @xmath196 $ ] , we can simply apply the reflection @xmath197 : @xmath198 the following example exemplifies this idea .
[ example : flow ] consider the isotropic hamiltonian @xmath199 .
our goal is to simulate [ cnot ] by a quantum circuit containing local gates and two - qubit gates generated by @xmath200 .
as shown in figure [ fig : flow ] , the flow generated by @xmath201 in the weyl chamber @xmath202 is @xmath203 $ ] , which evolves along @xmath204 for @xmath205 $ ] . in the weyl chamber @xmath202 , the point @xmath206 ( @xmath207 $ ] ) corresponds to [ cnot ] .
we then want to switch the flow from @xmath203 $ ] to @xmath208 $ ] at certain time instant so that the flow can reach the point @xmath206 . in order to do that
, we can simply apply the reflections @xmath209 and @xmath210 in series .
the corresponding local gate is thus @xmath211 and we have @xmath212 now consider the following quantum circuit ( 88,10 ) ( 0,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 ( 6,0)(19,10)[c]@xmath213 ( 25,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 ( 31,0)(10,10)[c]@xmath214 ( 41,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 ( 47,0)(19,10)[c]@xmath215 ( 66,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 ( 72,0)(10,10)[c]@xmath216 ( 82,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 the flow generated by this quantum circuit evolves along the line @xmath204 for @xmath217 $ ] , and then switches into a direction parallel to the line @xmath218 in the plane @xmath219 for @xmath220 . the matrix representation of this quantum circuit is @xmath221 hence the terminal point of the flow is @xmath222 $ ] .
if we choose @xmath223 and @xmath224 , the terminal point is none other than @xmath207 $ ] , and thus the quantum circuit simulates [ cnot ] . as shown in figure [ fig : flow ] , the flow generated by this quantum circuit is @xmath225 , which goes along the line @xmath226 first , and after hitting the point @xmath227 , it turns to the point @xmath206 along the line @xmath218 . since @xmath227 is nothing but @xmath7 , and @xmath228 , we arrive at the known result that [ cnot ] can be simulated by a circuit consisting of two @xmath7 and a local gate @xcite .
[ ] [ ] @xmath229 [ ] [ ] @xmath230 [ ] [ ] @xmath231 [ ] [ ] @xmath232 [ ] [ ] @xmath233 [ ] [ ] @xmath182 $ ] [ ] [ ] @xmath234}$ ] in the weyl chamber @xmath202.,title="fig : " ] we now derive the following theorem which asserts that when @xmath235 the quantum circuit ( [ eq:46 ] ) can simulate _ any _ arbitrary non - local two - qubit gate .
this theorem provides a geometric approach to construct a quantum circuit to simulate any arbitrary two - qubit gate from a two body interaction hamiltonian .
given a hamiltonian @xmath236 with @xmath237 , any arbitrary two - qubit gate @xmath104 can be simulated by the following quantum circuit ( 142,10 ) ( 0,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 ( 6,0)(10,10)[c]@xmath238 ( 16,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 ( 22,0)(18,10)[c]@xmath239 ( 40,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 ( 46,0)(10,10)[c]@xmath106 ( 56,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 ( 62,0)(18,10)[c]@xmath240 ( 80,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 ( 86,0)(10,10)[c]@xmath241 ( 96,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 ( 102,0)(18,10)[c]@xmath242 ( 120,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 ( 126,0)(10,10)[c]@xmath243 ( 136,0 ) ( 6,10 ) ( 0,2)(1,0)6 ( 0,8)(1,0)6 where @xmath244 are local gates . from cartan decomposition of @xmath0 in section [ sec : cartansu4 ] , any arbitrary two - qubit gate @xmath104 can be written in the following form : @xmath245 where @xmath246 and @xmath247 are local gates , and @xmath248 , @xmath249 , @xmath250 .
we also know that for any given @xmath251 , there exists a local gate @xmath41 such that @xmath252 , where @xmath253 , and @xmath254 .
therefore , the flow generated by @xmath255 on the 3-torus is @xmath256 $ ] .
the matrix representation of the above quantum circuit is @xmath257 let @xmath258 the quantum circuit ( [ eq:85 ] ) can be written as @xmath259 choose some appropriate local gates @xmath238 , @xmath106 , and @xmath241 such that @xmath260 and let @xmath261 .
it follows that the quantum circuit ( [ eq:85 ] ) is now @xmath262 to simulate the two - qubit gate @xmath128 in eq .
( [ eq:73 ] ) , we only need to solve the following equation : @xmath263 since @xmath264 we can always find a solution for eq .
( [ eq:55 ] ) .
therefore , the quantum circuit ( [ eq:85 ] ) can simulate any arbitrary two - qubit gate .
from the above constructive proof , it is clear that together with four appropriate local gates , we can simulate any arbitrary two - qubit gate by turning on a two - body interaction hamiltonian for at most three times .
also note that in the proof , the way to choose the local gates @xmath238 , @xmath106 , and @xmath241 is not unique .
there are many different ways to choose the local gates and time parameters so as to construct the quantum circuit that achieves the same two - qubit operation .
we therefore can pick the one that is optimal in terms of some cost index such as time .
in this paper we have derived a geometric approach to study the properties of non - local two - qubit operations , starting from the cartan decomposition of @xmath0 and making use of the weyl group .
we first showed that the geometric structure of non - local gates is a 3-torus . by further reducing the symmetry , the geometric representation of non - local gates
was seen to be conveniently visualized as a tetrahedron .
each point inside this tetrahedron corresponds to a different equivalent class of non - local gates .
we then investigated the properties of those two - qubit operations that can generate maximal entanglement .
we provided a proof of the condition of makhlin for perfect entanglers @xcite and then derived the corresponding geometric description of these gates within the tetrahedral representation .
it was found that exactly half of the non - local two - qubit operations result in maximal entanglement , corresponding to a seven - faced polyhedron with volume equal to one half of the tetrahedron .
lastly , we investigated the non - local operations that can be generated by a given hamiltonian .
we proved that given a two - body interaction hamiltonian , it is always possible to explicitly construct a quantum circuit for exact simulation of any arbitrary non - local two - qubit gate by turning on the two - body interaction for at most three times , together with four local gates .
this guarantees that a highly _ efficient _ simulation of non - local gates can be made with any hamiltonian consisting of arbitrary two - qubit interactions and allowing control of single qubit operations .
we thank the nsf for financial support under itr grant no .
eia-0204641 ( ss and kbw ) .
jv and kbw s effort is also sponsored by the defense advanced research projects agency ( darpa ) and air force laboratory , air force materiel command , usaf , under contract number f30602 - 01 - 2 - 0524 , and the office of naval research under grant no .
fdn 00014 - 01 - 1 - 0826 . | we study non - local two - qubit operations from a geometric perspective . by applying a cartan decomposition to @xmath0 , we find that the geometric structure of non - local gates is a 3-torus .
we derive the invariants for local transformations , and connect these local invariants to the coordinates of the 3-torus . since different points on the 3-torus may correspond to the same local equivalence class , we use the weyl group theory to reduce the symmetry .
we show that the local equivalence classes of two - qubit gates are in one - to - one correspondence with the points in a tetrahedron except on the base .
we then study the properties of perfect entanglers , that is , the two - qubit operations that can generate maximally entangled states from some initially separable states .
we provide criteria to determine whether a given two - qubit gate is a perfect entangler and establish a geometric description of perfect entanglers by making use of the tetrahedral representation of non - local gates .
we find that exactly half the non - local gates are perfect entanglers .
we also investigate the non - local operations generated by a given hamiltonian .
we first study the gates that can be directly generated by a hamiltonian
. then we explicitly construct a quantum circuit that contains at most three non - local gates generated by a two - body interaction hamiltonian , together with at most four local gates generated by single qubit terms .
we prove that such a quantum circuit can simulate any arbitrary two - qubit gate exactly , and hence it provides an efficient implementation of universal quantum computation and simulation . |
chern - simons ( cs ) theories are relevant in a quantum field theory context since they provide an alternative gauge - invariant procedure of mass generation @xcite . moreover , the high - temperature limit of quantum field theories in @xmath6 dimensions are effectively three dimensional and cs terms are precisely induced by fermions in @xmath7 dimensions through the parity anomaly @xcite .
cs actions play also a role in the analysis of interesting condensed matter phenomena @xcite-@xcite , the computation of topological invariants of 3-manifolds @xcite and they are connected with @xmath8 conformal field theories @xcite .
cs - higgs models differ drastically from theories in which solely a maxwell or yang - mills term governs the dynamics of the gauge fields .
in particular , at the classical level , axially symmetric ( vortex or string ) solutions to the equations of motion necessarily carry electric charge @xcite-@xcite which , in the non - abelian case , is quantized ( for a complete review on cs theories and planar physics see @xcite ) .
bogomolny equations for the non - abelian cs - higgs system were first obtained in @xcite , where it was shown that , as in the abelian case @xcite-@xcite , a sixth order potential has to be considered .
explicit vortex solutions were exhibited in @xcite , with the flux directed in the cartan subalgebra of the non - abelian group .
it is the purpose of the present work to find genuine non - abelian vortex configurations by proposing a more general ansatz as the one already considered for yang - mills - matter theories @xcite-@xcite with gauge group @xmath1 and @xmath9 flavors of fundamental matter multiplets .
the resulting vortex configurations can be characterized by non - abelian collective coordinates related to orientational degrees of freedom and , when @xmath4 , to infinitesimal variations of the transverse size .
we consider the truncated bosonic sector of the @xmath0 susy @xmath10 chern - simons - higgs action in @xmath11 dimensions ( the complete susy lagrangian can be found in @xcite ) @xmath12 \right\ } , \label{uno}\end{aligned}\ ] ] where @xmath13 , @xmath14 , @xmath15 are the structure constants of the non - abelian group , and the quantization condition implies @xmath16 .
our theory has a sixth - order potential which , forced by @xmath17 supersymmetry , allows bogomolny completion of the energy functional with all coupling constants at the bogomolny point ( see ref.@xcite and references therein for details on this point ) @xmath18 & = \frac{1}{16 \kappa_1 ^ 2 n^2 } \phi^\dag_f \phi^f \left(\phi^\dag_g\phi^g - n\xi\right)^2 + \frac{1}{4\kappa_2 ^ 2}\phi^\dag_f\tau^i\tau^j\phi^f \left(\phi^\dag_g\tau^i\phi^g\right ) \left(\phi^\dag_h\tau^j\phi^h\right)\nonumber\\ & - \frac{1}{4\kappa_1\kappa_2 n}\left(\phi^\dag_f\tau^i\phi^f\right)^2\left(\phi^\dag_g\phi^g - n\xi\right).\end{aligned}\ ] ] here , @xmath19 are lorentz indices , @xmath20 are the @xmath21 `` color '' group indices and @xmath22 are the anti - hermitian generators of @xmath21 .
the complex scalar multiplets @xmath23 , besides the color index @xmath24 , possess additional flavor index @xmath25 with @xmath3 , thus can be written as @xmath26 matrices .
the covariant derivatives and field strengths are defined as @xmath27}^0\ , & & \hspace{-4 cm } f_{\mu\nu}^i = \partial_{[\mu } a_{\nu ] } ^i + f^{ijk } a_{\mu}^j a_{\nu}^k \ ; .\end{aligned}\ ] ] up to gauge transformations , minima of the potential are given by @xmath28 in what follows we set , without loss of generality , @xmath29 .
the energy density is @xmath30\ ] ] and the euler - lagrange equations of motion of the theory are @xmath31 defining @xmath32 with @xmath33 , and using gauss law , we can write the energy as a sum of squares @xmath34^\dag \right . \nonumber\\ & \times \left [ d_0 \phi^f - i\epsilon \left ( \frac{1}{4\kappa_1 n}\left[\phi^\dag_g\phi^g - n \right ] \phi^f -\frac{1}{2\kappa_2 } \phi^\dag_g\tau^i\phi^g\tau^i\phi^f \right)\right ] \nonumber \\ & + \left . \vphantom{\frac{1}{4\kappa_1 } } \left(d_{-\epsilon}\phi^f\right)^\dag \left(d_{-\epsilon } \phi^f\right ) + \epsilon\sqrt{2n } f^0_{12 } \right\}\end{aligned}\ ] ] leading to the bogomolny equations .
in this section we investigate the so - called `` local @xmath35 string - type solutions '' as discussed for the yang - mills case in @xcite-@xcite ( see also @xcite ) .
we set @xmath5 so the matter fields can be arranged as a square matrix @xmath36 .
the lagrangian ( [ uno ] ) is then invariant under @xmath37 rotations , @xmath38 with @xmath39 and @xmath40 .
we start from the trivial vacuum in the asymmetric phase @xmath41 after a @xmath1 gauge transformation we obtain a singular vortex configuration which , for the scalar field , takes the form @xmath42 with @xmath43 concerning the gauge fields we have @xmath44 in such configuration , the @xmath35 center of the gauge group @xmath21 has been combined with @xmath45 elements to get a topologically stable string solution possessing both windings , in @xmath21 and in @xmath45 ( since @xmath46 , the topology is nontrivial ) .
these kind of topological objects are also called @xmath35 strings " . in the present case the configurations ( [ 15]),([20 ] ) represent a @xmath47 singular string .
a general @xmath48 vortex configuration can be obtained following the same method .
this suggests the following ansatz for the regular vortex configuration @xmath49 it should be noted that , contrary to the yang - mills - higgs case @xcite , the @xmath50 fields are nontrivial .
the boundary conditions for the fields are @xmath51 where we required the solution to be singled - valued at the origin , and to be a pure gauge at infinity . with this ansatz both the flux and the energy of these solutions
are quantized @xmath52 note that with our conventions @xmath53 is always a positive integer , and @xmath54 determines whether the flux is positive ( @xmath55 ) , or negative ( @xmath56 ) .
ansatz ( [ fi ] ) corresponds , for the @xmath57 case , to what is called an `` elementary string '' .
composite strings can be constructed by introducing windings in several diagonal elements in the scalar field and can be seen as the superposition of elementary ones .
substituting ansatz ( [ fi])-([superansatz ] ) into the bogomolny and gauss equations , we obtain the following system of non - linear first - order differential equations @xmath58 \label{non3}\\ & \frac{1}{r}\partial_r f^{n^2 - 1 } = \frac{1}{4 n \kappa_2 } \left[f_0 \left(\phi^2 - \phi^2_n\right ) + f_0^{n^2 - 1 } \left(\phi^2 + ( n - 1 ) \phi^2_n\right)\right]\label{non4}\\ & f_0 = \frac{\epsilon}{2\kappa_1}\left((n-1)\phi^2 + \phi_n^2 - n\right ) \ ; , \;\;\;\;\ ;
f_0^{n^2 - 1 } = \frac{\epsilon}{2\kappa_2 } \left(\phi^2 - \phi_n^2\right ) .
\label{non6}\end{aligned}\ ] ] note that equations ( [ non6 ] ) can be used to eliminate @xmath59 and @xmath60 in ( [ non3 ] ) and ( [ non4 ] ) . equations ( [ non1])-([non4 ] ) then correspond to the bogomolny equations written in the standard form . in terms of the profile functions ,
the energy takes the form @xmath61 the magnetic field ( which is a pseudoscalar in 2 + 1 dimensions ) and the electric field ( with component only in the radial direction ) take the form @xmath62 let us study the solutions of equations ( [ non1])-([non6 ] ) .
we can distinguish two cases : first , when the @xmath45 and @xmath21 coupling constants are equal and then the effective symmetry group is @xmath63 , and second , when the @xmath45 and @xmath21 coupling constants are different .
if the @xmath45 and @xmath21 coupling constants are equal @xmath65 , the system of equations ( [ non1])-([non6 ] ) decouples into @xmath66 where we have defined @xmath67 in terms of the new functions , the boundary conditions are @xmath68 since the field @xmath69 has no winding , it does not necessarily vanish at the origin so @xmath70 is an arbitrary parameter .
systems ( [ trivial ] ) and ( [ nontrivial ] ) are formally the same but the functions @xmath71 and @xmath72 obey different boundary conditions . remarkably , both systems coincide with those arising in the abelian case discussed in @xcite . for the pair @xmath71 ,
boundary conditions imply that @xmath73 the system ( [ nontrivial ] ) for the functions @xmath72 was solved numerically in @xcite .
we remark that these equations do not depend on @xmath74 . at the origin , @xmath75 approaches to zero as @xmath76 ( this fact will be important for the semi - local vortex ) .
( solid line ) , electric field @xmath77 ( dashed line ) , and energy density @xmath78 ( dotted line ) for local vortices with @xmath79 . the distance between the energy density maximum and the origin increases with the winding number @xmath53.,width=604 ] we show in figure [ bkk1 ] profiles of @xmath80 , @xmath77 and the energy density @xmath81 .
as usually happens in cs theories , the sixth order potential makes the maximum of the magnetic field to be away from the origin . when @xmath83 , the complete set of equations ( [ non1])-([non6 ] ) has to be solved numerically .
we used a relaxation method to find explicit numerical solutions .
the ratio of coupling constants @xmath84 is in fact the only independent parameter of the theory , since @xmath85 or @xmath86 can be absorbed by a rescaling .
in fact the energy can be expressed in terms of @xmath87 only .
we observe that as @xmath87 departs from @xmath88 , @xmath89 and @xmath90 tend to separate from each other as well , forcing @xmath69 to be non - constant . as @xmath87 goes from @xmath91 to @xmath92 , the difference @xmath93 changes sign , forcing the derivative of @xmath69 to change sign .
qualitatively , the behavior resembles the @xmath63 case ( in which both coupling constant coincide ) .
when varying the winding number @xmath53 of the vortex , the profile functions change in a similar way as they do in the @xmath63 case .
for equal @xmath53 and different @xmath74 , solutions do not change considerably .
we present in figure [ kk05 ] some solutions for the case @xmath94 .
we have discussed non - abelian vortex solutions in a @xmath95 theory which are usually called _ local _ vortex solutions . in order to have _ semilocal _ vortex solutions , those for which @xmath96 , one has to extend the matter content of the theory by adding @xmath97 extra flavors so that @xmath98 @xcite-@xcite .
for definiteness we take @xmath97 equal to @xmath74 ( but a general case can be equally treated ) .
then , the symmetry of the model is @xmath99 .
we call @xmath100 the extra matter fields . in this case
the trivial vacuum state is given by @xmath101 and it is invariant under the following transformation @xmath102 where we have chosen , as in the local case , a gauge element @xmath103 with @xmath104 a @xmath105 flavor block and we have called @xmath106 the @xmath105 flavor block acting on the extra @xmath100 mater fields . we follow now the same steps as in the local case : first we perform a rotation of the vacuum to find a singular vortex configuration and then propose an ansatz for a regular configuration which we write explicitly @xmath107 where @xmath108 is the @xmath105 identity matrix .
it should be stressed that more general ansatz could lead to solutions which exhaust the number @xmath109 of collective coordinates related to the transverse moduli space @xcite .
our ansatz corresponds to just two collective coordinates . substituting ( [ ansatzchi ] ) into the bogomolny equations , we arrive to a system of six differential equations for the fields @xmath69 , @xmath75 , @xmath110 , @xmath111 , @xmath89 and @xmath112 , and two constraints for the fields @xmath59 and @xmath60 .
the solutions to these equations will have the same energy and flux than the local ones ( [ flux - energy ] ) , provided they satisfy the same boundary conditions ( [ boundcond ] ) . at infinity
the solutions should reach the vacuum state , so we impose @xmath113 @xmath114 again we distinguish the case in which the @xmath45 and @xmath21 gauge coupling constants are equal from the one in which they are different . as it happens for local vortex , when @xmath116 the equations of motion decouple into two sets of independent equations .
one corresponds to a system with no winding @xmath117 and the other corresponds to a system with winding @xmath53 @xmath118 ( we have defined @xmath119 and @xmath120 as in ( [ nontrivial2 ] ) ) . as in the case of local strings ,
these equations coincide with those that arise in the abelian case @xcite .
the system ( [ semitrivial ] ) admits a trivial solution @xmath121 while ( [ seminontrivial ] ) can be solved numerically . combining the equations for @xmath111 and @xmath75
, we get @xmath122 where @xmath123 is a ( complex ) integration constant .
so , we finally have @xmath124 the same argument applied to @xmath110 and @xmath69 determines that @xmath125 , but since @xmath126 and @xmath127 , @xmath128 must vanish , and so @xmath129 everywhere .
the solutions we obtain are then the most general ones .
the energy is independent of the complex parameter @xmath123 , which is then associated with the `` size moduli '' .
one has has in general , @xmath130 complex parameters @xmath131 for arbitrary flavor @xmath96 . for elementary strings one can see that @xmath132 ( since we are considering @xmath133 , we have in this case just one complex parameter ) . from figure [ semikk1 - 2 ]
we see that the solutions spread when the parameter @xmath134 is incremented . as @xmath123 increases , since the flux is conserved , the extremum of the magnetic field must approach to the origin to compensate the spread towards spatial infinity .
a similar phenomenon occurs with the energy .
we consider now the case in which @xmath136 .
the same arguments as above lead to @xmath137 , so that the flavor group is in fact reduced from @xmath138 to @xmath139 .
similarly , the relation holds , and the differential equations do not decouple .
the system is very analogous to the system ( [ non1])-([non6 ] ) obtained for the local case , with the only difference that , in the equations for the @xmath89 s , the field @xmath75 gets locally scaled @xmath140 then , semi - local solutions arising from these equations are similar to those shown in figure [ kk05 ] for the local case , with the only difference that they are smoother since they decay as powers of @xmath141 at spatial infinity . in summary , our main task in this work was to solve the bps equations for a non - abelian chern - simons - higgs theory . by proposing an axially symmetric ansatz we obtained non - abelian vortex solutions and discussed their properties . a class of vortex solutions in non - abelian cs theories were already known @xcite-@xcite,@xcite , @xcite .
the solutions discussed here correspond to more fundamental vortices in the sense that they are genuinely non - abelian while the former correspond to @xmath35 vortices with the gauge flux in the cartan algebra of @xmath21 ) .
the model discussed here is indeed related to the one analyzed by @xcite-@xcite except that in our case the dynamics of the gauge fields is governed by a cs action instead of a yang - mills one .
this drastically changes the vortex properties , in particular forcing them to carry electric charge . in the case of local vortices ,
our solutions generalize those discussed in @xcite to the case in which the gauge group is @xmath1 , with the abelian and non - abelian sectors having different gauge coupling constants .
when both couplings are equal , the equations decouple into two sets of equations that coincide with those arising in the abelian case .
this is not the case when the couplings are different and the bps equations do not decouple .
nevertheless we were able to construct explicit solutions and discuss their properties .
further , we have also considered semi - local vortices , by allowing the flavor number @xmath2 to be larger than the color number @xmath142 . as already noted for the yang - mills - higgs system ,
the main feature in this case is that the solutions develop an additional moduli @xmath123 related to the vortex transverse size , thus modifying the asymptotic behavior of the fields , from exponential decay for the case of local vortices , to a decay as negative power of the radial coordinate for the semi - local ones .
interestingly enough , one can see from our explicit solutions how the size of the vortex grows with @xmath123 .
we would like to thank the sociedad cientifica argentina for hospitality .
we are grateful to acknowledge len aldrovandi for interesting discussions and comments .
this work is partially supported by conicet ( pip6160 ) , anpcyt ( pict 20204 ) , unlp , uba and cicba grants .
99 s. deser , r. jackiw and s. templeton , phys .
( 1982 ) 975 ; ann .
* 140 * ( 1982 ) 372 ; ann .
phys . * 185 * ( 1988 ) 406 , erratum ; ann .
* 281 * ( 2000 ) 409 .
a. n. redlich , phys .
* 52 * ( 1984 ) 18 ; phys .
d * 29 * ( 1984 ) 2366 .
a. polyakov , mod .
* a3 * ( 1988 ) 325 .
see f. wilczek , " fractional statistics and anyon superconductivity , _ singapore , singapore : world scientific ( 1990 ) _ and references therein .
see o. heinonen , `` composite fermions : a unified view of the quantum hall regime , '' _ singapore , singapore : world scientific ( 1998 ) _ and references therein .
e. witten , commun .
121 * ( 1989 ) 351 ; nucl.phys .
* b311 * ( 1989 ) 46 .
s. elitzur , g. w. moore , a. schwimmer and n. seiberg , nucl .
b * 326 * ( 1989 ) 108 .
s. k. paul and a. khare , phys . lett .
b * 174 * ( 1986 ) 420 [ erratum - ibid . * 177b * ( 1986 ) 453 ] .
h. j. de vega and f. a. schaposnik , phys .
* 56 * ( 1986 ) 2564 ; h. j. de vega and f. a. schaposnik , phys .
d * 34 * ( 1986 ) 3206 .
g. lozano , m. v. manias and f. a. schaposnik , phys .
d * 38 * , 601 ( 1988 ) .
j. hong , y. kim and p. y. pac , phys .
rev . lett .
* 64 * , 2230 ( 1990 ) .
r. jackiw and e. j. weinberg , phys .
* 64 * , 2234 ( 1990 ) .
r. jackiw , k. m. lee and e. j. weinberg , phys .
d * 42 * , 3488 ( 1990 ) .
l. f. cugliandolo , g. lozano , m. v. manias and f. a. schaposnik , mod .
lett . a * 6 * , 479 ( 1991 ) .
g. v. dunne , arxiv : hep - th/9902115 .
r. auzzi , s. bolognesi , j. evslin , k. konishi and a. yung , nucl .
b * 673 * ( 2003 ) 187 .
a. hanany and d. tong , jhep * 0307 * , 037 ( 2003 ) ; jhep * 0404 * , 066 ( 2004 ) .
a. gorsky , m. shifman and a. yung , phys .
d*71 * ( 2005 ) 045010 l. g. aldrovandi and f. a. schaposnik , arxiv : hep - th/0702209 .
s. j. j. gates and h. nishino , phys .
b * 281 * ( 1992 ) 72 ; h. nishino and s. j. j. gates , int .
j. mod .
* a8 * , 3371 ( 1993 ) .
f. a. schaposnik , arxiv : hep - th/0611028 .
a. achucarro and t. vachaspati , phys .
* 327 * ( 2000 ) 347 .
m. shifman and a. yung , phys .
d * 73 * ( 2006 ) 125012 .
a. khare , phys .
d * 46 * ( 1992 ) r2287 . | we consider the bosonic sector of a @xmath0 supersymmetric chern - simons - higgs theory in 2 + 1 dimensions .
the gauge group is @xmath1 and has @xmath2 flavors of fundamental matter fields .
the model supports non - abelian ( axially symmetric ) vortices when @xmath3 , which have internal ( orientational ) moduli . when @xmath4 , the solutions acquire additional collective coordinates parameterizing their transverse size .
we solve the bps equations numerically and obtain local ( @xmath5 ) and semi - local ( @xmath4 ) string solutions . |
the kontsevich integral of a knot in @xmath2 is a graph - valued invariant , which is characterized by a universal property , namely it is a universal @xmath3-valued vassiliev invariant of knots .
using the language of physics , the kontsevich integral is the feynmann diagram expansion ( i.e. , perturbative expansion ) of the chern - simons path integral , expanded around a trivial flat connection , @xcite .
this explains the shape of the graphs ( namely , they have univalent and trivalent vertices only ) , their vertex - orientation and their vassiliev degree , namely half the number of vertices
. in physics one often takes the logarithm of a feynmann diagram expansion , which is given by a series of connected graphs , and one studies the the terms with a fixed number of loops ( i.e. , with a fixed first betti number ) .
the purpose of this paper is to study the kontsevich integral , graded by the euler degree ( rather than the vassiliev degree , or the number of loops ) , to introduce a null - move on knots , and to relate the euler degree to the null - move .
this point of view explains in a natural way a rationality conjecture of the kontsevich integral ( formulated by the second author and proven by @xcite ) , the relation of the null - move to @xmath0-equivalence , the relation of the null - move to cyclic branched covers @xcite , and offers an opportunity to use vassiliev invariants as obstructions to the existence of seifert forms , @xcite .
we begin by explaining the euler expansion of the kontsevich integral @xmath4 , defined for links in @xmath2 by kontsevich , @xcite , and extended to an invariant of links in arbitrary closed 3-manifolds by the work of le - murakami - ohtsuki , @xcite .
our paper will focus on the above invariant @xmath5 of knots @xmath6 in s @xmath7 : @xmath8 where @xmath9 is the completed vector space over @xmath3 generated by unitrivalent graphs with vertex - orientations modulo the well - known antisymmetry @xmath10 and @xmath11 relations .
the graphs in question have a vassiliev degree ( given by half the number of vertices ) and the completion of @xmath9 refers to the vassiliev degree .
we now introduce a second grading on unitrivalent graphs : the _ euler degree _
@xmath12 of a unitrivalent graph @xmath13 is the number of trivalent vertices that remain when we shave - off all legs of @xmath13 .
it is easy to see that @xmath14 where @xmath15 is the euler characteristic of @xmath13 , which explains the naming of this degree .
the @xmath10 and @xmath11 relations are homogeneous with respect to the euler degree , thus we can let @xmath16 denote the euler degree @xmath1 part of the kontsevich integral .
notice that there are finitely many unitrivalent graphs with vassiliev degree @xmath1 , but infinitely many connected graphs with euler degree @xmath1 .
for example , wheels with @xmath1 legs have vassiliev degree @xmath1 but euler degree @xmath17 .
the kontsevich integral of a knot , graded by the vassiliev degree is a universal vassiliev . on the other hand ,
@xmath16 are not vassiliev invariants ; they are rather power series of vassiliev invariants . why consider the euler degree ?
a deep and unexpected geometric reason is the content of the next section .
a beautiful theory of goussarov and habiro is the study of the geometric notion of _ surgery on a clasper _ , see @xcite and @xcite ( in the latter claspers were called clovers ) . following the notation of @xcite , given a clasper @xmath13 in a 3-manifold @xmath19 ,
we let @xmath20 denote the result of surgery .
clasper surgery can be described in terms of twisting genus @xmath21 handlebodies in @xmath19 , or alternatively in terms of surgery on a framed link in @xmath19 .
we will refer the reader to @xcite for the definition and conventions of surgery on a clasper .
we will _ exclude _ claspers with no trivalent vertices , that is , ones that generate @xmath22-moves . in the present paper we are interested in pairs @xmath23 and claspers @xmath24
whose leaves are _ null homologous links _ in @xmath25 .
we will call such claspers @xmath18-_null _ , or simply _ null_.
the terminology is motivated by the fact that the leaves of such claspers are sent to @xmath17 under the map : @xmath26 . in order to motivate our interest in null claspers ,
recall the basic and fundamental principle : surgery on a clasper preserves the homology and ( when defined ) the linking form . applying that principle in the @xmath18-cover of a knot complement
, it follows that surgery on a null clasper preserves the blanchfield form , as we will see below .
surgery on null claspers describes a move on the set of knots in s. given this move , one can define in the usual fashion a notion of _ s _ and a dual notion of @xmath1-_equivalence _ , as was explained by goussarov and habiro .
explicitly , we can consider a decreasing filtration @xmath27 on the vector space generated by all pairs @xmath23 as follows : @xmath28 is generated by @xmath29 $ ] , where @xmath30 is a disjoint collection of null claspers in @xmath25 , and where @xmath31=\sum_{i \subset \{0,1\}^n } ( -1)^{|i| } ( m , k)_{g_i}\ ] ] ( where @xmath32 denotes the number of elements of @xmath22 and @xmath33 stands for the result of simultaneous surgery on claspers @xmath34 for all @xmath35 ) . a function @xmath36 is a _ of null - type _
@xmath1 iff @xmath37 .
we will discuss the general notion of @xmath1-null - equivalence in a later publication , at present we will consider the special case of @xmath38 .
two pairs @xmath23 and @xmath39 are _ null - equivalent _ iff one can be obtained from the other by a sequence of null moves . in order to motivate the next lemma ,
recall a result of matveev @xcite , who showed that two closed 3-manifolds @xmath7 and @xmath40 are equivalent under a sequence of clasper surgeries iff they have the same homology and linking form .
this answer is particularly pleasing since it is expressed in terms of abelian algebraic invariants . the following are equivalent : @xmath23 and @xmath39 * are null equivalent * are @xmath0-equivalent * have isometric blanchfield form . for a definition of @xmath0-_equivalence _ and _
blanchfield forms _ , see for example @xcite .
it is well - known that @xmath23 and @xmath39 are @xmath0-equivalent iff they have isometric blanchfield forms ; for an algebraic - topological proof of that combine levine and kearton , @xcite , or alternatively trotter , @xcite .
thus ( b ) is equivalent to ( c ) .
if @xmath39 is obtained from @xmath23 by a null move on a clasper @xmath13 , then @xmath13 lifts to the universal abelian cover @xmath41 of the knot complement @xmath42 .
the lift @xmath43 of @xmath13 is a union of claspers translated by by the group @xmath18 of deck transformations .
furthermore , @xmath44 is obtained from @xmath41 by surgery on @xmath43 .
since clasper surgery preserves the homology and linking form , it follows that @xmath39 and @xmath23 have isometric blanchfield forms , thus ( a ) implies ( c ) .
suppose now that @xmath39 is @xmath0-equivalent to @xmath23 .
then , by matveev s result , the @xmath7 can be obtained from @xmath2 by surgery on some null @xmath13 . moreover ,
if @xmath6 is a knot in @xmath7 that bounds a seifert surface @xmath45 , we can always arrange ( by an isotopy on @xmath13 ) that @xmath13 is disjoint from @xmath45 , thus @xmath13 is @xmath23-null .
thus , if @xmath23 and @xmath39 are @xmath0-equivalent pairs , modulo null moves we can assume that @xmath46 . in that case , naik and stanford show that @xmath0-equivalent pairs @xmath47 and @xmath48 are equivalent under a sequence of double @xmath49elta - moves , that is null - moves where all the leaves of the clasper bound disjoint disks , @xcite . thus , ( b ) implies ( a ) and the lemma follows .
it follows that if @xmath23 and @xmath39 are null equivalent , then they have the same alexander module , in particular the same alexander polynomial , and the same algebraic concordance invariants which were classified by levine @xcite .
there is a well - known similarity between @xmath50-valued linking forms of rational homology spheres and @xmath51$]-valued blanchfield forms . from this point of view , correspond to knots with trivial alexander polynomial .
this classical analogy extends to the world of s as follows : one can consider s of s ( using surgery on claspers ) and s of knots in s ( using surgery on null - claspers ) . in our paper
, we will translate well - known results about s and @xmath1-equivalence of s to the setting of knots with trivial alexander polynomial . are there any nontrivial s of null - type ?
the next result answers this , and reveals an unexpected relation between the null - move and the euler degree : for all @xmath1 , @xmath52 is a null - type @xmath1 invariant of pairs @xmath23 with values in @xmath53 . in the above
, @xmath54 stands for the subspace of @xmath9 generated by diagrams of euler - degree @xmath1 .
the above theorem has implications for the _ loop expansion _ of the kontsevich integral defined as follows .
the logarithm of the kontsevich integral @xmath55 takes values in the completed vector space ( with respect to the vassiliev degree ) @xmath56 generated by connected vertex - oriented unitrivalent graphs , modulo the @xmath10 and @xmath11 relations .
let us define the _ loop degree _ of a graph to be the number of loops , i.e. , the first betti number .
the @xmath10 and @xmath11 relations are homogeneous with respect to the loop degree , thus we can define @xmath57 to be the @xmath58-loop degree of @xmath59 . since a connected trivalent graph with @xmath60 loops has euler degree @xmath61 , and since @xmath59 is an additive invariant of pairs under connected sum , it follows that for all @xmath1 , @xmath57 is an additive null - type @xmath61 invariant of pairs @xmath23 with values in the loop - degree @xmath58 part of @xmath56 .
our next task is to study the graded quotients @xmath62 ; knowing this would essentially answer the question of how many null - s there are .
since null equivalence has more than one equivalence classes , we ought to study @xmath27 and @xmath63 on each null equivalence class .
this is formalized in the following way .
given a pair @xmath23 , let @xmath64 denote the subspace of @xmath27 generated by all pairs @xmath39 which are null equivalent to @xmath23 .
it is easy to see that @xmath65 when @xmath23 and @xmath39 are null equivalent and that we have a direct sum decomposition @xmath66 fixing @xmath23 , we can define @xmath67 accordingly .
the simplest case to study are the graded quotients @xmath68 for the unknot @xmath69 in @xmath2 .
this is equivalent to the study of null- of knots with trivial alexander polynomial .
it turns out in a perhaps unexpected way that @xmath68 can be described in terms of trivalent graphs with _
beads_. let us define these here .
let @xmath70 $ ] denote the group - ring of the integers .
it is a ring with involution @xmath71 given by @xmath72 and augmentation map @xmath73 given by @xmath74 .
consider a trivalent graph @xmath13 with oriented edges .
@xmath75-coloring _ of @xmath13 is an assignment of an element of @xmath75 to every edge of @xmath13 .
we will call the assignment of an edge @xmath76 , the _ bead _ of @xmath76 .
@xmath77 is the completed vector space over @xmath3 generated by pairs @xmath78 ( where @xmath13 is a trivalent graph , with oriented edges and vertex - orientation and @xmath79 is a @xmath75-coloring of @xmath13 ) , modulo the relations : @xmath10 , @xmath11 , linearity , orientation reversal , holonomy and graph automorphisms .
@xmath80 @xmath77 is graded by the euler degree of a graph , and the completion is with respect to the above grading .
we could have defined @xmath81 and @xmath77 over @xmath18 rather than over @xmath3 .
we will not use special notation to indicate this ; instead when we wish , we will state explicitly which coefficients we are using . over @xmath18 , there is degree - preserving map : @xmath82 which is onto , over @xmath83 $ ] . since @xmath84 , it follows that over @xmath83 $ ] , @xmath85 is a @xmath86-step filtration , i.e. , satisfies @xmath87 for all @xmath1 . in view of theorems [ thm.loop2 ] and [ thm.loop1 ] ,
it is natural to ask if there is any relation among the algebras @xmath9 and @xmath77 . in order to make contact with later work ,
we mention here an important _ hair _ map @xmath88 which is defined by replacing a bead @xmath89 by an exponential of hair : @xmath90 the kontsevich integral ( graded by the euler degree , and evaluated on alexander polynomial @xmath91 knots ) fails to be a universal ( with respect to the null - move ) because it takes values in the `` wrong '' space , namely @xmath9 rather than @xmath77 .
it is natural to conjecture the existence of an invariant @xmath92 that fits in a commutative diagram @xmath93{\a(\la ) } \arrow{s , r}{\hair } \\ \node{(\delta=1)-\mathrm{knots } } \arrow{e , t}{z}\arrow{ne , t}{\zrat } \node{\a(\ast ) } \end{diagram}\ ] ] this rationality conjecture was formulated by the second author ( in a version for all knots ) and was proven in @xcite . in an earlier version of the paper
, we formulated a conjectural relation of the 2-loop part of the @xmath92 invariant and the casson - walker invariant of cyclic branched coverings of a knot .
this relation has been settled by ( * ? ? ?
* corollary 1.4 ) . in the study of knot theory via surgery
, knots with alexander trivial polynomial are topologically ( but not smoothly ) slice , @xcite and also ( * ? ? ? * appendix ) .
the `` abelian '' and solvable invariants of knots with trivial alexander polynomial , such as their alexander module , casson - gordon invariants and all of their recent systematic obstructions of cochran - orr - teichner @xcite vanish .
however , already the euler degree @xmath86 part of the kontsevich integral @xmath94 ( which incidentally equals to the @xmath86-loop part of @xmath4 ) does not vanish on alexander polynomial @xmath91 knots .
the symbol of @xmath94 can be computed in terms of equivariant linking numbers ( see theorem [ thm.u2 ] ) and this gives good realization properties for @xmath94 .
the @xmath94 invariant offers an opportunity to settle an error in one of m. freedman s lemmas on knots with alexander polynomial @xmath91 , @xcite . to those who prefer moves that untie knots
, it might sound disappointing that the null - move fails to do so . on the other hand
, the geometric null - move describes a natural equivalence relation on knots , namely isometry of blanchfield forms .
concordance is another well - prized equivalence relation on knots .
if one could generate concordance in terms of surgery of some type of claspers , this could open the door for constructing a plethora of concordance invariants of knots .
it is known that surgery on certain claspers preserve concordance , @xcite , but it is also known that these moves do not generate concordance , @xcite .
the paper consists of five , largely independent , sections .
@xmath95 in section [ sec.loop ] we apply the topological calculus of claspers to the case of null - claspers which leads naturally to trivalent graphs with beads , and their relation to the graded quotients @xmath96 .
@xmath95 in section [ sub.zbrief ] we give a detailed study of the behavior of the aarhus integral under surgery on claspers .
in particular , counting arguments above the critical degree imply that the euler degree @xmath1 part of the kontsevich integral is a of null - type @xmath1 .
this provides a conceptual relation between the null - move and the euler degree of graphs .
this counting leads naturally to the study of the hairy strut part of the kontsevich integral .
struts correspond to linking numbers and the goal in the rest of the sections is to show that hairy struts correspond to equivariant linking numbers .
@xmath95 in section [ sec.abelian ] we review linking numbers and equivariant linking numbers , and give an axiomatic description of the latter , motivated by the calculus of claspers .
@xmath95 in the final section [ sub.kbehavior ] we show that the hairy strut part of the kontsevich integral satisfies the same axioms as the equivariant linking numbers , and as a consequence of a uniqueness result , the two are equal .
in this section we recall briefly the definition of surgery on claspers , with the notation of @xcite .
we refer the reader to @xcite for a detailed description . a framed graph @xmath13 in a 3-manifold @xmath7 is called a _ clasper of degree _
@xmath91 ( or simply , a _ y - graph _ ) , if it is the image of @xmath97 under a smooth embedding @xmath98 of a neighborhood @xmath19 of @xmath97 .
the embedding @xmath99 can be recovered from @xmath13 up to isotopy .
surgery on @xmath13 can be described either by surgery on the corresponding framed six component link , or in terms of cutting a neighborhood of @xmath13 , ( which is a handlebody of genus @xmath21 ) , twisting by a fixed diffeomorphism of its boundary and gluing back .
@xmath100 a clasper of higher degree is a ( thickening of ) an embedded trivalent graph , with some distinguished loops which we call _ leaves_. the degree of a clasper is the number of trivalent vertices , excluding those of the leaves .
surgery on claspers of higher degree is defined similarly .
for example , a clasper of degree @xmath86 and its corresponding surgery link are shown below : @xmath101 surgery on a clasper of degree @xmath1 can be described in terms of surgery on @xmath1 claspers of degree @xmath91 .
we will exclude claspers of degree @xmath17 , that is with no trivalent vertices . these correspond to @xmath22-moves in the language of goussarov and habiro . as we mentioned in the introduction ( see remark [ rem.linkingform ] ) , the study of the graded quotients @xmath96 is entirely analogous to the study of the graded quotients @xmath102 of s of s. for a detailed description of the latter invariants , we refer the reader to @xcite .
let us recall here some key ideas , for later use .
let @xmath103 denote the vector space over @xmath3 generated by s , and let @xmath104 denote the subspace of @xmath103 generated by @xmath105 $ ] for all disjoint collections of claspers @xmath106 in s @xmath7 , where @xmath107=\sum_{i \subset \{0,1\}^n } ( -1)^{|i| } m_{g_i}\ ] ] compare with equation .
a function @xmath108 is a _ of type _
@xmath1 iff @xmath109 .
let @xmath110 denote the graded quotients .
these quotients can be described in terms of a vector space @xmath111 defined as follows : @xmath111 is the completed vector space over @xmath3 generated by trivalent graphs with vertex - orientations , modulo the @xmath10 and @xmath11 relations . in @xcite we defined a degree - preserving map @xmath112 as follows : given an abstract vertex - oriented trivalent graph @xmath13 ,
choose an embedding @xmath113 that preserves its vertex - orientation .
recall that a vertex - orientation is a cyclic order of the set of three flags around a trivalent vertex .
an embedding @xmath114 gives rise at each vertex @xmath115 to a frame that consists of the tangent vectors of the images under @xmath114 of the three flags around @xmath116 . comparing this frame with the standard orientation of @xmath2
, we get a number @xmath117 at each vertex @xmath116 of @xmath13 .
we say that @xmath114 is orientation preserving if @xmath118
. we will consider the clasper @xmath119 of degree @xmath1 and the associated element @xmath120 \in \g_n ( s^3)$ ] .
lemma [ lem.slide ] implies that the element @xmath120 \in \g ( s^3)$ ] is independent of the embedding @xmath114 .
further , it was shown in ( * ? ? ?
* corollary 4.6 ) and ( * ? ? ?
* theorem 4.11 ) that the @xmath10 and the @xmath11 relations hold .
this defines the map over @xmath18 , which is obviously degree - preserving .
we now recall the following theorem of ( * ? ? ?
* theorem 4.13 ) : the map is onto , over @xmath83 $ ] .
let us comment on the theorem .
it is not obvious that the map is onto , since @xmath121 is generated by elements of the form @xmath105 $ ] for claspers @xmath13 of degree @xmath1 with leaves _ arbitrarily _ framed links , whereas the image of the map involves claspers whose leaves are linked in the pattern of @xmath17-framed hopf links .
let us briefly review a proof of theorem [ thm.ggpas3 ] , taken from @xcite , which adapts well to our later needs .
observe that each leaf of a clasper in @xmath2 is _
null homologous_. using lemmas [ lem.cut ] and lemma [ lem.slide ] we can write @xmath122 $ ] as a linear combination of @xmath123 $ ] such that each @xmath124 satisfies the following properties : * each leaf @xmath125 of @xmath124 is an unknot with framing @xmath17 or @xmath126 and bounds a disk @xmath127 .
* the disks @xmath127 of @xmath126-framed leaves @xmath125 are disjoint from each other , disjoint from @xmath124 , and disjoint from the disks of the @xmath17-framed leaves .
* each disk of a @xmath17-framed leaf intersects precisely one other disk of a @xmath17-framed leaf in a single clasp . using ( * ? ? ?
* lemma 4.8 ) , and working over @xmath83 $ ] , we may assume that @xmath124 as above has no @xmath126-framed leaves .
then , it follows that @xmath123 $ ] lies in the image of .
this completes the proof of theorem [ thm.ggpas3 ] .
if we fix a @xmath7 , we could define a filtration on the vector space @xmath128 generated by all homology spheres with a fixed linking form ( these are exactly @xmath17-equivalent to @xmath7 under a sequence of clasper moves ) and introduce a decreasing filtration on @xmath128 and graded quotients @xmath129 .
there is a degree - preserving map @xmath130 which is onto , over @xmath131 $ ] .
the proof uses the same reasoning as above together with the fact that every knot ( such as a leaf of a clasper ) in @xmath7 is null homologous in @xmath7 , once multipled by @xmath132 . in this section
we recall the key cutting and sliding lemmas . @xcite*(cutting a leaf )
* let @xmath133 and @xmath134 be the claspers of figure [ split3 ] in a handlebody @xmath135 embedded in a @xmath7 and let @xmath136 be a clasper of degree @xmath137 in the complement of @xmath135 . if @xmath138 , @xmath139 and @xmath140 then we have that @xmath107=[m , g']+[m , g '' ] \,\ , \promod \f_{m+1}(s^3).\ ] ] @xmath141 informally , we may think that we are splitting a leaf of @xmath13 into a connected sum .
@xcite*(sliding an edge ) * let @xmath142 and @xmath97 be the claspers of figure [ slide ] in a handlebody @xmath135 embedded in a manifold @xmath7 and let @xmath136 be a clasper of degree @xmath137 in the complement of @xmath135 . if @xmath143 and @xmath138 , then we have that @xmath144=[m , g ] \,\ , \promod \f_{m+1}(s^3).\ ] ] @xmath145 informally , we may think that we are sliding an edge of @xmath13 . our first goal in this section is to define the analogue of the map using trivalent graphs with beads .
this will be achieved by introducing _
finger moves _ of @xmath146-null claspers .
let us begin with an alternative description of the space @xmath77 .
consider a unitrivalent graph @xmath13 with vertex - orientation and edge orientation , and cut each edge to a pair of _ flags _
( i.e. , half - edges ) .
orient the two flags @xmath147 incident to an edge @xmath76 of @xmath13 as follows @xmath148 the ring @xmath149 is the polynomial ring over @xmath18 with generators @xmath150 for every flag @xmath76 of @xmath13 and relations : @xmath151 if the pair of flags @xmath152 is an edge of @xmath13 , and @xmath153 for all trivalent vertices @xmath116 of @xmath13 .
let us define an abelian group : @xmath154 where the sum is over all isomorphism classes of trivalent graphs with oriented edges and vertex - orientation . for a strut @xmath155 , a vortex @xmath156 and a theta graph @xmath157 , the associated rings are given by @xmath158 \\
\la_{\sfy } & \cong & \bz[t_1^{\pm 1},t_2^{\pm 1},t_3^{\pm 1}]/(t_1t_2t_3 - 1 ) \\
\la_{\theta } & \cong & \bz[t_1^{\pm 1},t_2^{\pm 1},t_3^{\pm 1}]/(t_1t_2t_3 - 1 , \sym_2 \times \sym_3)\end{aligned}\ ] ] where @xmath159 acts as a permutation on the @xmath160s and @xmath161 acts as an simultaneous involution of the powers of the @xmath160 .
consider an @xmath146-null clasper @xmath13 of degree @xmath1 , and let @xmath162 ( the superscipt stands for no leaves ) denote the abstract unitrivalent graph obtained from removing the leaves of @xmath13 .
then , there exists a geometric action @xmath163
\longto \g \so\ ] ] given by _
moves_. the action will be defined in terms of @xmath22-moves .
let us recall surgery on a clasper with no trivalent vertices , a so - called _
@xmath22-move _ : @xmath164 an @xmath22-move can be thought of as a right - handed _ finger move _ , or a right - handed dehn twist , and has an inverse given by a left - handed finger move ( indicated by a stroke in the opposite direction in @xcite )
. we will call right - handed @xmath22-moves positive , and left - handed ones negative . to define the action , without loss of generality
, we will assume that @xmath13 is a collection of claspers of degree @xmath91 .
further , we orient the flags of @xmath13 outward , towards their univalent vertices . consider a monomial @xmath165 where the product is over the set of flags of @xmath13 .
for each flag @xmath76 of @xmath13 choose a segment @xmath166 of @xmath69 , such that segments of different flags are nonintersecting .
for each flag @xmath76 of @xmath13 , choose a collection @xmath167 of @xmath168 @xmath22-claspers such that when @xmath169 , we have : @xmath170 let @xmath171 be the collection of @xmath22-claspers for all flags @xmath76 of @xmath13 , and let @xmath172 denote the image of @xmath13 after clasper surgery on @xmath173 .
this defines an element @xmath174 \in \g \so$ ] . we claim that this element depends on @xmath79 alone , and not on the intermediate choices of arcs and the @xmath22-claspers
indeed , lemma [ lem.slide ] implies that if an edge of @xmath175 passes through two oppositely oriented arcs of @xmath69 , resulting in a clasper @xmath124 , then @xmath176=[\so , g ' ] \in \g \so$ ] .
this implies easily our previous claim .
thus , we can define the action of @xmath79 on @xmath13 by @xmath177= [ \so , i_c \cdot g]$ ] .
it is easy to see that the relations of the ring @xmath178 hold .
we can now define map @xmath179 over @xmath18 as follows ( the superscipt stands for clasper ) : consider a trivalent graph @xmath13 with a vertex - orientation and an orientation of the edges , and let @xmath180 be an orientation preserving embedding in a small ball @xmath181 .
consider the element @xmath182 \in \g \so$ ] , as in the map .
using lemma [ lem.finger ] and the fact that @xmath13 is trivalent , we get a map @xmath183 induced by the action @xmath184
\to \g \so$ ] . just as in the map , the @xmath10 , @xmath11 and @xmath185 relations are preserved .
this defines the map .
we now give a version of the ring @xmath149 that uses half the number of generators : if @xmath13 is a vertex - oriented graph with oriented edges , let us define the ring @xmath186 over @xmath18 with generators @xmath150 for every edge @xmath76 of @xmath13 and relations : @xmath153 for all trivalent vertices @xmath116 of @xmath13 .
the next lemma identifies @xmath149 with @xmath186 : if @xmath13 has oriented edges , then we have a canonical isomorphism : @xmath187 if @xmath76 is an oriented edge , let @xmath152 denote the pair of flags such that @xmath76 has the same orientation as @xmath188 and opposite from @xmath189 .
this gives rise to a map @xmath190 and @xmath191 , which is an isomorphism .
we have a canonical isomorphism @xmath192 it follows immediately using lemma [ lem.lahalf ] above . combining the above lemma with the map , we get the desired map @xmath82 of theorem [ thm.loop1 ] .
( of theorem [ thm.loop1 ] ) we now claim that the proof of theorem [ thm.ggpas3 ] works without change , and proves theorem [ thm.loop1 ] .
indeed , lemmas [ lem.cut ] and [ lem.slide ] work for @xmath146-null claspers as stated .
furthermore , if @xmath193 is an @xmath146-null clasper , then each leaf of @xmath13 lies in the commutator group @xmath194=1 $ ] , where @xmath195 . since @xmath196=[[\pi,\pi],[\pi,\pi]]$ ] , this implies that each leaf of @xmath13 bounds a surface in @xmath197 whose bands are null homologous links in @xmath197 .
the surfaces for different leaves may intersect each other .
we can apply now the proof of theorem [ thm.ggpas3 ] exactly as stated , to conclude the proof of theorem [ thm.loop1 ] .
we end this section with an alternative description of the ring @xmath149 , and thus of the vector space of graphs @xmath77 , that was introduced by the second author @xcite and studied by p. vogel in unpublished work .
we thank p. vogel for explaining us his unpublished work . for a trivalent graph @xmath13 with oriented edges we have a canonical isomorphism @xmath198\ ] ] using lemma [ lem.lahalf ] , we will rather describe a canonical isomorphism @xmath199 $ ] . recall the exact sequence @xmath200 where @xmath201 and @xmath202 are the abelian groups of @xmath18-valued functions on the vertices and oriented edges of @xmath13 .
let @xmath203 .
this gives rise to the element @xmath204 where the product is taken over all edges of @xmath13 .
it is easy to see that this element depends on the image of @xmath79 in @xmath205 and gives rise to a map
@xmath206 \to \la_g$ ] .
it is easy to see that this map is an isomorphism .
let us define a vector space @xmath207 \cdot g \right)/(\as,\ihx , \aut)\ ] ] where the sum is over trivalent graphs with oriented edges .
lemma [ lem.aid1 ] and [ lem.rv ] imply that there are canonical isomorphisms : @xmath208 the map @xmath209 ( which is the composite of with the identification of the above corollary ) , can be defined more explicitly without appealing to finger moves .
indeed , observe that given an embedding @xmath210 , then taking linking number in @xmath2 of @xmath69 with @xmath91-cycles in @xmath211 , we get a canonical element @xmath212 . if @xmath213 are two embeddings of a trivalent graph of degree @xmath61 such that @xmath214 , then @xmath215=[\so,\psi(g ) ] \in \gl { 2n}\so$ ] .
our sliding lemma implies that an edge of @xmath13 can slide past two arcs of @xmath6 with opposite orientations .
it can also slide past another edge .
it is easy to see that this implies our result .
using the above lemma , and given a vertex - oriented trivalent graph @xmath13 and an element @xmath216 , let @xmath217 be any embedding with associated element @xmath79 .
this defines a map @xmath206 \cdot g
\to \g \so$ ] . the @xmath218 and @xmath185 relations are satisfied and this gives rise to a well - defined map @xmath219 given a cocommutative hopf algebra @xmath220 and an arbitrary graph @xmath13 with oriented edges , vogel associates an abelian group @xmath186 . in case @xmath221 $ ] with comultiplication @xmath222 and antipode @xmath223 the abelian group @xmath186 is the one given by definition [ def.lahalf ] . all the results of this section work without change if we replace the pair @xmath146 with a pair @xmath23 of a knot @xmath6 in a @xmath7 with trivial alexander polynomial .
in this section , we specialize the general principle of defining / calculating the aarhus invariant to the case of links obtained by surgery on claspers .
the reader is referred to @xcite for a detailed discussion of the aarhus integral . in the present paper
we will only be interested in pairs @xmath23 of knots @xmath6 in s @xmath7 , and @xmath224 of s @xmath19 . in case @xmath225
, we will be using the term `` kontsevich integral '' rather than lmo invariant or aarhus integral .
hopefully this will not cause any confusion or historical misunderstanding .
given a nondegenerate framed link @xmath226 in a @xmath2 ( i.e. , whose linking matrix is invertible over @xmath3 ) the aarhus invariant @xmath227 of @xmath228 is obtained from the kontsevich integral @xmath229 in the following way : * consider @xmath229 , an element of the completed ( with respect to the vassiliev degree ) vector space @xmath230 of chord diagrams on @xmath226-colored disjoint circles . * after some _ suitable basing _ of @xmath226 ( defined below ) we can lift @xmath229 to an element of the completed algebra @xmath231 of unitrivalent graphs on @xmath226-colored vertical segments . *
symmetrize the legs on each @xmath226-colored segment to convert @xmath229 to an element of the completed algebra @xmath232 of unitrivalent graphs with symmetric @xmath226-colored legs . *
separate the _ strut _ part @xmath233 from the other part @xmath234 : @xmath235 where @xmath236 contains no diagrams that contain an @xmath226-labeled strut component .
it turns out that the strut - part is related to the linking matrix of @xmath226 as follows : @xmath237 * glue the @xmath226-colored legs of the graphs of @xmath234 pairwise in all possible ways and multiply the created edges by the entries of the negative inverse linking matrix of @xmath226 . * finally renormalize by a factor that depends on the signature of the linking matrix of @xmath226 .
the end result , @xmath227 depends only on the 3-manifold @xmath228 and not on the surgery presentation @xmath226 of it or the basing of @xmath226 . what is a `` suitable basing of @xmath226 ? '' ideally , we wish we could choose a base point on each component of @xmath226 to convert @xmath226 into a union of @xmath226-labeled intervals . unfortunately , it is more complicated : a suitable basing of @xmath226 is a string - link representative of @xmath226 , equipped with _
relative scaling _
( i.e. , a parenthetization ) between the strings .
such objects were called _
q - tangles _ ( see @xcite ) by le - murakami , _ non - associative tangles _ by bar - natan ( see @xcite ) , and were described in the equivalent language of _ dotted morse links _ in ( * ? ? ?
* , section 3 ) .
notice that certain parts ( such as the @xmath226-labeled struts ) of the kontsevich integral of a based link @xmath226 are independent of the basing .
the above discussion of the aarhus integral also works when we start from pairs @xmath238 of nondegenerate framed links @xmath226 in @xmath7 .
the discussion also works in a relative case when we start from pairs @xmath239 ( with @xmath226 nondegenerate in a @xmath7 ) and do surgery on @xmath226 alone . in what follows
, we will assume silently that the links in question are suitably based , using parenthesized string - link representatives .
we will be interested in links that describe surgery on claspers .
consider a clasper @xmath13 of degree @xmath91 in @xmath2 ( a so - called ) .
surgery on the clasper @xmath13 can be described by surgery on a six component link @xmath240 associated to @xmath13 , where @xmath241 ( resp .
@xmath226 ) is the three component link that consists of the edges ( resp .
leaves ) of @xmath13 .
the linking matrix of @xmath242 and its negative inverse are given as follows : @xmath243 the six component link @xmath240 is partitioned in three blocks of two component links @xmath244 each for @xmath245 , the _ arms _ of @xmath13 .
a key feature of surgery on @xmath13 is the fact that surgery on any proper subset of the set of arms does not alter @xmath7 .
in other words , alternating @xmath7 with respect to surgery on @xmath13 equals to alternating @xmath7 with respect to surgery on all nine subsets of the set of arms @xmath246 .
that is , @xmath247)=z([s^3 , a]).\ ] ] due to the locality property of the kontsevich integral ( explained in @xcite and in a leisure way in ( * ? ? ?
* ii , section 4.2 ) ) , the nontrivial contributions to the right hand side of equation come from the part of @xmath248 that consists of graphs with legs _ touch _
( i.e. , are colored by ) all three arms of @xmath13 .
the above discussion generalizes to the case of an arbitrary disjoint union of claspers .
in this section we will explain briefly why the aarhus integral ( evaluated at s and graded by the euler degree ) is a universal .
we will use essentially the same ideas to prove theorem [ thm.loop2 ] in the following section .
the ideas are well - known and involve elementary counting arguments , see @xcite and also @xcite .
we begin by showing the following well - known proposition , which we could not find in the literature .
the euler degree @xmath1 part of @xmath249 is a type @xmath1 invariant of s with values in @xmath250 . since @xmath251
, it suffices to consider the case of even @xmath1 . recalling definition [ def.ihsfti ] ,
suppose that @xmath252 ( for @xmath253 ) is a collection of claspers in @xmath2 each of degree @xmath91 , and let @xmath254 denote the set of arms of @xmath13 .
equation and its following discussion implies that @xmath255)=z([s^3 , a])\ ] ] and that the nonzero contribution to the right hand side come from diagrams in @xmath256 that touch all arms .
thus , contributing diagrams have at least @xmath257 @xmath254-colored legs , to be glued pairwise .
since pairwise gluing needs an even number number of univalent vertices , it follows that we need at least @xmath258 @xmath254-colored legs .
notice that @xmath256 contains no struts .
thus , at most three @xmath254-colored legs meet at a vertex , and after gluing the @xmath254-colored legs we obtain trivalent graphs with at least @xmath259 trivalent vertices , in other words of euler degree at least @xmath260 .
thus , @xmath261)=0 \in \a_{2n}(\phi)$ ] , which implies that @xmath262 is a invariant of s of type @xmath61 with values in @xmath263 .
sometimes the above vanishing statement is called _ counting above the critical degree_. our next statement can be considered as _ counting on the critical degree_. we need a preliminary definition .
consider a clasper @xmath13 in @xmath2 of degree @xmath61 , and let @xmath264 denote the collection of degree @xmath91 claspers @xmath265 which are obtained by inserting a hopf link in the edges of @xmath13 .
then the _ complete contraction
_ @xmath266 of @xmath13 is defined to be the sum over all ways of gluing pairwise the legs of @xmath267 , where we multiply the resulting elements of @xmath111 by the product of the linking numbers of the contracted leaves .
if @xmath13 is a clasper of degree @xmath61 in @xmath2 , then @xmath268)=\la g \ra \in \a_{2n}(\phi)\ ] ] it suffices to consider a collection @xmath269 of claspers in @xmath2 each of degree @xmath91 .
let @xmath254 denote the set of arms of @xmath13 .
the counting argument of the above proposition shows that the contributions to @xmath261)=z_{2n}([s^3 , a])$ ] come from complete contractions of a disjoint union @xmath270 of @xmath61 _ vortices_. a vortex is the diagram @xmath156 , the next simplest unitrivalent graph after the strut .
furthermore , the @xmath271 legs of @xmath272 should touch all @xmath271 arms of @xmath13 .
in other words , there is a 1 - 1 correspondence between the legs of such @xmath272 and the arms of @xmath13 .
consider a leg @xmath125 of @xmath272 that touches an arm @xmath273 of @xmath13 .
if @xmath125 touches @xmath274 , then due to the restriction of the negative inverse linking matrix of @xmath13 ( see section [ sub.lmoy ] ) , it needs to be contracted to another leg of @xmath272 that touches @xmath275 . but this is impossible , since the legs of @xmath272 are in 1 - 1 correspondence with the arms of @xmath13 . thus , each leg of @xmath272 touches presicely one edge of @xmath13 . in particular ,
each component @xmath276 of @xmath272 is colored by three edges of @xmath13 .
notice that the set of edges of @xmath13 is an algebraically split link . given
a vortex colored by three edges of @xmath13 , the coefficient of it in the kontsevich integral equals to the _ triple milnor invariant _ , ( as is easy to show , see for example @xcite ) and vanishes unless all three edges are part of a degree @xmath91 clasper @xmath265 .
when the triple milnor invariant does not vanish , it equals to @xmath91 using the orientation of the clasper @xmath265 .
thus , the diagrams @xmath272 that contribute are a disjoint union of @xmath61 vortices @xmath277 and these vortices are in 1 - 1 correspondence with the set of claspers @xmath278 , in such a way that the legs of each vortex @xmath276 are colored by the edges of a unique clasper @xmath279 . after we glue the legs of such @xmath280 using the negative inverse linking matrix of @xmath13 , the result follows .
let us mention that the discussion of propositions [ prop.mku1 ] and [ prop.mku2 ] uses the unnormalized aarhus integral ; however since we are counting above the critical degree , we need only use the degree @xmath17 part of the normalization which equals to @xmath91 ; in other words we can forget about the normalization .
the above proposition is useful in realization properties of the @xmath262 invariant , but also in proving the following _ universal property _ : for all @xmath1 , the composite map of equation and proposition [ prop.mku2 ] @xmath281 is the identity .
since the map on the left is onto , it follows that the map is an isomorphism , over @xmath3 .
we may call the map @xmath282)$ ] ( where @xmath13 is of degree @xmath61 ) the _ symbol _ of @xmath262 . in the above propositions [ prop.mku1]-[prop.mku3 ] ,
@xmath2 can be replaced by any ( or even a s ) @xmath7 .
our goal in this section is to modify ( when needed ) the discussion of sections [ sub.lmobrief]-[sub.auni ] and deduce in a natural way the proof of theorem [ thm.loop2 ] .
it is clear that struts and vortices and their coefficients in the kontsevich integral play an important role .
as we will see presently , the same holds here , when we replace struts and vortices by their _ hairy _ analogues : @xmath283 we begin by considering an @xmath146-null clasper @xmath13 . we know already that we can compute @xmath284)$ ] from the kontsevich integral @xmath285 of the corresponding link by integrating along @xmath13-colored struts .
choose string - link representatives of @xmath286 with relative scaling .
we will denote edges and leaves of @xmath13 by @xmath241 and @xmath226 respectively .
the diagrams that appear in @xmath287 are _ hairy _ , in the sense that upon removal of all @xmath69-colored legs , they are simply unitrivalent graphs whose legs are colored by @xmath13 .
let us separate the hairy strut part @xmath288 from the other part @xmath289 : @xmath290 using the disjoint union multiplication , where @xmath291 contains no diagrams that contain an @xmath226-labeled hairy strut component .
let us write the hairy strut part as follows : @xmath292 where @xmath240 is the corresponding link of leaves and edges of @xmath13 and @xmath293 in the above equation , @xmath294 is the coefficient in @xmath295 of the hairy diagram appearing on the right .
we will call @xmath296 the _ hairy linking matrix_. the hairy linking matrix of @xmath13 and its negative inverse are given by @xmath297 since @xmath298 is an unlink and @xmath299 is the disjoint union of an unknot @xmath300 and the link @xmath301 , it suffices to consider the case of a hairy strut @xmath302 . in this case
, @xmath300 is a meridian of @xmath303 . the formula for the kontsevich integral of the long hopf link of @xcite ( applied to @xmath300 ) , together with the fact that @xmath303 has linking number zero with @xmath69 , imply that the hairy part @xmath302 vanishes .
@xmath304)\in \a(\ast_\o)$ ] can be computed from @xmath305 by gluing pairwise the legs of the @xmath13-colored graphs of @xmath306 using the negative inverse hairy linking matrix .
we have that @xmath307 in other words , the diagrams that contribute in @xmath304)$ ] are those whose components either lie in @xmath289 or are hairy struts of the shape @xmath308 . because of the restriction of the linking matrix , the pair of legs @xmath309 of a hairy strut of the above shape must be glued to a pair of legs labeled by @xmath310 as follows : @xmath311 the result of this gluing is equivalent to gluing pairs of @xmath312 colored legs using the negative inverse hairy linking matrix .
this concludes the proof of the lemma .
given this lemma , the proof of proposition [ prop.mku1 ] works without change and proves theorem [ thm.loop2 ] . furthermore , observe that the coefficients of hairy vortices with legs colored by the edges of @xmath13 and with nonzero number of hair vanish .
this is true since the borromean rings ( i.e. , the edges of @xmath13 ) form an unlink in the complement of @xmath69 .
thus , the proof of proposition [ prop.mku2 ] implies the following if @xmath13 is a @xmath146-null clasper of degree @xmath61 , then @xmath313)=\la g \ra \in \a_{2n}(\ast)\ ] ] where complete contractions are using the matrix of lemma [ lem.f ] instead and put the entries of it as beads on the edges that are created by the contraction of the legs .
let us remark that the analogue of proposition [ prop.mku3 ] does not hold since @xmath4 takes values in @xmath9 rather than in @xmath77 .
the invariant @xmath92 discussed in section [ sub.further ] takes values in @xmath77 , satisfies the analogue of [ prop.mku2 ] and is thus a universal @xmath3-valued invariant of alexander polynomial @xmath91 knots , with respect to the null - move .
in theorem [ thm.u2 ] we calculated the symbol @xmath314
\in \a_{2n}(\la)$ ] in terms of a complete contraction of @xmath13 that uses the the hairy linking matirx .
struts correspond to linking numbers and we will show that hairy struts correspond to equivariant linking numbers . the goal of this section is to discuss the latter .
recall that all links are oriented .
we begin by recalling the definition of the _ linking number _
@xmath315 of a link @xmath316 of two ordered components in @xmath2 .
since @xmath317 , there is an oriented surface @xmath318 that @xmath319 bounds .
we then define @xmath320\cdot [ l_2 ] \in \bz$ ] where @xmath321 is the intersection pairing .
since @xmath322 , the result is independent of the choice of surface @xmath318 .
the following lemma summarizes the well - known properties of the linking numbers of two component links in @xmath2 : @xmath315 does not depend on the ordering of the components of @xmath226 .
if a component @xmath303 of @xmath226 is a connected sum of @xmath323 and @xmath324 ( as in figure [ split3 ] ) , then @xmath325 * ( initial condition ) * @xmath326 * ( uniqueness ) * if a function of two - component links satisfies the symmetry , cutting , and initial conditions , then it equals to @xmath327 .
it is easy to see that linking numbers satisfy the above axioms
. for uniqueness , assume that @xmath328 is another such function , and let @xmath329 .
consider a link @xmath316 , and a surface @xmath318 that bounds @xmath319 .
using the cutting property , as in the proof of theorem [ thm.ggpas3 ] , it follows that @xmath330 is a linear combination of @xmath331 and @xmath332 , and hence zero .
the above definition of linking number can be extended to the case of an two component links @xmath226 in a @xmath7 . in that case
, the linking number takes values in @xmath333 , and proposition [ lem.mu ] continues to hold . the above definition of linking number can be extended to the case of a two - component link @xmath226 in a possibly open 3-manifold @xmath19 that satisfies @xmath334 . in this section
we review well - known results about equivariant linking numbers .
these results are useful to the _ surgery view _ of knots , ( see @xcite and also @xcite ) and will help us identify the hairy struts with equivariant linking numbers .
we begin with the following simple situation : consider a null homologous link @xmath316 of two ordered components in @xmath335 .
we will call such links @xmath226 @xmath336-_null_. consider the universal abelian cover @xmath337 corresponding to the natural map @xmath338 .
since @xmath226 is null homologous , it lifts to a link @xmath339 , which is invariant under the action by the group of deck transformations @xmath18 .
note that @xmath41 is an open 3-manifold diffeomorphic to @xmath340 , in particular @xmath341 . using remark [ rem.qlk2 ]
, we can define linking numbers between the components of @xmath339 .
consider an arc - basing @xmath342 of @xmath226 , that is an embedded arc in @xmath343 that begins in one component of @xmath226 and ends at the other , and is otherwise disjoint from @xmath226 .
then , we can consider a lift of @xmath344 to @xmath41 which is an arc - based two component link @xmath345 in @xmath41 .
if @xmath346 is an @xmath146-null arc - based link , we define the _ equivariant linking number _ by @xmath347 the above sum is finite .
moreover , since linking numbers ( in @xmath41 ) are invariant under the action of deck transformations , it follows that the above sum is independent of the choice of lift of @xmath344 to @xmath41 .
the following lemma summarizes the properties of the equivariant linking number of @xmath146-null arc - based links .
@xmath73 stands for the map @xmath348
. if @xmath349 is a permutation of the two components of @xmath226 then , we have @xmath350 * ( specialization ) * @xmath351 * ( cutting ) * suppose that a component @xmath303 of an @xmath336-null link @xmath346 is a connected sum of @xmath323 and @xmath324 ( as in figure [ split3 ] ) , and @xmath352 and @xmath353 are @xmath336-null .
then , @xmath354 * ( @xmath75-sliding ) * if @xmath355 denote the result of sliding the arc - basing of the first component of @xmath346 along an oriented arc - based curve @xmath0 , then @xmath356 where @xmath357 . if @xmath226 lies in a ball disjoint from @xmath69 , then @xmath358 * ( uniqueness ) * if a function of @xmath146-null arc - based links satisfies the symmetry , @xmath75-sliding , cutting and initial conditions , then it equals to the equivariant linking number . the symmetry , cutting and @xmath75-sliding properties follows immediately from lemma [ lem.mu ] .
the specialization property follows from the well - known fact that given a covering space @xmath359 and a cycle @xmath79 in @xmath343 that lifts to @xmath360 in @xmath41 , and a cycle @xmath361 in @xmath41 , then the intersection of @xmath360 with @xmath361 equals to the intersection of @xmath79 with the push - forward of @xmath361 in @xmath343 .
the uniqueness statement follows from the proof of theorem [ thm.loop1 ] given in section [ sub.graded ] .
the definition [ def.lkf ] and lemma [ lem.lkprops ] can be extended without change to the case of a null homologous arc - based two component link @xmath346 in @xmath25 where @xmath23 is a knot @xmath6 in a @xmath7 with trivial alexander polynomial .
notice that @xmath362 and the condition on the alexander polynomial ensures that @xmath363
. a further generalization of definition [ def.lkf ] and lemma [ lem.lkprops ] is possible to the case of a null homologous arc - based two component link @xmath346 in @xmath25 where @xmath23 is a knot @xmath6 in a @xmath7 . in that case
, @xmath364 lies in a _ localization _
@xmath365 of @xmath75 defined by @xmath366 this specializes to tha case of @xmath226 being a zero - framed parallel of a null homologous knot @xmath367 in @xmath368 and @xmath342 a short arc between the two components of @xmath226 . in this case , @xmath369 coincides with the self - linking @xmath370-function @xmath371 of kojima - nakanishi , @xcite .
the purpose of this section is to show that the hairy struts coincide with equivariant linking numbers .
we begin with a definition .
a _ special link _
@xmath372 is the union of @xmath69 and an @xmath146-null link .
note that a special link @xmath226 has a special component , namely @xmath69 .
in what follows , @xmath342 will refer to a _ disk - basing _
( i.e. , a string - link representative ) of @xmath226 . consider a disk - based special link @xmath346 , together with a choice of relative scale , and its kontsevich integral .
there is an algebra isomorphism @xmath373 with inverse the symmetrization map @xmath374 which is the average of all ways of placing symmetric @xmath226-colored legs on @xmath226-intervals . in what follows
, we will denote by @xmath229 the image of the kontsevich integral in either algebra , hopefully without causing confusion .
consider a homotopy quotient @xmath375 of the algebra @xmath376 where we quotient out by all nonforests ( i.e. , graphs with at least one connected component which is not a forest ) and by all _ flavored _ forests that contain a tree with at least two legs flavored by the same component of @xmath226 , see also @xcite and @xcite .
a _ flavoring _ of a graph is a decoration of its univalent vertices by a set of colors , which in our case is @xmath226
. we will be interested in the lie subalgebra @xmath377 of @xmath375 spanned by all connected diagrams .
it is spanned by hairy trees colored by @xmath378 such that each label of @xmath379 appears at most once .
notice that if @xmath380 and @xmath381 are @xmath254-colored and @xmath382-colored hairy trees then @xmath383= \begin{cases } 0 & \text{if $ |a \cup a'|\neq 1 $ } , \\
t\cdot_{i } t ' & \text{if $ a \cup a'=\{i\}$ } \end{cases}\ ] ] where @xmath384 is the result of _ grafting _ the trees @xmath380 and @xmath381 along their common leg @xmath385 .
let us define @xmath386 to be the image of @xmath229 in @xmath387 .
the group - like property of the kontsevich integral implies that @xmath388 lies in @xmath389 . in the case
@xmath226 has three components , @xmath377 has a basis consisting of the trees @xmath390 , for @xmath391 , together with the degree @xmath91 struts @xmath392 and @xmath393 , and that @xmath394 here the product ( and the commutator ) are taken with respect to the natural multiplication on @xmath231 .
it follows by the definition of the hairy linking matrix that for every special link @xmath226 , we have @xmath396 where we explicitly denote the dependence on the disk basing @xmath342 , and where the sum is over all special sublinks @xmath397 of @xmath226 that contain @xmath69 and two more components of @xmath379 ( with repetition )
. it will be more convenient to present the proof in the algebra @xmath231 .
let @xmath401 denote a change of scaling of a string - link @xmath349 , as shown in the following figure in case @xmath349 has three stands : @xmath402 the locality of the kontsevich integral implies that @xmath403 for an associator @xmath404 .
write @xmath405 for an element @xmath406 $ ] , where @xmath407 is the free - lie algebra @xmath407 of two generators @xmath408 . the following identity @xmath409 ( valid in a free lie algebra of two generators ) implies that @xmath410 .
if we project the above equality to the quotient @xmath411 where connected diagrams with two legs either both on the first strand or both on the second strand vanish , then it follows that we can replace @xmath412 $ ] by its image @xmath413/[\cl , [ \cl,\cl]]$ ] .
it is easy to see that @xmath414 $ ] for _ any _ associator @xmath404 .
now we can finish the proof of the lemma as follows .
consider a sting - link @xmath349 with relative scaling obtained from a disk - basing of a special link @xmath226 of three components and let @xmath397 be the one obtained by a change of relative scaling of @xmath226 .
projecting to @xmath415 , and using the fact that @xmath388 lies in the center of the lie algebra @xmath416 , it follows that @xmath417 .
consider the special link @xmath420 , ( this is an abbreviation for @xmath421 ) whose connected sum of the first two components gives @xmath346 .
how does the kontsevich integral of @xmath422 determine that of @xmath423 ?
the answer , though a bit complicated , is known by ( * ? ? ?
* part ii , proposition 5.4 ) .
following that notation , we have @xmath424 where @xmath425 is the operation that glues all @xmath426-colored legs of @xmath254 to those of @xmath427 ( assuming that the number of legs of color @xmath428 and of color @xmath429 in @xmath254 and @xmath427 match ; otherwise it is defined to be zero ) , and @xmath430 where @xmath431 is an ( infinite ) linear combination of of rooted trees with at least one trivalent vertex whose leaves are colored by @xmath432 and whose root is colored by @xmath319 .
we will call such trees @xmath433-trees
. the reader may consult ( * ? ? ? *
part ii , proposition 5.4 ) for the first few terms of @xmath434 , which are given by _ any _ baker - cambell - hausdorff formula , translated in terms of rooted trees . upon projecting the answer to the quotient @xmath435
, the above formula simplifies .
indeed , if we glue some disjoint union of trees of type @xmath436 that contain at least one trivalent vertex to some hairy @xmath437-colored trees , the resulting connected graph will either have nontrivial homology , or at least two labels of @xmath438 .
such graphs vanish in @xmath435 .
thus , when projecting the above formula to @xmath435 , we can assume that @xmath439 .
using the fact of how the kontsevich integral of a link determines that of its sublinks and the above , it follows that @xmath440 which , together with equation concludes the proof . in an artistic way , the next figure shows the result of a slide move ( compare also with figure [ finger ] ) that replaces @xmath319 by @xmath443 , where @xmath444 is the orientation reversed knot . @xmath445 as in the previous lemma , we have that @xmath446 where @xmath447 is the operation that replaces an @xmath0-colored leg to an @xmath448-colored one . upon projecting to @xmath449 ,
the above formula simplifies .
indeed , the @xmath450-colored hairy trees are ( after removal of the hair ) of two shapes @xmath156 and @xmath155 .
when we glue , a @xmath450-colored hairy @xmath451 vanishes in @xmath449 .
the remaining @xmath450-colored hairy trees are of the shapes @xmath452 as well as the hairless @xmath453 and @xmath454 .
when glued to @xmath455 and @xmath456 rooted trees , the ones of shapes @xmath457 vanish in @xmath458 , thus we remain it remains to consider only trees of shape @xmath459 and the hairless @xmath453 and @xmath454 .
when we glue these to @xmath455 and @xmath456 rooted trees , the only nonzero contribution comes from gluings like @xmath460 the above figure shows that each of the above gluings can be thought of as starting from a hairy graph of shape @xmath461 , and adding to it some additional hair , first on the left and then on the right ; each time multiplying the result by @xmath462 ( where @xmath463 ) and @xmath464 is the number of left and right added hair . on the other hand , adding hair is the same as commuting with @xmath392 ( as follows by equation ) . translating from gluings back to lie algebras , it follows that @xmath465 where @xmath466 means to replace the label @xmath91 by @xmath467 in the expression @xmath241 .
equation implies that @xmath468 which concludes the proof .
the symmetry follows by equation .
specialization follows from the fact that @xmath470 coincides with the coefficient of a strut in the kontsevich integral , which equals to @xmath471 .
the @xmath75-sliding property follows from lemma [ lem.fslide ] , the cutting property follows from lemma [ lem.fcut ] .
the initial condition property follows from the fact that the kontsevich integral is multiplicative for disjoint union of links , thus the only diagrams @xmath472 that contribute to the kontsevich integral ( and also in @xmath418 ) in this case is @xmath473 , which contributes the linking number of @xmath319 and @xmath438 .
since the coefficient @xmath474 of @xmath475 in the power series @xmath400 are given by a combination of milnor s invariants ( as follows by the work of habegger - masbaum @xcite for @xmath225 and forthcoming work of the first author for a general ) , we obtain that this generalizes a result of cochran , @xcite , who showed that a power series expansion of the kojima self - linking function was a generating function for a special class of repeated milnor invariants of type @xmath478 of a two component algebraically split link @xmath479 . in an earlier version of the paper , we also identified hairy struts with equivariant triple milnor linking numbers , for special links of four components . using this
, we gave an algorithm for computing the @xmath86-loop part ( or , equivalently , the euler degree @xmath86 part ) of the kontsevich integral of a knot with trivial alexander polynomial .
this involves a more delicate counting below the critical degree , in the language of section [ sub.lmoy ] .
for an easier digestion of the present paper , we prefer to come back to this matter in a future publication . * acknowledgement * we wish to thank d. bar - natan , j. conant , ttq le , j. levine , p. teichner and d. thurston for many stimulating conversations and the anonymous referee for pointing out several improvements of an earlier version .
le , _ an invariant of integral homology 3-spheres which is universal for all finite type invariants _ , soliton geometry and topology : on the crossroad , ams translations * 2 * eds .
v. buchstaber , s. novikov . | the kontsevich integral of a knot is a graph - valued invariant which ( when graded by the vassiliev degree of graphs ) is characterized by a universal property ; namely it is a universal vassiliev invariant of knots .
we introduce a second grading of the kontsevich integral , the euler degree , and a geometric null - move on the set of knots .
we explain the relation of the null - move to @xmath0-equivalence , and the relation to the euler grading of the kontsevich integral .
the null move leads in a natural way to the introduction of trivalent graphs with beads , and to a conjecture on a rational version of the kontsevich integral , formulated by the second author and proven in @xcite .
stavros 1991 _ mathematics classification .
_ primary 57n10 .
secondary 57m25 .
kontsevich integral , claspers , null - move , @xmath0-equivalence , blanchfield pairing , euler degree , beads , trivalent graphs , hair map , hairy struts , hairy vortices , finite type invariants , @xmath1-equivalence . ] |
th is thankful to i.o kulik ( department of physics , bilkent university ) for discussions and critical reading of the manuscript .
99 dj wineland , c monroe dm meekhof et al .
a * 454 * , 411 ( 1998 ) .
il chuang , n gershenfeld , mg kubinec et al .
a * 454 * , 447 ( 1998 ) .
y nakamura , ya pashkin and js tsai , nature * 398 * , 786 ( 1999 ) .
d deutsch , a barenco and a ekert , proc .
a * 449 * , 669 ( 1995 ) ; s lloyd , phys .
* 75 * , 346 ( 1995 ) ; dp divincenzo , phys .
a * 51 * , 1015 ( 1995 ) ; a barenco , proc .
a * 449 * , 679 ( 1995 ) .
j. clarke , an cleland , mh devoret et al . ,
science * 239 * , 992 ( 1988 ) ; r rouse , s han and je lukens , phys
* 75 * , 1614 ( 1995 ) .
jonathan r. friedman , vijay patel , w. chen , s.k .
tolpygo and j.e .
lukens , _ detection of a schrdinger s cat state in an rf - squid _ [ cond - mat/0004293 ] .
michael a. nielsen and isaac l. chuang , _ quantum computation and quantum information _
( cambridge univ .
press , 2000 ) .
yuriy makhlin , gerd schn and alexander shnirman , rev .
73 * , 357 ( 2001 ) .
mooij , t.p .
orlando , l. levitov et al . , science bf 285 , 1036 ( 1999 ) ; t.p .
orlando , j.e .
mooij , lin tian et al .
, phys . rev .
* b 60 * , 15398 ( 1999 ) .
l. tian , l.s .
levitov , c.h .
van der wal et al . in _
quantum mesoscopic phenomena and mesoscopic devices in microelectronics _
i.o . kulik and r. ellialtiolu , nato - asi series e. ( kluwer academic publ . 2000 ) .
leggett , s. chakravarty , a.t .
dorsey et al . , rev .
phys . * 59 * , 1 ( 1987 ) ; u. weiss , _ quantum dissipative systems _ ( world sci .
fig.1 : lifting the four - fold degeneracy stepwise in the proposed mechanism . in the ordering of the flux states shown in ( c )
[ see eq.([updownst.1 ] ) for the definition of the flux states ] with respect to the energy eigenvalues we considered @xmath9 . note that the initial four - fold degeneracy causes the energy centers indicated by the horizontal dashed lines to be independent of the couplings .
fig.2 : inductive coupling of the two rf - squid loops .
the dc - squids have two - junctions for balanced read - in and read - out the flux at each rf - squid loop independently .
although the details of the mutual field screening is not shown , the circuitry should be designed such that the i / o devices are decoupled from each other completely .
fig.3 : pseudo - degenerate energy eigenfunctions and the corresponding single well localized flux states .
the flux states @xmath149 are properly normalized by @xmath150 .
the parameters used are @xmath151 , @xmath50 , @xmath29 , @xmath30 .
fig.4 : lifting the degeneracy by tilting the symmetric double - well potential in the coupled squid pair .
the vertial scale is the numerically calculated eigen energies in ev units .
the horizontal variable corresponds to no tilt when @xmath50 and critical tilt when @xmath132 .
the tilting parameter @xmath152 simulates the coupling @xmath2 in the proposed mechanism in section a. the parameters used are @xmath151 , @xmath29 , @xmath30 and @xmath153 .
fig.5 : the transformation from the flux basis @xmath76 to the parity basis by two hadamard transformations .
after @xmath10 is applied on the flux basis , the states are @xmath154 ; + @xmath155 with the energies @xmath156 , + @xmath157 . | a novel mechanism is proposed for single and double qubit state manipulations in quantum computation with four - fold degenerate energy levels .
the principle is based on starting with a four - fold degeneracy , lifting it stepwise adiabatically by a set of control parameters and performing the quantum gate operations on non - degenerate states . a particular realization of the proposed mechanism
is suggested by using inductively coupled rf - squid loops in the recently observed macroscopic quantum tunneling regime where the energy eigen levels are directly connected with the measurable flux states .
the one qubit and two qubit controlled operations are demonstrated explicitly .
the appearance of the flux states also allows precise read - in and read - out operations by the measurement of the flux .
recent advances on the experimental efforts to demonstrate the fundamental single and two bit quantum gates by ion trap@xcite and nuclear magnetic resonance ( nmr ) experiments@xcite as well as the observation of the coherent rabi oscillations in cooper pair boxes@xcite are hinting at new challenges awaiting the realization of the quantum computational devices .
it has been shown theoretically that the basic gate operations based on the unitary transformations of the qubits can be performed by single and double qubit manipulations@xcite .
therefore it is a necessary first step that a proposed mechanism should demonstrate these fundamental operations as well as a successful read - in and read - out before any decoherence comes into play .
a next step which is not detailed in this letter is to construct arrays of independent qubits or coupled qubit pairs to perform parallel gate operations . in our suggested mechanism
the state space is four dimensional corresponding to the state space of a pair of qubits .
the paired qubit states are inseparable ; nevertheless , each qubit can be manipulated individually .
the mechanism also allows controlled operations on selected bits rather easily without affecting those bits that need to be unchanged .
a more detailed discussion of the proposed mechanism with its specific experimental realizations using rf - squid loops will be published elsewhere .
the mechanism is based upon coupling a pair of ideal two level systems with manifestly degenerate energy levels at the energy @xmath0 by two complex coupling parameters as indicated in fig.1a .
the coupling @xmath1 lifts the four - fold degeneracy to two - fold ( fig.1b ) .
the second coupling @xmath2 couples the two remaining degenerate levels in each pair and lifts the degeneracy completely ( fig.1c ) .
the states in the initial four - fold degenerate configuration at @xmath3 are labelled by @xmath4 .
we consider them to be the eigen states of some underlying non interacting system . in the qubit language below the first ( second ) bit
is @xmath5 .
the states @xmath6 are the components of the four dimensional row vector @xmath7 which is to be considered as the basis for the model hamiltonian .
the hamiltonian including the degeneracy breaking couplings is given in this four dimensional basis by the hermitian matrix @xmath8 without loss of generality we assumed that the couplings are _ real and positive _ and that they are applied adiabatically in the order @xmath1 first and @xmath2 second ( the order is commutative ) with @xmath9 .
the diagonalizing matrices for independent couplings are unitary transformations acting on the four dimensional basis .
these are @xmath10 for @xmath11 and @xmath12 for @xmath13 as given by @xmath14 when both couplings are turned on , the hamiltonian ( [ prl.1 ] ) is diagonalized by @xmath15 as @xmath16 which correspond to the levels in fig.1c . for arbitrary couplings @xmath17
the eigen energies are all nondegenerate and they are distributed symmetrically around the major center @xmath0 and the two minor ones @xmath18 .
the single and double qubit quantum gate operations are performed ideally by manipulating the strengths of at least one of the coupling constants and making use of the invariance of the major center under adiabatic changes of both couplings and of the two minor centers under that of @xmath2 . in fig.1c
we have indicated the eigenstates with up and down arrows . in the double - squid model considered below
they correspond to the flux states .
these states are connected with the initial basis states @xmath6 by @xmath19 note that the ordering of the @xmath6 basis is made with respect to the ordering of the energy eigen values of the flux states on the left hand side in ( [ updownst.1 ] ) which is different than the ordering used for constructing the hamiltonian matrix ( [ prl.1 ] ) . for later convenience
we introduce the short notation @xmath20 physically , this suggested four state mechanism can be realized by using inductively coupled rf - squids as depicted in fig .
2 .
the squids must operate in the macroscopic quantum tunneling ( mqt ) regime@xcite which requires temperatures on the order of mili kelvin for standard josephson junctions .
the four level system of fig.1a is constructed by preparing two identical squids a and b in which each qubit is represented , to a good approximation , by its two lowest energy eigen states ( denoted as @xmath21 and @xmath22 ) ; in the mqt regime these are nearly degenerate ( pseudo - degenerate ) in energy ( due to the fluctuations in the flux ) compared to the next ( third ) level . a qubit which is close to being ideal
is characterized by an isolated hypothetical two - level system with a high qubit quality ratio @xmath23 , where @xmath0 and @xmath24 are the two lowest levels and @xmath25 is the next ( well - separated ) level . in the case of manifest degeneracy , i.e. @xmath26 one obtains @xmath27 , the ideal case . for the realistic case here using coupled squids ,
we search for a pseudo degenerate configuration ( by performing a numerical search at different parameter values ) in which we obtain sufficiently high qubit quality factors on the order of @xmath28 . in fig.3
the calculated wavefunctions for the two lowest states of a symmetric double - well single - squid potential is plotted for the specific parameter values which are junction capacitance @xmath29 , the loop inductance @xmath30 and the potential modulation parameter@xcite @xmath31 where @xmath32 is the critical josephson current , @xmath33 is the superconducting current in the loop and @xmath34 is the superconducting flux quantum @xmath35 . in the numerical calculations the range of the junction and
the loop parameters are chosen such that the pseudo - degenerate eigen states are located in the double well , whereas the third level is above the central maximum of the double well potential . for the flux dominated regime the josephson energy @xmath36 is expected to be much larger than the charging energy @xmath37 in the junctions . considering that the sum of these energies is fixed ( @xmath38 is the total energy in the noninteracting squid loop ) , we find for our parameters @xmath39 for the symmetric configuration of the double well potential . if the symmetric configuration is tilted , the additional level splitting raises the charging energy . for the range of interest of the bias flux ( see fig .
4 ) a drop to @xmath40 in the above ratio is observed as a result of the tilting . the pseudo- degenerate single squid states @xmath21 and @xmath22 have respectively even and odd parities .
we will refer to these states as the one - bit parity basis .
the odd and even superpositions of the parity states are given by @xmath41 and @xmath42 .
these are the single - well localized flux states .
hence , we refer to them as the one - bit flux basis .
the single - well localized means that these one - bit states actually refer to a single squid where they are well confined to a single well of the one - squid double - well potential .
the calculated flux states corresponding to the chosen squid parameters above are also shown in fig.3 ( the solid triangles ) .
we now change from the single bit - single squid to the double bit- double squid configuration .
the case for two identical squid loops in the pseudo - degenerate regime with no mutual interaction is given by fig.1a .
inductively coupling these two identical squid loops in fig.2 by a weak inductance @xmath43 , lifts the four - fold degeneracy to two - fold degeneracy . in the coupled squid model the weak mutual inductance
can not be varied after fabrication .
this mutual inductance corresponds to the coupling @xmath1 ( fig.1 ) and we believe that @xmath44 is a reasonable range to consider in the experiment .
once the squids are manifactured , @xmath45 is fixed , therefore the four fold degenerate configuration in fig.1a is out of reach .
however we still have leverage on the second coupling .
this second coupling is the tilt of the symmetrical josephson potential of the two squids simultaneously deriving both potentials to unsymmetric configurations by means of an additionally applied flux @xmath46 ( the total flux is @xmath47 ) .
this is modelled by the second coupling @xmath2 which lifts the remaining two - fold degeneracy completely .
the combined effects of the mutual inductance , which we now consider to be @xmath48 , and the tilt can be calculated numerically by diagonalizing the inductively coupled two - squid hamiltonian .
the calculated eigenenergies are plotted against the tilting parameter @xmath49 in fig.4 .
the configuration @xmath50 ( i.e. @xmath51 ) corresponds to the symmetric squid potential .
the solution of the schrdinger equation for this potential and this coupled squid configuration has two closely spaced energy levels ( at the mentioned parameter range above ) comprising the best pseudo - degenerate approximation we ever found to the manifest double degeneracy in fig.1b .
at this point we read from the vertical axis in fig.4 that @xmath52 .
as the double - well potentials are tilted in parallel and simultaneously in both squids the shift obtained for each eigenenergy is approximately linear in the bias flux @xmath53 .
this indicates that the coupling can be primarily represented as a first order perturbation in @xmath53 .
a comparison of the fig.1 with the calculations shown in fig.4 indicates that the coupled rf - squid model is in essence an isolated four - level quantum system .
recently , schrdinger cat states have been obtained experimentally by manipulating these two couplings in the mqt regime.@xcite for the physically relevant range @xmath54 the centers of the splitting energies in fig.4 also respect the manifest conservation principle of the energy centers in fig.1 . in fact , in the numerical calculations we found a shift smaller than @xmath55 in the energy centers for the whole range @xmath56 . in this paper we will refer to two different representations of the basis states for two coupled squids .
the first , which is the even and the odd states of the mutually noninteracting squids , is represented by the vector @xmath57 as in section a. here in the state @xmath58 @xmath59 represents the parity state of the squid a and @xmath60 represents that of squid b. we will designate @xmath61 as the _ two - bit parity basis_. the second , which is the set of flux states for the coupled and tilted squids , i.e. both couplings are nonzero , is represented by @xmath62 which we refer to as the _ two - bit flux basis_. according to the hamiltonian ( [ prl.1 ] ) , the eigen levels in fig.1b corresponding to @xmath63 and @xmath64 are the doubly degenerate states @xmath65 . explicitly these are , @xmath66 and @xmath67 for @xmath68 , and , @xmath69 and @xmath70 for @xmath71 . if @xmath72 is also turned on as in fig.1c , the non - degenerate energy eigenstates become the two - bit flux states defined in eq.([updownst.1 ] ) and written explicitly as @xmath73 the choice of the basis in which the fundamental quantum gates are to be operated should be determined by the accessibility of the input and output states by the measurement mechanism .
this condition actually means that all information in the relative magnitudes and the phases of these states should be measurable .
the absolute values of the amplitudes of the flux states are directly measurable whereas their relative phase factors do not couple to the measurement ( see below ) . since it is crucial to be able to read and write the relative phase between the flux states , we choose the two - bit parity basis @xmath74 instead of the two - bit flux basis @xmath75 to perform the gate operations .
the flux basis is used for the i / o operations . before and after performing the logic
gates it is therefore needed to switch between the @xmath61 and the @xmath76 bases .
notice that the matrices @xmath10 and @xmath12 are two - bit hadamard transformations .
in particular , it can be verified that @xmath77 where @xmath78 is the @xmath79 unit matrix .
the matrix @xmath80 is the one bit hadamard transformation . the transformation from the natural eigenbasis @xmath76 to the parity basis @xmath61
can therefore be performed by two successive two - bit hadamard transformations . in this way
any phase information of a quantum state encoded in @xmath61 basis contributes to the relative amplitudes of the @xmath76 basis yielding direct accesibility to the measurement .
this transformation can be summarized as @xmath81 .
the effective result of this transformation is that each state in the parity basis is finaly assigned a distinct eigen energy as shown below .
this transformation is equivalent to effectively assigning each component @xmath6 of the @xmath61 basis an energy eigenvalue .
the logic operations are then performed on the @xmath61 basis .
we now demonstrate the fundamental single - qubit operations , i.e. the phase flip and the hadamard transformation .
_ single qubit operations : _ although the mechanism is based on four states , it is possible to identify two individual flux qubits and perform single bit operations on each one independently . in the parity basis ,
@xmath82 we assign the first bit as the control ( qubit no.1 ) and the second bit as the target ( qubit no.2 ) . by definition
, the single qubit operations are performed on the target in the four dimensional state space unconditionally from the state of the control bit . whereas the controlled operations ( two - qubit operations ) are performed when the control bit is in a particular , i.e ( @xmath83 or @xmath84 ) state . with this convention
we now start with the single qubit operations .
* i. * the phase flip is a special case of the phase evolution described by @xmath85 the phase flip is obtained in ( [ z.1a ] ) at @xmath86 which implies that the state of the second bit is changed as @xmath87 and @xmath88 independent from the state of the first bit .
it is shown below that this can be obtained as a direct sum of two controlled phase flips . the phase evolution @xmath89 in eq.([z.1a ] )
is achieved in the two coupled squid realization in two steps .
we switch the coupling configuration of the system to @xmath90 which makes the states @xmath91 and @xmath92 degenerate at the major center @xmath0 . by this switching , the remaining ( nondegenerate ) states ,
i.e. @xmath93 and @xmath94 are now at the energies @xmath95 and @xmath96 respectively . the four state system
is then permitted to freely evolve in this configuration .
degenerate states do not acquire any relative phase with respect to each other . at this first stage the phases acquired by the states
are described by @xmath97 where @xmath98 as the time dependent phase . in the second stage we set @xmath99 which turns the states @xmath100 and @xmath101 to become degenerate at the energy @xmath0 and @xmath91 and @xmath92 nondegenerate at energies @xmath95 and @xmath96 respectively .
an identical phase evolution is then performed on the states @xmath91 and @xmath102 .
the net transformation is given by @xmath103 which , after factoring out the overall phase @xmath104 is equivalent to eq.([z.1a ] ) .
the phase flip is obtained at @xmath86 corresponding to @xmath105 which can be described as @xmath106\ , ( { \cal c}_2 \to -{\cal c}_1 ) \nonumber \\ & \times & \bigl[-i\,e^{-i\,e_0\,t_1}\,e^{i{\bf h}_{{\cal c}_1,{\cal c}_1}\,t_1 } \bigr ] ( { \cal c}_1 \to { \cal c}_1)~. \label{z.3}\end{aligned}\ ] ] where the action of the operators ( matrices ) is defined rightwards .
notice that , eq.([z.3 ] ) is manifestly the composition of two controlled phase flips where these two operations commute . subtracting the overall phase , the first square bracket is equivalent to [ @xmath107 . this is the controlled phase flip @xmath87 and @xmath88 when the control bit is in the state @xmath84 .
similarly , the second bracket denotes the same controlled phase flip when the control bit is in the @xmath83 state .
hence , this method proves an efficient realization of all controlled operations as well ( see the double qubit operations below ) . _
double qubit operations _ : the controlled phase flip was considered previously .
it basically comprises the first or the second half of the @xmath121 operation in eq.([z.3 ] ) .
we will therefore not consider it here again .
the other crucial double qubit operation is the controlled - not gate . _
controlled not _ ideally , the gate can be obtained from and by@xcite @xmath122 it is much easier to obtain in this proposed mechanism , which can be done by a single controlled rotation by @xmath123 .
this can be written as @xmath124\ , \,r_x(\pi/2)\ , [ { \cal c}_2 \to { \cal c}_2 ] \label{cz.4}~.\ ] ] if one ignores the overall phase , eq.([cz.4 ] ) amounts to setting the coupling so that @xmath90 , applying the rf - field to induce the rotation @xmath125 , and finally switching back to the initial configuration of the @xmath1 and @xmath2 . the full matrix in the @xmath61 basis is @xmath126 which amounts to the flip of the second bit @xmath127 and @xmath128 when the first bit is in the @xmath84 state .
the time scale required for the gate operations can be rougly considered to be independent from the particular gate operation .
we consider the operational time as @xmath129 where @xmath130 describes the energy difference of the states at the configuration where the gate operation is performed .
for instance , a typical scale is the phase flip @xmath131 can be read off from fig.4 at the degeneracy point @xmath132 .
we find that @xmath133 which yields @xmath134 .
the switching time is possible for less than @xmath135 between two gate operations .
@xcite by comparing these two time scales , we estimate that the gate operations are dictated by the switching time .
as the dephasing mechanisms are concerned certain expected sources and their decoherence time scales have been previously analyzed in detail.@xcite squids in the flux regime are known to be insensitive to the charge fluctuations .
the decoherence time for the charge fluctuations in the flux dominated regime has been estimated@xcite to be @xmath136 . among the other sources
there are , the quasiparticle tunneling ( @xmath137 ) and electromagnetic radiation of the junction@xcite ( @xmath138 ) .
more substantial dephasing effects are known to arise from the near distance dipole - dipole interactions between the squids ( @xmath139 ) .
the relaxation time in the time dependence of the basis ( flux or parity ) states is the most crucial element for the i / o operations based on the density matrix measurements . for the parameters considered here
it can be calculated by the explicit expression given by leggett et al . and weiss ref.[11 ] for which we find @xmath140 .
the state of the coupled squids is observed before and after the computation by the flux measurement made by two dc - squids each independently facing one squid loop as shown in fig.2 . before any read - in or read - out made
, a back transformation is needed between the operational @xmath61 basis to the physical flux basis @xmath75 .
suppose that a particular logic operation is denoted by the matrix @xmath141 in the basis @xmath61 .
the formula @xmath142\,t\ , [ { \cal u}_2{\cal u}_1]^{-1}\,{\bf v } \label{backtrf}\ ] ] describes the effect of @xmath141 in the basis @xmath76 . here
@xmath141 is any desired composition of the single and double qubit operations .
suppose @xmath143 . using ( [ backtrf ] ) we find that the flux states are transformed by , @xmath144 we give a second example from @xmath145 . applying ( [ backtrf ] ) to this case we obtain @xmath146 the unique correspondence between the flux states and the parity basis permits
read - in and read - out operations to be performed by measuring the flux as shown in fig.2 .
the flux measurements simulate single event quantum mechanics .
suppose an output state to be measured is @xmath147 a single flux measurement yields one of the flux states in the output with the probability @xmath148 associated with that flux state .
the partial amplitudes of the flux states can therefore be inferred only after repeated measurements . in this context , the final measurement is similar to the quantum computation model using ion traps .
we demonstrated a new theoretical mechanism for quantum logic gate operations by manipulating degeneracy of a symmetric four level system by two variable coupling parameters .
a realization of this system using rf - squids in the flux regime is made in which only one of the couplings is variable .
physical examples in which both couplings vary are very likely to exist but currently unknown to us . in that case
an ideal realization of the proposed mechanism is possible with the particularly added feature of fig.1a namely storing the computed information by suppressing the energetic interference between the states .
the realization of the gates for the fundamental single and double qubit operations is also demonstrated . as the scalability of this squid realization
is concerned we have less to say at this moment . masking and evaporation techniques that are already well established can be used to fabricate arrays of the qubit pairs .
the coupling between different pairs can be realized in a similar way as those that are coupled as members of a single pair .
we continue our current effort in the direction of simulating two and three such pairs coupled inductively . |
the evolution of many complex systems in natural , economical , and social sciences is usually presented in the form of time series . in order to analyze time series , several of statistical measures have been introduced in the literature .
these include such concepts as probability distributions @xcite , autocorrelations @xcite , multi - fractals @xcite , complexity @xcite , or entropy densities @xcite .
recently , it has been pointed out that the _ transfer entropy _ ( te ) is a very useful instrument in quantifying statistical coherence between time evolving statistical systems @xcite . in particular , in schreiber s paper @xcite it was demonstrated that te is especially expedient when global properties of time series are analyzed .
prominent applications are in multivariate analysis of time series , including e.g. , study of multichannel physiological data or bivariate analysis of historical stock exchange indices .
methods based on te have substantial computational advantages which are particularly important in analyzing a large amount of data . in all past works , including @xcite , the emphasis has been on various generalizations of transfer entropies that were firmly rooted in the framework of shannon s information theory .
these , so called shannonian transfer entropies are , indeed , natural candidates due to their ability to quantify in a non - parametric and in explicitly non - symmetric way the flow of information between two time series .
an ideal testing ground for various te concepts are financial - market time series because of the immense amount of electronically recorded financial data .
recently , economy has become an active research area for physicists .
they have investigated stock markets using statistical - physics methods , such as the percolation theory , multifractals , spin - glass models , information theory , complex networks , path integrals , etc .. in this connection the name econophysics has been coined to denote this new hybrid field on the border between statistical physic and ( quantitative ) finance . in the framework of econophysics
it has became steadily evident that the market interactions are highly nonlinear , unstable , and long - ranged .
it has also became apparent that all agents ( e.g. , companies ) involved in a given stock market exhibit interconnectedness and correlations which represent important internal force of the market .
typically one uses correlation functions to study the internal cross - correlations between various market activities .
the correlation functions have , however , at least two limitations : first , they measure only linear relations , although it is clear that linear models do not faithfully reflect real market interactions .
second , all they determine is whether two time series ( e.g. , two stock - index series ) have correlated movement .
they , however , do not indicate which series affects which , or in other words , they do not provide any directional information about cause and effect .
some authors use such concepts as time - delayed correlation or time - delayed mutual information in order to construct asymmetric `` correlation '' matrices with inherent directionality .
this procedure is in many respects _ ad hoc _ as it does not provide any natural measure ( or quantifier ) of the information flow between involved series . in the present paper we study multivariate properties of stock - index time with the help of econophysics paradigm . in order to quantify the information flow between two or more stock indices we generalize schreibers shannonian transfer entropy to rnyi s information setting . with this
we demonstrate that the corresponding new transfer entropy provides more detailed information concerning the excess ( or lack ) of information in various parts of the underlying distribution resulting from updating the distribution on the condition that a second time series is known .
this is particularly relevant in the context of financial time series where the knowledge of tale - part ( or marginal ) events such as spikes or sudden jumps bears direct implications , e.g. , in various risk - reducing formulas in portfolio theory .
the paper is organized as follows : in section [ sec2 ] we provide some information - theoretic background on shannon and rnyi entropies ( re s ) . in particular , we identify the _ conditional _ rnyi s entropy with the information measure introduced in ref .
apart from satisfying the chain rule ( i.e. , rule of additivity of information ) the latter has many desirable properties that are to be expected from a conditional information measure .
another key concept the _ mutual _ rnyi entropy , is then introduced in a close analogy with shannon s case .
the ensuing properties are also discussed .
transfer _ entropy of schreiber is briefly reviewed in section [ sec4 ] .
there we also comment on effective transfer entropy of marchinski _
et all_. the core quantity of this work the rnyian _ transfer _ entropy ( rte ) , is motivated and derived in section [ sec7 ] .
in contrast to shannonian case , the rnyian transfer entropy is generally not positive semi - definite .
this is because re non - linearly emphasizes different parts of involved probability density functions ( pdf s ) . with the help of campbell s coding theorem we show that the rte rates a gain / loss in risk involved in a next - time - step behavior in a given stochastic process , say @xmath2 , resulting from learning a new information , namely historical behavior of another ( generally cross - correlated ) process , say @xmath3 . in this view
the rte can serve as a convenient _ rating _ factor of a _ riskiness _ in inter - connected markets .
we also show that rnyian transfer entropy allows to amend spurious effects caused by a finite size of a real data set which in shannon s context must be , otherwise , solved by means of the surrogate data technique and ensuing effective transfer entropy . in section [ sec10 ] we demonstrate the usefulness and formal consistency of rte by analyzing cross - correlations in various international stock markets . on a qualitative level we use
@xmath4 simultaneously recorded data points of the eleven stock exchange indices , sampled at a daily ( end - of - trading day ) rate to construct the _ heat maps _ and _ net flows _ for both shannon s and rnyi s information flows . on a quantitative level
we explicitly discuss time series from the dax and s&p500 market indices gathered on a minute - tick basis in the period from december 1990 till november 2009 in the german stock exchange market ( deutsche brse ) . presented calculations of rnyi and shannon transfer entropies
are based on symbolic coding computation with the open source software @xmath5 .
our numerical results imply that rte s among world markets are typically very asymmetric .
for instance , we show that there is a strong surplus of an information flow from the asia - pacific region to both europe and the u.s .
a surplus of the information flow can be also observed to exists from europe to the u.s . in this last case
the substantial volume of transferred information comes from tail - part ( i.e. , risky part ) of underlying asset distributions .
so , despite the fact that the u.s . contributes more than half of the world trading volume , this is not so with information flow .
further salient issues , such as dependence of rte on rnyi s @xmath0 parameter or on the data block length are numerically also investigated . in this connection
we find that the cross - correlation between dax and s&p500 has a long - time memory which is around 200 - 300 mins .
this should be contrasted with typical memory of stock returns which are of the order of seconds or maximally few minutes .
various remarks and generalizations are proposed in the concluding section [ sec12 ] . for reader
s convenience we give in appendix a a brief dictionary of market indices used in the main text and in appendix b we tabulate an explicit values of effective transfer entropies used in the construction of heat maps and net information flows .
in order to express numerically an amount of information that is shared or transferred between various data sets ( e.g. , two or more random processes ) , one commonly resorts to information theory and especially to the concept of entropy . in this section
we briefly review some essentials of shannon s and rnyi s entropy that will be needed in following sections .
the entropy concept was originally introduced by clausius @xcite in the framework of thermodynamics . by analyzing a carnot engine he was able to identify a new state function which never decreases in isolated systems .
the microphysical origin of clausius phenomenological entropy was clarified more than @xmath6 years later in works of boltzman and ( yet later ) gibbs who associated clausius entropy with the number of allowed microscopic states compatible with a given observed macrostate .
the ensuing _ boltzmann gibbs entropy _ reads @xmath7 where @xmath8 is boltzmann s constant , @xmath2 is the set of all accessible microstates compatible with whatever macroscopic observable ( state variable ) one controls and @xmath9 denotes the number of microstates .
it should be said that the passage from boltzmann gibbs to clausius entropy is established only when the conditional extremum @xmath10 of @xmath11 subject to the constraints imposed by observed state variables is inserted back into @xmath11 .
only when this _ maximal entropy prescription _ @xcite is utilized @xmath11 turns out to be a thermodynamic state function and not mere functional on a probability space . in information theory , on the other hand , the interest was in an optimal coding of a given source data . by _
optimal code _ is meant the shortest averaged code from which one can uniquely decode the source data .
optimality of coding was solved by shannon in his 1948 seminal paper @xcite . according to shannon s _
source coding theorem _
@xcite , the quantity @xmath12 corresponds to the averaged number of bits needed to optimally encode ( or zip " ) the source dataset @xmath2 with the source probability distribution @xmath13 . on a quantitative level ( [ ii.2.a ] )
represents ( in bits ) the minimal number of binary ( yes / no ) questions that brings us from our present state of knowledge about the system @xmath2 to the one of certainty @xcite .
it should be stressed that in shannon s formulation @xmath2 represents a discrete set ( e.g. , processes with discrete time ) , and this will be also the case here .
apart from the foregoing _ operational _ definitions , eq .
( [ ii.2.a ] ) has also several axiomatic underpinnings .
axiomatic approaches were advanced by shannon @xcite , khinchin @xcite , fadeev @xcite an others @xcite . the quantity ( [ ii.2.a ] )
has became known as shannon s entropy ( se ) .
there is an intimate connection between boltzmann
gibbs entropy and shannon s entropy .
in fact , thermodynamics can be viewed as a specific application of shannon s information theory : the thermodynamic entropy may be interpreted ( when rescaled to `` bit '' units ) as the amount of shannon information needed to define the detailed microscopic state of the system , that remains uncommunicated " by a description that is solely in terms of thermodynamic state variables @xcite . among important properties of se is its concavity in @xmath14 , i.e. for any pair of distributions @xmath14 and @xmath15 , and a real number @xmath16 holds @xmath17 eq . ( [ i.19.a ] ) follows from jensen s inequality and a convexity of @xmath18 for @xmath19 .
concavity is an important concept since it ensures that any maximizer found by the methods of the differential calculus yields an absolute maximum rather than a relative maximum or minimum or saddle point . at the same time it is just a sufficient ( i.e. , not necessary ) condition guarantying a unique maximizer .
it is often customary to denote se of the source @xmath2 as @xmath20 rather than @xmath21 .
note that se is generally not convex in @xmath2 !
it should be stressed that the entropy ( [ ii.2.a ] ) really represents a self - information : the information yielded by a random process about itself .
a step further from a self - information offers the _
joint entropy _ of two random variables @xmath2 and @xmath3 which is defined as @xmath22 and which represents the amount of information gained by observing jointly two ( generally dependent or correlated ) statistical events .
a further concept that will be needed here is the _ conditional entropy _ of @xmath2 given @xmath3 , which can be motivated as follows : let us have two statistical events @xmath2 and @xmath3 and let event @xmath3 has a sharp value @xmath23 , then the gain of information obtained by observing @xmath2 is @xmath24 here the conditional probability @xmath25 . for general random @xmath3 one defines the conditional entropy as the averaged shannon entropy yielded by @xmath2 under the assumption that the value of @xmath3 is known , i.e. @xmath26 from ( [ ii.a.5a ] ) , in particular , follows that @xmath27 identity ( [ ii.2.6.a ] ) is known as additivity ( or chain ) rule for shannon s entropy . in statistical thermodynamics
this rule allows to explain , e.g. , gibbs paradox .
applying eq .
( [ ii.2.6.a ] ) iteratively , we obtain : @xmath28 another relevant quantity that will be needed is the _ mutual information _ between @xmath2 and @xmath3 .
this is defined as : @xmath29 and can be equivalently written as @xmath30 this shows that the mutual information measures the average reduction in uncertainty ( i.e. , gain in information ) about @xmath2 resulting from observation of @xmath3 .
of course , the amount of information contained in @xmath2 about itself is just the shannon entropy : @xmath31 notice also that from eq .
( [ eq : mi ] ) follows @xmath32 and so @xmath2 provides the same amount of information on @xmath3 as @xmath3 does on @xmath2 . for this reasons
the mutual information is not a useful measure to quantify a flow of information .
in fact , the flow of information should be by its very definition directional . in the following we will also find useful the concept of _ conditional mutual entropy _ between @xmath2 and @xmath3 given @xmath33 which is defined as
@xmath34 the latter quantifies the averaged mutual information between @xmath2 and @xmath3 provided that @xmath33 is known .
applying ( [ eq : condmii ] ) and ( [ eq : mi2 ] ) iteratively we may write @xmath35 & = & \ i(x;y_1\cap \cdots
\cap y_n \cap z_1\cap \cdots \cap z_m)\nonumber \\ & - & \ i(x;z_1\cap \cdots
\cap z_m ) \ , .
\label{eq : condmiiii}\end{aligned}\ ] ] for further details on the basic concepts of shannon s information theory , we refer the reader to classical books , e.g. , ash @xcite and , more recently , csiszr and shields @xcite .
rnyi introduced in refs .
@xcite a one - parameter family of information measures presently known as _ rnyi entropies _ @xcite . in practice , however , only a singular name rnyi s entropy is used .
re of order @xmath0 @xmath36 of a distribution @xmath37 on a finite set @xmath38 is defined as @xmath39 for re ( [ renyi ] ) one can also formulate source coding theorem .
while in the shannon case the cost of a code - word is a linear function of the length so the optimal code has a minimal cost out of all codes , in the rnyi case the cost of a code - word is an exponential function of its length @xcite .
this is , in a nutshell , an essence of the so - called campbell s coding theorem ( cct ) . according to this re corresponds to the averaged number of bits needed to optimally encode the discrete source @xmath2 with the probability @xmath13 , provided that the codeword - lengths are exponentially weighted @xcite . from the form ( [ renyi ] ) one can easily see that for @xmath40 re depends more on the probabilities of the more probable values and less on the improbable ones .
this dependence is more pronounced for higher @xmath0 . on the other hand , for @xmath41 marginal events
are accentuated with decreasing @xmath0 . in this connection
we should also point out that campbell s coding theorem for re is equivalent to shannon s coding theorem for se provided one uses instead of @xmath42 the _ escort distribution _
@xcite : @xmath43 the pdf @xmath44 was first introduced by rnyi @xcite and in the physical context brought by beck , schlgl , kadanoff and others ( see , e.g. , refs .
@xcite ) . note ( cf .
[ fig2 ] ) for event probability @xmath45 and varying rnyi s parameter @xmath0 .
arrows indicate decreasing values of @xmath0 for @xmath41 ( dark arrow ) or increasing values of @xmath0 for @xmath40 ( lighter arrow).,width=377 ] ( 20,7 ) ( 105,65 ) @xmath45 ( -135,185)@xmath46 that for @xmath40 the escort distribution emphasizes the more probable events and de - emphasizes more improbable ones .
this trend is more pronounced for higher values of @xmath0 .
for @xmath41 the escort distribution accentuates more improbable ( i.e. , marginal or rare ) events .
this dependence is more pronounced for decreasing @xmath0 . this fact is clearly seen on fig . [ fig1 ] . :
@xmath47 . , width=302 ] ( 20,7 ) ( 90,80 ) @xmath0 ( -35,55 ) @xmath45 ( 125,155)@xmath48 so by choosing different @xmath0 we can scan " or probe " different parts of the involved pdf s .
it should be stressed that apart from cct , re has yet further operational definitions , e.g. , in the theory of guessing @xcite , in the buffer overflow problem @xcite or in the theory of error block coding @xcite .
re is also underpinned with various axiomatics @xcite . in particular
, it satisfies identical khinchin axioms @xcite as shannon s entropy save for the additivity axiom ( chain rule ) @xcite : @xmath49 where the conditional entropy @xmath50 is defined with the help of the escort distribution ( [ ii.15.a ] ) ( see , e.g. , refs .
@xcite ) . for @xmath51 re reduces to the shannon entropy : @xmath52 as one can easily verify with lhospital s rule .
we define the _ joint rnyi entropy _ ( or the _ joint entropy _ of order @xmath0 ) for two random variables @xmath38 and @xmath53 in a natural way as : @xmath54 the _ conditional entropy _ of order @xmath0 of @xmath2 given @xmath3 is similarly as in the shannon case defined as the averaged rnyi s entropy yielded by @xmath2 under the assumption that the value of @xmath3 is known . as shown in refs .
@xcite this has the form @xmath55 & = & \
\frac{1}{1-q}\log_2 \frac{\sum_{x\ \ !
\in \ \ ! { x } , \ \ ! y \ \ !
{ y}}p^{\,q}(x , y)}{\sum_{y\ \ !
\in \ \ ! { y } } q^{\,q}(y)}\ , .
\label{re.condentropy}\end{aligned}\ ] ] in this connection it should be mentioned that several alternative definitions of the conditional re exist ( see , e.g. , refs .
@xcite ) , but the formulation ( [ re.condentropy ] ) differs from other versions in a few important ways that will show up to be desirable in the following considerations .
the conditional entropy defined in ( [ re.condentropy ] ) has the following important properties , namely @xcite : : @xmath56 , where @xmath57 is a number of elements in @xmath2 , : : @xmath58 only when @xmath3 uniquely determines @xmath2 ( i.e. , no gain in information ) , : : @xmath59 , : : when @xmath2 and @xmath3 are independent then @xmath60 . unlike shannon s case one can
not , however , deduce that the equality @xmath61 implies independency between event @xmath2 and @xmath3 .
also the inequality @xmath62 ( i.e. , an extra knowledge about @xmath3 lessens our ignorance about @xmath2 ) does not hold here in general @xcite .
the latter two properties may seem as a serious flaw .
we will now argue that this is not the case and , in fact , it is even desirable .
first , in order to understand why @xmath61 does not imply independency between @xmath2 and @xmath3 we define the information - distribution function @xmath63 which represents the total probability caused by events with information content @xmath64 . with this
we have @xmath65 and thus @xmath66 taking the inverse laplace transform with the help of the so - called _
post s inversion formula _
@xcite we obtain @xmath67\right|_{q \ = \
k/(x \ln 2 ) + 1 } .
\label{iii.b.20a}\end{aligned}\ ] ] analogous relation holds also for @xmath68 and associated @xmath50 . as a result
we see that when working with @xmath69 of different orders we receive much more information on underlying distribution than when we restrict our investigation to only one @xmath0 ( e.g. , to only shannon s entropy ) .
in addition , eq . ( [ iii.b.20a ] ) indicates that we need all @xmath70 ( or equivalently all @xmath71 , see @xcite ) in order to uniquely identify the underlying pdf . in view of eq .
( [ iii.b.20a ] ) we see that the equality between @xmath50 and @xmath72 at some neighborhood of @xmath0 merely implies that @xmath73 for some @xmath74 .
this naturally does not ensure independency between @xmath2 and @xmath3 .
we need equality @xmath75 for all @xmath40 ( or for all @xmath71 ) in order to secure that @xmath73 holds for all @xmath74 which would in turn guarantee that @xmath76 .
therefore , all re with @xmath40 ( or all with @xmath71 ) are generally required to deduce from @xmath77 an independency between @xmath2 and @xmath3 . in order to understand the meaning of the inequality @xmath78 we first introduce the concept of mutual information .
the _ mutual information of order _
@xmath0 between @xmath2 and @xmath3 can be defined as ( cf .
( [ eq : mi2 ] ) ) @xmath79 & = & \ s_q^{(r)}(x)\ + \
s_q^{(r)}(y ) \ - \ s_q^{(r)}(x\cap y ) \
, , \label{re.mutinf}\end{aligned}\ ] ] which explicitly reads @xmath80 & = & \ \frac{1}{1-q}\log_2 \frac{\sum_{x\ \ !
\in \ \ ! { x } , \ \ ! y \ \ !
\in \ \ ! { y } } q^{\,q}(y ) p^{\,q}(x)}{\sum_{x\ \ ! \in \ \ ! { x } , \ \ !
\in \ \ ! { y}}q^{\,q}(y)p^{\,q}(x|y ) } \ , .
\label{ii.20a}\end{aligned}\ ] ] note that we have again the symmetry relation @xmath81 as well as the consistency condition @xmath82 .
so similarly as in the shannon case , rnyi s mutual information formally quantifies the average reduction in uncertainty ( i.e. , gain in information ) about @xmath2 that results from learning the value of @xmath3 , or vice versa . from eq .
( [ re.mutinf ] ) we see that the inequality in question , i.e. , @xmath83 implies @xmath84 .
according to ( [ ii.20a ] ) this can be violated only when @xmath85 & & \sum_{x\ \ !
\in \ \ ! { x } }
p^{\,q}(x ) \ < \sum_{x\ \ ! \in \ \ ! { x } } \langle { \mathcal{p}}^{\,q}(x|y)\rangle_q \;\;\;\;\;\ ; \mbox{for } \;\;\;\ ; 0 < q < 1\ , . \label{2.b.26a}\end{aligned}\ ] ] here @xmath86 is an average with respect to the escort distribution @xmath87 ( see eq .
( [ ii.15.a ] ) ) . by taking into account properties of the escort distribution
, we can deduce that @xmath88 when a larger probability events of @xmath2 receive by learning @xmath3 a lower value .
as for the marginal events of @xmath2 , these are by learning @xmath3 indeed enhanced , but the enhancement rate is smaller than the suppression rate of large probabilities . for instance , this happens when @xmath89 for @xmath90 the inequality ( [ 2.b.28a ] ) ensures that @xmath91 holds . the left inequality in ( [ 2.b.28a ] )
saturates when @xmath92 , see also fig .
[ fig2a ] . for @xmath41
is the situation analogous . here
properties of the escort distribution imply that @xmath93 when marginal events of @xmath2 get by learning @xmath3 a higher probability .
the suppression rate for large ( i.e. close - to - peak ) probabilities is now smaller than the enhancement rate of marginal events .
this happens , for example , for distributions , @xmath94 with @xmath95 fulfilling again the inequality ( [ 2.b.28a ] ) .
this can be also directly seen from fig .
[ fig2a ] when we revert the sign of @xmath96 .
when we set @xmath97 then both inequalities ( [ 2.b.26a ] ) are simultaneously satisfied yielding @xmath98 as it should .
in contrast to a shannonian case where the mutual information quantifies the average reduction in uncertainty resulting from observing / learning a further information , in the rnyi case we should use campbell s coding theorem in order to properly understand the meaning of @xmath96 .
( 20,7 ) ( 118,235)@xmath0 ( -100,85)@xmath99 according to the cct @xmath72 corresponds to the minimal average cost of a coded message with a non - linear ( exponential ) weighting / pricing of codeword - lengths . while according to shannon we never increase ignorance by learning @xmath3 ( i.e. , possible correlations between @xmath2 and @xmath3 can only reduce the entropy ) , in rnyi s setting extra knowledge about @xmath3 might easily increase the minimal price of coding @xmath2 because of the nonlinear pricing . since the cct penalizes long codewords which in shannon s coding have low probability , the price of the @xmath100 code may easily increase , as we have seen in examples ( [ ii.20aaa ] ) and ( [ ii.20aaaa ] ) . in the key context of financial time
series , the risk valuation of large changes such as spikes or sudden jumps is of a crucial importance , e.g. , in various risk - reducing formulas in portfolio theory .
the rle of campbell s pricing can in these cases be interpreted as a risk - rating method which puts an exponential premium on rare ( i.e. , risky ) asset fluctuations . from this point of view the mutual information @xmath99 represents a _ rating factor _ which rates a gain / loss in risk in @xmath2 resulting from learning a new information , namely information about @xmath3 .
the _ conditional mutual information _ of order @xmath0 between @xmath2 and @xmath3 given _ z _ is defined as @xmath101 note that because of a validity of the chain rule ( [ iii.b.15a ] ) , relations ( [ ii.a.8a ] ) and ( [ eq : condmiiii ] ) also hold true for the re . to close this section
, we shall stress that information entropies are primarily important because there are various coding theorems which endow them with an operational ( that is , experimental ) meaning , and not because of intuitively pleasing aspects of their definitions . while coding theorems do exist both for the shannon entropy and the rnyi entropy
there are ( as yet ) no such theorems for tsallis , kaniadakis , naudts and other currently popular entropies .
the information - theoretic significance of such entropies is thus not obvious . since the information - theoretic aspect of entropies is of a crucial importance here , we will in the following focus only on the se and the re .
as seen in section [ sec2b ] , the mutual information @xmath102 quantifies the decrease of uncertainty about @xmath2 caused by the knowledge of @xmath3 .
one could be thus tempted to use it as a measure of an informational transfer in general complex systems . a major problem , however , is that shannon s mutual information contains no inherent directionality since @xmath103 .
some early attempts tried to resolve this complication by artificially introducing the directionality via time - lagged random variables . in this way
one may define , for instance , the _ time - lagged mutual _ ( or _ directed kullback leibler _ ) _ information _ as @xmath104 the later describes the average gain of information when replacing the product probability @xmath105 by the joint probability @xmath106 .
so the information gained is due to cross - correlation effect between random variables @xmath107 and @xmath108 ( respectively , @xmath109 ) .
it was , however , pointed out in ref .
@xcite that prescriptions such as ( [ iii.a.1.a ] ) , though directional , also take into account some part of the information that is statically shared between the two random processes @xmath2 and @xmath3 . in other words , these prescriptions do not produce statistical dependences that truly originate only in the stochastic random process @xmath3 , but they do include the effects of a common history ( such as , for example , in the case of a common external driving force ) .
for this reason , schreiber introduced in ref .
@xcite the concept of ( shannonian ) transfer entropy ( ste ) .
the latter , apart from directionality , accounts only for the cross - correlations between statistical time series @xmath2 and @xmath3 whose genuine origin is in the source " process @xmath3 .
the essence of the approach is the following .
let us have two time sequences described by stochastic random variables @xmath107 and @xmath108 .
let us assume further that the time steps ( data ticks ) are discrete with the size of an elementary time lag @xmath110 and with @xmath111 ( @xmath112 is a reference time ) .
the transfer entropy @xmath113 can then be defined as @xmath114 & = & i(x_{t_{m+1 } } ; x_{t_1}\cap \cdots \cap
x_{t_{m}}\cap y_{t_{m - l+1}}\cap \cdots \cap y_{t_{m } } ) - i(x_{t_{m+1}};x_{t_1 } \cap \cdots \cap
x_{t_{m } } ) \ , .\nonumber \\ & & \mbox{\hspace{-15 mm } } \label{iii.a.23a}\end{aligned}\ ] ] the last line of ( [ iii.a.23a ] ) indicates that @xmath113 represents the following .
[ cols= " < , < " , ]
in this appendix we specify explicit values of effective transfer entropies that are employed in section [ sec10 ] .
these are calculated for alphabet with @xmath115 . j .- s .
yang , w. kwak , t. kaizoji and i .-
kim , _ eur .
j. _ * b61 * ( 2008 ) 389 ; k. matal , m. pal , h. salunkay and h.e .
stanley , _ europhys . lett .
_ * 66 * ( 2004 ) 909 ; h.e .
stanley , l.a.n .
amaral , x. gabaix , p. gopikrishnan and v. plerou , _
physica _ * a299 * ( 2001 ) 1 .
chaudy and j.b .
edinburgh mathematical notes _ * 43 * ( 1960 ) 7 ; for h. tverberg s , p.m. lee s or d.g .
kendall s axiomatics of shannon s entropy see e.g. , s. guias , _ information theory with applications _ ,
( mcgraw hill , new york , 1977 ) .
this exponential weighting is also known as a kolmogorov
nagumo averaging . while the linear averaging is given by @xmath118 , the exponential weighting ia defined as @xmath119 with @xmath120 .
the @xmath121 factor is known as campbell exponent .
one can map @xmath69 with @xmath122 to @xmath69 with @xmath123 via duality @xmath124 that exists between @xmath48 and @xmath37 .
in fact , we can observe that @xmath125 and @xmath126 .
so @xmath69 with @xmath122 and @xmath69 with @xmath127 carry equal amount of information .
y. liu , p. cizeau , m. meyer , c .- k .
peng , h.e .
stanley , _ physica _ * a245 * ( 1997 ) 437 ; _ physica _ * a245 * 441 ; y. liu , p. gopikrishnan , p. cizeau , m. mayer , c .- k .
peng , h.e .
stanley , _ phys .
* e60 * ( 1999 ) 1390 . | in this paper , we quantify the statistical coherence between financial time series by means of the rnyi entropy . with the help of campbell s coding theorem
we show that the rnyi entropy selectively emphasizes only certain sectors of the underlying empirical distribution while strongly suppressing others .
this accentuation is controlled with rnyi s parameter @xmath0 . to tackle the issue of the information flow between time series we formulate the concept of rnyi s transfer entropy as a measure of information that is transferred only between certain parts of underlying distributions .
this is particularly pertinent in financial time series where the knowledge of marginal events such as spikes or sudden jumps is of a crucial importance .
we apply the rnyian information flow to stock market time series from @xmath1 world stock indices as sampled at a daily rate in the time period 02.01.1990 - 31.12.2009 . corresponding _ heat maps _ and _ net information flows _ are represented graphically .
a detailed discussion of the transfer entropy between the dax and s&p500 indices based on minute tick data gathered in the period from 02.04.2008 to 11.09.2009 is also provided .
our analysis shows that the bivariate information flow between world markets is strongly asymmetric with a distinct information surplus flowing from the asia pacific region to both european and us markets .
an important yet less dramatic excess of information also flows from europe to the us .
this is particularly clearly seen from a careful analysis of rnyi information flow between the dax and s&p500 indices .
pacs numbers : 89.65.gh , 89.70.cf , 02.50.-r keywords : econophysics ; rnyi entropy ; information transfer ; financial time series # 1 # 1 |
includes the proofs of the propositions , and the analytical calculation of the msevalues for the considered examples .
let us assume that @xmath26 and @xmath41 are a msevalue and a msevector of @xmath12 , respectively .
this means they fulfill @xmath114 for all @xmath35 .
the result of the mapping of @xmath12 acting on a vector @xmath41 can be decomposed in one part parallel to the input state and another part @xmath115 perpendicular to the input state @xmath116 note that @xmath117 .
let us consider an arbitrary @xmath49 ( @xmath35 ) .
we get the following projection for the map @xmath118 : @xmath119 hence , the msevalue equations imply that for the solution holds the form @xmath120 where @xmath115 has the property given in eq . . the other way around
, we get from these equivalent transformations , that eq .
implies that @xmath26 and @xmath41 are a msevalue and a msevector of @xmath12 , respectively .
let us assume that @xmath26 and @xmath41 are a msevalue and a msevector of @xmath12 , respectively .
hence , eq . is fulfilled .
now , we consider an operator @xmath57 ( @xmath53 ) : @xmath121\left(\hat u_1\otimes\dots\otimes \hat u_k\right ) , \end{aligned}\ ] ] where @xmath54 are unitary transformations acting on @xmath122 ( @xmath35 ) .
we get for the transformed input vector @xmath123 where @xmath124 obviously fulfills eq . for @xmath125 and arbitrary @xmath126 . hence , eq . is the second form of the msevalue equation for @xmath57 with the transformed solutions : @xmath127 let us note that an inverse transformation of @xmath57 yields the inverse implication , @xmath128\left(\hat u_1\otimes\dots\otimes \hat u_k\right)^\dagger .
\end{aligned}\ ] ] we consider a rank one operator @xmath59 .
let us consider the schmidt decomposition @xcite of @xmath62 , with respect to a bipartite decomposition @xmath129 and @xmath130 , as @xmath131 with @xmath132 ( schmidt coefficients ) , @xmath133 ( schmidt rank ) , orthonormal @xmath134 , and orthonormal @xmath135 .
the msevalue equation in the first form for the @xmath58-th subsystem reads as @xmath136 with the abbreviation for the msevalue @xmath137 ( @xmath138 ) . here
, we are only interested in solutions with @xmath139 . from eq . , we get that the solution @xmath140 must have the form @xmath141 we may insert this result into the msevalue equation for the @xmath142-th party ( @xmath143 ) : @xmath144 where we used the abbreviation @xmath145 let us note that @xmath12 is a positive semi - definite operator which has a rank @xmath133 .
now , we consider the operator @xmath67 , where @xmath146 is a generalized w - state . the reduced operator with respect to the third subsystem is @xmath147 we may start with the calculation of the maximal msevalue for partial separability in the case @xmath148 .
since @xmath149 and @xmath150 are perpendicular vectors in @xmath151 , we immediately get @xmath152 similarly , this yields the maximal msevalues for the other decompositions : @xmath153 for @xmath154 and @xmath155 for @xmath156 . for the full separable case @xmath157 , we can easily get the obvious solutions of @xmath57 : @xmath158 a convenient parametrization of all other possible msevectors is @xmath159 here , we use the msevalue equations in the second form @xmath160 . since we require that @xmath161 and @xmath162 for all @xmath163 ( @xmath164 ) , we have @xmath165 this leads to the msevalue equation in the second form as @xmath166 where we multiplied the whole equation with the normalization constant @xmath167 .
we may decompose @xmath168 for @xmath164 in polar coordinates .
the individual basis components can be rewritten as @xmath169 where we introduced @xmath170 and @xmath171 ( @xmath164 ) . since @xmath172
, it follows from eq .
, that @xmath173 . eqs . and can be combined to @xmath174 note that this equation has a real solution @xmath175 , iff @xmath176 or , equivalently , @xmath177 .
more generally , this equation has a solution , iff @xmath178 in addition , we may resolve eqs . and as @xmath179
hence , we get quadratic equations and in @xmath26 and @xmath180 .
the difference of both equations gives the relation between @xmath26 and @xmath180 as @xmath181 this relation , we may insert into eq . which results in @xmath182g-|\lambda_3|^2\right ) .
\end{aligned}\ ] ] omitting @xmath183 , we get the root of this equation as @xmath184 finally , we conclude for the case of full separability that the maximal msevalue @xmath185 of @xmath12 is @xmath186 where we have taken the solutions in eq . into account .
it is of importance to mention that from the calculation of the msevector of @xmath187 follows , that we have to fulfill the requirements @xmath188 , @xmath189 , and @xmath190 .
let us consider the operator @xmath84 , with @xmath191 being a generalized ghz - state .
the reduced operator is @xmath192 first , we may consider a decomposition @xmath148 .
since @xmath57 is already given in a spectral decomposition , we get the non - zero msevalues @xmath193 and @xmath194
. similarly , we can argue for the partial separability with respect to @xmath156 and @xmath154 .
we get that the maximal msevalue , @xmath185 , for partial separability is @xmath195 second , we consider full separability @xmath157 . due to the reduction from @xmath12 to @xmath57 , we have to solve the msevalue equations of @xmath57 as given in eq . .
it is easy to check that @xmath196 is an msevector for the msevalue @xmath197 ( @xmath198 ) . for all other msevectors ,
we use the same parametrization as given in eq . .
now , the msevalue equation read as @xmath199 these equations are fulfilled , if @xmath200 for the msevalue @xmath201 , see the components in @xmath202 direction .
due to the fact that this @xmath26 is smaller or equal to @xmath197 , we get the maximal msevalue for full separability as @xmath203 let us consider the hamilton operator @xmath108 of a system of @xmath60 harmonic oscillators .
@xmath204 where we used the position operator vector @xmath205 and momentum operator vector @xmath206 .
the masses of the individual oscillators are given by @xmath207 and the total potential energy reads as @xmath208 , with : @xmath209 ( real valued ) ; @xmath210 ( symmetric ) ; and @xmath211 ( positive definite ) .
the well - known ground state is given by the normalized , gaussian wave - function @xmath212{\frac{\det \boldsymbol v}{\pi^n}}\exp\left[-\frac{1}{2}x^{\rm t}\boldsymbol v x \right],\quad
\text{with } \boldsymbol v=\boldsymbol v^{\rm t}. \end{aligned}\ ] ] in order to get @xmath213 , we need the first and second derivative : @xmath214 the resulting eigenvalue problem is @xmath215\psi_0(x ) + x^{\rm t}\left[-\frac{\hbar^2}{2}\boldsymbol v\boldsymbol m^{-1}\boldsymbol v+\frac{1}{2}\boldsymbol g\right ] x\,\psi_0(x ) = e_0\psi_0(x ) , \end{aligned}\ ] ] where the here used trace operation , @xmath216 , is acting on @xmath217 matrices .
the eigenvalue equation is solved by the ground state @xmath218 for @xmath219^{\frac{1}{2}}\boldsymbol m^{1/2}. \end{aligned}\ ] ] the minimal eigenvalue is @xmath220^{\frac{1}{2}}\right ) .
\end{aligned}\ ] ] as an intermediate step , it is useful to show that a minimal energy in our case is always obtained for states with zero mean values for @xmath221 and @xmath222 .
we have an energy given by @xmath223\right ) + \frac{1}{2}{\rm tr}\!\left(\boldsymbol g\left [ \langle(\delta \hat x)(\delta \hat x)^{\rm t}\rangle+\langle \hat x\rangle\langle \hat x\rangle^{\rm t } \right]\right)\\ = & \frac{1}{2}\langle \hat p\rangle^{\rm t } \boldsymbol m^{-1}\langle \hat p\rangle+\frac{1}{2}\langle \hat x\rangle^{\rm t } \boldsymbol g\langle \hat x\rangle + \frac{1}{2}{\rm tr}\!\left[\boldsymbol m^{-1}\,{\rm covar}(\hat p,\hat p^{\rm t})\right]+\frac{1}{2}{\rm tr}\!\left[\boldsymbol g\,{\rm covar}(\hat x,\hat x^{\rm t})\right ] .
\end{aligned}\ ] ] a local displacement operation gives the smallest value for the case @xmath224 , when considering the same covariance matrices .
now , we may study an arbitrary decomposition of @xmath225 : @xmath226 we define the position and momentum operator for the corresponding subspace as @xmath227 and @xmath228 , respectively .
the block - matrices @xmath229 and @xmath230 are similarily obtained from @xmath231 and @xmath232 , by ignoring all rows or columns which are not in @xmath28 or @xmath32 .
note that @xmath233 for @xmath234 . a product wave function of this decomposition @xmath235 reads in position representation as @xmath236 hence ,
we may write the energy operator as @xmath237 now the msevalue equations read for a particular choice of @xmath238 as @xmath239\psi_{(j)}\left(x_{(j)}\right ) , \end{aligned}\ ] ] together with the partially reduced operator @xmath240 where we used @xmath224 . now
the msevalue equation for the partition @xmath241 is an ordinary eigenvalue problem of a quadratic ( reduced ) hamiltonian : @xmath242\psi_{(j)}\left(x_{(j)}\right)\\ & + \left[\frac{1}{2}\sum_{j\neq j}\langle \hat p^{\rm t}_{(j)}\boldsymbol m_{(j , j)}^{-1 } \hat p_{(j)}\rangle+ \frac{1}{2}\sum_{j\neq j } \langle \hat x_{(j)}^{\rm t}\boldsymbol g_{(j , j ) } \hat x_{(j')}\rangle\right]\psi_{(j)}\left(x_{(j)}\right ) = e_0\left[(\mathcal i_j)_{j=1}^k\right]\psi_{(j)}\left(x_{(j)}\right ) .
\end{aligned}\ ] ] from our initial solution of the ground state of the full hamiltonian , we obtain the ground state of the reduced hamiltonian being the msevector in position representation , @xmath243{\frac{\det \boldsymbol
v_{\!j}}{\pi^{n_{\!j}}}}\exp\left[-\frac{1}{2}x_{(j)}^{\rm t}\boldsymbol v_{\!j } x_{(j ) } \right ] , \quad\text{for } \boldsymbol v_{\!j}=\frac{1}{\hbar}\boldsymbol m_{(j , j)}^{1/2}\left[\boldsymbol m_{(j , j)}^{-1/2}\boldsymbol g_{(j , j)}\boldsymbol m_{(j , j)}^{-1/2}\right]^{\frac{1}{2}}\boldsymbol m_{(j , j)}^{1/2 } , \end{aligned}\ ] ] with the cardinality @xmath244 of @xmath241 , @xmath245 .
this yields the corresponding minimal msevalue as @xmath246=\sum_{j=1}^k e_0\left[\mathcal i_j\right]=\sum_{j=1}^k\frac{\hbar}{2}\,{\rm tr}\!\left(\left[\boldsymbol m_{(j , j)}^{-1/2}\boldsymbol g_{(j , j)}\boldsymbol m_{(j , j)}^{-1/2}\right]^{\frac{1}{2}}\right ) .
\end{aligned}\ ] ] let us note , that we get for @xmath101 @xmath247=e_0 $ ] , cf .
eq . .
the example in the paper has the following properties : @xmath248{\rm diag}(\boldsymbol 1_1,\dots,\boldsymbol 1_n)-\left[n\gamma\right]\vec n\,\vec n^{\rm
t}\\ \vec n=&\frac{1}{\sqrt{n}}\left(\boldsymbol 1_i\right)_{i=1}^{n}\\ \boldsymbol 1_i=&\vec e_{i , x}\,\vec e_{i , x}^{\,\rm t}+\vec e_{i , y}\,\vec e_{i , y}^{\,\rm t}+\vec e_{i , z}\,\vec e_{i , z}^{\,\rm t}. \end{aligned}\ ] ] note that each component is a 3 dimensional matrix itself and the potential energy is @xmath249^{\rm t}\boldsymbol g\left[(\vec r_j)_{j=1}^n\right]$ ] . for a given @xmath28 , we have @xmath250= & \frac{\hbar}{2}\,{\rm tr}\!\left(\left[\boldsymbol m_{(j , j)}^{-1/2}\boldsymbol g_{(j , j)}\boldsymbol m_{(j , j)}^{-1/2}\right]^{\frac{1}{2}}\right)\\ \nonumber = & \frac{\hbar}{2\sqrt{m}}\,{\rm tr}\!\left(\left ( \left[m\omega^2+n\gamma\right]\left[{\rm diag}(\boldsymbol 1_1,\dots,\boldsymbol 1_n)-\vec n_{(j)}\,\vec n_{(j)}^{\rm t}\right]+\left[m\omega^2+\gamma(n-|\mathcal i_j|)\right]\vec n_{(j)}\,\vec n_{(j)}^{\rm t } \right)^{\frac{1}{2}}\right)\\ \nonumber = & \frac{\hbar}{2\sqrt{m}}\,{\rm tr}\!\left ( \left[m\omega^2+n\gamma\right]^{\frac{1}{2}}\left[{\rm diag}(\boldsymbol 1_i)_{i\in\mathcal i_j}-\vec n_{(j)}\,\vec n_{(j)}^{\rm t}\right ] + \left[m\omega^2+\gamma(n-|\mathcal i_j|)\right]^{\frac{1}{2}}\vec
n_{(j)}\,\vec n_{(j)}^{\rm t}\right)\\ = & \frac{\hbar\omega}{2}\left(\left[1+\frac{\gamma n}{m\omega^2}\right]^{\frac{1}{2}}\left[3|\mathcal i_j|-3\right ] + \left[1+\frac{\gamma ( n-|\mathcal i_j|)}{m\omega^2}\right]^{\frac{1}{2}}3\right ) , \end{aligned}\ ] ] where the second line of @xmath251 $ ] represents the spectral decomposition of @xmath252 . | we derive a set of algebraic equations , the so - called multipartite separability eigenvalue equations . based on their solutions ,
we introduce a universal method for the construction of multipartite entanglement witnesses .
we witness multipartite entanglement of @xmath0 coupled quantum oscillators , by solving our basic equations analytically .
this clearly demonstrates the feasibility of our method for studying ultrahigh orders of multipartite entanglement in complex quantum systems .
entanglement represents a fundamental quantum correlation between compound quantum systems . since the early days of quantum physics
this property has been used to illustrate the surprising consequences of the quantum description of nature @xcite .
moreover , entanglement plays a fundamental role in various applications and protocols in quantum information science @xcite . in multipartite systems
a separable state is a statistical mixture of product states @xcite .
a quantum state is entangled , whenever it can not be represented in this form .
various forms of multipartite entanglement are known @xcite .
the most prominent and nonequivalent forms of entangled multipartite quantum states are the ghz - state @xcite and the w - state @xcite , which have been generalized to so - called cluster and graph states @xcite .
another classification is given in terms of partial and full ( or genuine ) multipartite entanglement , for an introduction see e.g. @xcite .
beyond finite dimensional systems , multipartite quantum entanglement in continuous variable systems turns out to be a cumbersome problem . even in the case of gaussian states ,
there exist multipartite entangled states , which can not be distilled @xcite .
high orders of multipartite entanglement are of great interest , for example , in quantum metrology .
multipartite entanglement has been shown to be essential to reach the maximal sensitivity in metrological tasks @xcite . in this context , the quantum fisher information has been used to characterize the entanglement @xcite . the detection of entanglement is typically done via the construction of proper entanglement witnesses @xcite , being equivalent to the method of positive , but not completely positive maps .
a witness is an observable , which is non - negative for separable states , and it can have a negative expectation value for entangled states . for different kinds of entanglement , different types of witnesses have been considered : bipartite witnesses @xcite ; schmidt number witnesses @xcite ; and multipartite witnesses for partial and genuine entanglement @xcite .
a systematic approach for witnessing entanglement in complex quantum systems is missing yet .
recently , we considered the construction of bipartite entanglement witnesses with the so - called separability eigenvalue equations @xcite .
we have shown that the same equations need to be solved to obtain entanglement quasiprobabilities , which are nonpositive distributions if and only if the corresponding quantum state is entangled @xcite .
moreover , we have shown that the schmidt number witnesses can be obtained by solving the related schmidt number eigenvalue problem @xcite . in the present letter
we derive a set of algebraic equations , which yield the construction of arbitrary multipartite entanglement witnesses .
for these so - called multipartite separability eigenvalue equations , we will study some fundamental properties , which uncover the structure of multipartite entanglement .
examples are given to witness partial and full entanglement in multipartite composed systems , for pure and mixed quantum states in discrete and continuous variable systems . in the following , we consider a composed hilbert space @xmath1 .
it has been shown that , without loss of generality , we could assume that the individual subsystems are finite dimensional ones @xcite .
let us consider a partition @xmath2 of the index set @xmath3 .
a quantum state @xmath4 is separable for the given partition , if it can be written as a classical mixture of product states @xcite : @xmath5 with @xmath6 being a classical probability distribution and @xmath7 being the set of pure and normalized separable states .
if a quantum state @xmath8 can not be written in the form of eq . , it is referred to as multipartite entangled .
a multipartite entanglement witness for the given partition is a hermitian operator @xmath9 , with @xmath10 based on refs .
@xcite , it can readily be shown that any witness can be presented in the form @xmath11 where @xmath12 is a general hermitian operator and the function @xmath13 denotes the maximally attainable expectation value for separable states : @xmath14 the supremum is taken over all @xmath15 .
hence , we can formulate a necessary and sufficient entanglement criterion being equivalent to the witness criterion : a quantum state @xmath8 is entangled with respect to the partition @xmath16 , if and only if there exists a hermitian operator @xmath12 such that @xmath17 this means that the mean value of @xmath12 exceeds the boundary of mean values for separable states .
a replacement @xmath18 leads to a similar entanglement criterion , but with the greatest lower bound ( @xmath19 ) instead of the least upper bound ( @xmath20 ) : @xmath21 for both entanglement criteria , we have to solve the following optimization problem for an observable @xmath12 : @xmath22 where @xmath23 represents the function to be optimized , and @xmath24 is the normalization condition .
for such an optimization problem , we can apply the method of lagrangian multipliers . in our case , the optimization condition is @xmath25 where @xmath26 is the lagrangian multiplier and @xmath27 is the null vector in the subspace given by the partition @xmath28 .
the partial derivatives of @xmath23 can be computed as : @xmath29}{\partial \langle a_j|}\\ = & { \rm tr}_{\mathcal i_1}\dots { \rm tr}_{\mathcal i_{j-1}}{\rm tr}_{\mathcal i_{j+1}}\dots { \rm tr}_{\mathcal i_k}\left(\hat l \left[|a_1,\dots , a_{j-1}\rangle\langle a_1,\dots , a_{j-1}|\otimes\hat 1_{\mathcal i_j}\otimes |a_{j+1},\dots , a_k\rangle\langle a_{j+1},\dots , a_k|\right]\right ) |a_j\rangle\nonumber\\ = & \hat l_{a_1,\dots , a_{j-1},a_{j+1},\dots , a_k}|a_j\rangle .
\end{aligned}\ ] ] the case @xmath30 yields the derivatives of @xmath24 .
let us also note that we assumed that the indices of the sets @xmath28 are ordered in a form that all elements of @xmath31 are larger then the elements of @xmath32 for @xmath33 .
this assumption is justified by the fact that one can employ , without loss of generality , a permutation of the hilbert spaces @xmath34 to order them in the required form . the euler - lagrangian optimization condition in eq .
can be reformulated for all @xmath35 as @xmath36 where all eigenstates are normalized ones , @xmath37 .
in addition , we may evaluate the value of @xmath26 .
we can do this by multiplying eq .
with @xmath38 .
this yields @xmath39 hence , the lagrangian multiplier corresponds to an optimal expectation value of @xmath12 for separable states . in conclusion of this derivation , we get an algebraic problem whose solutions give all optimal expectation values . [ [ definition - msevalue - equations . ] ] definition : msevalue equations. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the equations @xmath40 are defined as the first form of the multipartite separability eigenvalue ( msevalue ) equations .
the value @xmath26 is denoted as the msevalue of @xmath12 , and the product vector @xmath41 is the corresponding multipartite separability eigenvector ( msevector ) . as a final conclusion from this derivation we get @xmath42 and condition is given by the infimum of all msevalues .
this means that all multipartite entanglement witnesses can be constructed from the solutions of the msevalue equations , @xmath43 in case we consider finite hilbert spaces ( @xmath44 ) , we can replace @xmath20 and @xmath19 by @xmath45 and @xmath46 , respectively .
the derived msevalue equations play a fundamental role for multipartite entanglement tests .
they give the possibility to construct arbitrary entanglement witnesses on the basis of the solution of an algebraic eigenvalue problem of an observable @xmath12 .
numerical and analytical methods originally developed to solve eigenvalue problems can be applied to handle the multipartite entanglement problem in quantum physics in a systematic way . before we apply our method ,
let us formulate some fundamental properties of the msevalue equations .
the proofs are given in the appendix .
[ [ proposition - second - form - of - msevalue - equations . ] ] proposition : second form of msevalue equations. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the hermitian operator @xmath12 has the msevalue @xmath26 for the msevector @xmath41 , if and only if it fulfills the second form of the msevalue equations @xmath47 with @xmath48 for all @xmath49 and @xmath35 .
@xmath50 this proposition transforms the coupled system of eigenvalue equations , which has been defined in the first form of the msevalue equations , into a single , but perturbed eigenvalue problem .
this second form yields several implications .
for example , if an eigenvector is a product vector @xmath41 , it is also an msevector with @xmath51 .
in addition , we also conclude that the operator @xmath12 yields a true entanglement witness , cf .
eq . , if and only if the eigenspace of the largest eigenvalue , does not contain a product vector . [ [ proposition - transformation - properties . ] ] proposition : transformation properties. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + a hermitian operator @xmath12 has a msevalue @xmath26 for the msevector @xmath41
. then , the operator @xmath52 \left(\hat u_1\otimes\dots\otimes \hat u_k\right ) , \end{aligned}\ ] ] with @xmath53 and @xmath54 being unitary operations acting locally on the partition @xmath28 , has the msevalue @xmath55 and the msevectors @xmath56 .
@xmath50 this transformation allows us to consider a whole class of witnesses , by solving the msevalue equation for a particular operator @xmath12 .
in addition , the shifting of @xmath12 to @xmath57 allows us to consider positive semidefinite operators only .
note that the invariance of the msevalues under local unitaries is of particular interest for the quantification of multipartite entanglement , see , e.g. , @xcite .
[ [ proposition - cascaded - structure . ] ] proposition : cascaded structure. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the nonzero solutions of an @xmath58-partite operator @xmath59 are identical to the solutions of an @xmath60-partite operator @xmath61 .
@xmath50 this property is quite surprising .
it shows us that all possible entanglement witnesses based on positive semidefinite operators @xmath12 of an @xmath60-partite system can be constructed by a few simple entanglement witnesses in a @xmath58-partite system , @xmath59 .
an arbitrary rank of @xmath12 can be achieved by choosing a state @xmath62 with the same schmidt rank for the bipartition @xmath63 and @xmath64 , cf . the appendix . in the following ,
we apply our method to analytically derive multipartite entanglement tests .
first , we may consider witnesses for prominent examples of states in a three qubit systems . in a second step ,
we apply our method to get a multipartite entanglement test in a complex continuous variable system .
let us consider a generalized tripartite w - state @xmath65 with @xmath66 , which defines the observable @xmath67 .
in the appendix , we solve the msevalue equation of @xmath12 .
this gives @xmath68 with @xmath69 hence , we can formulate the following multipartite entanglement conditions : a quantum state @xmath70 is partially entangled , if @xmath71 .
the corresponding entanglement witness is @xmath72 a quantum state @xmath70 is fully entangled , if @xmath73 .
the corresponding entanglement witness is @xmath74 in fig .
[ fig : wstate ] , we apply the considered witness to study the entanglement of a noisy w - state .
@xmath75 , is plotted for @xmath76 .
the boundary for partial or full separability is @xmath77 or @xmath78 , respectively .
the expectation value @xmath79 exceeds the boundary for full or partial separability as long as the mixing parameter is @xmath80 or @xmath81 , respectively .
, height=151 ] in a second step , a generalized ghz - state is given , @xmath82 together with @xmath83 , which yields an observable @xmath84 . from the appendix
, we get the maximal msevalues : @xmath85 hence , a state @xmath70 is genuinely tripartite entangled , if @xmath86 , see fig .
[ fig : ghzstate ] .
note that the corresponding witness @xmath87 can not discriminate between partially and fully entangled states , see eq .
, which is possible for the generalized w - state projection in the previous example . for a ghz - state mixed with white noise , @xmath88 ( @xmath89 ) ,
is plotted . the lower boundary for partial and full
entanglement is @xmath90 .
the expectation value @xmath79 exceeds this boundary which automatically implies genuine entanglement as long as the mixing parameter is @xmath91 . , height=151 ] as a proof of principle , we are going to test multipartite entanglement of a continuous variable system .
our considered example is a system of @xmath60 coupled harmonic oscillators .
the observable we are using to verify entanglement is the total energy of this system , @xmath92 , @xmath93 where @xmath94 denotes the coupling strength of the interaction , @xmath95 the position and @xmath96 the momentum operator .
let us note that we considered an even more general case in the appendix .
for the partition @xmath97 , we get the smallest msevalue of the hamiltonian as @xmath98=\frac{3}{2}\hbar\omega&\sum_{j=1}^k\left([n_j-1]\sqrt{1+n\frac{\gamma}{m\omega^2}}\right.\\ & + \left.\sqrt{1+[n - n_j]\frac{\gamma}{m\omega^2}}\,\right ) , \end{aligned}\ ] ] where @xmath99 is the number of subsystems in @xmath28 .
the resulting witness reads as @xmath100\hat 1- \hat h. \end{aligned}\ ] ] in the special case @xmath101 ( @xmath102 ) we get the true ground state energy of the system , @xmath103=\frac{3}{2}\hbar\omega\left(\sqrt{1+n\frac{\gamma}{m\omega^2}}(n-1)+1\right ) .
\end{aligned}\ ] ] in case of full separability , @xmath104 ( @xmath105 ) , we have a minimal energy of : @xmath106=\frac{3}{2}\hbar\omega n\sqrt{1+[n-1]\frac{\gamma}{m\omega^2}}. \end{aligned}\ ] ] in fig .
[ fig : gaussianstate ] , we plotted the corresponding entanglement test based on eq .
, for @xmath107 interacting oscillators . for the witnessing by the total energy @xmath108 , no information about the structure of the quantum states is needed . )
/\left(\frac{3}{2}\hbar\omega\right)$ ] between quantum states and the ground state of the system is plotted depending on the interaction strength .
the boundary @xmath109 corresponds to the minimal attainable energy for fully separable states .
all quantum states having an energy below @xmath110 ( @xmath111 ) reveal entanglement between more than 200 oscillators .
the quantum states in the range below @xmath112 ( @xmath113 ) exhibit entanglement distributed over more than 500 subsystems .
, height=207 ] in conclusion , we have derived an algebraic set of equations to construct arbitrary entanglement witnesses .
we studied some fundamental properties of these equations .
for example , they are invariant under local unitary transformations and have a cascaded structure .
the latter allows us to deduce all entanglement witnesses from elementary projections .
our method enables us to use all known procedures for solving eigenvalue problems to construct entanglement witnesses .
we applied our method to analytically identify full and partial entanglement of generalized , noisy ghz- and w - states .
moreover , we witnessed multipartite entanglement for a system of @xmath0 interacting oscillators , by analytical computation of the energetic boundaries of separable states .
this demonstrates the feasibility of our method for studying ultrahigh orders of multipartite entanglement , which are of fundamental interest for understanding the transition from the microscopic to the macroscopic world .
the authors are grateful to d. pagel for helpful comments .
this work was supported by the deutsche forschungsgemeinschaft through sfb 652 .
99 a. einstein , n. rosen , and b. podolsky , phys . rev . * 47 * , 777 ( 1935 ) .
e. schrdinger , naturwissenschaften * 23 * , 807 , ( 1935 ) .
m. a. nielsen and i. l. chuang , _ quantum computation and quantum information _
( cambridge university press , cambridge , 2000 ) .
r. horodecki , p. horodecki , m. horodecki , and k. horodecki , rev .
mod .
phys . * 81 * , 865 ( 2009 ) .
o. ghne and g. tth , phys . rep . *
474 * , 1 ( 2009 ) .
r. f. werner , phys .
rev .
a * 40 * , 4277 ( 1989 ) .
l. amico , r. fazio , a. osterloh , and v. vedral , rev .
mod .
phys . * 80 * , 517 ( 2008 ) .
j. grabowski , m. ku , and g. marmo , j. phys .
a : math .
theor . * 44 * , 175302 ( 2011 ) .
m. huber and j. i. de vicente , phys .
rev .
lett . * 110 * , 030501 ( 2013 ) .
f. levi and f. mintert , phys .
rev .
lett . * 110 * , 150402 ( 2013 ) .
d. m. greenberger , m. a. horne , and a. zeilinger , _ going beyond bell s theorem in bell s theorem , quantum theory , and conceptions of the universe _
( kluwer academic , dordrecht , 1989 ) .
w. dr , g. vidal , and j. i. cirac , phys .
rev .
a * 62 * , 062314 ( 2000 ) .
r. raussendorf , d. e. browne , and h. j. briegel , phys .
rev .
a * 68 * , 022312 ( 2003 ) .
h. j. briegel and r. raussendorf , phys .
rev .
lett . * 86 * , 910 ( 2001 ) .
r. f. werner and m. m. wolf , phys .
rev .
lett . * 86 * , 3658 ( 2001 ) .
m. horodecki , p. horodecki , and r. horodecki , phys .
lett .
a * 223 * , 1 ( 1996 ) .
m. horodecki , p. horodecki , and r. horodecki , phys .
lett .
a * 283 * , 1 ( 2001 ) .
b. m. terhal , phys .
lett . a * 271 * , 319 ( 2000 ) .
j. sperling and w. vogel , phys .
rev .
a * 79 * , 022318 ( 2009 ) .
b. m. terhal and p. horodecki , phys .
rev .
a * 61 * , 040301(r ) ( 2000 ) .
j. sperling and w. vogel , phys .
rev .
a * 83 * , 042315 ( 2011 ) ; d. pagel , h. fehske , j. sperling , and w. vogel , phys . rev .
a * 86 * , 052313 ( 2012 ) .
p. van loock and a. furusawa , phys .
rev .
a * 67 * , 052315 ( 2003 ) .
m. bourennane _
et al . _ ,
phys .
rev .
lett . * 92 * , 087902 ( 2004 ) .
j. eisert , f. g. s. l. brando , and k. m. r. audenaert , new j. phys .
* 9 * , 46 ( 2007 ) .
d. chruciski and a. kossakowski , j. phys .
a : math . theor . * 41 * , 145301 ( 2008 ) . c. gross , t. zibold , e. nicklas , j. estve , and m. k. oberthaler , nature ( london ) * 464 * , 1165 ( 2010 ) .
g. tth , phys .
rev .
a * 85 * , 022322 ( 2012 ) .
p. hyllus , w. laskowski , r. krischek , c. schwemmer , w. wieczorek , h. weinfurter , l. pezz , and a. smerzi , phys .
rev .
a * 85 * , 022321 ( 2012 ) .
r. krischek , c. schwemmer , w. wieczorek , h. weinfurter , p. hyllus , l. pezz , and a. smerzi , phys .
rev .
lett . * 107 * , 080504 ( 2011 ) .
j. sperling and w. vogel , phys .
rev .
a * 79 * , 042337 ( 2009 ) ; j. sperling and w. vogel , new j. phys .
* 14 * , 055026 ( 2012 ) . j. sperling and w. vogel , phys .
rev .
a * 79 * , 052313 ( 2009 ) .
g. tth , phys .
rev .
a * 71 * , 010301(r ) ( 2005 ) . |
jan oort @xcite based his historic hypothesis on the existence of a distant , spherical cloud of cometary bodies surrounding the solar system on the original @xmath1-distribution of nineteen near - parabolic comets with the best - determined ( at that time ) original orbits .
comet c/1890
f1 brooks belonged to this group .
it is distinguished by a long and rich series of observations : almost 2yr period covered by about 900 positional measurements .
a question of the distribution of the reciprocal original semimajor axes @xmath2 is frequently repeated in the literature not only to support the oort cloud ( oc ) hypothesis , but also to examine the density distribution of the oc , recognized as a reservoir of the cometary bodies that can potentially penetrate into the planetary system @xcite .
however , it was already established @xcite , that a significant fraction of investigated comets with the original @xmath1 within the so - called oort spike , @xmath3au@xmath4 , passed through the inner part of the solar system in their previous perihelion passage . therefore , the interpretation of the original @xmath1-distribution of the near - parabolic comets should be treated with caution . in particular
, one should take into consideration that a vast majority of comets from the oort spike suffered planetary perturbations that have changed significantly their semimajor axes during each deep passage through planetary zone . here
we focus our investigation on the dynamical history of oort spike comets in the period extending back to their previous perihelion passage .
it is the only direct method that allows us to separate the dynamically new comets of the oort spike from the dynamically old ones .
typical planetary perturbations defined by a change of @xmath1 during the passage within the inner part of the solar system are typically two to four times greater than the present estimate of the oort spike width . in consequence , after visiting the inner planetary system , a comet has an excessive chance to be outside the oort spike , leaving the solar system on a hyperbolic orbit , or moving on a significantly tighter orbit than previously .
since many comets discovered so far will escape from the solar system in the future , we also investigate their future evolution which gives a much broader perspective on the dynamical evolution of comets discovered as oort spike comets ( * ? ? ?
* hereafter paper 5 and references therein ) .
in addition to the above general arguments , there are also some specific objectives of the present investigation .
we chose comet c/1890
f1 brooks not only for its historic importance but also to develop and test our methods of data treatment for such a long - ago discovered comet . in the case of c/1890
f1 , we have an opportunity to repeat the reduction of astrometric observations using contemporary star catalogues , typically much more precise , particularly in terms of proper motions .
the ( @xmath5)-measurementsstar)-measurements ] ( @xmath6 in right ascension and @xmath7 in declination ) are fortunately published in original papers together with calculated apparent comet positions for almost all observations of this comet . apart from seven transit circle observations made in washington the only sources without explicit ( @xmath5)-differences are original papers reporting 99 observations from bordeaux observatory ( only one paper with five observations from bordeaux listed also ( @xmath5)-measurements )
. fortunately , these publications also consist of both , reference star identifications and their apparent places used for the reduction .
thereby , the reconstruction of the ( @xmath5)-measurements is quite straightforward .
for all these reasons , we also decided to show in this paper how the determination of the observed osculating orbitdepends on the adopted method of data processing ( how deep we are looking for sources of errors in the published data ) and our choice of the catalogue for reference star positions .
two methods and two catalogues have been used for this purpose : ppm star catalogue @xcite and tycho-2 star catalogue @xcite .
the first , fully automatic , method is very useful in cases in which we have at hand a list of astrometric cometary positions together with ( @xmath5)-measurements in right - ascension and declination , and we had not entered data on the reference stars originally used by observers .
it involves the reconstruction of hypothetical star positions on the basis of ( @xmath5)-measurements , and published cometary positions if they were given by observers .
otherwise , the cometary positions are estimated from preliminary orbital elements .
next , the re - reduction of these stars positions using ppm catalogue is carried out . finally , the positions of comets are calculated based on obtained present - day star positions and original ( @xmath5)-measurements . of course
, it can also be applied to a star catalogue other than ppm , but here the ppm catalogue is used for a comparison purpose .
this method was successfully used for many comets discovered long ago ( @xcite and @xcite ) , thus it is only very briefly described in section [ stars_in_ppm ] . in this investigation however , all the data originally published by observers were collected .
therefore , a much more in - depth analysis of the original data can be done and the second method described in this paper is devoted to such an investigation .
the tycho-2 catalogue is chosen as it is perceived to be one of the best modern catalogues for this purpose , particularly regarding proper motions . in the first step , we collected a list of mean coordinates of all reference stars used by all observers and then their contemporary astrometric data were automatically obtained . next , we calculated astrometric positions of the comet using ( @xmath5)-measurements .
finally , manual and time consuming part was to investigate the causes of large residuals obtained from this automated step for about 30 per cent of all observations .
about half of them resulted from the wrong star identifications and the rest came from some inconsistency or typographical errors in papers published by observers . more details about this second method are given in section [ section : stars_in_tycho-2 ] .
the paper consists of seven sections .
next section describes the observational material and previous orbital determinations of c/1890 f1 brooks that exist in literature . in section [ section : star_recalculation ] details of the method of reference star position recalculation using the procedures based on ppm and tycho-2 star catalogues are given , and a brief description of further steps of our data treatment is delivered .
a grid of osculating orbits based on both methods of reference star recalculations is described in section [ section : our_osculating_orbits ] , where also the analysis of differences between these results is presented .
the original and future orbits are discussed in section [ section : original_future_orbits ] .
the final result of this study , the most important from the dynamical point of view , is presented in section [ section : past_next_orbits ] , where we discuss what can be deduced about the origin of comet c/1890
f1 brooks and its future dynamical evolution . the summary and conclusions resulting from this research are outlined in the last section . [ cols="<,^,<,^ " , ]
many interesting aspects of the past and future orbit evolution of c/1890 f1 under stellar and galactic perturbations can be found in fig.[fig : nominal - c/1890f1-orbit ] . in this picture
the horizontal time axis extends from the previous perihelion passage through the observed apparition up to the next perihelion passage .
the left vertical axis is expressed in au and describes the osculating perihelion distance evolution ( q , green line ) , as well as the heliocentric distance plot ( @xmath8 , thin blue lines ) . due to the time scale of this picture ,
the heliocentric distance plots take a form of vertical blue lines exactly at perihelion passage moments .
the right vertical axis is expressed in degrees and describes changes in the osculating inclination ( @xmath9 , magenta line ) and in the argument of perihelion ( @xmath10 , red line ) .
both of these angular elements are expressed in the galactic frame .
all thick lines depict dynamical evolution under joint stellar and galactic perturbations , while the thin lines mark the evolution with the stellar perturbations excluded .
horizontal dashed lines draw attention to the beginning of the second and fourth quarter of @xmath10 , whose values ( 90and 270 ) are important from the point of view of the galactic perturbations ( crossing these lines coincides with perihelion distance minimum ) .
the vertical dashed lines show the closest approaches of a comet with the star or stellar system , which name is placed at the top of the picture .
it is worth mentioning that the timing of the stellar perturbation is not necessarily strictly aligned with this closest approach moment it strongly depends on the geometry , since the final heliocentric orbit change is a net effect of a stellar gravitational action on both a comet and the sun .
the action of gj 217.1 is a good example of a frequent self - cancelling stellar action : the perturbation gained during the approaching phase is then cancelled during the receding phase . during the presented 3.5myr interval of cometary evolution ,
the galactic and stellar perturbations are notable only in the perihelion distance , while being infinitesimally small in angular elements .
we have not identified any profound stellar perturbation in future motion of c/1890f1 since it will depart from the sun for only about 20000au and it will complete its future orbital revolution ( 1.1myr ) about 0.3myr before the predicted close flyby of the star gliese 710 .
other stars are too small and/or too distant to noticeably change the future orbit of this comet . and @xmath11 of 5001 vcs derived from the nominal orbit of the preferred solution for c/1890 f1 , stopped at their previous perihelion passage .
the centre of the green circle marks the nominal values .
small green dots overprinted on the marginal distributions show best - fitting gaussian distributions . ] and [ tab : previous_and_next ] for more details ) .
all known stellar and galactic perturbations were taken into account .
black lines depict nominal orbits evolution .
distributions of previous and next perihelia epochs are also shown . ]
c/1890 f1 is definitely the dynamically old comet .
even the most protruding clones in our swarm of vcs have the previous perihelion distance well below 10au . the osculating perihelion distance and the inverse semimajor axis distributions of c/1890 f1 for the moment of the previous perihelion passage ( nominally 2.5myr ago )
are presented in fig .
[ fig : joint - and - marginal ] .
it is worth mentioning that the presented @xmath1-distribution is still almost perfectly gaussian while the @xmath11-distribution significantly departs from the normal one , mainly as a result of galactic perturbations . the central , black joined distribution corresponds to the smaller cloud of 5001vcs presented in fig .
[ fig : original_swarms_dataii ] after the 1.73.8myr of backward dynamical evolution under the simultaneous galactic and stellar action ( different vcs have different orbital periods , a nominal value equals 2.5myr , the fastest vc was at the previous perihelion 1.7myr ago while the slowest one almost 3.8myr ago ) . the time - spread ( along horizontal axis ) and the space - spread ( along vertical axis ) of our swarms of vcs in previous and next perihelion
are clearly visualized in fig .
[ fig : fontanny ] .
we presented here changes in the barycentric distance of all 5001 vcs from two different solutions of orbit determination , both based on the data_iib set of observations .
the red ( darker ) swarm depicts the distribution evolving from the osculating orbit fitted to the weighted observations ( a5-solution ) while the green ( grey ) one depicts the evolution of swarm constructed without weighting ( a1-solution ) .
nominal orbits for both of our solutions are marked as black curves .
thus , this figure shows also how the data handling can change the results during the evolution to the previous / next perihelion distance ( see also table [ tab : previous_and_next ] ) .
the dynamical evolution of strmgren s orbit would be very similar to the evolution of our a1 orbital solution .
we conclude that our preferred a5-solution gives longer previous and next orbital periods of about 2.5 and 1.1myr respectively . as a result ,
galactic perturbations act more effective during the evolution , and for this solution we get a greater change of the perihelion distances between three consecutive perihelion passages analysed here .
a rich observational material of c/1890 f1 brooks allows to test whether it is productive to recalculate with the use of modern star catalogues the original positions of comets discovered long ago . during this study we compared two distinct algorithms : * automatic search for reference stars in the ppm star catalogue according to @xcite , which was frequently used by us when we only have ( + @xmath5)-measurements and data on the positions of comets , * automatic search for reference stars in the tycho-2 star catalogue on the basis of mean coordinates of reference stars used by all observers , and next , a detailed analysis of all the cases of large residuals in the observed comet s positions , resulting from the star search ( @xmath1230 per cent of all observations in the case of c/1890 f1 ) . the first method was successfully used several times before , and more recently by @xcite .
this method of automatic search for stars in ppm catalogue recalculates about 50 per cent of the existing data of c/1890 f1 .
the second method was developed for the present investigation . in the simbad database
we have successfully found all stars used by observers more than hundred years ago to calculate the comet s position in @xmath13 and/or @xmath14 from ( @xmath5)-measurements .
it means , that all positions of c/1890f1 were recalculated here using these original ( @xmath5)-measurements ( @xmath6 and/or @xmath7 ) .
we decided to use stellar positions and proper motions from the tycho-2 catalogue for that purpose .
it should be emphasized that thanks to the monumental publication by elis @xcite , our task has become much easier , though , it was still time consuming . searching for stars in the simbad database would be more difficult without the mentioned publication , particularly in the more complicated cases .
in addition , thanks to strmgren s deep analysis of data , it turned out that one should take into account corrections to @xmath7 measurements from bordeaux observatory ( we prefer linear version of these corrections as described in section [ section : stars_in_tycho-2 ] ) .
however , these corrections change the orbital solution on the significantly lower level than the use of modern stellar data . for example , a solution based on data recalculated in about 50 per cent ( ppm catalogue , data_ia / data_ib ) differs a lot from the respective solution based on fully recalculated data ( tycho-2 catalogue data_iia / data_iib ) than any two solutions based on data with and without bordeaux corrections .
moreover , the bordeaux corrections are also much less important than the method of data treatment .
this is clearly shown in fig .
[ fig : original_axes_all_models ] , where we can see that light - green histogram ( weighted data , corrections for bordeaux omitted ) is very similar to dark - green histogram ( weighted data , corrections for bordeaux applied ) , however , it differs significantly from the pair of yellow / orange histograms based on unweighted data .
next , we have shown that both search algorithms for contemporary data for the reference stars resulted in a significant reduction of the uncertainties of orbital elements in the case of c/1890 f1 brooks .
the comparison of our solutions with that derived by strmgren is presented in table [ tab : orbits_osculating ] and fig .
[ fig : osculating_swarms_dataiib ] . by analysing all these models
, we concluded that the orbital solution based on data_iib ( all positional observations were recalculated using tycho-2 catalogue and next weighted ) gives the most reliable , and also the most accurate osculating orbit .
using this most preferred orbital solution ( and its unweighted variant for comparison , see fig .
[ fig : fontanny ] ) we numerically followed a motion of c/1890 f1 for one orbital period to the past and to the future . starting from full swarms of original and future vcs orbits we obtained previous and next orbital elements of this comet together with their uncertainties .
all known stellar and galactic perturbations were fully taken into account but for comparison purposes we have also performed the calculations in which stellar perturbations were ignored .
this comparison ( see table [ tab : previous_and_next ] for details ) shows that none of known stellar perturbers significantly change the past and future motion of c/1890 f1 during the three successive perihelion passages analysed here .
it should be stressed that due to relatively short past orbital period of c/1890f1 brooks ( nominally 2.5myr ) it seems rather improbable that we have missed here some unknown but significant ( i.e. massive and/or slow moving ) and nearby stellar perturber .
three consecutive perihelion passages of this comet ( the time interval of about 3.6myr ) clearly indicate that this comet deeply penetrates the planetary region during each of them ( below 5au from the sun , see table [ tab : previous_and_next ] ) and easily crosses the so - called jupiter - saturn barrier .
therefore , we conclude that c/1890 f1 brooks is a dynamically old comet and additionally , as a result of negligible planetary perturbations during the observed apparition , it will remain a member of the oort spike during the next apparition .
note that @xcite treated this comet ( by definition ) as a dynamically new one and used it in support for his cometary cloud hypothesis .
currently , on the basis of analysis of 109 oort spike comets , we estimate that about 50 per cent of them are dynamically old .
however , the percentage of dynamically old comets grows up to almost 90 for comets with @xmath2 inside the range of 0.0000400.000100au@xmath4 @xcite .
thus , the past dynamical evolution of comet brooks having @xmath2 greater than @xmath15au@xmath4 is not surprising . on the other hand , for the reason of very weak planetary perturbations one can say that the dynamical behaviour of c/1890 f1 is unusual among small perihelion comets . in the near future ,
we plan to deal with all near - parabolic comets discussed by @xcite as well as all oort spike comets discovered in the years 19011950 to answer the question about their source from the previous perihelion perspective , and make our sample of near - parabolic comets with known previous and next orbital elements more complete .
it can also happen that some comets observed on a more tightly bound orbit ( @xmath16au@xmath4 ) will contribute to a future @xmath1-distribution as oort spike comets .
therefore the extension of our research to objects with @xmath1 similar to that shared by c/2013 v2 borisov ( @xmath17au@xmath4and @xmath18au@xmath4 , @xcite ) seems to be important to put our investigation in the broader framework of dynamical evolution of long - period comets .
some auxiliary material to this paper is available at ` http://ssdp.cbk.waw.pl/lpcs ` and ` http://apollo.astro.amu.edu.pl/wcp ` .
we are very grateful to ryszard gabryszewski for providing software and related assistance with the search for stars in ppm star catalogue .
we also thank the anonymous referee of this paper for valuable remarks .
this research has made use of nasa s astrophysics data system bibliographic services and of the the simbad database and vizier catalogue access tool , cds , strasbourg , france , and was partially supported from the project 2015/17/b / st9/01790 founded by national science centre in poland .
99 w. r. , 1890 , mnras , 50 , 375 dybczyski p. a. , 2001 , a&a , 375 , 643 dybczyski p. a. , krlikowska m. , 2011 , mnras , 416 , 51 p. a. , krlikowska m. , 2015 , mnras , 448 , 588 fernndez j. a. , ed .
2005 , comets - nature , dynamics , origin and their cosmological relevance astrophysics and space science library vol .
328 r. , 1997 , planetary and space science , 45 , 1653 hg e. , et al . , 2000 ,
a&a , 355 , g. w. , 2003 , cometography . a catalog of comets .
18001899 m. , 2014 , a&a , 567 , a126 krlikowska m. , dybczyski p. a. , 2010 , mnras , 404 , 1886 krlikowska m. , dybczyski p. a. , 2013 , mnras , 435 , 440 m. , sitarski g. , sotan a. m. , 2009 , mnras , 399 , 1964 m. , sitarski g. , pittich e. m. , szutowicz s. , ziokowski k. , rickman h. , gabryszewski r. , rickman b. , 2014 , a&a , 571 , a63 marsden b. g. , williams g. v. , 2008 , catalogue of cometary orbits 17th edition .
smithsonian astrophysical observatory , cambridge , mass .
marsden b. g. , sekanina z. , everhart e. , 1978 , aj , 83 , 64 nakano s. , 2015 , nakano note 2952 , url http://www.oaa.gr.jp/ oaacs / nk / nk2952.htm oort j. h. , 1950 , bull.astron.inst.nether . , 11 , 91 perryman m. a. c. , et al .
, 1997 , a&a , 323 , s. , bastian u. , 1988 , astronomy and astrophysics supplement series , 74 , 449 e. , 1948 , pub .
copenhagen obs .
, 146 sitarski g. , 1998 , acta astronomica , 48 , 547 e. , 1896 , berechnung der bahn des kometen 1890ii e. , 1914 , publ .
copenhagen , 19 , 189 superintendent of the us naval observatory ( probably phythian , r.l . ) , 1890 , aj , 10 , 63 yabushita s. , 1989 , aj , 97 , 262 n. , finch c. t. , girard t. m. , henden a. , bartlett j. l. , monet d. g. , zacharias m. i. , 2013 , aj , 145 , 44 | c/1890 f1 brooks belongs to a group of nineteen comets used by jan oort to support his famous hypothesis on the existence of a spherical cloud containing hundreds of billions of comets with orbits of semimajor axes between 50 and 150thousand au .
comet brooks stands out from this group because of a long series of astrometric observations as well as nearly two - year long observational arc .
rich observational material makes this comet an ideal target for testing the rationality of an effort to recalculate astrometric positions on the basis of original ( comet star)-measurements using modern star catalogues .
this paper presents the results of such new analysis based on two different methods : ( i ) automatic re - reduction based on cometary positions and the ( comet star)-measurements , and ( ii ) partially automatic re - reduction based on the contemporary data for originally used reference stars .
we show that both methods offer a significant reduction of orbital elements uncertainties . based on the most preferred orbital solution
, the dynamical evolution of comet brooks during three consecutive perihelion passages is discussed .
we conclude that c/1890 f1 is a dynamically old comet that passed the sun at a distance below @xmath0au during its previous perihelion passage .
furthermore , its next perihelion passage will be a little closer than during the 1890 - 1892 apparition .
c/1890 f1 is interesting also because it suffered extremely small planetary perturbations when it travelled through the planetary zone .
therefore , in the next passage through perihelion it will be once again a comet from the oort spike .
[ firstpage ] solar system : general , oort cloud , comets : c/1890 f1 brooks |
search for the higgs bosons and the precise measurements on their properties , such as the masses , the decay widths and the decay branching ratios , are the most important subjects to study the mechanism of the electroweak symmetry breaking @xcite . in the standard model ( sm ) , only one physical neutral higgs boson appears .
on the other hand , models with multiple higgs doublets have cp - even and odd neutral higgs bosons as well as charged higgs bosons , if cp is a good symmetry .
otherwise , the neutral higgs bosons do not have to carry any definite cp parity .
+ cp violation was observed in the neutral kaon system @xcite and the violation in b meson decays is strongly suggested by recent experiments @xcite .
in addition , cp violation constitutes one of the crucial ingredients for an efficient non
sm generation of the cosmological baryon asymmetry at the electroweak scale @xcite .
an appealing scheme of cp violation beyond the sm can be provided by models with an extended higgs sector where the cp asymmetry is broken by the ground state of the higgs potential .
it has recently been realized that , even though the higgs potential of the minimal supersymmetric sm ( mssm ) is cp invariant at the tree level , explicit cp violation in the mass matrices of the third generation squarks can induce sizable cp violation in the mssm higgs sector through loop corrections @xcite .
the cp violating phases for the third generation sfermions can be quite large , since they contribute to the electric dipole moments ( edm s ) of the electron and neutron only at the two loop level with no generation mixing in the sfermion sector @xcite . as a result , although a one loop effect , the induced cp violation in the mssm higgs sector can be large enough to significantly affect higgs phenomenology at present and future colliders @xcite . in this light
, it is very important to examine the possibility of observing these higgs bosons and investigating their properties in detail . +
a muon collider is one of the ideal machines to look for the higgs bosons and their properties , where @xmath6 pairs can be directly converted into neutral higgs
boson resonances @xcite .
polarized muons can be used to detect the higgs bosons and measure their properties such as masses , total widths , decay branching fractions and cp parities @xcite . in this paper , we present a general formalism and a detailed analysis for the higgs boson effects on the polarization observables of the production process @xmath0 @xmath7 and @xmath2 ) using the initial muon beam polarization along with the unpolarized final fermions and with the final fermion polarization configuration of equal helicity , respectively .
it has been recently pointed out that the amplitudes for two higgs states , a scalar and a pseudoscalar , can interfere with each other sizably if the helicities of the initial and final particles are properly fixed and if the mass difference of these higgs bosons is at most of the same order as their decay widths @xcite .
we extend the work comprehensively in order to consider higgs bosons of no definite cp parity in the mssm with explicit cp violation induced from the third generation squark sectors .
+ the remainder of this article is organized as follows .
section [ sec : explicit cp violation ] is devoted to a brief review of the explicit cp violation in the mssm higgs sector based on the work @xcite . in section [ sec :
production ] we present the helicity amplitudes of the production of a fermion pair in @xmath6 collisions and give a comprehensive classification of the observables according to the cp and cp@xmath8 transformation properties that are constructed by the initial muon beam polarization along with the final fermions of no polarization and of equal helicity , respectively . in section [ sec : result ] we make a detailed numerical analysis based on a given parameter set so as to get a concrete estimate of the usefulness of the observables . section [ sec : conclusion ] is devoted to a brief summary of our findings and to a conclusion .
the existence of the non trivial cp violating phases in the mssm is due to the breakdown of supersymmetry so that the cp violating phases appear in the soft breaking parameters and the mixing among sparticles due to the electroweak gauge symmetry breaking .
there are three well known sources of cp violation in the mssm @xcite .
the first is related to the two higgs boson doublets present in the model since both the @xmath9 parameter in the superpotential and the soft breaking parameter @xmath10 can be complex .
secondly , there are three more phases of the complex masses of the u(1)@xmath11 , su(2)@xmath12 and su(3)@xmath13 gauginos of the sm gauge group .
thirdly , the other cp violating phases originate from the flavor sector of the mssm lagrangian , either in the scalar soft mass matrices or the trilinear matrices . + the cp violating phases associated with the off - diagonal terms of the trilinear matrices are , however , strongly suppressed by the same mechanism required to suppress the flavor changing neutral current effects .
therefore , all these flavor changing cp violating phases are neglected in the present work such that the scalar soft mass matrices and trilinear parameters are flavor diagonal and the complex trilinear terms are proportional to the corresponding fermion yukawa couplings . clearly , the yukawa interactions of the third generation quarks and squarks play the most significant role in radiative corrections to the higgs sector . in this section ,
we give a brief review of the calculation @xcite of the higgs boson mass matrix based on the full one loop effective potential , valid for all values of the relevant third generation soft breaking parameters .
+ the mssm contains two higgs doublets @xmath14 , with hypercharges @xmath15 .
here we are only interested in the neutral components , which we write as @xmath16 where @xmath17 and @xmath18 are real fields .
the constant phase @xmath19 can be set to zero at tree level , but will in general become non zero once loop corrections are included .
+ the mass matrix of the neutral higgs bosons can be computed from the effective potential @xcite @xmath20,\end{aligned}\ ] ] where we have allowed the soft breaking parameter @xmath21 to be complex , and we have introduced the quantities @xmath22 where the symbols @xmath23 and @xmath24 stand for the su(2)@xmath12 and u(1)@xmath11 gauge couplings , respectively .
@xmath25 in eq .
( [ e2 ] ) is the renormalization scale ; the parameters of the tree
level potential , in particular the mass parameters @xmath26 and @xmath27 , are running parameters , taken at scale @xmath25 .
the potential ( [ e2 ] ) is then independent of @xmath25 , up to two loop corrections .
+ the matrix @xmath28 is the field
dependent mass matrix of all modes that couple to the higgs bosons .
the by far dominant contributions come from the third generation quarks and squarks .
the ( real ) masses of the former are given by @xmath29 where @xmath30 and @xmath31 are the bottom and top yukawa couplings .
the corresponding squark mass matrices can be written as @xmath32 \\ - h_t \left [ a_t h^0_2 + \mu^ * \left ( h_1 ^ 0 \right)^ * \right ] & m^2_{{\widetilde}u } + m_t^2 - \frac{g'^2}{6 } { \cal d } \end{array } \right ) ; $ } \nonumber\\ { \cal m}_{\tilde b}^2 & = \mbox{$ \left ( \begin{array}{cc } m^2_{{\widetilde}q } + m_b^2 + \frac{1}{8 } \left ( g^2 + \frac{g'^2}{3 } \right ) { \cal d } & - h_b^ * \left [ a_b^ * \left ( h_1 ^ 0 \right)^ * + \mu h_2 ^ 0 \right ] \\ - h_b \left [ a_b h_1 ^ 0 + \mu^ * \left ( h_2 ^ 0 \right)^ * \right ] & m^2_{{\widetilde}d } + m_b^2 + \frac{g'^2}{12 } { \cal d } \end{array } \right ) .
$ } \label{e4}\end{aligned}\ ] ] here , @xmath33 and @xmath34 are given by eq .
( [ e1 ] ) while @xmath35 and @xmath36 are as in eq .
( [ e3 ] ) and @xmath37 has been defined in eq .
( [ e5 ] ) . in eq .
( [ e4 ] ) @xmath38 and @xmath39 are real soft breaking parameters , @xmath40 and @xmath41 are complex soft breaking parameters , and @xmath9 is the complex supersymmetric higgs(ino ) mass parameter . + the calculation proceeds by plugging the field
dependent top / bottom(s)quark mass eigenvalues into the potential ( [ e2 ] ) . the mass matrix of the higgs bosons ( at vanishing external momentum )
is then given by the matrix of second derivatives of this potential , computed at its minimum . in order to make sure that we are indeed in the minimum of the potential
, we solve the stationarity relations , i.e. set the first derivatives of the potential to zero .
this allows us to , e.g. , express @xmath42 and @xmath43 as functions of the vacuum expectation values ( vevs ) and the remaining parameters appearing in the loop corrected higgs potential .
the equations @xmath44 and @xmath45 are linearly dependent , i.e. lead to only one constraint on parameters for @xmath46 ; the remaining vevs are defined through @xmath47 and @xmath48 .
the re phasing invariant quantity @xmath49 is then determined by @xmath50 , @xmath51 and @xmath52 as well as by the re phasing invariant quantities @xmath53 which describe the amount of cp violation in the squark mass matrices .
+ the mass matrix of the neutral higgs bosons can now be computed from the matrix of second derivatives of the potential ( [ e2 ] ) , where ( after taking the derivatives ) @xmath54 and @xmath55 are determined by the stationarity conditions .
the massless state @xmath56 is the would
be goldstone mode `` eaten '' by the longitudinal @xmath57 boson .
we are thus left with a squared mass matrix @xmath58 for the three states @xmath59 and @xmath60 .
this matrix is real and symmetric , i.e. it has 6 independent entries .
the diagonal entry for @xmath61 reads : @xmath62 and the cp violating entries of the mass matrix , which mix @xmath61 with @xmath63 and @xmath60 read : @xmath64 \right .
\\ & & \left .
\hskip 1 cm + \frac { m_b^2 \delta_{\tilde b } } { \cos \beta } \left [ -g({m^2_{\tilde{b}_1}},{m^2_{\tilde{b}_2 } } ) \left ( x_b + 2 { \left| h_b \right|^2}r_b ' \right ) + \left ( { \hat{g}^2}- 2 { \left| h_b \right|^2}\right ) \log
\frac { { m^2_{\tilde{b}_2 } } } { { m^2_{\tilde{b}_1 } } } \right ] \right\ } , \nonumber\\ \left . { \cal m}^2_h \right|_{a \phi_2 } & = & \frac { 3 } { 16 \pi^2 } \left\ { \frac { m_t^2 \delta_{\tilde t } } { \sin \beta } \left [ -g({m^2_{\tilde{t}_1}},{m^2_{\tilde{t}_2 } } ) \left ( x_t + 2 { \left| h_t \right|^2}r_t ' \right ) + \left({\hat{g}^2}-2{\left| h_t \right|^2}\right)\log\frac{{m^2_{\tilde{t}_2}}}{{m^2_{\tilde{t}_1 } } } \right]\right . \\ & & \left
. \hskip 1 cm + \frac { m_b^2 \delta_{\tilde b } } { \cos \beta } \left [ g({m^2_{\tilde{b}_1 } } , { m^2_{\tilde{b}_2 } } ) \left ( x_b { \tan \ ! \beta}- 2 { \left| h_b \right|^2}r_b \right ) - { \hat{g}^2}{\tan \ ! \beta}\log \frac{{m^2_{\tilde{b}_2}}}{{m^2_{\tilde{b}_1 } } } \right]\right\}. \nonumber\end{aligned}\ ] ] where @xmath65 and @xmath66 are as in eq .
( [ e9 ] ) and the function @xmath67 is given by @xmath68 the definition of the mass squared @xmath69 and the dimensionless quantities @xmath70 , @xmath71 and @xmath72 as well as the other cp preserving entries of the mass matrix squared @xmath73 can be found in ref .
as noted earlier , the size of these cp violating entries is controlled by @xmath65 and @xmath66 . + the real
, symmetric matrix @xmath74 can be diagonalized with an orthogonal rotation @xmath75 ; @xmath76 with the ordering of @xmath77 taken as a convention in the present work .
note that the loop corrected neutral higgs
boson sector is determined by fixing the values of various parameters ; @xmath78 , @xmath9 , @xmath41 , @xmath40 , a renormalization scale @xmath25 , @xmath79 , and the soft breaking third generation sfermion masses , @xmath80 , @xmath81 , and @xmath82 .
the radiatively induced phase @xmath19 is no more an independent parameter and it can be absorbed into the definition of the @xmath9 parameter so that the physically meaningful cp phases in the higgs sector are the phases of the re phasing invariant combinations @xmath83 and @xmath84 .
this neutral higgs
boson mixing changes the couplings of the higgs fields to fermions , gauge bosons , and higgs fields themselves so that the effects of cp violation in the higgs sector can be probed through various processes @xcite .
the couplings of the @xmath85 and the neutral gauge boson @xmath57 to fermions in the mssm is described by the same interaction lagrangian as in the sm : @xmath86 \gamma_\mu f\ , z^\mu\,,\end{aligned}\ ] ] with the chirality projection operators @xmath87 , and those of the neutral higgs
boson fields with leptons and quarks are described by the interaction lagrangians @xmath88\ell h_k -\frac{h_d}{\sqrt{2}}\sum_{k=1}^3 \bar{d}\left[o_{2k}-is_\beta\,o_{1k}\gamma_5\right ] d h_k\nonumber\\ & & -\frac{h_u}{\sqrt{2}}\sum_{k=1}^3 \bar{u}\left[o_{3k}- ic_\beta\,o_{1k}\gamma_5\right]u h_k\,,\end{aligned}\ ] ] where @xmath89 , @xmath90 , and @xmath91 are the lepton and quark yukawa couplings : @xmath92 respectively .
it is then clear that all the neutral higgs bosons couple dominantly to the third generation fermions @xmath2 , @xmath93 and @xmath94 and they couple to a muon about 200 times more strongly than to an electron the _ primary reason _ for having a muon collider .
+ ( 300,100)(0,0 ) ( 5,85)[]@xmath95 ( 10,75)(35,50 ) ( 35,50)(10,25 ) ( 5,15)[]@xmath96 ( 35,50)(75,50)4 ( 55,37)[]@xmath97 ( 75,50)(100,75 ) ( 105,85)[]@xmath98 ( 100,25)(75,50 ) ( 105,15)[]@xmath99 ( 185,85)[]@xmath95 ( 190,75)(215,50 ) ( 185,15)[]@xmath96 ( 215,50)(190,25 ) ( 215,50)(255,50)48 ( 235,37)[]@xmath100 ( 255,50)(280,75 ) ( 287,85)[]@xmath98 ( 280,25)(255,50 ) ( 287,15)[]@xmath99 + * figure 1 * : _ the mechanisms contributing to the process @xmath0 ; three spin0 neutral higgs boson exchanges and spin1 @xmath85 and @xmath57 exchanges . here
, the index @xmath101 is 1,2 or 3 . _
the fermion pair production process @xmath0 is generated by five @xmath102channel mechanisms : spin1 @xmath85 and @xmath57 exchanges and three spin0 neutral
boson exchanges , cf . fig . 1
. the transition matrix element of the process @xmath103
\left[\bar{u}(f ) p_b v(\bar{f})\right ] + \frac{v_{ab}}{s}\left[\bar{v}(\mu^+ ) \gamma_\mu p_a u(\mu^-)\right ] \left[\bar{u}(f)\gamma^\mu p_b v(\bar{f})\right ] \ , , \label{eq : ff production amplitude}\end{aligned}\ ] ] can be expressed in terms of the scalar and vector bilinear charges , @xmath104 and @xmath105 , classified according to the chiralities @xmath106 for the right/left handed chiralities of the associated muon and produced fermion currents , respectively . here
, the scalar bilinear charges are given by : @xmath107 \left[s^f_k -ib\ , \xi_\beta\ , o_{1k}\right ] -\frac{e^2m_\mu m_f}{4m^2_w s^2_w}\ , d_z(s)\ , t_{f3}\ , ab\,,\end{aligned}\ ] ] where @xmath108 and @xmath109 are the initial muon and final fermion yukawa couplings , respectively , and @xmath110 & & s^f_k = \left\{\begin{array}{l } o_{2k } \\
o_{3k } \end{array}\right . , \qquad\quad \xi_\beta = \left\{\begin{array}{ll } s_\beta & { \rm for}\ \ f = l\,\,{\rm or}\,\,d \\
c_\beta & { \rm for}\ \ f = u\ , \end{array}\right .
\,.\end{aligned}\ ] ] on the other hand , the vector bilinear charges @xmath105 have the contributions only from the @xmath85 and @xmath57 boson exchanges : @xmath111\,,\end{aligned}\ ] ] where @xmath112 is the electric charge of the fermion , @xmath113 denote the right / left handed couplings of the @xmath57 boson to the fermions , respectively : @xmath114 the vector bilinear charges @xmath105 are real in the approximation of neglecting the @xmath57 boson width @xmath115 , which is valid for the c.m .
energy away from the @xmath57 boson pole .
+ ( 300,100)(0,0 ) ( 15,10)(55,90 ) ( 55,90)(285,90 ) ( 285,90)(245,10 ) ( 245,10)(15,10 ) ( -3,50)[]@xmath95 ( 15,50)(145,50 ) ( 145,50)(140,54 ) ( 145,50)(140,47 ) ( 298,50)[]@xmath96 ( 155,50)(285,50 ) ( 155,50)(160,54 ) ( 155,50)(160,47 ) ( 150,50)3 ( 209,109)[]@xmath98 ( 150,50)(200,100 ) ( 200,100)(200,95 ) ( 200,100)(197,100 ) ( 93,-7)[]@xmath99 ( 150,50)(100,0 ) ( 100,0)(105,0 ) ( 100,0)(100,5 ) ( 90,65)[]@xmath116 ( 75,50)(20,60,90 ) ( 232,80)[]@xmath117 ( 235,50)(20,60,135 ) ( 62,79)[]@xmath118 ( 75,50)(95,90)3 ( 195,79)[]@xmath119 ( 235,50)(255,90)3 ( 190,62)[]@xmath120 ( 150,50)(30,0,45 ) ( 75,50)(75,80 ) ( 75,80)(71,77 ) ( 75,80)(79,77 ) ( 235,50)(205,80 ) ( 205,80)(210,80 ) ( 205,80)(205,75 ) + * figure 2 * : _ the schematic description of the production plane with the scattering angles @xmath120 and @xmath121 as well as the transverse polarization vectors @xmath122 and @xmath123 with the azimuthal angles @xmath116 and @xmath117 with respect to the scattering angle , respectively . _ defining the polar angle of the flight direction of the fermion @xmath98 with respect to the @xmath95 beam direction by @xmath120 ( see fig .
2 ) , the explicit form of the production amplitude ( [ eq : ff production amplitude ] ) can be evaluated in the helicity basis by the 2component spinor technique of ref .
@xcite . denoting the @xmath124 helicity by the first ( second ) index , the @xmath98 and @xmath99 helicities by the remaining two indices ,
then the general form of the helicity amplitude @xmath125 , consisting of a scalar helicity amplitude and a vector helicity amplitude , reads @xmath126 where the scalar helicity amplitude @xmath127 is given by @xmath128 and the vector helicity amplitude @xmath129 by @xmath130 respectively , with @xmath131 , @xmath132 and @xmath133 .
however , the muon mass @xmath134 ( @xmath135 mev ) is extremely small compared to the present experimental mass bound of approximately 100 gev on the lightest higgs boson so that one can safely neglect all the terms involving the kinematical muon mass boson contributions to the scalar bilinear charges @xmath104 , which are also proportional to the muon mass kinematically , should be neglected in the massless muon limit . ] . in the approximation ,
the scalar and vector helicity amplitudes can be written as : + * ( i ) scalar helicity amplitudes : * @xmath136 where @xmath137 . at
asymptotically high energies @xmath138 and @xmath139 , and the other scalar helicity amplitudes vanish .
+ * ( ii ) vector helicity amplitudes : * @xmath140 ( 1+\cos\theta)\,,\nonumber\\ & & \langle + -;-+\rangle_{_v } = + \left[\,\beta_-\,v_{++}+\beta_+\,v_{+-}\right ] ( 1-\cos\theta)\ , , \nonumber\\ & & \langle + -;--\rangle_{_v } = + \sqrt{\beta_+\beta_-}\left(v_{++}+v_{+-}\right)\ , \sin\theta\,,\nonumber\\ & & \langle -+;++\rangle_{_v } = -\sqrt{\beta_+\beta_-}\left(v_{-+}+v_{--}\right)\ , \sin\theta \,,\nonumber\\ & & \langle -+;+-\rangle_{_v } = + \left[\,\beta_+\,v_{-+}+\beta_-\,v_{--}\right ] ( 1-\cos\theta)\,,\nonumber\\ & & \langle -+;-+\rangle_{_v } = -\left[\,\beta_-\,v_{-+}+\beta_+\,v_{--}\right ] ( 1+\cos\theta)\ , , \nonumber\\ & & \langle -+;--\rangle_{_v } = + \sqrt{\beta_+\beta_-}\left(v_{-+}+v_{--}\right)\ , \sin\theta\ , , \end{aligned}\ ] ] and again the other vector helicity amplitudes vanish .
+ the cp transformation leads to the relation among the transition helicity amplitudes : @xmath141 or equivalently for the scalar and vector helicity amplitudes : @xmath142 only the simultaneous presence of the scalar and pseudoscalar couplings can lead to cp violation as can be explicitly checked in eqs .
( [ eq : scalar amplitude ] ) and ( [ eq : vector amplitude ] ) .
one additional useful classification is provided by the so called naive " time reversal @xmath8 ; under the cp@xmath8 transformations the helicity amplitudes are transformed as follows : @xmath143 we note that it is crucial to have finite @xmath57 or higgs boson widths for cp@xmath8 violation . in this light , it is very useful to take into account the cp and cp@xmath8 properties of any physical observable simultaneously so as to investigate not only cp violation itself but also the overlapping of any pair of three neutral higgs boson resonances effectively .
+ the matrix element squared for general ( longitudinal or transverse ) beam polarization can be computed either using standard trace techniques ( employing general spin projection operators ) , or from the helicity amplitudes by a suitable rotation @xcite from the helicity basis to a general spin basis . in the former case ,
neglecting the muon mass in the spin projection operators we can obtain the following approximated form for the @xmath144 projection operators @xmath145 equivalently in the helicity basis the polarization weighted matrix element squared is given by @xmath146\ , , \label{eq : trace}\end{aligned}\ ] ] where @xmath147 ( @xmath148 ) denotes the helicity amplitude for any given production process @xmath149 and the @xmath150 matrices @xmath151 are the polarization density matrices for the initial @xmath144 beams : @xmath152 p_t\,{\rm e}^{i\alpha } & 1-p_l \end{array}\right)\,,\qquad \rho^+=\frac{1}{2}\left(\begin{array}{cc } 1+\bar{p}_l & -\bar{p}_t\,{\rm e}^{i\bar{\alpha } } \\[1 mm ] -\bar{p}_t\,{\rm e}^{-i\bar{\alpha } } & 1-\bar{p}_l \end{array}\right)\,.\end{aligned}\ ] ] here , @xmath153 and @xmath154 are the longitudinal polarizations of the @xmath95 and @xmath96 beams , while @xmath122 and @xmath123 are the degrees of transverse polarization with @xmath116 and @xmath117 being the azimuthal angles between the transverse polarization vectors and the momentum vector of @xmath98 as shown in fig . 2 .
+ applying the projection operators ( [ eq : projection ] ) or / and evaluating the trace ( [ eq : trace ] ) leads to the polarized matrix element squared of the form @xmath155 & & \hskip 0.7 cm + \,(p_t\cos\alpha-\bar{p}_t\cos\bar{\alpha})\ , c_7 + \,(p_t\sin\alpha-\bar{p}_t\sin\bar{\alpha})\ , c_8 \nonumber \\[1 mm ] & & \hskip 0.7 cm + \,(p_l\bar{p}_t\cos\bar{\alpha}+\bar{p}_l p_t\cos\alpha)\,\ , c_9 \,+\,(p_l\bar{p}_t\sin\bar{\alpha}+\bar{p}_l p_t\sin\alpha)\,\ , c_{10 } \nonumber\\[1 mm ] & & \hskip 0.7 cm + \,(p_l\bar{p}_t\cos\bar{\alpha}-\bar{p}_l p_t\cos\alpha)\ , c_{11 } + \,(p_l\bar{p}_t\sin\bar{\alpha}-\bar{p}_l p_t\sin\alpha)\ , c_{12 } \nonumber \\[1 mm ] & & \hskip 0.7 cm + \,p_t\bar{p}_t\left[\,\cos(\alpha+\bar{\alpha})\ , c_{13 } + \sin(\alpha+\bar{\alpha})\ , c_{14}\right]\nonumber\\[1 mm ] & & \hskip 0.7 cm + \,p_t\bar{p}_t\left[\,\cos(\alpha-\bar{\alpha})\ , c_{15 } + \sin(\alpha-\bar{\alpha})\ , c_{16}\right ] , \label{eq : polarized cross section}\end{aligned}\ ] ] where the coefficients @xmath156 ( @xmath157 - @xmath158 ) are defined in terms of the helicity amplitudes by @xmath159 , \ \ c_2=\frac{1}{4}\sum_{\lambda,\bar{\lambda}=\pm } \left[|\langle + -;\lambda\bar{\lambda}\rangle|^2 - |\langle -+;\lambda\bar{\lambda}\rangle|^2\right ] , \nonumber\\ & & c_3=\frac{1}{4}\sum_{\lambda,\bar{\lambda}=\pm } \left[|\langle + + ; \lambda\bar{\lambda}\rangle|^2 + |\langle --;\lambda\bar{\lambda}\rangle|^2\right ] , \ \ c_4=\frac{1}{4}\sum_{\lambda,\bar{\lambda}=\pm } \left[|\langle + + ; \lambda\bar{\lambda}\rangle|^2 - |\langle --;\lambda\bar{\lambda}\rangle|^2\right ] , \nonumber\\ & & c_5=\frac{1}{4}\,{\re { \rm e}}\sum_{\lambda,\bar{\lambda}=\pm } \left(\langle + + ; \lambda\bar{\lambda}\rangle - \langle --;\lambda\bar{\lambda}\rangle\right ) \left(\langle -+;\lambda\bar{\lambda}\rangle - \langle + -;\lambda\bar{\lambda}\rangle\right)^ * , \nonumber\\ & & c_6=\frac{1}{4}\,{\im { \rm m}}\sum_{\lambda,\bar{\lambda}=\pm } \left(\langle + + ; \lambda\bar{\lambda}\rangle - \langle --;\lambda\bar{\lambda}\rangle\right ) \left(\langle
-+;\lambda\bar{\lambda}\rangle + \langle + -;\lambda\bar{\lambda}\rangle\right)^ * , \nonumber\\ & & c_7=\frac{1}{4}\,{\re { \rm e}}\sum_{\lambda,\bar{\lambda}=\pm } \left(\langle + + ; \lambda\bar{\lambda}\rangle + \langle
--;\lambda\bar{\lambda}\rangle\right ) \left(\langle -+;\lambda\bar{\lambda}\rangle + \langle + -;\lambda\bar{\lambda}\rangle\right)^ * , \nonumber\\ & & c_8=\frac{1}{4}\,{\im { \rm m}}\sum_{\lambda,\bar{\lambda}=\pm } \left(\langle + + ; \lambda\bar{\lambda}\rangle + \langle --;\lambda\bar{\lambda}\rangle\right ) \left(\langle -+;\lambda\bar{\lambda}\rangle - \langle + -;\lambda\bar{\lambda}\rangle\right)^ * , \nonumber\\ & & c_9=\frac{1}{4}\,{\re { \rm e}}\sum_{\lambda,\bar{\lambda}=\pm } \left(\langle + + ; \lambda\bar{\lambda}\rangle + \langle --;\lambda\bar{\lambda}\rangle\right ) \left(\langle -+;\lambda\bar{\lambda}\rangle - \langle + -;\lambda\bar{\lambda}\rangle\right)^ * , \nonumber\\ & & c_{10}=\frac{1}{4}\,{\im { \rm m}}\sum_{\lambda,\bar{\lambda}=\pm } \left(\langle + + ; \lambda\bar{\lambda}\rangle + \langle
--;\lambda\bar{\lambda}\rangle\right ) \left(\langle -+;\lambda\bar{\lambda}\rangle + \langle + -;\lambda\bar{\lambda}\rangle\right)^ * , \nonumber\\ & & c_{11}=-\frac{1}{4}\,{\re { \rm e}}\sum_{\lambda,\bar{\lambda}=\pm } \left(\langle + + ; \lambda\bar{\lambda}\rangle - \langle --;\lambda\bar{\lambda}\rangle\right ) \left(\langle
-+;\lambda\bar{\lambda}\rangle + \langle + -;\lambda\bar{\lambda}\rangle\right)^ * , \nonumber\\ & & c_{12}=\frac{1}{4}\,{\im { \rm m}}\sum_{\lambda,\bar{\lambda}=\pm } \left(\langle + + ; \lambda\bar{\lambda}\rangle - \langle --;\lambda\bar{\lambda}\rangle\right ) \left(\langle + -;\lambda\bar{\lambda}\rangle
- \langle -+;\lambda\bar{\lambda}\rangle\right)^ * , \nonumber\\ & & c_{13}=-\frac{1}{2}\,{\re { \rm e}}\sum_{\lambda,\bar{\lambda}=\pm }
\left[\langle -+;\lambda\bar{\lambda}\rangle \langle + -;\lambda\bar{\lambda}\rangle^*\right],\ \ c_{14}=\frac{1}{2}\,{\im { \rm
m}}\sum_{\lambda,\bar{\lambda}=\pm } \left[\langle -+;\lambda\bar{\lambda}\rangle \langle + -;\lambda\bar{\lambda}\rangle^*\right ] , \nonumber\\ & & c_{15}=-\frac{1}{2}\,{\re { \rm e}}\sum_{\lambda,\bar{\lambda}=\pm } \left[\langle --;\lambda\bar{\lambda}\rangle \langle + + ; \lambda\bar{\lambda}\rangle^*\right],\ \ c_{16}=\frac{1}{2}\,{\im
{ \rm m}}\sum_{\lambda,\bar{\lambda}=\pm } \left[\langle --;\lambda\bar{\lambda}\rangle \langle + + ; \lambda\bar{\lambda}\rangle^*\right ] .
\end{aligned}\ ] ] + @xmath160 + the production cross section is then given in terms of the distribution @xmath161 in eq .
( [ eq : polarized cross section ] ) by @xmath162 0.2 cm where @xmath163 is the color factor of the final fermion ; 3 for the top or bottom quark and 1 for the tau lepton .
the dependence of the distribution @xmath161 on the azimuthal angle @xmath121 of the production plane is encoded in the angles @xmath116 and @xmath117 .
if the azimuthal angle @xmath121 is measured with respect to the direction of the @xmath95 transverse polarization vector , the @xmath121 dependence can be exhibited explicitly by taking @xmath164 where @xmath165 is the _ rotational invariant _ difference @xmath166 of the azimuthal angles of the @xmath96 and @xmath95 transverse polarization vectors with respect to the production plane .
clearly , only the six observables @xmath167 can be measured independently of the azimuthal angle , but the other ten observables , in particular , the observables involving the sv correlations , require the reconstruction of the production plane . +
the cp transformation on the polarization vectors of the initial muon beams corresponds to the simultaneous exchanges : @xmath168 the cp relation ( [ eq : polarization cp ] ) of the polarization vectors leads to the classification that the first 3 terms and the distributions @xmath169 in eq .
( [ eq : polarized cross section ] ) are cp even while the other six distributions @xmath170 are cp odd . among the cp
odd observables , the three observables @xmath171 are cp@xmath8odd and the other three observables @xmath172 are cp@xmath8even . here , it will be noteworthy to emphasize again that one crucial requirement for having a large cp@xmath8odd observable is the presence of so called cp preserving re - scattering phases which can be provided by the higgs boson propagators with relatively large widths in the present work .
let us now express the coefficients @xmath173 ( @xmath174 to 16 ) in terms of the scalar and vector bilinear and quartic charges @xcite as well as the newly introduced notations : + @xmath175 which are to be employed for the interference between the scalar and vector contributions .
it is worthwhile to note that both @xmath176 and @xmath177 vanish in the cp invariant theory .
the sixteen polarization distributions can be then classified as follows : + * ( i ) scalar scalar ( ss ) correlations : * @xmath178\,=\frac{1}{2}\left[(1+\beta^2)\ , s_1 -(1-\beta^2)\ , s_2\right],\nonumber\\ & & c_{\,4}[--]\ , = \frac{1}{2}\left[(1+\beta^2)\ , s'_1 -(1-\beta^2)\ , s'_2\right],\nonumber\\ & & c_{15}[++]=\frac{1}{2}\left[(1+\beta^2)\ , s_4 - ( 1-\beta^2)\,s_5\right],\nonumber\\ & & c_{16}[-+]=\frac{1}{2}\left[(1+\beta^2)\ , s'_4 - ( 1-\beta^2)\,s'_5\right],\end{aligned}\ ] ] where the first and second signatures in the square brackets are for the cp and cp@xmath8 parities of the corresponding observable , respectively , and the relevant scalar quartic charges are defined as @xmath179\,,\nonumber\\ & & s'_1=\frac{1}{4 } \left[|s_{++}|^2+|s_{+-}|^2-|s_{-+}|^2-|s_{--}|^2\right]\,,\nonumber\\ & & s_2=\frac{1}{2}\,{\re { \rm e}}\left[s_{++}s^*_{+-}+s_{--}s^*_{-+}\right]\,,\quad s'_2=\frac{1}{2}\,{\re { \rm e}}\left[s_{++}s^*_{+-}-s_{--}s^*_{-+}\right]\,,\nonumber\\ & & s_4=\frac{1}{2}\,{\re { \rm e}}\left[s_{++}s^*_{-+}+s_{--}s^*_{+-}\right]\,,\quad s'_4=\frac{1}{2}\,{\im { \rm m}}\left[s_{++}s^*_{-+}-s_{--}s^*_{+-}\right]\,,\nonumber\\ & & s_5=\frac{1}{2}\,{\re { \rm e}}\left[s_{++}s^*_{--}+s_{+-}s^*_{-+}\right]\,,\quad s'_5=\frac{1}{2}\,{\im { \rm m}}\left[s_{++}s^*_{--}+ s_{+-}s^*_{-+}\right]\,,\end{aligned}\ ] ] + * ( ii ) vector vector ( vv ) correlations : * @xmath180\ , = ( 1+\beta^2\cos^2\theta)\ , v_1+(1-\beta^2)\ , v_2 + 2\beta\ , v_3\cos\theta,\nonumber\\ & & c_{\,2}[++]\ , = ( 1+\beta^2\cos^2\theta)\ , v'_1+(1-\beta^2)\ , v'_2 + 2\beta\ , v'_3\cos\theta,\nonumber\\ & & c_{13}[++]= \beta^2\sin^2\theta\ , v_4,\nonumber\\ & & c_{14}[+-]= \beta^2\sin^2\theta\ , v'_4,\end{aligned}\ ] ] where the quartic charges @xmath181 and @xmath182 are given by @xmath183\,,\nonumber\\ & & v'_3=\frac{1}{4 } \left[|v_{++}|^2-|v_{+-}|^2+|v_{-+}|^2-|v_{--}|^2\right]\,,\end{aligned}\ ] ] and the other vector quartic charges are defined in the same way as the scalar quartic charges with the notation @xmath184 replaced by @xmath185 everywhere .
we note that the scalar as well as vector quartic charges defined as an imaginary part of the bilinear
charge correlations might be non - vanishing only when there are complex cp violating couplings or / and cp preserving phases like re - scattering phases or finite widths of the intermediate particles .
so , if there are no cp preserving phases , non vanishing values of these quartic charges signal cp violation in the given process .
+ * ( iii ) scalar vector ( sv ) correlations : * @xmath186\ , = + \frac{m_f}{4\sqrt{s}}\,\beta\sin\theta\ , { \re { \rm e}}(w_3\,r^*_1),\quad c_{\,6}[+-]\,\ , = + \frac{m_f}{4\sqrt{s}}\,\beta\sin\theta\ , { \im { \rm m}}(w_1\,r^*_1),\nonumber\\ & & c_{\,7}[--]\ , = -\frac{m_f}{4\sqrt{s}}\,\beta\sin\theta\ , { \re { \rm e}}(w_1\,r_3^ * ) , \quad c_{\,8}[-+]\,\ , = -\frac{m_f}{4\sqrt{s}}\,\beta\sin\theta\ , { \im { \rm m}}(w_3\,r_3^ * ) , \nonumber\\ & & c_{\,9}[--]\ , = + \frac{m_f}{4\sqrt{s}}\,\beta\sin\theta\ , { \re { \rm e}}(w_3\,r_3^ * ) , \quad\ , c_{10}[-+ ] = + \frac{m_f}{4\sqrt{s}}\,\beta\sin\theta\ , { \im { \rm m}}(w_1\,r^*_3),\nonumber\\ & & c_{11}[++ ] = + \frac{m_f}{4\sqrt{s}}\,\beta\sin\theta\ , { \re { \rm e}}(w_1\,r^*_1),\quad c_{12}[+- ] = + \frac{m_f}{4\sqrt{s}}\,\beta\sin\theta\ , { \im { \rm m}}(w_3\,r_1^ * ) . \end{aligned}\ ] ] all the sv correlations are proportional to the mass of the final state fermion divided by the c.m
. energy so that they are strongly suppressed at high energies . in this light , they are sizable only in the production of the top
quark pair because of the largest fermionic mass .
various types of experimental observables by use of the initial muon polarization have been considered in the previous section for determining the cp character of the neutral mssm higgs bosons .
in addition , spin correlations of @xmath187 or @xmath188 in the final state can also probe the cp nature of the higgs boson , independently of or combined with initial beam polarization .
if the production plane or directly the momentum directions of two final fermions are determined , the secondary decays of the primary final state fermions can allow a complete analysis of their spin or helicity directions .
however , even if the production plane is not reconstructed , one can obtain the fermion pair polarization combination by statistically studying decay products from the correlated and polarized @xmath189 .
it is then natural that the production distribution needs to be integrated over the azimuthal angle @xmath121 .
+ the @xmath102channel higgs boson ( spin0 ) exchange populates the equal @xmath190 helicity combinations , i.e. @xmath191 .
this results in the correlations of @xmath189 polarization with @xmath192 by angular momentum conservation , while the sm background channels yield dominantly the @xmath193 or @xmath194 polarization combination .
in addition , the cp transformation interchanges the @xmath195 and @xmath196 helicity configurations . in the light of these two arguments combined with the one in the previous paragraph , it will be valuable to consider the @xmath121independent polarization observable and @xmath57exchange terms corresponding to the polarization combinations @xmath197 and @xmath198 of the initial @xmath6 beams . however , those terms turn out to be zero for the real vector and axial vector couplings in the approximation of neglecting the @xmath57boson width . ] : @xmath199 , \label{eq : difference}\end{aligned}\ ] ] where the distributions @xmath200 ( @xmath174 to 4 ) are defined in terms of the helicity amplitudes as @xmath201=\frac{1}{4}\sum_\lambda\ , \lambda \left[|\langle + + ; \lambda\lambda\rangle|^2 + |\langle --;\lambda\lambda\rangle|^2\right ] , \nonumber\\ & & d_2[++]=\frac{1}{4}\sum_\lambda\ , \lambda \left[|\langle + + ; \lambda\lambda\rangle|^2 - |\langle --;\lambda\lambda\rangle|^2\right ] , \nonumber\\ & & d_3[--]=-\frac{1}{2}\,{\re { \rm e}}\sum_\lambda\ , \lambda \left[\langle --;\lambda\lambda\rangle \langle + + ; \lambda\lambda\rangle^*\right],\nonumber\\ & & d_4[+-]=-\frac{1}{2}\,{\im { \rm m}}\sum_\lambda\ , \lambda \left[\langle --;\lambda\lambda\rangle \langle + + ; \lambda\lambda\rangle^*\right ] , \end{aligned}\ ] ] where the first and second signatures in the square brackets are for the cp and cp@xmath8 parities of the corresponding observable .
the cp@xmath8 parity of each observable plays a crucial role in determining the interference pattern of the higgs bosons among themselves and with the @xmath85 and @xmath57 bosons appearing in the observable ; every cp@xmath8even ( odd ) observable involving only the scalar contributions depends on the real ( imaginary ) part of the combination @xmath202 of each pair of two higgs boson propagators .
the cp@xmath8odd observable @xmath203 involving only the vector contributions is negligible because the @xmath57boson width effect is significantly small for the energy far from the @xmath57boson pole .
when the @xmath57boson width is neglected , all the vector couplings are real . as a result ,
the interference terms between the scalar and vector contributions depends on the real or imaginary parts of the higgs boson propagators @xmath204 itself .
+ for the sake of discussion in the following , let us introduce for the real and imaginary parts the abbreviated notations : @xmath205 \ , = \frac{(s - m^2_{h_k})(s - m^2_{h_l } ) + m_{h_k}m_{h_l}\gamma_{h_k}\gamma_{h_l } } { [ ( s - m^2_{h_k})^2+m^2_{h_k}\gamma^2_{h_k } ] [ ( s - m^2_{h_l})^2+m^2_{h_l}\gamma^2_{h_l } ] } , \nonumber\\ & & { \cal d}_{kl}\equiv{\im { \rm m}}[d_{h_k } d^*_{h_l } ] = \frac{(s - m^2_{h_k})\ , m_{h_l}\gamma_{h_l } -(s - m^2_{h_l})\ , m_{h_k}\gamma_{h_k } } { [ ( s - m^2_{h_k})^2+m^2_{h_k}\gamma^2_{h_k } ] [ ( s - m^2_{h_l})^2+m^2_{h_l}\gamma^2_{h_l}]}.\end{aligned}\ ] ] it is worthwhile to note two points ; ( a ) the denominator reveals a typical two pole structure so that the higgs boson contributions are greatly enhanced at the poles ; ( b ) the numerator of @xmath206 is negative in the middle of two resonances with a mass splitting larger than their typical widths but positive otherwise , while the numerator of @xmath207 is always positive ( negative ) if @xmath208 ( @xmath209 ) and linearly increasing ( decreasing ) if @xmath210 ( @xmath211 ) , respectively .
+ on the other hand , the sv interference terms depend on the real and imaginary parts of each higgs boson propagators @xmath204 , the form of which is given by @xmath212=\frac{s - m^2_{h_k}}{(s - m^2_{h_k})^2+m^2_{h_k}\gamma^2_{h_k } } , \qquad { \im { \rm m}}[d_{h_k}]=-\frac{m_{h_k}\gamma_{h_k}}{(s - m^2_{h_k})^2 + m^2_{h_k}\gamma^2_{h_k}}.\end{aligned}\ ] ] note that the real part changes its sign whenever the c.m .
energy crosses the pole , but the imaginary part is always negative . + combining the coefficients from the mixing matrix elements with those propagator
dependent parts enables us to make a straightforward qualitative understanding of the @xmath213 dependence of each observable .
this will be demonstrated in detail in the following section with a concrete numerical example .
we are now ready to present some numerical results .
it is known that loop induced cp violation in the higgs sector can only be large if both @xmath214 and @xmath215 ( or @xmath216 , if @xmath217 ) are sizable @xcite .
we therefore choose @xmath218 .
for definiteness we will present results only for the @xmath187 mode by taking @xmath219 when probing the region around two heavy higgs boson resonances , but we will give a qualitative description of probing the cp property of the lightest higgs boson .
we take the running yukawa couplings by including the qcd loop effects depending on the gluino mass properly . since for moderate values of @xmath79 the contributions from the ( s)bottom sector are still quite small , our results are not sensitive to @xmath220 and @xmath40
; we therefore fix @xmath221 , although different values for the su(2 ) doublet and singlet soft breaking squark masses , @xmath222 are allowed , and also take equal phases for @xmath41 and @xmath40 .
since we are basically interested in distinguishing the cp non invariant higgs sector from the cp invariant one , we take for the re - phasing invariant phase @xmath223 of @xmath224 two values ; 0 ( cp invariant case ) and @xmath225 ( maximally cp violating case ) . on the other hand ,
we fix for simplicity the other real mass parameters and couplings except for @xmath79 as follows : @xmath226 finally , we take for the top and bottom quark running masses @xmath227 gev and @xmath228 gev . for the given @xmath78
much larger than @xmath229 , two neutral higgs bosons have almost degenerate masses : @xmath230 as a result , a significant overlapping , i.e. interference between two higgs boson resonances is expected .
+ in the previous section , we have listed 16 polarization observables constructed solely by initial muon polarizations under the assumption that the production plane is ( at least statistically ) reconstructed , and 4 additional polarization observables by combining initial muon polarizations and final equal helicity configurations only for the case when the production plane does not have to be explicitly reconstructed .
for some polarization observables it is also important to experimentally identify the electric charges of the produced fermion pair from their decay products .
taking into account all these requirements naturally leads us to the following classification ; for the energy range around the heavy higgs boson resonances the production of a top quark pair is recommended to be used ( when the heavy higgs boson masses are larger than twice the top
quark mass ) , and for the energy close to the lightest higgs mass , the tau lepton or bottom quark pairs are recommended to be considered decay mode of the lightest higgs boson , even if it has the largest branching fraction , is not useful for certain observables because of the difficulty in determining the polarization and electric charge of the produced bottom quark and anti
quark experimentally . ]
+ in the mssm two heavy higgs bosons are ( almost ) degenerate with a mass splitting comparable to or less than their widths for @xmath231 as in the case of the parameter set ( [ eq : parameter set ] ) , and one of them is cp even and the other cp odd in the cp invariant theory . however , the cp violating neutral higgs boson mixing can cause a mass splitting larger than their typical widths and it can lead to several significant non vanishing cp odd observables , in particular , in the @xmath187 mode . in this section , we investigate in detail the possibility of probing those aspects through the observables classified in the previous section by considering the cases without and with direct reconstruction of the production plane separately .
if the production plane is not directly reconstructed but only the helicities of the produced top quark and anti - quark are statistically determined , all the interference between the scalar and vector contributions is averaged away over the azimuthal angle @xmath121 and only the observables , @xmath232 can be reconstructed .
+ figures [ fig : fig1 ] and [ fig : fig2 ] exhibit six independent observables of definite cp and cp@xmath8 parities near the region of the heavy higgs boson resonances ; @xmath233 , \ \
{ \rm ( b)}\ , \sigma_{_{lr}}[\,++ ] , \ \ { \rm ( c)/(d)}\,\ , \frac{\sigma_{_{rr}}\pm\sigma_{_{ll}}}{2}[\,\pm\pm ] , \ \ { \rm ( e)}\,\ , \sigma_{\vert}[++ ] , \ \ { \rm ( f)}\,\ , \sigma_{\bot}[-+],\end{aligned}\ ] ] for @xmath219 in figs .
[ fig : fig1 ] and [ fig : fig2 ] , respectively .
the observables obtained by controlling solely the muon and anti
muon polarizations are given by @xmath234\,,\nonumber\\ \sigma_{_{rr / ll } } & = & \frac{n_c\ , \beta}{8\,\pi\,s}\ , \,\left[\ , c_3\pm c_4\,\right]\ , , \nonumber\\ \sigma_{\,\vert/\bot}&= & \frac{n_c\ , \beta}{16\,\pi\,s}\,(\pm)\,c_{15/16}\ , .
\label{eq : observable1}\end{aligned}\ ] ] as shown in the frames ( a ) and ( b ) the spin1 ( lr ) and ( rl ) observables @xmath235 and @xmath236 in both figures are almost constant and the former is almost twice larger than the latter , showing the large left
right asymmetry . on the other hand , the average of the observables @xmath237 and @xmath238 showing the higgs boson contributions are peaked on each heavy higgs boson resonance and the size at each pole is comparable to that of the spin1 observable . in the cp invariant case ( @xmath239 ) with @xmath240 the parameter set ( [ eq : parameter set ] ) with @xmath240 accidentally generates two extremely degenerate higgs bosons as denoted by a single resonance line in the frame ( c ) as well as a single oscillating pattern in the frame ( e ) for @xmath241 of fig .
[ fig : fig1 ] unlike the case for @xmath242 with a finite mass splitting of about 3 gev .
the cp odd observables @xmath243 and @xmath244 are identically zero for both @xmath240 and @xmath245 as they ought to be . on the other hand , in the cp non invariant case a large mass splitting between two higgs boson resonances
is developed for @xmath240 , implying a large cp violating mixing between two higgs bosons , while the splitting in the case of @xmath242 is not so much enlarged .
a more quantitative understanding about the mass splitting can be obtained from eq .
( [ eq : higgs masses ] ) . in this case , the cp
odd observables are non vanishing .
note that observable @xmath244 are quite sizable on the poles , especially , in the case of @xmath240 .
+ the observable @xmath246 in the frame ( d ) is cp@xmath8odd , involving the propagator combination @xmath207 , but the observables @xmath241 and @xmath244 in the frames ( e ) and ( f ) are cp@xmath8even , involving the propagator combination @xmath206 . examining the analytic structure of those observables and taking into account the parameter set ( [ eq : parameter set ] ) , we find that ( i ) the coefficient related with the cp odd ( cp even ) resonance pole is negative ( positive ) , respectively ; ( ii ) for @xmath240 the lighter ( heavier ) resonance of the heavy higgs bosons is dominantly cp odd ( cp even ) in both the cp invariant and non invariant cases , respectively ; ( iii ) for @xmath242 the lighter resonance is dominantly cp even ( cp odd ) and the heavier one cp
odd ( cp even ) in the cp invariant ( non invariant ) case for @xmath239 ( @xmath225 ) , respectively .
combining all these features leads to a symmetric resonance pattern in the frame ( d ) and anti
symmetric patterns in the frames ( e ) and ( f ) in both figures and a remarkable sign flipping at each resonance pole in the frame ( e ) of fig .
[ fig : fig2 ] . + we present in fig .
[ fig : fig3 ] three observables is strongly suppressed in the present case due to the orthogonality of the neutral higgs mixing matrix @xmath75 : @xmath247 is very close to unity even for @xmath240 such that the product sum @xmath248 is approximately @xmath249 , i.e. vanishing in the present case because of @xmath250 forced by the antisymmetric combination @xmath251 with @xmath252 for @xmath253 . ] related with the difference between the cross sections with the @xmath195 and @xmath196 helicity configurations : @xmath254,\quad \delta_{\bot}[+-],\quad \delta_{\vert}[--]\ , , \label{eq : observable22}\end{aligned}\ ] ] near the heavy higgs boson resonances , the explicit form of which is obtained by integrating the distributions @xmath255 over the production angles as @xmath256\ , , \quad \delta_{\,\vert/\bot}= \frac{n_c\ , \beta}{16\,\pi\,s}\,d_{3/4}. \label{eq : observable2}\end{aligned}\ ] ] it should be noted that every observable @xmath200 ( @xmath174 to 4 ) involves the antisymmetric combination @xmath257 forcing @xmath250 ( @xmath258 ) . because two heavy higgs bosons exhibit almost a typical two state mixing for the parameter set ( [ eq : parameter set ] ) , there exist simply one coupling combination formed by the mixing matrix elements for each observable . as a result ,
the @xmath213 dependence of each observable is ( almost ) completely determined by that of the propagator combinations @xmath259 or @xmath260 , depending on whether the observable is cp@xmath8even or cp@xmath8odd . as two higgs bosons
are extremely degenerate in the cp invariant case with @xmath240 a single resonance peak ( solid line ) is shown in the first two upper frames of fig .
[ fig : fig3 ] .
however , we note in the upper part of fig .
[ fig : fig3 ] that ( i ) in the cp non invariant case ( @xmath261 ) a large mass splitting is developed ; and ( ii ) the cp@xmath8even observable @xmath262 changes its sign at each resonance pole as @xmath259 does , while the other two cp@xmath8odd observables have a two pole structure of the same sign . on the other hand ,
the mass splitting between two heavy higgs bosons for @xmath242 is not so different in the cp invariant and non invariant cases .
as explained in the case of @xmath240 , one can see a sign change at each resonance pole in the cp@xmath8even observable and a typical two pole structure in the cp@xmath8odd observables .
it is also noteworthy that the observable @xmath263 is negative ( positive ) for the cp invariant ( non invariant ) case . finally , the typical size of the cp
odd observables is very small compared to that of the cp even observables .
all the terms involving interference between the scalar and vector contributions requires a reasonable identification of the production plane and they are proportional to the mass of the final fermion .
the @xmath188 mode with two final neutrinos escaping detection can be used at high energies where the produced tau leptons are very energetic and therefore their decay products fly along the original tau direction to a very good approximation
. nevertheless , the small mass leads to very small interference effects . on the other hand , the production plane in the @xmath187 mode can be determined without big difficulty and the top
quark mass is almost 100 times larger than the tau lepton mass . in this light
, we will concentrate on the @xmath187 mode which could generate significant interference effects .
+ among the eight interference terms the four terms @xmath264 are cp
odd and the other four terms @xmath265 are cp even .
it is noteworthy that all the cp odd terms are proportional to @xmath176 while all the cp even observables to @xmath266 ; as a matter of fact the cp properties of the interference terms originates from those of @xmath266 and @xmath176 as can be worked out by their explicit form for the @xmath187 mode : @xmath267 we note that ( i ) @xmath266 describes the cp preserving mixing between two cp even states , @xmath63 and @xmath60 , so that the relevant observables can be used to select the dominantly cp even higgs boson states ; ( ii ) the coefficients @xmath268 and @xmath269 have an equal sign while the coefficients @xmath270 and @xmath271 have an opposite sign ; ( iii ) @xmath176 the cp violating mixing between @xmath60 and the cp
odd state @xmath61 .
in addition , the coefficients @xmath272 involving vector bilinear charges are ( almost ) real to a very good approximation so that the sv interference terms depends on the real and imaginary parts of each higgs boson propagator @xmath204 itself .
+ certainly we need to apply an appropriate @xmath121dependent weight function to extract each interference term ; for example , ( @xmath273 ) with @xmath274 taken can be used as a weight function to extract @xmath275 .
the extraction efficiency should be determined with the detailed information on various experimental machine parameters .
so , we will not provide any further detailed procedure to extract those observables , but concentrate ourselves on the observables @xmath173 themselves .
let us define the observables @xmath276 ( @xmath277 to @xmath278 ) as : @xmath279 so that @xmath276 has the same cp property as @xmath173 .
+ taking the maximal cp violating phase @xmath280 , we present in fig . [
fig : fig4 ] the cp odd observables @xmath281,\sigma_8[-+],\sigma_9[--],\sigma_{10}[-+]\}$ ] for the @xmath187 mode near the heavy higgs boson resonances for @xmath240 ( solid line ) and @xmath242 ( dashed line ) with the susy parameter set ( [ eq : parameter set ] ) .
the cp@xmath8even observables @xmath282 and @xmath283 changes their sign at each pole as the real part of each higgs boson propagator changes its sign . on the other hand ,
the fact that the coefficients @xmath284 and @xmath271 have an opposite sign is responsible for the sign flipping pattern at the poles , appearing in the left frames of fig .
[ fig : fig4 ] . on the whole , among those cp odd observables @xmath285 and @xmath283 are sizable while the other two observables are small in size .
+ figure [ fig : fig5 ] shows the cp even observables @xmath286,\sigma_6[+-],\sigma_{11}[++],\sigma_{12}[+-]\right\}$ ] for the @xmath187 mode near the heavy higgs boson resonances for @xmath239 ( solid line ) and @xmath280 ( dashed line ) with the susy parameter set ( [ eq : parameter set ] ) . in this case , we take @xmath240 causing a larger cp violating higgs boson mixing than @xmath242 . as explained before , the observables dependent on the coefficients @xmath287 single the dominantly cp even states out . as can be seen in every frame of fig .
[ fig : fig5 ] the heavier higgs boson of two heavy higgs bosons is ( dominantly ) cp even in the case of @xmath239 and @xmath225 , respectively .
the cp@xmath8even observables @xmath288 and @xmath289 show a sign change at the higgs boson resonance pole , while the cp@xmath8odd observables show a typical peak .
according to a detailed analysis @xcite of higgs boson decays , the width of the lightest higgs boson @xmath290 is of the order of mev for intermediate @xmath79 , which is smaller than a typical muon energy resolution .
the lightest higgs boson decays dominantly into @xmath291 and sub - dominantly into @xmath188 with the ( almost ) fixed branching fractions and the interference between the higgs exchange and the @xmath85 and @xmath57boson exchanges is negligible because of the small @xmath93 and @xmath94 masses compared to the higgs boson mass , even if the muon beam polarization is employed .
furthermore , @xmath292 at the tree level so that any further information on the cp property of the lightest higgs boson will not be obtained by measuring the difference between the final @xmath195 and @xmath196 helicity configurations helicities is , however , quite useful to reduce the continuum backgrounds from the @xmath85 and @xmath57 exchanges @xcite . ] . +
a convolution of the higgs boson production cross section with the beam energy distribution tends to reduce the production rate as given by the expression @xmath293 where @xmath294 is the muon beam energy resolution and the distribution @xmath295 is given by @xmath296\,.\end{aligned}\ ] ] every @xmath173 is proportional to the second power of @xmath297 and factored into the production and decay parts on the lightest higgs boson resonance pole .
so , apart from the factor depending on the beam energy resolution , the cp property of the lightest higgs boson can be directly investigated through the higgs boson resonance production by polarized muon and anti
muon collisions as well as through final fermion spin spin correlations with high efficiencies .
the reduction factor in eq .
( [ eq : lightest higgs ] ) clearly shows the importance of having a very good beam energy resolution in order to obtain the spin0 higgs exchange signal events clearly distinguished from the spin1 @xmath298exchange background events . since
several works @xcite along this line have been already done , we close this section to referring to the relevant recent works .
we have performed a systematic investigation of the production of a third generation fermion pair in polarized @xmath6 collisions so as to probe the explicit cp violation in the mssm higgs sector , induced radiatively by soft trilinear interactions related to squarks of the third generation .
we have classified all the observables for probing the cp property of the higgs bosons constructed by the initial muon beam polarization along with the final fermions of no polarization and of equal helicity , respectively .
the polarization observables have turned out to allow for complete determination of the cp property of the higgs bosons .
furthermore , we have found that the interference between the spin0 higgs boson and spin1 @xmath298 contributions can provide a powerful and independent means for the determination of the cp property of two heavy higgs boson in the top
quark pair production with the c.m .
energy near the ( almost ) degenerate heavy higgs boson resonances . on the other hand
, there is no sizable interference between the lightest higgs boson and spin1 @xmath298 contributions so that the cp property of the lightest higgs boson can be optimally measured on its pole by using the initial muon beam polarization or the final fermion spin
spin correlations with high efficiencies while keeping a very good beam energy resolution .
+ in conclusion , the third generation fermion
pair production in @xmath6 collisions equipped with initial muon beam polarization and final
fermion spin correlations provides a powerful probe of the cp property of the higgs bosons in the mssm with explicit cp violation .
s.y.c wishes to acknowledge financial support of the 1997 sughak program of the korea research foundation .
e.a thanks kias for the great hospitality extended to her while this work was being performed . j.h .
christenson , j.w .
cronin , v.l .
fitch and r. turlay , phys .
* 13 * , 138 ( 1964 ) ; for reviews , see , for example , p.k .
kabir , _ the cp puzzle _ , ( academic press , london and new york , 1968 ) ; w. grimus , fortschr . phys . * 36 * , 201 ( 1988 ) ; e.a .
paschos and u. trke , phys . rep . *
178 * , 147 ( 1989 ) ; b. winstein and l. wolfenstein , rev .
65 * , 1113 ( 1993 ) ; g.d . barr _
et al . _ ,
na31 collaboration , phys .
b317 * , 233 ( 1993 ) ; a. alavi
et al . _ ,
ktev collaboration , phys .
83 * , 22 ( 1999 ) ; v. fanti _ et al .
_ , na48 collaboration , phys . lett . *
b465 * , 335 ( 1999 ) .
t. ibrahim and p. nath , phys .
b * 418 * , 98 ( 1998 ) ; phys .
d * 57 * , 478 ( 1998 ) ; d * 58 * , 019901 ( 1998 ) ; t. falk and k.a .
olive , phys .
b * 439 * , 71 ( 1998 ) ; m. brhlik , g.j .
good and g.l .
kane , _ ibid .
_ d * 59 * , 115004 ( 1999 ) ; s. pokorski , j. rosiek and c.a .
savoy , nucl .
* b570 * , 81 ( 2000 ) .
a. pilaftsis , phys .
lett . * 77 * , 4996 ( 1997 ) , hep
ph/9603328 ; k.s .
babu , c. kolda , j. march russell and f. wilczek , phys . rev . * d59 * , 016004 ( 1999 ) , hep ph/9804355 ; s.y . choi and m. drees ,
phys . rev .
lett . * 81 * , 5509 ( 1998 ) , hep
ph/9808377 ; j.f .
gunion and j. pliszka , phys . lett .
* b444 * , 136 ( 1998 ) , hep ph/9809306 ; c.a .
boe , o.m .
ogreid , p. osland and j. zhang , eur .
j. * c9 * , 413 ( 1999 ) , hep
ph/9811505 ; b. grzadkowski , j.f .
gunion and j. kalinowski , phys . rev .
* d60 * , 075011 ( 1999 ) , hep ph/9902308 .
v. barger , m.s .
berger , j.f .
gunion , and t. han , phys .
lett . * 75 * , 1462 ( 1995 ) ; j.f .
gunion , in _ proceedings of the 5th international conference on physics beyond the standard model , balholm , norway , 1997 _ , edited by g. eigen , p. osland and b. stugu ( aip , woodbury , new york , 1997 ) , p. 234 .
s. dimopoulos and d. sutter , nucl .
* b 452 * ( 1995 ) 496 ; h. haber , proceedings of the 5th international conference on supersymmetries in physics ( susy97 ) , may 1997 , ed .
m. cveti and p. langacker , hep - ph/9709450 .
s. coleman and e. weinberg , phys . rev . *
d7 * , 1888 ( 1973 ) ; y. okada , m. yamaguchi and t. yanagida , prog .
. phys . * 85 * , 1 ( 1991 ) ; phys . lett . * b262 * , 54 ( 1991 ) ; j. ellis , g. ridolfi and f. zwirner , phys . lett . *
b257 * , 83 ( 1991 ) ; * b262 * , 477 ( 1991 ) . | we perform a systematic study of the production of a third generation fermion pair , @xmath0 for @xmath1 , and @xmath2 in the minimal supersymmetric standard model ( mssm ) with explicit cp violation , which is induced radiatively by soft trilinear interactions related to squarks of the third generation .
we classify all the observables for probing the cp property of the higgs bosons constructed by the initial muon beam polarization along with the unpolarized final fermions and with the final fermion polarization configuration of equal helicity , respectively .
the observables allow for complete determination of cp property of the neutral higgs bosons .
the interference between the higgs boson and gauge boson contributions also could provide a powerful method for the determination of the cp property of two heavy higgs bosons in the top
quark pair production near the energy region of the higgs boson resonances . for the lightest higgs boson mass there is no sizable interference between scalar and vector contributions for the determination of the cp property of the lightest higgs boson .
we give a detailed numerical analysis to show how the radiatively induced cp violation in the higgs sector of the mssm can be measured .
16.3 cm 20.5 cm -1 cm -0.5 cm + kias
p00021 + ocha pp156 + hep
ph/0005118 + + * probing the mssm higgs boson sector with explicit + cp violation through third generation fermion pair + production at muon colliders * + eri asakawa@xmath3 , s.y .
choi@xmath4 and jae sik lee@xmath5 @xmath3_graduate school of humanities and sciences , ochanomizu university , + 1 - 1 otsuka 2-chome , bunkyo , tokyo 112 - 8610 , japan _
+ @xmath4_department of physics , chonbuk national university , chonju 561756 , korea _ + @xmath5_korea institute for advanced study ,
seoul 130012 , korea _
pacs number(s ) : 12.60.jv , 11.30.er , 13.10.+q , 14.80.cp |
at a distance of @xmath7 pc ( menten et al . 2007 ; kim et al .
2008 ) , radio `` '' in the kleinmann - low ( kl ) nebula of orion is believed to be the nearest example of a massive young stellar object ( yso ) . is highly embedded and has no optical or infrared counterpart ;
at wavelengths of 8 and 22@xmath8 m greenhill et al .
( 2004a ) estimated continuum optical depths of @xmath9300 .
however , high - resolution 7-mm radio continuum observations have revealed an extended ( @xmath370 au across ) source that reid et al .
( 2007 ) proposed may be an ionized accretion disk surrounding an early b - type star .
in addition to its proximity , orion kl has the distinction of being one of only three star - forming regions known to exhibit sio maser emission ( hasegawa et al .
1986 ; zapata et al .
2009 ) , and these masers were definitively linked with by menten & reid ( 1995 ) .
sio masers offer key advantages for studying stellar sources in that they arise at small radii ( @xmath10 k , @xmath11 @xmath12 ) , are unaffected by extinction from dust and neutral gas , and can sample gas kinematics via emission line doppler shifts and proper motions .
moreover , their high brightness temperature makes them observable with extraordinarily high angular resolution using very long baseline interferometry ( vlbi ) .
indeed , previous vlbi observations have established that the sio masers lie at projected distances of @xmath320 - 100 au from ( greenhill et al .
1998 , 2004b ; doeleman et al . 1999 , 2004
; kim et al . 2008 ) .
these scales are of considerable interest , since they correspond to regions where accretion is expected to occur and where winds and outflows from ysos are expected to be launched and collimated . at present
, little is known about the dynamics of these regions in higher mass ysos ( @xmath13 ) owing to the dearth of direct observations with the requisite resolution .
consequently , persistent gaps in our knowledge of the formation process for ob stars have remained ( e.g. , bally & zinnecker 2005 ; zinnecker & yorke 2007 ; tan 2008 ) .
we have used the national radio astronomy observatory s very long baseline array ( vlba ) to monitor the sio maser emission surrounding at monthly intervals over several years ( see also greenhill et al .
our images have higher dynamic range and higher cadence than any observations of the masers to date . here
we showcase our results in the form of 19-epoch movies spanning 21 months .
these movies chronicle for the first time the kinematics and evolution of the molecular material @xmath320 - 100 au from a massive yso .
this is part of a series of papers examining on 10 - 1000 au scales based on vlba monitoring and complementary observations of other maser and thermal lines using the very large array ( vla ) and the green bank telescope ( e.g. , goddi et al . 2009a , b and in preparation ; greenhill et al .
, in preparation ; matthews et al . , in preparation ) .
together , these data are allowing us to forge a comprehensive new picture of and its role in shaping the orion kl region .
the @xmath0sio @xmath1=1 and @xmath1=2 , @xmath2=1 - 0 transitions toward ( @[email protected]@xmath16 , @xmath17=@xmath18@xmath19 ) were observed using the vlba between 2001 march 19 and 2002 december 10 .
the separation between observations was approximately one month , although no usable data were obtained during two months in mid-2002 ( table 1 ) .
observations were made with the 10 antennas of the vlba , together with a single antenna from the vla .
the @xmath1=1 and @xmath1=2 transitions were observed simultaneously , each in a single polarization ( rr ) with a 15.9 mhz bandwidth centered at 7.5 relative to the local standard of rest ( lsr ) .
the adopted rest frequencies for the @xmath1=1 and @xmath1=2 transitions were 43.122024 ghz and 42.820432 ghz , respectively .
the data were processed with the vlba correlator , yielding 512 spectral channels with a spacing of @xmath30.2 in each of the two ifs .
correlator dump times were 2.097 s. the total integration time on during each epoch ( excluding calibration measurements ) was @xmath36 hours for each transition .
we performed all data reduction using the astronomical imaging processing system ( aips ) .
because our science goals require high precision , high dynamic range measurements , we took particular care to minimize systematic errors in each stage of our reduction and to maintain accurate astrometric registration between the @xmath1=1 and @xmath1=2 data ( which are very sensitive to delay calibration errors because of the 300 mhz frequency separation between transitions ) . to insure uniform processing of each epoch
, we developed a partially automated pops script to facilitate the implementation of certain reduction steps .
these semi - automated steps were interspersed with frequent manual checks of the data .
corrections for errors in the _ a priori _ earth orientation parameter and station positions ( computed starting with the usno 2004b geodetic vlbi solution ) were applied to the data , followed by digital sampler corrections for the vlba correlator .
if uncorrected , these position errors will lead to phase and delay errors which subsequently can not be calibrated out .
next , a preliminary amplitude calibration was performed using the default gain curves and system temperatures provided by nrao .
subsequently , an improved amplitude scale and pointing corrections were computed based on comparison of a time series of total - power spectra to a template for each station .
one - time atmospheric delay and instrumental phase and delay offsets between the ifs were estimated from fringe fits to a one - minute portion of a calibrator scan ( 0528 + 134 or 4c39.25 ) , followed by a second fringe - fit to both calibrators ( including all times ) to solve for the residual delays and rates .
solutions for both ifs were derived independently . however , after ensuring that no systematic offsets were present between the two sets of solutions , the rate solutions from if2 ( @xmath1=1 ) were copied and applied to if1 ( @xmath1=2 ) in order to minimize phase drifts between the two ifs ( which may in turn lead to astrometric misalignments between the two frequencies ) .
the time series phase differences for the two ifs for each antenna were examined following these steps to verify that no phase wraps were present during the track .
the amplitude portion of the bandpass calibration was computed using the total power data for the calibrators 0528 + 134 and 4c39.25 , and the phase portion was derived using a complex polynomial fit to the cross - power data of the same calibrators .
time - dependent frequency shifts were then applied to the line data to compensate for changes caused by the earth s motion over the course of the observation . to refine the astrometric position of the maser and further constrain delay - like errors in the calibration that would affect the relative astrometry of the @xmath1=1 and @xmath1=2 emission , a strong , compact @xmath1=1 maser reference feature was identified in each epoch through examination of cross - power spectra on a variety of baselines . each feature selected ( see table 2 ) was one that showed roughly constant visibility amplitudes as function of baseline length and orientation .
a fringe fit was performed on this reference feature to determine fringe rates . following this ,
a fringe - rate analysis was performed to solve for the position of the reference feature relative to the _ a priori _ correlator position .
this analysis was performed outside of aips , using a non - linear least - squares algorithm to fit a sinusoid to the measured fringe rates as a function of time and simultaneously solve for a single vertical atmospheric delay error term in the correlator model for each station .
the position offsets determined by this method ( table 2 ) were then applied to both ifs .
typical uncertainties in ra and dec were 5 mas and 20 mas , respectively .
the proximity of source i to the celestial equator accounts for the larger uncertainty in the latter coordinate .
self - calibration was performed on the @xmath1=1 reference channel , first in phase only , then in both amplitude and phase . typically a total of @xmath325 - 30 iterations was required . because the source s position on the sky compromises the @xmath20-@xmath1 coverage ,
care was exercised to avoid artifacts in the clean component models used for each subsequent self - calibration iteration .
final solutions were applied to the full spectral line data sets for both the @xmath1=1 and @xmath1=2 transitions . finally , several additional iterations of self - calibration were used to further improve the @xmath1=2 solutions before imaging the full line data sets .
imaging and deconvolution of the fully calibrated sio data sets were performed using robust weighting ( briggs 1995 ) with @xmath21=0 .
clean boxes were placed around emission regions in each spectral channel individually to reduce the effects of clean biasing ( e.g. , cotton 2007 ) .
all channels were cleaned to a level of 20 mjy beam@xmath22 .
this depth of cleaning balanced sidelobe removal against the production of deconvolution artifacts that may result from over - cleaning .
these data are particularly susceptible to such artifacts as a result of a combination of the source s location on the sky and the limited @xmath20-@xmath1 coverage of the array .
the resulting deconvolved images have a field - of - view of @xmath3@xmath23 , and in all cases we used a restoring beam of [email protected] mas with a position angle of 0@xmath25 .
the 1@xmath26 rms noise in individual channels ranged from @xmath38 - 10 mjy beam@xmath22 for line - free channels ( consistent with expected thermal noise ) to @xmath39 - 13 mjy beam@xmath22 for channels with strong line emission .
our resulting imaging data have higher dynamic range ( up to @xmath38000 ) than any previously obtained vlbi maps of the sio masers and also recover a significantly larger fraction of the total sio flux ( @xmath27% ; matthews et al .
, in preparation ) .
zeroth , first , and second moment maps were computed from each @xmath1=1 or @xmath1=2 data cube ( using the aips task momnt ) ; combined moments containing both transitions were also produced .
consecutive epochs were registered by maximizing the cross - correlation between maps of the logarithm of the velocity - integrated intensity .
this method removes any component of motion caused by a proper motion of ( cf .
gmez et al .
2008 ; goddi et al .
, in preparation ) .
when registering maps separated in time by roughly one month , this method appears to be robust to within approximately one pixel ( 0.05 mas ) , although for some epochs , misalignments between the two sio transitions of up to @xmath30.3 mas may be present as a result of residual calibration uncertainties .
lclc 2001mar19 & 51987.5 & bg118a & mk + 2001apr16 & 52015.5 & bg118b & mk + 2001may18 & 52047.5 & bg118c & ... + 2001jun22 & 52082.5 & bg118d & ... + 2001jul22 & 52112.5 & bg118e & ... + 2001aug24 & 52145.5 & bg118f & sc , hn + 2001sep19 & 52171.5 & bg118 g & fd + 2001oct21 & 52203.5 & bg118h & ... + 2001nov19 & 52232.5 & bg118i & fd + 2001dec21 & 52264.5 & bg118j & ... + 2002jan27 & 52301.5 & bg118k & ... + 2002mar02 & 52335.5&bg118l & mk , y + 2002apr01 & 52365.5 & bg118 m & ... + 2002apr29@xmath28 & 52393.5 & bg118n & ... + 2002may27@xmath28 & 52421.5 & bg118o & ... + 2002jun28 & 52453.5 & bg118p & ... + 2002aug09 & 52495.5 & bg129a & br , y + 2002sep05 & 52522.5 & bg129b & hn + 2002oct07 & 52554.5 & bg129c & ...
+ 2002nov06 & 52584.5 & bg129d & ... + 2002dec10 & 52618.5 & bg129e & ...
+ lcrcc bg118a & 229 & + 13.4 & + 0.032 & + 0.015 + bg118b & 228 & + 13.6 & @xmath290.015 & @xmath290.054 + bg118c & 338 & @xmath2910.2 & + 0.087 & @xmath290.077 + bg118d@xmath30 & 336 & @xmath31 & + 0.098 & @xmath290.119 + bg118e & 336 & @xmath31 & + 0.098 & @xmath290.119 + bg118f & 333&@xmath32 & + 0.086 & @xmath290.134 + bg118
g & 225 & + 14.3 & + 0.022 & + 0.059 + bg118h & 334 & @xmath33 & + 0.101 & @xmath290.078 + bg118i & 333 & @xmath32 & + 0.096 & @xmath290.101 + bg118j@xmath28 & 334 & @xmath33 & + 0.080 & @xmath290.094 + bg118k & 335 & @xmath34 & + 0.082 & @xmath290.145 + bg118l & 198 & + 20.1 & @xmath290.050 & + 0.010 + bg118 m & 338&@xmath35 & + 0.086 & @xmath290.055 + bg118p & 335 & @xmath34 & + 0.092 & @xmath290.061 + bg129a@xmath36 & 335 & @xmath34 & + 0.093 & @xmath290.084 + bg129b & 340&@xmath37 & + 0.095 & @xmath290.106 + bg129c & 198 & + 20.1 & + 0.033 & @xmath290.025 + bg129d & 338 & @xmath35 & + 0.092 & @xmath290.084 + bg129e & 340 & @xmath37 & + 0.089 & @xmath290.082 +
total intensity images of the @xmath0sio @xmath1=1 and @xmath1=2 , @xmath2=1 - 0 emission toward are presented in fig .
[ fig : mom0 ] . to guide the reader ,
several key features of the maser distribution are also illustrated schematically in fig .
[ fig : cartoon ] . as seen in previous observations of with vlbi resolution ( greenhill et al .
1998 ; doeleman et al .
1999 ; kim et al .
2008 ) , the bulk of the sio emission is located within four `` arms '' of an x - shaped pattern , centered on the position of . along a position angle of @xmath38 , a @xmath314-au - thick dark band with well - defined inner boundaries is evident , inside of which no sio maser emission is detected .
this zone maintains a constant thickness over many months and harbors the disk - shaped 7-mm continuum source imaged by reid et al .
( 2007 ) .
another key feature of fig . [ fig : mom0 ] is the presence of two `` bridges '' of emission connecting the north and east arms and the south and west arms , respectively .
bridge emission was detected in the @xmath1=1 and/or @xmath1=2 transitions in all of our observational epochs .
the western bridge was always significantly brighter than the eastern bridge
the latter typically showing at most a few weak , isolated spots during any given epoch ( e.g. , fig .
[ fig : eastbridge ] ) . a counterpart to the western bridge
was first seen in the 1997 vlba images of doeleman et al .
( 2004 ) , and those authors noted that it could not be readily explained in the context of existing biconical outflow models for , which assumed that the symmetry axis of the outflow lay along the northwest / southeast direction ( cf .
greenhill et al . 1998 ; doeleman et al . 1999 ) .
indeed , this bridge emission is now a key piece of evidence demanding a new geometric and dynamical model for ( [ globvel ] & [ model ] ; see also greenhill et al .
2004b ; kim et al .
2008 ) .
comparison of the maps of the two sio transitions in fig .
[ fig : mom0 ] shows that on average , the @xmath1=2 emission arises at smaller projected distances from than the @xmath1=1 emission ( see also kim et al .
2008 ; goddi et al .
2009a ) , although there is considerable overlap . to better quantify this , in fig .
[ fig : histo ] we plot histograms of the number of discrete maser features identified in each of the two sio transitions ( see [ proper ] ) as a function of distance from source i. to minimize the impact of projection effects , only spots within the four `` arm '' regions are included .
we see that @xmath1=2 features dominate at smaller distances from the star , while the @xmath1=1 features become more prevalent at larger radii . beyond @xmath3945 au , @xmath1=2 emission fades , and only @xmath1=1 features are observed .
the radial offset , @xmath40 , between the locations where the peak number densities of @xmath1=1 and @xmath1=2 features occur is @xmath311 au .
this is substantially larger than the offsets observed in evolved stars ( @xmath411 au ; e.g. , desmurs et al . 2000 ; soria - ruiz et al .
2004 ; yi et al . 2005 ) .
furthermore , we find no evidence for variation in the relative distributions of the @xmath1=1 and @xmath1=2 masers during our monitoring .
this too is in contrast to the case of evolved stars , where the relative position of the two transitions may be seen to vary as a function of stellar phase ( i.e. , on timescales of months ; see gray & humphreys 2000 ; yi et al . 2005 ) .
nonetheless , we can not yet exclude that changes in the relative distributions of the @xmath1=1 and @xmath1=2 masers around may evolve on timescales of several years or more .
the presence of an overlap region for the sio @xmath1=1 and @xmath1=2 masers provides an interesting constraint on the dominant excitation mechanism for the masers .
purely radiative pumps ( e.g. , alcolea 1989 ; bujarrabal 1994a , b ) do not produce spatial coincidence of the @xmath1=1 and @xmath1=2 masers . on the other hand , collisional pumps lead to a large amount of overlap in the parameter space for these two transitions ( e.g. , lockett & elitzur 1992 ) , in agreement with our observations .
preliminary radiative transfer modeling by goddi et al .
( 2009a ) using a large velocity gradient code provides some insight into the physical conditions ( temperature , density ) required to produce the observed @xmath1=1 and @xmath1=2 maser distributions . using a model that included both collisional and radiative pumping , as well as physical parameters appropriate for the region , goddi et al
. found that the @xmath0sio @xmath1=2 maser will survive in a stronger , hotter radiation field than the @xmath1=1 maser and that it will be optimized at higher kinetic temperatures ( @xmath422000 k ) where it can become more strongly inverted than the @xmath1=1 transition .
this is consistent with the observation that @xmath1=2 masers tend to arise closer to than the @xmath1=1 masers .
the goddi et al . model also predicts overlap in the range of densities where the @xmath1=1 and @xmath1=2 transitions can occur ( i.e. , @xmath43 @xmath12 for @xmath1=1 and @xmath44 @xmath12 for @xmath1=2 , respectively ) .
this suggests that the mean gas densities within this overlap zone observed around should be @xmath45 @xmath12 .
more sophisticated modelling that treats the effects of maser saturation and line overlap will be presented by humphreys et al .
( in preparation ) . a combined map of both sio transitions , summed over all 19 observing epochs ( fig . [ fig : nest ] ) , reveals the trails of hundreds of individual maser features , underscoring that at vlbi resolutions , is the most complex maser source known .
individual maser features can persist for several months or more , and a significant fraction exhibit detectable motions in the plane of the sky on timescales of a few weeks ; see also [ proper ] ) .
in addition to the multitude of maser trails in fig . [
fig : nest ] , elongated , high surface brightness `` emission fronts '' are visible in three of the arms ( see also fig . [
fig : cartoon ] ) .
these fronts are more prominent in the @xmath1=1 transition ( cf .
[ fig : mom0 ] ) .
analogous features have been seen in previous high - resolution images of the sio masers ( greenhill et al .
1998 ; doeleman et al .
1999 , 2004 ) , but our new observations show that such structures can remain relatively stable in extent , position , and velocity over many months .
an animated movie ( fig . [ fig : nest ] , online version ) provides a `` 4-d '' view highlighting the month - to - month changes that give rise to the intricate structures seen in the 19-epoch summed image . while the gross morphology of the maser distribution persists over the length of the movie , significant temporal changes are evident in the positions and brightnesses of thousands of individual features , collectively giving the maser - emitting material an `` effervescent '' appearance .
further inspection reveals that the observed motions of the maser features exhibit certain systematic trends .
first , outward motions ( i.e. , away from ) are visible along each of the four arms , as well as in several isolated groups of features that are detached from the main arms and bridges ( see fig .
[ fig : cartoon ] ) .
radially outward motions are also observed within the western bridge region .
however , in addition , maser features along the western bridge clearly display a component of motion _ tangential _ to the bridge .
interestingly , material along the outer edge of the western bridge is seen moving to the northwest , while material along the inner edge moves in the opposite sense .
these patterns most likely arise from emitting material along the rim of a rotating , geometrically thick disk whose near side is tipped slightly to the southwest ( [ model ] ) .
motions within the eastern bridge also appear to be consistent with this picture , although features in this region typically persist for only one or two epochs , making it more difficult to establish systematic trends .
the bright , elongated emission fronts described above pose an interesting contrast to many of the other arm features .
although modest changes in the position angle and small - scale morphology of these fronts are visible from month to month , their general locations and extents remain fairly constant over nearly two years .
we also observe gas within these regions whose line - of - sight velocity dispersion ( @xmath31 - 5 ) is several times higher than the mean of the surrounding material ( @xmath30.3 ; fig . [
fig : veldisp ] ) . in fig .
[ fig : mom1 ] we show a first moment map obtained from the combined sio @xmath1=1 and @xmath1=2 emission on 2001 march 19 .
integrated spectra of the two transitions are also shown .
the integrated line profiles of both the @xmath1=1 and @xmath1=2 emission maintained their characteristic double - peaked shape throughout the course of our monitoring .
a more detailed analysis of the spectral line profiles and the temporal variability of the sio masers will be presented by matthews et al .
( in preparation ) .
consistent with earlier studies , we also see that the redshifted sio emission arises primarily from the north and west arms , while the blueshifted emission arises predominantly from south and east arms .
furthermore , our new high dynamic range images showcase some additional characteristics of the sio distribution that were only hinted at by previous investigations .
[ fig : mom1 ] further highlights the two bridges of emission connecting the south and west ( blue ) arms and the east and north ( red ) arms , respectively ( see also fig [ fig : mom0 ] & [ fig : nest ] ; greenhill et al .
we see that emission from both bridges occurs predominantly within a few of the systemic velocity and that a velocity gradient is present in the sense that the material in the northern end of the bridge has ( on average ) higher radial velocities compared with the material in the southern end .
we interpret these observations as signature of rotational motion about an axis oriented along the northeast / southwest direction . in this picture , the red and blue arms represent the receding and approaching sides , respectively , of the rotating structure ( see [ model ] ) .
another notable feature of fig .
[ fig : mom1 ] is the presence of clear systematic velocity gradients along each of the four arms i.e .
, emission at smaller projected distances from the source tends to have higher radial velocities than emission at larger projected radii .
the magnitude of the gradients is @xmath30.4 au@xmath22 .
while there had been some indication of velocity gradients along individual arms from earlier imaging studies ( doeleman et al .
2004 ; greenhill et al .
2004b ) , our new data show similar gradients along all four arms , implying that they result from a global rather than a localized phenomenon .
this finding has been confirmed independently by kim et al .
( 2008 ) using data from the vera array . as with the bridge masers , these velocity gradients are difficult to account for in any model of that does not include rotation about a northeast / southwest axis . an animated version of fig .
[ fig : mom1 ] ( online version ) displays the radial velocity fields derived from all 19 epochs of observations .
the first moment maps are somewhat noisier than the zeroth moment maps used to produce the animation in fig .
[ fig : mom0 ] , leading to visible north / south artifacts around bright features in some frames . nonetheless , fig .
[ fig : mom1 ] helps to highlight the evolution of some of the smallest and faintest detected features , particularly near the bridge regions .
[ fig : pm ] presents proper motions of individual maser features derived from our data . here
, positions of the maser features during each epoch were measured from two - dimensional gaussian fits to the sio total intensity images , while radial velocities were taken to be the intensity - weighted mean velocity at each feature s location .
our typical detection threshold was @xmath46 .
the total numbers of maser spots catalogued over 19 epochs were 27,857 in the @xmath1=1 line and 15,448 in the @xmath1=2 line , respectively .
the resulting @xmath1=1 and @xmath1=2 spot catalogues were then systematically searched to identify features that persisted over at least _ three _ epochs ( bridge regions ) or _ four _ epochs ( all other regions ) and appeared to move along linear trajectories in the plane of the sky . for each of the epochs ( beginning with bg118b ) an automated search of the preceding epoch was performed for possible counterparts to each of the identified maser spots .
the search radius was restricted to 1 mas ( corresponding to @xmath4735 ) , with the additional requirement that the radial velocity change between epochs was less than or equal to 0.2 .
the latter criterion corresponds to an acceleration @xmath48 cm s@xmath49
. for each candidate identified , a line was fit between the position of the initial spot and the candidate to delineate an initial predicted trajectory .
the subsequent epoch was then searched for spot candidates lying along the predicted path to within @xmath500.15 mas in @xmath51 and @xmath52 and @xmath500.2 in @xmath53 .
the former limit corresponds approximately to our registration uncertainty between epochs , while the latter is comparable to our velocity resolution .
each time a spot was found matching these criteria , a linear least squares fit ( in @xmath51 and @xmath52 ) was performed to the resulting set of three spots to refine the predicted trajectory . outside of the bridge regions , this procedure
was then repeated for the next epoch in the sequence . because the region around is so crowded with maser features
, there were a number of cases where more than one feature from a given epoch was identified as a candidate for falling on a particular trajectory .
for these cases , only the trajectory with the lowest @xmath54 value was used to compute a proper motion vector ( @xmath55 , @xmath56 , @xmath57 , pa ) for the final database .
the results of the above analysis are summarized in fig .
[ fig : pm ] . features persisting for more than 3 months ( bridge regions ) or 4 months ( all other regions ) are visible as overlapping arrows .
typical feature lifetimes range from less than one month to 3 or 4 months , with occasional features persisting as long as @xmath311 months .
a full catalogue and analysis of the sio maser proper motions will be presented elsewhere ( goddi et al . , in preparation ) , but here we describe a few key trends . the derived 3-d ( space ) velocities for individual features range from 5.3 to 25.3 , with a mean of 14.0 .
we assume a systemic velocity for @xmath58=5.5 .
we find no evidence for any systematic decrease in the magnitudes of the motions in the plane of the sky with increasing distance from the star , suggesting that the radial velocity gradients along the arms ( [ globvel ] ) are not simply the result of a decelerated outflow , but instead arise primarily from differential rotation ( see also kim et al .
2008 ) .
the proper motion vectors plotted in fig .
[ fig : pm ] were derived by conservatively assuming that the motions were linear over 3 - 4 months and that the radial component of the velocity did not change by more than @xmath500.2 from month to month .
these initial selection criteria may exclude features with significant accelerations .
a systematic search for accelerations in the spot motions will be presented in a future paper .
however , we note that we already see qualitative evidence of accelerated motions within our data based on the time - integrated distributions of features along the `` streamers '' and within several `` isolated '' regions ( e.g. , the features labeled `` a '' and `` b '' , respectively on fig . [ fig : pm ] ) .
two examples are highlighted in fig .
[ fig : streamer ] & [ fig : twizzler ] .
although the motions of _ individual _ maser clumps within the regions shown in figs . [ fig : streamer ] & [ fig : twizzler ] can not be distinguished from linear paths to within positional uncertainties , the time - integrated views reveal the appearance of multiple features over multiple epochs that cluster along curving arcs , as would be expected if the maser - emitting material were being shepherded along twisted magnetic field lines ( e.g. , banerjee & pudritz 2006 ; see also [ mhd ] ) .
other noteworthy features of fig .
[ fig : pm ] include the proper motions along the western bridge ( labeled `` c '' on fig .
[ fig : pm ] ) , which exhibit a component of radially outward motion in addition to two oppositely directed streams moving tangential to the bridge , as seen in the animated versions of fig .
[ fig : nest ] & fig .
[ fig : mom1 ] .
finally , as indicated by arrow `` d '' , in the north arm we find that the proper motion vectors twist by @xmath59 roughly half - way along the arm , again suggestive of material that may be following non - linear paths .
earlier vlbi observations of sio masers around established the x - shaped distribution of emission as well as the spatial separation between the red- and blue - shifted arms . based on these observations , greenhill et al .
( 1998 ) and doeleman et al .
( 1999 ) proposed that the sio masers arise from limb - brightened edges of a biconical outflow , oriented along a southeast / northwest direction .
however , our new , more sensitive observations are better explained by a new model namely that the masers are associated with an edge - on disk whose rotation axis is oriented along the northeast / southwest direction ( see also greenhill et al .
2004b ; kim et al .
support for this model comes from the radial velocity gradients along each of the four arms ( indicative of differential rotation ) , as well as the presence of the bridge emission and its associated velocity gradients ( [ globvel ] ) .
furthermore , the canting of the maser arms ( fig .
[ fig : nest ] ) suggests a reflexive symmetry about a plane whose position angle ( pa@xmath60 ) closely matches the ridge of 7-mm continuum emission observed by reid et al .
( 2007 ) on 2000 november 10 ( pa@xmath61 ) .
the presence of the well - defined dark band in fig .
[ fig : nest ] ( see 3.1.1 ) also suggests that the conditions favorable for excitation of the sio masers set in rather abruptly at a fixed scale height above a flattened , disk - like structure .
underlying the new kinematic model for is the assumption that the sio masers trace real , physical motions of gas clumps rather than , e.g. , illumination patterns or shocks transversing a fixed medium .
the contrast between the pattern of motions observed in the western bridge versus the arms ( [ proper ] ) strongly supports this interpretation , as shocked material would not be expected to have two apparent kinematic components ( both tangential and radially outward ) .
moreover , owing to gradients in temperature and density between regions close to the disk plane and the outer reaches of the wind , shock conditions are unlikely to be similar over the scales of tens of au from where maser motions are observed . finally , the multitude of linear maser trails seen in fig .
[ fig : nest ] also supports a kinematic interpretation for the sio masers ; features that move many times their characteristic sizes without significantly changing morphology would not be expected to arise as a shock front transverses a clumpy , inhomogeneous medium and are a strong indicator that we are tracing motions of individual clumps .
one of the longstanding controversies surrounding has been the mass of the central star .
while our present data can offer important new constraints on this quantity , obtaining a precise mass estimate is complicated by the likelihood that the maser - emitting material is not in purely keplerian rotation ( see also kim et al .
for example , some degree of turbulence is almost certain present in the gas ( e.g. , as evidenced by small - scale complexities in the radial velocity fields in the arm regions ) , and the outward motions of the maser clumps ( [ proper ] ) imply forces acting on the maser gas opposite to those of gravity ( e.g. , radiative and/or magnetic forces ) .
such effects can lead to underestimates of the enclosed mass ( e.g. , knigl & pudritz 2000 ; pitu et al .
2005 ; bujarrabal et al . 2005 ) .
moreover , the disk itself might contain up to a few solar masses of material ( reid et al .
2007 ) and thus have a non - negligible mass relative to the central source .
one means of estimating of the mass of comes from the observed transverse motions of maser features in the western bridge region ( fig .
[ fig : pm ] ) .
we assume that the bridge features lie at @xmath62 au in a circularly rotating disk ( i.e. , just outside the edge of ionized inner disk measured by reid et al .
2007 ) and that the disk has a nearly edge - on inclination to our line - of - sight ( @xmath63 ) . taking the mean space motion of the @xmath1=2 bridge features ( @xmath6413.5 ) then implies @xmath65 .
a second mass estimate can be derived by assuming that the material within the four arms is part of an outflowing wind ( see [ wind ] ) and therefore must be moving at or near escape velocity .
taking the mean space velocities of the @xmath1=2 masers within the four arms ( @xmath66 ) and a fiducial radius @xmath6725 au from ( roughly equal to the radius of the midpoint of the base of each arm from ) also implies a central mass of @xmath65 .
these kinematically - derived masses are somewhat smaller than the estimate of reid et al .
( 2007 ) based on the 7-mm radio continuum luminosity ( @xmath68 ) .
since a star with @xmath69 would be unable to produce an region and would not have sufficient luminosity to power the sio maser emission ( @xmath70 ; menten & reid 1995 ) , it therefore seems probable that is somewhat higher than the above kinematically determined values i.e .
, @xmath71-@xmath72 .
if magnetic fields are threading the disk and are responsible for powering the wind ( see [ mhd ] ) , then the models described by knigl & pudritz ( 2000 ) predict that only the material at the _ base _ of the wind is expected to exhibit keplerian rotation .
we therefore have measured the locations of the peak rotational velocity of the sio emission along the northeastern edge of the dark band that runs parallel to the disk midplane ( see fig .
[ fig : cartoon ] ) . taking a mean from the 19 epochs of data and from the red- and blue - shifted sides of the disk , we find @xmath7319 at @xmath74 au .
this implies @xmath13 .
if dust is mixed with the sio maser - emitting gas as proposed by elitzur ( 1982 ) , then radiation pressure on the grains could also influence the gas kinematics , assuming the grains can efficiently transfer some of their momentum to the gas .
this would again lead to observed velocities that are smaller than predicted by kepler s law for a given mass .
however , estimating the magnitude of this effect on the observed gas velocities will depend on the intrinsic stellar mass and luminosity , as well as the detailed properties of the grains all of which are uncertain ( e.g. , kwok 1975 ) . moreover , it is unclear whether dust could survive at the temperatures expected near the base of the wind ( @xmath752000 k ; goddi et al .
2009a ; see also below ) .
we emphasize that the interpretation of as a single , luminous yso with @xmath71-@xmath72 seems to provide the simplest explanation for _ both _ the radio continuum observations and the sio maser kinematics .
a binary of equivalent mass , or a less massive , less luminous star , would not be able to produce the observed radio continuum emission nor provide the necessary luminosity to power the sio masers .
on the other hand , a significantly more massive star would be difficult to reconcile with the observed sio kinematics ( i.e. , the transverse motions in the bridges , which we interpret as material orbiting the central mass , and the systematically outward motions in the arms , which we interpret as material traveling at or near escape velocity ) .
our results therefore seem to be in contradiction with the scenario recently proposed by gmez et al .
( 2008 ) , in which was ejected from a multiple stellar system @xmath3500 years ago and now comprises a tight binary with a mass in the range @xmath76 .
in addition to the problem of the discrepancy between our derived mass for source i and the value proposed by gmez et al .
( 2008 ) , accounting for the properties of source i s disk in such a picture ( e.g. , its size , density , and symmetry ) might also be problematic .
the passage of another star within @xmath77 au of ( see gmez et al . ) and the subsequent formation of a tight binary most likely would have disrupted any previously existing disk around ( e.g. , moeckel & bally 2006 ) , requiring formation of the existing disk structure within the past 500 yr . nonetheless , under certain conditions , such rapid disk regrowth might be possible via either bondi - hoyle accretion or the tidal shredding of the interloper . in the case of bondi - hoyle accretion
, the bondi radius of an accreting object of mass @xmath78 is defined as @xmath79 , where @xmath80 is the gravitational constant and @xmath81 is a characteristic velocity , which may be taken to be the motion of the object with respect to the surrounding gas ( e.g. , krumholz et al .
2006 ) . from goddi et al .
( in prep . ) , the motion of with respect to the ambient medium is @xmath8212 , implying @xmath8350 au for @xmath84 .
such a radius is comparable to the observed size of the disk . while the mass of the disk surrounding is uncertain ( see reid et al .
2007 ) , we can roughly estimate this quantity , @xmath85 , by assuming the disk shape is a flattened cylinder with @xmath8650 au and @xmath87=14 au ( [ morph ] ) and that the mean particle density is comparable to the value required to explain the sio maser emission ( @xmath88 @xmath12 ) .
assuming the bulk of the disk material is ionized , we adopt a mean molecular weight per particle of @xmath8=0.6 , implying @xmath89 .
given a time frame of 500 yr , this implies a minimum required mass flux @xmath90 yr@xmath22 ( comparable to the mass _ outflow _ rate estimated from the sio @xmath1=0 emission ; greenhill et al .
, in prep . ) . using equation 1 of krumholz et al .
( 2006 ) , we can now estimate the required ambient density to support this mass accretion rate as @xmath91^{-1}\approx 3.0\times10^{-17}$ ] g @xmath12 .
assuming the ambient material is purely molecular ( @xmath8=2.3 ) then implies an ambient particle density @xmath92 .
although the latter value is rather high , it is comparable to values previously measured for the orion hot core region ( e.g. , masson & mundy 1988 and references therein ) and thus may be roughly consistent with plausible values for this orion kl region .
we conclude that in the absence of more sophisticated calculations , we can not rule out disk augmentation or rebuilding via bondi - hoyle accretion may have occurred during the past 500 yr . however , we note that one additional caveat is that disks formed in this manner are predicted to have rather chaotic and asymmetric structures ( e.g. , krumholz et al . ) , in contrast to the case .
one additional scenario that might reconcile our new measurements with the findings of gmez et al .
( 2008 ) is the possibility that has recently formed or altered its disk by tidally shredding a lower mass interloper ( see davies et al .
if this interloper had a mass as high as @xmath93 , this would bring the total mass of the system in marginal agreement with values proposed by gmez et al .
furthermore , such an event might offer a natural explanation for the maser emission and outflows associated with ( see bally & zinnecker 2005 ) .
nonetheless , this picture would likely require a considerably closer passage between the two stars than estimated by gmez et al .
( i.e. , as close as a few tens of stellar radii ; see davies et al . ) .
in addition , the relatively large resulting mass of the disk may in turn require an implausibly luminous central star in order to account for the observed radio continuum emission ( see reid et al .
2007 ) . in summary , while our latest observations of present a compelling case for disk - mediated accretion in a massive yso , it is clear that further modeling will be required to explore possible interaction scenarios and to better constrain whether such an event is likely to have influenced its present disk and outflow properties .
new n - body simulations as well as further discussion of the interaction history of will be presented by goddi et al .
( in preparation ) . because sio maser emission is extremely rare around ysos , its origin in the case of has been another longstanding puzzle .
cunningham et al .
( 2005 ) proposed that the masers arise from a shear layer along the walls of a cavity that has been evacuated as a bipolar wind expands into a rotating , collapsing envelope .
this model is able to reproduce the magnitudes and directions of the proper motions in the arms as well as the frothy appearance of the sio - emitting material .
however , the observed breadth of the arms appears to be greater than expected for an interface region .
moreover , this model can not readily explain several other features of the sio masers , including the transverse motions observed in the inter - arm bridges , the presence of groups of maser features beyond the four main arms ( the `` isolated '' features in fig .
[ fig : cartoon ] ) , or the linear trajectories of features that persist unperturbed over many months ( fig .
[ fig : nest ] , [ fig : streamer ] , & [ fig : twizzler ] ) .
the cunningham et al . model also predicts higher densities along the outer edges of the arms ( their fig . 1 ) , which appears to be inconsistent with our observation that the @xmath1=2 masers ( which preferentially occur in higher density gas ) lie on average closer to than the @xmath1=1 masers ( [ morph ] ;
[ fig : histo ] ) .
wright et al . ( 1995 ) proposed a slightly different scenario namely that the sio masers arise from material ablated from the surface of an accretion disk by a wind or outflow .
because our current observations provide evidence for the presence of an accretion disk , we now favor some variant of this `` boiling disk '' picture .
constraining the driving mechanism for this wind will require detailed modeling , but here we comment briefly on the likely applicability of various classes of disk wind models to the case . as discussed by hollenbach et al .
( 1994 ) , ysos hot enough to produce an region are capable of mass - loss via photoevaporation of their disks as material is heated to temperatures in excess of the local escape temperature . in the `` weak wind '' case , an ionized flow is predicted to set in beyond the disk radius , @xmath94 , where the sound speed is roughly equal to the escape speed . for , if we take the sound speed as @xmath311 ( assuming @xmath95 k on the surface of the ionized disk ; reid et al .
2007 ) , a mean mass per particle of 1.13@xmath96 g within the ionized disk ( hollenbach et al . 1994 ) , and @xmath97 , this predicts @xmath98 cm roughly a factor of three larger than what is observed .
in addition to this discrepancy , the outflowing material is predicted to be mostly ionized , raising the problem of how to maintain sufficient quantities of dense , molecular material in the disk wind and how to account for the sio maser emission in the bridge regions . for hot stars , radiation pressure mediated by ultraviolet absorption line
opacity offers another means of powering a wind .
classically , such winds tend to have velocities too high ( @xmath99400 ) and densities too low ( @xmath100 g @xmath12 ) to readily account for the maser - emitting gas surrounding ( e.g. , lamers et al .
1995 ) . drew et al .
( 1998 ) showed that the presence of a rotating , circumstellar disk helps to produce a slower , denser wind along the equatorial direction .
however , for @xmath68 , the characteristic wind speeds at more oblique angles ( i.e. , angles consistent with the outward motions along the arms of ) still are an order of magnitude too high to match those of the sio masers ( see drew et al . ) .
furthermore , the high poloidal velocities predicted by this model are inconsistent with outflow speeds inferred from other line tracers , such as sio @xmath1=0 emission ( goddi et al .
2009a ; greenhill et al . , in preparation ) .
it thus appears that line - mediated radiation pressure is unlikely to play a significant role in powering the outward migrations of the sio maser - emitting gas around .
elitzur ( 1982 ) has suggested that gas and dust may be mixed within the zone from which the sio maser emission arises around .
this would be in contrast to the situation in the extended atmospheres of evolved stars , where the sio maser emission arises from a region just inside the dust - formation radius ( e.g. , reid & menten 1997 ) .
elitzur further showed that in such a case , dust - mediated radiation pressure alone might be able to account for the outflow of material from .
although we can not yet exclude this general class of model for driving the wind , several of the previous assumptions made by elitzur ( 1982 ) require revision as a result of more recent , high - resolution observations .
for example , elitzur assumed a luminosity for the of @xmath101 , which is likely an order of magnitude too high ( reid et al .
2007 ) and a mass - loss rate of @xmath102 yr@xmath22two orders of magnitude larger than the value recently derived by greenhill et al .
( in preparation ) based on sio @xmath1=0 measurements .
elitzur s model also depends upon energy input by turbulent motions of @xmath37 in order to maintain the gas at a higher temperature than the dust ( as otherwise the gas would rapidly thermalize to the dust blackbody temperature and population inversion could not be maintained ) .
such large turbulent motions are inconsistent with the low mean line - of - sight velocity dispersion of the maser - emitting material seen in our present data ( fig .
[ fig : veldisp ] ) as well as with the ordered velocity field of the sio masers ( fig .
[ fig : mom1 ] ) .
lastly , as already noted above , it would likely be difficult for grains to survive within @xmath103100 au from , particularly near the base of the wind , where temperatures may be @xmath992000 k ( goddi et al .
given these new developments , the dust - driven wind scenario for now faces a number of challenges .
apparent curved and helical trajectories of certain sio maser features ( [ morph ] & [ proper ] ) , strongly hint that magnetic fields may play a role in shaping the dynamics of the region .
for example , it would seem difficult to explain features such as the streamer emulating from the western bridge ( fig . [
fig : streamer ] ) in the absence of magnetic fields . while pressure gradients within the disk might act to bend the paths of outflowing material
, the observations of pronounced curvature at large vertical displacements from the plane would require that the pressure scale height of the disk is comparable to the observed vertical extent of the masers .
moreover , we see numerous proper motion trajectories that appear to be linear , including along the outer edges of the individual arms , contrary to what would be expected if pressure gradients were important .
we therefore suggest that the streamers are more likely to be comprised of gas clumps constrained to move along a magnetic field lines like `` beads on a wire '' ( see e.g. , blandford & payne 1982 ) .
previous evidence for a magnetic field associated with was provided by polarization measurements of the sio @xmath104 , @xmath2=1 - 0 and @xmath2=2 - 1 emission ( tsuboi et al .
1996 ; plambeck et al .
2003 ) . after correcting for faraday rotation , the position angle for the polarization vectors derived by plambeck et al .
( 57@xmath25 ) is consistent with magnetic field lines threading roughly perpendicular to the disk defined by the sio masers and 7-mm continuum emission and parallel to the outflow direction traced by the sio @xmath104 emission and the h@xmath105o masers ( see goddi et al .
2009a ; greenhill et al . , in preparation ) .
assuming a magnetic field is present , a magnetocentrifugal wind ( e.g. , blandford & payne 1982 ; knigl & pudritz 2000 ) would be a natural candidate for powering a disk wind from . in this scenario ,
gas clumps fragment from the disk , are swept outward along field lines by hydromagnetic forces ( e.g. , emmering et al .
1992 ) , and are induced to excite maser emission when irradiated , shocked , or heated by collisions .
magnetic phenomena may also provide an explanation for the elongated emission fronts visible in three of the arms .
for example , tangled magnetic field lines ( e.g. , kigure & shibata 2005 ; banerjee & pudritz 2006 ) or instabilities within a magnetically - driven flow ( e.g. , kim & ostriker 2000 ) might account for the elevated velocity dispersions and rope - like morphologies of these features ( see [ morph ] ) .
doeleman et al .
( 1999 ) originally proposed that these elongated structures might arise at the shocked interfaces of an outflow .
however , the observation that the emission fronts are not preferentially oriented perpendicular to the outflow direction ( as expected for shock fronts ) seems to argue against this interpretation .
it is believed that a necessary condition for launching a magnetically - powered wind is that the vertical magnetic field is close to equipartition ( e.g. , ferreira 2007)i.e . ,
@xmath106 , where @xmath107 is the gas number density , @xmath108 is the mean velocity of a gas molecule , @xmath109 is the magnetic field strength , and @xmath110 is the permeability of free space . assuming @xmath111 @xmath12 ( 3.1.1 ) and @xmath108=14.0 ( [ proper ] ) , the implied magnetic field strength for is @xmath30.3 g. one possible source for this field might be the original interstellar magnetic field threading the molecular cloud out of which was born .
based on measurements of oh 1665-mhz masers across a @xmath112 au region surrounding , cohen et al .
( 2006 ) derived field strengths of 1.8 to 16.3 mg .
since the oh masers are believed to arise from material with molecular hydrogen density @xmath113 @xmath12 ( gray et al .
1992 ) , and magnetic field strength is expected to scale as the square root of the gas density , it is plausible that field strengths of order the value predicted by assuming equipartition might now be present within the denser , sio - emitting gas . direct measurements of the magnetic field strength in the disk of would be of considerable interest , both for understanding its role during the accretion / outflow process and for providing new clues on the nature of magnetic fields in b - type stars during later evolutionary stages .
little is presently known about the range of magnetic field strengths present in early - type b stars on the main sequence owing to the difficulty of measuring fields @xmath1033 kg in rapidly rotating stars ( e.g. , landstreet 1992 ; schnerr et al .
2008 and references therein ) .
moreover , it has been suggested that magnetic fields with strengths as low as a few gauss and originating at the time of the star s formation might account for the origin of magnetized neutron stars ( ferrario & wickramasinghe 2005 ) .
we note that even if magnetohydrodynamic forces are not the main driving mechanism for the wind , magnetic forces may still play a key role in shaping the outflow , similar to what has been proposed for planetary nebulae ( e.g. , blackman 2008 and references therein ) .
in addition , magnetic fields may provide a mechanism to enhance the coupling between gas and dust grains , thereby increasing the efficiency of a possible dust - driven wind ( hartquist & havnes 1992 ) .
we have presented two multi - epoch movies chronicling the evolution of the @xmath0sio @xmath1=1 and @xmath1=2 , @xmath2=1 - 0 maser emission surrounding orion over nearly two years .
these movies comprise vlba observations with @xmath30.5 mas ( 0.2 au ) resolution and provide the most detailed view ever obtained of the dynamics and temporal evolution of molecular material within @xmath320 - 100 au of a massive yso .
we interpret the sio masers surrounding as arising from a wide - angle , bipolar wind that emanates from a rotating accretion disk viewed nearly edge - on .
we find evidence to support the suggestion that magnetic fields are playing a role in shaping and/or powering this wind .
the maser kinematics and proper motions , coupled with constraints from previous radio continuum observations , imply a mass for the central star of @xmath998@xmath114 .
our study provides compelling evidence that disk - mediated accretion and low - velocity , wide - angle winds are both key elements in the evolution of young stars in this mass range .
however , we can not exclude the possibility that a recent encounter has also played a role in shaping the properties of the disk and outflow .
we thank mark reid for valuable technical discussions and for supplying his fit@xmath115rates code .
we also acknowledge helpful discussions with mark krumholz and comments from our anonymous referee that have helped to improve our presentation . this project has been supported by nsf grant 0507478 and a visiting scientist appointment to ldm from the smithsonian astrophysical observatory .
the data presented here were part of nrao programs bg118 and bg129 .
greenhill , l. j. , reid , m. j. , chandler , c. j. , diamond , p. j. , & elitzur , m. 2004b , in star formation at high angular resolution , iau symp . 221 , asp conf .
series , ed .
m. g. burton , r. jayawardhana , & t. l. bourke , ( asp : san francisco ) , 155 hasegawa , t. , morita , k .-
okurmura , s. , kaifu , n. , suzuki , h. , ohishi , m. , hayashi , m. , & ukita , n. 1986 , in masers , molecules , and mass outflows in star forming regions , ed .
a. d. haschick , ( haystack observatory : westford ) , 275 | we present multi - epoch very long baseline array ( vlba ) imaging of the @xmath0sio @xmath1=1 and @xmath1=2 , @xmath2=1 - 0 maser emission toward the massive young stellar object ( yso ) orion .
both sio transitions were observed simultaneously with an angular resolution of @xmath30.5 mas ( @xmath30.2 au for @xmath4=414 pc ) and a spectral resolution of @xmath30.2 .
here we explore the global properties and kinematics of the emission through two 19-epoch animated movies spanning 21 months ( 2001 march 19 to 2002 december 10 ) .
these movies provide the most detailed view to date of the dynamics and temporal evolution of molecular material within @xmath320 - 100 au of a massive ( @xmath5 ) yso . as in previous studies , we find that the bulk of the sio masers surrounding lie in an x - shaped locus ; the emission in the south and east arms is predominantly blueshifted and emission in the north and west is predominantly redshifted .
in addition , bridges of intermediate - velocity emission are observed connecting the red and blue sides of the emission distribution . we have measured proper motions of over 1000 individual maser features and find that these motions are characterized by a combination of radially outward migrations along the four main maser - emitting arms and motions tangent to the intermediate - velocity bridges .
we interpret the sio masers as arising from a wide - angle bipolar wind emanating from a rotating , edge - on disk .
the detection of maser features along extended , curved filaments suggests that magnetic fields may play a role in launching and/or shaping the wind .
our observations appear to support a picture in which stars with masses as high as at least @xmath6 form via disk - mediated accretion .
however , we can not yet rule out that the disk may have been formed or altered following a recent close encounter . |
quantum information processing ( qip ) has been extensively studied for a qubit system which is a quantum extension of a bit , spanning two - dimensional hilbert space .
a qubit is realized by a electronic spin , a two - level atom , the polarization of a photon and a superconductor among others .
parallelly , much attentions have been paid to the qip of quantum continuous variable ( cv ) system which is a quantum extension of analog information in classical information theory .
cv physical systems such as a harmonic oscillator , a rotator and a light field are defined in infinitive - dimensional hilbert space .
while conversions of analog to digital ( a / d ) and digital to analog ( d / a ) are quite usual in information processing , qubit and cv systems are nearly always treated separately .
there have been some pilot works on how to entangle two separate qubits by an entangled gaussian field , the efficient of the transfer is not high @xcite .
we would propose a scheme of perfect transferring the entanglement in this paper .
the two - mode squeezed vacuum state @xmath0 , where @xmath1 with @xmath2 the squeezing parameter .
the entanglement of the state is @xmath3 @xmath4 .
the interaction between different systems can cause the transfer of entanglement between the systems .
the scheme of the system considered is that two individual qubits each interacting with one entangled part of the field . the whole system will evolve in the way of @xmath5 , where @xmath6 $ ] is the evolution operator in interaction picture , and @xmath7 @xmath8is the initial state of the cv system while @xmath9 is the initial state of the qubit system .
firstly suppose the model hamiltonian of entanglement transfer from cv system to qubit system or vice versa is @xmath10 where @xmath11 and @xmath12 are the photon annihilation and generation operators respectively , @xmath13 , @xmath14 and @xmath15 are operators which convert the atom ( the qubit ) from its excited state @xmath16 to ground state @xmath17 and from ground state to excited state respectively .
the hamiltonian ( [ wave1 ] ) can be considered as a kind of nonlinear jaynes - cummings model@xcite @xcite . then @xmath18\left| m,-\right\rangle = \cos ( m\omega t_1)\left| m,-\right\rangle -i\sin ( m\omega t_1)\left| m-1,+\right\rangle .$ ] if the interaction time @xmath19 is adjusted in such a way that @xmath20
then @xmath18\left| 2m,-\right\rangle = ( -1)^m\left| 2m,-\right\rangle $ ] and @xmath21\left| 2m+1,-\right\rangle = -i(-1)^m\left| 2m,+\right\rangle .$ ] apply the evolution operator @xmath22 $ ] to the state@xmath23 , we have @xmath24 with @xmath25 and @xmath26 it should be noticed that the state after evolution is a product state of cv system state and two qubit state .
the cv state @xmath27 has even number of photons in each mode .
we can separate the two qubit state @xmath28 from the combined state , then append another vacuum two qubit state @xmath29of @xmath30 partite to state @xmath31 , the new state will be @xmath32 we would design another interaction hamiltonian to assign the entanglement of cv state to two qubit state
. the hamiltonian will be @xmath33the evolution will be @xmath34 with the interaction time @xmath35 @xmath36 , and @xmath37 , @xmath38 then we move from the second two qubit to the vacuum state of the third two qubit of @xmath30 partite and so on .
the @xmath39 hamiltonian will be @xmath40 $ ] and interaction time @xmath41 the whole state reads @xmath42 with @xmath43 , @xmath44 the entanglement transferred to qubits system is @xmath45 \\ & = & \log \frac{1-\lambda ^{2^{k+1}}}{1-\lambda
^2}-(\frac{\lambda ^2}{% 1-\lambda ^2}-\frac{2^k\lambda ^{2^{k+1}}}{1-\lambda ^{2^{k+1}}})\log \lambda ^2 .
\nonumber\end{aligned}\ ] ] the entanglement remained at the cv system is @xmath46the total entanglement remains unchanged for each @xmath47 , @xmath48when @xmath49 @xmath50 , thus @xmath51 the entanglement transferred to the qubit system tends to @xmath52 .
the entanglement is perfectly transferred .
the entanglement transfer is depicted in fig.1 for different value of receiving qubit pair number @xmath47 .
in the reverse conversion , we have the initial state @xmath53 @xmath54 , where @xmath55 @xmath56 @xmath57 .
the process of entanglement transfer is to transfer firstly the higher qubit pair ( @xmath58 to the cv bipartite state then the lower .
the result of conversion will be @xmath59 @xmath60 @xmath61 @xmath62 @xmath63 @xmath64 where @xmath65 @xmath66 @xmath67 @xmath68 @xmath69 @xmath70with @xmath71 @xmath72and @xmath73 we denoted @xmath74 for later use .
the entanglement of the state @xmath63 is equal to that of a state @xmath75 @xmath76 @xmath77 , thus it is equal to the sum of entanglements of qubit pairs .
we have @xmath78 the conversion procedure will convert a general qubit pair product state @xmath79 into a continuous variable state @xmath80 while keeping the entanglement due to local unitary operations .
the process of reverse entanglement conversion is to convert firstly the highest two qubits @xmath81 to the cv system then the lower .
the combined state will evolve to @xmath82 where @xmath83 is the initial state of the field ( @xmath84 and @xmath85 bipartite ) . since @xmath86 the evolution time is so chosen that @xmath87 we choose @xmath88 as before
. then @xmath89 @xmath90 .
the first step evolution will be @xmath91 @xmath92 where @xmath93 $ ] as before .
the basis of @xmath94 are @xmath95 with @xmath96 or @xmath97 we see that all the information of qubit state @xmath98 is transferred to the field , leave the two qubit state a definite blank state .
moreover , the combined state is a direct product of the field and two qubit state , thus the @xmath99 qubit pair can be dropped after the evolution .
the next step is to transfer @xmath100to the remained field @xmath101 .
since when @xmath102 we have the state after the evolution will be @xmath103 the basis of @xmath104 are @xmath105 with @xmath106 or @xmath97 the quantum state of a pair of the two qubits @xmath107 @xmath98 are transferred to @xmath108 when all the two qubits are transferred to the field , we get at last a bipartite quantum cv state @xmath109 while leaving all the two qubit series in the lower energy level state @xmath110 thus the reverse conversion procedure will convert a general qubit pair product state @xmath111 into a continuous variable bipartite state @xmath112 while keeping the entropy of the whole state .
this is due to the fact that local unitary transformation does not change the entanglement .
we here divide the system into @xmath113 and @xmath114 subsystems .
the reverse conversion is perfect .
cv werner state is a mixture of the two mode squeezed vacuum state and the two mode thermal state @xcite @xmath115 where @xmath116 the sufficient conditions of inseparability and separability as well as other physical properties have been displayed .
denote partial transposed state of @xmath117 as @xmath118 the eigenvalues of @xmath119 are @xcite @xmath120 the logarithmic negativity ( ln ) as an entanglement measure of the state @xmath121 can be calculated @xcite .
the entanglement measured by ln is @xmath122 where @xmath123tr@xmath124 in the simplest case of @xmath125 ( although ln can be worked out for any values of the parameters ) , the inseparable condition is @xmath126we have @xmath127.\ ] ] the conversion of the cv werner state to a serial of two qubits system leads to the result state @xmath128 where @xmath129 and @xmath130 is the similar . for convenience ,
we denote @xmath17 and @xmath131 as @xmath132 and @xmath133 then the basis of @xmath134 system ( as well as @xmath135 system ) can be denoted as @xmath136 with @xmath137 the basis can be further simplified as @xmath138 with @xmath139 thus @xmath140 the partially transposed matrix @xmath141 has a block diagonal form with @xmath142blocks in one - dimensional subspaces spanned by vectors @xmath143 , @xmath144 and @xmath145 blocks in two - dimensional subspaces spanned by vectors @xmath146 , @xmath147 consequently , the eigenvalues of the partially transposed matrix @xmath141 can easily be calculated as roots of quadratic equations and read @xmath148 after the conversion , the entanglement measured by ln for the case of @xmath149 will be @xmath150.\ ] ] the entanglement transfer is shown in fig.2 for different value of receiving qubit pair number @xmath47 .
the entanglement of continuous variable system can be converted to the entanglement of a serial of qubit pairs .
the conversion error can be made arbitary low if the series is long enough . the reverse conversion is perfect as far as the different qubit pairs are not correlated before the reverse conversion . for a typical two mode squeezed vacuum state prepared in laboratary ( the squeezing parameter @xmath151 )
, the qubit system with eight pairs of qubits is sufficiently good in simulation of the original cvsystem .
this is indicated by our calculations of the entanglement transfer of two mode squeezed vaccum state or cv werner state . in quantum information theory ,
the entanglement of two qubits changes from @xmath152 to @xmath153 this flexibility makes the error of conversion quite small as we can see from the figures .
funding by the national natural science foundation of china ( grant no .
10575092,10347119 ) , zhejiang province natural science foundation ( grant no
. rc104265 ) and aqsiq of china ( grant no .
2004qk38 ) are gratefully acknowledged . | we investigate how entanglement can be transferred between continuous variable and qubit systems .
we find that a two - mode squeezed vacuum state and a continuous variable werner state can be converted to the product states of infinitive number of two - qubit states while keeping the entanglement .
the reverse process is also possible .
pacs : 03.67.mn;03.65.ud |
recently , it is believed that the quark - gluon plasma ( qgp ) just above the chiral and deconfinement phase transitions is an unexpectedly strongly interacting system , which is mainly based on the facts that the created matter at rhic behaves like a perfect fluid @xcite and that some mesonic bound states of heavy quarks can survive above the critical temperature ( @xmath3 ) from lattice qcd studies @xcite . because the fundamental degrees of freedom in qgp are quarks and gluons , it is also important to study their properties in such a strongly interacting system . here
, we investigate the quark spectrum just above @xmath3 of the chiral transition , focusing on the precursory soft modes .
it is known that the soft modes exist over a wide range of temperature above @xmath3 owing to a strong coupling nature between quarks @xcite . in this work ,
we show that they affect the quark spectrum significantly in a region just above @xmath3 .
in particular , it is shown that the quark spectrum form three - peaks at low energy and low momentum region near @xmath3 @xcite . in sec.[njl ] , using a chiral effective model , we show that the quark spectrum near @xmath3 forms the three - peak structure owing to a coupling with the soft modes .
such a three - peak structure is seen more clearly if the soft modes are replaced by an elementary massive boson @xcite . in sec.[ykw ] , a detailed analysis on the quark spectrum is given using yukawa models with a massless quark and a massive scalar ( pseudoscalar ) or vector ( axialvector ) boson .
we elucidate the mechanism of the formation of the three - peak structure through which the new quark collective excitations are realized in terms of the landau damping of a quark ( an antiquark ) induced by scattering with the thermally excited boson , which gives rise to mixing and hence a level repulsion between a quark ( antiquark ) and an antiquark hole ( quark hole ) .
brief summary and concluding remarks are given in sec.[sum ] .
to describe the chiral transition and the fluctuations , we first employ the two - flavor nambu jona - lasinio model in the chiral limit @xmath4,\end{aligned}\ ] ] with the coupling constant @xmath5 gev@xmath6 and the three dimensional cutoff @xmath7 mev taken from ref .
this model gives a second order phase transition at @xmath8mev for vanishing quark chemical potential .
the fluctuations of the chiral condensate are described by the quark - antiquark green function in the random phase approximation , @xmath9 where @xmath10 is the matsubara frequency for bosons and @xmath11 is the undressed quark - antiquark polarization function at one - loop . to evaluate strengths of the fluctuations ,
we employ the spectral function , @xmath12 , given by @xmath13 we show @xmath14 near @xmath3 in the left panel of fig .
[ fig : imd ] .
one can see that there appear pronounced peaks which denote the precursory soft modes @xcite .
the peak positions of the modes are approximately expressed as @xmath15 . a @xmath16-dependent ` mass '
@xmath17 becomes smaller as @xmath16 approaches @xmath3 , which means the softening at @xmath3 . in the quark - antiquark channel ( left ) and
the quasi - quark spectral function @xmath18 ( right ) for @xmath19 and @xmath20.,title="fig:",scaledwidth=45.0% ] in the quark - antiquark channel ( left ) and the quasi - quark spectral function @xmath18 ( right ) for @xmath19 and @xmath20.,title="fig:",scaledwidth=45.0% ] the effect of the soft modes on the quark spectrum is incorporated in the quark self - energy in a non - selfconsistent way , @xmath21 where @xmath22 is the free quark propagator with @xmath23 being the matsubara frequency for fermions .
the quasi - quark and quasi - antiquark spectral functions , @xmath24 , are obtained from the retarded self - energies , @xmath25 $ ] , respectively , i.e. @xmath26^{-1},\ ] ] with the analytic continuation @xmath27 and the projection operators @xmath28 .
the quasi - quark spectral function @xmath18 for @xmath19 mev and @xmath20 is plotted in the right panel of fig .
[ fig : imd ] .
we see a clear three - peak structure at low momentum , which exists even at @xmath29 @xcite .
although not shown in the figure , the quasi - antiquark spectrum , @xmath30 , has also a three - peak structure for a relation , @xmath31 .
the mechanism of the appearance of the three - peak structure in @xmath18 is as follows : the imaginary part of @xmath32 has two peaks at nonzero values of @xmath33 , which means that there exist two large damping modes of the quasi - quark there . from a kinematical consideration which is explained in detail in the following section , we see that one is a collision of a thermally excited antiquark and the quasi - quark creating the soft mode , and the other is a collision of the quasi - quark and the soft mode creating an on - shell quark .
both the processes are interpreted as a landau damping of the quasi - quark .
the point is that the quasi - quark is a mixed state between quarks and ` antiquark - holes ' which are annihilation of thermally excited antiquarks and have the positive quark number .
then , these damping modes cause a mixing between quarks and antiquark - holes .
this mixing mechanism can be described in terms of the resonant scattering as in the case of the color superconductivity @xcite , although a crucial difference arises owing to the different nature of the soft modes .
we can show that a coupling with the soft mode with _ a nonzero mass @xmath17 _ is essential for the appearance of the three - peak structure in the quark spectrum , as will be explained below in yukawa models .
as mentioned above , the soft modes of the chiral transition have the character of a well - defined elementary boson with a mass @xmath17 and a small width near @xmath3 .
thus it is seen that a coupling with an elementary massive boson is favorable for the three - peak structure of the quark spectrum . in this section , we show that a system composed of a massless quark and a massive boson , as described by yukawa models , exhibits the three - peak structure in the quark spectral function @xcite .
we first investigate the spectrum of a massless quark coupled with a massive scalar ( pseudoscalar ) boson at finite @xmath16 in a yukawa model : @xmath34 @xmath35 is the coupling constant and @xmath36 is the boson mass .
the quark self - energy in the imaginary time formalism at the one - loop order is expressed as @xmath37 where @xmath38^{-1}$ ] and @xmath39^{-1}$ ] are the matsubara green functions for the free quark and scalar boson , respectively , and @xmath40 and @xmath41 are the matsubara frequencies for the fermion and boson , respectively . for the coupling with a pseudoscalar boson , a factor @xmath42 is added in both sides of @xmath43 in eq .
( [ eq : tildesigma ] ) .
such a factor is , however , canceled out because it anti - commutes with @xmath43 and thus the self - energy becomes the same form as eq ( [ eq : tildesigma ] ) .
therefore , the following results and discussion in this section hold also for the coupling with a pseudoscalar boson . for several temperatures ( left ) and the kinetic processes in im@xmath44 .
@xmath45 represents the quasi - quark , @xmath46 on - shell free quarks and @xmath47 the on - shell scalar boson .
the incident on - shell particles are thermally excited particles .
, title="fig:",scaledwidth=54.0% ] for several temperatures ( left ) and the kinetic processes in im@xmath44 .
@xmath45 represents the quasi - quark , @xmath46 on - shell free quarks and @xmath47 the on - shell scalar boson .
the incident on - shell particles are thermally excited particles . , title="fig:",scaledwidth=40.0% ] after the matsubara summation and the analytic continuation , the retarded self - energy is obtained . in the left panel of fig .
[ fig : imsigma_p0 ] , we show im@xmath48 , which is independent of @xmath35 . we see that it vanishes at @xmath49 , irrespective of @xmath16 .
the vanishing decay rate at @xmath50 is owing to the suppression of the phase space ; the energy - momentum conservation requires the zero momentum of the on - shell ( anti)quark for each process at @xmath50 . around @xmath51 ,
on the other hand , the distribution functions suppress the decay rate , because the on - shell energies of the boson and the quark go to infinity as @xmath52 . at finite @xmath16 , the landau damping , the processes ( ii ) and ( iii ) in the right panel of fig .
[ fig : imsigma_p0 ] , comes to play and im@xmath53 have supports in the regions @xmath54 for ( ii ) and @xmath55 for ( iii ) .
these two peaks in these regions grow rapidly as @xmath16 is raised , and around @xmath56 , the landau damping is dominant .
( left ) and the quark spectral function @xmath57 ( right ) for @xmath58 .
the dotted straight line in the left panel denotes @xmath59 .
, title="fig:",scaledwidth=45.0% ] ( left ) and the quark spectral function @xmath57 ( right ) for @xmath58 .
the dotted straight line in the left panel denotes @xmath59 . ,
title="fig:",scaledwidth=52.0% ] the corresponding quark spectral function around for @xmath58 in the right panel of fig .
[ fig : self ] shows a peculiar behavior .
there appear broad peaks both in the positive- and negative - energy regions as well as the narrow peak around the origin for lower momenta : thus the spectral function in the low - momentum region has _ three peaks_. to understand the behavior of @xmath60
, we show @xmath61 for @xmath58 in the left panel of fig .
[ fig : self ] .
we see that the oscillatory change of the real part around @xmath62 is seen , reflecting the peaks in the imaginary part . it is noted that in general , a sharp peak in the spectral function can be formed at @xmath63 if the quasi - dispersion relation , @xmath64 , and a relation @xmath65 are hold . in order to find out the quasi - dispersion relation from fig .
[ fig : self ] , we draw the line @xmath59 . for vanishing momentum , we see that there appear five crossing points of re@xmath66 and @xmath59 .
the crossing points with the second and fourth largest @xmath33 , however , are located at the energies of the peak of @xmath67 and hence the spectral function does not form a peak there .
the other three crossing points correspond to the three peaks in the spectral function . for large momenta ,
the number of the crossing points decreases and eventually only one crossing point remains . at higher temperature ,
the strength of the peak around the origin is getting weaker , and eventually there remain the other two peaks which correspond to the normal quasi - quark and the plasmino in the htl approximation . to understand the physical mechanism responsible for the three - peak structure of @xmath60 , we first recall that there develop two peaks in @xmath68 which correspond to the decay processes ( ii ) and ( iii ) . the process ( ii )
is the annihilation process of the incoming quark @xmath45 and the thermally excited antiquark into a boson in the thermal bath , @xmath69 , and its inverse process .
two remarks are in order here .
first , the disappearance of an anti - quark implies the creation of a ` hole ' in the thermally excited anti - quark distribution@xcite .
second , the creation of bosons in a thermal bath is enhanced in comparison with the case in vacuum by a statistical factor of @xmath70 , which becomes large when @xmath16 is comparable to @xmath36 .
thus , we see that the process ( ii ) causes a virtual mixing between the quark and ` anti - quark hole ' states and as a result , the mixing is enhanced when @xmath71 .
the process ( iii ) is another decay process of a quasi - quark state @xmath45 , which is now understood to be a mixed state of quarks and antiquark - holes , into an on - shell quark via a collision with a thermally excited boson : @xmath72 and its inverse process .
these processes again give rise to a mixing of a quasi - quark and an anti - quark hole state .
the mechanism for the mixing of the quark and hole state can also be characterized as a _
resonant scattering _
@xcite , which was originally introduced to understand the non - fermi liquid behavior of fermions just above the critical temperature of the superconducting .
in fact , we have seen that the process ( ii ) includes a scattering process of the quark by a massive boson , thus creating a hole state in the thermally distributed anti - quark states : @xmath73 .
such a process is called resonant scattering@xcite .
note that the most probable process for finite @xmath16 involves the lowest energy state of the boson , i.e. , a rest boson with a energy @xmath36 .
the energy conservation law in the most probable case for the above process is @xmath74 , or equivalently , @xmath75 .
this equation actually represents the energy - momentum relation for the particles involved in the state mixing .
thus we see that the physical energy spectrum is obtained as a result of the level repulsion between the energies @xmath76 and @xmath77 in the perturbative picture .
this situation is schematically depicted in the upper right part of fig .
[ fig : reso ] .
similarly , the process ( iii ) includes the process @xmath78 , and the energy - momentum conservation law for this process for the most probable case is @xmath79 .
thus , the physical energy spectrum is obtained as a result of the level repulsion between the energies @xmath80 and @xmath81 .
this situation is also depicted in the lower - right part in fig .
[ fig : reso ] .
we thus find that at temperatures satisfying @xmath82 , owing to the finite boson mass , the level repulsions occur far from the origin and two gap - like structures in the quark spectrum are formed at positive and negative energies , as shown in fig .
[ fig : reso ] . at the high temperature limit , @xmath83 , or @xmath84 , the effect of the boson mass can be ignored , and the resonant scattering occurs only once at the origin ( @xmath85 ) , since the energy levels which are to be repelled cross only there .
then the quark spectrum has two peaks in the positive and negative energy region , which is realized in the htl approximation .
it was shown in the previous section that the quark spectral function possesses a three - peak structure near @xmath3 of the chiral transition when the chiral soft mode is incorporated into the quark self - energy .
recall that the soft modes behave like a massive elementary boson with a mass @xmath17 as @xmath16 approaches @xmath3 , i.e. @xmath86 , and hence the quark spectra in such a case are essentially the same as that treated in this section .
we also note that as @xmath16 is lowered toward @xmath3 , @xmath17 tends to vanish , and hence the ratio @xmath87 becomes large .
thus the quark spectrum approaches that in the @xmath88 limit of the yukawa model . in this subsection
, we investigate the quark spectrum using a yukawa model with a massive vector ( axial - vector ) boson and show that the three - peak structure of the quark spectral function is obtained .
we start from the following lagrangian , composed of a massless quark @xmath89 and vector boson field @xmath90 : @xmath91 with the field strength @xmath92 . at one - loop order ,
the quark self - energy in the imaginary time formalism is given by @xmath93 with the matsubara propagator for the massive vector boson , @xmath94 where @xmath95 .
for coupling with an axial - vector boson , as the self - energy has the same form as eq .
( [ eq : sigmavimag ] ) , the following results hold also . in the left panel of fig .
[ fig : imsigmap_p0 ] , we plot im@xmath96 at @xmath97 and @xmath98 .
we fix the coupling constant at @xmath99 .
the qualitative features of both parts for @xmath100 are quite similar to those in the left panel of fig .
[ fig : imsigma_p0 ] : there are two clear peaks in @xmath101 for @xmath100 and they grow rapidly as @xmath16 increases . for @xmath102 , it is seen that @xmath103 grows more rapidly than in the yukawa model with the scalar boson shown in fig .
[ fig : imsigma_p0 ] . for several temperatures ( left ) and the quark spectral function @xmath57 for @xmath104 ( right ) .
, scaledwidth=48.0% ] in the right panel of fig .
[ fig : imsigmap_p0 ] , we plot the quark spectral function for @xmath105 .
it is seen that there appears a three - peak structure , as found in the previous subsection .
this is quite natural , and it would be expected from the behavior of the self - energy .
a clear three - peak structure is formed at lower values of @xmath16 than in the scalar boson case , which comes from the difference of the matrix elements between the scalar and vector boson .
we have investigated the quark spectrum near but above the critical temperature of the chiral phase transition taking the effect of the fluctuations of the chiral order parameter into account .
we have shown that for @xmath106 the quark spectrum has a three - peak structure at low frequency and low momentum .
the mechanism of the formation of the three - peak structure is based on the fact that the soft modes which are composite system of quark - antiquarks acquire the character of the well - defined elementary bosonic excitation when @xmath16 is close to @xmath3 .
then , we have elucidated the essential mechanism of the formation of the three - peak structure quantitatively in yukawa models composed of a massless quark and a massive boson with a mass @xmath36 at finite @xmath16 , in which the mass of the boson field is varied by hand .
we have found that the quark spectral function at low frequency and low momentum has a three - peak structure at intermediate values of @xmath16 , i.e. for @xmath71 , which is independent of the species of the massive boson . among them , the two peaks have finite thermal masses and approach the normal quasi - quark and plasmino excitations in the htl approximation in the high @xmath16 limit , while the strength of the other peak around the origin becomes weaker and disappears in this limit .
we have discussed the fact that the three - peak structure originates from the landau damping , that is , the scattering processes of a quark and an antiquark hole of a thermally excited antiquark , and of an antiquark and a quark hole of a thermally excited quark .
they cause the formation of energy gaps in the quark spectrum , owing to the level mixing between the quark ( antiquark ) and the hole of the thermally excited antiquarks ( quarks ) .
this leads to a level repulsion or gap in the resultant physical spectrum .
these mixings can be understood in terms of the resonant scattering @xcite of the quasi - quarks off the bosons .
in particular , owing to the mass of the boson , the level repulsion due to the resonant scattering occurs _ twice _ at different points in the energy - momentum plane , leading to the three - peak structure of the spectral function .
this contrasts with the case of the quark spectrum coupled to a massless boson , such as a gauge boson , in which the resonant scattering occurs only at the origin in the energy - momentum plane and then leads to only two peaks , as in the spectrum in the htl approximation @xcite . the soft modes for the chiral transition behave like a massive elementary boson with a mass @xmath17 as @xmath16 approaches @xmath3
, i.e. @xmath107 , and hence the quark spectra in such a case are essentially the same as that treated in the yukawa models .
we also note that as @xmath16 is lowered toward @xmath3 , @xmath17 tends to vanish , and hence the ratio @xmath87 becomes large .
thus the quark spectrum approaches that in the @xmath88 limit of the yukawa models . in this work ,
we have considered only massless quarks .
a finite quark mass should affect the formation of the three - peak structure in the quark spectral function .
thus , the incorporation of the quark mass effect is important for describing the chiral transition precisely , because the order of the transition , and hence , the quark spectrum near the critical point are sensitive to the quark mass .
a finite quark mass also leads to a difference between the scalar ( vector ) and pseudoscalar ( axial - vector ) boson cases , and may suppress the three - peak structure .
detailed study of such a effect is now under way and will be reported elsewhere@xcite .
is supported by the 21st century coe program at nagoya university and the jsps grand - in - aid for scientific research ( # 18740140 ) .
is supported by a grand - in - aid for scientific research by monbu - kagakusho of japan ( # 17540240 ) .
this work is supported by the grand - in - aid for the 21st century coe
center for diversity and university in physics " of kyoto university . i. arsene et al .
( brahms collaboration ) , nucl .
phys . a * 757 * ( 2005 ) , 1 . + b. b. back et al . ( phobos collaboration ) , nucl . phys . a * 757 * ( 2005 ) , 28
. + j. adams et al .
( star collaboration ) , nucl .
a * 757 * ( 2005 ) , 102 .
+ k. adcox et al .
( phenix collaboration ) , nucl .
a * 757 * ( 2005 ) , 184 .
t. umeda , r. katayama , o. miyamura and h. matsufuru , int .
a * 16 * ( 2001 ) , 2215 .
+ m. asakawa and t. hatsuda , phys .
* 92 * ( 2004 ) , 012001 .
+ s. datta , f. karsch , p. petreczky and i. wetzorke , phys .
d * 69 * ( 2004 ) , 094507 .
+ t. umeda , k. nomura and h. matsufuru , eur .
j. c * 39s1 * ( 2005 ) , 9 .
+ a. jakovac , p. petreczky , k. petrov and a. velytsky , phys .
d * 75 * ( 2007 ) , 014506 .
+ g. aarts , c. allton , m. b. oktay , m. peardon and j. i. skullerud , arxiv:0705.2198 [ hep - lat ] . t. hatsuda and t. kunihiro , phys .
b * 145 * ( 1984 ) , 7 ; phys .
* 247 * ( 1994 ) , 221 .
m. kitazawa , t. kunihiro and y. nemoto , prog .
phys . * 117 * 103 ( 2007 ) , 103 .
m. kitazawa , t. kunihiro and y. nemoto , phys .
b * 631 * ( 2005 ) , 157 .
h. a. weldon , phys .
d * 40 * ( 1989 ) , 2410 ; physica a * 158 * ( 1989 ) , 169 ; phys .
d * 61 * ( 2000 ) , 036003 . | we investigate the quark spectrum near but above the critical temperature of the chiral transition , taking into account the precursory soft modes . it is found that there appear novel excitation spectra of quasi - quarks and quasi - antiquarks with a three - peak structure . by a detailed analysis on the formation of the three - peak structure using yukawa models
, it is shown that the new quark spectra originate from the mixing between a quark ( anti - quark ) and an antiquark hole ( quark hole ) caused by a resonant scattering of the quasi - quark with the soft modes which have a small but finite excitation energy with a small width near the critical temperature .
@xmath0yukio nemoto , @xmath1masakiyo kitazawa and @xmath2teiji kunihiro _ @xmath0department of physics , nagoya university , nagoya 464 - 8602 , japan _ _ @xmath1department of physics , osaka university , toyonaka , osaka 560 - 0043 , japan _ _ @xmath2yukawa institute for theoretical physics , kyoto university , kyoto 606 - 8502 , japan _ 1.3em |
the unified model of active galactic nuclei ( agn ; @xcite ) states that most of the observational differences between type-1 and type-2 seyfert - like galaxies arise from an orientation effect . according to this theory ,
the disappearance of ultraviolet ( uv ) and optical broad emission features at type-2 inclinations can be explained by the presence of an obscuring , circumnuclear material along the equatorial plane of the agn ( the so - called dusty torus ) that hides both the central engine and the photoionized broad line regions ( blrs ; the low ionization line lil blr and the highly ionized line hil blr ; @xcite ) . a type-2 viewing angle
can then be defined as a line of sight towards the central source that intercepts the equatorial dust , while type-1 inclination allows a direct view of the central engine .
the observational lack of type-1 agn with edge - on host galaxy @xcite suggests that dust along seyfert 1 galaxy discs may obscure the hil and lil blr and make the agn appear like seyfert 2s @xcite .
the number count of seyfert 1 objects is thus expected to be small , even if the fraction of type-1 against type-2 agn in the nearby universe still needs to be properly determined .
estimating the orientation of a large sample of agn is necessary to verify the assumptions of the unified model , and check whether all the differences between seyfert 1 and seyfert 2 objects can be explained by inclination or if morphological differences must also be taken into account . in this
regard , polarization has proven to be a solid tool to investigate the inner structure of agn .
the spectropolarimetric measurements of ngc 1068 by @xcite helped to identify electron and dust scattering as the main mechanisms producing a continuum polarization in radio - quiet agn .
going further , the extensive high - resolution , high signal - to - noise spectropolarimetric observation of the same agn by @xcite revealed the presence of highly polarized , broad , symmetric balmer and permitted fe ii lines .
the polarization spectrum was found to be closely similar to typical seyfert 1 galaxies , supporting the idea that seyfert 2 agn are hiding seyfert 1 core behind the dusty torus .
this discovery was a key argument in favour of a unified model of agn .
spectropolarimetry is thus a powerful method to probe the validity of any agn model , as the computed fluxes shall match both observational intensity , polarization percentage and polarization angle , reducing the number of free parameters / degeneracies @xcite . in order to model a peculiar source ,
the observer s viewing angle @xmath4 has to be set ( e.g. @xmath5 70@xmath2 for ngc 1068 ; @xcite ) to explore the resulting polarization @xcite .
the impact of the system s orientation on to the net polarization can lead to significantly different results , especially when the observer s line of sight matches the half - opening angle of the obscuring region @xcite . to be consistent with observation , an investigation of the model over a broad range of inclination must be undertaken .
however , any comparison between the observed polarization and the theoretical orientation of individual agn is hampered by the lack of a data base that combines inclination and polarization .
it is the aim of this paper to provide the first spectropolarimetric compendium of seyfert - like galaxies , gathering observed continuum polarizations from literature correlated with estimated inclinations of individual agn . in sect .
[ comp ] , i investigate the different observational and numerical techniques used to estimate the inclination of 53 objects , and match the sources with their archival uv / optical polarization measurement , whenever it is feasible . to illustrate the significance of this catalogue when comparing models to observations , in sect .
[ analysis ] , i pick four different , competitive agn models from the literature and analyse them in the framework of this compendium . in sect .
[ discussion ] , i review the successes and potential improvements of agn modelling , explore the problematic , high polarization levels of peculiar type-1 objects and discuss possible bias on the estimation of inclination . finally , conclusions are drawn in sect .
[ conclusion ] .
while polarimetric measurement of agn are difficult to obtain due to intrinsically low polarization degrees in type-1 objects @xcite and flux dilution by unpolarized starlight in type-2s @xcite , it is more straightforward to measure polarization than to estimate the inclination of the system , as orientation is not easily derived directly from observations . hence , the following sections present different techniques to infer the orientation of agn ( with potential caveats discussed in sect . [
discussion : bias ] ) , the selection criteria used to select / remove estimated inclinations , corrections that have to be made to most of the seyfert-2 agn and the final compilation of data . for the remaining of this paper , the terminology `` inclination of the system '' will refer to the nuclear , not the host , inclination .
method a. based on the tight correlation found by @xcite , @xcite and @xcite between the mass of the central supermassive black hole and the bulge velocity dispersion in nearby galaxies , @xcite and @xcite developed a method to derive the inclination angle of nearby seyfert 1 agn . by assuming a keplerian motion of the lil blr and a similar mass / velocity dispersion between type-1 agn and regular galaxies , they estimated the orientation angles @xmath4 for a variety of seyfert 1 objects , with associated errors calculated from the uncertainties of both the black hole mass ( obtained by reverberation mapping techniques ; @xcite ) and the measured velocity dispersion .
method b. the inclination estimation of the seyfert 1 galaxy eso 323-g077 comes from the optical spectropolarimetric measurement achieved by @xcite , who detected very high levels of linear polarization ( up to 7.5 % at 3600 ) . within the framework of the unified model ,
those levels are inconsistent with the polarization degrees produced by an object seen in the polar orientation @xcite .
@xcite argued that the system must be partially hidden by the dusty torus and tilted by @xmath6 45@xmath2 with respect to the observer s line of sight to produce such a high polarization degree .
the same method was previously applied to fairall 51 ( continuum polarization 4.12 % @xmath7 0.03 % ) by @xcite , who also derived an inclination of @xmath6 45@xmath2 .
method c. an increasing amount of x - ray bright , type-1 agn shows an asymmetrically blurred emission feature at 6.4 kev , associated with iron fluorescence in near - neutral material @xcite .
interestingly , the line broadening caused by doppler effects and gravitational plus transverse redshifts can be used to numerically probe the inclination of the system @xcite . in @xcite , this characteristic line profile
is equally fitted within a schwarzschild or a kerr metric ( even if recent modelling seems to favour maximally rotating black holes in the centre of type-1 agn ; @xcite ) , giving a mean seyfert 1 inclination of 30@xmath2 .
method d. constraints on the inclination of ngc 1097 are derived by @xcite , who applied an eccentric accretion ring model to the observed broad , double - peaked balmer emission lines . between 1991 and 1996 ,
the double - peaked h@xmath8 line of ngc 1097 evolved from a red - peak dominance @xcite to a nearly symmetrical profile @xcite and up to a blue - peak dominance @xcite .
this line profile evolution can be explained by a refinement of the precessing , planar , elliptical accretion - ring model developed by @xcite and @xcite , to fit the data using an eccentric accretion disc inclined by 34@xmath2 .
method e. @xcite measured the two - dimensional distribution and kinematics of the molecular , ionized , and highly ionized gas in the inner regions of a sample of radio - quiet , type-1 agn using high spatial resolution , near - infrared ( ir ) spectroscopy . based on a model developed by @xcite , they assumed that a gravitational well , created by the combined action of a central supermassive black hole and a distant stellar population , is driving the circular motion of a coplanar thin disc , reproducing the observed emission line gas kinematics . exploring four free parameters ( disc inclination , position angle of its major axis , black hole mass and mass - to - light ratio )
, @xcite statistically estimated the inclination of ngc 3227 , ngc 4151 and ngc 7469 using a bootstrap gaussian fit @xcite .
method f. to determine the inclination of a sample of nearby agn , @xcite explored the three - dimensional geometry and kinematics of the narrow - line regions ( nlrs ) of agn , observing both type-1 and type-2 objects .
resolved by [ o iii ] imaging and long - slit spectroscopy , most of the agn show a bi - conical structure which can be morphologically and kinematically constrained .
however , to extract information about the agn orientation , a kinematic model must be generated .
using uniform , hollow , bi - conical models with sharp edges , @xcite were able to statistically derive a set of morphological parameters ( including orientation ) for 17 objects out of 53 .
method g. relatively bright and situated in the nearby universe , ngc 1068 is an archetypal seyfert 2 galaxy , observed during the last fifty years .
taking advantage of past near and mid - ir photometric and interferometric observations @xcite , @xcite applied a three - dimensional radiative transfer code to a clumpy , dusty structure in order to reproduce the observed spectral energy distribution ( sed ) . among new constraints on the bolometric luminosity and the ir optical depth of the torus , @xcite estimated the overall inclination of ngc 1068 to be close to 70@xmath2 .
method h. @xcite examined the generation of c iv line profiles in broad absorption line ( bal ) quasars using a two - component wind model . by modelling a structure based on stellar wind laws and composed of axisymmetric , polar and equatorial outflows filled with 2-level atoms , they succeeded to reproduce a large set of bal profiles and concluded that the viewing angle to the wind is generally large .
unfortunately , degeneracies in line profile fitting do not allow stronger constraints .
method i. finally , @xcite investigated their own polarimetric and photometric observations of the type-2 quasar iras 13349 + 2438 in the context of an axisymmetric distribution of scatterers to explain the alignment of polarization with the major axis of the host galaxy . using a model of a dusty disc parallel to the plane of the galaxy , similar to a usual dusty torus , surrounding the continuum source and the lil blr
, they showed that both the observed polarization in the continuum and in the broad h@xmath8 line could be reproduced if the inclination of the observer is about 52@xmath2 with respect to the symmetry axis of their model .
once inclinations are obtained , i match them with uv / optical spectropolarimetric measurements , whenever it was possible . the methods presented in sect . [
comp : incl ] derive about 100 agn orientations but only 53 of them have published continuum polarization measurements . moreover , not all of the estimated inclinations are unique and a selection has to be done whenever two methods , or more , give different estimations for the same agn . tab .
[ table : reject ] lists discrepancies of duplicate inclinations .
reasons for the choice of a given inclination are discussed below . .
the inclination of 3c 120 found by @xcite , @xmath4 = 88@xmath9 , is rejected in favour of the estimation by @xcite , who found @xmath4 = 21@xmath10 , a result in a better agreement with the type-1 classification of 3c 120 emission @xcite . ] .
fairall 9 . due to the huge error bars derived by @xcite on the orientation of fairall 9 ( 89@xmath11 ) , covering the full permitted range of inclination for a type-2 object plus a fraction of the permitted range of type 1s ,
the viewing angle computed by @xcite is favored .
ic 4329a .
the inclination derived by @xcite in the case of ic 4329a ( 5.0@xmath2 ) is compatible within the error bars of the estimation computed by @xcite .
the latest ( 10@xmath12 ) is thus selected in order to concur with the two values .
mrk 279 . estimated inclinations of mrk 279 are rather different between @xcite , @xmath4 = 13.0@xmath2 , and @xcite , @xmath4 = 35.0@xmath2 , especially since they do not have overlapping error bars .
the inclination derived by @xcite is favoured as @xcite were not able to recover the measured stellar velocity dispersion of mrk 279 and had to estimate it from the [ o iii ] emission line , introducing another potential bias in their final agn orientation .
mrk 509 .
the inclination estimated by @xcite covers the whole range of inclination possible for an agn ( 89@xmath13 ) and therefore does not make much sense .
the inclination of mrk 509 by @xcite is thus selected .
ngc 3227 .
there are four different estimations for the viewing angle of ngc 3227 .
two have very large error bars @xcite overlapping the two others estimates by @xcite and @xcite . since the two later inclinations are very similar but from totally different estimation methods , they are likely to be representative of the real inclination of ngc 3227 .
hence i favour the one of @xcite , 14.2@xmath2 @xmath7 2.5@xmath2 , which agrees with the value found by @xcite , 15.0@xmath2 .
ngc 3516 .
the values estimated by @xcite , @xmath4 = 26@xmath14 , and @xcite , @xmath4 = 38.3@xmath2 @xmath7 7.6@xmath2 , are not overlapping but still very close to each other .
however , provided that @xcite had to artificially estimate the errors on the black hole mass for ngc 3516 while @xcite derived it from their simulation , the estimation made by @xcite is used .
ngc 3783 .
similarly to the case of mrk 279 , estimations of the orientation of ngc 3783 are rather different between @xcite and @xcite and , for the same reasons , the value derived by @xcite is selected . the inclination recovered by @xcite for ngc 3783 , covering the whole possible inclination range for type-1 objects ,
is discarded .
ngc 4051 .
both estimations made by @xcite , @xmath4 = 25@xmath15 , and @xcite , @xmath4
= 19.6@xmath16 , are compatible and their error bars nearly fully overlap .
the viewing angle taken from @xcite having slightly higher error bars , this value is chosen to fully concur with the estimations from @xcite and to be representative of the value derived by @xcite , 10.0@xmath2 . ngc 4151 .
the inclination angle of ngc 4151 derived by @xcite is not considered due to its huge error bars that cover two thirds of the possible agn inclinations ( 60@xmath17 ) .
the overlapping values , from different methods , found by @xcite , 9@xmath18 , and @xcite , 19.8@xmath2 @xmath7 2.9@xmath2 , exclude the fourth one derived by @xcite , 45.0@xmath2 .
finally , the estimation made by @xcite is preferred as it fully covers the potential inclination calculated by @xcite .
ngc 5548 .
similar to the case of mrk 509 , the inclination of ngc 5548 evaluated by @xcite , @xmath4 = 10@xmath19 , is rejected in favour of the one derived by @xcite , @xmath4 = 47.3@xmath20 .
ngc 7469 .
the viewing angle derived by @xcite is consistent with the one derived by @xcite , within the error bars of the later .
the inclination of ngc 7469 estimated by @xcite is not worth considering due to its huge error bars ( 20@xmath21 ) .
ngc 1068 . finally , the estimations of @xcite and @xcite are different and do not overlap . however , the inclination derived by @xcite is supported by the three - dimensional structure of the nuclear region of ngc 1068 reconstructed by @xcite using a different set of observations .
as @xcite derived a similar , @xmath6 70@xmath2 , inclination , the estimation of @xcite is thus favoured . similarly to the inclination estimates in the previous section , i have found for a very few number of objects several continuum polarization measurements achieved by different authors .
the level of polarization and the polarization position angle ( measured from north through east ) were coherent between the different observing campaigns , regardless of the epoch .
i have thus favoured polarimetric measurement from seyfert atlases @xcite . however , not all of these measurements are reliable .
most , if not all , type-2 agn are dominated by relatively large , unpolarized starlight fluxes @xcite . removing the contribution from old stellar populations
drastically increases the resulting continuum polarization but previous polarimetric type-2 atlases still recorded low , 1 3 % , polarization degrees @xcite .
such behaviour is in disagreement with the unified model , where radiation escapes from the inner parts of the obscuring equatorial torus by perpendicular scattering into the polar outflows , carrying a large amount of polarization @xcite .
high polarization percentages are thus expected , but not observed , for type-2 objects @xcite .
@xcite reported that even after starlight subtraction , the remaining continuum is dominated by another component responsible for dilution of the polarized flux , now identified as originating from starburst regions . to estimate the continuum polarization of a given object after corrections for the interstellar polarization and dilution by host galaxy starlight
, @xcite proposed to measure the equivalent width ( ew ) of the broad balmer lines in both intensity and polarized flux spectra , since polarized flux spectra have the advantage to suppress low polarization emission from starlight and narrow emission lines .
if no additional unpolarized , or very little polarized , continuum superimposes on the polarization originating from the scattered light alone , the intrinsic polarization in the line and the adjacent continuum should be equal .
this is the case for ngc 1068 , where the broad lines have the same polarization as the continuum @xcite .
unfortunately , nearly all the remaining measurements of type-2 agn polarization are likely to be biased downward . in
the following i revisit and , when necessary and possible , revise the estimated continuum polarization of type-2 objects to be included in the compendium . the best way to estimate the continuum polarization due only to
scattered light is to divide the polarized flux by the total flux across the broad emission lines , as suggested by @xcite and @xcite .
only the broad lines polarization is a reliable indicator of the polarization of the scattered component .
however , by definition , broad lines are not detected in the total flux spectra of type-2s . to overcome this problem ,
i compare the typical ew of type-1 polarized , broad , balmer emission lines h@xmath8 @xmath226563 ( ew @xmath23 400 ; @xcite ) and h@xmath24 @xmath224861 ( ew @xmath23 80 ; @xcite ) to the ew of the polarized h@xmath8 @xmath226563 and h@xmath24 @xmath224861 lines of the compendium - selected seyfert 2s .
this is a first - order approximation , but it is justified by the fact that the broad lines have the same ew in polarized flux in type-1 and type-2 agn @xcite . if the ratio is equal to unity , the continuum polarization is correct ; if larger than unity , the continuum must be revised by the same factor . in the case where no spectra are available to estimate the ew of broad emission lines in polarized flux
, the continuum polarization reported by previous seyfert atlases will be used as a lower limit @xcite .
i revisit the polarization of the circinus galaxy measured by @xcite , 1.9 % , using the only broad emission line detected in its polarized flux spectrum , namely h@xmath8 @xmath226563 ( h@xmath24 @xmath224861 being only marginally detected , @xmath6 2@xmath25 ) .
the ew , estimated using a lorentzian profile @xcite , is about 34 . from then
, the ratio between type-1 h@xmath8 ew ( @xmath23 400 ) and ew@xmath26 is equal to 11.8 and the resulting continuum polarization is equal to 22.4 % .
interestingly , @xcite have derived a similar result ( 25 % ) using a simple , numerical model applied to their own spectropolarimetric measurement of the circinus galaxy .
mrk 3 has been observed by @xcite , who found after correction for interstellar polarization and dilution by starlight a continuum polarization of 7.0 % . using h@xmath8 @xmath226563 and h@xmath24 @xmath224861 as diagnostic lines from @xcite ,
the ew of h@xmath8 is estimated to be @xmath6 360 and ew h@xmath24 @xmath6 65 .
the ratio are 1.11 and 1.23 , respectively .
the scattered light is thus expected to have an intrinsic polarization of 7.77 8.61 % .
the small bump around 6563 in the percentage of polarization spectrum of ngc 1667 was attributed to the red wing of the h@xmath8 + [ n ii ] profile by @xcite , who consequently stated that no polarized h@xmath8 emission have been detected . however , the polarization spectra are dominated by high noise levels , probably hiding the appearance of the broad wings of the line .
the ew of the polarized h@xmath8 @xmath226563 line is consequently very similar to the ew measured in the total flux spectrum , i.e. 14.3 .
the resulting ratio , 28.0 , can be used to account for an upper limit of the intrinsic continuum polarization , which is then set between 0.35 % @xcite and 9.8 % .
both h@xmath8 @xmath226563 and h@xmath24 @xmath224861 polarized broad lines can be detected in the polarized flux spectrum of ngc 4507 @xcite .
the estimated ew are 165 ( h@xmath8 ) and 30 ( h@xmath24 ) , raising the continuum polarization from its initial value ( 6.1 % ; @xcite ) to 14.8 16.3 % .
ngc 5506 has been observed by @xcite , who found a continuum polarization of 2.60 % @xmath7 0.41 % . from the spectropolarimetric measurements of a sample of nearby compton - thin ( @xmath27 10@xmath28 @xmath29 ) seyfert 2s , @xcite found no evidence for a broad h@xmath8 @xmath226563 line in ngc 5506 .
they argued that the cause of non - detection of the broad lines in the polarized spectrum of this agn can be due to an extended obscuring region rather than non - existence .
i estimate an upper limit on the ew for the polarized h@xmath8 @xmath226563 line ( @xmath6 10 ) looking at the total flux spectrum and set the corrected polarization degree between 2.6 % and 100 % .
@xcite measured the continuum polarization of ngc 7674 to be 3.8 % after removing the starlight contribution . from their spectra , i estimate the ew of the h@xmath8 @xmath226563 and h@xmath24 @xmath224861 lines to be 208 and 40 , respectively .
the ratio are 1.72 and 2.0 , increasing the intrinsic polarization of the scattered light to 6.54 7.6 % .
ngc 1068 has proven to be remarkable in a sense that the polarization degree of its broad emission lines is similar to that of the continuum .
this peculiar feature is also shared by the broad h@xmath8 line / continuum polarization in iras 13349 + 2438 and no revisions are necessary for these two objects .
the high level of polarization detected in mrk 78 by @xcite , 21.0 % @xmath7 9.0 % , is significantly different from the rest of the sample and might face the same physical conditions , if not suffering from an upward bias ( since polarization is a positive - definite quantity ) .
the absence of spectra make it impossible to verify this assumption .
the polarization spectra of the seven bal quasars extracted from @xcite , 0019 + 0107 , 0145 + 0416 , 02261024 , 0842 + 3431 , 1235 + 1453 , 1333 + 2840 and 1413 + 1143 , are too noisy to measure the ew of isolated broad emission lines in polarized flux .
the few broad emission lines detected in total flux are either merged with some other lines ( l@xmath8 + n v ; al iii + c iii ] ) , truncated or poorly resolved .
most of them are not even detected in polarized flux .
the reported polarization continuum will then be used as lower limits . unfortunately , there are no detailed polarization spectra published so far for mrk 34 , mrk 573 , mrk 1066 and ngc 5643
. over the 20 type-2 agn , 3 of them did not need any correction , 6 were revised and 11 can only be used as basic lower limits for the intrinsic polarization of the scattered light .
type-1 agn , once corrected for starlight contribution , do not suffer from an additional dilution component and are thus directly exploitable .
the detailed lists of the agn sampled , with inclination matched to continuum polarization , are given in tab .
[ table : type1 ] and tab .
[ table : type2 ] , for type-1 and type-2 objects respectively . the resulting compendium , comparing the polarization percentage @xmath3 versus the inclination @xmath4 for 53 objects , is presented in fig . [ fig : survey ] .
type-1 agn are shown in red , type-2 agn in violet . according to the inclination estimations ,
type-1 agn cover a range of orientation from @xmath4 = 0@xmath2 ( pole - on ) to @xmath4 = 45@xmath2 60@xmath2 , which is in agreement with the estimations of the overall half opening angle of the system , @xmath30 @xmath31 58@xmath2 , made by @xcite and @xcite .
type-1 objects exhibit low polarization degrees ( @xmath32 1 % ) , except for seven unusually , highly polarized sources ( eso 323-g077 , fairall 51 , ic 4329a , mcg-6 - 30 - 15 , mrk 486 , mrk 766 and mrk 1239 ) .
those high levels of polarization are similar to the detection of polarization degrees up to 4 % in mrk 231 @xcite , or in the warm infrared ultraluminous agn survey done by @xcite , but they still need to be explained ( see sect . [ discussion : type1 ] )
. however , most of the type-1 sources collected here follow the empirical ascertainment , started with the observational surveys of seyfert 1 agn realized by @xcite , @xcite and @xcite : type-1 agn predominantly show low levels of polarization , associated with polarization position angles roughly parallel to the radio axis of the system ( when radio axis measurements were available ) .
it is particularly interesting to note that , among the seven polar scattering dominated agn ( with `` perpendicular '' polarization ) identified , five exhibit a polarization degree higher than 1 % . if a perpendicular polarization position angle can be explained by an increase of the opacity of the polar outflows ( through which the observer s line of sight is passing ) , it is more difficult to explain such high @xmath3 at inclinations below 45@xmath2 @xcite . in the case of type-2 agn , objects show perpendicular polarization position angles associated with higher polarization percentages , not uncommonly @xmath31 7 % ( after first - order correction for the intrinsic continuum polarization ) . as electron - induced @xmath3 depends on the cosine square of the scattering angle ,
seyfert galaxies seen edge - on are expected to be intrinsically more polarized @xcite .
radiation scatters perpendicularly on to the ionized , polar outflows detected in type-2 agn , leading to higher polarization percentages than at polar inclinations , where the combination of forward scattering and dilution by the unobscured nucleus diminishes the net polarization .
a clear estimation of the average polarization percentage of seyfert 2 objects remains problematic since most of the quoted values are lower limits .
however , the trend is that @xmath3 seems to increase with inclination , with a possible decrease at extreme type-2 inclinations that has to be properly observed by future , rigorous measurements of the scattered light alone .
the overall inclination of seyfert 2 agn lies between 60@xmath2 and 90@xmath2 ( edge - on ) .
no clear polarization break is detected at the transition angles ( 45@xmath2 60@xmath2 ) between type-1 and type-2 objects ; @xmath3 increases continuously from pole - on to edge - on view .
note that the range of orientation that separates type-1 and type-2 seyfert galaxies is not artificially enhanced by the selection criteria ( see sect .
[ comp : tables ] ) , as none of the rejected inclinations cover the 45@xmath2
60@xmath2 range .
however , the reader is advised to note that possible bias in the estimation of @xmath4 may shift some objects to different inclination values .
eso 323-g077 ( type-1 ) , ngc 1365 ( type-1 ) , iras
13349 + 2438 ( type-2 ) and the circinus galaxy ( eso 97-g13 , type-2 ) are in the intermediate zone between the two classifications . while the type-2 classification of the circinus galaxy is undisputed ,
the cases of eso 323-g077 , ngc 1365 and iras 13349 + 2438 are more ambiguous .
spectropolarimetric observations achieved by @xcite showed that , at least , a fraction of the inner region of eso 323-g077 must be hidden behind the torus horizon , classifying this agn as a borderline seyfert 1 galaxy .
ngc 1365 is another intriguing , borderline object , showing rapid transition between compton - thin and compton - thick regimes due to x - ray eclipses @xcite .
such behaviour leads to a difficult classification of ngc 1365 , either type-1 @xcite , type-1.5 @xcite , type-1.8 @xcite or even type-2 @xcite .
the case of iras 13349 + 2438 is somewhat similar to the previous object .
the h@xmath24 line width measured by @xcite favours a type-2 classification , an argument in contradiction with @xcite who found that iras 13349 + 2438 shares many properties with narrow - line seyfert 1s , though they noted that this object must have a peculiar geometry .
such statement is supported by the spectropolarimetric observations realized by @xcite , who have shown that the observer s line of sight is probably intercepting part of the equatorial dusty material , classifying iras 13349 + 2438 as a borderline type-2 agn .
thus , from a polarimetric point of view and if its estimated inclination is correct , the circinus galaxy should be considered as a borderline seyfert 2 object .
a direct application of this polarization / inclination study concerns radiative transfer in numerical agn models . while more recent simulations tend to complexify in terms of morphology , composition and kinematic , testing the relevance of a model against observations is a significant consistency check . in the following section ,
i run monte carlo simulations on four different agn models from the literature and compare the polarimetric results to the compendium .
simulations of emission , multiple scattering and radiative coupling in complex agn environments were achieved using stokes @xcite , a public monte carlo code including scattering - induced polarization .
the results presented hereafter are representative of each different model , characterized by a unique set of parameters .
the resulting wavelength - independent polarization percentage is integrated over 2000 to 8000 . for consistency ,
the same input spectrum is used for all the models , namely an isotropic source emitting an unpolarized spectrum with a power - law sed @xmath33 ( @xmath8 = 1 ) .
finally , an important condition on the models investigated below is that , at least , they reproduce the expected polarization dichotomy ( i.e. parallel polarization position angles for type-1 inclinations and perpendicular polarization position angles for type-2s ) . according to the axisymmetric unified model @xcite , the central supermassive black hole and its accretion disc , that radiates most of its bolometric luminosity in the uv / optical bands , are obscured by an equatorial , dusty torus .
a common hypothesis is that the funnel of the torus collimates ejection winds in the form of a bi - conical polar outflow .
past spectropolarimetric models of agn , composed by a central irradiating source , a dusty torus and a bi - conical , electron - filled wind , showed that only perpendicular polarization ( with respect to the symmetry axis of the torus ) can emerge @xcite . to introduce the production of parallel polarization in polar viewing angles , a third , equatorial region lying between the torus and the source has been proposed @xcite .
this highly ionized , geometrically thin disc can be associated with the accretion flow between the torus and the blr @xcite and is necessary to reproduce the observed polarization dichotomy @xcite .
following the parametrization from @xcite , the central , unpolarized source is surrounded by an equatorial , scattering flared disc with a half - opening angle of 20@xmath2 with respect to the equatorial plane and a thomson optical depth in the @xmath34 band of @xmath35 = 1 . along the same plane ,
an optically thick ( @xmath36 1 ) , dusty torus , filled with a standard `` milky way '' dust mixture @xcite prevents radiation to escape along the equator .
an hourglass - shaped , electron - filled region ( @xmath37 = 0.3 ) accounts for the polar ejection flow . the torus and the collimated ionized wind sustain the same half - opening angle , 60@xmath2 with respect to the symmetry axis of the model , see fig .
[ fig:3comp ] ( top ) .
the reader may refer to @xcite for further details about the model .
the wavelength - integrated polarization spectrum of the three - component model is shown in fig .
[ fig:3comp ] ( bottom ) . from 0@xmath2 to 60@xmath2 ( where the transition between type-1 and type-2 classification occurs )
, the three - component model successfully reproduces the average polarization level expected from type-1 agn ( i.e. @xmath0 1 % ) as well as a parallel polarization position angle .
@xmath3 rises from pole - on view to intermediate inclination without exceeding 1 % .
when the observer s line of sight crosses the torus height , @xmath3 decreases .
this is due to the competition between parallel polarization produced by the equatorial , scattering disc and perpendicular polarization originating from the torus / polar regions , cancelling each other .
such behaviour is expected in any axisymmetric model and @xmath3 may decrease down to zero , depending on the global morphology of the system .
once the equatorial , electron - filled disc disappears behind the torus horizon , @xmath3 strongly increases , up to 30 % , as radiation becomes dominated by perpendicular , thomson scattering inside the polar outflows .
the resulting polarization is not high enough in the 45@xmath2
65@xmath2 range and becomes too strong at large inclinations to fit the majority of type-2 objects .
a three - component model produces too much polarization at extreme type-2 inclinations . a natural way to decrease
the amount of @xmath3 is to add an absorbing medium to the previous model .
we know from observations that , beyond the dust sublimation radius , the ionized outflows merge continuously with the dusty environment of the host galaxy , forming the so - called ( low - density ) nlr @xcite .
the next step is then to investigate , in the framework of this compendium , the polarization signature of a four - component model that includes a dust - filled , low opacity ( @xmath38 = 0.3 ) nlr @xcite .
i used the same parametrization as in sect .
[ analysis:3comp ] , with the nlr bi - cone sustaining the same half - opening angle as the ionized outflows ( fig .
[ fig:4comp ] , top ) .
the addition of nlr into the modelling of an agn does not strongly impact the overall polarization signature .
it slightly decreases the net polarization in type-1 viewing angles ( see fig .
[ fig:4comp ] , bottom ) but does not alter its polarization position angle due to the large opening angle of the torus .
however , the possibility to reach polarization degrees of few percents ( such as for ic 4329a , mcg-6 - 30 - 15 or mrk 766 ) becomes less likely due to absorption .
the transition between parallel and perpendicular polarization occurs at the same inclination but the net polarization is lower , due to absorption . finally , in comparison with observations
, @xmath3 is found to be still too small between 45@xmath39 65@xmath2 ( @xmath40 1 10 % ) .
optical and uv observations of the nlr of ngc 1068 @xcite revealed the presence of many knots of different luminosity in the outflowing gas that can be attributed to inhomogeneities of the medium .
similar results are found for other sources ( e.g. mrk 3 ; @xcite ) , strengthening the idea that agn outflows may not be a continuous flow @xcite . due to the torus compactness
, there is less direct evidence for a clumpy torus and most of the suppositions about the fragmented nature of the circumnuclear matter comes from numerical simulations @xcite .
i now investigate a model in which the ionization cones are fragmented , while maintaining a compact dusty torus and an equatorial , scattering disc responsible for the production of parallel polarization .
the equatorial disc and the circumnuclear dusty region retain the same morphological and composition parameters as in sect .
[ analysis:3comp ] and sect .
[ analysis:4comp ] .
the fragmented outflows now consist of 2000 electron - filled spheres of constant density ( @xmath41 = 0.3 ; @xcite ) and radius ( filling factor @xmath6 5 % ) .
the filling factor is evaluated by summing up the volume of the clumps and dividing the total by the volume of the same unfragmented bi - conical nlr . a schematic view of the model is presented in fig .
[ fig : frag ] ( top ) .
similarly to the three and four - component models , a clumpy agn ( fig .
[ fig : frag ] , bottom ) reproduces both the low polarization levels and the expected polarization position angle in type-1 orientations .
however , due to multiple scattering on the outflow s clumps increasing the polarization degree , and gaps along the type-1 line of sights that allow a direct view of the electron disc , the net polarization percentage is higher , up to 2 % for intermediate inclinations .
this level is still not sufficient to reproduce the observed polarization of highly polarized type-1 objects , but a fragmented medium enables higher polarization degrees for type-1 modelling .
fragmentation also impacts the inclination at which perpendicular polarization starts to dominate the production of parallel polarization , but the resulting transition inclination is not consistent with the observed polarization position angle of type-2 agn . at type-2 inclinations
, @xmath3 is much lower ( @xmath42 3 % ) than in previous modellings due to the enhanced escape probability from the outflows .
a fragmented model can match the lower limit on polarization of a large fraction of type-2 agn but fails to reproduce the high continuum polarization of ngc 1068 , mrk 78 , iras 13349 + 2438 or the circinus galaxy .
however , the cloudlet distribution is probably different for each individual object and should be adapted case by case , i.e. by increasing / decreasing the filling factor . by increasing the filling factor of the clumpy outflows ,
the model will start to behave like the agn model presented in sect .
[ analysis:3comp ] , strengthening the production of perpendicular polarization at type-2 viewing angles and matching higher polarization percentages .
the hydrostatic equilibrium hypothesis , postulated for the equatorial , toroidal region , is slowly evolving .
based on the pioneering work done by @xcite , a hydrodynamical scenario is now considered as an alternative to the usual dusty torus , involving clumps of dusty matter embedded in a hydromagnetic disc - born wind ( see @xcite , and references therein ) .
@xcite took advantage of this scenario to build a model which attempts to explain the broad and narrow absorption line regions , as well as the broad emission line region , of type-1 quasars . in its phenomenologically derived structure ,
a flow of warm , highly ionized matter ( whim ) arises from an accretion disc in a narrow range of radii , bent outward and driven into a radial direction by radiation pressure .
the model of @xcite was recently explored by @xcite , who proposed a number of adjustments to match observed polarization data in the uv and optical bands . to explore the consistency of the model described by @xcite and modified by @xcite , i plotted in fig . [
fig : elvis ] ( bottom ) the adjusted model proposed by @xcite .
the whim bending angle is set to 45@xmath2 and its collimation angle to 3@xmath2 .
it arises at a distance @xmath43 = 0.0032 pc from the central source and extends up to 0.032 pc .
the thomson optical depth at the outflow s base and inside the conical , outflowing direction are , respectively , set to @xmath44 = 0.02 and @xmath45 = 2 .
a failed wind , composed of cold dust , is self - shielded from the continuum source by the whim .
the dusty outflow sustains a half - opening angle of 51@xmath2 with respect to the symmetry axis of the system , and a collimation angle of 3@xmath2 .
its opacity along the equator is set to @xmath46 = 4 .
refer to @xcite for details about this choice of parameters . from fig .
[ fig : elvis ] ( bottom ) , it can be seen that the continuum polarization arising at type-1 inclinations follow the same trend as the three previous models : @xmath3 reaches a maximum value of 1 % and can not account for the highly polarized type-1 objects of the compendium .
a local diminution of @xmath3 appears at @xmath4 = 42@xmath2 , when the observer s line of sight crosses the outflowing material .
the polarization position angle then switches from parallel to perpendicular , with respect to the projected symmetry axis of the system . while the transition between parallel and perpendicular polarization occurs at a smaller @xmath4 in comparison with previous modelling , it is still coherent with the polarization position angle measurements of ngc 5548 and eso 323-g077 , which exhibit perpendicular polarization at @xmath4 = 47.3@xmath20 and @xmath4 = 45@xmath2 , respectively .
beyond 54@xmath2 , when the observer s line of sight no longer passes through the radial outflows , @xmath3 reaches 10 %
11 % then slowly decreases because of the overwhelming impact of dust absorption along the equatorial direction .
the polarization predicted by the line - driven wind model fits nearly all the observational , highly polarized type-2 objects in the 45@xmath2
65@xmath2 range and can account for objects with lower @xmath3 at extreme inclinations .
it is noteworthy that while a disc - born wind is by far the closest model to observations , one must be cautious as type-2 polarizations have first - order corrections and estimated inclinations are subject to potential biases .
the polarization - versus - inclination study presented in this paper allows a test of the relevance of four different agn models from the literature .
all of them successfully reproduce both the observed polarization dichotomy and the average polarization percentage of type-1 agn , but strongly differ in the 45@xmath2 90@xmath2 inclination range .
it is then easier to discriminate between several agn models at intermediate inclinations .
models composed of uniform , homogeneous reprocessing regions ( an equatorial scattering disc , a dusty torus and a pair of collimated cones , with the possible addition of dusty nlr ) tend to create high perpendicular polarization degrees for type-2 agn , while the polarization produced by a model with fragmented polar outflows do not extend farther than 3 % in the same orientation range .
moreover , in the case of the model with clumpy ionization cones , the transition between parallel and perpendicular polarization happens at typical type-2 inclinations , which is in disagreement with observations .
a refinement of the clumpy model is necessary .
a deeper analysis of fragmented media is ongoing , targeting equatorial scattering discs , lil and hil blr , tori , ionization cones and nlr .
preliminary results show that the polarization degree at type-2 inclinations can rise up to few tens of percent for dense cloudlet distributions .
the transition between parallel and perpendicular polarization is correlated with the half - opening angle of the torus , and a fragmented circumnuclear region with a half - opening angle of 45@xmath2 can produce a switch between parallel and perpendicular polarization position angle at @xmath6 60@xmath2 .
the exploration of the parameter space of the models ( optical depth , filling factor , covering factor ... ) will be considered .
the model of @xcite is undoubtedly the closest to observation , as it produces both high and low polarization degrees at type-2 viewing angle , strengthening the hypothesis that at least some undermined fraction of agn components are wind - like structures . even by varying the parameters , none of the model can reach the high polarization levels of the inventoried type-1s eso 323-g077 ( 7.5 % ) , fairall 51 ( 4.12 % @xmath7 0.03 % ) , ic 4329a ( 5.80 % @xmath7 0.26 % ) , mcg-6 - 30 - 15 ( 4.06 % @xmath7 0.45 % ) , mrk 486 ( 3.40 % @xmath7 0.14 % ) , mrk 766 ( 3.10 % @xmath7 0.80 % ) , mrk 1239 ( 4.09 % @xmath7 0.14 % ) or mrk 231 ( @xmath6 4 % ; @xcite ) .
results presented in sect . [ comp : survey ] show that their inclination ranges from 0@xmath2 to 45@xmath2 , indicating that scattering at large angles between the photon source , the polar winds and the observer is unlikely to be responsible for all the atypically high continuum polarization observed . as the optical thickness of agn polar outflows
is estimated to be relatively small ( @xmath47 1 ) , the major contribution to scattering - induced type-1 polarization comes from the equatorial scattering disc .
@xcite showed that the resulting polarization percentage from an equatorial , electron - filled disc is fairly low , independently of its half opening angle and thomson opacity .
a major challenge to numerical models of agn is to increase the net polarization degree at type-1 viewing angles for isolated cases .
it is even more challenging taking into account that the polarization position angle of five out of the seven objects ( eso 323-g077 , fairall 51 , mrk 486 , mrk 766 and ngc 3227 ) is found to be perpendicular to their radio axis , similarly to mrk 231 @xcite .
there are several potential ways to strengthen the net polarization of the models .
one can consider perpendicular scattering between the source , a reprocessing region located along the equatorial plane , and the observer
. a promising target could be the hil and lil blr .
considering the constraints on the hil and lil blr structure derived by @xcite , using kinematic measurements of emission lines in four nearby agn , one can estimate the half - opening angle of the emission line region . according to @xcite , the hils ( i.e. emitted close to the photoionizing source ) originate from a medium with half - opening angle 3.8@xmath2
26@xmath2 from the equatorial plane , while the lils are created in a structure with a half - opening angle 11@xmath2 60@xmath2 .
further tests must be achieved to explore the polarization position angle and the amount of parallel polarization that hil and lil blr can generate but it is unlikely that scattering within an axisymmetrical model can reach up to few percents at type-1 inclinations . as stated by @xcite ,
breaking the symmetric pattern of irradiation can help to understand the velocity dependence of broad emission line variability detected in many agn .
strong off - axis flares could then explain the observed , extremely asymmetric , balmer lines with broad peak redshifted or blueshifted by thousands of km@xmath48s@xmath49 @xcite , and produce higher polarization degrees even at polar orientation .
it is important to test the off - axis flare model as , if correct , the azimuthal phase of the continuum source would play a critical role in the measurement of the supermassive black hole mass .
@xcite have recently started the investigation of the off - axis irradiation theory by looking at the velocity dependence of the polarization of the broad emission lines .
preliminary results , compared to spectropolarimetric data for type-1 agn from the literature , indicate that both the degree and position angle of polarization should be affected by asymmetrical emission .
the net polarization percentage of the optical continuum is also slightly stronger .
further modelling will be achieved to explore how far optical and uv continuum polarization can be strengthened by temporary off - axis irradiation .
finally , highly polarized agn are not uncommon in type-1 radio - loud objects , with @xmath3 up to 45.5 % @xmath7 0.9 % @xcite .
the net polarization is quite variable from the radio to the optical band , often on short time - scales , pointing towards well - ordered magnetic fields surrounding a spatially small emitting region .
while most of the observed polarization of radio - quiet agn is undoubtedly originating from scattering off small particles @xcite , the high , optical polarization of blazing quasars is thought to be associated with doppler - boosted synchrotron emission from relativistic jets pointing towards us .
the polarization position angle of blazar cores is usually perpendicular to the jet axis while a parallel component is detected for emerging superluminal knots @xcite .
if electron and dust reprocessing appear to be unable to reach @xmath3 @xmath50 3 % along poloidal directions , the correct interpretation might lie somewhere in the middle .
the presence of a sub - parsec , aborted jet has yet to be proven but could explain the time variability and spectra of narrow line seyfert 1 galaxies @xcite and potentially create perpendicular polarization degree up to a few percent , while the lines and continuum polarization would be still mainly produced by reprocessing . if an undetermined fraction of the total polarization of highly polarized type-1 agn indeed originates from synchrotron emission , the net polarization could be expected to vary , but with a much smaller amplitude than for synchrotron - dominated , radio - loud objects .
long term monitoring of radio - quiet , highly - polarized type-1 agn would then help to evaluate the fraction of polarization arising from reprocessing and from synchrotron emission .
estimations made by @xcite and @xcite are primarily based on the assumptions that seyfert 1 and normal galaxies follow the same black hole mass = 5 was used by @xcite . ] - bulge velocity dispersion correlation , and that the lil blr are in pure keplerian rotation , coplanar to the system inclination . while detected , the motion and the morphology of the lil blr remain uncertain @xcite .
there are no strong constraints from the emission line profiles as a wide variety of kinematic models are able to reproduce the non - gaussian profiles detected in agn @xcite .
the technique of velocity - resolved reverberation mapping @xcite is a step forward and tends to rule out any significant outflow from the agn , while detecting a slight inflow @xcite and fast keplerian motion .
fitting the distorted red wing of the fe k@xmath8 fluorescent line in x - ray bright , type-1 agn can lead to possible bias .
the procedure used by @xcite assumes that the asymmetrical broadening of the iron line is caused by doppler and general relativistic effects close to the central black hole ; however , a competitive mechanism was proposed by @xcite and @xcite . in this scenario ,
line broadening occurs at larger distances from the accretion disc , where a distribution of cold , absorbing gas blocks a fraction of the initial continuum .
transmitted and scattered radiation through the cloudlet environment finally carves out the distorted red wing .
if distant absorption dominates relativistic effects , the estimated inclinations might then be questionable .
the method developed by @xcite , based on the work achieved by @xcite and @xcite , relies on the nature of the nlr kinematics to determine the orientation of the system .
one of their fundamental hypothesis is consistent with the unified model : agn are axisymmetrical objects . in this picture
, the nlr structure sustains the same symmetry axis as the dusty torus , which is coplanar with the accretion disc .
thus , determining the inclination of the nlr is equivalent to determining the orientation of the whole system .
however , a recent ir interferometric campaign carried out by @xcite found that the extended outflows of ngc 1068 are likely to be inclined by 18@xmath2 with respect to the obscuring torus axis .
if this trend is confirmed , and observed for other seyfert - like galaxies , the overall agn picture will become more complex . finally , the two - component model produced by @xcite shows degeneracies between the various parameter combinations .
it disallows the characterization of the outflow geometry in quasars showing broad absorption features , and weakens the constraints brought on the inclination of the system .
the viewing angles of the outflow derived by the authors are quite large , a conclusion shared by @xcite and @xcite , based on optical polarization surveys of bal quasars , but contested by several other authors @xcite . in particular
, @xcite showed that two - thirds of the bal quasars they observed using h@xmath24 line width as a diagnostic are well represented by objects with gas flowing along polar directions .
the first match of 53 agn inclinations with their intrinsic continuum polarization originating from electron and dust scattering is presented in this paper .
different techniques to retrieve the nuclear orientation of type-1 and type-2 seyfert galaxies were presented and discussed , highlighting their potential caveats .
the continuum polarization of several seyfert-2s was corrected using broad h@xmath8 and h@xmath24 line polarization as a reliable indicator of the true polarization of the scattered light , and lower limits were put for the remaining agn whose polarization spectra were either noise - saturated or unpublished .
the resulting compendium is in agreement with past observational / theoretical literature , and warrants additional conclusions and remarks the following .
* seyfert 1 agn are associated with low polarization degrees , @xmath0 1 % , and predominantly characterized by a polarization position angle parallel to the projected radio axis of the system .
the inclination of type-1 objects ranges from 0@xmath2 to 60@xmath2 . *
seven type-1s have been identified as polar scattering dominated agn , i.e. showing a perpendicular polarization position angle . among them , five have an atypical continuum polarization higher than 1 % , mostly associated with 10@xmath2 45@xmath2 inclinations . as
scattering - induced polarization is unlikely to produce such high polarization degrees at type-1 orientation , a more elaborate scenario must be considered . * after correction , seyfert 2 agn show polarization degrees higher than 7 % and perpendicular polarization position angle .
unfortunately , most of the objects have only lower limits .
the inclination of type-2 objects is ranging from 47@xmath2 to 90@xmath2 . *
the transition between type-1 and type-2 agn occurs between 45@xmath2 60@xmath2 .
this range of inclination is likely to include agn classified as borderline objects , where the observer s line of sight crosses the horizon of the equatorial dusty medium .
four objects lie in this range , three of them ( eso 323-g077 , ngc 1365 and iras 13349 + 2438 ) being already considered as borderline seyfert galaxies from spectroscopic observations . if the estimated inclination of the fourth object ( the circinus galaxy ) is correct , it should be considered as another borderline agn . *
the usual axisymmetric agn models have difficulties to reproduce the trend of polarization with inclination . fragmenting the reprocessing regions
is helpful to cover a wide range of continuum polarization but a disc - born wind model is found to be already quite close from observations.a fine tuning of the line - driven disc wind could easily match a substantial fraction of the reported measurements .
problems determining the inclination of agn must be taken into consideration but , despite potential caveats , the associated continuum polarization lies within the margins of past empirical results , consolidating the basis of the compendium .
it is then important for future models , as a consistency check , to reproduce the average continuum polarization , the polarization dichotomy and the transition between type-1 and type-2 classification ( 45@xmath39 60@xmath2 ) . by improving the quality of the methods to determine the inclination of agn and properly removing the contribution of both stellar and starburst light in future polarimetric measurement of type-2 objects
, it will possible to bring very strong constraints on the morphology , composition and kinematics of agn . to achieve this goal ,
a new uv / optical spectropolarimetric atlas of seyfert 2s is necessary .
the author is grateful to the referee ski antonucci for his useful and constructive comments on the manuscript .
i also acknowledge the academy of sciences of the czech republic for its hospitality , and the french grant anr-11-js56 - 013 - 01 of the project polioptix and the cost action mp1104 for financial support .
99 alexander , d. m. ; heisler , c. a. ; young , s. ; lumsden , s. l. ; hough , j. h. ; bailey , j. a. 2000 , mnras , 313 , 815 alloin , d. , edmunds , m. g. , lindblad , p. o. , & pagel , b. e. j. 1981 , a&a , 101 , 377 antonucci , r. r. j. 1984 , apj , 278 , 499 antonucci , r. r. j. , & miller , j. s. 1985 , apj , 297 , 621 antonucci , r. 1993 , ara&a , 31 , 473 antonucci , r. , hurt , t. , & miller , j. 1994 , apj , 430 , 210 antonucci , r. 2002a , agn surveys , proceedings of iau colloquium 184 , asp conference proceedings , vol . 284 .
edited by r.f .
green , e.ye .
khachikian , and d.b .
san francisco , ca : astronomical society of the pacific , 2002 , p. 147 .
antonucci , r. 2002b , astrophysical spectropolarimetry .
proceedings of the xii canary islands winter school of astrophysics , 2000 , edited by j. trujillo - bueno , f. moreno - insertis , and f. snchez .
cambridge , uk : cambridge university press , isbn 0 - 521 - 80998 - 3 , 2002 , p. 151
- 175 bambi , c. 2011 , modern physics letters a , 26 , 2453 bambi , c. 2013 , the astronomical review , 8 , 010000 barth , a. j. , filippenko , a. v. , & moran , e. c. 1999 , apj , 525 , 673 barvainis , r. , & lonsdale , c. 1997 , aj , 113 , 144 batcheldor , d. ; robinson , a. ; axon , d. j. ; young , s. ; quinn , s. ; smith , j. e. ; hough , j. ; alexander , d. m. 2011 , apj , 738 , 90 berriman , g. 1989 , apj , 345 , 713 berriman , g. , schmidt , g. d. , west , s. c. , & stockman , h. s. 1990 , apjs , 74 , 869 blandford , r. d. , & mckee , c. f. 1982 , apj , 255 , 419 bon , e. , gavrilovi , n. , la mura , g. , & popovi , l. .
2009 , newar , 53 , 121 borguet , b. , & hutsemkers , d. 2010 , a&a , 515 , a22 brandt , w. n. , fabian , a. c. , & pounds , k. a. 1996 , mnras , 278 , 326 brindle , c. ; hough , j. h. ; bailey , j. a. ; axon , d. j. ; ward , m. j. ; sparks , w. b. ; mclean , i. s. 1990 , mnras , 244 , 577 capetti , a. , axon , d. j. , macchetto , f. , sparks , w. b. , & boksenberg , a. 1995 , apj , 446 , 155 capetti , a. , macchetto , f. , axon , d. j. , sparks , w. b. , & boksenberg , a. 1995 , apj , 448 , 600 capetti , a. , axon , d. j. , macchetto , f. , sparks , w. b. , & boksenberg , a. 1996 , apj , 469 , 554 capetti , a. , macchetto , f. d. , & lattanzi , m. g. 1997 , ap&ss , 248 , 245 capetti , a. , axon , d. j. , macchetto , f. d. , marconi , a. , & winge , c. 1999 , mmsai , 70 , 41 chun , m .- s . 1982 , journal of korean astronomical society , 15 , 41 crenshaw , d. m. , kraemer , s. b. , hutchings , j. b. , et al .
2000 , aj , 120 , 1731 darcangelo , f. d. , marscher , a. p. , jorstad , s. g. , et al .
2009 , apj , 697 , 985 dai , h .- f .
, & wang , t .- g . 2008 , chjaa , 8 , 245 das , v. , crenshaw , d. m. , hutchings , j. b. , et al .
2005 , aj , 130 , 945 das , v. , crenshaw , d. m. , kraemer , s. b. , & deo , r. p. 2006
, aj , 132 , 620 efron , b. 1979 , bootstrap methods : another look at the jackknife .
7 , 1 elitzur , m. , & shlosman , i. 2006 , apjl , 648 , l101 evans , i. n. ; ford , h. c. ; kinney , a. l. ; antonucci , r. r. j. ; armus , l. ; caganoff , s. 1991 , apjl , 369 , l27 elvis , m. 2000 , apj , 545 , 63 eracleous , m. , livio , m. , halpern , j. p. , & storchi - bergmann , t. 1995 , apj , 438 , 610 fabian , a. c. , rees , m. j. , stella , l. , & white , n. e. 1989 , mnras , 238 , 729 ferrarese , l. , & merritt , d. 2000 , apjl , 539 , l9 fischer , t. c. , crenshaw , d. m. , kraemer , s. b. , & schmitt , h. r. 2013 , arxiv:1308.4129 gallagher , s. c. ; schmidt , gary d. ; smith , paul s. ; brandt , w. n. ; chartas , g. ; hylton , shavonne ; hines , d. c. ; brotherton , m. s. 2005 , apj , 633 , 71 gaskell , c. m. 1988 , apj , 325 , 114 gaskell , c. m. , klimek , e. s. , & nazarova , l. s. 2007 , arxiv:0711.1025 gaskell , c. m. 2010 , arxiv:1008.1057 gaskell , c. m. 2011 , baltic astronomy , 20 , 392 gebhardt , k. , bender , r. , bower , g. , et al .
2000 , apjl , 539 , l13 ghisellini , g. , haardt , f. , & matt , g. 2004 , a&a , 413 , 535 goodrich , r. w. , & miller , j. s. 1994 , apj , 434 , 82 goosmann , r. w. , & gaskell , c. m. 2007 , a&a , 465 , 129 goosmann , r. w. , & matt , g. 2011 , mnras , 415 , 3119 goosmann , r. w. , gaskell , c. m. , & marin , f. 2013 , arxiv:1311.2249 hicks , e. k. s. , & malkan , m. a. 2008 , apjs , 174 , 31 hines , d. c. 1994 , ph.d .
thesis , texas university ho , l. c. , filippenko , a. v. , & sargent , w. l. 1995 , apjs , 98 , 477 ho , l. 1999 , assl , 234 , 157 hnig , s. f. , beckert , t. , ohnaka , k. , & weigelt , g. 2007 , the central engine of active galactic nuclei , 373 , 487 inoue , h. , & matsumoto , c. 2003 , pasj , 55 , 625 jaffe , w. , meisenheimer , k. , rttgering , h. j. a. , et al .
2004 , nature , 429 , 47 keel , w. c. 1980 , aj , 85 , 198 kartje , j. f. 1995 , apj , 452 , 565 kay , l. e. 1994 , apj , 430 , 196 kishimoto , m. 1999 , apj , 518 , 676 kollatschny , w. , & zetzl , m. 2012 , arxiv:1211.3065 kollatschny , w. , & zetzl , m. 2013 , a&a , 558 , a26 lawrence , a. , & elvis , m. 1982 , apj , 256 , 410 lee , j. c. , kriss , g. a. , chakravorty , s. , et al .
2013 , mnras , 430 , 2650 lumsden , s. l.,alexander , d. m. , & hough , j. h. 2004 , mnras , 348 , 1451 macchetto , f. ; marconi , a. ; axon , d. j. ; capetti , a. ; sparks , w. ; crane , p. 1997
, apj , 489 , 579 maiolino , r. , & rieke , g. h. 1995 , apj , 454 , 95 marin , f. , goosmann , r. w. , gaskell , c. m. , porquet , d. , & doviak , m. 2012a , a&a , 548 , a121 marin , f. , & goosmann , r. w. 2012 , sf2a-2012 : proceedings of the annual meeting of the french society of astronomy and astrophysics . eds . : s. boissier , p. de laverny , n. nardetto , r. samadi , d. valls - gabaud and h. wozniak , pp.587 - 590 marin , f. , goosmann , r. , & doviak , m. 2012 , journal of physics conference series , 372 , 012065 marin , f. ; goosmann , r. w. ; doviak , m. ; muleri , f. ; porquet , d. ; grosso , n. ; karas , v. ; matt , g. 2012b , mnras , 426 , l101 marin , f. , & goosmann , r. w. 2013 , mnras , 436 , 2522 marin , f. , & goosmann , r. w. 2013b , sf2a-2013 : proceedings of the annual meeting of the french society of astronomy and astrophysics . eds . : l. cambresy , f. martins , e. nuss , a. palacios , pp.475 - 478 marin , f. , & goosmann , r. w. 2013c , sf2a-2013 : proceedings of the annual meeting of the french society of astronomy and astrophysics .
l. cambresy , f. martins , e. nuss , a. palacios , pp.479 - 482 marin , f. , & tamborra , f. 2013 , asr , in press , arxiv:1309.1684 marin , f. ; porquet , d. ; goosmann , r. w. ; doviak , m. ; muleri , f. ; grosso , n. ; karas , v. 2013 , mnras , 436 , 1615 marscher , alan p. ; jorstad , svetlana g. ; gmez , jos - luis ; aller , margo f. ; tersranta , harri ; lister , matthew l. ; stirling , alastair m. 2002 , nature , 417 , 625 martin , p. g. , thompson , i. b. , maza , j. , & angel , j. r. p. 1983
, apj , 266 , 470 mathis , j. s. , rumpl , w. , & nordsieck , k. h. 1977 , apj , 217 , 425 mead , a. r. g. ; ballard , k. r. ; brand , p. w. j. l. ; hough , j. h. ; brindle , c. ; bailey , j. a. 1990 , a&as , 83 , 183 merritt , d. , & ferrarese , l. 2001 , apj , 547 , 140 moran , e. c. , barth , a. j. , kay , l. e. , & filippenko , a. v. 2000 , apjl , 540 , l73 miller , j. s. , & antonucci , r. r. j. 1983 , apjl , 271 , l7 miller , j. s. , & goodrich , r. w. 1990 , apj , 355 , 456 miller , l. , turner , t. j. , & reeves , j. n. 2008 , a&a , 483 , 437 miller , l. , turner , t. j. , & reeves , j. n. 2009 , mnras , 399 , l69 miller , l. , & turner , t. j. 2013 , apjl , 773 , l5 nandra , k. , george , i. m. , mushotzky , r. f. , turner , t. j. , & yaqoob , t. 1997 , apj , 477 , 602 nenkova , m. , ivezi , . , & elitzur , m. 2002 , apjl , 570 , l9 ogle , p. m. ; cohen , m. h. ; miller , j. s. ; tran , h. d. ; goodrich , r. w. ; martel , a. r. 1999 , apjs , 125 , 1 ogle , p. m. , brookings , t. , canizares , c. r. , lee , j. c. , & marshall , h. l. 2003 , a&a , 402 , 849 oliva , e. , marconi , a. , cimatti , a. , & alighieri , s. d. s. 1998 , a&a , 329 , l21 osterbrock , d. e. 1978 , proceedings of the national academy of sciences of the united states of america , volume 75 , issue 2 , pp .
540 - 544 osterbrock , d. e. , & martel , a. 1993 , apj , 414 , 552 packham , c. , young , s. , hough , j. h. , axon , d. j. , & bailey , j. a. 1997 , mnras , 288 , 375 peterson , b. m. 2006 , physics of active galactic nuclei at all scales , 693 , 77 pier , e. a. , & krolik , j. h. 1992 , apj , 401 , 99 pier , e. a. , & krolik , j. h. 1993 , apj , 418 , 673 punsly , b. , & zhang , s. 2010 , apj , 725 , 1928 raban , d. , jaffe , w. , rttgering , h. , meisenheimer , k. , & tristram , k. r. w. 2009 , mnras , 394 , 1325 reeves , j. n. , fabian , a. c. , kataoka , j. , et al . 2006 , astronomische nachrichten , 327 , 1 risaliti , g. , elvis , m. , fabbiano , g. , baldi , a. , & zezas , a. 2005 , apjl , 623 , l93 risaliti , g. , harrison , f. a. , madsen , k. k. , et al . 2013 , nature , 494 , 449 rowan - robinson , m. 1977 , apj , 213 , 635 rush , b. , malkan , m. a. , & spinoglio , l. 1993 , apjs , 89 , 1 schmid , h. m. ; appenzeller , i. ; camenzind , m. ; dietrich , m. ; heidt , j. ; schild , h. ; wagner , s. 2001 , a&a , 372 , 59 schmid , h. m. , appenzeller , i. , & burch , u. 2003 , aap , 404 , 505 schmidt , g. d. , & miller , j. s. 1985 , apj , 290 , 517 schmidt , g. d. , & hines , d. c. 1999 , apj , 512 , 125 schulz , h. , knake , a. , & schmidt - kaler , t. 1994 , a&a , 288 , 425 smith , p. s. , schmidt , g. d. , hines , d. c. , cutri , r. m. , & nelson , b. o. 2002b , apj , 569 , 23 smith , j. e. ; young , s. ; robinson , a. ; corbett , e. a. ; giannuzzo , m. e. ; axon , d. j. ; hough , j. h. 2002a , mnras , 335 , 773 smith , j. e. ; robinson , a. ; alexander , d. m. ; young , s. ; axon , d. j. ; corbett , elizabeth a. 2004 , mnras , 350 , 140 stockman , h. s. , angel , j. r. p. , & miley , g. k. 1979 , apjl , 227 , l55 storchi - bergmann , t. , baldwin , j. a. , & wilson , a. s. 1993 , apjl , 410 , l11 storchi - bergmann , thaisa ; eracleous , michael ; livio , mario ; wilson , andrew s. ; filippenko , alexei v. ; halpern , jules p. 1995 , apj , 443 , 617 storchi - bergmann , thaisa ; eracleous , michael ; ruiz , maria teresa ; livio , mario ; wilson , andrew s. ; filippenko , alexei v. 1997 , apj , 489 , 87 tran , h. d. 1995 , apj , 440 , 565 tran , h. d. 1995 , apj , 440 , 578 tran , h. d. 1995 , apj , 440 , 597 veron , p. , lindblad , p. o. , zuiderwijk , e. j. , veron , m. p. , & adam , g. 1980 , a&a , 87 , 245 wills , beverley j. ; wills , d. ; evans , neal j. , ii ; natta , a. ; thompson , k. l. ; breger , m. ; sitko , m. l. 1992 , apj , 400 , 96 wittkowski , m. ; kervella , p. ; arsenault , r. ; paresce , f. ; beckert , t. ; weigelt , g. 2004 , a&a , 418 , l39 wu , x .- b . , & han , j. l. 2001 , apjl , 561 , l59 young , s. , hough , j. h. , axon , d. j. , & bailey , j. a. 1997 , iau colloq . 163 : accretion phenomena and related outflows , ed .
d. t. wickramasinghe ; g. v. bicknell ; and l. ferrario , 121 , 837 young , s. 2000 , mnras , 312 , 567 zhang , t .- z . , &
wu , x .- b .
2002 , chjaa , 2 , 487 | the anisotropic nature of active galactic nuclei ( agn ) is thought to be responsible for the observational differences between type-1 ( pole - on ) and type-2 ( edge - on ) nearby seyfert - like galaxies . in this picture , the detection of emission and/or absorption features is directly correlated to the inclination of the system .
the agn structure can be further probed by using the geometry - sensitive technique of polarimetry , yet the pairing between observed polarization and seyfert type remains poorly examined .
based on archival data , i report here the first compilation of 53 estimated agn inclinations matched with ultraviolet / optical continuum polarization measurements .
corrections , based on the polarization of broad emission lines , are applied to the sample of seyfert-2 agn to remove dilution by starburst light and derive information about the scattered continuum alone .
the resulting compendium agrees with past empirical results , i.e. type-1 agn show low polarization degrees ( @xmath0 1 % ) predominantly associated with a polarization position angle parallel to the projected radio axis of the system , while type-2 objects show stronger polarization percentages ( @xmath1 7 % ) with perpendicular polarization angles .
the transition between type-1 and type-2 inclination occurs between 45@xmath2 and 60@xmath2 without noticeable impact on @xmath3 .
the compendium is further used as a test to investigate the relevance of four agn models .
while an agn model with fragmented regions matches observations better than uniform models , a structure with a failed dusty wind along the equator and disc - born , ionized , polar outflows is by far closer to observations .
however , although the models correctly reproduce the observed dichotomy between parallel and perpendicular polarization , as well as correct polarization percentages at type-2 inclinations , further work is needed to account for some highly polarized type-1 agn [ firstpage ] polarization radiative transfer scattering atlases galaxies : active galaxies : structure . |
following the measurement of the third lepton mixing angle , the so - called reactor angle @xmath4 @xcite , neutrino physics has entered the precision era .
indeed all three lepton mixing angles are expected to be measured with increasing precision over the coming years , with forthcoming accurate measurements expected for both the atmospheric angle @xmath5 and the solar angle @xmath6 .
first hints of the cp - violating ( cpv ) phase @xmath7 have also been reported in global fits @xcite , and rapid progress can be expected with the next generation of oscillation experiments .
the measurement of the reactor angle has had a major impact on models of neutrino mass and mixing , ruling out at a stroke models based on tribimaximal ( tbm ) lepton mixing @xcite , although , as we shall discuss in this paper , these patterns may survive in the neutrino sector , if charged lepton corrections are included .
such tbm patterns can arise from `` direct '' models @xcite , in which the full klein symmetry ( @xmath8 generators ) of the neutrino mass matrix as well as the @xmath9 symmetry of the charged lepton mass matrix are subgroups of an underlying discrete family symmetry .
alternatively , tbm mixing can arise from `` indirect '' models based on constrained sequential dominance ( csd ) @xcite with special family symmetry breaking vacuum alignments . in response to
the experimental data , many different model building directions capable of accounting for the reactor angle have emerged , as recently reviewed in refs .
@xcite and @xcite .
the viability of these ideas can only be established by comparison with experiment , and a tractable approach to test large classes of models is to identify generic types of prediction associated with these models .
a promising example of such a signature can be found in lepton mixing sum rules , which relate the three lepton mixing angles to the cpv oscillation phase @xmath7 , or more precisely to @xmath10 .
indeed , given the precisely measured values of the mixing angles , they can be regarded as predictions for @xmath10 , to be tested in future experiments .
lepton mixing sum rules arise from two distinct types of scenarios and lead to two different types , referred to as atmospheric and solar sum rules @xcite . _
atmospheric sum rules _
@xcite arise from a variety of `` semi - direct '' models in which only half of the klein symmetry emerges from the discrete family symmetry , classified in terms of finite von dyck groups , with charged lepton mixing controlled by the @xmath9 generator @xcite .
for example , such models can lead to trimaximal-1 ( tm1 ) or trimaximal-2 ( tm2 ) mixing , in which the first or second column of the tbmmixing matrix is preserved , eq .
( [ tm1 ] ) and eq .
( [ tm2 ] ) respectively , @xmath11 the atmospheric sum rule @xmath12 can be derived from these conditions @xcite , where @xmath13 , @xmath14 and @xmath15 for tm1 and @xmath16 for tm2 .
the study of correlations of this type , and their application to the discrimination between underlying models , has been shown to be a realistic aim for a next - generation superbeam experiment @xcite .
it was first shown in refs .
@xcite and @xcite that a@xmath0 generally leads to a `` semi - direct model '' which predicts tm2 mixing with the second atmospheric sum rule , while the indirect csd2 model with special family symmetry breaking vacuum alignments @xmath17 and @xmath18 in ref .
@xcite predicts tm1 mixing and the first atmospheric sum rule .
in fact the tm1 atmospheric sum rule arises from all generalised versions of csd@xmath19 , based on the vacuum alignments @xmath17 and @xmath20 for integer @xmath21 @xcite , since such alignments are orthogonal to the first column of the tbm matrix , @xmath22 , and hence predict tm1 mixing .
the pontecorvo - maki - nakagawa - sakata ( pmns ) matrix @xmath23 can be expressed as the product of the diagonalising matrices of the neutrino and charged lepton mass terms , @xmath24 and @xmath25 , respectively , @xmath26 _ solar sum rules _
@xcite arise in models in which a leading - order mixing matrix @xmath24 is corrected by a small basis change from the charged leptons . in the first models of this type @xmath25 had a cabibbo - like form : if we denote the angles which parameterise @xmath27 by @xmath28 , the angles obey @xmath29 .
these scenarios are motivated by grand unified theories ( guts ) where the approximately diagonal charged lepton mass matrix is related to the down - type quark mass matrix , together with the assumption that quark mixing arises predominantly from the down - type quark sector .
indeed this was the case in the csd model where solar sum rules were first proposed @xcite . in the context of `` direct '' models , solar sum rules arise when the full klein group continues to emerge from the discrete family symmetry ( leading for example to tbm mixing in the neutrino sector ) while the @xmath9 generator which governs the charged leptons is broken .
in the simple case where only @xmath30 are non - zero , with @xmath31 , the charged - lepton corrections do not change the third row of the neutrino mixing matrix , and solar sum rules can be derived from the conditions @xcite | u_1 | = s^_12s^_23 , | u_2 | = c^_12s^_23 , | u_3 | = c^_23 , [ sol ] where @xmath32 and @xmath33 .
for example , with tbm neutrino mixing @xmath34 , @xmath35 and @xmath36 , @xmath37 the solar sum rule @xmath38 can be derived from these conditions @xcite , where @xmath39 .
it is clear that the conditions on @xmath40 and @xmath41 in eq .
( [ sol2 ] ) are identical to the corresponding conditions for tm1 and tm2 mixing , respectively .
however the conditions on the other elements of the pmns mixing matrix are different , so the resulting atmospheric and solar sum rules will also be different . in this paper
, we extend the above derivation of the solar sum rule to the more general case where not only @xmath30 are non - zero but also @xmath42 is allowed to be non - zero and all complex phases are kept arbitrary , while still keeping @xmath43 . as a result we shall find the remarkable condition , = = t^_12 .
[ sol3 ] of course , the condition in eq .
( [ sol3 ] ) can be trivially derived from eq .
( [ sol ] ) , assuming @xmath44 .
the notable feature is that eq .
( [ sol3 ] ) also holds independently of @xmath42 and of all complex phases . however , eq . ( [ sol ] )
involves further relations which only hold for @xmath44 .
these further relations can be used to eliminate the atmospheric angle , providing a solar sum rule which is more restrictive than that coming from eq .
( [ sol3 ] ) alone .
nevertheless , we shall continue to refer to the relation in eq .
( [ sol3 ] ) as a solar sum rule , since it is satisfied even when @xmath44 , as in eq .
( [ sol ] ) , and is distinct from the atmospheric sum rules discussed earlier . the solar sum rule in eq .
( [ sol3 ] ) may be cast as a prediction for @xmath10 , as a function of the measured mixing angles and @xmath45 , = , [ sol4 ] an expression which had been derived previously using an alternative argument in refs .
@xcite and @xcite .
we also highlight a second remarkable feature of eq .
( [ sol4 ] ) , namely that it is not only independent of @xmath42 but also of @xmath46 .
the sum rule in eq .
( [ sol4 ] ) is specified by only fixing the value of @xmath47 .
therefore , we can enumerate the viable models of this type by deriving the values of @xmath48 associated with those leading - order mixing patterns with @xmath49 which are derivable from considerations of symmetry . in this article
, we shall show that this leads us to four well - motivated solar sum rules : one based on tbmmixing @xcite where @xmath35 , one based on bimaximal ( bm ) mixing @xcite where @xmath50 and two patterns based on versions of golden ratio mixing including gr1with @xmath51 @xcite and gr3with @xmath52 @xcite , where @xmath53 is the golden ratio .
we shall also discuss the viability of two leading - order patterns which have been invoked in the literature called gr2with @xmath54 @xcite , and hexagonal ( hex ) mixing with @xmath55 @xcite . for each viable prediction
we perform a study of the scope to test the sum rule in eq .
( [ sol4 ] ) within the current experimental programme . over the next few decades
, significant new information will be provided on the leptonic mixing matrix from two main sources : the next generation of medium - baseline reactor ( mr ) experiments and long - baseline wide - band superbeams ( wbb ) .
the mr programme primarily seeks to measure the interference between atmospheric and solar neutrino oscillations at baseline distances of around @xmath56@xmath57 km .
these facilities have been shown to be sensitive to the mass hierarchy @xcite .
there are two main experiments working towards a mr facility , both building on successful measurements of @xmath58 at a shorter baseline : juno @xcite and reno-50 @xcite .
the wbb experiments can be seen as complementary to the mr proposals .
collaborations such as lbne @xcite and lbno @xcite intend to construct a high - power long - baseline neutrino and antineutrino beam which can exploit matter effects and large statistics for the primary aim of constraining the cpv phase @xmath7 .
the combination of mr and wbb facilities will provide new levels of precision in the neutrino sector , with @xmath6 and @xmath58 being probed to the level of percent by mr experiments , and @xmath7 being constrained by dedicated wbb facilties .
this complementarity offers for the first time the possibility of experimentally testing relations such as eq .
( [ sol4 ] ) . in this work
, we shall simulate illustrative mr and wbb facilities with an aim to exploring how their complementarity can be used to constrain the models of charged - lepton corrections .
the idea of correcting a leading - order neutrino mixing pattern by contributions from the charged leptons has recently been revisited @xcite .
our work goes beyond these analyses in three ways .
firstly , we present a novel derivation of the correlation in eq .
( [ sol4 ] ) in a more general setting , showing it to be the consequence of the simpler relation eq .
( [ sol3 ] ) , which helps to highlight its relationship to the earlier solar sum rules .
secondly , we systematically derive the possible leading - order mixing patterns from considerations of residual symmetry , finding a small well - motivated set .
thirdly , we present the results of simulations assessing the potential to constrain the solar sum rules from two upcoming complementary oscillation experiments : a superbeam and a reactor facility .
the layout of the remainder of the paper is as follows : in section [ sec : derivation ] , we present a simple derivation of the correlations induced by charged - lepton corrections .
we then systematically identify the viable leading - order neutrino mixing matrices , and comment on their relation to the underlying flavour symmetry .
section [ sec : numerics ] is devoted to our numerical study .
we first consider the currently allowed parameter spaces of these correlations , then we present the details and results of our simulations of a superbeam and reactor experiment , showing how these can be used to test these relations .
we comment on the case where @xmath59 in section [ sec : beyond ] , and discuss renormalisation group effects in section [ sec : rge ] .
finally , section [ sec : conclusion ] concludes the paper .
in the first subsection , we present a simple derivation of the solar sum rule of eq .
( [ sol4 ] ) .
then in later subsections we discuss the leading - order mixing patterns which one encounters in the considered class of models .
we shall find that there are only four well - motivated patterns of interest , whose relation to model building will be discussed . in the equations that follow , superscripts are attached to quantities which are naturally associated with the neutrinos or the charged leptons ( _ e.g._@xmath60 and @xmath61 ) , whilst physical parameters go without . assuming .
we comment on one example in section [ sec : beyond ] . ]
@xmath62 , the pmns matrix is given by the product of five unitary matrices @xmath63 the three right - most matrices describe the neutrino sector , and are parameterised by @xmath64 and @xmath65 is a diagonal matrix of uni - modular complex numbers .
the two unitary matrices on the left of eq .
( [ 123 ] ) characterise the charged - lepton corrections , and will be allowed to include extra complex phases , @xmath66 with these definitions , it is simple enough to compute the explicit form of the pmns matrix .
however , our derivation focuses only on the first two elements of the bottom row of the physical pmns matrix , which are found to be @xmath67 by comparing eq .
( [ ucorr ] ) to the pdg parameterisation of @xmath23 @xcite , we find the relations between the physical parameters and our internal parameters , @xmath68 and @xmath69 , we are left with a correlation between observable parameters and the value of the neutrino mixing parameter @xmath48 , = = t^_12 .
[ sum1 ] this correlation will be referred to as the solar mixing sum rule .
it can be viewed as a predictive statement about the physical cpv phase : squaring both sides of eq .
( [ sum1 ] ) and solving for @xmath10 leads us to the expression in eq .
( [ sol4 ] ) , which we repeat below , = .
[ sol4p ] an equivalent correlation has been derived previously using a lengthier argument in refs .
@xcite and @xcite . understanding its application to specific models , its compatibility with global data and its potential use as a signature of new physics will be the focus of the rest of this article .
the correlation in eq .
( [ sol4 ] ) is in fact the full non - linear version of a more familiar first - order relation .
we collect a number of phenomenologically interesting approximations in appendix [ sec : simpapp ] . if we expand eq .
( [ sol4 ] ) in a small parameter @xmath70 , assumed to control the deviation from a leading - order neutrino mixing pattern with maximal atmospheric mixing , @xmath71 we find the well - known first - order relation @xcite , @xmath72 the validity of this approximation is dependent upon the severity of the assumptions in eq .
( [ eq : approx_parameters_linear ] ) .
this can only be assessed on a model dependent basis ; however , in fig .
[ fig : cosdelta_error ] we show the size of the error @xmath73 which is introduced by the linear approximation for the patterns which we will derive in subsections [ sec : fully ] and [ sec : partially ] .
apart from the patterns denoted gr3 and bm ( which we will argue in the following section are strongly disfavoured by current data ) , the error approximately satisfies @xmath74 . as deviations of this size
are expected to be close to the attainable precision at a next - generation oscillation facility , all subsequent numerical work will use the full correlations in eq .
( [ sol4 ] ) .
induce an error of less than @xmath75 ( @xmath76 ) for @xmath77 ( @xmath78 ) of the range of @xmath7 .
therefore the linearized expressions well describe the correlation to the precision of the first phases of the next - generation of superbeams , which expect a sensitivity of @xmath79@xmath80 ; however , the full expressions will be necessary in the subsequent phases , where precisions are expected to be @xmath81@xmath82 . ] [ cols="^,^ " , ]
neutrino mixing patterns in which @xmath84 have also been predicted in the literature , and can likewise give rise to solar sum rules . however , eq . ( [ sol4 ] )
does not apply in such scenarios , and we must work on a case by case basis .
examples of patterns of this type can be found in the fully - specified patterns in ref .
@xcite . of these patterns ,
we identify 10 unique mixing matrices from the sporadic patterns which meet our criteria on the mixing angles and have @xmath85 , along with an infinite subset of the family @xmath86 . as an example
, we shall derive another sum rule from one member of @xmath86 characterised by a purely imaginary parameter @xmath87 , cf .
( 150 ) of ref .
this pattern had been previously studied in a grand - unified model of flavour based on the group @xmath88 @xcite , where it was known as the bi - trimaximal mixing ( btm ) pattern , u^ _
= a_+ & & a_- + - & & + a_- & - & a_+ , with @xmath89 .
multiplication of the charged lepton mixing @xmath90 from the left yields the pmns matrix ( up to majorana and unphysical phases ) u = a_+c^e_12 + s^e_12e^-i^e_12 & c^e_12 - s^e_12e^-i^e_12 & a_-c^e_12 - s^e_12e^-i^e_12 + a_+s^e_12e^i^e_12 - c^e_12 & s^e_12e^i^e_12 + c^e_12 & a_-s^e_12e^i^e_12 + c^e_12 + a_-&- & a_+ .
the two free continuous parameters @xmath91 and @xmath92 will control the four physical parameters @xmath6 , @xmath58 , @xmath5 and @xmath7 .
therefore we expect two sum rules .
they can be derived easily by comparing the ( square of the ) absolute value of the @xmath93 entry with the corresponding entry in the pdg parameterisation of @xmath23 . from @xmath94
we find the first exact sum rule involving the atmospheric angle c_13 c_23 = a_+ .
[ eq : bt1 ] similarly we get from @xmath95 c_23 ^ 2 s_12 ^ 2 s_13 ^ 2 + s_23 ^ 2 c_12 ^ 2 + 2 s_23 c_23s_12c_12 s_13 = .
[ eq : bt2a ] solving eq .
( [ eq : bt1 ] ) for @xmath5 and inserting the result into eq .
( [ eq : bt2a ] ) gives rise to the sum rule involving @xmath6 , @xmath58 and @xmath7 , = .
[ eq : bt2 ] satisfying both of these constraints simultaneously is very difficult with the known global data on the mixing angles : @xmath96 is only well defined for large values of @xmath58 and small values of @xmath6 , but the constraint of eq .
( [ eq : bt1 ] ) requires that @xmath58 and @xmath5 are at the low - valued extremes of their allowed parameter space .
this tension may be alleviated by introducing further corrections to these predictions , for example the renormalisation effects .
in this analysis we have ignored the effects of renormalisation group ( rg ) corrections to mixing angles .
although this is generally a good approximation , it is useful to be aware of the typical magnitudes of such corrections and when they might be important . in this section ,
we briefly review such issues . for
previous discussion of rg corrections in this context see _
e.g. _ ref .
@xcite for a discussion in case of cabibbo - like charged lepton correction to bm mixing and charged lepton corrections to tbm in ref .
@xcite . in the effective theory the rg correction to @xmath6 , which is generically the largest ,
is described by the following renormalisation group equation @xcite @xmath97 with @xmath98 in the minimal supersymmetric standard model ( mssm ) and @xmath99 in the standard model ( sm ) . in the mssm
large values of @xmath100 lead to an enhancement of the rg running via the yukawa coupling @xmath101 , where @xmath102gev .
there is no such enhancement in the sm . in order to estimate the size of the rg corrections , we use the exact analytic rg equation for @xmath6 fixing all parameters at their respective best fit values .
we do not include the running of the other parameters .
the running of the solar mixing angle from the electroweak scale to the gut scale @xmath103gev is given to leading order by @xmath104 .
it depends on the absolute neutrino mass scale . in the sm the corrections are generally small . using
the exact becomes relevant in eq .
( [ eq : rgt12approx ] ) and the exact expression of @xmath105 has to be used .
] one - loop formula for @xmath6 given in ref .
@xcite , we obtain a conservative estimate by taking a quasi - degenerate mass spectrum with vanishing phases : @xmath106 ( @xmath107 ) for @xmath108 ev ( @xmath109 ) . experimentally , the expected sensitivity for determining @xmath6 is about @xmath110 @xcite , i.e. @xmath111 . in the context of the sm
it is therefore justified to neglect rg effects on the solar sum rules discussed in this paper . on the other hand , rg corrections
are typically bigger in the mssm .
we have considered 4 different cases : normal and inverted mass ordering and a majorana phase difference @xmath112 of @xmath113 and @xmath114 . with @xmath115 ,
rg effects become more and more relevant for increasing absolute neutrino mass scales and increasing values of @xmath116 . for an inverted mass ordering , the lower bound on @xmath117
entails relevant rg corrections for @xmath118 . in the case of a normal mass ordering ,
rg corrections have to be taken into account if @xmath119 , where this bound is decreasing with increasing @xmath117 , _ e.g._with @xmath120 one has @xmath121 .
a non - vanishing majorana phase difference generically suppresses rg running . for @xmath122 , the leading term in eq .
( [ eq : rgt12approx ] ) is proportional to @xmath123 , such that it
_ decreases _ with increasing absolute neutrino mass scale .
rg corrections to the solar sum rules can be neglected for @xmath124 provided the neutrino mass spectrum is not quasi - degenerate .
we have presented a succinct derivation of solar lepton mixing sum rules , arising from simple patterns of neutrino mixing with @xmath125 , enforced by discrete family symmetry and corrected by a rather generic charged lepton mixing matrix , assuming only that @xmath126 . from our derivation we have expressed the result as the ratio of the absolute magnitude of two pmns matrix elements , given in terms of @xmath127 , namely @xmath128 .
when expanded in terms of the three pmns mixing angles and the cpv oscillation phase @xmath7 , the resulting solar sum rule may be cast in terms of a prediction for @xmath10 which depends only on @xmath127 .
we have considered in detail the resulting solar mixing sum rules , arising from four particularly well - motivated cases of neutrino mixing , which can be derived from discrete family symmetries , namely : bm mixing where @xmath50 , tbm mixing where @xmath35 , and two patterns based on versions of golden ratio mixing including gr1with @xmath51 and gr3with @xmath129 , where @xmath130 is the golden ratio .
we have also fully discussed two leading - order patterns which have been invoked in the literature , but which can not be enforced by any simple discrete family symmetry , called gr2with @xmath131 and hexmixing with @xmath55 .
it turns out that two of the above six cases , namely bmand gr3 , are almost excluded by current data , so in the phenomenological study , we have focused on the remaining four viable cases , namely tbm , gr1 , gr2and hex , of which only the first two ( tbmand gr1 ) are well founded by symmetry arguments .
the predictions for @xmath10 for all these cases are summarised in fig .
[ fig : mccosdelta ] .
for the four viable cases , we performed a simulation of a next - generation superbeam experiment , based on lbno , and a future reactor experiment , based on juno , to see how well the sum rules can be tested .
for example , in fig .
[ fig : solar_sum_rules ] we show the allowed regions of true parameter space in the @xmath132 plane , following a 6-year medium - baseline reactor experiment and a decade of running with a wbb and a 35 kton detector .
we have seen that the ability to constrain solar sum rules relies crucially on the complementary sensitivities of both reactor and superbeam experiments , and that , acting together , these facilities will be capable of significantly restricting the allowed parameter space of the models associated with solar mixing sum rules .
it is possible that such experiments could exclude all of the models considered in this paper , which would be the case if , for example @xmath133 .
if this occurs , the theoretical approach of explaining lepton mixing as a result of charged - lepton corrections to simple symmetry - driven patterns of neutrino mixing would become strongly disfavoured .
we acknowledge support from the european union fp7 itn - invisibles ( marie curie actions , pitn- ga-2011 - 289442 ) .
the work of cl is supported by the deutsche forschungsgemeinschaft within the research unit for 1873 ( quark flavour physics and effective field theories ) .
ms acknowledges support by the australian research council . this work has been additionally supported by the european research council under erc grant `` numass '' ( fp7-ideas - erc erc - cg 617143 ) .
in this appendix we show that the exact sum rule in eq .
( [ sol4 ] ) reduces to the well - known leading - order sum rule to an accuracy of a few percent . if we drop terms proportional to @xmath134 in eq .
( [ sol4 ] ) , we obtain the approximate sum rule , @xmath135 this sum rule can be written to leading order in @xmath58 as , .
[ sum5 ] if we write @xmath136 , then to leading order in @xmath137 , @xmath138 hence eq .
( [ sum5 ] ) becomes , to leading order in @xmath137 and @xmath58 , _ 12-^_12 .
[ sum6 ] if we write @xmath139 , then to leading order in @xmath140 , @xmath137 and @xmath58 we find , @xmath141 which is the usual well - known leading - order solar sum rule @xcite .
the corrections to this linearised sum rule are of order @xmath142 , @xmath143 and @xmath144 .
for example , the second order correction from the reactor angle is @xmath145 .
the formulae preceding eq .
( [ sum7 ] ) may be used for a more accurate approximate description of the sum rule .
for example , eq . ( [ sum6 ] ) could be used to better account for the atmospheric mixing angle deviating from maximal .
the possibility that the full @xmath146 symmetry of the neutrino mass term is a subgroup of the flavour symmetry was considered from a bottom - up perspective in ref . @xcite .
it is a simple extension of the authors previous constraints ( see ref .
@xcite for details ) : each neutrino generator @xmath147 fixes one column of the mixing matrix using the formulae , @xmath148 where @xmath149 such that @xmath150 and @xmath151 with the requirement that @xmath152 and @xmath153 ( @xmath154 and @xmath155 ) are coprime , and @xmath156 . neglecting for the time being the case @xmath157 or @xmath158 , the constraint of unitarity implies that fixing two columns of the mixing matrix using these formulae fully specifies the pattern .
each column is given by a permutation of @xmath159 , where the @xmath160 parameter for the first and second column are given , respectively , by @xmath161 we can generate the consistent mixing patterns by choosing those @xmath162 and @xmath163 parameters which generate finite groups .
considering only patterns related to von dyck groups with irreducible triplets , _
i.e._a@xmath0 , s@xmath0 and a@xmath1 , the result is the set of numbers listed in eq .
( [ eq : choices ] ) .
as we are assuming a single @xmath164 residual symmetry in the charged - lepton sector , we must also take a common value of @xmath153 , _ i.e. _ @xmath165 .
up to row and column permutations , there are two ways of relatively aligning the elements of the two fixed columns . in one case , we choose the unique element fixed by the two real constraints to be in the same row , @xmath166 in the other case , we choose these elements to be misaligned , @xmath167 however , patterns of the form given in eq .
( [ pmns_bad_pattern ] ) are not possible if the residual symmetry in the charged - lepton sector has only a single generator : it is the choice of @xmath168 which specifies the row @xmath169 of the @xmath160 parameter . under our assumption of a single cyclic charged - lepton symmetry , @xmath164 , we only need to consider patterns of the form of eq .
( [ pmns_pattern ] ) and its row and column permutations . up to now ,
we have refrained from discussing the case related to dihedral groups where @xmath170 .
in fact , such a scenario does not give rise to any new eligible neutrino patterns .
if @xmath153 or @xmath155 take the value @xmath171 , the symmetry constraints from the generators of this subgroup in eq .
( [ eq : app_hs_const ] ) do not necessarily fix a column of matrix elements , leaving parts of the leading - order mixing matrix unspecified .
therefore , we will only consider the choice which does indeed fix a column completely ; it is given by @xmath172 with @xmath173 and yields @xmath174 then , the only way to fully specify the mixing matrix is to either apply this constraint twice , or to apply this constraint in conjunction with one related to a@xmath0 , s@xmath0 or a@xmath1 . by scanning through the row and column permutations of eq .
( [ pmns_pattern ] ) , we find that the only viable choices are ( unsurprisingly ) when the dihedral constraint is applied once and is used to fix @xmath175 ; however , all patterns of this type reproduce patterns already discussed previously .
f. p. an _ et al . _ [ daya - bay collaboration ] , phys .
* 108 * ( 2012 ) 171803 [ arxiv:1203.1669 ] ; j. k. ahn _
[ reno collaboration ] , phys .
* 108 * ( 2012 ) 191802 [ arxiv:1204.0626 ] .
m. c. gonzalez - garcia , m. maltoni and t. schwetz , arxiv:1409.5439 .
d. v. forero , m. tortola and j. w. f. valle , arxiv:1405.7540 .
f. capozzi , g. l. fogli , e. lisi , a. marrone , d. montanino and a. palazzo , phys .
d * 89 * ( 2014 ) 093018 [ arxiv:1312.2878 ] .
p. f. harrison , d. h. perkins and w. g. scott , phys .
b * 530 * ( 2002 ) 167 [ hep - ph/0202074 ] .
s. f. king and c. luhn , jhep * 0910 * ( 2009 ) 093 [ arxiv:0908.1897 ] . s. f. king , jhep * 0508 * ( 2005 ) 105 [ hep - ph/0506297 ] . s. f. king and c. luhn , rept .
* 76 * ( 2013 ) 056201 [ arxiv:1301.1340 ] .
s. f. king , a. merle , s. morisi , y. shimizu and m. tanimoto , new j. phys .
* 16 * ( 2014 ) 045018 [ arxiv:1402.4271 ] .
s. f. king , phys .
b * 659 * ( 2008 ) 244 [ arxiv:0710.0530 ] .
d. hernandez and a. y. smirnov , phys .
d * 86 * ( 2012 ) 053014 [ arxiv:1204.0445 ] .
d. hernandez and a. y. smirnov , phys .
d * 87 * ( 2013 ) 5 , 053005 [ arxiv:1212.2149 ] .
p. ballett , s. f. king , c. luhn , s. pascoli and m. a. schmidt , phys .
d * 89 * ( 2014 ) 016016 [ arxiv:1308.4314 ] .
s. antusch , s. f. king , c. luhn and m. spinrath , nucl .
b * 856 * ( 2012 ) 328 [ arxiv:1108.4278 ] .
s. f. king , jhep * 1307 * ( 2013 ) 137 [ arxiv:1304.6264 ] ; s. f. king , phys .
b * 724 * ( 2013 ) 92 [ arxiv:1305.4846 ] .
i. masina , phys .
b * 633 * ( 2006 ) 134 [ hep - ph/0508031 ] .
s. antusch and s. f. king , phys .
b * 631 * ( 2005 ) 42 [ hep - ph/0508044 ] .
s. antusch , p. huber , s. f. king and t. schwetz , jhep * 0704 * ( 2007 ) 060 [ hep - ph/0702286 ] .
d. marzocca , s. t. petcov , a. romanino and m. c. sevilla , jhep * 1305 * ( 2013 ) 073 [ arxiv:1302.0423 ] .
s. t. petcov , arxiv:1405.6006 ; m. fukugita , m. tanimoto and t. yanagida , phys .
d * 57 * ( 1998 ) 4429 [ hep - ph/9709388 ] ; v. d. barger , s. pakvasa , t. j. weiler and k. whisnant , phys .
b * 437 * ( 1998 ) 107 [ hep - ph/9806387 ] ; s. davidson and s. f. king , phys .
b * 445 * ( 1998 ) 191 [ hep - ph/9808296 ] .
a. datta , f. s. ling and p. ramond , nucl .
b * 671 * ( 2003 ) 383 [ hep - ph/0306002 ] ; l. l. everett and a. j. stuart , phys .
d * 79 * ( 2009 ) 085005 [ arxiv:0812.1057 ] ; f. feruglio and a. paris , jhep * 1103 * ( 2011 ) 101 [ arxiv:1101.0393 ] . c. s. lam , phys .
d * 83 * ( 2011 ) 113002 [ arxiv:1104.0055 ] .
r. de adelhart toorop , f. feruglio and c. hagedorn , nucl .
b * 858 * ( 2012 ) 437 [ arxiv:1112.1340 ] .
w. rodejohann , phys .
b * 671 * ( 2009 ) 267 [ arxiv:0810.5239 ] ; a. adulpravitchai , a. blum and w. rodejohann , new j. phys .
* 11 * ( 2009 ) 063026 [ arxiv:0903.0531 ] . c. h. albright , a. dueck and w. rodejohann , eur .
j. c * 70 * ( 2010 ) 1099 [ arxiv:1004.2798 ] .
s. t. petcov and m. piai , phys .
b * 533 * ( 2002 ) 94 [ hep - ph/0112074 ] .
x. qian , d. a. dwyer , r. d. mckeown , p. vogel , w. wang and c. zhang , phys .
d * 87 * ( 2013 ) 3 ,
033005 [ arxiv:1208.1551 ] . s. -f .
ge , k. hagiwara , n. okamura and y. takaesu , jhep * 1305 * ( 2013 ) 131 [ arxiv:1210.8141 ] .
y. -f .
li , j. cao , y. wang and l. zhan , phys .
d * 88 * ( 2013 ) 1 , 013008 [ arxiv:1303.6733 ] .
w. wang , talk presented at invisibles 13 at lumley castle , durham on 16/07/2013 .
j. park , aip conf .
* 1604 * ( 2014 ) 421 . j. park , pos neutel * 2013 * ( 2013 ) 076 .
c. adams _ et al .
_ [ lbne collaboration ] , arxiv:1307.7335 .
s. agarwalla : _ et al . _
[ laguna - lbno collaboration ] , arxiv:1312.6520 .
p. ballett , s. f. king , c. luhn , s. pascoli and m. a. schmidt , arxiv:1406.0308 . s. gollu , k. n. deepthi and r. mohanta , mod .
. lett .
a * 28 * ( 2013 ) 31 , 1350131 [ arxiv:1303.3393 ] .
sruthilaya m. , soumya c. , k. n. deepthi and r. mohanta , arxiv:1408.4392 .
i. girardi , s. t. petcov and a. v. titov , arxiv:1410.8056 . k. a. olive _ et al .
_ [ particle data group collaboration ] , chin .
c * 38 * ( 2014 ) 090001 .
p. coloma , a. donini , e. fernandez - martinez and p. hernandez , jhep * 1206 * ( 2012 ) 073 [ arxiv:1203.5651 ] .
a. de gouvea _ et al . _
[ intensity frontier neutrino working group collaboration ] , arxiv:1310.4340 .
r. m. fonseca and w. grimus , jhep * 1409 * ( 2014 ) 033 [ arxiv:1405.3678 ] .
a. blum , c. hagedorn and m. lindner , phys .
d * 77 * ( 2008 ) 076004 [ arxiv:0709.3450 ] .
f. capozzi , e. lisi and a. marrone , phys .
d * 89 * ( 2014 ) 013001 [ arxiv:1309.1638 ] .
m. grassi , j. evslin , e. ciuffoli and x. zhang , arxiv:1401.7796 .
a. stahl , c. wiebusch , a. m. guler , m. kamiscioglu , r. sever , a. u. yilmazer , c. gunes and d. yilmaz _ et al .
_ , cern - spsc-2012 - 021 .
a. longhin , arxiv:1206.4294 .
p. huber , m. lindner and w. winter , comput .
commun .
* 167 * ( 2005 ) 195 [ hep - ph/0407333 ] .
p. huber , j. kopp , m. lindner , m. rolinec and w. winter , comput .
commun .
* 177 * ( 2007 ) 432 [ hep - ph/0701187 ] .
s. f. king , c. luhn and a. j. stuart , nucl .
b * 867 * ( 2013 ) 203 [ arxiv:1207.5741 ] .
m. a. schmidt and a. y. smirnov , phys .
d * 74 * ( 2006 ) 113003 [ hep - ph/0607232 ] .
s. boudjemaa and s. f. king , phys .
d * 79 * ( 2009 ) 033001 [ arxiv:0808.2782 ] . s. antusch , j. kersten , m. lindner and m. ratz , nucl . phys .
b * 674 * ( 2003 ) 401 [ hep - ph/0305273 ] . | small discrete family symmetries such as s@xmath0 , a@xmath0 or a@xmath1 may lead to simple leading - order predictions for the neutrino mixing matrix such as the bimaximal , tribimaximal or golden ratio mixing patterns , which may be brought into agreement with experimental data with the help of corrections from the charged - lepton sector .
such scenarios generally lead to relations among the parameters of the physical leptonic mixing matrix known as solar lepton mixing sum rules . in this article , we present a simple derivation of such solar sum rules , valid for arbitrary neutrino and charged lepton mixing angles and phases , assuming only @xmath2 .
we discuss four leading - order neutrino mixing matrices with @xmath3 which are well motivated from family symmetry considerations .
we then perform a phenomenological analysis of the scope to test the resulting four solar sum rules , highlighting the complementarity between next - generation neutrino oscillation experiments such as the reactor experiment juno and a superbeam experiment .
= 1 + |
the understanding of the @xmath0-process contribution to the isotopic abundances of heavy elements in the solar system is fundamental to disentangle between several nucleosynthesis processes that have competed during the evolution of the milky way .
the @xmath0-process abundances observed in the solar system are the result of a complex galactic chemical evolution process , which mainly accounts for the pollution of several agb generations with different initial masses and metallicities .
it was shown that agb stars with low initial masses ( @xmath4 = 1.5 and 3 @xmath5 ) , half solar metallicity , and a specific @xmath6c - pocket choice ( called case st ) reproduce the @xmath7 component of the @xmath0-process ( arlandini et al .
1999 ) , which synthesized about half of the elements from sr to pb .
this approximation still provides strong information about the @xmath0-process contribution to isotopes in the region between @xmath8ba and @xmath9pb ( see discussion in bisterzo et al .
+ two additional @xmath0-process components should be considered to reproduce the solar abundances of light neutron capture isotopes up to sr ( the @xmath10 component ) and the stable isotopes at the termination point of the @xmath0-path , @xmath11pb and @xmath12bi ( the @xmath13 component ) .
the @xmath10 process occurs in massive stars during core he and shell c burning and partly produces neutron capture isotopes lighter than a @xmath14 90 ( @xmath15sr @xmath1410% ; lower contribution to y and zr isotopes ; see e.g. , pignatari et al .
agb stars with low metallicity and low initial mass synthesize about half of solar @xmath11pb , the so called @xmath13 component ( see gallino et al .
1998 ; travaglio et al . 2001 ) .
the @xmath6c in the pocket is primary , which means it is directly synthesized in the star independently of the initial composition .
therefore , by decreasing the metallicity , the number of free neutrons per iron seed becomes so large to overcome the first and the second @xmath0-peaks ( sr - y - zr and ba - la - ce , respectively ) , to feed directly @xmath11pb ( and @xmath12bi ) . actually , the understanding of the origin of light @xmath0-process isotopes sr , y and zr is more enigmatic .
gce models by travaglio et al .
( 2004 ) found that agb yields underestimate the solar abundance of sr , y , zr , and the solar composition of isotopes from @xmath16sr to @xmath17xe by about 2030% ( including the @xmath0-only @xmath18mo , @xmath19ru , @xmath20pd , @xmath21cd , @xmath22sn , @xmath23te and @xmath17xe ) .
the @xmath10 process can not compensate the missing solar abundance , because it mainly produces isotopes up to sr , with a negligible contribution up to @xmath17xe .
spectroscopic observations of peculiar metal - poor stars showing large enhancement in @xmath24-process elements neutron capture process is ascribed to explosive nucleosynthesis phases of massive stars . ]
suggest that @xmath1410% of solar sr - y - zr is due to the @xmath24-process .
however , both @xmath0- and @xmath24-processes are not sufficient to explain the solar observations of light neutron capture elements .
travaglio et al .
( 2004 ) hypothesized the existence of an additional process of unknown origin , called by the authors lepp ( light element primary process since it was supposed to be of primary origin ) .
several scenarios have been recently explored , both involving the secondary @xmath0-process in massive stars ( e.g. , @xmath25-component by pignatari et al .
2013 , which may explain the missing @xmath0-process component ) or primary @xmath24-process during the advanced phases of explosive nucleosynthesis ( see review thielemann et al . 2011 , which may account of complementary @xmath24-contributions ) . therefore , even if promising theoretical improvements related to the explosive phases of massive stars and core collapse supernovae , as well as recent spectroscopic investigations ( roederer et al . 2012 ; hansen et al .
2012 ) have been made , a fully understanding about the origin of the neutron capture elements from sr up to xe is still lacking .
we aim to investigate the effects of a new set of agb yields ( with updated nuclear cross section network and solar abundances ) on the solar @xmath0-process composition ( section 2 ) .
we focus the analysis on the variations of the solar - system @xmath0-process gce predictions by adopting updated agb yields , and by testing the effects of different prescriptions on nuclear cross sections .
the galactic evolution is computed as function of time up to the present epoch ( @xmath26 = 13.73 @xmath27 0.12 by wmap ) , following the three zones of the galaxy , halo , thick and thin disk .
we adopted the yields by rauscher et al .
( 2002 ) and travaglio et al .
( 2004b ) for sne ii and ia , respectively .
major revisions involve the solar system meteoritic abundances by lodders , palme and gail ( 2009 ) , a neutron capture cross section network with the most recent published measurements , as well as a larger set of agb yields that extend toward lower initial mass ( down to @xmath4 = 1.3 @xmath5 ) .
we started from the agb models presented by bisterzo et al .
( 2010 ) , which were based on the franec code by straniero et al .
( 1995 , 2000 , 2003 ) . +
a series of thousands of new agb models have been run , for a total of 28 metallicities from [ fe / h ] = + 0.2 down to @xmath283.6 ( most of them focused on the metallicity range between solar and [ fe / h ] = @xmath281.6 , where the isotopes of the three @xmath0-peaks are largely produced , see travaglio et al . 1999 , 2001 , 2004 ;
serminato et al .
yields of a set of five agb models with low initial masses ( @xmath4 = 1.3 , 1.4 , 1.5 , 2 , and 3 @xmath5 ) and two agb models with intermediate initial masses ( @xmath4 = 5 and 7 @xmath5 ) have been interpolated / extrapolated over the whole metallicity and mass range ( hereafter lms refers to the mass range between 1.3 @xmath29 @xmath30 @xmath31 4 , and ims to 4 @xmath29 @xmath30 @xmath31 8) .
this assures a sufficiently high accuracy in the agb mass and metallicity ranges . particularly challenging for the @xmath0-process
is the understanding of the formation of the @xmath6c - pocket , specifically the mass fraction of @xmath6c and @xmath32n in the pocket and the mass involved .
both uncertainties largely affect the @xmath0-yields ( see e.g. , straniero et al .
2006 ; herwig 2005 ) .
current full evolutionary agb models still adopt a free parametrization to reproduce the @xmath6c - pocket , by means of overshooting ( e.g. , herwig 2000 , herwig et al .
2011 ; karakas 2010 ) or other better physically justified prescriptions ( e.g. , new fruity models by cristallo et al .
2009 , 2011 , and references therein ; or mixing produced by magnetic fields , busso et al .
as suggested by the @xmath0-process spread observed at a given metallicity in different stellar populations ( e.g. , post - agb , ba , ch and cemp - s stars ; sneden et al .
2008 ; kppeler et al 2011 ) , a range of the @xmath0-process efficiency strengths is needed .
the theoretical reason of this spread observed among @xmath0-elements is still under investigation ( agb initial mass , magnetic fields , gravitational waves , or rotation , see piersanti , cristallo and straniero 2013 ) . + as discussed by bisterzo et al .
( 2010 ) , we artificially introduce the @xmath6c - pocket in our post - process agb calculations , and we treated it as a free parameter kept constant pulse by pulse .
starting from the @xmath6c - pocket st case , similar to that adopted by gallino et al .
( 1998 ) , we multiply or divide the @xmath6c ( and @xmath32n ) abundances in the pocket by different factors .
we considered an accurate weighted average of the @xmath6c - pocket efficiencies in order to reproduce @xmath14100% of solar @xmath33sm and the other @xmath0-only isotopes heavier than @xmath34 @xmath14 90 .
note that @xmath33sm has well defined solar abundance , and it is the @xmath0-only isotope less affected by branchings and nuclear cross section uncertainties as well . a second reaction , the @xmath1ne(@xmath2 , n)@xmath3 mg , starts to partially burn at @xmath35 = 3 during thermal pulses and produces an efficient neutron burst mainly affecting isotopes close to the branching points . results are shown in figure 1 .
the @xmath0-only isotopes are indicated by solid circles .
different symbols have been used for isotopes that receive additional contributions : @xmath36xe , @xmath37gd , and @xmath38er ( open squares ) , which have a non - negligible p contribution ; @xmath39lu ( open triangle ) , a long - lived isotope ( 3.8 x 10@xmath40 yr ) , which decays into @xmath39hf ; @xmath41os ( open triangle ) , which is affected by the long - lived decay of @xmath41re ( 4.1 x 10@xmath40 yr ) ; @xmath42ta ( open circle ) , which also receives contributions from the p process and from @xmath43-nucleus interactions in massive stars ; @xmath11pb ( filled big square ) , half of which is produced by the strong-@xmath0 component .
we compare the @xmath0-process predictions , calculated with the gce model at the epoch of the solar system formation ( @xmath44 = 9.17 gyr from the beginning of the birth of the universe ) , with the meteoritic abundances by lodders et al .
( 2009 ) . as shown in the @xmath45 @xmath46 , lms ( black symbols )
reproduce almost all the solar system @xmath0-only isotopes between @xmath34 = 140 and 210 .
an additional small contribution ( @xmath3110% ) comes from ims for isotopes with @xmath34 @xmath31 140 .
the total @xmath0-percentages ( lms + ims ) is represented by blue symbols .
ims agbs play a minor role in the galactic enrichment , because their he - intershell is smaller than in lms by one order of magnitude , with an uncertain formation of the @xmath6c - pocket and a less efficient third dredge - up .
the @xmath1ne(@xmath2 , n)@xmath3 mg neutron source is efficiently activated in ims due to the high temperature reached at the bottom of the thermal pulses ( @xmath35 = 3.5 ) .
the peak neutron density reached in ims easily allows a strong overproduction of @xmath16kr , @xmath47rb , and @xmath18zr , three neutron rich isotopes affected by the branchings at @xmath48kr and @xmath49zr , which are very sensitive to the neutron density .
+ particularly large is the abundance of @xmath18zr ( big asterisk ) .
note that @xmath18zr is strongly sensitive to the number of thermal pulses experienced by the agb models , as well as to the @xmath1ne(@xmath2 , n)@xmath3 mg rate adopted . in general , an over - prediction of @xmath18zr may suggest a low number of thermal pulses for ims . as discussed by bisterzo et al .
( 2010 ) , we assume a strong mass loss for ims , which allows a total of 24 thermal pulses . under this hypothesis ,
ims produce about 30% of solar @xmath18zr , while an additional 70% comes from lms ( figure 1 , @xmath45 @xmath46 ) .
@xmath18zr is the most neutron rich stable zr isotope , and a contribution of 100% from agb is surely overestimated .
moreover , it disagrees with recent @xmath50-process predictions by travaglio et al .
( 2011 ) , which estimate an additional non - negligible contribution to @xmath18zr by snia ( up to 30% ) . as a further improvement
, we computed a new set of agb yields that includes a new evaluation of the @xmath1ne(@xmath2 , n)@xmath3 mg rate , which takes into account the most recent experimental information .
the recommended value we are suggesting is very close to jaeger et al .
( 2001 ) and agrees with the recent determination of longland et al .
( 2012 ) . at agb temperature
( @xmath35 @xmath14 2.5 to 3.5 ) the new @xmath1ne(@xmath2 , n)@xmath3 mg rate is about a factor of 2 lower than our rate used so far ( corresponding to the lower limit suggested by kppeler et al .
1994)ne(@xmath2 , @xmath51)@xmath52 mg . specifically , at the temperature of interest of agbs , between @xmath35 = 2.5 to 3.5 , this rate is almost unchanged . ] .
the related new solar - system abundances are displayed in figure 1 , @xmath53 @xmath46 , red symbols .
as expected , major differences are shown close to the branchings points .
in particular , @xmath1450% of solar @xmath18zr is produced by agbs , in better agreement with expectations .
+ updated gce calculations plausibly reproduce within the uncertainties all @xmath0-only isotopes with @xmath34 @xmath54 130 , and confirm the missing 20% solar @xmath0-contribution of @xmath0-only isotopic abundances between @xmath34 = 96 130 found by travaglio et al .
variations with respect to the results presented by kppeler et al .
( 2011 ; their fig . 15 , bottom panel )
are mainly due to new solar abundances , recent neutron capture cross section measurements , and the new evaluation of the @xmath1ne(@xmath2 , n)@xmath3 mg rate . + note that @xmath55pt and @xmath56hg are affected by large uncertainties : concerning @xmath55pt , the neutron capture cross section of @xmath57os and @xmath55ir evaluated theoretically at 22% , the extrapolation of the @xmath55ir measurement in stellar conditions ( see discussion by rauscher et al . 2012 ) , as well as the old measurement of the @xmath55pt(n , @xmath51 ) reaction , with 20% of uncertainty at 30 kev ( bao et al .
2000 ) ; moreover , hg is too volatile for a reliable experimental determination of the solar abundance ( lodders et al .
2009 estimated an uncertainty of 20% ) .
@xmath9pb ( and all pb isotopes ) have well determined neutron capture cross sections ( see kadonis ) , but it is strongly affected by the branching at @xmath9tl , with variations of @xmath1410% .
updated solar @xmath0-process abundances of some selected elements are compared with previous gce computations ( travaglio et al .
2001 , 2004 ) in table [ tab1 ] . marginal differences ( @xmath315% ) are seen in general . la and ce are among the few exceptions :
the larger @xmath0-contribution obtained by this work ( + 12% and + 6% ) is the consequence of the new @xmath58la(n , @xmath51)@xmath59la rate measured by winckler et al .
( 2006 ) . a more detailed comparison and a discussion concerning the most relevant updated information will be provided in a future work .
5cm*13c & & & & & & & & & & & & & & & & & & & & & el & s(2001 - 2004 ) & s(this work ) & r(this work ) & & & & & & & & & sr & 71 & 67 & & & & & & & & & & y & 69 & 70 & & & & & & & & & & zr & 65 & 64 & & & & & & & & & & ba & 80 & 83 & 17 & & & & & & & & & la & 61 & 73 & 27 & & & & & & & & & ce & 75 & 81 & 19 & & & & & & & & & pr & 47 & 49 & 51 & & & & & & & & & nd & 54 & 56 & 44 & & & & & & & & & sm & 30 & 31 & 69 & & & & & & & & & eu & 6 & 6 & 94 & & & & & & & & & pb & 91 & 88 & 12 s.b .
wishing to acknowledge jina for financial support under nd fund @xmath60201387 and 305387 .
9 lodders , k. , palme , h. , gail , h .-
2009 , in trmper j. e. , ed .
, landolt- brnstein group vi astronomy and astrophysics numerical data and functional relationships in science and technology , 4b : solar system , 4.4 .
springer - verlag , berlin ( doi:10.1007/978 - 3 - 540 - 88055 - 4 - 34 ) | we study the @xmath0-process abundances at the epoch of the solar - system formation as the outcome of nucleosynthesis occurring in agb stars of various masses and metallicities .
the calculations have been performed with the galactic chemical evolution ( gce ) model presented by travaglio et al .
( 1999 , 2004 ) . with respect to previous works
, we used updated solar meteoritic abundances , a neutron capture cross section network that includes the most recent measurements , and we implemented the @xmath0-process yields with an extended range of agb initial masses .
the new set of agb yields includes a new evaluation of the @xmath1ne(@xmath2 , n)@xmath3 mg rate , which takes into account the most recent experimental information . |
the result that drives this paper is @xmath5 where @xmath6 and @xmath7 are relatively prime positive integers and @xmath8 is the multiplicative inverse of @xmath9 .
the lhs is complicated by the irregularity arising from the floor function , while the rhs is complicated by the presence of a modular inverse ; therein lies the beauty and utility of . before stating the general result of which is a special case ( theorem [ thm : maintheorem ] )
, we need to introduce some notation .
we set @xmath10 , and whenever the range of a summation is not written explicitly , it is to be taken over all of @xmath11 , the integers modulo @xmath7 : @xmath12 we use the fourier transform @xmath13 ( @xmath14 is called the @xmath15-th fourier coefficient ) , the fourier inversion formula @xmath16 convolution @xmath17 and the interchange - of - summations result @xmath18 also , let @xmath19\!\right]$}}\,r : = \begin{cases } 0 & \text{$p$ is false;}\\ r & \text{$p$ is true.}\end{cases}\ ] ] note that @xmath20\!\right]$}}r$ ] is defined even if @xmath21 is not . when @xmath22 , we omit it from the notation . we also adopt the common practice of identifying a multiset with its indicator function , i.e. , @xmath23 is the multiplicity of @xmath24 in the multiset @xmath25 .
we distinguish the rational beatty sets ( @xmath26 are any integers , and @xmath27 any real number ) @xmath28 usually , @xmath7 will be fixed and in this situation we omit it from the notation .
we will always assume that @xmath27 is an integer is an integer . to see this ,
let @xmath29 be a value of @xmath30 for which the fractional part @xmath31 is minimal ; then @xmath32 . ] , and when @xmath33 we omit it from our notation .
note that the density of the set @xmath34 is @xmath2 .
note that @xmath34 consists of @xmath6 congruence classes modulo @xmath7 , and so @xmath34 is naturally considered as a subset of @xmath11 . if the @xmath6 points were perfectly evenly distributed around @xmath11 ( as happens if @xmath35 ) , then the fourier transform would be 0 except at multiples of the difference between points .
thus , one naturally expects that @xmath36 will be small except when @xmath37 is near an integer .
this is confirmed by figure 1 , which shows the fourier transform of @xmath38 , and figures 2 and 3 , which show @xmath39 for small relatively prime @xmath40 ( @xmath27 is irrelevant )
. theorem [ thm : maintheorem ] gives an explicit formula for @xmath41 which quantifies the validity of this expectation .
( 216,450 ) ( 0,225 ) ( @xmath42 ) , shown in the complex plane and labeled by @xmath15 .
the top graph shows all 120 points .
the bottom graph , which resembles a spider , shows only those closest to 1.,title="fig : " ] ( 0,0 ) ( @xmath42 ) , shown in the complex plane and labeled by @xmath15 .
the top graph shows all 120 points .
the bottom graph , which resembles a spider , shows only those closest to 1.,title="fig : " ] [ thm : maintheorem ] let @xmath43 , @xmath44 be integers with @xmath45 , let @xmath46 satisfy @xmath47 , and let @xmath27 be any integer .
then @xmath48 and for @xmath49 @xmath50\!\right]$}}\ , g\ , \frac{1-\omega^j}{1-\omega^{j\bar p } } \ , \omega^{-jr},\\ \left|\widehat{{{\mathcal b}}_{p , r}^q}(j ) \right| & = { \mbox{$\left[\!\left [ g|j \right]\!\right]$}}\ , g\ , \left|\frac{\sin ( \pi j /q)}{\sin ( \pi j \bar p /q)}\right| .
\notag \end{aligned}\ ] ] ( 360,120 ) ( 0,0 ) for @xmath51 , @xmath52 , and @xmath53,title="fig : " ] [ fig : firstcoeff ] ( 360,120 ) ( 0,0 ) , for @xmath51 , @xmath54 , @xmath55.,title="fig : " ] [ fig : secondcoeff ] ( 360,120 ) ( 0,0 ) , for @xmath51 , @xmath54 , @xmath56.,title="fig : " ] figure 2 shows @xmath57 for relatively prime @xmath6 and @xmath7 with @xmath58 and @xmath59 .
figures 3 and 4 show the first and second coefficients when @xmath6 and @xmath7 are relatively prime , @xmath60 , and @xmath61 .
there are three symmetries visible to the naked eye .
the first ( from figure 2 ) is that @xmath62 is periodic with period @xmath7 .
this is a consequence of the fact @xmath63 , which we prove along the way to proving theorem [ thm : maintheorem ] .
the second is that @xmath64 which is seen in the pictures as a symmetry about @xmath65 . this is a consequence of a theorem of fraenkel ( corollary [ cor : fptrational ] below ) which states that the complement of a rational beatty set is a rational beatty set .
we give a new proof of this in section [ sec : fpt ] .
the third symmetry is that the graphs on @xmath66 $ ] , on @xmath67 $ ] , etc .
, seem to be quite similar .
this is essentially the symmetry of the continued - fraction map @xmath68 . for each rational @xmath2
, there is a unique finite sequence @xmath69 $ ] of integers with the properties : for @xmath70 , @xmath71 ; for @xmath72 , if @xmath73 then @xmath74 is positive ; @xmath75 ; and @xmath76 this is the nearest - integer continued fraction ( commonly abbreviated nicf ) .
compare figure 3 with figure 5 , which shows the points @xmath77 : the points in figure 5 are located precisely at the bottom of the `` cups '' in figure 3 .
we remark that while eq .
connects beatty sequences with density @xmath2 directly to the inverse of @xmath6 modulo @xmath7 , the direct connection between both objects and continued fractions is well - studied .
we also note that the inverse of @xmath6 modulo @xmath7 has arisen independently in the recent work of simpson @xcite , in which he uses generating functions to prove necessary and sufficient conditions for two rational beatty sequences to not intersect .
( 360,120 ) ( 0,0 ) , where @xmath69 $ ] is the nicf of @xmath2,title="fig : " ] the main result in the study of beatty sequences was discovered by lord rayleigh : if @xmath78 is irrational and @xmath79 , then the sets @xmath80 , @xmath81 partition @xmath82 . in the 1950s skolem
extended to this non - homogeneous sets , and in 1969 fraenkel corrected skolem s work and extended it to include rational @xmath78 .
we direct the reader to @xcite for the general rational / irrational statement , an elementary proof , and the history of fraenkel s partition theorem .
the rational case ( corollary [ cor : fptrational ] ) is an easy consequence of theorem [ thm : maintheorem ] .
[ cor : fptrational ] the sets @xmath83 and @xmath84 , where @xmath85 are integers , partition @xmath86 if and only if @xmath87 and @xmath88 attempts to extend this to more than two sequences have had some success , but a general statement remains elusive . in the early 1970s
( see @xcite ) , fraenkel was led to conjecture that there was essentially only one way to partition @xmath86 into beatty sets with distinct densities . if the sets @xmath89 partition @xmath86 , @xmath4 , and @xmath90 , then @xmath91 . the conjecture has been proven in case any @xmath78 is irrational @xcite , @xmath92 @xcite , or @xmath93
@xcite , and in several other less - easily - stated circumstances .
the article of r. tijdeman @xcite contains an excellent survey of the progress on fraenkel s conjecture .
we offer a stronger conjecture , and will apply theorem [ thm : maintheorem ] to prove some special cases .
we say that sets @xmath94 are a _
perfect @xmath95-fold covering of @xmath86 _ if @xmath96 for all @xmath97 , and simply that they are a _ perfect covering _ if we do not wish to specify @xmath95 .
let @xmath98 , and let @xmath99 be distinct integers with @xmath100 , , and with no proper subset @xmath101 $ ] having @xmath102 .
let @xmath103 be arbitrary integers .
the sets @xmath104 ( @xmath105 ) are a perfect covering if and only if : 1 .
@xmath7 is odd , and @xmath106 is the order of 2 modulo @xmath7 ; 2 .
the sets can be renumbered so that there are integers @xmath107 ( relatively prime to @xmath7 with modular inverse @xmath108 ) and @xmath109 such that for @xmath105 : @xmath110 this is admittedly not as pithy as fraenkel s conjecture , but we hope that its greater generality will shed new light on an old problem . since @xmath111 , we may assume without loss of generality that @xmath112 .
we use theorem [ thm : maintheorem ] to prove the `` if '' claim of the cfc , and in fact we show that if the covering ( with all @xmath112 ) is a perfect @xmath95-fold covering , then @xmath95 is the number of ones in the binary expansion of @xmath113 .
we prove the `` only if '' part of the cfc under the additional hypothesis that @xmath114 , and also under the additional hypotheses that @xmath115 for some @xmath116 and that @xmath7 is sufficiently large .
we suspect that if @xmath4 sequences @xmath117 are a perfect covering and some @xmath118 is irrational , then two of the @xmath118 s are in fact equal .
graham s proof @xcite of this in the 1-covering case does not extend easily to multiple coverings .
we have not investigated this further , and at any rate this guess does not fall within the scope of the present article .
we will use the following consequence of theorem [ thm : maintheorem ] several times .
[ cor : coveringcriterion ] the sequences @xmath104 ( @xmath105 ) are a perfect @xmath95-fold covering of @xmath86 if and only if @xmath119 and for @xmath120 ( with @xmath121 ) , @xmath122\!\right]$ } } g_k \frac{\omega^{-jr_k}}{1-\omega^{j \bar p_k}}.\ ] ]
we will give a kernel function @xmath123 , whose transform we can compute , such that @xmath124 where @xmath125 will also have an easily computed transform .
taking the fourier transform of this equation gives @xmath126 , which is the same as @xmath127 .
the zeroth coefficient can be dealt with immediately , and thereafter we demonstrate that we may assume that @xmath33 , that @xmath51 , and that @xmath60 . then , we define @xmath128 and @xmath21 , show that holds , and compute the transforms of @xmath21 and @xmath128 to complete the proof .
we have @xmath129 for @xmath130 , this gives the first claim of theorem [ thm : maintheorem ] : @xmath131 .
since we assume that @xmath27 is an integer , we also have @xmath132 thus , it is sufficient to work with @xmath33
. now we wish to show that we may take @xmath6 and @xmath7 to be relatively prime . if @xmath133 is periodic with period @xmath134 , and @xmath135 and @xmath136 are the induced functions on @xmath137 and @xmath138 , then @xmath139\!\right]$ } } \widehat{s_\ell}(j / k).\ ] ] thus , if @xmath140 and @xmath141 with @xmath142 , then @xmath143\!\right]$ } } \widehat{{{\mathcal b}}_a^b}(j / g).\ ] ] if we assume for the moment that we have proved theorem [ thm : maintheorem ] in the relatively prime case , then we have @xmath144\!\right]$ } } \frac{1-(e^{2\pi i /b})^{j / g}}{1-(e^{2\pi
i \bar{a } /b})^{j / g } } \\ & = g { \mbox{$\left[\!\left [ g|j \right]\!\right]$ } } \frac{1-(e^{2\pi i /(bg)})^{j}}{1-(e^{2\pi i \bar{a } /(bg)})^{j}}\\ & = g { \mbox{$\left[\!\left [ g|j \right]\!\right]$ } } \frac{1-\omega^j}{1-\omega^{j\bar{a } } } \end{aligned}\ ] ] where @xmath145 and @xmath146 , the inverse of @xmath147 modulo @xmath148 satisfies @xmath149 .
thus , it is sufficient to work with relatively prime @xmath6 and @xmath7 .
we may assume that @xmath60 since @xmath150 and so the fourier coefficients ( except the zeroth ) are sensitive only to @xmath9 . to see ,
observe that the function @xmath151 takes on only the values @xmath152 and @xmath153 .
moreover , @xmath154 ( with @xmath155 ) if and only if there is an integer @xmath30 with @xmath156 , which is equivalent to @xmath157 this happens if and only if the fractional part @xmath158 is 0 or is strictly larger than @xmath159 .
thus , the property @xmath160 can be described entirely in terms of the fractional part @xmath161 , which depends only on @xmath162 .
now , we take @xmath128 to be the set @xmath163 $ ] , and set @xmath164 .
we use the following two properties of the beatty set @xmath165 ( with @xmath60 ) .
we call the first property `` duality '' : an integer @xmath166 if and only if the fractional part of @xmath167 is 0 or strictly greater than @xmath168 . to prove this simply observe that @xmath166 if there is an integer @xmath30 with @xmath169 , which we rearrange as @xmath170
. this happens exactly if the fractional part of @xmath171 is 0 or strictly greater than @xmath168 .
the second property is called `` balance '' : for all real numbers @xmath172 , the number @xmath173 is either @xmath174 or @xmath175 . to prove this ,
observe that we are counting the integers @xmath30 with @xmath176 , which is equivalent to @xmath177 .
since we only care about integral @xmath30 , we can write this as @xmath178 . clearly there are @xmath179 such @xmath30 , and this is @xmath180 where the @xmath181 s are both in @xmath182 .
it follows that the true value is an integer which is strictly less than 1 away from @xmath183 , i.e. , either @xmath174 or @xmath175 .
we will show that @xmath184\!\right]$}},\ ] ] but first we show that @xmath185 .
since @xmath186 counts the number of elements of @xmath187 in @xmath188 and @xmath189 counts those in @xmath190 $ ] , we need to show that @xmath191 and @xmath192 .
obviously @xmath193 , so the substance here is that @xmath194 . by duality , @xmath195 if and only if the fractional part of @xmath196 is 0 or strictly greater than @xmath168 . since @xmath197 , the fractional part of @xmath198 is @xmath199 , which is neither 0 nor strictly greater than @xmath168 ( using that @xmath40 are relatively prime )
. we now show by evaluating @xmath200 in two ways .
first , every @xmath201 contributes to @xmath202 .
thus @xmath203 second , by the balance property of beatty sets , we know that @xmath125 is either @xmath204 or @xmath205 , and in particular @xmath206 .
thus @xmath207 reducing and modulo @xmath7 tells us that @xmath208 . since @xmath185 , we know that @xmath209 , whence eq . .
theorem [ thm : maintheorem ] now follows from the straightforward calculations ( for @xmath210 ) @xmath211 and @xmath212\!\right]$ } } { \mbox{$\left[\!\left [ -\bar{p } < x - y \le 0 \right]\!\right]$ } } \\ & = \ , \sum_{y}{\mbox{$\left[\!\left [ y \in { { \mathcal b}}_{p } \right]\!\right]$ } } { \mbox{$\left[\!\left [ y\in [ x , x+\bar{p } ) \right]\!\right]$ } } \\ & = \
, \left|{{\mathcal b}}_{p } \cap [ x , x+\bar{p})\right| & = : r(x ) .
\end{aligned}\ ] ] there is another interesting way to finish the proof of theorem [ thm : maintheorem ] .
it plays on another expression of the `` balance '' property of beatty sets : for fixed @xmath213 , the difference @xmath214 is either @xmath215 or @xmath216 .
assume that @xmath210 , @xmath33 , and @xmath217 are relatively prime and positive .
let @xmath218 be the inverse of @xmath7 modulo @xmath6 , and let @xmath8 be the inverse of @xmath6 modulo @xmath7 .
we will use the identity @xmath219 set @xmath220 , and @xmath221 we compute @xmath222 in two ways .
first , the sum telescopes to @xmath223 second , note that @xmath224 is either @xmath225 or @xmath226 , say there are @xmath227 integers @xmath30 inclusively between 0 and @xmath228 with @xmath229 , and @xmath230 integers @xmath30 with @xmath231 .
we have @xmath232 equating these two evaluations modulo @xmath6 ( and using @xmath233 ) , we find that @xmath234 . by direct arithmetic ,
@xmath235 , and so for @xmath236 , @xmath237 .
we now use this information directly ( set @xmath238 ) : @xmath239 solving this equation yields @xmath240
by corollary [ cor : coveringcriterion ] , we can assume that @xmath87 ; we need only show that @xmath241\!\right]$ } } \,g_1\ , \frac{\omega^{-jr_1}}{1-\omega^{j \bar p_1}}\ ; + \;{\mbox{$\left[\!\left [ g_2 | j \right]\!\right]$ } } \,g_2 \,\frac{\omega^{-jr_2}}{1-\omega^{j \bar p_2 } } \label{equ:2pieces}\ ] ] is satisfied for @xmath242 if and only if @xmath243 .
we first assume that holds for all @xmath244 .
in particular , we set @xmath245 . since @xmath87 , we have @xmath246 , and simplifies to @xmath247 rearranging this gives @xmath248 since @xmath249 , @xmath250 , and @xmath251 , eq .
becomes @xmath252 whence @xmath253 .
we can read this argument from the bottom up to see the other half of `` if and only if '' .
[ lem : constructions ] if @xmath254 , @xmath255 , @xmath256 , and @xmath257 , then @xmath258 ( @xmath105 ) is a perfect covering .
we have @xmath259 so the first equation of the covering criterion is satisfied for some @xmath95 .
our hypotheses imply that @xmath260 and @xmath261 , so we need only to show that @xmath262 holds for @xmath263 any @xmath7-th root of unity except 1 . since @xmath264
, we can bring the summands in over a common denominator : @xmath265 and we see that it will suffice to show that @xmath266 is zero . but @xmath267 where @xmath268 consists of those integers whose binary expansions have the form @xmath269 with @xmath270 and @xmath271 . thus @xmath272 and since @xmath273 is a multiple of @xmath7 , this is zero .
[ thm : multiplicity ] suppose that @xmath274 ( @xmath105 ) satisfy conditions ( i ) and ( ii ) of the cfc , and that that @xmath100 for @xmath105 .
the @xmath106 sequences @xmath104 are a perfect @xmath95-fold covering , where @xmath95 is the number of ones in the binary expansion of @xmath275 .
this is equivalent to something by the covering criterion : the zeroth coefficient criterion becomes @xmath276 , and the non - zero coefficient part of the criterion , with an appropriate choice of @xmath15 and multiplying by @xmath277 , becomes equivalent to lemma [ lem : constructions ] .
thus , what remains to be proved is that @xmath278 .
we have @xmath279 so @xmath25 is definitely a multiple of @xmath7 .
note that @xmath280 defined by @xmath281 also satisfies @xmath282 .
if the binary expansion of @xmath283 is @xmath284 , then the binary expansion of @xmath280 is @xmath285 .
it follows that @xmath286 , where @xmath287 is the number of 1 s in the binary expansion of any of the @xmath280 , in particular , in the expansion of @xmath288 . on the other hand , @xmath289 .
consequently , @xmath290 .
according to fraenkel s conjecture , the values of @xmath7 for which there is a nontrivial perfect 1-covering by @xmath291 beatty sets are of the form @xmath273 . as a consequence of the preceding result
, we can identify those @xmath7 which allow for a nontrivial perfect 2-covering .
if @xmath7 is of the form @xmath292 , or @xmath293 for @xmath294 , then there is a perfect @xmath295-covering by at least three beatty sets with period @xmath7 and distinct densities . in this proof ,
we abbreviate @xmath296 as simply @xmath297 . from theorem [ thm :
multiplicity ] , we know that @xmath7 has a perfect @xmath295-covering if there is a @xmath298 with @xmath299 such that @xmath300 for some @xmath301 , where @xmath106 is the order of @xmath295 modulo @xmath7 .
let @xmath302 .
since the fractions @xmath303 are both reduced we see that @xmath304 using the elementary identity @xmath305 , we now have @xmath306 there are two possibilities : @xmath307 or @xmath308 .
in the first case , we find that @xmath309 and so @xmath310 .
it is clear that any value of @xmath311 will work here . in the second case
, we must have @xmath312 .
thus , in this case @xmath313 by hypothesis , i.e. , @xmath314 . hence , if we let @xmath315 then @xmath316 for some @xmath317 .
this implies that @xmath318 for some @xmath294 .
this completes the proof .
if our covering fraenkel conjecture is correct , then these are all the values of @xmath7 for which there exist non - trivial perfect @xmath295-coverings . besides the values of the form @xmath273 , the other values less than @xmath319 given by these expressions are : 5 , 21 , 51 , 85 , 341 , 455 , 819 , 1365 , 3855 , 5461 , 13107 , 21845 , 29127 , 31775 , 87381 , 209715 , 258111 , 349525 , and 986895 . the computer algebra systems
_ mathematica 5.0 _ and _ maple 7.0 _ will not automatically simplify the expressions ] to be 0 .
the authors are unaware of any source which documents the mathematical failures of these often - used rarely - cited closed - source programs . ]
@xmath320 @xmath321 to 0 , although they can be coaxed into verifying these identities by first replacing @xmath322 with @xmath323 .
but even this algebraification does not enable the ca systems to verify @xmath324 we will use the construction of perfect covers and theorem [ thm : maintheorem ] ( with some further manipulation ) to generate these and other trigonometric identities . if @xmath254 and @xmath255 , then ( from the proof of lemma [ lem : constructions ] ) @xmath325 thus , @xmath326 two of the identities above are given by this with @xmath327 and @xmath328 ( using @xmath329 ) .
we note that jager & lenstra @xcite showed that all linear dependencies over @xmath330 of @xmath331 have this form when @xmath7 is a prime .
let us take a closer look at our basic identity , which we can expand as follows : @xmath332 for any @xmath213 with @xmath333 .
thus , @xmath334 let @xmath335 denote the ( multi-)set @xmath336 .
thus , interchanging order of summation , we have @xmath337 we can rewrite the lhs summand as @xmath338 where we have taken @xmath339 thus , @xmath340 hence , we need to understand the sums @xmath341 to begin with , if @xmath342 then every term @xmath343 in @xmath25 will also have its conjugate @xmath344 in @xmath25 as well .
this happens exactly when the order of @xmath295 modulo @xmath7 is even , and in this case we have @xmath345 but for the trivial reason that every term in the sum occurs with its negative ! in general , each @xmath346 contains @xmath27 terms , so that @xmath347 is the sum of @xmath213 blocks of @xmath27 powers of @xmath263 .
if the union of these blocks is a perfect @xmath95-covering of @xmath348 then @xmath347 is just equal to @xmath349 .
for example , for @xmath350 , we have @xmath351 .
thus , @xmath352 so that @xmath353 another simple case where this happens is when @xmath354 is prime , 2 has order @xmath355 modulo @xmath7 , and 1 , 3 and 5 are in distinct @xmath356 . in this case , @xmath357 and @xmath358 are disjoint , so that their union is @xmath359 , which implies that @xmath360
. examples of this occur for @xmath361 , etc .
note that the real part of @xmath362 is equal to the real part of @xmath363 .
hence , if the @xmath364 form a complete set of disjoint @xmath365 , then in forming a perfect covering of @xmath359 , we an use either @xmath366 or @xmath367 interchangeably , if we only want to control the real part of @xmath347 .
for example , for @xmath368 , 2 has order 11 modulo 89 , and the complete set of disjoint @xmath365 is @xmath369 , @xmath370 , @xmath370 , @xmath371 , @xmath372 , @xmath373 , @xmath374 , and @xmath375 .
however , one can check that @xmath376 , @xmath377 , @xmath378 and @xmath379 .
thus , @xmath380 this implies ( as usual ) the unlikely identity @xmath381 using these ideas ( and other extensions thereof ) , many other results of this type can be derived but we will not pursue these here .
we assume in this section that the @xmath382 are distinct , @xmath100 , define , assume that @xmath383 , @xmath384 , that the @xmath106 sequences @xmath258 are a perfect @xmath95-fold covering , and that there is no pair @xmath385 with @xmath386 ( this is weaker than the hypotheses of the cfc ) .
let , and let @xmath30 be the multiplicity of @xmath387 in @xmath388 .
[ lem : few_gi ] if @xmath210 is a multiple of one of @xmath389 , then it is a multiple of at least three of them .
using this @xmath15 in the second displayed equation in corollary [ cor : coveringcriterion ] , we have @xmath390\!\right]$ } } \frac{\omega^{-jr_k}}{1-\omega^{j\bar p_k}}.\ ] ] clearly this sum can not have only one nonzero term .
suppose that it has exactly two , say @xmath391 and @xmath392 . if @xmath393 , then @xmath394 is a multiple of only one of @xmath395 , which can not happen ( set @xmath396 ) .
thus without loss of generality we may assume that @xmath397 .
we have @xmath398 multiply by @xmath399 and clear denominators to get ( setting @xmath400 ) @xmath401 if four complex numbers with the same modulus sum to 0 , then we can the split the four into two pairs , each of which sums to 0 .
our first case is @xmath402 , @xmath403 , which is the same as @xmath404 .
it follows that @xmath405 , and we multiply this congruence by @xmath406 ( a multiple of @xmath407 ) to get @xmath408 thus @xmath409 , and since @xmath100 , we actually have @xmath410 .
this contradicts our hypothesis that the @xmath382 are distinct .
our second case is @xmath411 , @xmath412 , which forces @xmath7 to be even and which is the equivalent to @xmath413 , @xmath414 .
it follows that @xmath415 , which we handled above .
our third case is @xmath416 , @xmath417 , and this is the same as @xmath418 , @xmath419 .
combining these gives @xmath420 .
multiply this equation by @xmath406 ( a multiple of @xmath407 ) to get @xmath421 whence @xmath422 . since @xmath100 , we actually have @xmath87 .
this contradicts our hypothesis that there is no pair of
s which sum to @xmath7 .
[ lem : q.bound ] if @xmath423 , then @xmath424 ; if @xmath425 , then @xmath426 ; if @xmath427 then @xmath428 ; if @xmath429 then @xmath430 , and in general @xmath431 set @xmath432 in corollary [ cor : coveringcriterion ] , and subtract @xmath433 from both sides , to get @xmath434\!\right]$}}\big ) \,{{\mbox{$\left[\!\left [ g_i|p_k \right]\!\right]$ } } \,g_i}\ , \frac{\omega^{-p_kr_i}}{1-\omega^{p_k\bar p_i}}.\ ] ] taking the absolute value of each side , and using the triangle inequality and the identity @xmath435 we get @xmath436\!\right]$ } } g_i}{|\sin(\pi p_k \bar p_i /q)|}.\ ] ] for all @xmath437 $ ] .
suppose that our numbering has @xmath438 for @xmath439 $ ] .
we have the inequalities for @xmath440 $ ] @xmath441 by replacing @xmath7 with @xmath442 , we can assume without loss of generality that @xmath443 ( the bound we find for @xmath7 will in truth be a bound for @xmath442 ) .
we wish to show that if @xmath7 is large enough , then the rhs must be small for some choice of @xmath116 .
let @xmath444 be the distance from @xmath24 to the nearest multiple of @xmath7 , and let @xmath445 satisfy @xmath446 . consider the directed graph with vertices @xmath447 , with an edge from @xmath448 to @xmath449 if @xmath450 . every finite directed graph contains either a sink ( a point with no out - edges ) or a cycle .
if @xmath451 is a cycle , then @xmath452 and @xmath453 therefore , @xmath454 . if @xmath382 is a sink , then all ( @xmath455 ) of @xmath456 are strictly greater than @xmath445 .
since @xmath457 are distinct and there is no solution to @xmath386 , the @xmath30 values @xmath458 are also distinct .
thus , @xmath459 using the approximation @xmath460 for small @xmath24 , one sees that bounds @xmath7 above .
more precisely , if the graph has a cycle , then @xmath461 , and otherwise @xmath7 is bounded by . for @xmath462
, we calculate that @xmath463 , and consequently @xmath7 is at most @xmath464 , @xmath465 , @xmath466 , @xmath467 , respectively . to prove the `` in general '' statement , we need only work with @xmath468 .
let @xmath469 .
we define the graph as above , and handle a cycle in the same way . if @xmath382 is a sink , then becomes @xmath470 where we have used the inequalities @xmath471 for @xmath472 .
we note that @xmath473 so that our inequality can be weakened to read @xmath474 this is inconsistent for any @xmath475 , @xmath476 and @xmath477 .
[ lem : g=1 ] if @xmath114 , then @xmath443 .
renumber so that @xmath478 . if all the @xmath479 are equal , then @xmath443 since @xmath480 .
thus we may assume that @xmath481 . by lemma [ lem : few_gi ] with @xmath482 , @xmath483 .
since @xmath481 , we may assume that @xmath484 , so that @xmath485
. clearly @xmath387 divides @xmath486 by definition , and @xmath487 by lemma [ lem : few_gi ] ( with @xmath488 and @xmath489 , respectively ) . since @xmath480
, we have shown that @xmath443 .
the cfc is true for @xmath114 .
first , observe that by lemma [ lem : g=1 ] we may restrict our attention to sequences @xmath258 with @xmath490 relatively prime to @xmath7 .
moreover , as a consequence of corollary [ cor : coveringcriterion ] , the sequences @xmath491 are a perfect covering if and only the sequences @xmath492 are a perfect covering .
thus , we may assume that @xmath493 and @xmath494 .
corollary [ cor : coveringcriterion ] also tells us that @xmath495 , and with @xmath496 that @xmath497\!\right]$ } } \frac{\omega^{-r_k}}{1-\omega^{\bar p_k}}.\ ] ] this implies that @xmath498 in this expression there are @xmath499 degrees of freedom : @xmath7 , @xmath500 , @xmath501 , @xmath502 , @xmath503 . without loss of generality @xmath504 . also , the hypotheses of the cfc imply that there is no solution to @xmath386 , and lemma [ lem : q.bound ] ( with lemma [ lem : g=1 ] ) implies @xmath505 .
there are only 346 @xmath106-tuples @xmath506 with @xmath114 , @xmath507 @xmath508 , no solution to @xmath386 , and satisfying .
we refine our search by noting that if @xmath509 , then @xmath510 must also be on our list of 346 ( this is equivalent to taking values of @xmath15 other than 1 in deriving the inequality ) .
this pares the list down to a single tuple for @xmath511 , a single tuple for @xmath512 , and 10 tuples for @xmath513 .
the tuples predicted by the cfc are on these lists , and the remaining 9 tuples are eliminated by an exhaustive search for @xmath514 such that @xmath515 the only perfect coverings with 5 or fewer beatty sequences are those predicted by the covering fraenkel conjecture .
we remark that we may similarly reduce the @xmath516 case of fraenkel s conjecture ( but not the cfc ) to a finite computation .
for example , if @xmath517 , then we may argue as in the @xmath513 case of lemma [ lem : g=1 ] that @xmath443 , and so by lemma [ lem : q.bound ] we get the bound @xmath518 . if @xmath519 ( so that @xmath424 ) , then we may ( using lemma [ lem : few_gi ] ) renumber so that @xmath520 and @xmath521 , with @xmath522 . since the @xmath382 are distinct , we have @xmath523 and @xmath524
. thus @xmath525 , a contradiction .
in contrast , for @xmath526 , we arrive at the consistent inequalities @xmath426 and @xmath527 .
we envision a non - computational proof of fraenkel s conjecture along the following lines .
suppose that @xmath104 ( @xmath105 ) partition @xmath86 , with @xmath528 , and suppose that @xmath106 is minimal .
let @xmath529 .
now , suppose that is larger than 1 .
the covering criterion with @xmath482 yields @xmath530 let @xmath531 be those @xmath532 for which @xmath533 ( there are @xmath534 of them by lemma [ lem : few_gi ] , and @xmath481 since @xmath480 ) , and let @xmath535 be the negatives of the @xmath536 for which @xmath533 . also replace @xmath145 with @xmath537 , and we get @xmath538 this seems to imply that the sum vanishes for @xmath24 _ any _ @xmath442-th root of unity , which would imply that these sequences alone form a perfect covering , whence @xmath539\!\right]$ } } \ge q$ ] .
this is impossible since @xmath481 .
thus , the following conjecture implies that , in the present setting , @xmath443 .
[ cnj : rationalfunction ] suppose that @xmath540 , with @xmath541 for all @xmath116 , and let @xmath542 be arbitrary integers . if the function @xmath543 vanishes at @xmath544 , then @xmath545 .
this in turn simplifies the covering criterion substantially .
considering the absolute value of the covering criterion as in the proof of lemma [ lem : q.bound ] we found that for @xmath546 ( roughly ) , @xmath116 could be chosen to make a particular inequality invalid .
conjecture [ cnj : strongmartin ] posits that the @xmath547 bound can be improved to @xmath548 , and barring the single exception of @xmath549 , it can be improved to @xmath550 .
this is supported by computational investigations . ]
the following conjecture becomes relevant .
[ cnj : strongmartin ] suppose that @xmath551 are distinct and relatively prime to @xmath552 , with @xmath553 , and for each @xmath440 $ ] @xmath554 then @xmath549 and @xmath555 .
at this point , we would have shown that a counterexample to fraenkel s conjecture ( with @xmath106 sequences ) must have @xmath556 .
we envision handling this situation combinatorially , probably in conjunction with tijdeman s combinatorial restrictions .
he @xcite*lemma 4 notes that if @xmath258 and @xmath557 are disjoint , then either @xmath558 or @xmath559 , and so @xmath560 also , his main lemma @xcite*lemma 3 can be strengthened ( using the same proof , but in terms of beatty sequences instead of balanced sequences ) to provide the powerful restriction on the @xmath382 s in a counterexample with minimal @xmath106 : @xmath561 tijdeman used these two lemmas ( and some casework ) to show that @xmath562 .
thus , the remaining situation has many sequences with small ( but spread out ) @xmath382 s and quite small @xmath7 .
ideally , one would like arithmetic conditions on @xmath563 for the sequences @xmath564 to be a perfect covering , without assuming that the @xmath78 are distinct .
morikawa has given such conditions for a small number of sequences to be a perfect 1-cover , see @xcite for a brief description of morikawa s work and citations for his many papers on the topic .
the covering fraenkel conjecture that we have advanced is another step in this direction . a more ambitious step would be to replace the condition `` with no proper subset
@xmath101 $ ] having @xmath102 '' condition with the weaker condition `` with no proper subset of the sequences being a perfect covering . ''
we note that conjecture [ cnj : rationalfunction ] can likely be strengthened : suppose that @xmath540 , with @xmath541 for all @xmath116 , and with no subset of the @xmath565 s summing to a multiple of @xmath7 , and let @xmath542 be arbitrary integers .
if the function @xmath543 vanishes at @xmath544 , then it vanishes at all @xmath7-th roots of unity except @xmath566 .
joe buhler notes that @xmath567 vanishes at primitive 15-th roots of unity , but not at the primitive 5-th roots of unity , and thus the condition on sums of subsets of the @xmath565 s is necessary .
the authors wish to thank the national science foundation for supporting mathematics in general and the authors in particular .
we thank professor greg martin of the university of british columbia for supplying the main idea underlying lemma [ lem : q.bound ] . | the set @xmath0 ( with integers @xmath1 ) is a beatty set with density @xmath2 .
we derive a formula for the fourier transform @xmath3 a. s. fraenkel conjectured that there is essentially one way to partition the integers into @xmath4 beatty sets with distinct densities .
we conjecture a generalization of this , and use fourier methods to prove several special cases of our generalized conjecture . * ams mathematics subject classification * : 42a16 , 11b50 , 11l99 * keywords * : beatty set , discrete fourier transform , fraenkel s conjecture , perfect covering |
let us start by stating the result and a few consequences .
let @xmath0 be an algebraic variety which is complete over a field @xmath1 , and let @xmath2 be any effective cartier divisor on @xmath3 we denote by @xmath4 the line bundle associated to @xmath5 with @xmath6 it is a very interesting problem to know if the ring @xmath7 is a finitely generated @xmath1-algebra .
it was o. zariski who worked out the case of algebraic surfaces in order to solve the riemann - roch problem , and gave examples where @xmath8 is not finitely generated , see @xcite . in this article
we prove that the ring @xmath8 is finitely generated when @xmath0 is a complete toric variety ( perhaps singular ) .
the ring @xmath8 appears in many interesting problems , for example , if @xmath0 is a nonsingular projective variety , then the finite generation of @xmath8 implies that the series @xmath9 is rational .
it was asked by s. d. cutkosky and v. srinivas @xcite if this series is rational when @xmath2 is a nef divisor .
our result gives trivially a positive answer in the case of a complete toric variety , and @xmath2 any effective cartier divisor .
furthermore , the rationality of the series also allows us to compute , at least theoretically , the dimension of @xmath10 in terms of @xmath11 .
this is the riemann - roch problem , see for example @xcite and @xcite .
observe that these dimensions are not given by the riemann - roch theorem since the variety @xmath0 can be singular , and the line bundle @xmath4 is not necessarily generated by its global sections .
we would like to make a final remark .
the ring @xmath8 is a subring of the ring @xmath12 , which is finitely generated in the case of toric varieties .
in fact , d. cox proves in @xcite that the ring @xmath13 is a polynomial ring graded by the the monoid of effective divisors classes in the chow group @xmath14 of @xmath0 , where @xmath15 is the dimension of @xmath0 .
in this section we prove that the ring @xmath8 , which was defined by equation ( [ ring ] ) , is a finitely generated @xmath1-algebra . throughout this section @xmath0
means a complete toric variety over a field @xmath16 and @xmath2 a cartier divisor in @xmath3 the construction of the cone @xmath17 , in the proof of the theorem , is a well know construction , see for example @xcite .
we start by recalling an important lemma that will be use in the proof of the theorem .
we first observe that we can consider our divisor @xmath2 to be @xmath22-invariant . for the following sequence
is exact , see @xcite @xmath23 where @xmath24 is the group of @xmath22-invariant cartier divisors and @xmath25 is the dual lattice of @xmath26
. then we can write @xmath2 as @xmath27 with @xmath28 the set of invariant divisors .
we consider the convex rational polyhedron @xmath29 in @xmath30 defined as @xmath31 where @xmath32 is the the first element in the lattice appearing in the divisor @xmath33 .
we know that generators for the space @xmath34 are given by the elements of @xmath35 , which is a finite set because @xmath0 is a complete variety , this also implies that @xmath36 is a rational convex polytope .
let us embed the k - dimensional @xmath37-vector space @xmath38 into the ( k+1)-dimensional @xmath37-vector space @xmath39 as the hyperplane with equation @xmath40 , where @xmath41 are coordinates for @xmath42 .
denote by @xmath17 the ( @xmath43)-dimensional cone in @xmath42 generated by the rays starting at the origin and passing through the vertices of @xmath36 ( see figure below ) .
it follows from the definition that @xmath44 , and this implies that the intersection of the hyperplane @xmath45 with the cone @xmath17 is just the polyhedron @xmath29 .
in other words , the cone @xmath17 is the cone associated to the ring @xmath8 , in the sense that any element of the ring @xmath8 is a finite linear combination of integral points of @xmath46 the theorem follows since the semigroup @xmath47 is finitely generated by lemma [ gordan ] ( gordan ) .
i thank very much v. srinivas for suggesting me this problem and for many wonderful and fruitful conversations .
i also would like to thank a. king for bringing to my attention the article of d. cox , and to the referee for pointing out the article of v. batyrev . | in this article we prove that for any complete toric variety , and for any cartier divisor , the ring of global sections of multiples of the line bundle associated to the divisor is finitely generated . |
with the rapid growth of mobile demands and the ever - increasing use of smart phones , mobile network has been one of the fastest growing technologies that impact major aspects of our life @xcite . in recent years ,
the 4 g mobile communication system is being deployed worldwide , leading to a rapid growth in the mobile network capacity , which further dramatically stimulates the mobile demands . as mobile network evolves from voice - oriented to media - oriented@xcite , incremental improvements of networks
can no longer keep the pace with the increase in mobile data demands , since there exist several challenges that the current mobile network can hardly address . _
1 ) convergence of heterogeneous wireless networks_. there exist various heterogeneous wireless networks , i.e. pluralistic standards such as gsm , umts , lte and wlan , and they will coexist for a long time @xcite . however , as current wireless operators typically deploy their networks by vertically constructed method , these heterogeneous networks can hardly interconnect with each other , which makes operators incapable of efficiently optimizing and dynamically scheduling from the global perspective .
for example , although there may be many mobile networks around us , we may not be able to access the most appropriate one or select multiple networks simultaneously .
the only thing we can do is to access one specific network all the time , even if this network performs quite poorly .
_ 2 ) efficient utilization of wireless resources_. as a result of the difficulty in achieving the convergence of heterogeneous networks , many network devices are not fully utilized and plenty of wireless resources are wasted , while generating enormous amounts of energy consumption .
while several big carriers , including at&t , verizon , t - mobile and sprint , say that in the next few years they may not have enough spectrum to meet the demands for mobile data , many scientists and engineers , including martin cooper , father of the mobile phone , are convinced that the main reason for spectrum crisis is that the spectrum resources are not fully utilized @xcite .
_ 3 ) network innovations_. proliferating mobile demands compel wireless operators to build sustained innovations to continuously boost their network performances .
however , tightly coupling with specific hardware and lack of flexible control interfaces , the current mobile network can hardly provide a fast track for technological innovations .
_ 4 ) wireless applications and services_. wireless services proliferate significantly , e.g. big data @xcite , and different kinds of services require very different network characteristics . unfortunately , the current mobile network often can only support these wide - ranging and very different services with the same network characteristics and , consequently , it may provide users with poor quality of service ( qos ) and poor quality of experience ( qoe ) .
because the current mobile network is vertically constructed over the years , it generally becomes ossified and closed , and has difficulty to meet the above mentioned challenges .
this calls for the research and deployment of new generation of mobile network , specifically , 5 g , which is capable of providing a more open and efficient mobile network .
in particular , 5 g should benefit from the convergence of heterogeneous networks , facilitate network evolution , boost the network performance , enhance users qos and qoe , and at the same time save resources and energy .
these objectives require mobile networks to offer much more efficient controlling and user data processing functions across the network layer down to physical layer by jointly considering the both layers together .
however , since the physical layer technologies in mobile networks are much more complicated than in the wired scenarios , implementing efficient control and programmable user planes across the network layer and the physical layer is extremely challenging .
the problem gets worse as the mobile environment becomes more complex , e.g. high mobility , frequent handovers , and heavy interferences .
consequently , the key technologies of 5 g system are still open and attract increasingly attentions @xcite . software defined network ( sdn ) , an innovative paradigm , advocates separating the control plane and data plane , and abstracting the control functions of the network into a logically centralized control plane , referred to as sdn controller @xcite .
aiming to establish a new paradigm for 5 g mobile network , in this article , we propose a cross - layer software - defined 5 g network architecture . in our architecture , the concept of sdn controller is extended to take the control functions of the physical layer into consideration as well , not just those of the network layer .
furthermore , we also introduce the novel cloud computing pool which considers the programmable user data processing functions of both the network layer and the physical layer jointly . specifically , by abstracting the control functions of the both layers , we introduce a logically centralized programmable control plane oriented towards both network and physical layers , through which we achieve the fine - grained controlling and flexible programmability of the both layers .
for example , we can control the packets forwarding in network layer as well as direct the beamforming and interference canceling in physical layer .
the cloud computing pool proposed in our architecture provides powerful computing capability to implement the baseband data processing of multiple heterogeneous networks , which efficiently improves the convergence of heterogeneous wireless networks .
moreover , the programmable scheme in our architecture flexibly supports the network evolution and the deployment of any potential technological innovations which is particularly important as the key technologies of network and physical layers of 5 g .
we also discuss the main technical challenges of our architecture , including the fine - grained control strategies oriented to the both network and physical layers , network virtualization that supports the customizability , and programmability that is beneficial to network evolution .
qualitative simulations are deployed to validate the advantages of our proposal .
our study demonstrates that our proposed 5 g architecture is capable of achieving its design goals , namely , convergence of heterogeneous networks , fine - grained controllability , efficient programmability for network evolution , and network and service customizability , and therefore it offers an architecture design of future mobile network .
the remainder of this article is organized as follows .
we first offer a rethinking of mobile network , and this is followed by the introduction to sdn .
next , we present our cross - layer software - defined 5 g architecture and analyze its features and system feasibility
. then we interpret the key technologies and challenges of our architecture .
several quantitative performance results are provided .
finally , we conclude the article by summarizing our contribution .
the challenges to the current mobile network are mostly caused by the ossification , closeness , and vertically constructed features . simply evolving the traditional networking architecture
can hardly meet the requirements of future mobile networks .
we rethink the problems of the current mobile networking and envisage how the future mobile network looks like , and consequently propose the design goals of our architecture .
* _ ossified _ _ vs _ _
controllable_*. controlling in the current mobile network is limited and ossified .
it is quite difficult to dynamically adjust the control strategies when the network status changes quickly .
for example , many users with high mobility may go through numerous and frequent handovers , especially when the handovers occur among heterogeneous networks , which needs more fine - grained and dynamical controlling .
therefore , efficiently controlling is a fundamental function that the future network has to provide . *
_ closed _ _ vs _ _ programmable_*. technological innovations in mobile network need to be continuously developed in order to meet the demands .
unfortunately , mobile network technologies are usually solidified in the hardware .
for example , the unequal error protection ( uep ) , a physical layer technology , can provide higher video quality while consuming less spectrum .
but to make such performance enhancing changes typically require replacing the hardware due to the tightly coupling with the hardware in the form of specific asics designed for each protocol of the lte network @xcite .
in particular , dynamically deploying specific technologies based on the run - time network status becomes especially difficult .
therefore , it is necessary and highly desired to introduce greater programmability into mobile networks . *
_ vertical _ _ vs _ _ converging_*. the vertical - constructed method has dominated the mobile networking almost since the first generation mobile communication system was conceived , and it has brought great successes to mobile communication . however , as the scale of mobile network is quickly enlarging and the network becomes ever - increasingly dense , traditional vertical - constructed networking severely restricts the development of mobile networks .
for example , 3 g , lte and wlan can hardly interconnect , and the operators are incapable of controlling the network from the global perspective .
thus , enabling efficiently convergence of heterogeneous networks must be supported in the future mobile network .
* _ unified _ _ vs _ _ evolvable_*. multiple mobile protocols and pluralistic standards will coexist for a long time @xcite .
each of them possesses the distinctive advantage for specific services or situations . for example
, 3 g provides better voice services and wlan suits for video applications .
the diversified world stimulates pluralistic demands , and thus it is almost impossible to find one unified network with solidified technologies to fullfill all the expectations .
consequently , the future mobile network must support smooth evolution for pluralistic standards .
based on the above analysis , we can envisage the basic characteristics of the future mobile network and naturally define the _ * design goals * _ of our proposed 5 g architecture , which are summarized as follows .
* beneficial to the convergence of heterogeneous networks ; * fine - grained controllability for mobile networks ; * high programmability and supply open interfaces for network technological innovations ; * high evolvability and fast deployment ; * customizability of networks and services .
sdn has become one of the hottest topics in computer networking . by separating the control plane and the data plane
, sdn abstracts the control functions of the network layer into a logically centralized control plane , known as the sdn controller @xcite .
this sdn controller makes the rules and control the behaviors of data plane devices from the global perspective .
openflow @xcite is a most common realization of sdn .
many manufacturers , such as hp , nec , ibm , produce the commercial openflow switch , and several kinds openflow controllers , e.g. nox and floodlight , are available .
the openflow protocol is sustainably released by open networking foundation ( onf ) , and sdn is becoming ever - increasingly popular in networking today . [ cols=">,<,>,<,>,<,>,<,>,<",options="header " , ] in recent years , several researchers attempt to extend sdn into wireless and mobile networks .
openradio @xcite mostly focuses on the programmable data plane design of software - defined wireless network , but this work does not propose the holistic network structure with the control plane architecture .
@xcite present a software - defined cellular network architecture , cellsdn , that covers both access network and core network . however , this work is oriented to the lte network , and it can hardly be adapted to the circumstances of heterogeneous networks .
softran @xcite is a software - defined control plane of the radio access network by introducing a virtual big - base station , but this architecture mostly focuses on the one - tier situation , consisting of micro - cells , which is not suitable for the scenarios of heterogeneous networks .
openrf @xcite and softcell @xcite focus on the software defined wifi network and core network respectively
. meanwhile , mobiflow @xcite and openroad @xcite focus on the transport network that is similar to wired network and offer no special support for mobile access network .
moreover , almost all the existing works , except for openradio and openrf , neglect the vital programmability features .
table [ tab1 ] compares our proposed architecture with these existing works .
in order to address the challenges and to achieve the goals introduced in the previous sections , we propose a cross - layer software - defined 5 g architecture , as depicted in fig . [
fig : architecture ] .
we introduce a logically centralized control plane , which makes the rules and controls the behaviors of network devices as well as provides appropriate application program interfaces ( apis ) for various applications . to achieve the efficient controlling of both the network layer and physical layer , we design a fine - grained control strategy
which is compatible with the both layers .
because the limitation of poor computing capability in baseband data processing and the lack of interconnection severely impede the traditionally designed physical layer s controllability , we introduce a novel cloud based baseband data processing pool .
the radio transceiver devices and the baseband data processing devices are completely separated .
we also introduce the programmability feature to make the network more evolvable and flexible for technological innovations .
our architecture contains four main parts : wireless access layer , cloud computing resource pool , centralized control plane , and application plane .
wireless access layer consists of a large number of physical remote radio units ( prrus ) distributed at various locations .
it enables several virtual rrus ( vrrus ) corresponding to different mobile communication protocols coexisting in one shared prru by radio frequency virtualization technology , to efficiently support fast deployment and the convergence of heterogeneous networks .
for example , one prru can fast deploy one 3gpp vrru and one wlan vrru ( vap ) simultaneously , according to the requirements .
cloud computing resource pool is comprised of a large amount of physical processors with high - performance computing and storage capability as well as high speed links . the current mobile network adopts the vertically constructed networking , in which one rru is tightly coupled with one specific base band unit ( bbu ) . to loosen this inefficient coupling , in our architecture , the traditional bbus ,
base station controllers ( bscs ) and routers are replaced by virtual ones ( vbbus , vbscs , and vrouters ) which are implemented in the shared physical processors through virtualization .
these shared physical processors can simultaneously run multiple vbbus , vbscs , and vrouters by allocating appropriate computing resources and storage resources . specifically , each vbbu can deploy appropriate wireless protocol softwares to implement the physical layer programmability .
the corresponding virtual elements ( vbscs , vbbus , vrouters , and vrrus ) constitute a complete wireless network with specific protocols .
our cloud based data processing pool significantly improves the computing capacity of baseband processing and overcomes the interconnection difficulty of traditional mobile networks , which greatly benefits the convergence of heterogeneous networks . to guarantee the run - time properties ,
the wireless access pool is connected to the cloud computing resource pool through high speed optical links .
the wireless access layer and the cloud computing resource pool comprise the data plane of our architecture .
the centralized control plane is the `` brain '' of our architecture by abstracting and combining the control functions of both the network and physical layers .
the control plane can acquire the configurations and real - time status of the both layers to make the decision from the global perspective .
we partition the control plane into several functional submodules . _
1 ) flow scheduler _ , which is related to the network layer , is similar to the wired sdn controller .
it makes the polices and schedules the behaviors of each virtual forwarding element , i.e. guiding them to process the packets .
_ 2 ) wireless scheduler _ , which is oriented to the physical layer , is the `` highlight '' in our architecture .
it can deploy appropriate wireless protocol softwares into each vbbu .
moreover , it also dynamically schedules the physical layer actions , for example , beamforming , power control , interference cancelling , massive mimo and carrier aggregation .
_ 3 ) virtualization _ , which is an indispensable element of sdn , makes it possible to customize networks and services .
specifically , virtualization scheduler manages the strategies of virtualization .
we introduce three kinds of virtualization flow space virtualization , cloud level virtualization , and spectrum level virtualization that conspicuously improve the network performance from different angles .
_ 4 ) resource scheduler_. because there are various resources in our architecture , including spectrum , computing , and storage , it is necessary to efficiently allocate these resources to different virtual elements , which is implemented in resource scheduler .
_ 5 ) global - view database_. all the above - mentioned submodules need to know the static configurations and dynamic status of the whole network and , therefore we introduce a database with global view to collect these information .
the application plane contains a variety of network applications , each of which utilizes the apis abstracted by the control plane to guarantee its qos and qoe .
we now emphasize that our proposed cross - layer software - defined 5 g architecture achieves the design goals set out in section [ s2 ] .
firstly , we abandon the traditional vertically constructed networking method and introduce a cloud computing pool to process the baseband data of heterogeneous networks
. therefore , our architecture conspicuously benefits the convergence of heterogeneous networks . with our architecture
, the mobile network can easily acquire real - time global network view , which enables its centralized control plane to efficiently control the entire network by considering both the network and physical layers as a whole .
this makes the network much more open and controllable .
the flexible programmability proposed not only provides the opportunities for fast deployment of technological innovations but also makes the network more evolvable .
our proposed three types of virtualization further enhance the customizability of the networks and services .
recent years , sdn and network virtualization attract increasingly attentions from both the academic and industrial communities .
many prototypes , testbeds and products have been developed and produced , in particular , openflow controller , switch and protocols , as well as geni network virtualization platform .
furthermore , data center network and cloud computing also stimulate fast development of sdn and virtualization technologies .
moreover , there are plenty of prototypes and testbeds for software defined radio ( sdr ) @xcite as well as its extended version , cognitive radio @xcite , which offer considerable programmability .
we also note that china mobile ltd .
positively focuses on the research and development of the cloud based mobile network architecture , referred to as cran @xcite . based on these progresses
, we believe that our proposed architecture is quite feasible .
the first technical challenge is to design the fine - grained control strategy , and we divide the problem into two subproblems : jointly optimizing strategy and individual control strategies . the jointly optimizing strategy makes the decision to achieve the efficient controlling according to the static configurations and real - time status of both the network and physical layers .
consider for example that the qos for a user at cell boundary deteriorates .
we may need to determine the appropriate routing and bandwidth requirements as well as to schedule the beamforming and interference cancelling .
after the decision has been made , the control plane decomposes the decision into two parts corresponding to the two layers , and then leverages the flow based `` match - action '' strategy for the both layers to make rules and control behaviors of each virtual element , as illustrated in fig .
[ fig : control ] .
correspondingly , each data plane device integrates a sdn agent , which complies with the control strategy , resolves the control flow , and communicates with the control plane .
virtual elements constantly report the real - time status to the control plane , which may then dynamically refresh the global - view database . since the responsibilities of the network layer and physical layer are distinctly different , the individual control strategies for the two layers are different , which defines the second subproblem of our fine - grained control strategy . _
network layer_. since the packet header of each flow possesses several function fields called _ match fields _ , such as ip address and mac address , the data plane devices can process the arriving packets by checking the match fields . specifically
, when a virtual element receives a packet , it first checks whether this flow matches its control rules .
if so , it executes the corresponding actions .
otherwise , the packets will be dropped or sent to the control plane .
an example can be interpreted as : * if * dest ip = xx.xx.xx.xx , * then * forward to vbbu1 with speed s1 .
_ physical layer_. although the physical layer controlling also adopts the `` match - action '' strategy , the match fields , rules , and actions are quite different from the network layer
. since key technologies of 5 g physical layer are still open , the format of match fields should be elastic .
for example , the basic information filed should contain the encoding mode , modulation type , carrier information , _
etc_. additionally , the power control field may present the power control information , while the beamforming field and the interference cancelling field should provide some corresponding information .
the control plane dynamically makes the decision according to the run - time status , and then sends control flows to the vbbus .
the `` actions '' in vbbus refers to the appropriate physical layer technologies .
for example , when qos of a user deteriorates , the control plane may launch beamforming , while when the users at cell boundaries encounter severe inter - cell interference , the control plane may start interference cancelling .
network virtualization as a promising means of the future network enables multiple concurrent virtual networks to run on the shared substrate resources @xcite .
network virtualization is an indispensable part of sdn , while sdn architecture facilitates network virtualization .
network virtualization is responsible for customizing the networks and services .
therefore , how the virtualization scheme is designed directly influences the system performance and resource utilization .
there are three types of virtualization , flow space virtualization , cloud level virtualization , and spectrum level virtualization , in our architecture , and we now elaborate each of them . _
flow space virtualization_. different operators or network services require different network characteristics . as the control strategy in our architecture
is flow based , flow space can be divided into multiple slices , each of which corresponds to one virtual network . as shown in fig .
[ fig : virtualization ] , we introduce a flow space abstracting layer . by abstracting the control functions of virtual elements ,
this layer slices the flow space according to the requirements of operators or services , and provides each flow slice with one controller
. then virtual networks are generated , and each virtual network is managed by its corresponding controller .
the virtual networks share the same underlying data plane devices . in vbsc or vrouter ,
different virtual networks have different rules and run different actions .
operators often need different vbbus due to the specific wireless protocols .
_ cloud level virtualization_. the control plane first creates vbbus , vbscs and vrouters by virtualizing physical processors and allocating appropriate computing and storage resources .
then it establishes the forwarding and data processing rules for virtual elements .
moreover , it deploys the corresponding wireless protocol softwares in vbbus by utilizing the programmability . _ spectrum level virtualization_. it refers to the virtualization of spectrum by radio frequency virtualization technology , which enables several vrrus with different wireless protocols to coexist in one shared prru .
spectrum level virtualization extends the virtualization scope , lessens the networking difficulty and saves the cost .
the physical layer programmability faces several challenges including how to guarantee sufficient computing capability and how to efficiently implement physical layer technologies .
our proposed cloud based baseband processing pool provides powerful computing capability . in the physical layer , although different wireless protocols operate quite different from each other , they always share some same modules , such as modulation , coding and interleaving . inspired from sdr , we modularize wireless protocols and establish a wireless module library in vbbus .
vbbus can select different modules , and then connect these selected modules to implement the specific protocol . as 5 g physical layer technologies are still open , our proposed programmability provides the flexibility for technological innovations , benefits the future network evolution , as well as enables fast deployment and customization of wireless protocols .
although in the current stage the architecture proposed has not been deployed in a real scenario , we try to evaluate the performance with real traces from a large cellular deployment .
we show that our proposed architecture significantly improves the energy efficiency and qos utility .
to truly reflect the performance advantages , we use the real datasets named `` sitefinder '' which is released by ofcom on behalf of uk s government@xcite . these datasets record massive base stations information supplied by the mobile network operators .
the information contains location information ( latitude and longitude ) , operator information , base station type , transmission power , communication protocol ( gsm , umts , and tetra ) , frequency band , antenna information , and etc . considering that the geographical area covered by these datasets is too wide , we select @xmath0 geographic area from manchester .
consequently , there are 159 gsm base stations as well as 255 umts base stations deployed by five operators i.e. o2 , orange , three , t - mobile , and vodafone . in recent years , many researchers focus on the energy efficiency in mobile communications and , as a result , lots of energy - efficient mechanisms are presented
. however , a form of stalemate exists : these efficient mechanisms are difficult to implement and deploy in real scenarios since the network functions are highly solidified in the hardware and the network devices are inflexible to control .
meanwhile , our proposed architecture provides flexible programmability and centralized controlling , which makes it easy to implement these energy - efficient schemes in real scenarios . considering that network traffic at midnight is low , i.e. tide effect ,
we propose two energy - efficient strategies : 1 ) the centralized control plane intelligently turn off the base stations on behalf of individual operator , referred to as iit strategy ; 2 ) control plane turn off the base stations from global perspective , i.e. cooperation cross operators , referred to as cit strategy .
it is notable that both of these two strategies needs to guarantee the whole area coverage .
[ fig : energy ] depict the energy - saving performance of gsm and umts . as coverage range of each base station increases ,
both strategies turn more base stations and save more energy .
for example , when coverage range is @xmath1 , 39 percent base stations can be turned off and 35 percent energy can be save by utilizing iit strategy for umts .
moreover , as figures shown , apparently , the cit strategy achieves much more energy - efficient benefits than iit since cit allows the control plane to optimize the energy - efficient solution across multiple mobile operators . in the current mobile network , every end user has no choice but to access one specific network
. they may not be able to access the most appropriate one or select multiple networks simultaneously , which leads to poor qos .
in contrast , in our architecture , all the network devices are programmable and controlled by the centralized control plane .
consequently , the control plane may guide the end users to access the most appropriate networks , by which qos can be improved .
we assume that there are 10,000 end users and each one runs an ongoing mobile service .
we define the qos utility between each pair of @xmath2end user , base station@xmath3 , and assume the utility follows the uniform distribution in the range of @xmath4 $ ] by normalizing .
the location ( latitude and longitude ) of each end user is randomly generated .
we introduce four types of strategies : 1 ) current strategy .
each business is bound to one specific network .
2 ) intra operator optimizing ( intra - opr ) : the control plane guides each end user to access the base station , belonging to the specific operator , with the largest qos utility .
3 ) inter operator optimizing ( inter - opr ) : breaking the limit of one specific operator .
4 ) breaking the limit of operator as well as one communication standard ( heter - stra ) , i.e. it allows the end users to the access the base station among several heterogeneous networks ( in this senario : cross umts and gsm ) . as fig .
[ fig : qos](a ) shows , both intra - opr and inter - opr obtain much more qos utility than current strategy , and this advantage becomes much more obvious as the coverage range increases .
after that , we consider the heterogeneous scenario and combine @xmath5 base stations .
[ fig : qos](b ) confirms the perform advantages of heter - stra .
in this article , we have proposed a cross - layer software - defined 5 g architecture , which addresses some of the key technical challenges facing the future mobile network .
our architecture is able to control the mobile network from the global perspective through the fine - grained controlling for the both network and physical layers .
moreover , the cloud computing pool in our architecture provides powerful computing capability to implement the baseband data processing of multiple heterogeneous networks , while the programmability feature efficiently supports the network evolution and technological innovations .
the key technologies we introduced to our architecture include fine - grained control strategy , network virtualization , and programmability , and the key features of our proposed 5 g architecture are controllability , programmability , evolvability and customizability . since the key technologies of 5 g are still open , we believe our flexible architecture is capable of supporting the 5 g standards that eventually emerge and efficiently satisfying the requirements of mobile demands .
chen b x. ( 2012 ) carriers warn of crisis in mobile spectrum .
the new york times .
accessed 17 april 2012 bansal m , mehlman j , katti s , levis p ( 2012 ) openradio : a programmable wireless dataplane . in : proceedings of the first workshop on hot topics in software defined networks , ser .
hotsdn 12 , new york , usa , pp 109 - 114 gudipati a , perry d , li l e , katti s ( 2013 ) softran : software defined radio access network . in : proceedings of the second acm sigcomm workshop on hot topics in software defined networking , ser .
hotsdn 13 , new york , usa , pp 25 - 30 jin x , li l e , vanbever l , rexford j ( 2013 ) softcell : scalable and flexible cellular core network architecture . in : proceedings of the ninth acm conference on emerging networking experiments and technologies , ser .
conext 13 , new york , usa , pp 163 - 174 yap k k , sherwood r , kobayashi m , huang t y , chan m , handigol n , mckeown n , parulkar g ( 2010 ) blueprint for introducing innovation into wireless mobile networks . in : proceedings of the second acm sigcomm workshop on virtualized infrastructure systems and architectures , ser .
visa 10 , new york , usa , pp 25 - 32 | in the past few decades , the world has witnessed a rapid growth in mobile communication and reaped great benefits from it . even though the fourth generation ( 4 g ) mobile communication system is just being deployed worldwide
, proliferating mobile demands call for newer wireless communication technologies with even better performance .
consequently , the fifth generation ( 5 g ) system is already emerging in the research field .
however , simply evolving the current mobile networks can hardly meet such great expectations , because over the years the infrastructures have generally become ossified , closed , and vertically constructed .
aiming to establish a new paradigm for 5 g mobile networks , in this article , we propose a cross - layer software - defined 5 g network architecture . by jointly considering both the network layer and the physical layer together ,
we establish the two software - defined programmable components , the control plane and the cloud computing pool , which enable an effective control of the mobile network from the global perspective and benefit technological innovations .
specifically , by the cross - layer design for software - defining , the logically centralized and programmable control plane abstracts the control functions from the network layer down to the physical layer , through which we achieve the fine - grained controlling of mobile network , while the cloud computing pool provides powerful computing capability to implement the baseband data processing of multiple heterogeneous networks .
the flexible programmability feature of our architecture makes it convenient to deploy cross - layer technological innovations and benefits the network evolution .
we discuss the main challenges of our architecture , including the fine - grained control strategies , network virtualization , and programmability .
the architecture significantly benefits the convergence towards heterogeneous networks and it enables much more controllable , programmable and evolvable mobile networks .
qualitative simulations validate these performance advantages . |
alice ( a large ion collider experiment ) is a general - purpose heavy - ion detector at the cern lhc ( large hadron collider ) .
it has been designed in order to fulfill the requirements to track and identify particles from very low ( @xmath5100 mev/@xmath6 ) up to quite high ( @xmath5100 gev/@xmath6 ) transverse momenta in an environment with large charged - particle multiplicities as in the case of central lead - lead ( pb pb ) collisions at the lhc . ; the alice experiment , shown in figure [ fig : alice ] , consists of a central - barrel detector and several forward detector systems .
the central system covers the mid - rapidity region ( @xmath13 0.9 ) over the full azimuthal angle .
it is installed inside a large solenoidal magnet providing a moderate magnetic field of 0.5 t. it includes a six - layer high - resolution inner - tracking system ( its ) , a large - volume time - projection chamber ( tpc ) and electron and charged - hadron identification detectors which exploit transition - radiation ( trd ) and time - of - flight ( tof ) techniques , respectively .
small - area systems for high-@xmath0 particle - identification ( hmpid ) , photon and neutral - meson measurements ( phos ) and jet reconstruction ( emcal ) complement the central barrel .
thanks to these unique features the experiment is able to identify hadrons in a wide momentum range by employing different detection systems and techniques , as discussed in section [ sec : pid ] .
the detectors which cover larger rapidity regions include a single - arm muon spectrometer covering the pseudorapidity range -4.0 @xmath14 -2.4 and several smaller detectors ( vzero , tzero , fmd , zdc , and pmd ) for triggering , multiplicity measurements and centrality determination .
a detailed description of the alice detector layout and of its subsystems can be found in @xcite . since november 2009 when the first collisions at the lhc occurred , the alice experiment has collected proton - proton data at several centre - of - mass energies ( @xmath9 = 0.9 , 2.76 , 7 and 8 tev ) . during the first two lhc heavy - ion runs , in 2010 and 2011 ,
the detector recorded pb
pb collisions at a centre - of - mass energy per nucleon pair of @xmath10 = 2.76 tev and could profit from of an integrated luminosity of about 10 @xmath15 and 100 @xmath15 , respectively .
it is worth stressing the outstanding performance of the lhc complex : the instant luminosity exceeded @xmath16 in the second run , higher than the design value .
proton - lead ( p pb ) collision data were also collected with the alice detector .
this occurred during a short run performed in september 2012 in preparation for the main p
pb run at the beginning of 2013 .
eight pairs of bunches collided in the alice interaction region , providing a luminosity of about @xmath17 .
the configuration ( 4 tev protons colliding with fully stripped @xmath18pb ions at @xmath19 tev ) resulted in interactions at @xmath10 = 5.02 tev in the nucleon - nucleon centre - of - mass system , which moves with a rapidity of @xmath20 = 0.465 .
/d@xmath8 in the its ( top - left ) and in the tpc ( top - right ) .
the continuous curves represent the bethe - bloch parametrization .
( bottom - left ) particle velocity @xmath21 measured with tof as a function of momentum .
( bottom - right ) cherenkov angle measured in the hmpid as a function of the track momentum.,title="fig : " ] /d@xmath8 in the its ( top - left ) and in the tpc ( top - right ) .
the continuous curves represent the bethe - bloch parametrization .
( bottom - left ) particle velocity @xmath21 measured with tof as a function of momentum .
( bottom - right ) cherenkov angle measured in the hmpid as a function of the track momentum.,title="fig : " ] /d@xmath8 in the its ( top - left ) and in the tpc ( top - right ) .
the continuous curves represent the bethe - bloch parametrization .
( bottom - left ) particle velocity @xmath21 measured with tof as a function of momentum .
( bottom - right ) cherenkov angle measured in the hmpid as a function of the track momentum.,title="fig : " ] /d@xmath8 in the its ( top - left ) and in the tpc ( top - right ) .
the continuous curves represent the bethe - bloch parametrization .
( bottom - left ) particle velocity @xmath21 measured with tof as a function of momentum .
( bottom - right ) cherenkov angle measured in the hmpid as a function of the track momentum.,title="fig : " ] ) of protons in the range between 0.6 gev/@xmath6 and 0.65 gev/@xmath6 in 0 - 5% most central pb
pb collisions together with the monte carlo templates which are fitted to the data . ] in this section the main particle - identification ( pid ) detectors relevant to the analyses presented in this paper are briefly discussed .
a detailed review of the alice experiment and of its pid capabilities can be found in @xcite .
the its is the innermost detector system , a six - layer silicon detector located at radii between 4 and 43 cm .
four of the six layers provide specific energy loss d@xmath7/d@xmath8 measurements and are used for particle identification in the non - relativistic ( @xmath22 ) region . by using the its as a standalone tracker it is possible to reconstruct and identify low - momentum particles ( below 200 mev / c ) not reaching the main tracking systems ( figure [ fig : pid ] top - left ) .
the tpc is the main central - barrel tracking detector of alice .
it provides three - dimensional hit information and specific energy - loss measurements with up to 159 samples . with the measured particle momentum and
@xmath23d@xmath7/d@xmath24 the particle type can be determined by comparing the measurements with the bethe - bloch expectation ( figure [ fig : pid ] top - right ) .
the tof detector is a large - area array of multigap resistive plate chambers ( mrpc ) and covers the central pseudorapidity region ( @xmath25 0.9 , full azimuth ) .
particle identification is performed by matching momentum and trajectory - length measurements performed by the tracking system with the time - of - flight information provided by the tof system .
the overall time - of - flight resolution is measured to be about 85 ps in pb pb collisions ( about 120 ps in pp collisions ) and it is determined by the time resolution of the detector itself and by the start - time resolution ( figure [ fig : pid ] bottom - left ) .
the hmpid detector consists of seven identical proximity focusing rich ( ring imaging cherenkov ) counters .
photon detection is performed using the proportional multiwire chambers coupled to pad - segmented csi photocathode .
particle identification is obtained by means of the measurement of the cherenkov angle allowing the separation of pions and kaons between 1 gev/@xmath6 and 3 gev/@xmath6 and protons from 1.5 gev/@xmath6 up to 5 gev/@xmath6 ( figure [ fig : pid ] ) .
the transverse momentum spectra of primary @xmath26 , @xmath27 , @xmath28 and @xmath29 are measured at mid - rapidity ( @xmath4 ) combining the techniques and detectors described above .
primary particles are defined as prompt particles produced in the collision and all decay daughters , except products from weak decays of strange particles . the contribution from the feed - down of weakly - decaying particles to @xmath26 , @xmath28 and @xmath29 and from protons emitted from secondary interactions with material
are subtracted by fitting the data using monte carlo templates of the dca distributions ( figure [ fig : dca ] ) .
particles can also be identified in alice via their characteristics decay topology or invariant mass .
this , combined with the direct identification of the decay daughters , allows one to reconstruct weakly - decaying particles and hadronic resonances with a good signal - to - background ratio .
( top ) and @xmath30 production ratio in pp collisions compared with pythia monte carlo predictions and nlo calculations.,title="fig : " ] ( top ) and @xmath30 production ratio in pp collisions compared with pythia monte carlo predictions and nlo calculations.,title="fig : " ] as a function of @xmath0 for pb
pb collisions at @xmath10 = 2.76 tev for centrality 0 - 5% ( left ) and 20 - 40% ( right ) .
statistical ( vertical error bars ) and systematic ( gray and colored boxes ) are shown for the charged pion raa .
the gray boxes contains the common systematic error related to pp normalization to inel and @xmath31.,title="fig : " ] as a function of @xmath0 for pb
pb collisions at @xmath10 = 2.76 tev for centrality 0 - 5% ( left ) and 20 - 40% ( right ) .
statistical ( vertical error bars ) and systematic ( gray and colored boxes ) are shown for the charged pion raa .
the gray boxes contains the common systematic error related to pp normalization to inel and @xmath31.,title="fig : " ] of charged pions compared to the @xmath32 of unidentified charged particles as a function of @xmath0 for different centrality classes .
statistical ( vertical error bars ) and systematic ( gray and colored boxes ) errors are shown for the charged pions .
the colored boxes contain the common systematic uncertainty related to the number of binary collisions and the pp normalization to inel .
only statistical errors are shown for the unidentified charged @xmath32 . ]
alice has measured the production yields of primary charged pions , kaon and ( anti)protons in a wide momentum range and in several colliding systems .
the measurements have been performed in proton - proton collisions at several centre - of - mass energies ( @xmath9 = 0.9 , 2.76 and 7 tev ) and in pb
pb collisions at @xmath10 = 2.76 tev as a function of collision centrality . in pb
pb collisions the transverse momentum @xmath0 distributions and yields are compared to previous results at rhic and expectations from hydrodynamic and thermal model .
the results obtained for central pb
pb collisions are shown in figure [ fig : spectrapbpb ] and [ fig : ratiospbpb ] .
the spectral shapes are harder than those observed at rhic , indicating an increase of the radial flow velocity with the centre - of - mass energy . the radial flow at the lhc is found to be about 10% higher than at rhic energy .
the hydrodynamic models shown in figure [ fig : spectrapbpb ] give for central collisions a fair description of the data , describing the experimental spectra within 20% .
this supports a hydrodynamic interpretation of the transverse momentum distribution in central collisions at the lhc .
further details of this anaysis and the models used to compare with the data can be found in @xcite . while the k/@xmath11 integrated production ratio is observed to be in line with lower energy measurements and predictions from the thermal model , both the p/@xmath11 and the @xmath33 ratios are lower than those at rhic and significantly lower ( a factor @xmath5 1.5 and 1.35 , respectively ) than predictions .
a possible explanation of these deviations from the thermal - model predictions may be re - interactions in the hadronic phase due to large cross sections for antibaryon - baryon annihilation @xcite .
the @xmath0-dependent yield of charged kaons and protons normalized to charged pions are shown in figure [ fig : ratiospp ] for pp collisions at @xmath9 = 2.76 and 7 tev where they are compared with theoretical model predictions . within the experimental uncertainties
no energy dependence is observed in the data .
the observed production ratios are not reproduced by nlo calculations @xcite .
pythia monte carlo generator @xcite underpredicts the proton - to - pion ratio at intermediate @xmath0 .
pion , kaon and ( anti)proton production in pb
pb collisions have been compared to that in pp collisions and all show a suppression pattern which is similar to that of inclusive charged hadrons at high momenta ( @xmath0 above @xmath34 10 gev/@xmath6 ) , as shown in figure [ fig : raa ] and [ fig : pionraa ] @xcite .
this seems to suggest that the dense medium formed in pb
pb collisions does not affect the fragmentation .
a similar conclusion can be drawn from the proton - to - pion ratio measured in pb
pb collisions ( figure [ fig : baryonmeson ] and [ fig : baryonmeson2 ] ) : for intermediate momenta ( 37 gev/@xmath6 ) it exibits a relatively strong enhancement , by a factor of 3 compared to that in pp collisions at @xmath35 3 gev/@xmath6 and returns to the value in pp collisions at higher momenta ( @xmath0 above @xmath34 10 gev/@xmath6 ) @xcite . a similar observation was also reported for the @xmath36 ratio and possible explanations have been proposed which include particle production via quark recombination @xcite .
particle production in proton - lead ( p pb ) collisions allows one to study and understand qcd at low parton fractional momentum @xmath8 and high gluon density .
it is moreover expected to be sensitive to nuclear effects in the initial state .
for this reason p pb measurements provide an essential reference tool to discriminate between initial and final state effects and they are crucial for the studies and the understanding of deconfined matter created in nucleus - nucleus collisions .
= 5.02 tev compared to measurements in central ( 0 - 5% ) and peripheral ( 70 - 80% ) pb pb collisions at @xmath10 = 2.76 tev . ]
the measurement of the transverse momentum @xmath0 distributions of charged particle in p
pb collisions were reported already in @xcite .
it was previously shown that the production of charged hadrons in central pb
pb collisions at the lhc is strongly suppressed @xcite .
the suppression remains substantial up to 100 gev/@xmath6 and is also seen in reconstructed jets @xcite .
proton - lead collisions provide a control experiment to clearly establish whether the initial state of the colliding nuclei plays a role in the observed high-@xmath0 hadron production in pb
pb collisions . in order to quantify nuclear effects ,
the @xmath0-differential yield relative to the proton - proton reference , the so - called nuclear modification factor , is calculated .
the nuclear modification factor is expected to be unity for hard processes which exhibit binary collision scaling .
this has been recently confirmed in pb
pb collisions at the lhc by the measurements of direct photon @xcite , z@xmath37 @xcite and w@xmath2 @xcite , observables which are not affected by hot qcd matter . in figure
[ fig : rppb ] the measurement of the nuclear modification factor in p
pb collisions r@xmath38 is compared to that in central ( 0 - 5% centrality ) and peripheral ( 7080% ) pb pb collisions r@xmath39 .
r@xmath38 is observed to be consistent with unity for transverse momenta higher than about 2 gev/@xmath6 .
this important measurement demonstrates that the strong suppression observed in central pb
pb collisions at the lhc is not due to an initial - state effect , but it is a final state effect related to the hot matter created in high - energy heavy - ion reactions .
alice has obtained so far a wealth of physics results both from the analysis of proton - proton collision data and from the first two lhc heavy - ion runs .
the transverse momentum spectra of @xmath1 , @xmath40 , @xmath41 and @xmath42 have been measured with alice in several colliding systems and energies at the lhc demonstrating the excellent pid capabilities of the experiment .
data in pp collisions show no evident @xmath9 dependence in hadron production ratios . in pb
pb collisions @xmath43 integrated ratio is significantly lower than statistical model predictions with a chemical freeze - out temperature @xmath44 mev .
the average transverse momenta and the transverse momentum spectra indicate a @xmath510% stronger radial flow than at rhic energies . in the intermediate transverse momentum @xmath0 region , an enhancement of the baryon - to - meson ratio
is observed .
the maximum of the ratio is shifted to higher @xmath0 with respect to rhic measurements .
the results of the measurements of charged pion , kaon , protons and antiproton production at high @xmath0 were also presented .
the nuclear modification factors for these species are similar in magnitude , suggesting that the medium does not significantly affect fragmentation .
first results from a short pilot run with proton - lead beams have been also reported .
the measurement of the charged - particle transverse momentum @xmath0 spectra and nuclear modification factor in p
pb collisions at @xmath10 = 5.02 tev , covering 0.5 @xmath45 20 gev/@xmath6 , show a nuclear modification factor consistent with unity for @xmath46 2 gev / c .
this measurement indicates that the strong suppression of hadron production at high @xmath0 observed at the lhc in pb
pb collisions is not due to an initial - state effect , but is the fingerprint of jet quenching in hot qcd matter .
9 k. aamodt _ et al . _
[ alice collaboration ] , `` the alice experiment at the cern lhc '' , jinst * 3 * ( 2008 ) s08002 .
g. alessandro , ( ed . ) _ et al . _
[ alice collaboration ] , `` alice : physics performance report , volume ii '' , j. phys .
g * 32 * ( 2006 ) 1295 .
b. abelev _ et al . _
[ alice collaboration ] , `` pseudorapidity density of charged particles @xmath41-pb collisions at @xmath47 tev '' , arxiv:1210.3615 [ nucl - ex ] .
b. alver , m. baker , c. loizides and p. steinberg , `` the phobos glauber monte carlo '' , arxiv:0805.4411 [ nucl - ex ]
. b. abelev _ et al . _
[ alice collaboration ] , `` pion , kaon , and proton production in central pb
pb collisions at @xmath48 tev '' , phys .
* 109 * ( 2012 ) 252301 [ arxiv:1208.1974 [ hep - ex ] ] .
j. steinheimer , j. aichelin and m. bleicher , `` non - thermal @xmath49 ratio at lhc as a consequence of hadronic final state interactions '' , arxiv:1203.5302 [ nucl - th ] .
f. becattini , m. bleicher , t. kollegger , m. mitrovski , t. schuster and r. stock , `` hadronization and hadronic freeze - out in relativistic nuclear collisions '' , phys .
c * 85 * ( 2012 ) 044921 [ arxiv:1201.6349 [ nucl - th ] ] .
y. pan and s. pratt , `` baryon annihilation in heavy ion collisions '' , arxiv:1210.1577 [ nucl - th ] .
r. sassot , p. zurita and m. stratmann , `` inclusive hadron production in the cern - lhc era '' , phys . rev .
d * 82 * ( 2010 ) 074011 [ arxiv:1008.0540 [ hep - ph ] ] .
z. skands , phys .
d * 82 * ( 2010 ) 074018 [ arxiv:1005.3457 [ hep - ph ] ] .
a. ortiz velasquez [ alice collaboration ] , `` production of pions , kaons and protons at high @xmath50 in @xmath51 tev pb - pb collisions '' , arxiv:1210.6995 [ hep - ex ] .
r. j. fries , b. muller , c. nonaka and s. a. bass , `` hadronization in heavy ion collisions : recombination and fragmentation of partons '' , phys .
* 90 * ( 2003 ) 202303 [ nucl - th/0301087 ] .
b. abelev _ et al . _
[ alice collaboration ] , `` transverse momentum distribution and nuclear modification factor of charged particles in @xmath41-pb collisions at @xmath47 tev '' , arxiv:1210.4520 [ nucl - ex ] . k. aamodt _ et al . _
[ alice collaboration ] , `` suppression of charged particle production at large transverse momentum in central pb
pb collisions at @xmath48 tev '' , phys .
b * 696 * ( 2011 ) 30 [ arxiv:1012.1004 [ nucl - ex ] ] .
b. abelev _ et al . _
[ alice collaboration ] , `` centrality dependence of charged particle production at large transverse momentum in pb
pb collisions at @xmath52 tev '' , [ arxiv:1208.2711 [ hep - ex ] ] . g. aad _ et al . _ [ atlas collaboration ] , `` measurement of the jet radius and transverse momentum dependence of inclusive jet suppression in lead - lead collisions at @xmath51 tev with the atlas detector '' , arxiv:1208.1967 [ hep - ex ] . s. chatrchyan _ et al . _
[ cms collaboration ] , `` measurement of isolated photon production in @xmath53 and pbpb collisions at @xmath51 tev '' , phys .
b * 710 * ( 2012 ) 256 [ arxiv:1201.3093 [ nucl - ex ] ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , `` study of z boson production in pbpb collisions at nucleon - nucleon centre of mass energy = 2.76 tev '' , phys .
* 106 * ( 2011 ) 212301 [ arxiv:1102.5435 [ nucl - ex ] ] . s. chatrchyan _ et al . _
[ cms collaboration ] , `` study of @xmath54 boson production in pbpb and @xmath53 collisions at @xmath51 tev '' , phys .
b * 715 * ( 2012 ) 66 [ arxiv:1205.6334 [ nucl - ex ] ] . | thanks to the unique detector design adopted to fulfill tracking and particle - identification ( pid ) requirements ( e.g. low momentum cut - off and low material budget ) , the alice experiment provides significant information about hadron production both in pp and pb pb collisions . in particular , the @xmath0-differential and integrated production yields of identified particles play a key role in the study of the collective and thermal properties of the matter formed in high - energy heavy - ion collisions . furthermore
, the production of high-@xmath0 particles provides insights into the property of the hot medium created in such collisions and the in - medium energy - loss mechanisms .
transverse momentum spectra of @xmath1 , k@xmath2 , p and @xmath3 are measured at mid - rapidity ( @xmath4 ) over a wide momentum range , from @xmath5 100 mev/@xmath6 up to @xmath5 20 gev/@xmath6 .
the measurements are performed exploiting the d@xmath7/d@xmath8 in silicon and gas , the time - of - flight and the ring - imaging cherenkov particle - identification techniques , which will be briefly reviewed in this report .
the current results on light - flavour charged - hadron production will be presented for pp collisions at @xmath9 = 0.9 , 2.76 and 7 tev and for pb
pb collisions at @xmath10 = 2.76 tev .
integrated production yields , transverse momentum spectra and particle ratios in pp are discussed as a function of the collision energy and compared to previous experiments and commonly - used monte carlo models .
pb collisions at the lhc feature the highest radial flow ever observed and an unexpectedly low p/@xmath11 production ratio . the results are presented as a function of collision centrality and compared to rhic data in au au collisions at @xmath10 = 200 gev and predictions from thermal and hydrodynamic models .
the nuclear modification factor ( @xmath12 ) of identified hadrons will also be discussed and compared to unidentified charged particles and theoretical predictions .
this is observed to be identical for all particle species at high-@xmath0 . |
atomic ensembles are key ingredients in many theoretical and experimental schemes whose aim is the implementation of quantum information protocols @xcite , or the generation of paired photons with non - classical correlations @xcite . in these `` writing - reading '' schemes , a weak classical field ( pump pulse ) interacts with an atomic ensemble , which leads to the spontaneous emission of a stokes photon .
since the stokes photon and the atomic ensemble are highly correlated , the projection of the stokes photon heralds the generation of an atomic state that is a coherent superposition of all possible states of the ensemble where only one atom has been excited , the so - called collective atomic state @xcite .
the selection of a specific direction of stokes photons emission is an important issue , since i ) one aims at choosing a direction that maximizes the flux of generated photons , and ii ) the direction in which the photons are detected might determine the specific quantum state of the atomic ensemble .
it has been shown that in the case of a room - temperature atomic cloud , where atoms are considered to move fast within the cloud , stokes photons are emitted within a small cone around the direction of propagation of the pump beam @xcite , whereas in the case where the atoms are considered to be fixed in their positions ( _ cold atomic clouds _ ) , stokes photons have no preferred direction of emission @xcite , always that it is not forbidden by the transition matrix elements . even though , in most experiments , the emitted photon is detected at a small angle ( @xmath1 ) @xcite .
these results consider only the angular distribution of emitted photons in two limiting cases : when the atoms are either moving very fast ( high temperature ) or completely fixed ( low temperature ) within the cloud .
notwithstanding , the transition between these two cases has not been explored yet . in this paper , we construct a model to describe the angular distribution of emitted stokes photons as a function of the temperature of the atomic cloud .
the importance of our result resides in the fact that we can readily develop a new technique where the measurement of the width of the emission cone can be used to determine the temperature of the atomic cloud .
this is made possible by unveiling the close relationship that exists between the range of possible directions of emission , and the _ which - way _ information available about where the photon originated , i.e. , knowledge of the position of the atom that emitted the stokes photon during the interaction .
one of the attributes of this new technique is that , unlike commonly used time - of - flight ( tof ) measurements @xcite , the atomic cloud is not destroyed during each measurement .
this new technique is thus added to the group of non - destructive measurements such as resonance fluorescence spectrum analysis @xcite , recoil - induced resonances @xcite , and transient four - wave mixing @xcite . also ,
since this technique does not require any additional elements in the experimental setup , its implementation can be easily carried out .
let us consider a cloud of n identical three - level atoms in a @xmath0-configuration ( fig .
[ figure1 ] ) .
the cloud is illuminated by a laser pulse that couples the transition @xmath2 with a detuning @xmath3 .
the spontaneous decay of the atom ( @xmath4 ) leads to the generation of a photon with different wavelength ( stokes photon ) . here
we investigate two important features that , as we will see later , depend on the temperature of the cloud .
first , the angular distribution of the spontaneously emitted photons , and second , the heralded generation of the symmetric collective atomic state after detection of the stokes photon . the pump beam , which is a slowly - varying classical field propagating along the @xmath5-direction , with a rayleigh range much larger than the length of the atomic cloud , writes @xmath6 where @xmath7 is the central frequency , @xmath8 is velocity of light , @xmath9 describes the transverse spatial shape of the pump beam and @xmath10 its temporal shape .
the stokes field , which is described quantum mechanically , writes @xmath11 where @xmath12 is the annihilation operator , @xmath13 is the wavevector of the stokes photon and @xmath14 its frequency . in the interaction picture , and after adiabatically eliminating the upper level @xmath15 , the interaction of the light field with the atomic cloud can be described by the hamiltonian @xcite @xmath16 where @xmath17 is the transition operator for the _ _ i__th atom , @xmath18 is the vector position of the _ _ i__th atom , @xmath19 is the coupling coefficient of the transition , @xmath20 and @xmath21 , with @xmath22 being the transition frequency between states @xmath23 and @xmath24 . before the interaction , we consider that all the atoms are in the ground state and that there are no photons in the optical modes , i.e , @xmath25 .
then , considering that the pump field is weak enough , we can make use of first - order perturbation theory to write the state of the system as @xmath26 where @xmath27 .
we assume that @xmath28 becomes independent of the frequency and that @xmath29 has the same value for all allowed directions of emission of the stokes photons .
the use of perturbation theory is motivated by experiments in which a weak pump pulse and a short interaction time are used in order to guarantee that the probability of creating more than one excitation in the collective atomic state is very low @xcite .
this weak - pumping condition makes a perturbative approach suitable for describing a realistic situation . in the case of cold atomic ensembles , since the atoms are considered to be fixed in their positions , we can directly use eq .
( [ state_atoms2 ] ) to obtain that the probability of emitting a photon in a given direction @xmath30 is equal for all directions , i.e. , there is no preferred direction of emission @xcite , independently of the specific shape of the atomic cloud , always that it is not forbidden by the transition matrix elements . in the case of hot atomic ensembles , due to the fact that during the light - atom interaction time the atoms are moving fast ,
an average value over all positions @xmath31 should be performed @xcite .
the resulting state becomes @xmath32 where the average value over all positions is expressed as @xmath33 , with @xmath34 being the atomic distribution function .
( [ state_hot_atoms ] ) , one can show that the stokes photons are emitted in a small cone around the forward direction ( see @xcite for a detailed calculation ) , whose width depends on the particular spatial shape of the atomic cloud .
notice that , in this case , the photon and atomic degrees of freedom can be decoupled , and the quantum state of the atoms is the so - called symmetric collective atomic state , i.e. , @xmath35 . equation ( [ state_atoms2 ] ) does not show any temperature dependence , so in its present form can not be used to describe the transition between the two limiting scenarios considered so far : warm and cold atomic ensembles . in order to model the temperature dependence
, we introduce a function that describes the movement of each atom around its mean position @xmath31 , @xmath36},\ ] ] where @xmath37 determines the radius of the area over which the atoms can move during the interaction time .
it depends on the pump pulse duration ( @xmath38 ) , and on the speed ( @xmath39 ) most likely to be possessed by any atom of the system . here
@xmath40 is the mass of the atom , @xmath41 is the boltzmann constant and @xmath42 is the temperature of the atomic ensemble .
notice that the origin of @xmath43 lies in the maxwell@xmath44boltzmann distribution .
this distribution is assumed , because it has been shown that the maxwell - boltzmann distribution provides an accurate description of the motion of atoms at temperatures above tenths of @xmath45 @xcite .
hence , eq . ( 6 ) is useful for describing the motion of atoms undergoing a transition from the hot to the cold condition , provided that the lowest temperature values are above tenths of @xmath45 . making use of the function given in eq .
( [ function_new ] ) to rewrite eq .
( [ state_atoms2 ] ) , the temperature - dependent quantum state of the system atoms - photon can be written as @xmath46 where @xmath47 is the volume of the cloud .
note that , in the limit where @xmath48 , the function given in eq .
( [ function_new ] ) tends to a dirac delta function , and we recover the state of the system described by eq .
( [ state_atoms2 ] ) . in order to obtain the angular distribution of the emitted stokes photons
, we trace out the atomic variables of the density matrix of the system ( @xmath49 ) . neglecting the vacuum contribution ,
the reduced density matrix of the photon state writes @xmath50 where @xmath51 considering that the atoms are contained in a cell with transversal dimensions
@xmath52 , @xmath53 and length @xmath54 , we can solve eq .
( 9 ) to obtain @xmath55 where @xmath56}^{2 } \right ] } \nonumber \\ & & \times { \left\lbrace \mathrm{erf}{\left [ -\frac{1}{2}\alpha^{3}{\left ( 2x_{i}-ik_{x}a^{2}-\frac{l_{x}}{\alpha^{2 } } \right ) } \right ] } - \mathrm{erf}{\left [ -\frac{1}{2}\alpha^{3}{\left ( 2x_{i}-ik_{x}a^{2}+\frac{l_{x}}{\alpha^{2 } } \right ) } \right ] } \right\rbrace } , \\ \omega(z_{i},k_{z } ) & = & \exp{\left [ -\frac{1}{4}k_{z}^{2}a^{2 } - ik_{z}z_{j } \right ] } \nonumber \\ & & \times { \left\lbrace \mathrm{erf}{\left [ -\frac{1}{2a}{\left ( 2z_{i}-ik_{z}a^{2}-l_{z } \right ) } \right]}-\mathrm{erf}{\left [ -\frac{1}{2a}{\left ( 2z_{i}-ik_{z}a^{2}+l_{z } \right ) } \right ] } \right\rbrace}.\end{aligned}\ ] ] notice that the presence of the error function ( erf(x ) ) in eq .
( 10 ) is due to the integration over the finite volume ( v ) of the cell that contains the atoms .
we now find that the probability of emitting a stokes photon in the direction @xmath57 is given by the diagonal terms of the density matrix ( [ density_matrix_stokes ] ) , @xmath58 where the normalization condition writes @xmath59 .
in general , since the atomic cloud contains a large atom number density , the atomic summation can be rewritten as @xmath60 .
note that eq .
( 14 ) needs to be solved numerically due to the presence of the error function in eq .
notwithstanding , since the functions of the spatial variables are separated ( as can be seen in eq .
( 10 ) ) , the numerical integration of eq .
( [ probability_of_emission ] ) can be easily performed .
we have calculated the angular distribution of the emitted stokes photons considering an ensemble of @xmath61rb atoms contained in a pencil - shaped cell with transversal dimensions : @xmath62 mm , and length @xmath63 mm .
the atoms are illuminated by a pump pulse with a transversal shape given by @xmath64 , where @xmath65 mm is the beam waist of the pump beam .
the level configuration of the atoms is set to @xmath66 for the excited level @xmath15 , and the zeeman - splitting levels @xmath67 and @xmath68 for the @xmath23 and @xmath24 states , respectively .
@xmath69k ; dashed line : @xmath70 k ; dotted line : @xmath71 k and dash - dotted line : @xmath72 k. in all cases , the pump pulse duration is set to @xmath73 ns.,width=264 ] figure 2 shows the angular distribution ( normalized to the maximum ) of emitted stokes photons as a function of the angle ( @xmath74 ) between the direction of the pump and the emitted photon ( as shown in fig .
3(a ) ) . in the low temperature limit
, the spontaneous emission of stokes photons has no preferred direction .
this result agrees with ref .
@xcite , in which stokes photons are said to be emitted into @xmath75 steradian .
in contrast , as the temperature of the cloud is increased , the probability distribution narrows around @xmath76 , evidencing the fact that , in the case of warm atomic ensembles , stokes photons are emitted preferentially along the direction of the pump , as it has been experimentally observed , for instance in @xcite .
the results shown in fig .
2 can be understood in terms of the which - way information left in the atoms after emitting a stokes photon , i.e. , information about the position of the atom that emitted the photon .
on one hand , for the case of cold atoms , due to the fact that they are fixed , one can obtain , in principle , information about the position of the atom that emitted the photon .
in this situation , the possible paths of the stokes photon will not interfere , because which - way information has been left in the atomic ensemble .
this can be clearly seen from eq .
( 14 ) , which for cold atomic ensembles takes the form @xmath77 equation ( 15 ) shows that emission of stokes photons from a cold atomic ensemble has no preferred direction .
on the other hand , notice that for the case of hot atomic ensembles , eq . ( 6 ) is a constant inside the integration volume , so we can write eq .
( 14 ) , by means of the large atom number density relation , as @xmath78 we see from eq .
( 16 ) that interference between the possible paths of the stokes photon is now restored , because which - way information has been erased by the movement of the atoms in the cloud .
interestingly , this which - way information effect has been also observed , for instance , in the context of second - order interference of single photons @xcite .
+ it is important to highlight that , it is not the actual acquisition of information from the system which determines the rise of interference , but that in principle , it would be possible to obtain such information @xcite .
the close relationship between the width of the emission cone and the temperature of the atomic ensemble allows us to put forward a new technique to determine the temperature of atomic clouds .
the proposed experimental setup consists of an array of detectors ( or a movable detector ) that would be able to detect stokes photons along different directions , as depicted in fig .
in this way , by measuring the width of the emission cone , we can make use of eq .
( [ probability_of_emission ] ) to retrieve information about the temperature of the atomic ensemble .
figure 3(b ) shows the temperature of the atomic ensemble as a function of the full width at half maximum ( fwhm ) of the emission cone .
notice that , by choosing a sufficiently short pulse , the relationship between the emission cone width and the temperature gets smoother .
this can be useful for a better discrimination of the width of the emission cone , enhancing thus the precision of the technique .
also , notice that this technique is not based on the ballistic expansion of the atomic cloud @xcite , so each measurement can be performed without destroying it .
we can also use eq . ( 7 ) to describe how the generation of the symmetric state depends on the temperature of the atomic cloud . when the stokes photon is detected in an arbitrary direction @xmath30 ,
i.e. , is projected into the state @xmath79 , the corresponding quantum state of the atomic cloud is @xmath80 + + + + in the case where the atoms are barely moving during the interaction time , only a small fraction of generated photons ( those in a small angle around the pump beam direction ) corresponds to the symmetric state , even though photons are emitted in a larger emission cone .
this can clearly be seen by comparing figs . [
figure4](a ) and ( b ) , where we plot the projection @xmath81 . when the atoms are let to move within the cloud by increasing the temperature , the _ which - way _ information is erased and the emission cone gets narrower ( see fig . [ figure4](g ) ) . in this case , as it can be seen from fig .
[ figure4](h ) , photons emitted in all allowed possible directions are in the symmetric state .
notwithstanding , we are again forced to consider small emission angles around the pump beam direction to enhance the flux of detected stokes photons .
therefore , in all cases , photons should be collected in a small cone around the direction of propagation of the pump beam , if the goal is to generate the symmetric atomic state .
but as fig .
[ figure4 ] shows , the reason behind this restriction depends on the temperature of the atomic ensemble .
in this paper , we have presented a new technique to measure the temperature of atomic ensembles , based on the relationship between the temperature of the ensemble and the width of the emission cone of the spontaneously emitted stokes photons .
we have also shown that heralded generation of the collective symmetric atomic state requires the detection of the heralding stokes photon in a narrow cone around the direction of the exciting pulse .
for cold atomic ensembles , this is the only direction of emission that guarantees the generation of such state , while for warm ensembles , it is the direction with the highest efficiency .
this work was supported by projects fis2010 - 14831 and fet - open 255914 ( phorbitech ) .
this work has also been supported by fundacio privada cellex barcelona .
we thank j. svozilk and g. puentes for helpful discussions .
duan , m. d. lukin , j. i. cirac , and p. zoller , nature ( london ) * 414 * , 413 ( 2001 ) .
duan , phys .
lett . * 88 * , 170402 ( 2002 ) . c. h. van der wal , m. d. eisaman , a. andr , r. l. walsworth , d. f. phillips , a. s. zibrov , and m. d. lukin , science * 301 * , 196 ( 2003 ) .
a. kuzmich , w. p. bowen , a. d. boozer , a. boca , c. w. chou , l .- m .
duan , and h. j. kimble , nature ( london ) * 423 * , 731 ( 2003 ) .
n. sangouard , c. simon , h. de riedmatten , and n. gisin , rev .
* 83 * , 33 ( 2011 ) .
l. m. duan , j. i. cirac and p. zoller , phys .
a * 66 * , 023818 ( 2002 ) .
m. o. scully and m. s. zubairy , science * 301 * , 181 ( 2003 ) .
m. o. scully , e. s. fry , c. h. raymond ooi , and k. wdkiewicz , phys .
lett . * 96 * , 010501 ( 2006 ) .
d. n. matsukevich , t. chanelire , m. bhattacharya , s .- y .
lan , s. d. jenkins , t. a. b. kennedy , and a. kuzmich , phys .
lett . * 95 * , 040405 ( 2005 ) .
d. felinto , c. w. chou , h. de riedmatten , s. v. polyakov , and h. j. kimble , phys .
a * 72 * , 053809 ( 2005 ) .
r. inoue , n. kanai , t. yonehara , y. miyamoto , m. koashi , and m. kozuma , phys .
a * 74 * , 053809 ( 2006 ) .
p. d. lett , r. n. watts , c. i. westbrook , w. d. phillips , p. l. gould , and h. j. metcalf , phys .
lett . * 61 * , 169 ( 1988 ) .
d. lett , w. d. phillips , s. l. rolston , c. e. tanner , r. n. watts , and c. i. westbrook , j. opt .
b * 6 * , 2084 ( 1989 ) . c. i. westbrook ,
r. n. watts , c. e. tanner , s. l. rolston , w. d. phillips , p. d. lett , and p. l. gould , phys .
lett . * 65 * , 33 ( 1990 ) .
j. y. courtois , g. grynberg , b. lounis , and p. verkerk , phys .
lett . * 72 * , 3017 ( 1994 ) .
m. mitsunaga , m. yamashita , m. koashi , and n. imoto , opt
. lett . * 23 * , 840 ( 1998 ) .
j. h. eberly , j. phys .
b : at . mol .
. phys . * 39 * , s599 ( 2006 ) .
x. y. zou , l. j. wang , and l. mandel , phys .
lett . * 67 * , 318 ( 1991 ) .
l. mandel , rev .
* 71 * , s274 ( 1999 ) . | we unveil the relationship existing between the temperature of an ensemble of three - level atoms in a @xmath0 configuration , and the width of the emission cone of stokes photons that are spontaneously emitted when atoms are excited by an optical pulse .
this relationship , which is based on the amount of _ which - way _ information available about where the stokes photon originated during the interaction , allows us to put forward a scheme to determine the temperature of atomic clouds by measuring the width of the emission cone . unlike the commonly used time - of - flight measurements , with this technique , the atomic cloud is not destroyed during each measurement .
pacs numbers : : 42.50.ct , 03.67.-a |
the d - wave machine is a quantum computer of the quantum annealing type @xcite .
it is designed to solve classical optimization problems , namely to optimize a cost function of many free parameters and problems , which can be mapped on the classical ising model @xcite .
the way to achieve the solution is by the algorithm of quantum annealing : slow variation of external experimental parameters , so that , as the adiabatic theorem says , the system develops with time adiabatically in its energetically ground state . at the end of the quantum annealing time the d - wave presents the solution of the optimization problem . in this sense
the d - wave machine is different from all other attempts to construct a quantum computer .
+ the expectation that quantum computation will be faster than classical is not confirmed , but nevertheless there is sound proof that the d - wave machine operates as a quantum computer .
numerous evidences , ranging from spectoscopic data @xcite to theoretical simulations @xcite , give sound basis to accept its quantum nature .
this macroscopic construction , consisting of hundred to more than 500 flux qubits on a chip , interacting with each other and subjected to transverse magnetic fields , behaves coherently according to schrdinger s time dependent equation and performs coherently quantum operations over long time of seconds and even minutes .
+ the experimental observation is that a single flux qubit , the building block of the d - wave device , a macroscopic object with @xmath3 cooper pairs and dimensions of the order of @xmath4 , loses coherence within nanoseconds in the photon field in ramsey interference experiments @xcite .
( more recent experiments with single flux qubits show coherence time of the order of microseconds @xcite , however they are not comparable with the technique used by the d - wave . ) a single flux qubit , consisting of a supercunducting loop ( nb or al ) , intercepted by the josephson junction ( al oxide ) , is a two - state system with one state the persistent current running clockwise and the other state with the persistent current anticlockwise . in an external transverse magnetic field
the two current states are no more degenerate .
the illumination with short ( ns ) near resonant photon pulse in the radiofrequency range is followed by oscillations between the two current states , called ramsey fringes , as the qubit is left to develop free with time .
the experimental observation is : the probability for switching between the two current states decays with time and reaches a finite value for the excitation probability larger than zero .
this observation is interpreted as loss of coherence and the suggested causes for the damping of the amplitude of the oscillations range from experimental imperfections and poor control of the electromagnetic environment of the flux qubit , to external and internal noises of different nature and the involvement of various dissipative channels ( charge noise , photon noise , quasiparticle excitations above the superconducting gap , radiative and dielectric loss , etc . ) .
+ the boltzmann factor , with the experimental excitation energy 6.6 ghz and the experimental temperature in the millikelvin range ( 25 mk ) , is of the order of @xmath5 .
hence , the oscillations in the ramsey interference experiments can not be due to thermal excitation .
stochastic approaches , providing quantum master equation for the rate of excitation of a two - level system in an environment of quantum harmonic oscillators , yield an exponential decay of the excitation probability . using the parameters from the experiments , zero limiting value of the excitation probability and no oscillatory behaviour with time is the result @xcite .
this is in complete contrast with the experimental observations . in a calculation based on a quantum stochastics
approach the experimental observation of non - zero final excitation probability can be reproduced , however , using much higher temperature than the experimental temperature .
+ the observation in an experiment with a single flux qubit , where its time development is not free but occurs in the presence of permanent illumination with radiofrequency laser , shows rabi oscillations of the population of the first excited state , driven by the continuous laser illumination of variable duration and _ immediately _ followed by the measurement with the readout pulse ( ref .
@xcite , p. 56 , fig . 3.12 ) .
( this is in contrast to the ramsey interference experiment where radiofrequency laser pulses of nanosecond duration are used , followed by a time interval of free development of the flux qubit . )
the exponential decay of the amplitude of the rabi oscillations , achieving a constant non - zero value at time of the order of 80 ns , is attributed to decoherence due to dynamic low - frequency noise , which is treated within a quantum stochastics approach and a quantum master equation .
this approach is most often used to simulate non - specified noise due to an environment . however , a much higher temperature of the environment , compared with the experimental temperature of the order of mk , has been used to reproduce the decay of the rabi oscillations of a single flux qubit @xcite . + a non - oscillatory exponential decay of the occupation of the excitated state to zero value has indeed been observed with a single flux qubit ( ref .
@xcite , p. 50
3.10 ) illuminated with long photon pulses as a function of the delay time after the rf pulse .
this experimental procedure , using long time illumination , allows to establish the equilibrium occupation of the excitated state and to follow its time development .
as it is expected , if boltzmann statistics dominates the distribution over the states of the flux qubit , the probability for occupation of the excited state decays exponentially without oscillations , approaching the value zero at delay time of the order of 80 ns .
this trend can be fitted to bloch - redfield like master equation for a two - level system in thermal equilibrium with the electromagnetic modes .
+ decoherence theory would claim that the oscillations in the ramsey interference experiment with the single flux qubit are the result of collapse in one or the other current states .
but then , if collapse in a single flux qubit occurs on the time scale of 20 ns , how can the collapses in an interacting ensemble of five hundred flux qubits be avoided to enable the coherent operation of the d - wave quantum computer ?
+ we must conclude that quantum stochastic approaches and decoherence theory can not provide the explanation of the long coherence time of the d - wave machine , being unable to describe the time development of a single flux qubit .
what the stochastic approach provides for a single flux qubit is simply a statistical boltzmann behaviour at temperature significantly higher than that used in the experiments .
+ the problem we focus on is the discrepancy of 8 orders of magnitude in the coherence time of a single flux qubit and the coherent operation of the d - wave machine consisting of many flux qubits . we solve the time dependent schrdinger equation with a hamiltonian including the entanglement of the current states to gravonons , existing in high dimensional spacetime .
the quantum theory lets copenhagen quantum mechanics emerge , without postulating collapse , and provides a deterministic definite outcome of a
, hence the name emerging quantum mechanics @xcite .
it has been successfully used to solve the problem of localization of quantum particles and their appearance as classical particles @xcite , the problem of quantum diffusion of atoms on solid surfaces @xcite in a telegraph - signal like dynamics @xcite and wave - particle duality in matter wave diffraction @xcite .
+ within the theory of emerging quantum mechanics we explain the apparent short coherence time of 20 ns of a single flux qubit ( section [ hamiltonian1qubit ] ) and the quantum behaviour of the d - wave machine ( section [ hamiltonian4qubit ] ) over a period of time 8 orders of magnitude longer than that of a single flux qubit .
a tentative suggestion is made to understand the temperature dependence of the success probability of the d - wave computer .
success probability is the probability to achieve the correct final result of the posed optimization problem .
experiment with a 16 flux qubit d - wave machine shows a higher success probability at temperature 20 - 100 mk than at lower temperature @xcite , which is in total contrast with expectations for the temperature dependence of the performance of quantum computers .
the environmental excitations are the gravonons , massive bosons which emerge in the limit of weak and local gravitational interaction in high dimensional spacetime ( 11d ) @xcite .
the coupling to gravonons is effective only within spacetime deformations called warp resonances .
the model for the non - perturbed gravonons @xmath6 is a harmonic oscillator , whereas the gravonons perturbed by matter and force fields @xmath7 are the solution of the @xmath8 model described in ref .
+ gravonons are bosons and every boson can be treated mathematically as a harmonic oscillator with an effective parabolic potential corresponding to a generalized coordinate describing collective motion . in 11 dimesional spacetime
this is a 10d parabolic protential with eigenstates in 10 spacial dimensions .
the model for the non - perturbed gravonons @xmath6 is a harmonic oscillator , whereas the gravonons perturbed by matter and force fields @xmath7 are the solution of the @xmath8 model described in @xcite .
the effect of entanglement with the matter fields means excitation from the non - perturbed gravonon states @xmath6 in @xmath7 and deexcitation from @xmath7 into @xmath6 , i.e. it is a scattering induced within the gravonon continuum .
+ the hamiltonian for the single flux qubit plus photon is : @xmath9 \nonumber \\ & + & \omega_{phot}b_{phot}^+b_{phot}+\sum_{j=1}^2\left [ \varepsilon_{grav_j}c_{grav_j}^+c_{grav_j } + \sum_{\kappa_j } \varepsilon_{\kappa_j}c_{\kappa_j}^+c_{\kappa_j}\right ] \nonumber \\ & + & v_{phot , qubit}(a_{qubit_1}^+a_{qubit_2}+a_{qubit_2}^+a_{qubit_1})(b_{phot}^++b_{phot } ) \nonumber \\ & + & \sum_{j=1}^2 \sum_{\kappa_j}\left [ w_{grav_j , w_j}n_{w_j}c_{grav_j}^+c_{\kappa_j}+w_{w_j , grav_j}n_{w_j } c_{\kappa_j}^+c_{grav_j}\right].\end{aligned}\ ] ] the meaning of the symbols is @xmath10 : regions in the josephson junction ( warp resonances ) , where cooper pairs couple to gravonons ; @xmath11 : creation , annihilation and number operator for persistent current in the state @xmath12 in the josephson junction ; @xmath13 : creation and annihilation operator for persistent current in the state @xmath14 in the flux qubit ; @xmath15 : creation and annihilation operator for the photon ; @xmath16 : creation and annihilation operator for the lowest energy gravonon in the potential , which is not perturbed by interaction between the matter field and the photon field ; @xmath17 : creation and annihilation operator for gravonons in the deformed manifold . + the terms in the hamiltonian eq .
( [ hamiltsinglequbit ] ) describe : * the two current states in the qubit ( first term in the second line of eq .
[ hamiltsinglequbit ] ) , * in regions of the josephson junction where the cooper pairs couple to the gravonons @xmath10 ( second term in the second line of eq . [ hamiltsinglequbit ] ) * and their interaction @xmath18 ( third term in the second line of eq . [ hamiltsinglequbit ] ) .
* term , describing the photon field ( first term in the third line of eq .
[ hamiltsinglequbit ] ) , * the gravonons in the initial wave packet ( second term in the third line of eq .
[ hamiltsinglequbit ] ) , as well as excited gravonons due to interactions with the supercurrents in the josephson junction ( third term in the third line of eq .
[ hamiltsinglequbit ] ) . *
the interaction of the supercurrent with the photon is the standard coupling between a photon and matter , i.e. the common dipole coupling term ( fourth line of eq . [ hamiltsinglequbit ] ) . * the last term in eq .
( [ hamiltsinglequbit ] ) is the coupling between gravonons and the supercurrents of the quadrupole coupling kind .
the time dependent schrdinger equation is solved for the time development of the total wave functional : @xmath19 which is represented in the basis of field configurations .
these are tensor products of the field states of the supercurrent , the photon field and the gravonon field . in our theory
the interaction with the environment is included in the hamiltonian , _ the environment is accounted for , when we solve the time dependent schrdinger equation_. in contrast , stochastic approaches , starting from the liouville equation for the dynamics of the density matrix with the total hamiltonian of the system plus environment , revert to an equation where the system alone is included in the hamiltonian , whereas the effects of the environment are in the lindblad term @xcite .
+ the fields in the system are the persistent current of the qubit @xmath20 , the photon field @xmath21 , the gravonon field comprising @xmath22 , @xmath23 and @xmath24 .
the notation for the field configurations is @xmath25 , for instance @xmath26 for the initial wave packet , @xmath27 for the switched current configuration plus photon , etc . according to the assumed principle that only these warped field configurations which arise from the flat field configurations ( non - warped configurations ) via first order transitions are taken into account , only configurations of the kind @xmath28 and @xmath29
are involved . + the physical nature of the gravonons as massive bosons arising from gravitons , as well as the derivation of an effective schrdinger equation in high dimensional spacetime _ yielding gravonons similar to the common quantum particles as its solution
_ , are described in detail in ref .
the gravonons are the only quanta which reside not only in four dimensional spacetime , but in the additional compactified hidden spacial dimensions .
the decisive features of the gravonons are the high density of the gravonon quanta and weak coupling to the matter fields .
+ why do we need gravonons at all energies for the single flux qubit - plus - photon calculation and not in the case of matter wave diffraction experiment ( cf .
@xcite ) ? in the single flux qubit - plus - photon calculation we need gravonon bands at different energies ( off - the - shell of the initial configuration ) because due to the local interactions the local configurations are split into states with higher and lower energy relative to that of the initial field configuration , and the states in the three dimensional subsystem are no more degenerate ( left panel in fig . [ spectrdistr1000 ] ) .
the solution of the time dependent schrdinger equation gives the time dependent total wave functional as a superposition of the basis configurations , varying with time .
the initial wave packet in interaction with the rest configurations , but not with the gravonons , splits into two components below and above its energy , corresponding mostly to the two current configurations , clockwise and anticlockwise . with the entanglement with the gravonons
they acquire finite broadening and shifts as it is shown in the left panel of fig .
[ spectrdistr1000 ] .
each group of eigenstates has entangled to the nearly degenerate part of the gravonon continuum .
oscillations between the eigenstates below and above the energy of the initial wave packet occur in the absence of entanglement with the gravonons .
this is the situation for short times . since the total wave functional
has not yet acquired the components with the excited gravonons , it can be written as : @xmath30 however , the entanglement with the gravonons develops with time .
the weights of warped field configurations with the persistent current clockwise , as well as those with the persistent current in the anticlockwise direction , increase with time : @xmath31 at the start of the time development just the terms in eq .
( [ wavefnctnograv ] ) dominate the total wave functional which explains the oscillations between the two current configurations . at @xmath32 ns , when the entanglement to the gravonons has fully developed , only the warped field configurations in the total wave functional eq . ( [ wavefnctgrav ] ) contribute , i.e. @xmath33 tends to zero .
they saturate at some final and finite values , i.e. the coefficients in the expansion of @xmath34 in the basis field configurations do not change with time any more . at 50 ns current clockwise and
current anticlockwise are completely entangled with the gravonons .
while the field configurations with excited gravonon components dominate the total wave functional the persistent currents can not switch direction .
the reason not to change with time is that , due to entanglement with the gravonons , the components and their contributions to the total superimposed quantum configuration can not change while the gravonons are in the hidden spacial dimensions .
recurrence of the gravonons back in three dimensional space , disentanglement from the near degenerate gravonon band the local current configuration is originally entangled with , and entanglement with the gravonon band , nearly degenerate with the current configuration of opposite direction , are needed to change the current configuration in the flux qubit .
as these local field configurations are not energetically degenerate , the change of current requires energy .
+ at @xmath32 ns the wave functional is still a superposition of many basis field configurations , but the expansion coefficients do not change with time .
finally we have a schrdinger cat configuration , a superposition of field configurations , and not a tensor product , which , however , does not change with time as it can be seen in fig . [ spectrdistr1000 ] ( red curve in the right panel ) . *
the single flux qubit is in a superposition of the two persistent current configurations .
* in the time before the freezing of the current distribution in the flux qubit occurs , of course , the oscillations between the two current configurations are still discernible as it is shown by the red curve in the right panel of fig .
[ spectrdistr1000 ] , though with damped amplitude . at time @xmath32 ns a single flux qubit is frozen in a configuration in which the relative contributions of current clockwise and current anticlockwise in the superposition do not change . at that time , according to our theory , the flux qubit is maximally entangled with the gravonon environment in a coherent state . it should be stressed that in our theory we do not have a new decoherence mechanism due to interaction with the gravonons , the gravonons are the environment which warrants the coherent development of the flux qubit , even suppressing the purturbing effect of other environmental excitations in three dimensional space , as it is shown in section [ gravononphonon ] . + the theoretical result in the right panel of fig .
[ spectrdistr1000 ] is in satisfactory agreement with the experimental ramsey fringes in ref .
@xcite ( fig .
this result might be wrongly associated with what is called collapse in the copenhagen interpretation of quantum mechanics .
we call it choice
. it can not be regarded as collapse since the total wave functional is not a tensor product , it is a superposition of the two current configurations . in the theory of emerging quantum mechanics , in contrast to copenhagen quantum mechanics ,
the choice is a solution of the deterministic time dependent schrdinger equation , whereas collapse in copenhagen quantum mechanics is postulated as random statistically occurring , and it is not the result of any well defined dynamics . at time 50 ns and beyond the flux qubit
is not described by a statistical distribution over the two current configurations , as the stochastic approach yields @xcite .
the single flux qubit is not collapsed in a classical state .
it is in a quantum superposition of the two current configurations , it is in a coherent superposition of field configurations which , however , does not change with time .
+ the reason for the observed decay of the amplitude of current switching in ramsey interferometry in a single flux qubit is not decoherence , which implies entanglement to environmental excitations in four dimensional spacetime and relies on the development of the reduced density matrix into quasi - diagonal form .
the exponential damping of the amplitude of the ramsey fringes with time is due to the development of the entanglement of the local current configurations+photon with the gravonons leading to a coherent quantum configuration which , though it represents a quantum superimposition of the two current configurations , can not change with time as long as the entanglement with the gravonons is effective . in this case a reduced density matrix ( which is not needed in the theory of emerging quantum mechanics ) need not acquire the diagonal form , required by decoherence theory and interpreted with born s probabilistic concept .
+ this is a rather astonishing result as the flux qubit is a macroscopic object which does not collapse in a classical state but retains its quantum behaviour for all the time preceding recurrence of the gravonons in three dimensional space .
but it is not a unexpected result .
how can otherwise the d - wave machine operate as a quantum computer , if the single flux qubits it consists of lose coherence and collapse in classical states in the course of the operation ? +
the theoretical model of the d - wave machine consists of four flux qubits as it is shown in the left panel of fig .
[ model4qubits ] , interacting with one another and subjected to external transverse magnetic fields which vary with time .
the hamiltonian for quantum annealing @xmath35 is now time dependent since the transverse external magnetic fields on the flux qubits vary with time : @xmath36 four terms describing the non - interacting flux qubits are taken from the hamiltonian in eq .
( [ hamiltsinglequbit ] ) , however , exempting the interaction with the photon field in the fourth line of this equation .
gravonon coupling is introduced only within qubit ( the last term in eq . [ hamilt4qubits ] ) .
@xmath35 includes terms describing the interaction between the flux qubits : @xmath37 and the interaction with the external magnetic fields : @xmath38 @xmath39 and @xmath40 are the pauli matrices .
the transverse magnetic fields are the ones which can switch the direction of the supercurrent in a given flux qubit and drive the superconducting current in a superposion of currents in two opposite directions .
the manipulation with time of the magnetic fields represents the quantum annealing procedure .
the aim is by slowly reducing the external transverse magnetic fields @xmath41 and , hence , the contribution in the hamiltonian of the additional term to the non - perturbed ising hamiltonian , to ensure an adiabatic time development of the system towards the ground state of the ising hamiltonian , which is the solution of the optimization problem .
+ the pauli matrices @xmath39 and @xmath40 act on the current states @xmath42 and @xmath43 : @xmath44 @xmath45 @xmath46 @xmath47 the time dependent schrdinger equation is solved with a time dependent hamiltonian : @xmath48\psi(0).\ ] ] the computational basis consists of 16 flat field configurations denoted by 4 labels , which indicate the state of each qubit by 1 for current in the clockwise direction and -1 for current in the anticlockwise direction : @xmath49 they represent eigenstates of the ising hamiltonian . the notation for the gravonons are omitted , being in all cases the ground configurations @xmath50 with gravonons in one qubit only . thus without gravonons we have a @xmath51 matrix of configurations to diagonalize to get the eigenenergies of the 4 flux qubit system
. the lowest eigenenergies and their variation with time can be seen in the right panel of fig .
[ model4qubits ] .
+ at @xmath52 the effect of @xmath53 is to mix in all computational basis configurations , leading to the ground configuration of the 4 flux qubit system in the presence of the transverse magnetic fields . at the annealing time @xmath54 the external local transverse magnetic fields are switched off , i.e. @xmath55 and the 4 flux qubit system provides the solution of the ising hamiltonian eq .
( [ hamilt4qubits ] ) with @xmath56 .
the development with time will be adiabatic if @xmath57 change slowly with time , leaving the system in the adiabatic time dependent global ground configuration , as it is the case shown in the right panel of fig .
[ model4qubits ] .
if @xmath57 change fast with time the 4 flux qubit system may develop diabatically with time with landau - zener transitions in excited eigenstates destroying the adiabaticity .
this development is avoided by suitable choice of the time variation of the external transverse magnetic fields in the experiment as well as in the theory .
+ the ising states , solution of the hamiltonian at the end of the quantum annealing , superimpose to build the eigenstates of the hamiltonian with the transverse magnetic fields .
the eigenstates possess sharp energies until the entanglement with the gravonons is included . and according to the adiabatic theorem the 4 flux qubit system , if it starts in the ground configuration , will develop with time in the global ground configuration . in experiment as well as in our theory
an artificial ground configuration is prepared by the choice of the transverse magnetic fields .
however , the entanglement with the gravonons may eventually destroy the adiabatic time development .
is it so indeed ?
with entanglement to gravonons the eigenstates of the 4 flux qubit system broaden into resonances within the gravonon continuum they couple to .
as the coupling to the gravonons is very weak and local , the broadening is orders of magnitude smaller than the energy differences between the lowest eigenstates of the 4 flux qubit system even at the time of avoided crossing ( cf .
the inset in the right panel of fig .
[ model4qubits ] showing the gap between the ground state and the first excited state configurations ) .
hence , the energy broadening , due to entanglement to gravonons , can not lead to superpositions of the eigenstates of the 4 flux qubit system , which would destroy the adiabaticity .
+ in analogy with the single flux qubit case , while entangled with the gravonons each flux qubit is frozen after 50 ns in a configuration , in which the relative contributions of the two current states in this flux qubit does not change with time any more .
each flux qubit is stabilized in this superposition , which is typical for the global ground state configuration .
hence , the global ground state of the d - wave machine can not change either . once in the ground configuration
, the system of many flux qubits will stay in the ground configuration for ever .
entanglement to gravonons stabilizes the global ground state configuration and helps the coherent adiabatic time development of the d - wave machine towards the final solution . + thermal excitations can not be avoided even at the millikelvin temperatures at which the d - wave machine operates .
it was experimentally demonstrated in ref .
@xcite for a 16 qubit d - wave machine that the success probability is temperature dependent : at temperature in the range 20 - 100 mk it is higher than at lower temperature , in contrast to other attempts for quantum computing .
+ let a phonon be excited in just one of the flux qubits , e.g. the third flux qubit in fig .
[ model4qubits ] .
terms are added to the quantum annealing hamiltonian @xmath35 eq .
( [ hamilt4qubits ] ) for the phonon and its interaction with the supercurrent of the kind : @xmath58 the term @xmath59 is analogous to the jaynes - cummings hamiltonian of a two - level atomic system , coupling to the phonon field with coupling strength @xmath60 . @xmath61 and @xmath62
are creation and annihilation operators for the phonon mode and @xmath63 is the phonon frequency . with the definition of @xmath59 in eq .
( [ hamilt - phon ] ) the effect of the phonons is to account just for switching the supercurrent in the third flux qubit .
+ the result of a sudden perturbation of the 4 flux qubit system by a single phonon mode is destroyed adiabaticity .
figure [ diab - v ] shows that a fast entanglement to the phonon field around @xmath64 ms perturbs the distribution of the supercurrents in the third flux qubit , which is typical for the global ground configuration of the 4 flux qubit system .
dramatic and nonrepairable arbitrary switches between the currents in the flux qubit occur which means that the global ground configuration of the 4 flux qubit system is changed , hence its adiabatic time development is destroyed and can not be recovered . the coupling to the phonon leads to destroyed adiabatic time development in the global ground configuration of the 4 flux qubit system , as it can be seen in the plot on the left panel in fig .
[ diab - v ] .
the quantity on the vertical axis is the number of the eigenstate of the non - perturbed 4 qubit system and not its energy .
the areas of the points in the plot scale with the weight of the eigenstates of the 4 flux qubit system , which get involved in the time dependent wave packet @xmath34 . when the perturbation by the phonon is switched on excited field configurations of the 4 flux qubit system gain weight and a significant redistribution from the global ground configuration over several excited configurations occurs .
thus , coupling to a phonon mode destroys the adiabatic time development of the system in an irrepairable way . in the josephson junction
the tunnelling currents entangle with gravonons in high spacial dimensions .
the result is suppression of the effect of the phonons , and retained coherence of the system in the adiabatic ground configuration , as it can be seen in fig .
[ adiabatic - vph - vgrav ] .
the length of the symbols in the plot scales with the weight of the initial wave packet in the gravonon band , i.e. it is the spectral distribution of the initial wave packet in the gravonon continuum .
the entanglement with the gravonons quenches the transitions between the low lying eigenstates of the 4 flux qubit system due to the perturbation by the phonon .
the spectral distribution of the initial wave packet in the gravonon continuum is not affected by the excitation of the phonon in one of the flux qubits . as a consequence
the many - qubit ground configuration is not modified by the interaction with the phonons .
the 4 flux qubit system develops adiabatically in the global ground configuration for long time significantly exceding the life time of the phonon .
the gravonons suppress the effect of the phonon . while being entangled with the gravonons in the hidden dimensions , the current can not switch direction .
entanglement to gravonons stabilizes the adiabatic ground configuration of the d - wave machine .
+ the noteworthy observation for the d - wave machine is that it achieves higher success probability at slightly higher temperature compared to lower temperature @xcite .
the temperature , in contrast to expectations , improves the performance of the d - wave machine .
higher temperature means phonon excitations , leading to current flips .
this experiment shows , however , that the d - wave quantum computer functions better when it is slightly warmed up .
+ our results also show that the phonons , i.e. temperature can destroy the d - wave coherence by flipping the current direction in one flux qubit ( cf . section [ phononspinflip ] ) .
but phonon excitation means enhanced atom vibrations . we know from the dynamics of chemical reactions that atom vibrations , i.e. temperature increase , are beneficial for the attenuation of the effective activation barriers for chemical reactions .
similar argument can be used to explain the effect of temperature on the effective barrier height for tunnelling from the initial configuration into the global energy minimum configuration of the d - wave machine .
the barrier heights for state tunnelling between different minima depend upon the atom vibrations and , hence , upon the temperature . for small temperature increase
the effective barrier heights can get lower and then the success probability will get higher for shorter annealing times .
of course , the reverse effect of destroyed superconductivity as the temperature further increases , leads to the attenuation of the success probability .
the present study of a quantum computer consisting of numerous flux qubits uses emerging quantum mechanics , a method based on schrdinger s time dependent quantum mechanics which accounts for the entanglement of fields in four dimensional spacetime with gravonons living in high dimensional spacetime @xcite .
the astonishing result is that a single flux qubit and the d - wave computer , both macroscopic objects , retain coherence and behave as quantum objects according to the laws of quantum mechanics with long coherence time of the order of minutes .
the clue to this result is the entanglement of the persistent currents in the flux qubits with the gravonons , the massive quanta of the gravitational field , which live in high spacial dimensions .
the necessary condition for this result is weak coupling of the local quantum fields to an environmental continuum of high density of states , which is satisfied only by the gravonons in high spacial dimensions .
+ the major results of the present study of the d - wave quantum computer can be summarized : * the coherence time observed experimentally of a single flux qubit is explained , obtaining acceptable agreement with experiment .
the explanation is based on entanglement with the gravonon continuum . *
the entanglement to gravonons also explains the coherence time of the d - wave quantum computer . * the entanglement to gravonons also explains why phonons do not perturb the quantum annealing process of the d - wave machine and it lets the temperature dependence of the success probability of the d - wave quantum computer appear comprehensible .
these results as well as the experimental observations on the d - wave quantum computer can not be reproduced and explained either by stochastic quantum approaches or within conventional decoherence theory in four dimensions .
this has , however , not been explicitely demonstrated in the paper so far .
+ emerging quantum mechanics reproduces experimental observations of very different nature , where orders of magnitude discrepancies have been found by computations based on conventional quantum mechanics in three dimensional space .
they include quantum diffusion of adsorbates on solid surfaces @xcite , and double slit diffraction experiments with massive molecules @xcite . in these cases localization of quantum particles in three dimensional space
is the result of the entanglement with the gravonons which live in high spacetime dimensions .
999 chiorescu i , nakamura y , harmans c j p m and mooij j e 2003 _ science _ * 299 * 1869 johnson m w et al . 2011 _ nature _ * 473 * 194 dickson n g et al .
2013 nature communications 4:1903 , doi : 10.1038/ncomms2920 boixo s , rnnow t f , isakov v , wang z , wecker d , lidar d a , martinis j m and troyer m 2014 _ nature physics _ * 10 * 218 ; arxiv : 1304.4595v2 [ quant - ph ] 21 jul 2013 www.dwavesys.com ising i 1924 _ contribution to the theory of ferromagnetism _ phd thesis hamburg lanting t et al .
2014 _ phys rev x _ * 4 * 021041 ; arxiv : 1401.3500 [ quant - ph ] 15 jan 2014 wang l , rnnow t f , boixo s , isakov s v , wang z , wecker d , lidar d a , martinis j m and troyer m 2013 arxiv : 1305.5837v1 [ quant - ph ] 24 may albash t , vinci w , mishra a , warburton p a , lidar d a 2014 arxiv : 1403.4228v3 [ quant - ph ] 18 jul bylander j , gustavsson s , yan f , yoshihara f , harrabi k , fitch g , cory d g , nakamura y , tsai j - s and oliver w d 2011 _ nature physics _ * 7 * 565 stern m , catelani g , kubo y , greezes c , bienfait a , vion d , esteve d and bertet p 2014 _ phys .
lett . _ * 113 * 123601 caldeira a o and leggett a j 1981 _ phys .
* 46 * 211 ; caldeira a o and leggett a j 1983 _ ann .
ny _ * 149 * 374 ; leggett aj , chakravarty s , dorsey at , fisher mpa , garg a and zwerger w 1987 _ rev
phys . _ * 59 * 1 cidrim a , dos santos fe and caldeira a o 2014 _ phys rev _ a * 90 * 052102 longobardi l 2010 _ studies of quantum transitions of magnetic flux in a rf squid qubit _ lap lambert ac . publ .
gmbh , saarbrcken doyen g and drakova d 2014 arxiv : 1408.2716 [ quant - ph ] 12 aug doyen g and drakova d 2011 _ j. physics : conf . ser . _ * 306 * 012033 drakova d and doyen g 2011 _ j. physics : conf
. ser . _ * 306 * 012068 drakova d and doyen g 2013 _ j. physics : conf
. ser . _ * 442 * 012049 drakova d and doyen g 2012 arxiv : 1204.5606 [ quant - ph ] 25 apr doyen g and drakova d 2013 _ j. physics : conf .
ser . _ * 442 * 012032 doyen g and drakova d presented at dice2014 , submitted for publication in j. phys .
lindblad g 1976 _ commun .
phys . _ * 48 * 119 | extensive experiments have demonstrated quantum behaviour in the long - time operation of the d - wave quantum computer .
the decoherence time of a single flux qubit is reported to be on the order of nanoseconds @xcite , which is much shorter than the time required to carry out a computation on the timescale of seconds @xcite .
previous judgements of whether the d - wave device should be thought of as a quantum computer have been based on correlations of the input - output behaviour of the d - wave machine with a quantum model , called simulated quantum annealing , or classical models , called simulated annealing and classical spin dynamics @xcite .
explanations for a factor of @xmath0 discrepancy between the single flux qubit decoherence time and the long - time coherent quantum behaviour of many integrated flux qubits of the d - wave device have not been offered so far . in our contribution
we investigate a model of four qubits with one qubit coupled to a phonon and ( optionally ) to environmental particles of high density of states , called gravonons .
the calculations indicate that when no gravonons are present , the current in the qubit is flipped at some time and adiabatic evolution is discontinued .
the time dependent wave functional becomes a non - correctable superposition of many excited states .
the results demonstrate the possibility of effectively suppressing the current flip and allowing for continued adiabatic evolution when the entanglement to gravonons is included .
this adiabatic evolution is , however , a coherent evolution in high dimensional spacetime and can not be understood as a solution of schrdinger s time dependent equation in four dimensional spacetime .
compared to schrdinger s time development , the evolution is considerably slowed down , though still adiabatic .
the properties of our model reflect correctly the experimentally found behaviour of the d - wave machine and explain the factor of @xmath0 discrepancy between decoherence time and quantum computation time .
the observation and our explanation are in anology to the @xmath0 discrepancy factor found , when comparing experimental results on adsorbate quantum diffusion rate with predictions of schrdinger s time dependent equation , which can also be resolved in a model with the coupling to gravonons included .
+ d. drakova@xmath1 and g. doyen@xmath2 + @xmath1 faculty of chemistry , university of sofia , bulgaria + url : [email protected] + @xmath2 ludwig - maximilians universitt , mnchen , germany + url : [email protected] |
let @xmath0 be the set of all simple graphs with @xmath1 vertices and @xmath2 edges and let @xmath3 denote the sum of the squares of the degrees , @xmath4 , of the vertices of @xmath5 .
the purpose of this paper is to finish the solution of an old problem : 1 . what is the maximum value of @xmath6 , for a graph @xmath5 in @xmath0 ?
2 . for which graphs @xmath5 in @xmath0
is the maximum value of @xmath6 attained ? throughout , we say that a graph @xmath5 is _ optimal _ in @xmath0 , if @xmath6 is maximum and we denote this maximum value by @xmath9 .
these problems were first investigated by katz @xcite in 1971 and by r. ahlswede and g.o.h .
katona @xcite in 1978 . in his review of the paper by ahlswede and katona .
p. erds @xcite commented that
the solution is more difficult than one would expect . "
ahlswede and katona were interested in an equivalent form of the problem : they wanted to find the maximum number of pairs of different edges that have a common vertex .
in other words , they want to maximize the number of edges in the line graph @xmath10 as @xmath5 ranges over @xmath0 .
that these two formulations of the problem are equivalent follows from an examination of the vertex - edge incidence matrix @xmath11 for a graph @xmath7 : @xmath12 where @xmath13 is the adjacency matrix of the line graph of @xmath5 .
thus @xmath14 .
( @xmath15 is twice the number of edges in the line graph of @xmath5 . )
ahlswede and katona showed that the maximum value @xmath9 is always attained at one or both of two special graphs in @xmath0 .
they called the first of the two special graphs a _
quasi - complete _ graph .
the quasi - complete graph in @xmath0 has the largest possible complete subgraph @xmath16 .
let @xmath17 be the unique integers such that @xmath18 the quasi - complete graph in @xmath0 , which is denoted by @xmath19 , is obtained from the complete graph on the @xmath20 vertices @xmath21 by adding @xmath22 vertices @xmath23 , and the edges @xmath24 , @xmath25 .
the other special graph in @xmath0 is the _ quasi - star _ , which we denote by @xmath26 .
this graph has as many dominant vertices as possible .
( a _ dominant vertex _ is one with maximum degree @xmath27 . )
perhaps the easiest way to describe @xmath26 is to say that it is the graph complement of @xmath28 , where @xmath29 .
define the function @xmath30 to be the sum of the squares of the degree sequence of the quasi - complete graph in @xmath0 , and define @xmath31 to be the sum of the squares of the degree sequence of the quasi - star graph in @xmath0 .
the value of @xmath30 can be computed as follows : let @xmath32 , with @xmath33 .
the degree sequence of the quasi - complete graph in @xmath0 is @xmath34 hence @xmath35 since @xmath26 is the complement of @xmath28 , it is straightforward to show that @xmath36 from which it follows that , for fixed @xmath1 , the function @xmath37 is point - symmetric about the middle of the interval @xmath38 .
in other words , @xmath39 it also follows from equation ( [ eqn : scconnection ] ) that @xmath19 is optimal in @xmath0 if and only if @xmath40 is optimal in @xmath41 .
this allows us to restrict our attention to values of @xmath2 in the interval @xmath42 $ ] or equivalently the interval @xmath43 $ ] . on occasion , we will do so but we will always state results for all values of @xmath2 . as the midpoint of the range of values for @xmath2 plays a recurring role in what follows ,
we denote it by @xmath44 and define @xmath45 to be the integer such that @xmath46 to state the results of @xcite we need one more notion , that of the distance from @xmath47 to @xmath48 .
write @xmath49 we are now ready to summarize the results of @xcite : theorem 2 @xcite : : @xmath9 is the larger of the two values @xmath30 and @xmath31 .
theorem 3 @xcite : : @xmath50 if @xmath51 and @xmath52 if @xmath53 lemma 8 @xcite : : if @xmath54 , or @xmath55 , then @xmath56 if @xmath57 and @xmath58 , then there exists an @xmath59 with @xmath60 such that @xmath61 ahlswede and katona pose some open questions at the end of @xcite .
some strange number - theoretic combinatorial questions arise .
what is the relative density of the numbers @xmath1 for which @xmath62 [ @xmath50 for all @xmath63 and @xmath52 for all @xmath64 ? " this is the point of departure for our paper .
our first main result , theorem [ thm : main1 ] , strengthens ahlswede and katona s theorem 2 ; not only does the maximum value of @xmath6 occur at either the quasi - star or quasi - complete graph in @xmath0 , but all optimal graphs in @xmath0 are related to the quasi - star or quasi - complete graphs via their so - called diagonal sequence . as a result of their relationship to the quasi - star and quasi - complete graphs , all optimal graphs can be and are described in our second main result , theorem [ thm : main2 ] .
our third main result , theorem [ thm : main3 ] , is a refinement of lemma 8 in @xcite .
theorem [ thm : main3 ] characterizes the values of @xmath1 and @xmath2 for which @xmath65 and gives an explicit expression for the value @xmath59 in lemma 8 of @xcite .
finally , the strange number - theoretic combinatorial " aspects of the problem , mentioned by ahlswede and katona , turn out to be pell s equation @xmath66 .
corollary [ cor : density ] answers the density question posed by ahlswede and katona .
we have just recently learned that wagner and wang @xcite have independently answered this question as well .
their approach is similar to ours , as they also find an expression for @xmath59 in lemma 8 of @xcite . before stating the new results , we summarize work on the problem that followed @xcite .
a generalization of the problem of maximizing the sum of the squares of the degree sequence was investigated by katz @xcite in 1971 and r. aharoni @xcite in 1980 .
katz s problem was to maximize the sum of the elements in @xmath67 , where @xmath68 runs over all @xmath69-square matrices of size @xmath70 with precisely @xmath71 ones .
he found the maxima and the matrices for which the maxima are attained for the special cases where there are @xmath72 ones or where there are @xmath73 ones in the @xmath69-matrix .
aharoni @xcite extended katz s results for general @xmath71 and showed that the maximum is achieved at one of four possible forms for @xmath68 .
if @xmath68 is a symmetric @xmath69-matrix , with zeros on the diagonal , then @xmath68 is the adjacency matrix @xmath74 for a graph @xmath5 .
now let @xmath5 be a graph in @xmath0 .
then the adjacency matrix @xmath74 of @xmath5 is a @xmath75 @xmath69-matrix with @xmath76 ones . but
@xmath74 satisfies two additional restrictions : @xmath74 is symmetric , and all diagonal entries are zero .
however , the sum of all entries in @xmath77 is precisely @xmath78 . thus our problem is essentially the same as aharoni s in that
both ask for the maximum of the sum of the elements in @xmath67 .
the graph - theory problem simply restricts the set of @xmath69-matrices to those with @xmath76 ones that are symmetric and have zeros on the diagonal .
olpp @xcite , apparently unaware of the work of ahlswede and katona , reproved the basic result that @xmath79 , but his results are stated in the context of two - colorings of a graph .
he investigates a question of goodman @xcite : maximize the number of monochromatic triangles in a two - coloring of the complete graph with a fixed number of vertices and a fixed number of red edges .
olpp shows that goodman s problem is equivalent to finding the two - coloring that maximizes the sum of squares of the red - degrees of the vertices .
of course , a two - coloring of the complete graph on @xmath1 vertices gives rise to two graphs on @xmath1 vertices : the graph @xmath5 whose edges are colored red , and its complement @xmath80 .
so goodman s problem is to find the maximum value of @xmath6 for @xmath7 .
olpp @xcite shows that either the quasi - star or the quasi - complete graph is optimal in @xmath0 , but he does not discuss which of the two values @xmath81 is larger .
he leaves this question unanswered and he does not attempt to identify all optimal graphs in @xmath0 . in 1999 ,
peled , pedreschi , and sterbini @xcite showed that the only possible graphs for which the maximum value is attained are the so - called threshold graphs .
the main result in @xcite is that all optimal graphs are in one of six classes of threshold graphs .
they end with the remark , further questions suggested by this work are the existence and uniqueness of the [ graphs in @xmath0 ] in each class , and the precise optimality conditions . "
also in 1999 , byer @xcite approached the problem in yet another equivalent context : he studied the maximum number of paths of length two over all graphs in @xmath0 .
every path of length two in @xmath5 represents an edge in the line graph @xmath10 , so this problem is equivalent to studying the graphs that achieve @xmath9 . for each @xmath8
, byer shows that there are at most six graphs in @xmath0 that achieve the maximum .
these maximal graphs come from among six general types of graphs for which there is at most one of each type in @xmath0 .
he also extended his results to the problem of finding the maximum number of monochromatic triangles ( or any other fixed connected graph with 3 edges ) among two - colorings of the complete graph on @xmath1 vertices , where exactly @xmath2 edges are colored red .
however , byer did not discuss how to compute @xmath9 , or how to determine when any of the six graphs is optimal . in section
[ sec : statements of main results ] , we have unified some of the earlier work on this problem by using partitions , threshold graphs , and the idea of a diagonal sequence .
all optimal graphs come from a class of special graphs called _ threshold
_ graphs . the quasi - star and quasi - complete graphs are just two among the many threshold graphs in @xmath0 .
the adjacency matrix of a threshold graph has a special form .
the upper - triangular part of the adjacency matrix of a threshold graph is left justified and the number of zeros in each row of the upper - triangular part of the adjacency matrix does not decrease .
we will show adjacency matrices using @xmath83 " for the main diagonal , an empty circle @xmath84 " for the zero entries , and a black dot , @xmath85 " for the entries equal to one .
for example , the graph @xmath5 whose adjacency matrix is shown in figure [ fig : example643](a ) is a threshold graph in @xmath86 with degree sequence @xmath87 . by looking at the upper - triangular part of the adjacency matrix
, we can associate the distinct partition @xmath88 of 13 with the graph . in general
, the _ threshold _ graph @xmath89 corresponding to a distinct partition @xmath90 of @xmath2 , all of whose parts are less than @xmath1 , is the graph with an adjacency matrix whose upper - triangular part is left - justified and contains @xmath91 ones in row @xmath92 .
thus the threshold graphs in @xmath0 are in one - to - one correspondence with the set of distinct partitions , @xmath93 of @xmath2 with all parts less than @xmath1 : @xmath94 we denote the adjacency matrix of the threshold graph @xmath95 corresponding to the distinct partition @xmath96 by @xmath97 .
peled , pedreschi , and sterbini @xcite showed that all optimal graphs in a graph class @xmath0 must be threshold graphs .
@xcite [ lem : optarethres ] if @xmath5 is an optimal graph in @xmath0 , then @xmath5 is a threshold graph .
thus we can limit the search for optimal graphs to the threshold graphs .
actually , a much larger class of functions , including the power functions , @xmath98 for @xmath99 , on the degrees of a graph are maximized only at threshold graphs .
in fact , every schur convex function of the degrees is maximized only at the threshold graphs .
the reason is that the degree sequences of threshold graphs are maximal with respect to the majorization order among all graphical sequences .
see @xcite for a discussion of majorization and schur convex functions and @xcite for a discussion of the degree sequences of threshold graphs . to state the first main theorem
, we must now digress to describe the diagonal sequence of a threshold graph in the graph class @xmath0 .
returning to the example in figure [ fig : example643](a ) corresponding to the distinct partition @xmath100 , we superimpose diagonal lines on the adjacency matrix @xmath97 for the threshold graph @xmath95 as shown in figure [ fig : example643](b ) .
the number of black dots in the upper triangular part of the adjacency matrix on each of the diagonal lines is called the _ diagonal sequence _ of the partition @xmath96 ( or of the threshold graph @xmath95 ) .
the diagonal sequence for @xmath96 is denoted by @xmath101 and for @xmath102 shown in figure [ fig : example643 ] , @xmath103 .
the value of @xmath104 is determined by the diagonal sequence of @xmath96 .
[ lem : diagonal ] let @xmath96 be a distinct partition in @xmath93 with diagonal sequence @xmath105 . then @xmath104 is the dot product @xmath106 for example , if @xmath102 as in figure [ fig : example643 ] , then @xmath107 which equals the sum of squares of the degree sequence @xmath108 of the graph @xmath95 .
theorem 2 in @xcite guarantees that one ( or both ) of the graphs @xmath109 must be optimal in @xmath0 .
but there may be other optimal graphs in @xmath0 , as the next example shows .
the quasi - complete graph @xmath110 , which corresponds to the distinct partition @xmath111 is optimal in @xmath112 .
the threshold graph @xmath113 , corresponding to the distinct partition @xmath114 is also optimal in @xmath112 , but is neither quasi - star in @xmath112 nor quasi - complete in @xmath115 for any @xmath1 .
the adjacency matrices for these two graphs are shown in figure [ fig : adjequivgraphs ] .
they have the same diagonal sequence @xmath116 and both are optimal . we know that either the quasi - star or the quasi - complete graph in @xmath0 is optimal and that any threshold graph with the same diagonal sequence as an optimal graph is also optimal .
in fact , the converse is also true .
indeed , the relationship between the optimal graphs and the quasi - star and quasi - complete graphs in a graph class @xmath0 is described in our first main theorem .
[ thm : main1 ]
let @xmath5 be an optimal graph in @xmath0 .
then @xmath117 is a threshold graph for some partition @xmath118 and the diagonal sequence @xmath101 is equal to the diagonal sequence of either the quasi - star graph or the quasi - complete graph in @xmath0 .
theorem [ thm : main1 ] is stronger than lemma 8 @xcite because it characterizes _ all _ optimal graphs in @xmath0 . in section [ sec : optimalgraphs ]
we describe all optimal graphs in detail .
every optimal graph in @xmath0 is a threshold graph , @xmath95 , corresponding to a partition @xmath96 in @xmath93 .
so we extend the terminology and say that the partition @xmath96 is _ optimal _ in @xmath93 , if its threshold graph @xmath95 is optimal in @xmath0 .
we say that the partition @xmath118 is the _ quasi - star partition _
, if @xmath95 is the quasi - star graph in @xmath0 .
similarly , @xmath118 is the _ quasi - complete partition _ , if @xmath95 is the quasi - complete graph in @xmath0 .
we now describe the quasi - star and quasi - complete partitions in @xmath93
. first , the quasi - complete graphs .
let @xmath1 be a positive integer and @xmath2 an integer such that @xmath38 .
there exists unique integers @xmath20 and @xmath71 such that @xmath119 the partition @xmath120 corresponds to the quasi - complete threshold graph @xmath19 in @xmath93 .
the symbol @xmath121 means that @xmath71 is missing . to describe the quasi - star partition @xmath122 in @xmath93 ,
let @xmath123 be the unique integers such that @xmath124 then the partition @xmath125 corresponds to the quasi - star graph @xmath26 in @xmath93
. in general , there may be many partitions with the same diagonal sequence as @xmath126 or @xmath122 .
for example , if @xmath127 , then @xmath128 and all of the partitions in figure [ fig : examplev14e28 ] have the same diagonal sequence , @xmath129 .
but none of the threshold graphs corresponding to the partitions in figure [ fig : examplev14e28 ] is optimal . indeed ,
if the quasi - complete graph is optimal in @xmath93 , then there are at most three partitions in @xmath93 with the same diagonal sequence as the quasi - complete graph .
the same is true for the quasi - star partition .
if the quasi - star partition is optimal in @xmath93 , then there are at most three partition in @xmath93 having the same diagonal sequence as the quasi - star partition . as a consequence ,
there are at most six optimal partitions in @xmath93 and so at most six optimal graphs in @xmath0 .
our second main result , theorem [ thm : main2 ] , entails theorem [ thm : main1 ] ; it describes the optimal partitions in @xmath0 in detail .
the six partitions described in theorem [ thm : main2 ] correspond to the six graphs determined by byer in @xcite .
however , we give precise conditions to determine when each of these partitions is optimal .
[ thm : main2 ] let @xmath1 be a positive integer and @xmath2 an integer such that @xmath130 .
let @xmath131 be the unique integers satisfying @xmath132 and @xmath133 then every optimal partition @xmath96 in @xmath93 is one of the following six partitions : 1.1 : : @xmath134 , the quasi - star partition for @xmath2 , 1.2 : : @xmath135 , if @xmath136 , 1.3 : : @xmath137 , if @xmath138 and @xmath139 , 2.1 : : @xmath140 , the quasi - complete partition for @xmath2 , 2.2 : : @xmath141 , if @xmath142 , 2.3 : : @xmath143 , if @xmath144 and @xmath139 . on the other hand , partitions @xmath145 and @xmath146 always exist and at least one of them is optimal .
furthermore , @xmath147 and @xmath148 ( if they exist ) have the same diagonal sequence as @xmath145 , and if @xmath149 , then they are all optimal . similarly , @xmath150 and @xmath151 ( if they exist ) have the same diagonal sequence as @xmath146 , and if @xmath152 , then they are all optimal . a few words of explanation are in order regarding the notation for the optimal partitions in theorem [ thm : main2 ] . if @xmath153 , then @xmath154 , and @xmath155 . if @xmath156 , then @xmath157 , and @xmath158 ; further , if @xmath138 , then @xmath159
. in all other cases
@xmath160 and then @xmath145 , @xmath147 , and @xmath148 are properly defined .
if @xmath161 or @xmath162 , then both partitions in 1.1 and 1.2 would be equal to @xmath163 and @xmath164 respectively .
so the condition @xmath165 merely ensures that @xmath166 .
a similar remark holds for the partitions in 2.1 and 2.2 . by definition the partitions @xmath145 and @xmath148 are always distinct ; the same holds for partitions @xmath146 and @xmath151 . in general the partitions
@xmath167 described in items 1.1 - 1.3 and 2.1 - 2.3 ( and their corresponding threshold graphs ) are all different .
all the exceptions are illustrated in figure [ fig : hexagon ] and are as follows : for any @xmath1 , if @xmath168 or @xmath169 then @xmath170 .
for any @xmath139 , if @xmath171 or @xmath172 , then @xmath173 and @xmath174 .
if @xmath175 then @xmath176 and @xmath177 . finally , if @xmath178 or @xmath179 , then @xmath180 . similarly ,
if @xmath181 or @xmath182 , then @xmath183 .
for @xmath184 and @xmath185 , all the partitions @xmath167 are pairwise distinct ( when they exist ) . in the next section ,
we determine the pairs @xmath8 having a prescribed number of optimal partitions ( and hence graphs ) in @xmath0 . in principle , a given pair @xmath8 , could have between one and six optimal partitions .
it is easy to see that there are infinitely many pairs @xmath8 with only one optimal partition ( either the quasi - star or the quasi - complete ) .
for example the pair @xmath186 only has the quasi - complete partition .
similarly , there are infinitely many pairs with exactly two optimal partitions and this can be achieved in many different ways .
for instance if @xmath187 and @xmath188 , then @xmath189 , @xmath190 , and @xmath191 ( c.f .
corollary [ cor : strictuniform ] ) .
thus only the partitions @xmath145 and @xmath147 are optimal .
the interesting question is the existence of pairs with 3,4,5 , or 6 optimal partitions .
often , both partitions @xmath147 and @xmath148 in theorem [ thm : main2 ] exist for the same pair @xmath8 ; however it turns out that this almost never happens when they are optimal partitions .
more precisely , [ thm : family5_6 ] if @xmath147 and @xmath148 are optimal partitions then @xmath192 or @xmath193 . similarly ,
if @xmath150 and @xmath151 are optimal partitions then @xmath194 or @xmath193 . furthermore , the pair @xmath193 is the only one with six optimal partitions , there are no pairs with five .
if there are more than two optimal partitions for a pair @xmath8 , then @xmath65 , that is , both the quasi - complete and the quasi - star partitions must be optimal . in the next two results , we describe two infinite families of partitions in @xmath93 , and hence graph classes @xmath0 , for which there are exactly three ( four ) optimal partitions .
the fact that they are infinite is proved in section [ sec : pell ] .
[ thm : family3 ] let @xmath195 and @xmath20 be positive integers that satisfy the pell s equation @xmath196 and let @xmath197
. then ( using the notation of theorem [ thm : main2 ] ) , @xmath198 , @xmath199 , @xmath200 , and there are exactly three optimal partitions in @xmath93 , namely @xmath201 the partitions @xmath202 , and @xmath151 do not exist .
[ thm : family4 ] let @xmath203 and @xmath20 be positive integers that satisfy the pell s equation @xmath204 and @xmath205 . then ( using the notation of theorem [ thm : main2 ] ) , @xmath206 , @xmath207 , and there are exactly four optimal partitions in @xmath93 , namely @xmath208 the partitions @xmath147 and @xmath150 do not exist . in this section , we compare @xmath31 and @xmath30 .
the main result of the section , theorem [ thm : main3 ] , is a theorem very much like lemma 8 of @xcite , with the addition that our results give conditions for equality of the two functions . if @xmath209 , then @xmath65 for all @xmath1 .
of course , if @xmath210 , @xmath211 and @xmath212 , or @xmath213 and @xmath214 , there is only one graph in the graph class @xmath0 . if @xmath215 and @xmath216 , then there are two graphs in the graph class @xmath217 : the path @xmath218 and the partial matching @xmath219 , with degree sequences @xmath220 and @xmath221 , respectively .
the path is optimal as @xmath222 and @xmath223 .
but the path is both the quasi - star and the quasi - complete graph in @xmath217 .
if @xmath171 and @xmath216 , then the quasi - star graph has degree sequence @xmath224 and the quasi - complete graph is a triangle with degree sequence @xmath225 .
since @xmath226 for both of these graphs , both are optimal .
similarly , @xmath65 for @xmath227 for @xmath228 .
now , we consider the cases where @xmath229 . figures [ fig : v25 ] , [ fig : v15 ] , [ fig : v17 ] , and [ fig : v23 ] show the values of the difference @xmath37 .
when the graph is above the horizontal axis , @xmath31 is strictly larger than @xmath30 and so the quasi - star graph is optimal and the quasi - complete is not optimal . and
when the graph is on the horizontal axis , @xmath65 and both the quasi - star and the quasi - complete graph are optimal . since the function @xmath37 is central symmetric , we shall consider only the values of @xmath2 from @xmath230 to the midpoint , @xmath48 , of the interval @xmath231 $ ] .
figure [ fig : v25 ] shows that @xmath232 for all values of @xmath2 : @xmath233 .
so , when @xmath234 , the quasi - star graph is optimal for @xmath235 and the quasi - complete graph is not optimal . for @xmath236 , the quasi - star and the quasi - complete graphs are both optimal .
figure [ fig : v15 ] shows that @xmath237 for @xmath238 and @xmath239 .
but @xmath240 . so the quasi - star graph is optimal and the quasi - complete graph is not optimal for all @xmath241 except for @xmath242 . both the quasi - star and the quasi - complete graphs are optimal in @xmath243 .
figure [ fig : v17 ] shows that @xmath244 for @xmath245 , @xmath246 , @xmath247 for @xmath248 , and @xmath249 .
finally , figure [ fig : v23 ] shows that @xmath250 for @xmath251 , but @xmath252 for @xmath253 .
these four examples exhibit the types of behavior of the function @xmath37 , for fixed @xmath1 .
the main thing that determines this behavior is the quadratic function @xmath254 ( the integer @xmath45 depends on @xmath1 . ) for example , if @xmath255 , then @xmath256 for all values of @xmath257 . to describe the behavior of @xmath37 for @xmath258
, we need to define @xmath259 where @xmath260 our third main theorem is the following : [ thm : main3 ] let @xmath1 be a positive integer 1 . if @xmath261 , then @xmath262 @xmath65 if and only if @xmath263 , or @xmath264 and @xmath265 .
2 . if @xmath266 , then @xmath267 @xmath65 if and only if @xmath268 3 .
if @xmath269 , then @xmath262 @xmath65 if and only if @xmath270 the conditions in theorem [ thm : main3 ] involving the quantity @xmath271 simplify and refine the conditions in @xcite involving @xmath272 and @xmath273
. the condition @xmath54 in lemma 8 of @xcite can be removed and the result restated in terms of the sign of the quantity @xmath274 . while @xcite considers only the two cases @xmath275 and @xmath255 , we analyze the case @xmath269 separately .
it is apparent from theorem [ thm : main3 ] that @xmath276 for @xmath277 if @xmath278 is large enough . indeed ,
ahlswede and katona ( * ? ? ?
* theorem 3 ) show this for @xmath279 , thus establishing an inequality that holds for all values of @xmath1 regardless of the sign of @xmath271 .
we improve this result and show that the inequality holds when @xmath280 .
[ cor : uniform ] let @xmath281 .
then @xmath282 for all @xmath283 and @xmath284 for all @xmath285 .
furthermore , the constant @xmath286 can not be replaced by a smaller value .
theorem 3 in @xcite can be improved in another way .
the inequalities are actually strict .
[ cor : strictuniform ] @xmath191 for @xmath287 and @xmath288 for @xmath289 .
we now turn to the questions asked in @xcite : what is the relative density of the positive integers @xmath1 for which @xmath50 for @xmath63 ?
of course , @xmath50 for @xmath290 if and only if @xmath52 for @xmath291 .
[ cor : density ] let @xmath292 be a positive integer and let @xmath293 denote the number of integers @xmath1 in the interval @xmath294 $ ] such that @xmath295 for all @xmath290 .
then @xmath296 the diagonal sequence for a threshold graph helps explain the behavior of the difference @xmath37 for fixed @xmath1 and @xmath297 . from figures [ fig :
v25 ] , [ fig : v15 ] , [ fig : v17 ] , and [ fig : v23 ] , we see that @xmath37 , regarded as a function of @xmath2 , is piecewise linear and the ends of the intervals on which the function is linear occur at @xmath298 and @xmath299 for @xmath300 .
we prove this fact in lemma [ lem : linearity ] . for now , we present an example .
take @xmath301 , for example .
figure [ fig : v15 ] shows linear behavior on the intervals @xmath302 $ ] , @xmath303 $ ] , @xmath304 $ ] , @xmath305 $ ] , @xmath306 $ ] , @xmath307 $ ] , and @xmath308 $ ] .
there are 14 binomial coefficients @xmath309 for @xmath310 : @xmath311 the complements with respect to @xmath312 are @xmath313 the union of these two sets of integers coincide with the end points for the intervals on which @xmath314 is linear . in this case , the function is linear on the 27 intervals with end points : @xmath315 these special values of @xmath2 correspond to special types of quasi - star and quasi - complete graphs . if @xmath298 , then the quasi - complete graph @xmath19 is the sum of a complete graph on @xmath71 vertices and @xmath316 isolated vertices . for example , if @xmath301 and @xmath317 , and @xmath318 , then the upper - triangular part of the adjacency matrix for @xmath319 is shown on the left in figure [ fig : v15e36789 ] . and
if @xmath299 , then the quasi - star graph @xmath26 has @xmath71 dominant vertices and none of the other @xmath316 vertices are adjacent to each other .
for example , the lower triangular part of the adjacency matrix for the quasi - star graph with @xmath301 , @xmath320 , and @xmath321 , is shown on the right in figure [ fig : v15e36789 ] . [ cols= " < ,
< , < , < " , ] as additional dots are added to the adjacency matrices for the quasi - complete graphs with @xmath322 , the value of @xmath323 increases by @xmath324
. and the value of @xmath325 increases by @xmath326 . thus , the difference _
increases _ by a constant amount of @xmath327 .
indeed , the diagonal lines are a distance of five apart .
hence the graph of @xmath314 for @xmath328 is linear with a slope of @xmath327 .
but for @xmath329 , the adjacency matrix for the quasi - star graph has an additional dot on the diagonal corresponding to @xmath330 , whereas the adjacency matrix for the quasi - complete graph has an additional dot on the diagonal corresponding to @xmath331 .
so @xmath332 _ decreases _ by @xmath327 .
the decrease of @xmath327 continues until the adjacency matrix for the quasi - complete graph contains a complete column at @xmath242 .
then the next matrix for @xmath333 has an additional dot in the first row and next column and the slope changes again .
returning for a moment to the threshold graph @xmath95 from figure [ fig : example643 ] , which corresponds to the distinct partition @xmath102 , we see the graph complement shown with the white dots .
counting white dots in the rows from bottom to top and from the left to the diagonal , we have 7,5,2,1 .
these same numbers appear in columns reading from right to left and then top to the diagonal .
so if @xmath95 is the threshold graph associated with @xmath96 , then the set - wise complement of @xmath96 ( @xmath334 ) in the set @xmath335 corresponds to the threshold graph @xmath336the complement of @xmath95 .
that is , @xmath337 the diagonal sequence allows us to evaluate the sum of squares of the degree sequence of a threshold graph .
each black dot contributes a certain amount to the sum of squares .
the amount depends on the location of the black dot in the adjacency matrix .
in fact all of the dots on a particular diagonal line contribute the same amount to the sum of squares . for @xmath338 ,
the value of a black dot in position @xmath339 is given by the entry in the following matrix : @xmath340\ ] ] this follows from the fact that a sum of consecutive odd integers is a square .
so to get the sum of squares @xmath104 of the degrees of the threshold graph associated with the distinct partition @xmath96 , sum the values in the numerical matrix above that occur in the positions with black dots . of course , an adjacency matrix is symmetric .
so if we use only the black dots in the upper triangular part , then we must replace the @xmath339-entry in the upper - triangular part of the matrix above with the sum of the @xmath339- and the @xmath341-entry , which gives the following matrix : @xmath342.\ ] ] thus , @xmath343 .
lemma [ lem : diagonal ] is proved .
there is a variation of the formula for @xmath104 in lemma [ lem : diagonal ] that is useful in the proof of lemma [ lem : vminoneqc ] .
we have seen that each black dot in the adjacency matrix for a threshold graph contributes a summand , depending on the location of the black dot in the matrix @xmath428 in ( [ eqn : evenmatrix ] ) .
for example , if @xmath429 , then the part of @xmath430 that corresponds to the black dots in the adjacency matrix @xmath97 for @xmath96 is @xmath431 thus @xmath432 .
now if we index the rows and columns of the adjacency matrix starting with zero instead of one , then the integer appearing in the matrix @xmath430 at entry @xmath339 is just @xmath433 .
so we can compute @xmath104 by adding all of the positions @xmath339 corresponding to the positions of black dots in the upper - triangular part of the adjacency matrix of @xmath95 .
what are the positions of the black dots in the adjacency matrix for the threshold graph corresponding to a partition @xmath434 ? the positions corresponding to @xmath435 are @xmath436 and the positions corresponding to @xmath437 are @xmath438 in general , the positions corresponding to @xmath439 in @xmath96 are @xmath440 we use these facts in the proof of lemma [ lem : vminoneqc ] .
let @xmath441 be the quasi - complete partition in @xmath442 , where @xmath443 and @xmath444 .
we deal with the cases @xmath445 , @xmath347 , and @xmath446 separately .
specifically , we show that if @xmath447 is optimal , then @xmath347 and @xmath448 with @xmath348 .
arguments for the cases are given below . since @xmath450
then @xmath451 . we show that @xmath452 is not optimal . in this case , the adjacency matrix for @xmath96 has the following form : let @xmath454 the parts of @xmath455 are distinct and decreasing since @xmath453 . thus @xmath456 . the adjacency matrices @xmath97 and @xmath457 each have @xmath2 black dots , many of which appear in the same positions .
but there are differences .
using the fact that @xmath458 , the first row of the following list shows the positions in which a black dot appears in @xmath97 but not in @xmath457 . and
the second row shows the positions in which a black dot appears in @xmath457 but not in @xmath97 : @xmath459 for each of the positions in the list , except the last ones , the sum of the coordinates for the positions is the same in the first row as it is in the second row . but the coordinates of the last pair in the first row sum to @xmath460 whereas the coordinates of the last pair in the second row sum to @xmath461 .
it follows that @xmath462 .
thus , @xmath96 is not optimal .
let @xmath464 .
since @xmath465 , the partition @xmath455 is in @xmath93 .
the positions of the black dots in the adjacency matrices @xmath97 and @xmath457 are the same but with only two exceptions .
there is a black dot in position @xmath466 in @xmath96 but not in @xmath455 , and there is a black dot in position @xmath467 in @xmath455 but not in @xmath96 . since @xmath468 , @xmath96 is not optimal . since @xmath470 ,
. we will show that if @xmath348 , then @xmath96 has the same diagonal sequence as the quasi - complete partition . and if @xmath472 , then @xmath96 is not optimal .
the adjacency matrix for @xmath96 is of the following form : @xmath473 the quasi - complete partition in @xmath0 is @xmath474 , where @xmath475 . to see this , notice that @xmath476 for @xmath475 . since @xmath348 and @xmath465 , then @xmath477 and @xmath456 . to see that @xmath96 and @xmath455 have the same diagonal sequence ,
we again make a list of the positions in which there is a black dot in @xmath97 but not in @xmath457 ( the top row below ) , and the positions in which there is a black dot in @xmath457 but not in @xmath97 ( the bottom row below ) : @xmath478 each position in the top row is in the same diagonal as the corresponding position in the bottom row , that is , @xmath479 . thus the diagonal sequences @xmath480 . in this case , let @xmath482 . and
since @xmath483 , the parts of @xmath455 are distinct and decreasing .
that is , @xmath456 .
using the fact that @xmath484 , we again list the positions in which there is a black dot in @xmath97 but not in @xmath457 ( the top row below ) , and the positions in which the is a black dot in @xmath457 but not in @xmath97 : @xmath485 all of the positions but the last in the top row are on the same diagonal as the corresponding position in the bottom row : @xmath486 .
but in the last positions we have @xmath487 and @xmath488 .
thus @xmath489 and so @xmath96 is not optimal .
we will show that @xmath491 is not optimal .
the adjacency matrix for @xmath96 has the following form : @xmath492 there are two cases .
let @xmath493 , where @xmath494 .
then @xmath495 because @xmath496 and @xmath465 .
we show that @xmath456 and @xmath497 . in order for @xmath455 to be in @xmath93 ,
the sum of the parts in @xmath455 must equal the sum of the parts in @xmath96 : @xmath498 and the parts of @xmath455 must be distinct and decreasing : @xmath499 the first inequality holds because @xmath500 .
the last two inequalities hold because @xmath501 .
thus @xmath456 . the top row below lists the positions where there is a black dot in @xmath97 but not in @xmath457 ; the bottom row lists the positions where there is a black dot in @xmath457 but not in @xmath97 : @xmath502 since @xmath495 then the lists above are non - empty . thus , to ensure that @xmath503 , it is sufficient to show that for each @xmath504 , position @xmath505 is in a diagonal to the left of position @xmath506 .
that is , @xmath507-[0+(v - t)]=2c+r - v - j,\ ] ] or equivalently , @xmath508 the inequality @xmath509 holds because @xmath510 , and @xmath511 holds because @xmath512 .
it follows that @xmath96 is not an optimal partition .
again we show that @xmath513 is not optimal .
let @xmath514 the sum of the parts in @xmath96 equals the sum of the parts in @xmath455
. and the partition @xmath455 is decreasing : @xmath515 the first three inequalities follow from the assumption that @xmath516 . and the fourth inequality holds because @xmath453 .
so @xmath456 .
the adjacency matrices @xmath97 and @xmath457 differ as follows .
the top rows of the following two lists contain the positions where there is a black dot in @xmath97 but not in @xmath457 ; the bottom row lists the positions where there is a black dot in @xmath457 but not in @xmath97 .
@xmath517 each position , @xmath518 ( @xmath519 ) , in the top row in list 1 is in the same diagonal as the corresponding position , @xmath520 , in the bottom row of list 1 .
each position , @xmath521 ( @xmath522 ) , in the top row of list 2 is in a diagonal to the left of the corresponding position , @xmath523 in the bottom row of list 2 .
indeed , @xmath524 . and
since @xmath525 , list 2 is not empty .
it follows that @xmath526 and so @xmath96 is not a optimal partition .
the proof of lemma [ lem : vminoneqc ] is complete .
theorem [ thm : main1 ] is an immediate consequence of theorem [ thm : main2 ] ( and lemmas [ lem : optarethres ] and [ lem : diagonal ] ) . and
theorem [ thm : main2 ] can be proved using the following central lemma : [ lem : vminoneqc ] let @xmath344 be an optimal partition in @xmath93 , where @xmath345 and @xmath346
. then @xmath347 and @xmath348 so that @xmath349 we defer the proof of lemma [ lem : vminoneqc ] until section [ sec : proofvminoneqc ] and proceed now with the proof of theorem [ thm : main2 ] .
the proof of theorem [ thm : main2 ] is an induction on @xmath1 .
let @xmath96 be an optimal partition in @xmath93 , then @xmath334 is optimal in @xmath350 .
one of the partitions , @xmath351 contains the part @xmath27 .
we may assume without loss of generality that @xmath352 , where @xmath353 is a partition in @xmath354 .
the cases where @xmath353 is a decreasing partition of @xmath355 and @xmath356 will be considered later .
for now we shall assume that @xmath357 .
since @xmath96 is optimal , it follows that @xmath353 is optimal and hence by the induction hypothesis , @xmath353 is one of the following partitions in @xmath354 : 1.1a : : @xmath358 , the quasi - star partition for @xmath359 , 1.2a : : @xmath360 , if @xmath361 , 1.3a : : @xmath362 , if @xmath138 , 2.1a : : @xmath363 , the quasi - complete partition for @xmath359 , 2.2a : : @xmath364 , if @xmath365 , 2.3a : : @xmath366 , if @xmath367 , where @xmath368 in symbols , @xmath369 , for one of the partitions @xmath370 above . for each partition , @xmath370
, we will show that @xmath371 for one of the six partitions , @xmath372 , in the statement of theorem [ thm : main2 ] .
the first three cases are obvious : @xmath373 next suppose that @xmath374 .
the partitions @xmath375 and @xmath376 do not exist unless certain conditions on @xmath377 , and @xmath1 are met .
and whenever those conditions are met , the partition @xmath378 is also optimal .
thus @xmath379 is optimal .
also , since @xmath380 , then @xmath381 .
there are two cases : @xmath382 .
if @xmath383 , then @xmath375 does not exist and @xmath384 if @xmath385 , then by lemma [ lem : vminoneqc ] , @xmath386 , with @xmath387 and @xmath388 .
we will show that @xmath389 and @xmath390 .
the above inequalities imply that @xmath391 but @xmath20 is the unique integer satisfying @xmath392 .
thus @xmath389 .
it follows that @xmath393 and so @xmath394 .
we now consider the cases 2.1a , 2.2a , and 2.3a individually .
actually , @xmath375 does not exist since @xmath395 .
if @xmath396 , then @xmath397 since @xmath398 . this contradicts the assumption that @xmath353 is a partition of an integer greater than 3 .
therefore @xmath399 since @xmath395 and @xmath389 .
now since @xmath400 we have @xmath401 finally , if @xmath353 is a decreasing partition of @xmath355 or @xmath356 , then either @xmath402 , or @xmath403 , or @xmath404 for some @xmath405 .
now , we prove that @xmath406 and @xmath407 ( if they exist ) have the same diagonal sequence as @xmath408 ( which always exists ) .
this in turn implies ( by using the duality argument mentioned in section [ sec : diagonallemma ] ) that @xmath409 and @xmath410 also have the same diagonal sequence as @xmath146 ( which always exists ) .
we use the following observation .
if we index the rows and columns of the adjacency matrix @xmath411 starting at zero instead of one , then two positions @xmath339 and @xmath412 are in the same diagonal if and only if the sum of their entries are equal , that is , @xmath413 . if @xmath147 exists then @xmath414 . applying the previous argument to @xmath408 and @xmath406 , we observe that the top row of the following lists shows the positions where there is a black dot in @xmath415 but not in @xmath416 and the bottom row shows the positions where there is a black dot in @xmath416 but not in @xmath415.@xmath417 each position in the top row is in the same diagonal as the corresponding position in the second row
. thus the number of positions per diagonal is the same in @xmath408 as in @xmath406 .
that is , @xmath418 . similarly , if @xmath148 exists then @xmath419 . to show that @xmath420 note that the only position where there is a black dot in @xmath421 but not in @xmath422 is @xmath423 , and the only position where there is a black dot in @xmath424 but not in @xmath425 is @xmath426 .
since these positions are in the same diagonal then @xmath427 .
theorem [ thm : main2 ] is proved .
we first prove theorem [ thm : family5_6 ] .
if @xmath406 and @xmath407 are optimal partitions , then according to theorem [ thm : main2 ] , @xmath804 , @xmath805 , and so @xmath806 .
in addition the quasi - star partition is optimal , that is , @xmath149 .
thus by corollary [ cor : strictuniform ] , either @xmath807 or @xmath808 .
if @xmath809 and since @xmath138 , then @xmath810 contradicting @xmath811 .
thus @xmath812 .
since @xmath813 and @xmath814 , then @xmath815 therefore @xmath816 . in this range of @xmath1 , the only pairs @xmath8 that satisfy all the required inequalities are @xmath192 or @xmath193 . using the relation between a graph and its complement described below equation ( [ eqn : scconnection ] ) , we conclude that if @xmath409 and @xmath151 are optimal partitions , then @xmath194 or @xmath193 . as a consequence
, we see that the pair @xmath193 is the only candidate to have six different optimal partitions .
this is in fact is the case .
the six graphs and partitions are depicted in figure [ fig : sixoptimal ] .
we note here that byer @xcite also observed that the pair @xmath817 yields six different optimal graphs .
another consequence is that the pairs @xmath182 and @xmath179 are the only candidates to have five different optimal partitions . for the pair @xmath182
, the partitions @xmath818 and @xmath409 all exist and are optimal .
however , @xmath819 . thus the pair @xmath182 only has four distinct optimal partitions . similarly , for the pair @xmath179 the partitions
@xmath820 and @xmath410 all exist and are optimal , but @xmath821 .
so there are no pairs with five optimal partitions , and thus all other pairs have at most four optimal partitions .
moreover , @xmath65 is a necessary condition to have more than two optimal partitions , since any pair other than @xmath182 or @xmath179 must satisfy that both @xmath145 and @xmath822 are optimal .
the proof of theorem [ thm : family5_6 ] is complete .
in theorem [ thm : family3 ] , @xmath823 and thus @xmath198 .
note that , if @xmath195 and @xmath20 satisfy equation ( [ eqn : family3 ] ) , then @xmath824 , and so @xmath825 .
thus @xmath826 with @xmath827 , that is , @xmath828 and @xmath829 .
hence , @xmath830 and @xmath831 ( which always exist ) are different because @xmath832 .
the partition @xmath833 exists because @xmath834 , and it is different to @xmath822 because @xmath835 ( @xmath836 by definition ) .
finally , the partitions @xmath837 and @xmath410 do not exist because @xmath838 and @xmath839 , respectively .
theorem [ thm : family3 ] is proved .
now , if @xmath1 and @xmath20 satisfy equation ( [ eqn : family4 ] ) , then @xmath840 . moreover , since @xmath203 , then @xmath841
hence , in theorem [ thm : family4 ] , @xmath842 , with @xmath843 because @xmath844 .
that is , @xmath207 and @xmath845 .
thus @xmath846 , and @xmath847 all exist and are different because @xmath848 does not yield a solution to ( [ eqn : family4 ] ) .
also @xmath406 and @xmath409 do not exist because @xmath849 .
theorem [ thm : family4 ] is proved .
the notation in this section changes a little from that used in section [ sec : intro ] . in section [ sec : intro ] , we write @xmath527 with @xmath33 . here
, we let @xmath528 so that @xmath529 with @xmath530 .
then equation ( [ eqn : cve ] ) is equivalent to @xmath531 before proceeding , we should say that the abuse of notation in @xmath532 should not cause confusion as it will be clear which set of parameters @xmath8 _ vs. _ @xmath533 are being used . also notice that if we were to expand the range of @xmath292 to @xmath534 , that is allow @xmath535 , then the representation of @xmath2 in equation ( [ eqn : repre ] ) is not unique : @xmath536 but the value of @xmath30 is the same in either case : @xmath537 thus we may take @xmath534 .
we begin the proofs now . at the beginning of section [ sec : qsvsqc ]
, we showed that @xmath65 for @xmath209 . also note that , when @xmath48 is an integer , @xmath538 .
we now compare @xmath31 with @xmath30 for @xmath539 .
the first task is to show that @xmath540 for all but a few values of @xmath2 that are close to @xmath48 .
we start by finding upper and lower bounds on @xmath31 and @xmath30 .
define @xmath541 the first lemma shows that @xmath542 is an upper bound for @xmath30 and leads to an upper bound for @xmath31 .
arguments used here to obtain upper and lower bounds are similar to those in @xcite .
[ lem : upperbs ] for @xmath543 @xmath544 it is clearly enough to prove the first inequality .
the second one is trivially obtained from equation ( [ eqn : scconnection ] ) linking the values of @xmath31 and @xmath30 .
we prove the inequality in each interval @xmath545 .
so fix @xmath546 for now .
we make yet another change of variables to get rid off the square root in the above expression of @xmath547 .
set @xmath548 , for @xmath549 .
then @xmath550 which is easily seen to be positive for all @xmath546 and all @xmath549 .
now define @xmath551 the next lemma shows that @xmath552 is a lower bound for @xmath30 and leads to a lower bound for @xmath31 .
[ lem : lowerbs ] for @xmath553 @xmath554 as above , set @xmath548 , @xmath555 , and @xmath556 , @xmath557 .
then @xmath558 this expression is easily seen to be positive for @xmath559 .
we are now ready to prove that @xmath540 for @xmath560 for all but a few small values and some values close to @xmath48 .
< v ] assume @xmath561 .
for @xmath562 we have @xmath563 .
as we showed above in lemma [ lem : upperbs ] , @xmath564 is an upper bound on @xmath30 for all @xmath565 .
furthermore , it is easy to see that for @xmath566 we have @xmath567 .
in fact , the quasi - star graph is optimal for @xmath568 .
the rest is then straightforward . for @xmath569 ,
we have @xmath570 taking square roots and rearranging some terms proves the result .
[ lemma:0.55 ] assume @xmath561 .
for @xmath571 we have @xmath572 assume that @xmath290 .
let @xmath573 with @xmath574 . by lemmas [ lem : upperbs ] and [ lem : lowerbs ]
, we have @xmath575 we focus on the first two terms .
set @xmath576 by considering a real variable @xmath577 , it is easy to see that @xmath578 , @xmath579 , and @xmath580 on the interval in question . thus @xmath581 is concave down on @xmath582 .
we are comparing @xmath581 with the line @xmath583 on the interval @xmath584 $ ] .
the concavity of @xmath581 allows to check only the end points . for @xmath585
, we need to check @xmath586 it is messy and elementary to verify this inequality for @xmath587 . for @xmath588 we need to check @xmath589 this inequality holds for @xmath590 .
this time the calculations are a bit messier , yet still elementary . for @xmath591
, we verify the result directly by calculation . in section 1
, we introduced the value @xmath592 .
we now define @xmath593 the next lemma shows that those binomial coefficients and their complements are all we need to consider .
[ lemma : e1f2 ] @xmath594 . as a consequence @xmath191 for all @xmath595 .
we need a small result on the relationship between @xmath272 and @xmath1 first .
the upper bound will be used later in this section .
[ lem : kzerobound ] @xmath596 . since @xmath597 , we have @xmath598 thus @xmath599 that is , @xmath600 the result follows using @xmath601 and @xmath602 .
note that @xmath603 and @xmath604 .
hence , it is enough to show that @xmath605 .
this follows from the previous lemma for @xmath606 .
for @xmath607 , we verify the statement by direct calculation .
next , we show that the difference function @xmath608 is piecewise linear on the intervals induced by the binomial coefficients @xmath609 , @xmath610 , and their complements @xmath611 , @xmath610 . in section [ sec : piecelinearity ] , we show a specific example .
[ lem : linearity ] as a function of @xmath2 , the function @xmath612 is linear on the interval @xmath613 .
the line has slope @xmath614 if @xmath615 with @xmath33 , then it is easy to see from equation ( [ eqn : cve ] ) that @xmath616 using equations ( [ eqn : scconnection ] ) and ( [ eqn : ckt ] ) , we find that , if @xmath617 , @xmath618 , then @xmath619 we now have @xmath620 the conclusion follows . since we already know that @xmath621 for @xmath622 , and @xmath623 for @xmath624 or @xmath48 , we can now focus on the interval @xmath625
the only binomial coefficients or complements of binomial coefficients that can fall into this interval are @xmath626 and @xmath627 .
there are two possible arrangements we need to consider 1 .
@xmath628 and 2 . @xmath629 .
the next result deals with the first arrangement . if @xmath630 , then @xmath255 .
furthermore , @xmath282 for @xmath290 with equality if and only if @xmath624 or @xmath48 ; or @xmath631 and @xmath632 .
@xmath633 implies @xmath634 . by lemma [ lem : kzerobound ]
, we conclude that for @xmath635 , @xmath636 for smaller values , we verify that @xmath255 by direct calculation . if @xmath637 in equation ( [ eqn : ckt ] ) , and since @xmath638 , then @xmath639 and @xmath640 . using equation ( [ eqn : scconnection ] ) , @xmath641 .
similarly , since @xmath642 , then for @xmath643 in equation ( [ eqn : ckt ] ) , we have @xmath644 and @xmath645 . again , using equation ( [ eqn : scconnection ] ) , @xmath646 notice that @xmath647 since both factors in ( [ diffve0 ] ) are even and differ by 2 .
equality occurs if and only if @xmath648 or @xmath649 . finally , observe that @xmath650 and @xmath651 by lemmas [ lemma:0.55 ] and [ lemma : e1f2 ] , and @xmath652 and @xmath653 are both less than @xmath627 .
hence @xmath654 for @xmath655 $ ] follows from the piecewise linearity of @xmath612 .
the rest follows from lemma [ lemma:0.55 ]
. now we deal with the case @xmath656 .
there are three cases depending on the sign of @xmath271 .
all these cases require the following fact . if @xmath656 , then for @xmath657 in equation ( [ eqn : ckt ] ) , @xmath639 and @xmath658 . since @xmath659 , for @xmath660 in equation ( [ eqn : ckt ] ) , @xmath639 and @xmath661 .
thus , using equation ( [ eqn : scconnection ] ) , @xmath662 whenever @xmath663 .
this automatically gives the sign of @xmath612 near @xmath48 . by the piecewise linearity of @xmath612 given by lemma [ lem : linearity ] , the only thing remaining is to investigate the sign of @xmath664 .
assume @xmath656 and @xmath255 .
then @xmath282 for @xmath290 , with equality if and only if @xmath665 .
first , note that @xmath666 , since @xmath667 occurs only if @xmath668 and thus @xmath669 . for @xmath670 , by equation ( [ eqn : diffnearm ] ) ,
@xmath671 . furthermore ,
if @xmath637 in equation ( [ eqn : ckt ] ) , then @xmath672 and @xmath673 .
thus , by equation ( [ eqn : scconnection ] ) , @xmath674 and @xmath675 the first factor is positive because @xmath656 .
the second factor is positive for @xmath676 .
this follows from the fact that @xmath677 by lemma [ lem : kzerobound ] , and @xmath678 because @xmath679 : for @xmath680 , @xmath681 since @xmath682 , then @xmath683 for @xmath676 . the only case left to verify satisfying the conditions of this lemma is @xmath684 . in this case ,
@xmath685 and @xmath686 .
the previous two lemmas provide a proof of part 1 of theorem [ thm : main3 ] .
assume @xmath656 and @xmath269 .
then @xmath282 for @xmath290 with equality if and only if @xmath687 . for @xmath688 , by equation ( [ eqn : diffnearm ] ) , @xmath689 .
as in the previous lemma , for @xmath690 @xmath691 and thus @xmath683 .
the only value of @xmath692 satisfying the conditions of this lemma is @xmath693 with @xmath694 , and @xmath695 .
the previous lemma provides a proof for part 3 of theorem [ thm : main3 ] .
assume @xmath629 and @xmath258 .
then @xmath282 for @xmath696 and @xmath284 for @xmath697 with equality if and only if @xmath698 . for @xmath670 , by equation ( [ eqn : diffnearm ] ) , @xmath699 .
this time it is possible that @xmath700 . in this case , by lemmas [ lemma:0.55 ] and [ lemma : e1f2 ] , we know that @xmath701 .
also , @xmath668 and @xmath702 , implying @xmath703 and @xmath621 for all @xmath704 .
if @xmath679 , by lemma [ lem : linearity ] , @xmath612 is linear as a function of @xmath2 on the interval @xmath705 $ ] .
let @xmath706 be the slope of this line .
since @xmath707 , then @xmath639 and @xmath708 in lemma [ lem : linearity ] . thus @xmath709 and @xmath710 .
the line through the two points @xmath711 and @xmath712 crosses the @xmath713-axis at @xmath714 .
we now show that @xmath715 , which in turn proves that @xmath683 .
we have @xmath716 since @xmath717 and @xmath718 , then @xmath719 which is equivalent to @xmath720 .
thus @xmath721 by equation ( [ eqn : r_0 first ] ) . to prove @xmath722 ,
according to equation ( [ eqn : r_0 second ] ) , we need to show @xmath723 after multiplying by @xmath724 , the last inequality becomes @xmath725 which is equivalent to @xmath726 since @xmath727 we know that @xmath728 . also , inequality ( [ eqn : v_k0 ] ) is equivalent to @xmath729 .
multiplying these two inequalities yields the result .
the previous lemma provides a proof of part 2 of theorem [ thm : main3 ] .
the expression for @xmath714 is sometimes an integer .
those @xmath730 for which @xmath714 is an integer are 14 , 17 , 21 , 120 , 224 , 309 , 376 , 393 , 428 , 461 , 529 , 648 , 697 , and 801 . in the remaining part of this section , we prove corollaries [ cor : uniform ] and [ cor : strictuniform ] .
assume that @xmath718 and @xmath258 .
then @xmath731 where @xmath281 .
we show that @xmath731 for @xmath718 .
recall that @xmath732 thus we need to show @xmath733 define the function @xmath734 .
the interval for @xmath713 is limited by the condition that @xmath258 which implies that @xmath735 furthermore , since @xmath717 , we know that @xmath736 . we show that @xmath737 is increasing on @xmath738 $ ] .
note that , since @xmath718 , @xmath739 for @xmath740 .
hence @xmath737 is concave up on @xmath741 .
furthermore @xmath742 for @xmath743 , and hence @xmath744 for @xmath743 .
the only values of @xmath1 greater than 4 and smaller than 11 for which @xmath258 are @xmath745 .
the result is easily verified in those two cases .
how good is the bound @xmath731 ?
suppose there is a parameter @xmath746 such that @xmath747 with @xmath748 .
assume @xmath749 .
there are infinitely many values of @xmath1 for which this is true ( see section [ sec : pell ] ) . in all of those cases @xmath750 .
we have the following @xmath751 as @xmath752 .
thus @xmath753 and hence @xmath286 the greatest number for which the bound on @xmath754 holds .
since @xmath282 for all @xmath755 , we proved corollary [ cor : uniform ] .
to prove corollary [ cor : strictuniform ] , we need to investigate the other non - trivial case of equality in theorem [ thm : main3 ] .
it occurs when @xmath631 and @xmath756 .
notice that this implies @xmath757 there are infinitely many values of @xmath1 such that @xmath648 , and infinitely many values of @xmath1 such that @xmath758 ( see section [ sec : pell ] ) .
thus the best we can say is that @xmath191 for all @xmath759 , and corollary [ cor : strictuniform ] is proved .
recall that for each @xmath1 , @xmath760 is the unique positive integer such that @xmath761 it follows that @xmath762 let us restrict our attention to the parts of the hyperbolas @xmath763 that occupy the first quadrant as shown in figure [ fig : hyperb ] . then each lattice point
, @xmath764 is in the closed region bounded by @xmath765 below and @xmath766 above .
furthermore , the sign of the quadratic form @xmath767 determines whether the quasi - star graph is optimal in @xmath0 for all @xmath290 . by theorem [ thm : main3 ] , if @xmath768 , then @xmath282 ( and the quasi - star graph is optimal ) for @xmath290 .
thus , if the lattice point @xmath769 is between @xmath766 and the hyperbola @xmath770 then the quasi - star graph is optimal in @xmath0 for all @xmath771 .
but if the lattice point @xmath764 is between @xmath772 and @xmath765 , then there exists a value of @xmath2 in the interval @xmath773 such that the quasi - complete graph is optimal and the quasi - star graph is not optimal .
of course , if the lattice point @xmath764 is on @xmath772 , then the quasi - star graph is optimal for all @xmath290 but the quasi - complete graph is also optimal for @xmath774 .
apparently , the density limit @xmath775 from corollary [ cor : density ] depends on the density of lattice points @xmath769 in the region between @xmath766 and @xmath772
. we can give a heuristic argument to suggest that the limit is @xmath776 .
the asymptotes for the three hyperbolas are @xmath777 and intersect the @xmath20-axis at @xmath778 the horizontal distance between @xmath779 and @xmath780 is @xmath781 and the horizontal distance between @xmath779 and @xmath68 is @xmath782 to make the plausibility argument rigorous , we need a theorem of weyl ( * ? ? ?
* satz 13 , page 334 ) , @xcite : for any real number @xmath783 , let @xmath784 denote the fractional part of @xmath783 .
that is , @xmath784 is the unique number in the half - open interval @xmath785 such that @xmath786 is an integer .
now let @xmath746 be an irrational real number .
then the sequence @xmath787 , @xmath788 , is uniformly distributed in the interval @xmath785 . in our problem
, the point @xmath789 is between the hyperbolas @xmath765 and @xmath766 and , with few exceptions , @xmath789 is also between the asymptotes @xmath780 and @xmath779 . to be precise ,
suppose that @xmath764 satisfies inequalities ( [ eqn : between ] ) .
we need an easy fact from number theory here .
namely that @xmath790 for all odd integers @xmath791 .
thus @xmath792 unless @xmath793 .
( these are the exceptions . ) but for all other points @xmath764 we have @xmath794 thus @xmath795 and so @xmath796 next , consider the condition @xmath797 , which is equivalent to @xmath798 unless @xmath799 , @xmath797 is equivalent to @xmath800 to summarize , if @xmath764 does not satisfy either of these pell s equations @xmath801 then @xmath797 if and only if @xmath802 from weyl s theorem , we know that the fractional part in the above inequality is uniformly distributed in the interval @xmath785 . since the density of the values of @xmath1 for which @xmath764 is a solution to one of the pell s equations above is zero , then @xmath803 .
the proof of corollary [ cor : density ] is complete .
pell s equation @xmath850 where @xmath851 , appears several times in this paper .
for example , a condition for the equality of @xmath31 and @xmath30 in theorem [ thm : main3 ] involves the pell s equation @xmath852 . and in theorem [ thm : family4 ] , we have @xmath853 . there are infinitely many solutions to each of these equations . in each instance , @xmath854 and @xmath855 in equation ( [ eqn : pell ] ) are positive odd integers and @xmath856 .
the following lemma describes the solutions to the fundamental pell s equation .
it follows from the lemma that if @xmath859 is a solution to equation ( [ eqn : pell-1 ] ) , then both @xmath854 and @xmath855 are odd .
we list the first several solutions to equation ( [ eqn : pell-1 ] ) : @xmath860 now let us consider the equation @xmath861 from theorem [ thm : family3 ] . since all of the positive solutions @xmath859 consist of odd integers , the pair @xmath769 defined by @xmath862 are integers and satisfy equation ( [ eqn : family3 ] ) .
thus there is an infinite family of values for @xmath195 such that there are exactly 3 optimal partitions in @xmath93 where @xmath197 .
the following is a list of the first three values of @xmath863 in this family : @xmath864 next , consider equation ( [ eqn : family4 ] ) from theorem [ thm : family4 ] and the corresponding pell s equation : @xmath865 a simple argument using the norm function , @xmath866 shows that all positive integral solutions are given by @xmath867 where @xmath70 is a nonnegative integer .
the first several solutions are @xmath868 thus the pairs @xmath769 , defined by @xmath869 satisfy equation ( [ eqn : family4 ] ) .
the first three members , @xmath870 of this infinite family of partitions @xmath93 with @xmath203 , @xmath871 , and exactly 4 optimal partitions are : all its positive integral solutions are given by @xmath877 where @xmath70 is a nonnegative integer .
the first several solutions are @xmath878 the proof of corollary [ cor : strictuniform ] requires infinitely many solutions to the pell s equation @xmath879 and infinitely many solutions to the pell s equation @xmath880 all positive integral solutions to ( [ eqn : pellcor2a ] ) are given by @xmath881 where @xmath70 is a nonnegative integer .
the first several solutions are @xmath882 we have shown that equation ( [ eqn : pellcor2b ] ) has infinitely many solutions , as it is the same equation that appears in theorem [ thm : family3 ] .
however , in corollary [ cor : strictuniform ] , @xmath20 must be @xmath272 , the unique integer that satisfies inequality ( [ eqn : k0 ] ) .
this condition is also necessary for equations ( [ eqn : pellmain3 ] ) , ( [ eqn : pellcor1 ] ) , and ( [ eqn : pellcor2a ] ) .
in other words , we must show that for @xmath1 large enough , every solution @xmath769 to one of the equations ( [ eqn : pellmain3 ] ) , ( [ eqn : pellcor1 ] ) , or ( [ eqn : pellcor2a ] ) , satisfies inequality ( [ eqn : k0 ] ) .
we do this only for equation ( [ eqn : pellmain3 ] ) as all other cases are similar .
suppose that @xmath769 is a solution to equation ( [ eqn : pellmain3 ] ) with @xmath883
. then @xmath884 .
inequality ( [ eqn : k0 ] ) consists of two parts , the first of which is @xmath885 to prove this part , we compute @xmath886 | let @xmath0 be the set of all simple graphs with @xmath1 vertices and @xmath2 edges and let @xmath3 denote the sum of the squares of the degrees , @xmath4 , of the vertices of @xmath5 .
it is known that the maximum value of @xmath6 for @xmath7 occurs at one or both of two special graphs in @xmath0the quasi - star graph or the quasi - complete graph . for each pair
@xmath8 , we determine which of these two graphs has the larger value of @xmath6 .
we also determine all pairs @xmath8 for which the values of @xmath6 are the same for the quasi - star and the quasi - complete graph .
in addition to the quasi - star and quasi - complete graphs , we find all other graphs in @xmath0 for which the maximum value of @xmath6 is attained .
density questions posed by previous authors are examined . *
ams subject classification : * 05c07 , 05c35 + * key words : * graph , degree sequence , threshold graph , pell s equation , partition , density |
it is already three decades since the first detection of solar neutrinos .
it was realized from the very beginning that the observed rate at the homestake experiment @xcite was far lower than the theoretical expectation based on the standard solar model @xcite with the implicit assumption that neutrinos created in the solar interior reach the earth unchanged , i.e. they are massless and have only standard properties and interactions . in the first two decades of solar neutrino research ,
the problem consisted only of the discrepancy between theoretical expectations based upon solar model calculations and the observations of the capture rate in the chlorine solar neutrino experiment .
this discrepancy led to a change in the original goal of using solar neutrinos to probe the properties of the solar interior towards the study of the properties of the neutrino itself . from the experimental point of view
much progress has been done in recent years .
we now have available the results of five experiments , the original chlorine experiment at homestake @xcite , the radio chemical gallium experiments on pp neutrinos , gallex @xcite and sage @xcite , and the water cherenkov detectors kamiokande @xcite and super
kamiokande @xcite .
the latter has been able not only to confirm the original detection of solar neutrinos at lower rates than predicted by standard solar models , but also to demonstrate directly that the neutrinos come from the sun by showing that recoil electrons are scattered in the direction along the sun - earth axis .
we now have good information on the time dependence of the event rates during the day and night , as well as a measurement of the recoil electron energy spectrum .
after 825 days of operation , super
kamiokande has also presented preliminary results on the seasonal variation of the neutrino event rates , an issue which will become important in discriminating the msw scenario from the possibility of neutrino oscillations in vacuum @xcite . on the other hand
there has been improvement on solar modelling and nuclear cross sections .
for example , helioseismological observations have now established that diffusion is occurring and by now most solar models incorporate the effects of helium and heavy element diffusion @xcite .
the quality of the experiments themselves and the robustness of the theory make us confident that in order to describe the data one must depart from the standard model ( sm ) of particle physics interactions , by endowing neutrinos with new properties . in theories beyond the standard model of particle physics neutrinos
may naturally have exotic properties such as non - orthonormality @xcite , flavour - changing interactions @xcite , transition magnetic moments @xcite and neutrino decays @xcite , the most generic is the existence of mass .
while many of these may play a role in neutrino propagation and therefore in the explanation of the data @xcite it is undeniable that the most generic and popular explanation of the solar neutrino anomaly is in terms of neutrino masses and mixing leading to neutrino oscillations either in _ vacuum _ @xcite or via the matter - enhanced _ msw mechanism
_ @xcite .
in this paper we study the implications of the present data on solar neutrinos in the framework of the two - flavor msw solutions to the solar neutrino problem .
we perform a global fit to the full data set corresponding to 825 days of data of the super kamiokande experiment as well as to chlorine , gallex and sage . in our analysis
we use as ssm the latest results from the most accurate calculation of neutrino fluxes by bahcall and pinsonneault @xcite which incorporates the new normalization for the low energy cross section @xmath17 indicated by the recent studies at the institute of nuclear theory @xcite .
we have also considered the possibility of departing from the ssm of bp98 by allowing a free normalization of the @xmath11 flux and @xmath12 neutrino fluxes and we present the results we obtain when we treat these normalization as free parameters .
we combined the measured total event rates at chlorine , gallium and super kamiokande experiments with the super kamiokande data on the zenith angle dependence , energy spectrum and seasonal variation of the events .
the goal of such analysis is not only to compare the quality of the solutions to the solar neutrino anomaly in terms of flavour oscillations of @xmath0 into active or sterile neutrinos but also to study the weight of different observables on the determination of the underlying neutrino physics parameters as emphasized in ref .
the outline of the paper is as follows . in sec .
[ sec : data ] we present the basic elements that enter into our calculation of the observables and the definitions used in the statistical combination of the data .
section [ sec : fits ] contains our results of the allowed ( or excluded ) regions of oscillation parameters from the analyses of the different observables .
the results on the allowed regions from the combined analysis of the total event rates is contained in sec .
[ fit : rates ] .
the constraints arising from the super kamiokande searches for day night variation of the event rates are discussed in sec .
[ fit : dn ] . in sec .
[ fit : spec ] we discuss the information which can be extracted from the distortion of the recoil electron energy spectrum measured by super kamiokande .
the restrictions arising from the preliminary super kamiokande data on the seasonal variation of the event rates are studied in sec .
[ fit : sea ] . in sec .
[ fit : total ] we present our results from the global fit to the full data set and we determine the allowed range of oscillation parameters which are consistent with all the data . finally in sec .
[ conclu ] we discuss the possible implications of our results for future investigations .
our results show that for oscillation into active neutrinos although the data on the total event rates favours the sma angle solution , once the full data set is included both sma and lma give an equally good fit to the data .
we find that the best
fit points for the combined analysis are @xmath1 ev@xmath2 and @xmath3 with @xmath4 d. o. f. and @xmath18 ev@xmath2 and @xmath6 with @xmath7 d. o. f. we note that in contrast with the earlier 504-day study of ref . @xcite our results indicate that the lma solution is not only allowed , but actually slightly preferred .
the existence of hints that the lma msw solution could be correct was also discussed in ref.@xcite . on the other hand
we find good quantitative agreement with the recent results of @xcite as well as qualitative agreement with the old results of @xcite based on a smaller sample .
we show that seasonal effects may be no - negligible in the lower part of the lma region , without conflict with the negative hints of a day - night variation .
in particular the best - fit lma solution predicts a seasonal effect of 8.5% , 6% of which is due to the eccentricity of the earth s orbit . for conversions into sterile neutrinos
we find that only the sma solution is possible with best fit point @xmath19 ev@xmath2 and @xmath20 and @xmath10 d. o. f. we also consider departures of the standard solar model ( ssm ) of bahcall and pinsonneault 1998 ( bp98 ) by allowing arbitrary @xmath11 and @xmath12 fluxes .
these modifications do not affect significantly the oscillation parameters .
we find that the best fit is obtained for @xmath13 and @xmath14 for the sma solution both for conversions into active or sterile neutrinos and @xmath15 and @xmath16 for the lma solution .
in order to study the possible values of neutrino masses and mixing for the msw solution to the solar neutrino problem , we have used data on the total event rates measured at the chlorine experiment in homestake @xcite , at the two gallium experiments gallex and sage @xcite and at the water cerenkov detector super kamiokande . apart from total event rates we have in this case the zenith angle distribution of the events ,
the electron recoil energy spectrum and the seasonal distribution of events , all measured with their recent 825-day data sample @xcite .
we first describe our calculation of the different observables . for simplicity
we consider the two - neutrino mixing case @xmath21 where @xmath22 can label either an active , @xmath23 , or sterile neutrino , @xmath24 . in order to account for earth regeneration effects , we have determined the solar neutrino survival probability @xmath25 in the usual way , assuming that the neutrino state arriving at the earth is an incoherent mixture of the @xmath26 and @xmath27 mass eigenstates .
@xmath28 where @xmath29 is the probability that a solar neutrino , that is created as @xmath30 , leaves the sun as a mass eigenstate @xmath31 , and @xmath32 is the probability that a neutrino which enters the earth as @xmath31 arrives at the detector as @xmath0 .
similar definitions apply to @xmath33 and @xmath34 .
the quantity @xmath35 is given , after discarding the fast oscillating terms , as @xmath36\ ] ] where @xmath37 denotes the improved landau - zener probability @xcite and @xmath38 is the mixing angle in matter at the neutrino production point . in our calculations of the expected event rates
we have averaged this probability with respect to the production point .
the electron and neutron number density in the sun and the production point distribution were taken from ref .
@xcite . in order to obtain the conversion probabilities in the earth , @xmath39
, we integrate the evolution equation in matter assuming a step function profile of the earth matter density ( the earth as consisting of mantle and core of constant densities equal to the corresponding average densities , @xmath40 @xmath41 and @xmath42 @xmath41 ) . to convert from the mass density to electron and neutron number density
we use the charge to nucleon ratio @xmath43 for the mantle and @xmath44 for the core . in the notation of ref .
@xcite , we obtain for @xmath45 @xmath46 where @xmath47 is the mixing angle in vacuum and the earth matter effect is included in the formulas for @xmath48 and @xmath49 , which can be found in ref .
@xcite . @xmath34 depends on the amount of earth matter travelled by the neutrino on its way to the detector , or , in other words , on its arrival direction which is usually parametrized in terms of the nadir angle , @xmath50 , of the sun at the detector site . due to this effect
the survival probability is , in general , time dependent .
this earth regeneration effect is important in the study of the zenith angle distribution of events as well as in their seasonal variation @xcite .
here we update previous analyses of solar neutrino data @xcite by including the recent 825-day super kamiokande data sample .
we perform a dedicated analysis of the seasonal variation data and we anticipate its future role in discriminating between different solutions of the solar neutrino anomaly .
working in the context of the bp98 standard solar model of ref .
@xcite we also allow for a free normalization of the @xmath11 flux and of the hep flux . in our statistical treatment of the data we follow closely the analysis of ref .
@xcite with the updated uncertanties given in refs .
@xcite . in our study we use the measured rates shown in table [ rates ] . for the combined fit
we adopt the @xmath51 definition : @xmath52 where @xmath53 is the theoretical prediction of the event rate in detector @xmath54 and @xmath55 is the measured rate .
the error matrix @xmath56 contains not only the theoretical uncertainties but also the experimental errors , both systematic and statistical .
the general expression of the expected event rate in the presence of oscillations in experiment @xmath54 is given by @xmath53 : @xmath57 \nonumber,\end{aligned}\ ] ] where @xmath58 is the neutrino energy , @xmath59 is the total neutrino flux and @xmath60 is the neutrino energy spectrum ( normalized to 1 ) from the solar nuclear reaction @xmath61
@xcite with the normalization given in ref .
@xcite . here
@xmath62 ( @xmath63 ) is the @xmath0 ( @xmath64 , @xmath65 ) interaction cross section in the standard model @xcite with the target corresponding to experiment @xmath54 , and @xmath66 is the time averaged @xmath0 survival probability . for the chlorine and gallium experiments we use
improved cross sections @xmath67 @xmath68 from ref .
@xcite . for the super
kamiokande experiment we calculate the expected signal with the corrected cross section given in sec .
[ data : spec ] .
the expected signal in the absence of oscillations , @xmath69 , can be obtained from eq.([ratesth ] ) by substituting @xmath70 . in table [ rates ] we also give the expected rates at the different experiments which we obtain using the fluxes of ref .
@xcite .
as already mentioned , in the msw picture the expected event rates can be different when the neutrinos travel through the earth due to the @xmath0 regeneration effect @xcite . as a result in certain regions of the oscillation parameters
the expected event rates depend on the zenith angle of the sun as observed from the experiment site , since this determines the amount of earth matter crossed by the neutrino on its way to the detector .
the super kamiokande collaboration has studied the dependence of the event rates with the period of time along the day and the night .
they present their results in the form of a zenith angle distribution of events . in our analysis
we have used the experimental results from the super kamiokande collaboration on the zenith angle distribution of events taken on 5 night periods and the day averaged value , shown in table [ daynight ] which we graphically reduced from ref . @xcite we define @xmath51 for the zenith angle data as : @xmath71 where we have neglected the possible correlation between the errors of the different angular bins which could arise from systematical uncertainties .
the factor @xmath72 is included in order to avoid over - counting the data on the total event rate which is already included in @xmath73 .
we compute the expected event rate in the period @xmath54 in the presence of oscillations as , @xmath74 \nonumber\end{aligned}\ ] ] where @xmath75 measures the yearly averaged length of the period @xmath54 normalized to 1 , so @xmath76 .500 , .086 , .091 , .113 , .111 , .099 for the day and five night periods respectively .
super kamiokande has also presented their results on the day - night variation in the form of a day - night asymmetry , @xmath77 since the information included in the zenith angle dependence already contains the day - night asymmetry , we have not added the asymmetry as an independent observable in our fit .
notice also , that being a ratio of event rates , the asymmetry is not a gaussian - distributed observable and therefore should not be included in a @xmath51 analysis .
the super - kamiokande collaboration has also measured the recoil electron energy spectrum . in
their published analysis @xcite after 504 days of operation they present their results for energies above 6.5 mev using the low energy ( le ) analysis in which the recoil energy spectrum is divided into 16 bins , 15 bins of 0.5 mev energy width and the last bin containing all events with energy in the range 14 mev to 20 mev . below 6.5 mev
the background of the le analysis increases very fast as the energy decreases .
kamiokande has designed a new super low energy ( sle ) analysis in order to reject this background more efficiently so as to be able to lower their threshold down to 5.5 mev . in their 825-day data
@xcite they have used the sle method and they present results for two additional bins with energies between 5.5 mev and 6.5 mev . in our study
we use the experimental results from the super kamiokande collaboration on the recoil electron spectrum on the 18 energy bins including the results from the le analysis for the 16 bins above 6.5 mev and the results from the sle analysis for the two low energy bins below 6.5 mev , shown in table [ spectrum ] .
notice that in table [ spectrum ] we have symmetrized the errors to be included in our @xmath51 analysis .
we have explicitly checked that the exclusion region is very insensitive to this symmetrization .
we define @xmath51 for the spectrum as @xmath78 where @xmath79 again , we introduce a normalization factor @xmath80 in order to avoid double - counting with the data on the total event rate which is already included in @xmath73 . notice that in our definition of @xmath81 we introduce the correlations amongst the different systematic errors in the form of a non - diagonal error matrix in analogy to our previous analysis of the total rates .
these correlations take into account the systematic uncertainties related to the absolute energy scale and energy resolution , which were not yet available at the time the analysis of ref .
@xcite was performed .
note that our procedure is different from that used by the super - kamiokande collaboration .
however we will see in sec .
[ fit : spec ] that both methods give very similar results for the exclusion regions .
this provides a good test of the robustness of the results of the fits , in the sense that they do not depend on the details of the statistical analysis .
the general expression of the expected rate in the presence of oscillations @xmath82 in a bin , is given from eq.([ratesth ] ) but integrating within the corresponding electron recoil energy bin and taking into account that the finite energy resolution implies that the _ measured _ kinetic energy @xmath83 of the scattered electron is distributed around the _ true _ kinetic energy @xmath84 according to a resolution function @xmath85 of the form @xcite : @xmath86\ , \ ] ] where @xmath87 and @xmath88 mev for super
kamiokande @xcite .
on the other hand , the distribution of the true kinetic energy @xmath84 for an interacting neutrino of energy @xmath58 is dictated by the differential cross section @xmath89 , that we take from @xcite .
the kinematic limits are : @xmath90 for assigned values of @xmath91 , @xmath92 , and @xmath93 , the corrected cross section @xmath94 @xmath68 is given as @xmath95 in fig .
[ spectrum ] we show our results for the recoil electron spectrum in the absence of oscillations and compare it with the expectations from the super kamiokande monte - carlo .
we see that the agreement is excellent . in this figure
no normalization has been included .
after 825 days of operation , super
kamiokande has also presented preliminary results on the seasonal variation of the neutrino event rates @xcite which seem to hint at a seasonal variation of the data , especially for recoil electron energies above 11.5 mev ( see table [ seasonal ] ) . as discussed in ref .
@xcite the expected msw event rates do exhibit a seasonal effect . in the lma region
such dependence can be expected mainly due to the different night duration throughout the year at the experimental site and also due to the different averaged earth densities crossed by the neutrino during the night periods , which lead to a seasonal - dependent @xmath0 regeneration effect in the earth . on the other hand , in the sma region ,
earth matter effects are due to resonant conversion of neutrinos in the earth which is only possible when neutrinos travel both through the mantle and the core @xcite .
thus we find that in the sma region the seasonal variation is associated with the fact that at super
kamiokande site only in october - march nights there are neutrinos arriving at sufficiently low zenith angle to satisfy the resonant condition .
we define @xmath51 for the seasonal variation data as , @xmath96 where , as before , the normalization factor @xmath97 is introduced to avoid double - counting .
taking into account the relative position of the super kamiokande setup in each period of the year , we calculate the distribution of the events as @xmath98 here @xmath99 months and @xmath100 is obtained from eq.([ratesth ] ) but using the time
dependent survival probabilities and integrating the recoil electron energy above 11.5 mev .
notice that , unlike in the 708 days data sample , in order to compare our results with the recent data on the seasonal dependence of the event rates from the super kamiokande collaboration for the 825 data sample , we have not included the geometrical seasonal neutrino flux variation due to the variation of the sun - earth distance arising from the earth s orbit eccentricity because the new super
kamiokande data is already corrected for this geometrical variation .
we now turn to the results of our fits with the observables described above .
we have obtained the regions of allowed oscillation parameters @xmath101-@xmath102 by obtaining the minimum @xmath51 and imposing the condition @xmath103+@xmath104 where @xmath105 for 90% ( 99% ) cl regions .
we first determine the allowed range of oscillation parameters using only the total event rates of the chlorine , gallium and super kamiokande experiments .
the average event rates for these experiments are summarized in table [ rates ] .
we have not included in our analysis the kamiokande data @xcite as it is well in agreement with the results from the super kamiokande experiment and the precision of this last one is much higher @xcite . for the gallium experiments we have used the weighted average of the results from gallex @xcite and sage @xcite detectors . using the predicted fluxes from the bp98 model
the @xmath51 for the total event rates is @xmath106 for 3 d. o. f. this means that the ssm together with the sm of particle interactions can explain the observed data with a probability lower than @xmath107 ! in the case of active - active neutrino oscillations we find that the best
fit point is obtained for the sma solution with @xmath108 what implies that the solution is acceptable with a 55% cl .
there are two more local minima of @xmath51 .
one is the lma solution with the best fit for @xmath109 which is acceptable with at 91% cl .
the other is the low solution with best fit point @xmath110 which is only acceptable at 99% cl . in the case of active - sterile neutrino oscillations
fit point is obtained for the sma solution with @xmath111 acceptable at 90% cl .
the lma and low solutions are not acceptable for oscillation into sterile neutrinos . in those regions
@xmath112 implying that they are excluded at the 99.999% cl .
unlike active neutrinos which lead to events in the super kamiokande detector by interacting via neutral current with the electrons , sterile neutrinos do not contribute to the super kamiokande event rates .
therefore a larger survival probability for @xmath11 neutrinos is needed to accommodate the measured rate . as a consequence a larger contribution from @xmath11 neutrinos to the chlorine and gallium experiments is expected , so that the small measured rate in chlorine can only be accommodated if no @xmath113 neutrinos are present in the flux .
this is only possible in the sma solution region , since in the lma and low regions the suppression of @xmath113 neutrinos is not enough . in fig .
[ fig : rates ] we show the 90% and 99% cl allowed regions in the plane @xmath101-@xmath102 . the best fit points in each region are marked .
we find that as far as the analysis of the total rates is concerned , there is no substantial change in the best
fit points in the three regions as compared to the previous most recent analysis including the 504 days of super kamiokande data @xcite .
we have also considered the possibility of departing from the ssm of bp98 by allowing a free normalization of the @xmath11 flux and we treat this normalization as a free parameter @xmath114 in our analysis .
figure [ fig : ratesbo ] shows the allowed regions in the msw parameter space when @xmath114 is allowed to take arbitrary values .
the best fit sma solution is obtained for @xmath115 the best fit for the lma solution occurs at : @xmath116 and the low solution has its best fit point at @xmath117 and therefore coincides with the one obtained in eq.([lowrates ] ) . in the case of active - sterile neutrino oscillations
fit point is obtained for the sma solution : @xmath118 for all solutions we find that the main effect of considering a free normalization of the @xmath11 flux is upon the quality of the fits , as measured by the depth of the @xmath51 . next comes the position of the best
fit points , mainly a reduction in the value of the mixing angle .
the allowed regions are considerably enlarged as can be seen by comparing figs .
[ fig : rates ] and [ fig : ratesbo ] .
we now study the constraints on the oscillation parameters from the super kamiokande collaboration measurement of the zenith angle distribution of events .
we use the data taken on 5 night periods and the day averaged value shown in table [ daynight ] which we graphically reduced from ref . @xcite .
considering only the zenith angle data ( not including the information from the total rates ) , the best
fit point in the case of active - active neutrino oscillations is obtained for @xmath119 ev@xmath2 and @xmath120 with @xmath121 for 3 d. o. f. and in the case of active - sterile neutrino oscillations is obtained for @xmath122 ev@xmath2 and @xmath123 with @xmath124 . with these values
we calculate the excluded region of parameters at the 99% cl , shown in fig [ fig : ratesdn ] .
notice that the zenith angle data favours the lma solution of the solar neutrino problem .
when combining the information from both total rates and zenith angle data , we obtain that in the case of active - active neutrino oscillations the best
fit point is still obtained for the sma solution with @xmath125 for 6 d. o. f. which is acceptable with a 56% cl .
however the lma solution becomes relatively better with a local minimum at @xmath126 valid at 70% cl .
this difference arises from the fact that , although small , some effect is observed in the zenith angle dependence which points towards a larger event rate during the night than during the day , and that this difference is constant during the night as expected for the lma solution @xcite . in the sma solution , however , the enhancement is expected to occur mainly in the fifth night @xcite .
the low solution is almost un modified and presents the best
fit point at @xmath127 which is acceptable at 95% cl . in the case of active - sterile neutrino oscillations
fit point is obtained for the sma solution with @xmath128 valid at 77% cl .
figure [ fig : ratesdn ] shows the regions excluded at 99% cl by the zenith angle data alone , together with the 90% and 99% cl allowed regions in the plane @xmath101-@xmath102 from the combined analysis of rates plus zenith angle data .
the best fit points in each region are indicated . the main difference with respect to fig .
[ fig : rates ] is observed in the lma allowed region which is cut from below , as the expected day - night variation is too large for smaller neutrino mass differences .
we now present the results of the study of the recoil electron spectrum data observed in super kamiokande . using the method described in sec .
[ data : spec ] we obtain that the @xmath51 for the undistorted energy spectrum ( standard model case ) is 20.1 for 17 d. o. f .. this corresponds to an agreement of the measured with the expected energy shape at the 27% cl .
this value depends on the degree of correlation allowed between the different errors . in this way , reducing the correlation in the error matrix the agreement decreases down to 13% .
this suggests that the correlations amongst the different systematic uncertainties related to the energy resolution will play an important role in the analysis of the energy spectrum .
indeed in our definition of @xmath81 we introduce the correlations amongst the different systematic errors in the form of a non - diagonal error matrix in analogy to our previous analysis of the total rates .
these correlations take into account the systematic uncertainties related to the energy resolution .
we believe that these might constitute the main difference between our treatment of the spectrum data and the earlier one presented in ref .
@xcite , when this experimental information was still unavailable .
we find that the best fit to the spectrum in the msw plane for the case of active - active neutrino oscillations is obtained for @xmath129 ev@xmath2 and @xmath130 with @xmath131 .
for the case of active - sterile neutrino oscillations we get @xmath132 ev@xmath2 and @xmath133 with @xmath134 . with these values
we obtain the excluded region of parameters at the 99% cl shown in fig .
[ fig : ratesspec ] .
we see that our results for the exclusion regions are quantitatively very similar to those of the super - kamiokande collaboration @xcite , even though our procedure is different from theirs .
this provides a good test of the robustness of the results of the fit of the spectrum , in the sense that they do not depend on the details of the statistical analysis .
in contrast we note that our results are different from those of ref .
@xcite .
when combining the information from both total rates and the recoil energy spectrum data we obtain that in the case of active - active neutrino oscillations both sma and lma solutions lead to fits to the data if similar quality . in this way , the best
fit for the lma solution is @xmath135 while for the sma solution we find @xmath136 for 18 .
d. o. f. which are acceptable at 83% .
finally for the low solution we find @xmath137 with agreement with data only at the 8.5% cl . in the case of active - sterile neutrino oscillations the best
fit point is obtained for the sma solution with @xmath138 acceptable at 90% . in fig .
[ fig : ratesspec ] we plot the excluded region at 99% cl by the energy spectrum data together with the 90% and 99% cl allowed regions in the plane @xmath101-@xmath102 from the combined analysis . the best
fit points in each region are marked .
the main point here is that the oscillation hypothesis does not improve considerably the fit to the energy spectrum as compared to the no - oscillation hypothesis . in this connection
it has been suggested @xcite that better description can be obtained by allowing a larger flux of hep neutrinos as they contribute mainly to the end part of the spectrum . in order to account for this possibility we have also analysed the data allowing for a free normalization of the @xmath11 and @xmath12 fluxes , treating them as a free parameters @xmath114 and @xmath139 correspondingly .
when doing so we find that the no - oscillation hypothesis gives @xmath51 = 17.4 for 15 d. o. f. for @xmath140 and @xmath141 .
when combining the information from both total rates and the recoil energy spectrum data in the case of active - active neutrino oscillations we obtain that for the lma : @xmath142 for 16 d. o. f. which is acceptable at 66% while for the sma solution @xmath143 which is acceptable at 75% cl .
finally for the low solution we find @xmath144 acceptable at 93% . in the case of active - sterile neutrino oscillations the best
fit point is obtained for the sma solution : @xmath145 which is acceptable at 89% cl . in fig .
[ fig : specnorm ] we display the normalized expected energy spectra for sma , lma solutions for active - active oscillations and for no - oscillation with non - standard @xmath11 and @xmath12 fluxes .
recently super kamiokande has also presented preliminary results which seem to hint for a seasonal variation of the event rates , especially for recoil electron energy above 11.5 mev .
as explained in sec .
[ data : sea ] the expected msw event rates do exhibit a seasonal effect due to the the seasonal - dependent @xmath0 regeneration effect in the earth .
we explore here the constraints on the msw oscillation parameters which can be extracted from the seasonal dependence data given in table [ seasonal ] . considering only the seasonal variation data above 11.5 mev ( allowing a free normalization for the @xmath11 flux ) we find that the the ssm yields a value of @xmath146 for 7 d. o. f. this shows that the data is still not precise enough to enable one to draw any definitive conclusion .
however one can still obtain some preliminary information on the msw parameters from the analysis of these data . in this way , when allowing oscillations into active flavours we obtain the best
fit point for @xmath147 ev@xmath2 and @xmath148 with @xmath149 for 5 d. o. f. and in the case of active - sterile neutrino oscillations best fit is obtained for @xmath150 ev@xmath2 and @xmath151 with @xmath152 . in fig .
[ fig : season ] we show the exclusion regions at 95% and 99% cl .
the larger area represents the _ allowed _ region at 95% cl .
being an allowed region , it means that at 95% cl , some effect is observed .
the darker area shows the small excluded region at 99% cl .
since the seasonal variation of the event rates in the msw region is due to neutrino regeneration in the earth one expects a correlation with the day - night variation which arises from the same origin .
this correlation is observed as the 95% allowed regions in fig .
[ fig : season ] have a large overlap with the 99% excluded region from the observed zenith angle dependence in fig . [
fig : ratesdn ] .
notice however that the upper part of the 95% cl allowed region for the oscillation in active flavours ( larger @xmath101 and larger mixing angles ) is still in agreement with the observed zenith angle dependence .
thus , should the seasonal effect be confirmed in the higher energy part of the spectrum , it will favour the lma solution to the solar neutrino problem .
we now present our results for the simultaneous fits to all the available data and observables . in the combination
we define the global @xmath51 as the sum of the different @xmath51 defined above . in principle such analysis should be taken with a grain of salt as these pieces of information are not fully independent ; in fact , they are just different projections of the double differential spectrum of events as a function of time and energy .
thus in our combination we are neglecting possible correlations between the uncertainties in the energy and time dependence of the event rates . from the full data sample
we obtain that the for the ssm of ref .
@xcite @xmath153 for 32 d. o. f. so the probability of explaining the full data sample as a statistical fluctuation of the the ssm together with the sm of particle interactions is smaller than @xmath154 . in the msw oscillation region
we obtain that for oscillations into active flavours the best fits are obtained for the lma solution with @xmath155 for 30 d. o. f. which implies that the solution is acceptable at 77% cl . , while in the sma region the local best fit point is @xmath156 valid at the 83% cl .. the global analysis still presents a minimum in the low region with best fit point : @xmath157 which is acceptable at 90% cl .
the results of the global analysis for the case of of active - sterile neutrino oscillations , show the fit is slightly worse in this case than in the active - active oscillation scenario .
fit point is obtained for the sma solution with @xmath158 acceptable at 90% cl . in fig .
[ fig : ratesdnss ] we show the 90% ( lighter ) and 99% ( darker ) cl allowed regions in the plane @xmath101-@xmath159 .
best fit points in each regions are also indicated . by comparing figs .
[ fig : ratesdnss ] and [ fig : rates ] we see the effect of the inclusion of the full data from super - kamiokande on both time and energy dependence of the event rates . for oscillations into active neutrinos ,
the larger modification is in the lma solution region which has its lower part cut by the day - night variation data while the upper part is suppressed by the data on the recoiled energy spectrum .
the best fit point is also shifted towards a larger mass difference by a more than a factor 2 .
as for the sma solution region the position of the best fit point is shifted towards a slightly smaller mixing angle , while the size of the region at the 90% cl is reduced . at 99% cl the allowed sma region is very little modified . for oscillations into sterile neutrinos the best fit point is also shifted towards a smaller mixing angle .
we finally study the allowed parameter space with the normalization of the @xmath11 and @xmath12 fluxes are left free .
figure [ fig : ratesdnssbo ] shows the allowed regions in the msw parameter space when the normalizations are allowed to take arbitrary values .
the best fit for the lma solution occurs at @xmath160 which is acceptable at 64% cl .
the best fit sma solution is obtained for @xmath161 which is acceptable at 80% cl . , the best fit for
the low solution occurs at @xmath162 which is acceptable at 90% cl . ,
in the case of active - sterile neutrino oscillations @xmath163 which is acceptable at 85% cl .
, for all solutions we find that the main effect of considering free normalization of the @xmath11 and @xmath12 fluxes is upon the position of the best - fit points , mainly a reduction in the value of the mixing angle .
the allowed regions are also enlarged as can be seen by comparing figs .
[ fig : ratesdnss ] and [ fig : ratesdnssbo ] .
we have presented an updated global analysis of two - flavor msw solutions to the solar neutrino problem using the full data set corresponding to the 825-day super kamiokande sample plus chlorine , gallex and sage experiments .
in addition to all measured total event rates we included all super kamiokande data on the zenith angle dependence , energy spectrum and seasonal variation of the events .
we have given a comparison of the quality of different solutions of the solar neutrino anomaly in terms of msw conversions of @xmath0 into active and sterile neutrinos .
for the case of conversions into active neutrinos we have found that once the full data set is included both sma and lma solutions give an equally good fit to the data .
we find that the best
fit points for the combined analysis are @xmath164 ev@xmath2 and @xmath165 with @xmath4 d. o. f and @xmath166 ev@xmath2 and @xmath167 with @xmath7 d. o. f. in contrast with the earlier 504-day study of bahcall - krastev - smirnov our results indicate that the lma solution is slightly preferred . although small , there is a hint for seasonality in the data and this should therefore take into account in future studies , as we have indicated here .
defining the seasonal variation ( in percent ) as @xmath168 where @xmath169 is the expected event rate during the winter ( summer ) period , we find that the seasonal effect may still reach 11% in the lower part of the allowed lma region ( 6% comes from the variation expected from the geometric effect due to the eccentricity of the earth s orbit ) , without conflict with the negative hints of a day - night variation .
such seasonal dependence is correlated with the day - night effect and this in turn can be used in order to discriminate the msw from the vacuum oscillation solution to the solar neutrino anomaly .
we have performed a numerical study of this correlation @xcite which generalizes the estimate presented by smirnov at the 1999 edition of moriond for a constant earth density .
our results show that due to the earth matter profile the seasonal variation can be substantially enhanced or suppressed as compared to the expected value obtained with an average earth density .
one must notice , however , that the exact values of masses and mixing for which a significant enhancement is possible may depend on the precise model of the earth density profile .
the numerical results presented above were obtained using the step function profile and thus numerical differences with the results obtained with , for instance the prem model @xcite can be expected for a given point in the msw plane . with the results from the combined analysis one can also predict the expected event rates at future experiments .
for example , we find that the average oscillation probabilities for @xmath170 neutrinos in the different allowed regions for msw oscillations into active neutrinos is @xmath171 @xmath172 these results imply that , at borexino we expect a suppression on the number of events as compared with the predictions of the ssm of @xmath173 @xmath174 for conversions into sterile neutrinos only the sma solution is possible with best fit point @xmath8 ev@xmath2 and @xmath175 and @xmath10 d. o. f which leads to @xmath176 and one expects a larger suppression of events at borexino @xmath177 as a way to improve the description of the observed recoil electron energy spectrum we have also considered the effect of departing from the ssm of bahcall and pinsonneault 1998 by allowing arbitrary @xmath11 and @xmath12 fluxes .
our results show that this additional freedom does not lead to a significant modification of the oscillation parameters .
the best fit is obtained for @xmath13 and @xmath14 for the sma solution both for conversions into active or sterile neutrinos and @xmath15 and @xmath16 for the lma solution .
m. c. g .-
is thankful to the instituto de fisica teorica of unesp and to the cern theory division for their kind hospitality during her visits .
we thank comments by john bahcall , vernon barger , venya berezinsky , plamen krastev and alexei smirnov .
this work was supported by spanish dgicyt under grant pb95 - 1077 , by the european union tmr network erbfmrxct960090 and by brazilian funding agencies cnpq , fapesp and the pronex program .
99 r. davis , jr , d. s. harmer , and k. c. hoffman , phys .
rev . lett . * 20 * , 1205 ( 1968 ) j. n. bahcall , n. a. bahcall , and g. shaviv , phys .
lett . * 20 * , 1209 ( 1968 ) ; j. n. bahcall , r. davis , jr .
, sicence * 191 * , 264 ( 1976 ) . b. t. cleveland _ et al .
j. * 496 * , 505 ( 1998 ) .
talks by m. c. gonzalez - garcia and a. yu .
smirnov , proceedings of _ international workshop on particles in astrophysics and cosmology : om theory to observation _ valencia , may 3 - 8 , 1999 , to be published by nucl .
supplements , ed .
v. berezinsky , g. raffelt and j. w. f. valle ( http://flamenco.uv.es//v99.html ) j.n .
bahcall , n. cabibbo and a. yahil , phys .
* 28 * , 316 ( 1972 ) ; j. schechter and j. w. f. valle , phys . rev .
* d25 * , 774 ( 1982 ) ; j. w. f. valle , phys .
131b * , 87 ( 1983 ) ; g.b .
gelmini and j.w .
valle , phys .
142b * , 181 ( 1984 ) .
j. w. f. valle , phys .
* 199b * , 432 ( 1987 ) ; p. langacker and d. london , phys . rev . *
d38 * , 907 ( 1988 ) ; m.m .
guzzo , a. masiero and s.t .
petcov , phys .
b260 * , 154 ( 1991 ) ; e. roulet , phys . rev . * d44 * , 935 ( 1991 ) ; p.i .
krastev and j.n .
bahcall , hep - ph/9703267 ; e.k .
akhmedov , phys . lett .
* b213 * , 64 ( 1988 ) ; c. lim and w.j .
marciano , phys .
* d37 * , 1368 ( 1988 ) ; j.n .
bahcall , s.t .
petcov , s. toshev and j.w .
valle , phys .
* 181b * , 369 ( 1986 ) ; a. acker and s. pakvasa , phys . lett . *
b320 * , 320 ( 1994 ) .
gribov and b.m .
pontecorvo , phys . lett . * 28b * , 493 ( 1969 ) ; v. barger , k. whisnant , r.j.n .
phillips , phys . rev . *
d24 * , 538 ( 1981 ) ; s.l . glashow and l.m .
krauss , phys .
190b * , 199 ( 1987 ) ; v. barger , r.j .
phillips and k. whisnant , phys .
lett . * 65 * , 3084 ( 1990 ) ; s.l .
glashow , p.j .
kernan and l.m .
krauss , phys .
b445 * , 412 ( 1999 ) ; v. berezinsky , g. fiorentini and m. lissia , hep - ph/9811352 and hep - ph/9904225 ; for theoretical models see r.n .
mohapatra and j. w. f. valle , phys .
* 177b * , 47 ( 1986 ) .
j. bouchez _ et .
_ , z. phys .
* c32 * , 499 ( 1986 ) ; s. p. mikheyev and a. yu .
smirnov , _
86 massive neutrinos in astrophysics and in particle physics _ , proceedings of the sixth moriond workshop , edited by o. fackler and j. tr@xmath178n thanh v@xmath178n ( editions frontires , gif - sur - yvette , 1986 ) , pp . 355 ; s.p .
mikheyev and a.yu .
smirnov , sov .
30 ( 1987 ) 759 - 790 ; a. dar _ et .
phys . rev .
* d 35 * ( 1987 ) 3607 ; e. d. carlson , phys . rev . * d34 * , 1454 ( 1986 ) ; a.j . baltz and j. weneser ,
phys . rev .
* d50 * , 5971 ( 1994 ) ; a. j. baltz and j. weneser , phys . rev .
* d51 * , 3960 ( 1994 ) ; p. i. krastev , hep - ph/9610339 ; q.y .
liu , m. maris and s.t .
petcov , phys .
rev . * d56 * , 5991 ( 1997 ) ; m. maris and s.t .
petcov , phys .
* d56 * , 7444 ( 1997 ) ; j.n . bahcall and p.i .
krastev , phys . rev . *
c56 * , 2839 ( 1997 ) ; a. j. baltz and j. weneser , phys . rev . *
d35 * , 528 ( 1987 ) ; a. j. baltz and j. weneser , phys . rev . * d37 * , 3364 ( 1988 ) . | 0.5 cm we present an updated global analysis of two - flavor msw solutions to the solar neutrino problem .
we perform a fit to the full data set corresponding to the 825-day super kamiokande data sample as well as to chlorine , gallex and sage experiments . in our analysis
we use all measured total event rates as well as all super kamiokande data on the zenith angle dependence , energy spectrum and seasonal variation of the events .
we compare the quality of the solutions of the solar neutrino anomaly in terms of conversions of @xmath0 into active or sterile neutrinos . for the case of conversions into active neutrinos
we find that , although the data on the total event rates favours the small mixing angle ( sma ) solution , once the full data set is included both sma and large mixing angle ( lma ) solutions give an equally good fit to the data .
we find that the best
fit points for the combined analysis are @xmath1 ev@xmath2 and @xmath3 with @xmath4 d. o. f and @xmath5 ev@xmath2 and @xmath6 with @xmath7 d. o. f. in contrast with the earlier 504-day study of bahcall - krastev - smirnov our results indicate that the lma solution is not only allowed , but slightly preferred . on the other hand we show that seasonal effects , although small , may still reach 11% in the lower part of the lma region , without conflict with the negative hints of a day - night variation ( 6% is due to the eccentricity of the earth s orbit ) .
in particular the best - fit lma solution predicts a seasonal effect of 8.5% . for conversions into sterile neutrinos
only the sma solution is possible with best fit point @xmath8 ev@xmath2 and @xmath9 and @xmath10 d. o. f. we also consider departures of the standard solar model ( ssm ) of bahcall and pinsonneault 1998 ( bp98 ) by allowing arbitrary @xmath11 and @xmath12 fluxes .
these modifications do not alter significantly the oscillation parameters .
the best fit is obtained for @xmath13 and @xmath14 for the sma solution both for conversions into active or sterile neutrinos and @xmath15 and @xmath16 for the lma solution . |
the cosmic infrared background ( cirb ) is the aggregate of the short wavelength radiation from the era of structure formation following the decoupling of matter and radiation in the early universe , redshifted to near infrared wavelengths by cosmological redshifting and far infrared wavelengths by dust - reprocessing , i.e. absorption and re - emission of starlight by intervening dust .
particle decay models also allow for a contribution to the cirb from photons produced in the decay of weakly interacting massive particles such as big bang relic neutrinos @xcite . the diffuse infrared background experiment ( dirbe ) on the cosmic background explorer ( cobe , see @xcite ) , which observed the entire sky in 10 infrared wavelengths from 1.25 to 240 @xmath2 m , was primarily intended to measure the cirb .
a determination of the cirb turned out to be exceedingly difficult due to bright foreground contamination including galactic starlight and the strong thermal emission and reflection of sunlight by dust in the solar neighborhood in many of the dirbe bands , especially in the near infrared ( nir ) .
the dirbe team estimated this zodiacal contribution by fitting a computed brightness , based on a physical model of the interplanetary dust ( ipd ) cloud , to the weekly variations in the dirbe brightness due to cobe s changing position in the solar system as the earth orbited the sun @xcite .
the modeled brightness was computed by integrating a source function , multiplied by a three dimensional density distribution function , along the line of sight . using this model for zodiacal light subtraction ,
the dirbe team succeeded in measuring an isotropic cirb of 25 @xmath0 7 nw m@xmath4 sr@xmath1 ( 1170 @xmath0 328 kjy sr@xmath1 ) and 14 @xmath0 3 nw m@xmath4 sr@xmath1 ( 1123 @xmath0 241 kjy sr@xmath1 ) at 140 and 240 @xmath2 m respectively @xcite . in the nir bands , they reported only upper limits on the cirb . using the `` very strong no - zodi principle '' of @xcite , another model of the zodiacal light was determined and is described in @xcite and @xcite . after creating dirbe all sky maps with
this new model subtracted , the problem of subtracting galactic starlight in the nir was addressed by @xcite at 2.2 and 3.5 @xmath2 m by directly observing galactic stars in a @xmath5 region of the sky , smoothing the observed intensities to the dirbe pixel size and subtracting the resulting intensities from dirbe maps with the @xcite zodiacal light model subtracted .
@xcite used this same procedure of subtracting galactic starlight from the dirbe maps using stellar fluxes from the 2mass 2nd incremental release point source catalog ( psc ) in four regions in the north and south galactic poles .
@xcite showed that the dirbe 3.5 @xmath2 m intensity was strongly correlated with the dirbe 2.2 @xmath2 m intensity at high galactic latitudes and used this correlation to convert a lower limit on the 2.2 @xmath2 m cirb , based on galaxy counts , into a lower limit on the 3.5 @xmath2 m cirb .
@xcite extended the @xcite analysis to 13 regions at 2.2 @xmath2 m and combined the @xcite and @xcite techniques to obtain estimates of the cirb for those 13 regions at 3.5 @xmath2 m . in this work
we extend the @xcite analysis to 40 new regions to further constrain the cirb at 1.25 , 2.2 and 3.5 @xmath2 m .
the two main datasets in this paper are the zodi - subtacted dirbe maps and the 2mass all - sky point source catalog ( psc ) .
dirbe has a large @xmath6 square beam with a diagonal of @xmath7 .
pixel intensities in the dirbe maps are averages of all observations made while the beam was centered in a given pixel in the cobe quadrilateralized spherical cube ( csc ) projection . due to the large beam size ,
bright stars outside of a particular pixel will , depending on the exact center position and position angle of the beam , occasionally affect the observed brightness in that pixel . thus a thick buffer ring is needed around any studied field to keep bright stars outside the field from influencing the measured dirbe intensity .
the resulting inefficiency was minimized by using the darkest possible @xmath8 circular regions which have the largest possible area : perimeter ratio .
a list of dark spots was created by smoothing the 3.5 @xmath2 m dirbe zodiacal subtracted mission average ( zsma ) map to 1/64th it s original resolution , reducing the 393,216 @xmath9 pixels to 6144 pixels approximately @xmath10 using a straight average of nearest neighbor pixels and sorting the resulting low resolution map .
the 40 darkest regions not used in previous analyses were used .
however , the dirbe zsma project data set only uses a fraction of the dirbe data since extreme solar elongations were dropped .
therefore , a set of mission averaged zodiacal subtracted dirbe maps were created by @xcite using the zodiacal light model of @xcite .
the physical model is similar to the @xcite model , with an added constraint .
the `` very strong no - zodi principle , '' described in @xcite and @xcite , requires that at high galactic latitudes , at 25 @xmath2 m , the background emission should be isotropic and adds a single pseudo - observation of zero emission to the nearly @xmath11 observations used to fit the time variation in the weekly maps to the model brightness . at 1.25 and 2.2 @xmath2 m , no correction for interstellar dust emission is needed , while at 3.5 @xmath2 m , there is a very small correction @xcite .
the @xmath12 pixel in these maps provides the dirbe data @xmath13 . using the nasa / ipac infrared science archive ( irsa )
, fluxes from all stars in each of the 40 regions brighter than k = 14 in the all - sky release of the 2mass psc were obtained .
again due to the large beam size and effectively random distribution of the beam center and position angle , these fluxes were converted into intensities by smearing with a @xmath6 square beam with a center uniformly distributed in the dirbe pixel and orientation uniformly distributed in position angle , to obtain the cataloged star contribution , @xmath14 , to the dirbe intensity .
this smearing process is described fully in section [ analysis ] .
for the j - band contribution , only stars with j and k less than 14 were used .
this dual wavelength magnitude selection is essentially equivalent to a simple @xmath15 14 selection @xcite .
the zodi - subtracted dirbe intensity in the @xmath12 pixel , @xmath13 , should be the sum of the cataloged stars , @xmath14 ; the faint stars , @xmath16 , assumed constant for all pixels in the @xmath17 region ; and the cirb , c , which is isotropic , i.e. @xmath18 the cataloged star contribution was computed using the `` smearing '' described in @xcite and @xcite , @xmath19 where @xmath20 is the solid angle of the dirbe beam , @xmath21 is the flux of the @xmath22 star and @xmath23 is the probability of the @xmath22 star affecting the @xmath12 pixel under the assumptions that over the many observations in which the dirbe beam was centered in the @xmath12 pixel , the beam center was uniformly distributed within the pixel and the beam orientation is uniformly distributed in position angle . due to the random distribution of the orientation of the square @xmath6 beam , the probablility as a function of the angular distance , r[(@xmath24 ) ] , between the beam center ( @xmath25 ) and a particular star at ( @xmath26 ) is @xmath27 where @xmath28 is the width of the beam ( 0.7 deg ) and , @xmath29 \ , .\ ] ] then , to account for the random position of the beam center within the pixel , this probability must be averaged over the area of the pixel by integrating p(r ) over the solid angle of the pixel and dividing by the pixel solid angle , @xmath30 , so that @xmath31 uncertainties in @xmath14 were calculated as in @xcite : @xmath32{s_j}^2 \ , .\ ] ] the first term is noise due to stars near the edge of the dirbe beam which will``flicker '' in and out of the beam when observed with various centers and position angles .
the second term is the flux error with an added allowance for variation in the fluxes between the dirbe and 2mass observations as in @xcite , with the modification that the allowance for variable stars was reduced from 0.1 to 0.001 , reducing this allowance from @xmath3 = 0.34 to @xmath3 = 0.03 magnitudes .
statistically , the @xmath3 s computed with the @xcite value of 0.1 were too large to be justifiable . upon dividing the residuals from the linear fit to the dirbe vs. 2mass intensities ( described below ) by the computed @xmath3 s
, we noticed that the residual/@xmath3 at all pixels was less than one .
the new value of 0.001 gives a statistically more reasonable distribution of residual/@xmath3 values which is described at the end of this section .
the error estimates for all stars brighter than k = 5.5 or j = 6.5 were set to @xmath0 1 mag to effectively remove pixels affected by confusion due to saturation from the final analysis .
stars with reported null uncertainties were assigned an uncertainty of @xmath0 0.5 mag .
since the error in the dirbe data is negligible @xcite , all of the error comes from the calculation of @xmath33 and is ascribed to @xmath34 for the fits .
figures [ kbdz ] , [ jbdz ] and [ lbdz ] show plots of @xmath13 vs. @xmath14 for all pixels in each of the 40 regions in k , j , and l respectively where the point sizes are inversely proportional to the above @xmath3 s . the fits in k ,
j and l have slopes of @xmath35 = 0.88 , @xmath36 = 0.97 and @xmath37 = 0.43 respectively with 40 independent intercepts in each band , derived using a weighted median procedure , i.e. finding the values of @xmath38 and dz(0 ) that minimize the sum : @xmath39 derived intercepts for each field are given in tables [ ktable ] , [ jtable ] and [ ltable ] for k , j , and l respectively .
the contributions from stars fainter than the 14th magnitude were evaluated statistically by fitting a power series of the form @xmath40(m ) = @xmath41 to counts of 2mass stars in each of the 40 regions , binned into 3 one - magnitude bins centered on m= 11.5 , 12.5 and 13.5 .
the fits resulted in 40 individual @xmath42 and @xmath43 and @xmath44 where any @xmath45 results in a converging flux contribution .
the intensity contribution from faint stars in the @xmath17 region with solid angle @xmath46 is then @xmath47 or , in the limit of infinitely fine bins , @xmath48 which was computed analytically . at l ,
the faint source contribution in each region is the 2.2 @xmath2 m value multiplied by the calibration ratio of 0.491 .
these f@xmath49 are listed in tables [ ktable ] , [ jtable ] and [ ltable ] for k , j and l. an uncertainty of 20% of the total prediction is assigned to this correction , and is listed in table [ error ] under `` faint source . ''
the cirb in each region is then the derived intercept , dz(0 ) , minus the faint star contribution f@xmath49 , c = dz(0 ) - f. these values are also listed in tables [ ktable ] , [ jtable ] and [ ltable ] . the mean of these cirb estimates are 14.59 @xmath0 0.14 , 8.83 @xmath0 0.51 and 15.57 @xmath0 0.20 kjy sr@xmath1 for k , j and l. these standard deviations of the means are listed in table [ error ] as `` scatter . ''
2mass magnitudes at k and j were converted into fluxes using f@xmath50(k ) = 614 jy and f@xmath50(j ) = 1512 jy which were derived by @xcite and @xcite . however , the derived slopes at k and j indicate dirbe fluxes for a zero magnitude 2mass star of 540 and 1467 jy respectively . the ratio of the calibration factors at 3.5 and 2.2 @xmath2 m is 0.491 , consistent with those found by @xcite , @xcite and @xcite .
uncertainties in the cirb due to calibration error were estimated using the change in the median dz(0 ) when the slopes , @xmath38 , were forced to change by @xmath05% at j and l , or @xmath010% at k due to the large difference @xcite between the fitted value of 0.88 and the expected value of 1 .
the change in the medians are @xmath51 2.24 , 1.77 , and 0.60 kjy sr@xmath1 at k , j and l respectively and are listed in table [ error ] under `` calibration . ''
there is a small correction for faint galaxies that appear in the 2mass psc .
these have been subtracted along with the galactic stars , but should be included in the cirb .
@xcite estimates this correction is 0.1 and 0.05 kjy sr@xmath1 at 2.2 and 1.25 @xmath2 m .
the 0.1 kjy sr@xmath1 correction at 2.2 @xmath2 m implies a 0.05 kjy sr@xmath1 correction at 3.5 @xmath2 m due to the relative calibration factor of 0.491 .
these corrections have been added back after taking the mean of c = dz(0 ) - f in the 40 regions .
thus , the final reported values of the cirb are the mean of the values c in tables [ ktable ] , [ jtable ] and [ ltable ] plus this correction .
an uncertainty of 100% of this correction is included in table [ error ] under `` galaxies . ''
@xcite adopt an uncertainty of 5% of the zodiacal intensity at the ecliptic poles .
these uncertainties are listed in table [ error ] under `` zodiacal . '' after adding errors in quadrature , we obtain a cirb of 14.69 @xmath0 4.49 kjy sr@xmath1 at 2.2 @xmath2 m , a weak limit of 8.88 @xmath0 6.26 kjy sr@xmath1 at 1.25 @xmath2 m and a cirb of 15.62 @xmath0 3.34 kjy sr@xmath1 at 3.5 @xmath2 m . figures [ kres ] , [ jres ] , and [ lres ] show histograms of the residuals @xmath52 for all 2971 pixels in k , j and l with interquartile ranges of 2.85 , 3.57 and 2.11 respectively . dividing these residuals by @xmath53 at each pixel gives a non - gaussian distribution which is tightly packed near zero with a few pixels extending out to large values . the number of pixels with @xmath54 less than \{0.5 , 1.0 , 2.0 , 3.0 } at k , j and l are \{2424 , 2832 , 2930 , 2951 } , \{2588 , 2873 , 2955 , 2955 } and \{2590 , 2867 , 2947 , 2955}. for comparison , with a gaussian distribution , one would expect these numbers to be , \{1138 , 2028 , 2835 , 2963 } ; more spread out in the center , with fewer pixels in the extended tail .
we have used a robust , least sum of absolute values fitting method , which , by it s nature , is insensitive to the few pixels with large residual/@xmath3 .
thus the fit for the slopes , @xmath38 , and intercepts , dz(0 ) , is to the pixels in the narrow core , and the outliers have little effect on their derived values and the final cirb values .
figures [ kecl ] , [ jecl ] , and [ lecl ] show the derived intercepts , dz(0 ) , vs. ecliptic latitude in the three bands . we see here the same trends with ecliptic latitude as were apparent in @xcite . while the k - band intercepts appear reasonably independent of ecliptic latitude , there is a strong trend in j and a slight negative trend in the l - band .
the zodiacal light is fainter at 3.5 @xmath2 m than at 2.2 @xmath2 m , and so the stronger dependence on @xmath55 at l than at k may seem surprising .
however , at 3.5 @xmath2 m , we begin to see thermal emission from the interplanetary dust , in addition to the scattered sunlight . from a modeling standpoint , this gives another free parameter which should provide a better fit , but from a physical standpoint , we are likely seeing a more complicated emission / scattering pattern on the sky which is more difficult to model correctly .
there is an overall scaling factor at each wavelength in the zodiacal light model , but the parameters that determine the physical shape of the dust cloud were fit simultaneously to observations at 8 dirbe bands .
the trend with ecliptic latitude at both j and l indicates a problem with the modeled shape of the cloud .
the better fit to the scattered sunlight at k could be a coincidence , rather than better modeling in that band .
this remaining solar system dependence in two of the three analyzed bands is evidence of a problem with the zodiacal light model that still prevents us from claiming a detection at 1.25 @xmath2 m .
the model may be improved by requiring that these dirbe minus 2mass intercepts be ecliptic independent simultaneously in all three bands . improvements to the zodiacal light model will be addressed in future work .
the results of this dirbe minus 2mass subtraction in these 40 regions of the sky give a statistically significant isotropic background of 14.69 @xmath0 4.49 kjy sr@xmath1 at 2.2 @xmath2 m and 15.62 @xmath0 3.34 kjy sr@xmath1 at 3.5 @xmath2 m where the uncertainty has not been significantly reduced since the dominant sources of error are systematic
. these results are consistent with earlier results , summarized in table [ oldvals ] , including @xcite , @xcite , @xcite and with the 13 similarly analyzed regions from @xcite all of which used the same zodiacal light model considered here @xcite . in the j - band
, we have weak limit on the 1.25 @xmath2 m cirb of 8.88 @xmath0 6.26 kjy sr@xmath1 .
this is also consistent with the 1.25 @xmath2 m values in table [ oldvals ] , which were reported in @xcite and @xcite , also using the same zodiacal light model used here .
the @xcite zodiacal light model gives a zodiacal intensity at the ecliptic pole 3.9 , 9.2 and 4.0 kjy sr@xmath1 lower at k , j and l. allowing for this difference in the models , these results are also consistent with those reported in @xcite and @xcite at 1.25 and 2.2 @xmath2 m and @xcite at 2.2 and 3.5 @xmath2 m , also shown in table [ oldvals ] .
the cumulative light from galaxies is a strict lower limit on the cirb .
a determination of the total contribution of resolved galaxies to the cirb was determined via galaxy number counts by @xcite . using the infrared array camera ( irac ) on the spitzer space telescope @xcite ,
surveys were done of the botes region , the extended groth strip and a deep image surrounding the qso hs 1700 + 6416 . after integrating the light from galaxies from the 10th to the 21st magnitudes , a total integrated intensity of 5.4 nw m@xmath4 sr@xmath1 ( 6.5 kjy sr@xmath1 ) at 3.6 @xmath2 m is reported , which is less than half of the cirb determined at that wavelength by this and other similar studies .
this 3.6 @xmath2 m value is the final entry in table [ oldvals ] .
this discrepancy may be partially resolved by improving the photometry of these and other spitzer surveys .
but , the dominant source of error in these dirbe minus 2mass measurements comes from the estimation of the zodiacal light , and based on the strong ecliptic dependence of our j - band results , it will likely take an improvement of the zodiacal light models to make significant progress in determining whether we can resolve this discrepancy with improvements in data analysis , or if we truly require some exotic diffuse source as suggested by @xcite .
it is also interesting to note another indirect constraint on the cirb from the attenuation of tev @xmath56-rays by @xmath57 pair production through collisions with cirb photons .
a recent attempt by @xcite to fit the tev spectrum of the blazar h1426 + 428 found that the spectrum can not be fit using only the integrated light from galaxies .
their best fit uses the wright ( 2001 ) determination of the cirb favoring the higher values determined from direct observations of the total sky brightness .
however , h.e.s.s .
observations of the blazars h2356 - 309 and 1es 1101 - 232 by @xcite suggest that the cirb can not be much higher than the lower limits from galaxy counts .
the @xcite 1.25 and 2.2 @xmath2 m values and the @xcite 3.5 @xmath2 m value were used by @xcite in fitting a model of the extragalactic background light ( ebl ) needed for estimating @xmath56-ray attenuation , resulting in their p1.0 ebl model which is the top curve in figure [ p1.0 ] .
this model gives ( 11,18,16 ) kjy sr@xmath1 at 1.25 , 2.2 and 3.5 @xmath2 m and assuming a power law blazar spectrum ( @xmath58 [ photons cm@xmath4 s@xmath1 tev@xmath1 ] @xmath61 @xmath62 ) results in a power law index for the spectrum of the blazar 1es 1101 - 232 of @xmath63 = -0.1 .
the model must be scaled down by a factor of 0.45 to the p0.45 model ( lower curve in figure [ p1.0 ] ) to give a power law index of at least 1.5 , which is considered by @xcite to be the lowest acceptable value .
@xcite show however that an ebl model , also based on the @xcite values at 1.25 and 2.2 @xmath2 m , but with a steeper decline from 4 to 10 @xmath2 m results in a power law index of @xmath63 = + 0.5 .
they consider @xmath63 = 0.6 to be the lowest acceptable index based on physical considerations and suggest then that while the dirbe minus 2mass cirb values at 1.25 and 2.2 @xmath2 m from @xcite do require a hard spectrum , and the lower limits from galaxy counts are favored , they , and the slightly lower values reported here ( ( 9,15,16 ) kjy sr@xmath1 ) , can not be ruled out based on the current h.e.s.s .
while the implications of @xmath56-ray attenuation for cirb measurement are still limited by the small number of observed sources , this independent limit on the cirb will only improve as more blazars are observed by h.e.s.s . and
eventually , veritas .
there is still a substantial difference between the @xcite lower limits from galaxy counts and the intensities determined here .
progress can be made in resolving this discrepancy with improvements in the photometry of survey data from _
spitzer , as well as improvements in the zodiacal light models and further data on gamma - ray attenuation .
however , the dominant source of error in directly measuring the cirb , the model based subtraction of the zodiacal light , will not be significantly reduced with currently available data .
a directly measured map of the zodiacal light , which would have to be observed from outside the bulk of the ipd cloud , beyond about 3au , would allow accurate removal of the zodiacal light from the dirbe maps and thus an accurate , direct measurement of the cosmic infrared background .
we have shown here that the subtraction of catalogued stars from low resolution maps works well , thus an instrument with a field of view of a few square degrees and a resolution of a few arcminutes would suffice .
the main requirements would be sensitivity to extremely low surface brightnesses , down to less than 1 nw m@xmath4 sr@xmath1 for good signal to noise , and an accurate , absolute flux calibration .
such an instrument could be a camera on a probe to one of the outer planets .
it would also be useful to have observations from different positions with respect to the ipd cloud .
this could be accomplished either by observing the same fields at widely different solar elongations during a long lived mission as the craft orbits the sun , or by observing during the cruise from 1 to 3 au as the dust density decreases . while we will continue to improve our understanding of the cirb in the mean time , a space mission of this type
will ultimately be required .
_ the cobe datasets were developed by the nasa goddard space flight center under the direction of the cobe science working group and were provided by the nssdc .
this publication makes use of the data product from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center , funded by the national aeronautics and space administration and the national science foundation .
this research has also made use of the nasa/ ipac infrared science archive , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration aharonian , a. g. et al .
2006 , nature 440 1018 - 1021 arendt , r. et al .
1998 , , 507 , 74 boggess , n. w. et al .
1992 , , 397 , 420 bond , j. , carr , b. & hogan , c. 1986 , , 306 , 428 cambresy , l. , reach , w. , beichman , c. & jarrett , t. 2001 , , 555 , 563 dwek , e. & arendt , r. 1998 , , 508 , l9 eisenhardt , p. , et al .
2004 , , 154 , 48e fazio , g. , et al .
2004 , , 154 , 39f gorjian , v. , wright , e. , & chary , r. 2000 , , 536 , 550 hauser , m. et al .
1998 , 508 , 25 kelsall , t. , et al .
1998 , , 508 , 44 mapelli , m. , salvaterra , r. & ferrara , a. 2004 , pos bdmh2004 012 astro - ph/0411134 mapelli , m. , salvaterra , r. & ferrara , a. 2006 , new astronomy 11 420 - 430 matsumoto , t. 2000 , the institute of space and astronautical science report sp no .
14 , p. 179
matsumoto , t. et al .
2005 , , 626 , 31 wright , e. 1997 baas , 29 , 1354 wright , e. 1998 , 496 , 1 wright , e. & reese , e. 2000 , , 545 , 43 wright , e. 2001 , , 553 , 538 wright , e. & johnson , b. 2001 , astro - ph/0107205 v2 ccccccc 204.4 & 67.8 & 20.9 & 79 & 18.83 & 4.11 & 14.72 + 293.9 & -72.5 & -46.2 & 74 & 19.71 & 4.55 & 15.16 + 191.5 & 72.1 & 26.6 & 75 & 19.43 & 3.83 & 15.60 + 227.8 & 75.1 & 19.6 & 75 & 19.37 & 3.70 & 15.67 + 150.5 & 67.6 & 41.0 & 73 & 19.11 & 3.86 & 15.25 + 193.5 & 76.9 & 27.5 & 75 & 18.23 & 3.68 & 14.55 + 208.7 & 78.3 & 24.7 & 77 & 20.89 & 3.72 & 17.17 + 146.7 & 69.3 & 41.6 & 75 & 19.09 & 3.88 & 15.21 + 150.6 & 72.6 & 38.9 & 72 & 19.18 & 3.90 & 15.28 + 164.0 & 79.0 & 33.2 & 77 & 17.72 & 3.73 & 13.99 + 158.3 & 81.1 & 38.2 & 74 & 20.53 & 3.77 & 16.76 + 127.6 & 53.8 & 57.0 & 76 & 19.03 & 4.93 & 14.10 + 143.2 & 79.4 & 36.5 & 73 & 18.41 & 3.62 & 14.79 + 166.0 & 86.3 & 31.2 & 71 & 16.98 & 3.67 & 13.31 + 133.5 & 80.8 & 36.7 & 73 & 18.00 & 3.63 & 14.37 + 317.5 & 76.8 & 19.7 & 74 & 18.24 & 3.96 & 14.28 + 115.4 & 65.7 & 52.4 & 74 & 19.40 & 4.18 & 15.22 + 355.1 & 81.7 & 27.7 & 75 & 18.29 & 4.07 & 14.22 + 43.5 & 82.9 & 34.1 & 73 & 18.05 & 3.88 & 14.17 + 115.8 & 56.4 & 60.2 & 76 & 18.99 & 4.74 & 14.25 + 72.2 & 81.9 & 36.9 & 73 & 18.55 & 3.89 & 14.66 + 15.3 & 81.9 & 30.9 & 77 & 18.15 & 3.85 & 14.30 + 79.3 & 79.1 & 40.4 & 72 & 17.78 & 3.90 & 13.88 + 347.9 & 76.1 & 24.5 & 74 & 18.85 & 4.10 & 14.75 + 49.7 & 79.4 & 36.5 & 74 & 18.32 & 4.67 & 13.65 + 111.6 & 57.0 & 60.6 & 75 & 18.97 & 4.61 & 14.36 + 93.3 & 70.0 & 49.5 & 72 & 16.84 & 4.02 & 12.82 + 103.0 & 57.8 & 61.5 & 73 & 19.89 & 4.80 & 15.09 + 42.2 & 72.6 & 39.0 & 72 & 18.98 & 4.23 & 14.75 + 56.4 & 68.5 & 44.7 & 75 & 18.28 & 4.25 & 14.03 + 83.8 & 63.3 & 55.3 & 73 & 18.65 & 4.44 & 14.21 + 109.1 & 46.2 & 70.8 & 73 & 19.66 & 5.75 & 13.91 + 84.7 & 60.4 & 58.2 & 75 & 18.48 & 4.73 & 13.75 + 74.3 & 62.1 & 54.5 & 76 & 18.76 & 4.52 & 14.24 + 44.6 & 64.0 & 43.2 & 76 & 19.31 & 4.61 & 14.70 + 69.8 & 61.2 & 53.8 & 76 & 17.79 & 4.82 & 12.97 + 45.1 & 59.8 & 44.4 & 71 & 20.29 & 5.07 & 15.22 + 252.6 & -71.8 & -45.9 & 74 & 18.93 & 4.42 & 14.51 + 255.8 & -59.2 & -57.6 & 77 & 20.02 & 5.13 & 14.89 + 254.1 & -53.0 & -61.4 & 74 & 20.27 & 5.60 & 14.67 + ccccccc 204.4 & 67.8 & 20.9 & 79 & 8.77 & 5.60 & 3.17 + 293.9 & -72.5 & -46.2 & 74 & 16.21 & 6.20 & 10.01 + 191.5 & 72.1 & 26.6 & 75 & 12.35 & 5.38 & 6.97 + 227.8 & 75.1 & 19.6 & 75 & 7.94 & 5.12 & 2.82 + 150.5 & 67.6 & 41.0 & 73 & 15.55 & 5.32 & 10.23 + 193.5 & 76.9 & 27.5 & 75 & 10.14 & 4.92 & 5.22 + 208.7 & 78.3 & 24.7 & 77 & 11.95 & 5.11 & 6.84 + 146.7 & 69.3 & 41.6 & 75 & 15.50 & 5.40 & 10.10 + 150.6 & 72.6 & 38.9 & 72 & 14.26 & 5.29 & 8.97 + 164.0 & 79.0 & 33.2 & 77 & 11.65 & 5.14 & 6.51 + 158.3 & 81.1 & 38.2 & 74 & 16.27 & 5.17 & 11.10 + 127.6 & 53.8 & 57.0 & 76 & 18.06 & 6.74 & 11.32 + 143.2 & 79.4 & 36.5 & 73 & 12.90 & 4.95 & 7.95 + 166.0 & 86.3 & 31.2 & 71 & 10.38 & 5.01 & 5.37 + 133.5 & 80.8 & 36.7 & 73 & 11.44 & 4.89 & 6.55 + 317.5 & 76.8 & 19.7 & 74 & 7.20 & 5.45 & 1.75 + 115.4 & 65.7 & 52.4 & 74 & 18.38 & 5.81 & 12.57 + 355.1 & 81.7 & 27.7 & 75 & 10.38 & 5.51 & 4.87 + 43.5 & 82.9 & 34.1 & 73 & 12.10 & 5.41 & 6.69 + 115.8 & 56.4 & 60.2 & 76 & 18.19 & 6.44 & 11.75 + 72.2 & 81.9 & 36.9 & 73 & 12.22 & 5.30 & 6.92 + 15.3 & 81.9 & 30.9 & 77 & 11.05 & 5.14 & 5.91 + 79.3 & 79.1 & 40.4 & 72 & 13.80 & 5.11 & 8.69 + 347.9 & 76.1 & 24.5 & 74 & 8.69 & 5.54 & 3.15 + 49.7 & 79.4 & 36.5 & 74 & 12.37 & 6.50 & 5.87 + 111.6 & 57.0 & 60.6 & 75 & 18.00 & 6.23 & 11.77 + 93.3 & 70.0 & 49.5 & 72 & 15.01 & 5.56 & 9.45 + 103.0 & 57.8 & 61.5 & 73 & 18.95 & 6.57 & 12.38 + 42.2 & 72.6 & 39.0 & 72 & 14.35 & 5.97 & 8.38 + 56.4 & 68.5 & 44.7 & 75 & 16.42 & 5.82 & 10.60 + 83.8 & 63.3 & 55.3 & 73 & 18.89 & 5.98 & 12.91 + 109.1 & 46.2 & 70.8 & 73 & 20.99 & 8.11 & 12.88 + 84.7 & 60.4 & 58.2 & 75 & 19.38 & 6.43 & 12.95 + 74.3 & 62.1 & 54.5 & 76 & 17.71 & 6.28 & 11.43 + 44.6 & 64.0 & 43.2 & 76 & 17.78 & 6.42 & 11.36 + 69.8 & 61.2 & 53.8 & 76 & 17.37 & 6.64 & 10.73 + 45.1 & 59.8 & 44.4 & 71 & 18.61 & 6.91 & 11.70 + 252.6 & -71.8 & -45.9 & 74 & 16.19 & 6.12 & 10.07 + 255.8 & -59.2 & -57.6 & 77 & 19.50 & 7.05 & 12.45 + 254.1 & -53.0 & -61.4 & 74 & 20.69 & 7.76 & 12.93 + ccccccc 204.4 & 67.8 & 20.9 & 79 & 19.54 & 2.00 & 17.54 + 293.9 & -72.5 & -46.2 & 74 & 18.35 & 2.21 & 16.14 + 191.5 & 72.1 & 26.6 & 75 & 19.50 & 1.86 & 17.64 + 227.8 & 75.1 & 19.6 & 75 & 20.18 & 1.80 & 18.38 + 150.5 & 67.6 & 41.0 & 73 & 18.07 & 1.88 & 16.19 + 193.5 & 76.9 & 27.5 & 75 & 18.78 & 1.79 & 16.99 + 208.7 & 78.3 & 24.7 & 77 & 20.18 & 1.81 & 18.37 + 146.7 & 69.3 & 41.6 & 75 & 17.44 & 1.89 & 15.55 + 150.6 & 72.6 & 38.9 & 72 & 17.77 & 1.90 & 15.87 + 164.0 & 79.0 & 33.2 & 77 & 18.02 & 1.81 & 16.21 + 158.3 & 81.1 & 38.2 & 74 & 18.37 & 1.84 & 16.53 + 127.6 & 53.8 & 57.0 & 76 & 16.46 & 2.40 & 14.06 + 143.2 & 79.4 & 36.5 & 73 & 17.70 & 1.76 & 15.94 + 166.0 & 86.3 & 31.2 & 71 & 17.17 & 1.79 & 15.38 + 133.5 & 80.8 & 36.7 & 73 & 17.60 & 1.77 & 15.83 + 317.5 & 76.8 & 19.7 & 74 & 18.76 & 1.93 & 16.83 + 115.4 & 65.7 & 52.4 & 74 & 17.30 & 2.03 & 15.27 + 355.1 & 81.7 & 27.7 & 75 & 19.11 & 1.98 & 17.13 + 43.5 & 82.9 & 34.1 & 73 & 17.74 & 1.89 & 15.85 + 115.8 & 56.4 & 60.2 & 76 & 16.68 & 2.31 & 14.37 + 72.2 & 81.9 & 36.9 & 73 & 17.46 & 1.90 & 15.56 + 15.3 & 81.9 & 30.9 & 77 & 17.65 & 1.87 & 15.78 + 79.3 & 79.1 & 40.4 & 72 & 16.60 & 1.90 & 14.70 + 347.9 & 76.1 & 24.5 & 74 & 18.71 & 1.99 & 16.72 + 49.7 & 79.4 & 36.5 & 74 & 16.85 & 2.27 & 14.58 + 111.6 & 57.0 & 60.6 & 75 & 16.02 & 2.24 & 13.78 + 93.3 & 70.0 & 49.5 & 72 & 15.86 & 1.96 & 13.90 + 103.0 & 57.8 & 61.5 & 73 & 16.52 & 2.34 & 14.18 + 42.2 & 72.6 & 39.0 & 72 & 17.72 & 2.06 & 15.66 + 56.4 & 68.5 & 44.7 & 75 & 16.26 & 2.07 & 14.19 + 83.8 & 63.3 & 55.3 & 73 & 16.54 & 2.16 & 14.38 + 109.1 & 46.2 & 70.8 & 73 & 16.30 & 2.80 & 13.50 + 84.7 & 60.4 & 58.2 & 75 & 17.07 & 2.30 & 14.77 + 74.3 & 62.1 & 54.5 & 76 & 16.22 & 2.20 & 14.02 + 44.6 & 64.0 & 43.2 & 76 & 17.36 & 2.24 & 15.12 + 69.8 & 61.2 & 53.8 & 76 & 16.09 & 2.34 & 13.75 + 45.1 & 59.8 & 44.4 & 71 & 17.57 & 2.47 & 15.10 + 252.6 & -71.8 & -45.9 & 74 & 18.90 & 2.15 & 16.75 + 255.8 & -59.2 & -57.6 & 77 & 17.91 & 2.50 & 15.41 + 254.1 & -53.0 & -61.4 & 74 & 17.76 & 2.73 & 15.03 + lcccl @xcite & & 16.4 @xmath64 4.4 & 12.8 @xmath0 3.8 & @xcite + @xcite & & 16.9 @xmath644.4 & 14.4 @xmath643.7 & @xcite + @xcite & 12 @xmath07 & 14.8 @xmath0 4.6 & & @xcite + @xcite & 10.1 @xmath0 7.4 & 17.6 @xmath0 4.4 & 16.1 @xmath0 4 & @xcite + _ this work & _ 8.9 @xmath646.3 & _ 14.7 @xmath0 4.5 & _ 15.6 @xmath0 3.3&_@xcite + @xcite & 25 @xmath0 6.3 & 20.5 @xmath64 3.7 & & @xcite + @xcite & 22.9 @xmath0 7.0 & 20.4 @xmath64 4.9 & & @xcite + @xcite & & 22.3 @xmath64 4 & 16.9 @xmath64 3.5 & @xcite + @xcite & & & @xmath65 6.5 & n.a . ( galaxy counts ) + _ _ _ _ _ lccc scatter & 0.51 & 0.14 & 0.20 + faint source & 1.17 & 0.85 & 0.42 + galaxies & 0.05 & 0.10 & 0.05 + calibration & 1.77 & 2.24 & 0.60 + zodiacal & 5.87 & 3.79 & 3.25 + quadrature sum & 6.26 & 4.49 & 3.34 + | with the release of the 2mass all - sky point source catalog , stellar fluxes from 2mass are used to remove the contribution due to galactic stars from the intensity measured by dirbe in 40 new regions in the north and south galactic polar caps . after subtracting the interplanetary and galactic foregrounds , a consistent residual intensity of 14.69 @xmath0 4.49 kjy sr@xmath1 at 2.2
@xmath2 m is found . allowing for a constant calibration factor between the dirbe 3.5 @xmath2 m and the 2mass 2.2 @xmath2 m fluxes
, a similar analysis leaves a residual intensity of 15.62 @xmath0 3.34 kjy sr@xmath1 at 3.5 @xmath2 m .
the intercepts of the dirbe minus 2mass correlation at 1.25 @xmath2 m show more scatter and are a smaller fraction of the foreground , leading to a still weak limit on the cirb of 8.88 @xmath0 6.26 kjy sr@xmath1 ( 1 @xmath3 ) . |
the layered cobalt oxides na@xmath1coo@xmath2 have attracted much attention in recent years , because of a rich variety of orders competing for the ground state . for @xmath4=0.75
, the material shows a large thermoelectric power @xcite , while the @xmath4=0.5 compound is a charge ordered insulator with @xmath13=50 k @xcite .
the magnetic properties are different for different na contents @xcite . at the side with @xmath4
larger than 0.5 , the dc susceptibility shows a curie - weiss temperature ( @xmath14 ) variation , while at the low- @xmath4 side the susceptibility is @xmath14-independent @xcite .
most interestingly , superconductivity with @xmath154.5 k emerges around @xmath4=0.3 when water molecules are intercalated in between the co layers @xcite .
it has been reported that @xmath6 can be tuned by varying @xmath4 in the range of @xmath160.38 @xcite .
one of the unsettled issues on the superconductivity is the symmetry of the gap function .
theoretically , a chiral(@xmath17 ) superconducting state was first proposed on the basis of rvb scenario @xcite . in a previous study , we found that the spin - lattice relaxation rate @xmath9 shows no coherence peak below @xmath6 , and follows a @xmath18 variation with @xmath19= 2.2@xmath51 at low @xmath14 @xcite .
this suggests that the superconductivity is of non - bcs type , but it has been pointed out that the data are insufficient to distinguish between some exotic states @xcite .
it has been shown that both chiral and pure @xmath20-wave states in the presence of disorder @xcite , or an odd - frequency @xmath21-wave state @xcite can be compatible with such @xmath14-variation of @xmath22 .
since identification of the gap function is the first step toward the understanding of the superconductivity , more work is needed .
another issue is how the electron correlations in the normal state evolve with doping , which should help understanding the mechanism for the superconductivity . in the octahedral crystal electric field environment , two @xmath23 orbitals of co
are located much higher in energy above @xmath24 orbitals .
the local density approximation ( lda ) band calculation shows that @xcite , among the three @xmath24 orbitals , the @xmath25 ( @xmath26 , where @xmath27-axis is the threefold rotation axis ) orbit lies higher in energy than the remaining two @xmath28 orbitals .
if one takes this picture @xcite , then the co@xmath29 state without na has one electron in the @xmath25 orbital ( spin 1/2 ) , with the @xmath28 orbitals fully occupied .
this single - orbital situation resembles the undoped cuprates ; adding na dopes electrons into the @xmath25 orbital .
in contrast , multiple - band models were also proposed @xcite , which draw similarities to the manganese oxides , another important class of strongly correlated materials . in this letter
, we address the above two issues by studying the evolution of the system with doping using nqr ( nuclear quadrupole resonance ) technique .
we present intensive data on the superconducting and the normal states over a wide range of na content .
the na@xmath30coo@xmath31@xmath32yh@xmath31o powder @xcite was synthesized following ref .
the precursor of na@xmath33coo@xmath2 was immersed in the br@xmath31/ch@xmath34cn solution to de - intercalate na .
the amount of br@xmath31 is listed in table 1 , where x means the theoretical value necessary to remove all the na .
the na content of the final product was determined by the ion coupled plasma method .
x - ray diffraction indicates that the samples are single phase except the 5x ( @xmath350.34 ) and 10x ( @xmath350.32 ) samples in which minor secondary phases are present .
the c - axis elongates as the na content decreases , in agreement with previous report @xcite .
@xmath6 was determined from the ac susceptibility and the @xmath9 measurements .
the previously reported sample @xcite was made with 10x br amount in a separate run and turned out to have an @xmath350.31 .
all other samples were made in the same run .
@xmath0co nqr measurements were carried out by using a phase - coherent spectrometer .
the nuclear magnetization decay curve is excellently fitted to the theoretical curves @xcite , with a unique @xmath22 component .
@*7l br amount & na content @xmath4 & c - axis length ( ) & a - axis length ( ) & @xmath6 ( k ) + 5x & 0.34 & 19.50 & 2.822 & 2.65 + 10x & 0.32 & 19.62 & 2.822 & 3.60 + 10x @xcite & 0.31 & 19.59 & 2.819 & 3.70 + 20x & 0.28 & 19.73 & 2.817 & 4.75 + 40x & 0.25 & 19.80 & 2.817 & 4.70 + 100x & 0.26 & 19.79 & 2.815 & 4.60 + figure 1(a ) shows a typical example of the three nqr transition lines arising from the nuclear spin @xmath36=7/2 of @xmath0co .
figure 1(b ) shows the @xmath37 3/2@xmath38@xmath375/2 transition line for various na contents .
firstly , the peak frequency increases as @xmath4 decreases down to @xmath4=0.28 .
analysis of three transition lines allows us to find that the @xmath39 increases as the na content decreases down to @xmath4=0.28 , but the asymmetry parameter @xmath40 decreases monotonically with decreasing na content , as seen in fig .
1(c ) and ( d ) , respectively . here
@xmath41 and @xmath40 are defined as @xmath42 , @xmath43 , with @xmath44 being the nuclear quadrupole moment and @xmath45 being the electric field gradient at the position of the nucleus @xcite .
the decrease of @xmath39 with increasing @xmath4 is consistent with doping electrons into the @xmath25 orbit @xcite .
secondly , for the 5x ( @xmath350.34 ) and 10x ( @xmath350.32 ) samples , there appear multiple peaks for the transition
. however , the additional peaks with smaller amplitude are not due to the minor impurity phase detected in the x - ray chart .
rather , we find that @xmath9 measured in different peaks differs only by 10% in magnitude and shows the identical temperature variation ; a clear decreases was seen below @xmath6 .
these peaks may have arisen from the ordering of na ions @xcite which are located above and below co. more detailed discussion will be published elsewhere @xcite .
we note in passing that the multiple peaks seen in the @xmath37 3/2@xmath38@xmath375/2 transition are not resolvable in the @xmath37 5/2@xmath38@xmath377/2 transition .
figure 2 shows the temperature dependence of the quantity @xmath46 for various @xmath4 .
all measurements were done at the @xmath37 3/2@xmath38@xmath375/2 transition . for the 5x and 10x samples ,
@xmath9 was measured at the main peak shown in fig .
two trends can be seen in fig .
2 . first , the absolute value increases with decreasing @xmath4 .
second , @xmath46 increases upon lowering @xmath14 , indicating the importance of spin correlations ; the smaller the @xmath4 the steeper the increase , indicating that the spin correlations become stronger as @xmath4 is reduced .
what is the nature of the spin correlations ?
one good way to tell is through inspection of the knight shift , which measures the spin susceptibility and , unlike the bulk susceptibility , is not affected by the presence of magnetic impurities . in a superconducting single crystal with @xmath350.3 , we found that the shift is @xmath14-invariant from @xmath6 up to @xmath14=120 k @xcite , which is in good agreement with the report of ning _
et al _ @xcite .
these results rule out the possibility of ferromagnetic spin fluctuations with wave vector @xmath47=0 , since in that case the knight shift would follow the temperature dependence of @xmath46 because both of them would be dominated by the susceptibility at wave vector @xmath48 0 . the results thus indicate that the spin fluctuations peak at finite wave vector .
the most likely candidate is antiferromagnetic ( af ) correlations @xcite , although its detailed @xmath47-dependence awaits clarification by future experiments , in particular neutron scattering .
below we analyze our data in terms of af correlations , in an attempt to quantify the discussion . in a general form , @xmath46
is written as @xmath49 in the @xmath500 limit , where @xmath51 is the hyperfine coupling constant , and @xmath52 is the imaginary part of the @xmath47-dependent , dynamical susceptibility .
if one assumes that there is a peak around the af wave vector @xmath44 , then one may have the following approximation : @xmath53 . here
@xmath54 denotes the contribution from wave vectors other than @xmath44 , which usually is dominated by the uniform susceptibility @xmath55 . for the af contribution
, we use the formula for a two - dimensional nearly antiferromagnet , namely , ( @xmath56 ( ref .
@xcite ) . we find that such model fits the data quite successfully , as depicted by the curves in fig .
the resulting fitting parameters are shown in figure 3 .
note that a small value of @xmath57 means that a system is close to the af instability ( @xmath57=0 ) .
the enhancement of the spin correlation ( decrease of @xmath57 ) with decreasing @xmath4 , as is the case in doped copper oxide high-@xmath6 superconductors , can be naturally understood in the single - orbital picture @xcite described earlier .
in contrast , the multiple - band theory predicts na content - insensitive , ferromagnetic spin correlations @xcite . turning to the quantity @xmath54 , its increase with decreasing @xmath4 can be attributed to the increase of density of state ( dos ) at the fermi level .
this result is consistent with the lda band calculation which found that the dominant fermi surface is a _ hole - like _ one and that the dos increases as the electron number is reduced @xcite .
this aspect is different from the case of doped cuprates where the dos decreases as decreasing carrier density .
we summarize our results in fig .
as @xmath4 decreases from 0.34 , @xmath6 increases , but reaches to a plateau of @xmath54.7 k for @xmath580.28 .
our results for high value of @xmath4 are in good agreement with the report by schaak _
et al _ @xcite .
a plateau has also been reported recently by chen _
et al _ and milne _ et al _ @xcite .
it is interesting to note that as the parameter @xmath57 decreases and @xmath59 increases , namely , as the electron correlation becomes stronger , @xmath6 increases .
next , we turn to the superconducting state . as can be seen already in fig .
2 , no coherence peak was seen just below @xmath6 in all samples , irrespective of na content and/or @xmath6 value
. however , the low-@xmath14 behavior is sample dependent .
we first show in fig .
4 the results for the sample with @xmath80.26 .
it can be seen that @xmath9 follows a @xmath10 variation down to the lowest temperature measured ( @xmath60 ) .
this result indicates unambiguously that there exists line - nodes in the gap function .
as we noted in previous publications @xcite , the gap function with line nodes generates an energy ( @xmath61)-linear dos at low @xmath61 which results in a @xmath18 ( @xmath19=3 ) dependence of @xmath9 . in the previous sample , @xmath9 did not follow an exact @xmath10 variation and became proportional to @xmath14 below @xmath62 , @xcite .
it has been proposed that many exotic superconducting states can be compatible with the data @xcite .
calculations incorporating impurity / disorder scattering showed that a @xmath63+i@xmath64 state which has a full gap is more consistent with the previous data @xcite .
our present result completely rules out the @xmath63+i@xmath64 state proposed by the @xmath65 model in an isotropic triangular lattice @xcite .
also , the odd frequency state proposed @xcite is gapless @xcite , and is incompatible with our finding in the @xmath4=0.26 sample .
however , if anisotropy of the @xmath66 term is incorporated , the theories could be compatible with our experimental results .
in fact , it has been recently shown that a small anisotropy in @xmath66 ( 10@xmath520% ) lifts the degeneracy between @xmath63 and @xmath64 states , leading to a stabilization of a pure @xmath20-wave state @xcite .
in other samples with either larger or smaller @xmath4 , however , @xmath9 deviates from the @xmath10 relation below @xmath14=2.0 k. to see this more clearly , we plot the @xmath14-dependence of @xmath9 for all samples in fig .
5 . in the inset of the figure
is shown the normalized @xmath9 by its value at @xmath67 against the reduced temperature . what is responsible for the deviation from the @xmath10 variation in samples with @xmath4 either larger or smaller than 0.26 ?
disorder or impurity is an unlikely candidate , since the samples with @xmath350.28 and @xmath350.25 have very similar or even higher @xmath6 and nqr line - width as the @xmath8 0.26 sample .
clearly , this unprecedented property deserves further investigation , both experimentally and theoretically .
meanwhile , low - lying excitations due to nearby competing orders may be a possible candidate . for larger @xmath4 ,
it is known that charge ordering @xcite occurs
. the na order could also give rise to the anisotropy of the hopping integral @xmath66 mentioned earlier .
for smaller @xmath4 , on the other hand , it has been proposed that different charge / magnetic order may occur at @xmath4=1/4 @xcite .
in conclusion , we have presented intensive @xmath0co- nqr measurements on the cobalt oxide superconductors na@xmath1coo@xmath2@xmath3h@xmath2o over a wide na content range @xmath4=0.25@xmath5 0.34 .
we find that @xmath6 increases with decreasing @xmath4 but reaches to a plateau for @xmath70.28 . in the superconducting state ,
the spin - lattice relaxation rate @xmath9 shows a @xmath10 variation down to @xmath11 for @xmath80.26 , which indicates unambiguously the presence of line nodes in the superconducting gap function . for both larger or smaller @xmath4
, however , @xmath9 deviates from the @xmath10 dependence even through the @xmath6 value is similar .
this unprecedented finding shows the richness of the physics of this new class of superconductors , and any pertinent theory must explain this peculiar property . in the normal state
, the temperature dependence of @xmath46 indicates that the spin correlations at finite wave - vector become stronger upon decreasing @xmath4 , while the dos increases .
this aspect is different from the case of doped cuprates but can be understood in terms of a single - orbital picture suggested on the basis of lda calculation .
we thank y. sun for x - ray chart analysis , g. baskaran , y. fuseya , h. harima , y. kitaoka , g.khaliullin , c.t .
lin , k. miyake , m. ogata and y. yanase for helpful discussions .
this work was supported in part by research grants from mext , nsf , the t. l. l. temple foundation , the john and rebecca moores endowment , and doe .
30 i. terasaki , y. sasago and k. uchinokura , phys . rev . *
b 56 * , r12685 ( 1997 ) .
yayu wang , n. rogado , r.j .
cava and n.p .
ong , nature * 423 * , 425 ( 2003 ) .
m. l. foo _
et al _ , phys .
lett . * 92 * , 247001 ( 2004 ) . c. barnhard _
et al _ , phys .
lett . * 93 * , 167003 ( 2004 ) .
a. t. boothroyd _
et al _ , phys .
lett . * 92 * , 197201 ( 2004 ) .
k. takada _
et al _ , nature * 422 * , 53 ( 2003 ) .
et al _ , nature * 424 * , 527 ( 2003 ) .
g. baskaran , phys .
lett . * 91 * , 097003 ( 2003 ) .
wang , q .- h . , d .- h . lee and p.a .
lee , phys . rev . * b 69 * , 092504 ( 2004 ) . m. ogata , j. phys .
. jpn . * 72 * , 1839 ( 2003 ) .
t. fujimoto , g. - q. zheng , y. kitaoka , r.l .
meng , j. cmaidalka and c.w .
chu , phys .
lett . * 92 * , 047004 ( 2004 ) .
et al _ , phys . rev .
* b 68 * , 212504 ( 2003 ) q - h wang and z.d .
wang , phys . rev . *
b 69 * , 092502 ( 2004 ) .
et al _ , phys .
rev . lett.*93 * , 097005 ( 2004 ) .
d. j. singh , phys .
rev . * b61 * , 13397 ( 2000 ) .
l. j. zou , j.l .
wang and z. zeng , phys . rev . * b69 * , 132505 ( 2004 ) .
j. kunes , k.w .
lee and w.e .
pickett , phys . rev . * b 70 * , 045104 ( 2004 ) .
g.khaliullin , w.koshibae , and s.maekawa , phys .
lett . 93,176401 ( 2004 ) .
y. yanase , m. mochizuki and m. ogata , j. phys .
jpn . * 74 * , 430 ( 2005 ) .
b. lorenz _
et al _ , phys . rev . *
b 68 * , 132504 ( 2003 ) .
a. abragam , _ the principles of nuclear magnetism _
( oxford univ . press , london , 1961 ) .
t. watanabe , h. yokoyama , y. tanaka , j. inoue and m. ogata , cond - mat/0411711 .
the knight shift contains substantial contribution from the spin susceptibility ( g. - q. zheng , k. matano , d. p. chen and c. t. lin , to be published ) .
et al _ , phys .
. lett . * 93 * , 237201 ( 2004 ) .
g. gasparovic _
et al _ , cond - mat/0508158 .
t. moriya , k. ueda , y. takahashi , j. phys .
jpn . * 59 * , 2905 ( 1990 ) .
p. zhang , w. luo , v. crespi , m. l. cohen , and s. louie1 , phys . rev . * b 70 * , 085108 ( 2004 ) .
i.i . mazin and m.d .
johannes , cond - mat/0506536 .
y. fuseya , h. kohno and k. miyake , j. phys .
jpn , * 72 * , 2914 ( 2003 ) .
g. baskaran , cond - mat/0306569 | + we report systematic measurements using @xmath0co nuclear quadrupole resonance ( nqr ) technique on the cobalt oxide superconductors na@xmath1coo@xmath2@xmath3h@xmath2o over a wide na content range @xmath4=0.25@xmath5 0.34 .
we find that @xmath6 increases with decreasing @xmath4 but reaches to a plateau for @xmath70.28 . in the sample with @xmath80.26 ,
the spin - lattice relaxation rate @xmath9 shows a @xmath10 variation below @xmath6 and down to @xmath11 , which unambiguously indicates the presence of line nodes in the superconducting ( sc ) gap function .
however , for larger or smaller @xmath4 , @xmath9 deviates from the @xmath10 variation below @xmath12 2 k even though the @xmath6 ( @xmath54.7 k ) is similar , which suggests an unusual evolution of the sc state . in the normal state , the spin correlations at a finite wave vector become stronger upon decreasing @xmath4 , and the density of states at the fermi level increases with decreasing @xmath4 , which can be understood in terms of a single - orbital picture suggested on the basis of lda calculation . |
trans - neptunian objects ( tnos ) are a dynamically diverse population of minor planets in the outer solar system .
a striking feature of the observed tnos is the significant number of objects found in mean motion resonance with neptune .
neptune s population of primordially captured resonant objects provides an important constraint on solar system formation and giant planet migration scenarios ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
but to understand these constraints on the early solar system , we first need to know the current resonant populations and orbital distributions .
identifying members of particular resonances is straightforward ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but using the observed set of resonant tnos to infer the intrinsic number and distribution of resonant objects is difficult due to complicated observational biases induced by the resonant orbital dynamics @xcite . here , we present the first set of 31 secure and 8 insecure resonant tnos detected by the outer solar system origins survey ( ossos ) , which was designed to produce detections with well - characterized biases @xcite .
ossos is a large program on the canada - france hawaii telescope surveying eight @xmath1 deg@xmath2 fields , some near the invariable plane and some at moderate latitudes from the invariable plane , for tnos down to a limiting magnitude of @xmath3 in r - band .
observations began in spring 2013 and will continue through early 2017 ( see @xcite for a full description of ossos ) .
two of the primary science goals for ossos are measuring the relative populations of neptune s mean motion resonances and modeling the detailed orbital distributions inside the resonances .
the most current observational constraints on both the distributions and number of tnos in neptune s most prominent resonances come from the results of the canada france ecliptic plane survey ( cfeps ; * ? ? ?
* hereafter referred to as g12 ) .
population estimates for some of neptune s resonances have also been modeled based on the deep ecliptic survey ( des ; * ? ? ?
ossos will offer an improvement on these previous constraints because it is optimized for resonant detections ( especially for the 3:2 resonance ) and includes off - invariable plane blocks to better probe inclination distributions . here
we report on the characterized resonant object detections from the first two of the eight ossos observational blocks : 13ao , an off - invariable plane block with a characterization limit of @xmath4 , and 13ae , a block overlapping the ecliptic and invariable planes with a characterization limit of @xmath5 .
the characterization limit is the faintest magnitude for which the detection efficiency of the survey is well - measured and for which all objects are tracked ( see @xcite for more details ) .
figure [ f : oe_blocks ] shows the location of the 13ao and 13ae blocks relative to a model ( g12 ) of neptune s 3:2 mean motion resonance .
the 13ao block is centered about @xmath6 above the ecliptic plane at the trailing ortho - neptune point ( @xmath7 in longitude behind neptune ) , and the 13ae block is at @xmath8 ecliptic latitude @xmath9 farther from neptune .
the full description of these blocks and the ossos observational methods are detailed in @xcite .
the 13ao block yielded 18 securely resonant tnos out of 36 characterized detections , and the 13ae block yielded 13 securely resonant tnos out of 50 characterized detections ; securely resonant objects are ones where the orbit - fit uncertainties fall within the width of the resonance ( see section [ s : detections ] and appendix [ a : orbits ] for the full list of resonant ossos detections and discussion of the classification procedure ) .
the larger yield of resonant objects in the 13ao block reflects both its favorable placement off the invariable plane near the center of the 3:2 resonance ( see section [ s:32 ] ) as well as its slightly fainter characterization limit .
our detections include secure 3:2 , 5:2 , 2:1 , 7:3 , and 7:4 resonant objects as well as insecure 5:3 , 8:5 , 18:11 , 16:9 , 15:8 , 13:5 , and 11:4 detections .
the 3:2 , 5:2 , and 2:1 resonances contain sufficiently many detections to model their populations . in this paper
, we use these detections and the survey s known biases to place constraints on the number and distribution of objects in these resonances .
we then discus how our constraints compare to the current theoretical understanding of the origins and dynamics of these tnos .
current models propose three possible pathways by which tnos may be captured into resonance .
first , they may have been captured by neptune as it smoothly migrated outward on a roughly circular orbit @xcite , driven by interactions between neptune and a primordial planetesimal disk @xcite .
a disk with the majority of its mass in planetesimals @xmath10100 km in radius can produce migration smooth enough for resonance capture @xcite , although formation of large planetesimals by the streaming instability @xcite , if efficient , could render planetesimal - driven migration too stochastic .
capture by a smoothly migrating neptune produces some objects that are deeply embedded in the resonance , having resonant angles ( see section [ s : methods ] ) that librate with low amplitude ( e.g. , * ? ? ?
given capture by smooth migration , the distribution of libration centers ( see section [ s:21 ] ) among 2:1 resonant objects serves as a speedometer , measuring the timescale of neptune s primordial orbital evolution @xcite .
smooth migration models predict that the 5:2 resonance captures fewer objects than the 3:2 and 2:1 @xcite and have difficulty producing the large inclinations observed in the kuiper belt , though @xcite recently suggest that transient resonant sticking and loss during slow migration may resolve the latter difficulty .
second , resonant objects could be the most stable remnants of a dynamically excited population that filled phase space in the outer solar system @xcite as a result of early dynamical instability among the giant planets ( e.g. , * ? ? ?
* ; * ? ? ?
the phase space volume of each resonance in which objects can have small libration amplitudes is limited , so this type of model preferentially produces larger - amplitude librators . because neptune spends time with high eccentricity
, such a scenario must be tuned to avoid disruption of the observed dynamically unexcited cold classical " tnos @xcite .
models of capture following dynamical instability may produce high inclination tnos more effectively than standard smooth migration models , but they still under - predict observations @xcite . like smooth migration , these models do not predict a large 5:2 population compared to the 3:2 and 2:1 populations ( * ? ? ?
* g12 ) .
third , resonant objects need not be primordial .
objects currently scattering off of neptune can be captured into resonance temporarily ( e.g. , * ? ? ?
* ; * ? ? ?
these marginally stable objects tend to have large libration amplitudes and may be a productive source of objects in distant resonances such as the 5:2 . inspired by the differences between these three emplacement mechanisms
, we focus our dynamical modeling on the libration amplitude distribution in the 3:2 resonance , the distribution of libration centers in the 2:1 resonance , and the relative abundance of objects in the 5:2 compared to the 3:2 and 2:1 .
finally , we emphasize that comparison of dynamical models of resonance capture with the current resonant populations must take into account the evolution of resonant orbits over the age of the solar system .
numerous theoretical studies of the current dynamics and stability of neptune s resonances ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) provide insight into this evolution .
we use these studies to inform our constructed orbital models .
farther from neptune.,width=307 ] we provide a brief description of neptune s resonances as well as our methods for modeling the resonant populations in section [ s : methods ] .
a table of the ossos resonant tno detections from the 13ae and 13ao blocks and a description of how we determine resonance membership is provided in section [ s : detections ] .
we model the distribution and total number of 3:2 objects based on the ossos detections in section [ s:32 ] , demonstrating that this population contains members with lower libration amplitudes than previously seen . in section [ s:52 ]
, we model the 5:2 resonance and confirm the unexpectedly large population of 5:2 resonant objects reported in g12 .
section [ s:21 ] presents a model of the 2:1 resonance and provides an improved constraint on the relative number of symmetric and asymmetric librators .
we summarize our population estimates and compare with previous results in section [ s : pop ] , and we comment on the populations of resonances for which we have insecure detections in section [ s : other ] .
section [ s : sum ] summarizes our key results .
after briefly introducing mean motion resonances ( section [ s : mmr ] ) , we summarize the challenges presented by detection bias for measuring resonant populations ( section [ s : bias ] ) . to circumvent these biases
, we employ the ossos survey simulator to test models of the resonant populations .
the survey simulator can only test models it can not produce them and we describe our choice of models in section [ ss : ssmod ] .
neptune s mean motion resonances ( which we call @xmath11 resonances , with @xmath12 for external resonances ) have resonant angles , @xmath13 , given by @xmath14 where @xmath15 , and @xmath16 are the mean longitude , longitude of perihelion , and longitude of ascending node ( the subscripts @xmath17 and @xmath18 refer to the elements of a tno and neptune ) , and @xmath19 are integers with the constraint that @xmath20 .
objects in a mean motion resonance have values of @xmath13 that librate around a central value with an amplitude defined as @xmath21 . for small eccentricity ( @xmath22 ) and inclination ( @xmath23 ) , the strength of the resonant terms in the disturbing function are proportional to @xmath24 @xcite , and resonances with small @xmath25 are generally stronger than those with larger @xmath25 .
tnos typically have eccentricities and inclinations much larger than neptune s , so we will ignore resonant angles involving @xmath26 and @xmath27 . likewise the resonant angles involving the inclination of the tno are typically less important than those involving the eccentricity because inclination resonances are at least second order in @xmath28 . throughout the rest of this work , we will generally consider this simplified resonance angle : @xmath29 with a few exceptions noted in table [ t : detections ] and section [ s : other ] . in most cases , such as in the 3:2 and 5:2 resonances ,
this resonant angle librates around @xmath30 .
the topology of n:1 exterior resonances allows for resonant orbits with more than one center of libration ; the 2:1 resonance has two so - called asymmetric libration centers near @xmath31 and @xmath32 ( the exact centers are eccentricity dependent ) in addition to the symmetric libration center at @xmath30 .
the libration of @xmath13 around specific values means that objects in resonance will come to perihelion at specific offsets from neptune s current mean longitude .
when a tno is at perihelion , its mean anomaly ( @xmath33 ) is 0 , so @xmath34 . substituting this into equation [ eq : phi ] shows that at perihelion : @xmath35 some resonances contain a subcomponent of objects also in the kozai resonance ; these objects exhibit libration of the argument of perihelion , @xmath36 , in addition to libration of the resonant angle @xmath13 .
this libration causes coupled variations in @xmath22 and @xmath23 such that the quantity @xmath37 is preserved .
outside of mean motion resonances , libration of @xmath38 only occurs at very large inclinations in the trans - neptunian region @xcite , but inside mean motion resonances kozai libration can occur at much smaller inclinations . in the 3:2 resonance , kozai libration can occur even at very low inclinations @xcite and a significant number of observed 3:2 objects are known to be in the kozai resonance , including pluto .
kozai resonance has also been observed for members of the 7:4 , 5:3 , and 2:1 resonances @xcite . in the 3:2 resonance ,
the libration of @xmath38 occurs around values of @xmath7 and @xmath39 with typical amplitudes of @xmath40 and typical libration periods of several myr . in order to be detected by ossos
, a tno must be in the survey s field of view , brighter than the limiting magnitude of the field , and moving at a rate of motion detectable by the survey s moving object detection pipeline ( see @xcite for more details ) ; because the ossos observing strategy is optimized to detect the motion of objects at distances between @xmath41 au , the first two criteria are the primary source of detection biases for the resonant objects .
the intrinsic brightness distribution of tnos with absolute magnitudes brighter than @xmath42 is generally well - modeled as an exponential in @xmath43 ( discussed in section [ ss : ssmod ] ) , meaning there are increasing numbers of objects at increasing @xmath43 ( decreasing brightness ) . for a population of tnos on eccentric orbits , this means that most detections will be made for faint , large-@xmath43 tnos near their perihelion .
consequently , populations containing preferentially fainter objects must have preferentially higher eccentricities to produce the same number of detections .
furthermore , given that resonant tnos come to perihelion at preferred longitudes relative to neptune ( equation [ eq : peri ] ) , this means that the placement of the field in longitude relative to neptune produces biases toward and against certain resonances .
objects in n:2 resonances librating about @xmath44 will preferentially come to perihelion at the ortho - neptune points ( @xmath45 away from neptune ) ; asymmetric n:1 librators will come to perihelion at various longitudes ahead or behind neptune , depending on the value of the libration center ( see figure 1 in g12 for an illustration of perihelion locations for various resonances ) .
the ossos 13ao and 13ae blocks are @xmath46 and @xmath47 behind neptune , which favors the detection of n:2 objects as well as asymmetric librators in the 2:1 resonance s trailing libration center . these biases for the 3:2 , 5:2 , and 2:1 resonances will be discussed in later sections .
similarly , latitude placement of the observing blocks relative to the ecliptic plane produces biases in inclination for tnos .
the 13ae block ( @xmath8 ecliptic latitude ) favors the detection of low-@xmath23 tnos because these tnos spend most of their time near the ecliptic plane , while in the 13ao block it is not possible to detect objects with inclinations smaller than the field s ecliptic latitude of @xmath48 . for resonances such as the 3:2
, the kozai subcomponent of the resonance introduces an additional observational bias ; the libration of @xmath38 means that kozai resonant objects come to perihelion at preferred ecliptic latitudes in addition to preferred longitudes with respect the neptune ( equation [ eq : peri ] ) .
the biases induced by the kozai resonance for the 3:2 population are discussed in detail by @xcite .
to account for these observational biases in our modeling , we use the ossos survey simulator .
we use the ossos detections of resonant objects combined with the ossos survey simulator to construct and test models of neptune s resonant populations .
the survey simulator is described in @xcite .
its premise is as follows : given a procedure ( i.e model ) for generating the position and brightness of resonant objects on the sky , the simulator repeatedly generates objects and then checks whether they would have been detected by the survey .
the simulator stops when the desired number of simulated detections is achieved .
when the model agrees with observations , the sets of real and simulated detected objects should have similar absolute magnitudes and orbital properties .
the intrinsic number of objects in a resonance ( i.e. a population estimate for the input model ) corresponds to the number of detected and undetected objects the survey simulator had to generate ( down to a specified absolute magnitude @xmath43 ) in order to match the real number of detections .
we run the survey simulator many times for each model with different random number generator seeds ; this allows us to build a distribution of population estimates and a large sample of simulated detections .
we then run statistical tests to determine whether the model provides simulated detections that are a good match to the real detections ; these tests are discussed later in this section as well as in appendix [ a : stats ] .
a resonant object s orbit is uniquely determined by its semi - major axis , @xmath49 , eccentricity , @xmath22 , inclination , @xmath23 , mean anomaly , @xmath50 , longitude of ascending node , @xmath16 , resonance angle @xmath13 , and epoch , @xmath51 , for the given value of @xmath33 .
following g12 , we construct a set of models for each resonant population by parameterizing the intrinsic distributions of @xmath49 , @xmath22 , @xmath23 , @xmath13 , and absolute magnitude @xmath43 . for each simulated object
, the simulator draws @xmath49 , @xmath22 , @xmath23 , @xmath13 , and @xmath43 from these models and then constructs the remaining orbital elements based on constraints from the resonant condition ( equation [ eq : phi ] ) .
we choose a uniformly - distributed random value for @xmath33 to reflect that the object s specific position within its orbit is random in time , and we draw @xmath49 randomly from a uniform distribution spanning the approximate resonance width .
appendix [ a : resmodeling ] provides the values used for the resonance widths , though we note that our results are not affected by this complication ; because the resonance widths are small , choosing a fixed @xmath49 for each resonance would produce equivalent results . for objects not experiencing kozai oscillations , the orientation of the orbit s plane relative to the ecliptic plane is not coupled to the resonance , so we also choose @xmath16 randomly from a uniform distribution . for the 3:2 population
, we include an additional parameter for the fraction of the population in the kozai resonance .
our procedure for selecting the orbital elements of these objects is described in section [ s:32 ] and appendix [ as:32 ] . in sections [ s:32 ] through [ s:21 ] and
appendix [ a : resmodeling ] we outline the exact models used , but the general form of the parameterized models in @xmath43 , @xmath22 , and @xmath23 is the same for each resonance .
we represent the cumulative luminosity distribution as an exponential in @xmath43 with logarithmic slope @xmath52 : @xmath53 where @xmath54 is the number of objects having magnitudes between a reference @xmath55 and @xmath43 .
this form models the absolute magnitude distribution well for @xmath56 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* g12 ) , but is not expected to work well for intrinsically fainter objects ( see section [ ss:32hd ] ) .
we model the differential eccentricity distribution as a gaussian centered on @xmath57 with a width @xmath58 : @xmath59 where @xmath60 is the number of objects with eccentricities between @xmath22 and @xmath61 .
this is a convenient form that acceptably describes populations with a typical eccentricity and a roughly symmetrical eccentricity dispersion .
following @xcite , we model the differential inclination distribution as a gaussian with width @xmath62 multiplied by @xmath63 : @xmath64 where @xmath65 is the number of objects with inclinations between @xmath23 and @xmath66 .
the @xmath13 distribution and treatment of the kozai resonance are specific to each resonance .
however for the 3:2 and 5:2 resonances , which have only one libration center @xmath30 , the @xmath13 distribution may be uniquely specified by a distribution of libration amplitudes , @xmath67 , about that center .
we approximate the time evolution of @xmath13 for an individual object as the oscillation of a simple harmonic oscillator with amplitude @xmath67 @xcite .
the instantaneous value of @xmath13 for a simulated object is then @xmath68 where @xmath51 is a random number distributed uniformly between 0 and 1 .
small - amplitude libration is well - approximated by a simple harmonic oscillator , while for large @xmath67 the angular evolution near the extrema of libration ( where @xmath69 changes sign ) slows less in full numerical simulations than equation [ eq : phitime ] implies .
this means that compared to full numerical simulations , equation [ eq : phitime ] slightly underestimates the likelihood that objects will be observed @xmath70 from neptune ( perihelion for @xmath71 ) and slightly overestimates the likelihood of finding objects at angles corresponding to the extrema of libration . however , in appendix [ as:32 ] we demonstrate that for all plutinos observed by ossos , full simulations of the resonant angle evolution do not deviate from equation [ eq : phitime ] enough to meaningfully affect our results . for resonances with a single libration center ,
we follow g12 and model the distribution of libration amplitudes as a triangle that starts at @xmath72 , rises linearly to a central value @xmath73 and then linearly falls to zero at the upper stability boundary for @xmath74 ( @xmath75 in the case of the 3:2 , @xcite ) .
this triangle need not be symmetric .
a triangular @xmath76 distribution is not an arbitrary choice ; theoretical studies of resonant phase space and of the dynamical capture and the evolution of plutinos often result in @xmath76 distributions that are roughly triangular in shape ( e.g. , * ? ? ?
* ; * ? ? ?
this outcome may be understood qualitatively as the result of shrinking phase space volumes at small libration amplitudes and increased dynamical instability at large amplitudes .
for example ,
plutinos with @xmath77 are not stable on gyr timescales @xcite .
when comparing real and simulated detections , we consider the following observables for each object : absolute magnitude @xmath43 , eccentricity @xmath22 , inclination @xmath23 , heliocentric distance at detection @xmath78 , and libration amplitude @xmath76 ( see section [ s : detections ] and appendix [ a : orbits ] for discussion of how @xmath76 is determined for the observed objects ) . because we are modeling each resonance separately , we do not consider the semimajor axis distribution within the resonance ; the small variations in @xmath49 for each object compared to the exact resonant value do not affect observability , so the @xmath49 distribution is not a useful model test . for plutinos , we also compare the observed and simulated fraction of objects in the kozai resonance ( see section [ ss:32kz ] ) . for resonances such as the 2:1 with symmetric and asymmetric libration centers , we compare the observed and simulated fractions of objects in each libration island ( see section [ s:21 ] ) .
we do not compare the @xmath16 , @xmath33 , or @xmath79 distributions of the simulated and real detections because these angles are related by @xmath13 .
we also do not compare the @xmath38 distributions because this angle is evenly distributed except in the case of kozai resonance . for each model test , we have two goals : ( 1 ) to determine the range of model parameters that provide acceptable matches with the data and ( 2 ) to determine the model parameters that best fit the data .
we note that our statistical approach is limited by computational feasibility .
ideally we would like to perform a maximum likelihood calculation , but the nature of the observational biases means we can not analytically calculate the detection probabilities ; they must instead be numerically determined by running the survey simulator .
given the wide range of possible models for the populations we are investigating , using the survey simulator to perform a maximum likelihood calculation is not currently feasible ( see appendix [ a : stats ] for a detailed explanation ) .
the observational biases affecting the @xmath23 and @xmath76 distributions are relatively independent of each other and of the chosen @xmath43 and @xmath22 distributions ; to reduce the complexity of model testing , we consider each of these observables separately as a one - dimensional distribution .
following @xcite and g12 , we use the anderson - darling ( ad ) test to identify a range of acceptable model parameters for these two distributions ( goal 1 above ) .
the test statistic described in appendix [ a : stats]is a weighted measure of the difference between two cumulative distributions . for each set of model parameters ,
we determine whether the set of @xmath23 and @xmath76 values for the real detections could be drawn from the simulated detections as follows : we generate a large number of synthetic detections for each model and calculate the ad statistic for the real detections compared to the model distribution .
we then determine the significance of that value of the ad statistic for the @xmath18 real objects by randomly drawing subsamples of @xmath18 synthetic detections from the model detections and calculating the ad statistic for these subsamples ( i.e. bootstrapping ) .
we reject a model if the ad statistic for the real detections is larger than the ad statistic for @xmath80 or more of the model subsamples compared to the model itself .
this procedure yields our @xmath80 confidence limits on the acceptable parameters ( @xmath62 and @xmath73 ) for our orbital model .
we note that this bootstrapping is required to produce confidence limits from the ad statistic because our distributions are not gaussian .
the ad test is a model rejection test , so to get a most probable values of @xmath62 and @xmath73 ( goal 2 ) , we must employ a different procedure .
we numerically generate one - dimensional probability distributions in @xmath23 and @xmath76 for each allowed value of the parameters @xmath62 and @xmath73 , calculate the probability of detecting the observed objects for each parameter value , and select the values of @xmath62 and @xmath73 that maximize this probability . see appendix [ a : stats ] for more details on how this calculation is done .
we note that a bootstrapping procedure to determine the significance of the calculated probabilities of @xmath62 and @xmath73 yield 95% confidence limits on those values that are very similar to the 95% confidence limits based on the ad statistic .
the observational biases that affect the @xmath43 , @xmath22 , and @xmath78 distributions are coupled such that the best parameters for the @xmath43 and @xmath22 distributions can not be determined independently of each other ( see discussions in @xcite and g12 ) . in these cases
we calculate the one - dimensional ad statistic for the observed @xmath43 , @xmath22 , and @xmath78 distributions compared to the model s synthetic detections .
following @xcite and @xcite , instead of calculating the significance of each of these statistics individually , we calculate the significance of the sum of the observed distribution s @xmath43 , @xmath22 , and @xmath78 ad statistics relative to the same sum for the model compared to itself .
we reject combinations of @xmath52 , @xmath58 , and @xmath57 for which the summed ad statistic is larger than @xmath80 of the summed statistics for the subsets of synthetic detections . to determine our preferred values of @xmath52 , @xmath58 , and @xmath57 within those @xmath80 confidence limits , we use the sum of their one - dimensional @xmath81-square values and select the @xmath52 , @xmath58 , and @xmath57 that minimizes this sum .
these values will not necessarily be the true , most probable values of @xmath52 , @xmath58 , and @xmath57 for our parameterized models because the one - dimensional @xmath81-square values do not account for how well the data fits the model in three - dimensional @xmath43 , @xmath22 , and @xmath78 space . correctly determining the most probable values of @xmath52 , @xmath58 , and @xmath57
is computationally too expensive for the wide range of parameter space we must explore ( see discussion in appendix [ a : stats ] ) .
we use the classification scheme outlined in @xcite to determine which ossos detections are resonant : a best - fit orbit for each ossos detection is computed using the @xcite algorithm and then a search around the best - fit orbit is done to find the maximum and minimum acceptable semimajor axis orbits . following @xcite , an orbit is deemed an acceptable fit to the observations if it meets two conditions : ( 1 ) the worst residual when comparing the observed astrometric position of the objects to the positions predicted by the orbit are not more than 1.5 times the worst residual for the best - fit orbit and ( 2 ) the rms residual is not more than 1.5 times the best - fit orbit s rms residual . the best - fit , minimum-@xmath49 , and maximum-@xmath49 orbits are integrated forward in time to look for resonant behavior ( defined as libration of a resonance angle described by equation [ eq : full_phi ] ) on @xmath82 year timescales ; we check all potential resonances with @xmath83 within 2% of the best - fit orbit s semimajor axis .
the resonant objects usually require more precise orbit fits than non - resonant objects in order to achieve secure classifications ( meaning all three orbits are resonant ) because uncertainty in an object s semimajor axis leads to uncertainty in the libration amplitude .
this classification procedure yielded 21 secure 3:2 objects , 4 secure 2:1 objects , 4 secure 5:2 objects , 1 secure 7:4 object , and 1 secure 7:3 object .
we also have 2 insecure 5:3 objects , and 1 insecure detection in each of the 11:4 , 8:5 , 18:11 , 16:9 , 15:8 , and 13:5 resonances .
these objects are listed in table [ t : detections ] along with their best - fit orbital parameters with uncertainties .
the listed uncertainty in @xmath49 is the @xmath84 uncertainty calculated from the @xcite orbit - fit covariance matrix ; the uncertainties in @xmath22 and @xmath23 are all small and rather than list them , these parameters have been reported to the appropriate number of significant figures .
the uncertainty in the libration amplitude is obtained by integrating 250 clones of each object s best - fit orbit ( obtained from the covariance matrix ) , measuring their @xmath76 distribution , and calculating the @xmath84 ( 68% ) confidence range ; see appendix [ a : orbits ] for a full discussion of the classification scheme and determination of the @xmath76 distributions .
many of the best - fit orbits @xmath76 distributions are asymmetric around the best - fit orbit s @xmath76 , and in these cases the ` @xmath84 ' uncertainties listed in table [ t : detections ] actually represent hard upper or lower limits to the value of @xmath76 ( see figure [ af : aphi - dist ] in appendix [ a : orbits ] ) ; these instances are marked in the table by asterisks .
we note that in many cases the orbit - fit uncertainties , and especially the libration amplitude uncertainties , are quite small even though the total arc length on the observations is only @xmath85 months ; this is due to the optimized observing schedule and accurate astrometry @xcite .
the libration amplitude uncertainties listed in table [ t : detections ] are comparable to or smaller than those determined for other tnos with significantly longer observational arcs ( see , for example , @xcite ) .
l l l l l r l l l l l o3e02 & 2013 gh137 & 3:2 & @xmath86 & 0.2282 & 13.468 & 31.1 & @xmath87 & 23.4 & 8.32 & + o3e03 & 2013 ge137 & 3:2 & @xmath88 & 0.257 & 3.866 & 31.1 & @xmath89 & 23.8 & 8.70 & + o3e04 & 2013 gj137 & 3:2 & @xmath90 & 0.265 & 16.873 & 32.1 & @xmath91 & 23.5 & 8.25 & kozai : @xmath92 + o3e06 & 2013 gl137 & 3:2 & @xmath93 & 0.199 & 10.440 & 34.4 & @xmath94 & 23.8 & 8.52 & + o3e07 & 2013 gg137 & 3:2 & @xmath95 & 0.136 & 2.931 & 35.2 & @xmath96 & 24.2 & 8.52 & + o3e08 & 2013 gd137 & 3:2 & @xmath97 & 0.1035 & 6.943 & 35.4 & @xmath98 & 23.7 & 8.45 & + o3e12 & 2013 gf137 & 3:2 & @xmath99 & 0.1567 & 14.680 & 37.3 & @xmath100 & 23.7 & 8.11 & + o3e41 & 2013 gk137 & 3:2 & @xmath101 & 0.178 & 9.879 & 45.6 & @xmath102 & 23.8 & 7.28 & + o3o02 & 2013 jc65 & 3:2 & @xmath103 & 0.2939 & 16.409 & 28.2 & @xmath104 & 23.8 & 9.11 & kozai : @xmath105 + o3o03 & 2013 jh65 & 3:2 & @xmath106 & 0.286 & 7.533 & 29.5 & @xmath107 & 24.0 & 9.22 & + o3o04 & 2013 jg65 & 3:2 & @xmath108 & 0.2492 & 15.934 & 30.0 & @xmath109 & 23.7 & 8.79 & kozai : @xmath110 + o3o05 & 2013 jk65 & 3:2 & @xmath111 & 0.2564 & 20.045 & 30.1 & @xmath112 & 24.3 & 9.44 & + o3o06 & 2013 jz64 & 3:2 & @xmath113 & 0.2325 & 10.128 & 30.4 & @xmath114 & 23.8 & 8.82 & kozai : @xmath115 + o3o08 & 2013 je65 & 3:2 & @xmath116 & 0.278 & 8.048 & 31.7 & @xmath117 & 24.3 & 9.15 & kozai : @xmath110 + o3o09 & 2013 jb65 & 3:2 & @xmath118 & 0.1889 & 24.898 & 32.0 & @xmath119 & 23.4 & 8.13 & + o3o10 & 2013 jf65 & 3:2 & @xmath120 & 0.1764 & 8.315 & 32.5 & @xmath121 & 24.2 & 8.94 & + o3o12 & 2013 ja65 & 3:2 & @xmath122 & 0.1488 & 10.223 & 33.8 & @xmath123 & 24.2 & 8.78 & + o3o13 & 2013 jl65 & 3:2 & @xmath124 & 0.230 & 7.251 & 34.7 & @xmath125 & 24.4 & 8.79 & + o3o15 & 2013 jd65 & 3:2 & @xmath126 & 0.0937 & 13.015 & 35.7 & @xmath127 & 23.7 & 7.90 & + o3o20pd & 2007 jf43 & 3:2 & @xmath128 & 0.186 & 15.080 & 38.3 & @xmath129 & 21.2 & 5.27 & + o3o27 & 2013 jj65 & 3:2 & @xmath130 & 0.256 & 19.814 & 41.0 & @xmath131 & 23.5 & 7.22 & + + o3e05 & 2013 gw136 & 2:1 & @xmath132 & 0.3440 & 6.660 & 33.0 & @xmath133 & 22.7 & 7.42 & asym .
@xmath134 + o3e55 & 2013 gx136 & 2:1 & @xmath135 & 0.2519 & 1.100 & 37.0 & @xmath136 & 23.9 & 7.67 & + o3o18 & 2013 je64 & 2:1 & @xmath137 & 0.284 & 8.335 & 36.1 & @xmath138 & 23.6 & 7.94 & asym .
@xmath139 + o3o33 & 2013 jj64 & 2:1 & @xmath140 & 0.082 & 7.650 & 41.0 & @xmath141 & 24.0 & 7.27 & + + o3e09 & 2013 gy136 & 5:2 & @xmath142 & 0.4143 & 10.877 & 35.8 & @xmath143 & 23.1 & 7.32 & + o3e48 & 2013 gs136 & 5:2 & @xmath144 & 0.3855 & 6.978 & 35.5 & @xmath145 & 23.9 & 8.50 & + o3o07 & 2013 jf64 & 5:2 & @xmath146 & 0.4497 & 8.785 & 30.5 & @xmath147 & 24.1 & 8.97 & + o3o11 & 2013 jk64 & 5:2 & @xmath148 & 0.4081 & 11.078 & 33.1 & @xmath149 & 23.0 & 7.69 & + + o3e19 & 2013 gr136 & 7:4 & @xmath150 & 0.0767 & 1.645 & 41.0 & @xmath151 & 23.8 & 7.20 & + + o3o19 & 2013 jn64 & 7:3 & @xmath152 & 0.287 & 7.740 & 38.1 & @xmath153 & 24.1 & 7.96 & + + o3e17 & 2013 gv136 & 8:5(i ) & @xmath154 & 0.035 & 7.452 & 40.6 & & 24.3 & 7.85 & insecure + & & & & & & & & & & @xmath155 of clones res for @xmath156 myr + o3o32 & 2013 jg64 & 18:11(ih ) & @xmath157 & 0.111 & 18.208 & 45.7 & & 24.1 & 7.48 & insecure , mixed argument + & & & & & & & & & & best - fit + 20% of clones res + & & & & & & & & & & @xmath158 + o3e52 & 2013 gt136 & 5:3(i ) & @xmath159 & 0.1540 & 12.112 & 48.9 & & 24.1 & 7.11 & insecure + & & & & & & & & & & 40% of clones res for @xmath156 myr + o3o25 & 2013 jm64 & 5:3(ih ) & @xmath160 & 0.047 & 7.287 & 40.6 & & 24.1 & 7.98 & insecure + & & & & & & & & & & 75% of clones res for @xmath156 myr + o3e13 & 2013 gu136 & 16:9(ih)@xmath161 & @xmath162 & 0.169 & 8.318 & 37.9 & & 23.6 & 7.86 & insecure + & & & & & & & & & & 50% of clones res for @xmath156 myr + o3e49 & 2013 hr156 & 15:8(i ) & @xmath163 & 0.1890 & 20.412 & 38.1 & @xmath164 & 23.6 & 7.72 & insecure , mixed argument + & & & & & & & & & & 85% of clones res + & & & & & & & & & & @xmath165 + o3o29 & 2013 jl64 & 13:5(ih)@xmath161&@xmath166 & 0.368 & 27.671 & 41.5 & & 23.6 & 7.03 & insecure + & & & & & & & & & & 35% of clones res + o3o34 & 2013 jh64 & 11:4(i ) & @xmath167 & 0.382 & 13.731 & 50.8 & @xmath168 & 22.8 & 5.60 & insecure + & & & & & & & & & & 50% of clones res + [
t : detections ]
there are 21 characterized 3:2 objects from the ossos 13ao and 13ae blocks ( listed in table [ t : detections ] ) .
our sample of 3:2 objects is sufficiently large to place some constraints on the libration amplitude distribution of plutinos .
this distribution is of interest because it is likely to reflect plutino capture histories ( see section [ s : intro ] ) .
we begin by presenting maps illustrating the sensitivity of the ossos 13ae / o blocks as a function of phase space in the 3:2 resonance ( section [ ss : sensitivity ] ) . in the following sections
, we use the survey simulator to constrain a parameterized model of the underlying 3:2 population . the population s @xmath23 and @xmath67 distributions are modeled independently ( section [ ss:32iaphi ] ) , while @xmath43 and @xmath22 must be constrained together ( section [ ss:32hd ] ) .
section [ ss:32kz ] presents constraints on the kozai fraction , and we summarize and report a population estimate for the plutinos in section [ ss:32sum ] .
figure [ f:32vis ] shows how the sensitivity of the 13ao and 13ae blocks to plutinos varies in @xmath169 and @xmath170 phase space with the actual detections over - plotted in white ; the relative visibilities are calculated using the survey simulator to simulate detections from a 3:2 population with uniform underlying distributions in the displayed ranges of @xmath22 , @xmath23 , and @xmath76 and a single exponential @xmath43 distribution with a slope @xmath171 ( see section [ ss:32hd ] )
. of note in this figure is the survey s sensitivity to moderately inclined , low-@xmath76 plutinos due to the placement of the 13ao block near the libration center of the resonance and @xmath172@xmath173 above the invariable plane .
the low-@xmath76 plutino phase space was not well explored by cfeps , and g12 found that the @xmath76 distribution of the plutinos could be acceptably modeled with no @xmath174 component ; this was also consistent with the scarcity of observed plutinos in the minor planet center database with @xmath174 ( @xcite report one such object ) .
ossos has detected 3 moderately inclined plutinos with @xmath174 in 13ao block , showing that the low-@xmath76 part of the resonance is populated ; we expect to further constrain the low - inclination , low-@xmath76 populations in an upcoming ecliptic block ( 15bd , see * ? ? ?
* ) pointed near the center of the 3:2 resonance on the other side of neptune . and @xmath169 plutino phase space for the ossos 13ao and 13ae blocks assuming uniform underlying distributions in orbital elements and an exponential @xmath43 magnitude distribution with a slope @xmath171 .
the white dots show the ossos detections .
the fact that the real detections do not cluster in the regions of high sensitivity simply indicates that the peaks of the intrinsic distributions lie at different values and that ( unsurprisingly ) a uniform underlying distribution for the population does not match the observations.,width=614 ] we use the 21 ossos detections and the survey simulator to constrain acceptable models for the intrinsic plutino @xmath23 and @xmath76 distributions , as described in section [ ss : ssmod ] and appendix [ a : stats ] .
we model the inclination distribution using equation [ eq : i ] .
the ad test ( appendix [ sec - ad ] ) identifies an acceptable match between the inclinations of synthetic and real ossos detections for @xmath175 at the 95% confidence level .
this range is consistent with previous observational estimates of the plutino inclination width : @xmath176 ( * ? ? ?
* -sigma confidence range ) , @xmath177 ( * ? ? ?
* -sigma confidence range ) , @xmath178 ( g12 , 95% confidence range ) , and @xmath179 ( * ? ? ?
* % confidence range ) .
we use a maximum likelihood approach ( appendix [ sec - maxlike ] ) to determine a best - fit value of @xmath180 for equation [ eq : i ] , although the probability distribution is quite flat in the range @xmath181 .
we also tested the acceptability of a gaussian inclination distribution of the form @xmath182 which was used by @xcite . using the ad test , equation [ eq :
i - offset - gaussian ] is a non - rejectable model for the plutino inclination distribution at 95% confidence for @xmath183 with @xmath62 ranging from @xmath184 at @xmath185 .
however , a maximum likelihood comparison shows that based on the ossos detections an offset gaussian ( equation [ eq : i - offset - gaussian ] ) is not a better description of the plutino inclination distribution than one centered on @xmath186 ( equation [ eq : i ] ) , so we confine ourselves to the single parameter model .
these results depend only weakly on the values for other model parameters , justifying our independent modeling of the @xmath23 distribution .
figure [ f : nocouple ] displays , as an example , the lack of coupling between the inclination and absolute magnitude distributions ; we show that the bootstrapped ad probability for a range of @xmath62 values does not significantly change for @xmath43 distributions with slopes @xmath187 and @xmath188 ( values near the extreme ends of the 95% confidence limits for @xmath52 that we find in section [ ss:32hd ] ) . for two different @xmath43 distributions .
the rejectable range of @xmath62 ( ad probability below 0.05 ) does not change much when comparing two very different @xmath43 distributions .
, width=403 ] as discussed in section [ ss : ssmod ] , we model the libration amplitude distribution as a triangle starting at a lower limit , @xmath189 , peaking at @xmath190 , and returning to zero at a maximum , @xmath191 .
the ossos plutinos have @xmath76 in the range @xmath192 , which constrains our choice of @xmath189 and @xmath74 .
we ran a suite of models through the survey simulator with @xmath193 , @xmath194 , and @xmath195 .
we find that @xmath196 , @xmath197 , and @xmath198 provides the best match to the observed libration amplitudes ( maximum likelihood ) , although the probability distribution in these parameters is quite flat .
using the ad test , we can not rule out any values of @xmath189 or @xmath191 in our tested ranges . at 95% confidence
we can constrain @xmath190 to be in the range @xmath199 . this range in @xmath190 , although wide , represents a rigorous constraint on the libration amplitude distribution ; detections from the remaining six ossos blocks should substantially improve this constraint .
though multiple emplacement mechanisms could produce a libration amplitude distribution with multiple components , a single component model provides an acceptable fit to current data .
we find a distribution of @xmath76 that , though mostly consistent with results of cfeps ( g12 ) , contains additional objects with lower libration amplitudes than previously reported .
future ossos blocks will provide additional 3:2 detections that will further constrain the @xmath76 distribution .
we ran a suite of survey simulations for plutino populations with a wide range of parameters for the eccentricity and @xmath43 distributions described by equations [ eq : e ] and [ eq : spl ] , respectively . because detection biases couple these distributions ( section [ s : bias ] ) , we model them together .
our results are presented in figure [ f:32edhmap ] .
the best - fit model , as measured by our summed chi - squared statistic ( appendix [ sec - cs ] ) , is @xmath171 , @xmath200 and @xmath201 , in agreement with the g12 results and derived from an observational sample that is completely independent from cfeps .
statistic for the @xmath22 , @xmath43 , and @xmath78 distributions .
lines : rejected parameter values using the summed ad statistic for the @xmath22 , @xmath43 , and @xmath78 distributions at the 99% confidence level ( solid white curves ) and the 95% confidence level ( dashed white curves ) .
our favored model parameters ( based on minimizing the summed @xmath202 statistic ) are shown by the black dots .
each panel is a 2-dimensional cut in our 3-dimensional parameter space search . for each panel , we fix one parameter at its favored value and show the goodness of fit map for the other two parameters ( for example , in the top panel , @xmath52 is fixed at 0.9 to show the allowed range in @xmath58 and @xmath57 for that value of @xmath52 ) .
, width=595 ] the 21 ossos plutinos are acceptably modeled by a single exponential in @xmath43 with a slope @xmath203 . this is somewhat surprising given that previous surveys have shown that the dynamically excited tno populations are not well - modeled by a single exponential .
recently @xcite found that these populations can be modeled by a broken exponential @xmath43 distribution with a bright - end slope @xmath171 that breaks to a faint - end slope @xmath204 at @xmath205 .
@xcite and @xcite find that the scattering population shows evidence of a divot ( a deficit of objects rather than a simple change in slope ) in the @xmath43 distribution near @xmath206 , corresponding to @xmath207 .
@xcite rejects a single exponential @xmath43 distribution for the plutinos , finding evidence for either a divot near @xmath208 or a break to a shallow slope at @xmath209 .
based on just the ossos sample , we can not rule out a single exponential despite being sensitive to plutinos with @xmath210 where the divot or change in slope has been proposed . to examine the conflicting conclusions between ossos and the @xcite results about the possibility of a single exponential all the way down to @xmath211
, we generated 100 sets of 21 synthetic ossos detections for @xcite s preferred divot model .
we then tested how many of these 100 synthetic ` observed ' data sets would be able to reject our best - fit single exponential @xmath43 distribution .
we find that if the real plutinos follow @xcite s nominal divot distribution , a sample of 21 detected in the two ossos blocks would reject a single exponential @xmath212 of the time .
so while we find no evidence of a transition in the ossos sample , this could just be due to our small sample size .
we note , however , that the placement of the ossos blocks means we were most sensitive to large-@xmath43 objects with low libration amplitudes , which differs from @xcite s survey .
figure [ f:32hvis ] shows the detectability of plutinos in the 13ao and 13ae blocks as a function of @xmath213 and of @xmath76 and @xmath213 .
many of the large-@xmath43 ossos objects have @xmath214 , a previously sparsely observed part of the resonance s phase space .
it would be very interesting if the low @xmath76 plutinos have a different @xmath43 distribution than the larger @xmath76 plutinos ; @xcite found some evidence for this in their analysis of the observed plutinos .
different dynamical capture mechanisms populate different parts of the resonance , so it is not impossible that the low and high @xmath76 plutinos were captured from different parts of the primordial tno population .
better statistics afforded by the upcoming ossos blocks will further test this idea . for the ossos 13ao and 13ae blocks assuming uniform underlying distributions .
the gray histogram shows the actual number of detected plutinos as a function of @xmath43 .
right panel : color coded relative visibility of plutinos as a function of both @xmath43 and @xmath67 for the ossos 13ao and 13ae blocks assuming uniform underlying distributions .
the white dots show the ossos plutino detections .
the sensitivity to 60 degree libration amplitude is due to the location of 13ae block , which favors detection of plutinos with @xmath76 somewhat larger than @xmath215.,width=595 ] finally , we model the fraction of plutinos that are also in the kozai resonance .
our dataset of 21 plutinos contains 5 kozai oscillators . within the survey simulator ,
these objects are generated separately from the other plutinos because they occupy a distinct phase space within the resonance . to account for this
we follow the procedure outlined in g12 and @xcite which uses an approximate kozai resonant hamiltonian @xcite to select values of @xmath22 , @xmath23 , and @xmath38 that correspond to kozai libration of various amplitudes within the resonance .
a kozai plutino s @xmath43 and @xmath76 are selected the same way as for the non - kozai plutinos ( we assume that kozai and non - kozai plutinos share a single libration amplitude distribution , which is sufficient to model current data ) .
our procedure for choosing the other orbital parameters for kozai plutinos is described in appendix [ as:32 ] .
we ran a suite of survey simulations varying the intrinsic kozai fraction ( @xmath216 ) from 0 - 1 to determine the probability of detecting 5 kozai plutinos in an sample of 21 plutino detections for each value of @xmath216 . to reproduce the 5 ossos 3:2 kozai plutinos more than 5% of the time
, we find that the kozai fraction must be @xmath217 ( @xmath218 at 99% confidence ) .
an intrinsic kozai fraction of 0.2 has the highest probability of reproducing the ossos detections .
this is in reasonable agreement with the @xmath219 ( @xmath100.33 at 95% confidence ) determined by cfeps ( g12 ) .
as discussed in @xcite , different resonant capture scenarios predict different values for @xmath216 ; the first two ossos blocks have already narrowed the range of allowable @xmath216 compared to the cfeps results , and we expect the future blocks to provide an even better determination of the intrinsic kozai fraction . our nominal best - fit values for the parameters in our plutino model are @xmath171 , @xmath200 , @xmath201 , @xmath180 , @xmath220 , and a triangular @xmath76 distribution that goes from @xmath221 with a peak at @xmath222 .
figure [ f:32cdf ] shows this distribution compared to the actual ossos detections ; there is generally good agreement in the one - dimensional distributions in @xmath23 , @xmath22 , @xmath76 , @xmath213 , and distance at discovery between the synthetic detections and the actual ossos detections . using our best fit model
, we estimate that the 3:2 resonance contains a population of @xmath223 objects with @xmath224 ( see section [ s : pop ] for more details ) .
the independent ossos data sample yields best - fit orbital parameters and a total population estimate for the plutinos that are in good agreement with the cfeps results ( g12 ) . , @xmath22 , @xmath76 , @xmath213 , and distance at discovery for the observed 13ao and 13ae block plutinos ( red dots ) , the intrinsic plutino population for our nominal plutino model ( gray dashed lines ) and for the synthetic detections from our nominal plutino model ( black lines ) .
the differences between the intrinsic models and the synthetic detections show the effects of the observational biases.,width=595 ]
one of the surprising results from cfeps was that the population of the 5:2 resonance was found to be nearly as large as the population of the 3:2 resonance ( g12 ) .
this is unexpected because planetary migration models do not predict efficient capture into the 5:2 resonance ( e.g. , * ? ? ?
* ) and capture following dynamical instability ( e.g. , * ? ? ?
* ) likewise predicts a smaller 5:2 population relative to the 3:2 .
so far , ossos has detected 4 objects in the 5:2 resonance at @xmath225 au .
given that the libration behavior of 5:2 resonant objects is similar to that of the 3:2 , where objects at exact resonance come to perihelion at the ortho - neptune points , the 13ao and 13ae blocks show a similar visibility profile for the 5:2 resonance ( figure [ f:52vis ] ) as for the plutinos ( figure [ f:32vis ] ) .
the major difference between these two resonances is the much lower sensitivity to low - eccentricity 5:2 objects because it is a more distant resonance .
the right panel of figure [ f:52vis ] shows contour lines in eccentricity below which the probability of observing an object from an eccentricity distribution uniform in the range @xmath226 drops below 5% and 1% assuming an underlying @xmath43 distribution with a slope @xmath171 ; we do nt expect a uniform eccentricity distribution , but this does demonstrate that ossos is not particularly sensitive to 5:2 objects with @xmath227 . and @xmath169 5:2 phase space for the ossos 13ao and 13ae blocks assuming uniform underlying distributions .
the white dots show the ossos detections . in the right panel ,
the solid and dashed lines show the eccentricities below which visibility drops to @xmath228% and @xmath229% respectively for an @xmath43 distribution with @xmath171 . as in figure
[ f:32vis ] , the fact that the real detections do not cluster in the regions of high sensitivity simply indicates that a uniform underlying distribution in @xmath230 , and @xmath76 does not match the observations.,width=595 ] we use a parameterized orbital model for the 5:2 resonance identical to that for the non - kozai plutinos .
we ran a suite of survey simulations to place limits on the parameterized @xmath23 , @xmath22 , and @xmath43 distributions .
given the small number of detections , we used a single , triangular @xmath76 distribution that ranged from @xmath231 with a peak at @xmath222 ; this provided a statistically adequate representation of the ossos 5:2 detections and is similar to the @xmath76 distribution used in g12 for this population .
the upper limit for libration in the 5:2 resonance ( from both observations and numerical integrations ) appears to be @xmath232 @xcite , but the extension of the @xmath76 distribution above @xmath233 is not necessary to describe the ossos 5:2 detections ; with the future ossos blocks , we expect more 5:2 detections and will explore the upper limit for the @xmath76 distribution .
we find the inclination distribution can be modeled using equation [ eq : i ] with a most - probable ( maximum likelihood ) width of @xmath234 . at 95% confidence using the ad statistic , the width ranges from @xmath235 in agreement with the width of @xmath236 for the 5:2 from g12 . and
center @xmath57 assuming an underlying @xmath213 distribution of slope 0.9 .
the lines indicate values rejectable at 99% ( solid white ) and 95% ( dashed white ) confidence.,width=403 ] as discussed for the plutinos , the @xmath22 and @xmath43 distributions ca nt be determined independently from each other .
we find that a single exponential is an adequate model for the 4 ossos detections ; reasonable eccentricity distributions can provide acceptable matches for the observed @xmath22 , @xmath43 , and heliocentric distance at discovery for slopes in the range @xmath237 with no strongly preferred value ( based on a summed chi - square statistic ) . with only 4 detections , it is not surprising that we do nt have a strong constraint on @xmath52 . assuming the 5:2 population has the same @xmath43 distribution as the plutinos
, we can constrain the allowable range of eccentricity distribution parameters ( equation [ eq : e ] ) . figure
[ f:52ewec ] shows the significance levels of the summed ad statistic for the @xmath78 , @xmath22 , and @xmath43 distributions of the 4 ossos 5:2 detections compared to simulated detections for a range of @xmath57 and @xmath58 values . because the observed 5:2 objects have a narrow range in @xmath22 of @xmath238 , the least - rejectable eccentricity distribution has @xmath239 and @xmath240 .
however this is not likely to be a good representation of the true 5:2 eccentricity distribution ; there are 5:2 objects with @xmath241 in the mpc database ( also listed in * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) which invalidates such a strongly peaked @xmath22 distribution centered at @xmath239 .
as figure [ f:52ewec ] shows , the ossos detections do not rule out @xmath22-distributions with smaller @xmath57 and larger @xmath58 , a result that is consistent with the findings of g12 ; however , the insensitivity of the ossos 2013ao / e blocks to 5:2 objects with @xmath242 makes this distribution difficult to constrain . if we limit our model to @xmath0 , we find that the ossos observations can be adequately reproduced by a uniform eccentricity distribution in the range @xmath243 .
we use this restricted @xmath22 range to model the total intrinsic population of the 5:2 with the understanding that this makes our population estimate a lower limit because we know that the @xmath244 region is occupied . for our best - fit model
applied to ossos data alone , we find that the 5:2 resonance contains @xmath245 objects with @xmath224 and @xmath246 ( see section [ s : pop ] for more details ) .
the 2:1 is the strongest of the n:1 resonances . in the 2:1
, symmetric librators have a resonant angle @xmath13 ( see section [ s : methods ] ) which , like that for all 3:2 objects , librates about @xmath247 .
asymmetric librators instead librate about a center near @xmath31 or @xmath32 .
@xcite studied the current dynamics of the 2:1 resonance , determining how the libration centers and amplitudes change with eccentricity and how the stability of the resonance is affected by inclination . @xcite also studied the stability of 2:1 phase space .
determining how the current 2:1 resonant objects are split between the symmetric , leading asymmetric , and trailing asymmetric libration islands is of particular interest for determining how this resonance became populated ; @xcite and @xcite demonstrated that neptune s migration speed affects the probability of capture into the leading or trailing asymmetric libration centers , with higher speed migration favoring the trailing island . in this section
we describe how we use the first two ossos blocks to constrain the fraction of symmetric 2:1 librators .
we then use the combined ossos and cfeps observations to provide a well - characterized constraint on the trailing - to - leading ratio in the 2:1 resonance ; as we discuss later , the combined data set is used for this constraint because the first two ossos blocks were only sensitive to trailing 2:1 asymmetric librators .
because of the more complicated phase space of the 2:1 resonance compared to the 3:2 or 5:2 resonances , we do not have a simple parameterized @xmath76 distribution for the 2:1 .
both the libration centers and the allowable range of @xmath76 for the asymmetric islands are @xmath22-dependent .
we also only have 4 ossos detections , so an overly complicated model is not warranted .
we base our 2:1 model on the results of @xcite , who published a plot of libration centers and maxiumum libration amplitudes as a function of @xmath22 . to generate a 2:1 population ,
we first decide if an object is symmetric or asymmetric .
if it is symmetric , we select @xmath22 from a uniform range @xmath248 and @xmath76 from a uniform range @xmath249 ; these ranges correspond to the regions of relatively stable libration found in theoretical and numerical experiments @xcite .
for asymmetric librators , we select @xmath22 uniformly from @xmath250 . for the chosen value of @xmath22
, we choose the libration center from @xcite and then assign a libration amplitude uniformly from @xmath251 .
the inclinations are randomly selected from a gaussian inclination distribution described by equation [ eq : i ] . from just the 4 ossos detections ,
we find that the above simplified model for the 2:1 resonance ( only slightly modified from the cfeps 2:1 model of g12 ) is consistent with the observations .
we find that the inclination distribution width must be @xmath252 at 95% confidence with a most - probable value of @xmath253 , independently confirming g12 s conclusion that the 2:1 population is significantly colder in inclination than either the 3:2 or the 5:2 .
we note that there are a few observed 2:1 objects in the mpc database with inclinations in the range @xmath254 .
most of these high inclination 2:1 objects appear to be large amplitude symmetric librators ( see for example table 1 in @xcite ) .
@xcite showed that high inclination symmetric librators are not stable on gyr timescales ; this perhaps indicates that these observed large inclination 2:1 objects ( a population not yet detected by ossos ) are only temporarily stuck to the 2:1 resonance rather than primordial members .
we will explore the possibility of a population of higher - inclination , temporary 2:1 objects in addition to the low-@xmath23 ( presumably primordial ) 2:1 population with future ossos observations .
based on the fact that half of the ossos 2:1 objects are symmetric librators , we can place a weak limit on the intrinsic fraction of symmetric 2:1 objects , @xmath255 .
for our parameterized model of the 2:1 resonance , we tested intrinsic symmetric fractions ranging from 0.05 - 0.95 .
for each tested @xmath255 we can determine the probability of drawing 4 synthetic observed objects with a @xmath256 .
this probability allows us to rule out @xmath257 at the 99% confidence level and @xmath258 and @xmath259 at the 95% confidence level . to further constrain the allowable range of @xmath260 ,
we repeat this calculation with the 9 combined cfeps and ossos 2:1 detections while additionally considering the division of the asymmetric librators between the leading and trailing libration centers .
the two ossos blocks both point toward the trailing libration center , them fairly insensitive to the leading / trailing fraction .
this is evident in figure [ f:21vis ] , which shows the relative visibility of all three libration islands in @xmath22-@xmath76 and @xmath23-@xmath76 phase space ; the probability of detecting a leading asymmetric 2:1 object in the ossos 13ao or 13ae blocks is nearly 0 .
additional ossos blocks will cover the leading center , but for now we can use the cfeps detections in addition to the ossos 13ao and 13ae block detections because cfeps covered both libration centers ( g12 ) . of the 9 combined ossos and cfeps 2:1 detections ,
3 are symmetric librators and 6 asymmetric ; 5 of the asymmetric detections are in the trailing libration island and 1 is in the leading island .
we ran a suite of ossos+cfeps survey simulations for a wide range of intrinsic symmetric fractions , @xmath261 , and a wide range of the intrinsic fraction of asymmetric librators in the leading libration center , @xmath262 .
the left panel of figure [ f:21maps ] shows the probability of drawing a sample from the synthetic detections for each combination of @xmath255 and @xmath263 that matches the observed symmetric fraction , @xmath264 ; the right panel shows the probability of drawing a sample with @xmath264 and @xmath265 . using the combined ossos and cfeps detections
, we have the constraint that @xmath266 at the 99% confidence level and @xmath267 at the 95% confidence level ; the fraction of asymmetric objects in the leading libration center is constrained to be @xmath268 at the 99% confidence level and @xmath269 at the 95% confidence level . to obtain a population estimate for the 2:1 resonance , we assume that the population is evenly split between symmetric and asymmetric librators and that the asymmetric librators are evenly split between the leading and trailing islands ( @xmath270 and @xmath271 ) ; we also assume the @xmath43 distribution has @xmath171 . using this model and just the ossos data ,
the 2:1 resonance is estimated to contain @xmath272 objects with @xmath273 ( see section [ s : pop ] for more details ) . and @xmath170 2:1 phase space for the ossos 13ao and 13ae blocks assuming an even split between the leading asymmetric , trailing asymmetric , and symmetric libration centers as well as uniform @xmath22 , @xmath23 , and @xmath76 distributions within the resonant phase space of all three libration islands .
the white dots show the ossos detections.,width=595 ]
we have modeled the 3:2 , 5:2 , and 2:1 resonances based on the first set of ossos detections . from this
independent data set , we find that the orbital and @xmath43 distributions for these resonances are consistent with those found by cfeps ( g12 ) .
taking our nominal orbital models based on the ossos detections , we can construct population estimates for these three resonances . to do this we run @xmath274 instances of the survey simulator for our orbital models of each resonance and determine how many objects with @xmath213 less than some limiting value must be generated to match the 21 plutino detections , the four 2:1 detections , and the four 5:2 detections in the 13ao and 13ae blocks . to facilitate comparison to the population estimates from cfeps , we choose a limiting magnitude @xmath275 to compare to their @xmath276 .
this assumes the resonant populations have colors @xmath277 , which matches the value ( within photometric uncertainties ) for the plutino population @xcite based on the @xmath278 colors of the cfeps 3:2 objects @xcite ; there is an ongoing gemini program @xcite that will quantify the colors for the ossos @xmath279 resonant objects , so in future analysis this color assumption might be improved .
our population estimates are listed in table [ t : pops ] and shown in figure [ f : pops ] .
all three population estimates overlap with the 95% confidence bounds on the cfeps ( g12 ) estimates , although our median number of plutinos and our lower limit for the 5:2 population are both smaller than the cfeps estimates and our 2:1 population is slightly larger .
our 2:1 population estimate is much more uncertain that the cfeps estimate despite the roughly equal numbers of ossos and cfeps 2:1 detections ; this is due to the restricted longitude range of the first two ossos blocks ( both trailing neptune ) compared to the wider longitude ranges of the cfeps observations .
c l l r r 3:2 & eq . [ eq : e ] , @xmath200 , @xmath280 & eq .
[ eq : i ] , @xmath281 & @xmath223 & @xmath282 + 5:2 & uniform @xmath283 & eq . [ eq : i ] , @xmath284 & @xmath245 & @xmath285 + 2:1 & sym : uniform @xmath286 & eq .
[ eq : i ] , @xmath287 & @xmath272 & @xmath288 + & asym : uniform @xmath289 & & & + [ t : pops ] we can also compare our population estimates to those from the deep ecliptic survey ( des ) given in @xcite .
the des observations were done in the vr filter , so we must assume a value of @xmath290 in order to compare the population estimates .
@xcite assumed @xmath291 for comparing the des population estimates to the cfeps population estimates and found that the two sets of population estimates for the 3:2 population were discrepant .
however , we find that @xmath292 is a better estimate of the color conversion for the resonant objects based on a comparison of h values in the two color filters for resonant objects observed by both surveys ( see appendix [ a : colors ] ) .
using the measured @xmath293 color rather than @xmath294 erases the discrepancy between the des and cfeps 3:2 population estimates reported by @xcite .
figure [ f : descompare ] shows the data from figure 1 in @xcite along with the resonant population estimates from table 3 in g12 and table [ t : pops ] of this work ( taking @xmath293 and @xmath277 ) .
we find that the cfeps and ossos population estimates for the 3:2 , 5:2 , and 2:1 resonances are in very good agreement with the intrinsically faint ( large h ) des population estimate for the 3:2 resonance and overlap with the des 2:1 and 5:2 population estimates .
anchored at the @xmath295 values from g12 for cfeps and the @xmath296 values from table [ t : pops ] ; the estimated 95% confidence limits are shown as dashed lines for both cfeps and ossos .
the arrows indicate the approximate range in @xmath297 where cfeps and ossos had detections for each resonance ; each purple dot for the des results is an individual detection and thus shows the des observed range of @xmath297 .
we assume color conversions of @xmath293 and @xmath277.,width=576 ] our median population estimates imply that intrinsic ratio of 3:2 / 5:2 / 2:1 objects is 1.5 / 1.1 / 1 , compared to 3.5 / 3.2 / 1 from the cfeps population estimates ( g12 ) and 1.4 / 0.7 / 1 from the deep ecliptic survey population estimates @xcite ( note that because the color conversion is assumed to be the same for all the resonances , the des population ratios are independent of the assumed @xmath298 color ) . combining the ossos and cfeps detections to obtain population estimates for our nominal resonance models ( also listed in table [ t : pops ] and shown in figure [ f : pops ] ) , we find a ratio of 2.5 / 2 / 1 . the uncertainties in our population estimates from
just the ossos data are currently too large to conclusively determine whether the 5:2 is more populated than the 2:1 or as populated as the 3:2 , but the ossos detections are consistent with a large population in the 5:2 resonance .
additionally , we have used an artificially restricted eccentricity range for the 5:2 resonance due to our insensitivity to @xmath244 objects , so the real 5:2 population is likely to be larger than our estimate .
in the ossos 13ao and 13ae blocks there are detections in nine other resonances : the 8:5 , 18:11 , 5:3 , 16:9 , 15:8 , 7:3 , 7:4 , 13:5 , and 11:4 resonances . of these detections , only the 7:4 and 7:3 detections are securely resonant as defined by the @xcite classification scheme .
we integrated many clones of each insecure resonant detection to determine the probability that the objects are resonant ; these probabilities are listed in table [ t : detections ] .
two of the insecure resonant detections have best - fit orbits that show libration of a mixed resonant argument .
ossos object o3o32 shows libration of the angle @xmath158 and object o3e49 shows libration of the angle @xmath165 .
single and/or insecure detections are not enough to characterize the structure of a resonance or provide a well - constrained population estimate , but we can check whether our single secure detections for the 7:4 and 7:3 resonances are consistent with the g12 models and population estimates for these resonances . to test the 7:3 and 7:4 resonance models , we ran the g12 parameterized models through the ossos survey simulator to generate 10000 simulated detections for the 7:4 and 7:3 . in both cases the observed characteristics of the real ossos detections ( @xmath22 , @xmath23 , @xmath78 , @xmath213 , and @xmath76 ) fall within the 95% bounds of the synthetic detections , indicating that the g12 models are consistent with the ossos detections .
we also generated population estimates for these models of the 7:4 and 7:3 resonances .
for the 7:4 resonance the median population of objects with @xmath224 is 1000 with a 95% confidence range of @xmath299 which agrees with the g12 95% confidence estimate of @xmath300 objects with @xmath301 ( assuming a @xmath277 ) . for the 7:3 resonance ,
the median population of objects with @xmath224 is 4000 with a 95% confidence range of @xmath302 again in agreement with the g12 95% confidence estimate of @xmath303 objects with @xmath301 .
testing of the other g12 resonance models and population estimates will be presented in future papers as more ossos blocks are completed and the orbits of the remaining insecure 13ae and 13ao resonant detections are improved by follow - up observations .
we have presented the detections of resonant objects from the first two of the eight ossos observational blocks .
the ossos detections of 3:2 , 5:2 , and 2:1 resonant objects are broadly consistent with the resonance models and population estimates found by cfeps ( g12 ) .
this verification of cfeps results with an entirely independent dataset inspires additional confidence in the results from the cfeps / ossos survey characterization method .
our primary results are as follows : * our population estimates are listed in table [ t : pops ] .
these values are consistent with cfeps population estimates within the uncertainties ( g12 ) .
we find that given a modified empirical color conversion , the des population estimates @xcite are also consistent with these results within our 95% confidence intervals .
* ossos detections of several very low amplitude 3:2 objects require a refinement of the cfeps plutino model , extending the libration amplitude distribution to lower values .
lower amplitude librators are produced more efficiently in models appealing to capture during smooth migration of neptune than in models which fill kuiper belt phase space ( for example during dynamical upheaval of the giant planets ) and leave behind resonant populations because the resonances are regions of dynamical stability .
additional dynamical modeling is required to determine whether our low - amplitude librators provide evidence for a population of migration - captured resonant objects in the 3:2 .
our low - amplitude detections were enabled by placement of the 13ao block @xmath17210 degrees from one of the ortho - neptune perihelion locations .
future ossos blocks will improve our characterization of this population .
figure [ f : visfull32 ] displays an estimated visibility map for the 3:2 resonance given the full ossos survey .
* we find no evidence of the h - magnitude distribution transition suggested by @xcite in our population of 3:2 objects .
however , we find that if such a transition is present , our small sample of objects would reject a single slope @xmath43 model only @xmath30480% of the time .
the increased sample size provided by future ossos blocks will place better constraints on the @xmath43 distribution of the plutinos . *
the ossos 5:2 detections confirm the finding in g12 that this resonance is more populated than expected based on existing models for the dynamical history of the outer solar system . after restricting ourselves to the eccentricity range visible in the ossos blocks ( @xmath0 ) , we independently verify that the total population of the 5:2 is at least as large as that of the 2:1 and possibly as large as that of the 3:2 . given this confirmation
, future models of dynamical emplacement of kuiper belt objects must produce a large population in the 5:2 .
the addition of future ossos detections will reduce the large uncertainty in our resonant population estimates and allow a more precise measurement of the 3:2 / 5:2 / 2:1 population ratios .
* we have confirmed that the inclination distribution of the 2:1 resonance is much colder than those of the 5:2 and 3:2 .
this result might indicate that a larger fraction of 2:1 objects were caught in resonance from a dynamically unexcited reservoir .
we speculate that a larger fraction of 2:1 objects may have been caught during migration of neptune from the objects originally located in the region of the observed cold classical kuiper belt , although this scenario might not be consistent with the wide range of colors seen for 2:1 objects compared to the cold classicals by @xcite .
we will investigate this speculation in future modeling work .
* using the combined cfeps and ossos 2:1 detections , we have placed new , more restrictive constraints on both the fraction of the 2:1 resonant objects that are symmetric librators as well as the ratio of leading to trailing asymmetric librators .
our current limits do not substantially constrain histories of resonance capture during migration , but future ossos blocks will be sensitive to both leading and trailing asymmetric librators ( see figure [ f : visfull21 ] for an estimated 2:1 visibility map for the full survey ) .
if the populations of leading and trailing librators are significantly different , this difference may be identifiable by the full ossos sample . , @xmath23 , and @xmath76 and a slope of @xmath171 for the @xmath43 distribution ( see section [ s:32 ] and figure [ f:32vis ] for comparison).,width=576 ] , @xmath23 , and @xmath76 within the resonant phase space for each libration island and a slope of @xmath171 for the @xmath43 distribution ( see section [ s:21 ] and figure [ f:21vis ] for comparison).,width=576 ] ossos observed two blocks leading neptune from late 2013-late 2015 , and this should slightly more than double the resonant sample once the data is fully analyzed ; this will allow an improvement of the current analysis ( especially for the 2:1 ) .
the second half of ossos will produce orbits by early 2017 , and will cumulatively provide @xmath305 times the current 13ae / o block sample ; the multiple is @xmath306 due to filter upgrades and seeing improvements occurring in a now - vented dome at the cfht telescope , both of which improve magnitude depth .
throughout this paper , we describe the @xmath43 distribution and dynamical properties of the underlying resonant population using simple parametrized models .
given the set of objects observed by ossos , we would like to determine which values of the models parameters are most probable , identify the range of parameter values that reasonably match the data , and verify that our simple models can fit the data sufficiently well that more complicated models are not required . to achieve the first two goals , we would in principle like to calculate , in multiple dimensions , the relative likelihoods of observing our detected objects given each set of model parameters . given a uniform ( uninformative ) prior , these relative likelihoods are equivalent to the bayesian posterior distribution or , in other words , the relative likelihoods of each set of model parameters . the sufficiency of our models
could then be assessed by using the most probable model parameters and comparing the probability computed for the observed data set to the distribution of probabilities produced by synthetic data sets .
this procedure is described in section [ sec - maxlike ] . in practice , this full calculation is difficult because significant computational resources are required to evaluate the observational biases in our data . fortunately , our inferred distributions for the inclination and libration amplitude do not depend substantially on the inferred distributions of other parameters ( see section [ ss:32iaphi ] ) .
we can therefore employ a maximum likelihood calculation for the inclination distribution and libration amplitude distributions using 1-dimensional models , fixing all other distributions to a set of acceptable parameter values . the inferred distance of detection , absolute magnitude , and eccentricity distributions , however , are correlated and must be treated together .
even this three dimensional space is very computationally expensive to probe with high resolution using a relative likelihood calculation ( section [ sec - maxlike ] ) , so we compromise by combining an anderson - darling rejectability statistic and a @xmath81-squared calculation intended to assess goodness - of - fit , described in sections [ sec - ad ] and [ sec - cs ] , respectively . from bayes theorem , the probability of a model @xmath307 given our data set @xmath308 is @xmath309 .
our model , @xmath307 , consists of a set of parameters that characterize the distributions of eccentricity ( @xmath22 ) , inclination ( @xmath23 ) , libration amplitude ( @xmath13 ) , absolute magnitude ( @xmath43 ) , and heliocentric distance at discovery ( @xmath78 ) for the kbo population in a specified resonance . in all cases , the logarithmic scale of our model parameters
we therefore assume uniform ( uninformative ) priors on our parameterized models @xmath310 , so that @xmath311 to calculate the relative likelihoods of each model ( each set of parameterized distributions ) given our set of observed objects @xmath308 , we would therefore like to know the probability of observing @xmath308 given each possible set of model parameters .
the nature of the observational biases that must be applied to our models means that we can not analytically calculate @xmath312 . instead
, we must estimate this probability numerically . to do so
, we use the ossos survey simulator .
the survey simulator produces a set of synthetic detections given a model of the underlying population .
if the survey simulator could be run an infinite number of times , it would ( when normalized ) translate each model ( or set of parameter values ) into a probability distribution for the properties of a detected object , @xmath313 , given the model and a specified total number of detections ; the total number of objects in the model itself is a free parameter in our parameterized models , and it is allowed to vary so that the model produces a total number of detections equal to the actual number of detections .
the probability of observing @xmath18 detected objects with a given set of @xmath314 is then @xmath315 the most probable model is the one that maximizes @xmath312 .
in practice , we can only run the survey simulator a finite number of times per set of model parameters , so our calculation of @xmath316 must be binned in @xmath22 , @xmath23 , @xmath13 , @xmath43 , and @xmath78 . because we are calculating @xmath317 numerically from synthetic detections ,
there is an uncertainty due to poisson noise in the number of detections expected by the model . in bin @xmath318
, we estimate this uncertainty to be @xmath319 where @xmath320 is the number of synthetic detections in bin @xmath318 and we have used @xmath321 as shorthand for the values of @xmath322 in each bin .
we then propagate this uncertainty through the calculation of @xmath312 ( equation [ eq : pda ] ) by adding them in quadrature .
two models can only be distinguished if their relative probabilities @xmath323 differ by more than the errors in their @xmath323 . in principle
, we would like to for each set of parameter values
produce sufficiently many simulated detections that we could map this five - dimensional probability distribution , allowing us to directly compute the relative probabilities of the observed data for a large sets of parameter values spanning their full range and allowing for correlations in the parameters . in practice ,
running the survey simulator a sufficient number of times to do this is prohibitively computationally expensive given the size of the currently allowed parameter space ( see @xcite for another discussion of this ) ; when ossos is complete , the allowed parameter space will be smaller and there will likely be enough real detections to make the expenditure of computational resources worthwhile .
due to the computational expense , in this paper we only apply this approach in a few cases . for our parameterized inclination distribution
, we can do this calculation in one dimension ( inclination ) to find the most probable inclination width ; we can limit ourselves to one dimension because the observability of the inclination distribution is fairy independent of the other orbital element distributions and the @xmath43 distribution ( see figure [ f : nocouple ] ) . in that case
, we fix the eccentricity , @xmath324 , and libration amplitude distributions and run a suite of survey simulations over a wide range of inclination widths to find which width , @xmath62 , maximizes the probability @xmath325 .
a similar calculation is done for the parameterized libration amplitude distribution .
these one - dimensional probability calculations only require @xmath326 synthetic detections per model iteration to adequately sample the binned probability distribution with fractional uncertainties @xmath327 in @xmath316 for almost all bins that contain a real detection . to assess model sufficiency for the inclination distribution ( and analogously for the libration amplitude distribution )
, for each @xmath62 , we produce a large set of synthetic data sets of size @xmath18 drawn from equation [ eq : i ] .
we calculate @xmath312 for each synthetic data set using equation [ eq : pda ] .
we then determine whether @xmath312 calculated for the observed data is consistent with the distribution of @xmath312 generated by the synthetic data .
consistency is defined as lying in the middle 95% of the cumulative distribution ( removing the bottom 2.5% and top 2.5% ) .
we find that the range of values for @xmath62 accepted by this procedure is similar to the range of values not rejected by the anderson - darling test , described in appendix [ sec - ad ] . because of the coupled observational biases , calculating relative probabilities for our different @xmath43 distribution / eccentricity models requires us to bin the probability distribution in three dimensions : eccentricity , distance at discovery , and absolute magnitude
. for reasonable bin sizes , determining the probability distribution to fractional uncertainties @xmath328 in most bins , and @xmath329 in all bins requires of order @xmath82 synthetic detections per model iteration . for this reason
, we only calculate @xmath312 for a few of our favored models in order to compare their relative probabilities .
we use the anderson - darling test statistic ( ad statistic ) outlined by nist . the ad statistic is defined as @xmath330 where there are
@xmath18 ordered data points , @xmath331 , and @xmath332 is given by : @xmath333}\ ] ] where @xmath334 is the cumulative distribution function of the model being tested .
larger values for @xmath335 correspond to cumulative distributions that are more different .
the significance of the value of @xmath335 is determined by calculating the expected distribution of @xmath335 by repeatedly drawing samples of @xmath18 random points from the model distribution and computing their ad statistic .
a model is rejected at p% confidence if the observed value of the ad statistic is larger than p% of the model s subsample ad statistics .
the above calculation is for a one - dimensional distribution . to extend to testing multiple distributions at once
, we linearly add the ad statistic for each one - dimensional distribution @xcite .
the significance of the summed ad statistic is then determined just as above .
one measure of the goodness - of - fit for any particular parameterized model is the chi - square statistic : @xmath336 where @xmath337 is the observed frequency in bin @xmath338 and @xmath339 is the expected frequency in bin @xmath338 given the model @xcite .
the exact value of @xmath202 will depend on the choice of bins for the data , but if the binning is the same across multiple models , a smaller value of @xmath202 indicates a better fit to the observations . as with the ad statistic ,
a simplistic way to extend this to multiple dimensions is to linearly add the one - dimensional values of @xmath202 .
we do not have enough observational data points to calculate a meaningful @xmath202 by binning in multiple dimensions simultaneously .
the determination of the best - fit orbit and classification status of each ossos detection follows the procedure outlined in @xcite . a barycentric best - fit orbit is determined based on all available ossos astrometric observations using the @xcite algorithm
this is the best - fit orbit listed in table [ t : detections ] ; the listed uncertainties are taken from the diagonal elements of the best - fit orbit s covariance matrix . to determine the classification status of an object , we integrate three clones of the observed object to determine the dynamical behavior : the best - fit orbit and then orbits with the maximum and minimum deviations in semimajor axis that are still consistent with the observations . to generate the minimum- and maximum-@xmath49 clones ,
we perform a search for acceptable orbits in @xcite s ( @xmath340 ) coordinate system , with a maximum variation in each coordinate of @xmath341 as determined from the diagonal elements of the ( @xmath340 ) covariance matrix .
an orbit fit is deemed consistent with the observations if the worst residual between its predictions and the observed astrometric positions is not larger than 1.5 times the best - fit orbit s worst residual and if the rms residual is not more than 1.5 times the best - fit orbit s rms residual .
the @xcite orbit fitting code does provide a semimajor axis uncertainty ( @xmath342 ) , but @xcite chose to use the residuals to determine the minimum- and maximum-@xmath49 orbits because doing so provides a better estimate of the true uncertainty when the observational arc is short or when there might be systematic errors in the astrometry . in these cases , it is not unusual for the maximum- and minimum-@xmath49 orbits to differ from the best - fit value by significantly more than @xmath343 .
this is because the nominal 1-@xmath342 uncertainty for the best - fit orbit is calculated from the diagonals of the barycentric orbital elements covariance matrix , which is produced by assuming an linearized conversion from the ( @xmath340 ) coordinate system to orbital elements ; in cases where the orbital uncertainties are still large , this conversion can become inaccurate @xcite , leading to the diagonal elements underestimating the uncertainty in @xmath49 .
the accuracy of the ossos astrometry means that for most of the objects listed in table [ t : detections ] , the minimum- and maximum-@xmath49 orbits converge to within @xmath343 of the best - fit @xmath49 even with a relatively short observational arc ; however , we still use the @xcite procedure for assessing the acceptable range in @xmath49 . the best - fit , minimum-@xmath49 , and maximum-@xmath49 orbits
are integrated forward in time for 10 myr using swift @xcite under the influence of the sun and the four giant planets .
we check for libration of any resonance angle ( defined by equation [ eq : full_phi ] ) for all @xmath11 resonances with @xmath344 and @xmath345 within 2% of the best - fit @xmath49 .
an object is securely resonant if all three clones librate for more than half of the 10 myr integration .
insecure resonant objects ( indicated by ` i ' in table [ t : detections ] ) have a best - fit orbit that is resonant but at least one other clone that does meet this criterion .
objects listed as ` ih ' in table [ t : detections ] had to be classified by hand as resonant because the libration behavior of the clones was too messy to be correctly identified as libration by the automated code . to better assess the likelihood that an insecure object is resonant and to determine the uncertainty in the libration amplitude for securely resonant objects , we integrated a distribution of clones for each resonant detection .
we generated 250 orbits by using the full ( @xmath340 ) covariance matrix to produce gaussian deviations from the best - fit orbit s ( @xmath340 ) .
we integrated all 250 clones for 10 myr and measured the libration amplitude , @xmath346 , over ten 1 myr windows and averaged the ten @xmath76 values .
the resulting distribution in @xmath76 is often not symmetric around the best - fit orbit s @xmath76 . in cases
where the distribution is roughly symmetric about the best - fit @xmath76 , we use the cumulative distribution in @xmath76 to find the @xmath347 values that bracket the central 67% of the distribution . in cases where it is highly asymmetric about the best - fit
( see figure [ af : aphi - dist ] ) , we take the @xmath347 uncertainty to be the range from the minimum @xmath76 to the point in the cumulative distribution where 67% of the clones
have smaller @xmath76 ( or the reverse if the asymmetric peak in the @xmath76 distribution occurs at large @xmath76 ) . in some cases where
the libration behavior is not well - behaved and the test particles slip in and out of libration , the time histories are examined manually to determine the percentage of the clones that are resonant for at least half of the 10 myr simulation ; in these instances no value of @xmath76 is given in table [ t : detections ] , and just the percentage of resonant clones is listed . for two plutinos .
the left panel shows ossos plutino o3o15 , where the @xmath76 distribution is nearly symmetric around the best - fit orbit s @xmath76 .
the right panel shows o3o20pd , which has a very asymmetric @xmath76 distribution . in both panels the best - fit orbit s @xmath76
is indicated by a black circle and the arrows show what we are calling the @xmath347 uncertainties.,width=595 ]
this section gives a more detailed accounting of the modeling of each resonant population within the ossos survey simulator . in the plutino section below , we also outline some validation testing we have performed to ensure that the simplified , parameterized models of the resonances are adequate representations of the resonant populations given the current data . in all cases below , we select orbital elements for the epoch jd 2453157.5 , at which time neptune s mean longitude is @xmath348 ( the survey simulator propagates the orbits to the appropriate epoch for each ossos observation ) .
the first step in generating plutinos within the survey simulator is to choose whether the plutino is also in the kozai resonance . for non - kozai plutinos , the procedure for choosing
an orbit is as follows : * @xmath49 is randomly chosen uniformly in the range @xmath349 au * @xmath22 is chosen randomly from equation [ eq : e ] * @xmath23 is chosen randomly from equation [ eq : i ] * the object s @xmath76 is chosen randomly from the specified distribution ( see section [ ss:32iaphi ] ) * @xmath13 is then given by @xmath350 where @xmath51 is a random number distributed uniformly from @xmath351 * @xmath33 is randomly chosen uniformly in the range @xmath352 * @xmath16 is randomly chosen uniformly in the range @xmath352 * @xmath38 is fully constrained by the above choices and is given by @xmath353 the object s absolute magnitude is then chosen randomly from equation [ eq : spl ] , fully specifying the object s position and brightness . the procedure for selecting @xmath49 and @xmath22
is simplified slightly compared to that in g12 and @xcite , who included the shape of the resonance s @xmath354 phase space in this selection by rejecting ( and re - selecting ) @xmath49 and @xmath22 in instances where the selection falls outside the resonance boundaries ; we find that this complication is unnecessary because the exact value of @xmath49 does not change the observability of an object in the survey simulator ( in fact , the results of our modeling would not change if we completely eliminated the variations in @xmath49 and just assigned every plutino @xmath355 au ) .
this procedure is simplified compared to real resonant dynamics , so we have performed a few basic tests to ensure that it provides an adequate representation of the plutino population .
the two major simplifications made above are ( 1 ) the independent selection of @xmath22 , @xmath23 , and @xmath76 and ( 2 ) using a simple harmonic oscillator to represent the time evolution of @xmath13 ( equation [ eq : phitime ] ) while ignoring the small changes in osculating orbital elements ( particularly @xmath22 and @xmath23 ) that occur over a resonant cycle .
to test the implications of the first simplifying assumption , we constructed various plutino populations that have identical 1-dimensional distributions of @xmath22 , @xmath23 , and @xmath76 , but within the generating procedure we imposed different relationships between these distributions to see how the relationship affects the distribution of simulated detections and the total population estimates generated by the survey simulator .
an example : depending on how an object becomes trapped in resonance , there can be a relationship between its eccentricity and its libration amplitude ( e.g. , * ? ? ?
a correlation between libration amplitude and eccentricity in the real population of plutinos could introduce observational biases that would not be properly accounted for in a survey simulator model that treats those two distributions independently .
the result could be a model where the real and simulated detections match in terms of the individual , 1-dimensional distributions but has an inaccurate total population estimate because the correlation was not modeled .
figure [ f : correlation_example ] shows two plutino models with identical intrinsic @xmath22 and @xmath76 distributions : one model has a population in which low libration amplitude simulated objects have higher eccentricities than large libration amplitude objects and the other model has no correlation . in this case
there are small differences between the simulated detected @xmath76 distributions for the two models and a small difference in the total population estimate ; note that the pattern of small differences here will depend on the sky coverage of a given survey .
however the differences are much smaller than the model uncertainties given a hypothetical observed sample size of 100 .
a test of correlations between @xmath23 and @xmath76 produces similar results .
given the current sample size of 21 ossos detections and 24 cfeps detections for the plutinos ( and smaller sample sizes for all the other resonances ) , the simplified , independent treatment of the @xmath22 , @xmath23 , and @xmath76 distributions is adequate .
population estimates.,width=576 ] to examine our second simplifying assumption , we tested whether more accurately generating objects with positions and velocities consistent with real resonant behavior can affect the results of model testing . within the survey simulator ,
resonant objects are assigned values of ( @xmath49,@xmath22,@xmath23,@xmath16,@xmath38,@xmath13 ) randomly from within the desired distributions , and then those values are interpreted as instantaneous osculating elements that can be translated into a sky position . in reality , however , an object s osculating @xmath49,@xmath22,@xmath23 vary depending on where the object is in the resonant cycle , as shown in fig [ f : res_cycle ] .
this means there is a relationship between the exact current osculating @xmath49,@xmath22,@xmath23 and the value of @xmath13 ( which determines @xmath16 , @xmath38 , and @xmath33 ) that is not correctly modeled when these parameters are chosen independently .
the time weighting of @xmath13 for a real population can also differ from the simple harmonic oscillator model we use in the survey simulator .
figure [ f : res_cycle ] shows the variation of @xmath49 , @xmath22 , and @xmath13 for two different plutinos from a numerical integration ; the left panel shows a plutino with moderate @xmath76 and sinusoidal variations while the right panel shows a larger @xmath76 plutino with non - sinusoidal variations .
the non - sinusoidal @xmath13 variations in the real plutinos means that the survey simulator with its simplified sinusoidal time - weighting is not correctly accounting for the amount of time a real plutino spends at specific values of @xmath13 ( which determines where on the sky relative to neptune the object s perihelion most often occurs ) . to see
if more accurate time - weighting of @xmath13 and the inclusion of the associated variations in @xmath49 and @xmath22 would significantly alter our conclusions about the plutino population , we performed short numerical integrations of @xmath356 plutinos whose time histories can be used to generate plutinos within the survey simulator .
we then took our nominal plutino @xmath43 , @xmath22 , and @xmath76 model from section [ s:32 ] and generated objects in the survey simulator either by our standard procedure or by selecting objects orbital elements from the numerical simulations . in this planar test case
, we find that sampling from the numerical integrations does not meaningfully change the distribution or number of simulated observed objects .
there are also no disallowed combinations of @xmath22 , @xmath23 , and @xmath76 in the planar case ( i.e. all values of @xmath22 exist in resonant phase space for all values of @xmath76 ) , so our independent treatment of these distributions is not artificially generating plutinos that would actually be in a non - resonant portion of the phase space . ,
@xmath22 , and @xmath13.,width=595 ] the procedure for kozai plutinos is different because the kozai resonance presents an additional constraint on the orbit due to the libration of @xmath38 around either @xmath7 or @xmath39 . in the idealized three body problem , where we just consider
the evolution of a test particle under the influence of the sun and neptune , a plutino librating in the kozai resonance will follow a closed path in @xmath22-@xmath38 space ( see for example figure 4 in @xcite ) .
this trajectory represents a path in @xmath22-@xmath38 space with a constant value of the resonant hamiltonian ( see for example the expression for the approximate hamiltonian given in @xcite )
. sets of hamiltonian level curves ( and thus libration trajectories ) can be grouped according to the value of the z component ( normal to the reference plane ) of the orbit s angular momentum , @xmath357 .
@xmath358 is a preserved quantity , meaning that a test particle s eccentricity and inclination evolution are coupled as the test particle traces out its trajectory in @xmath22-@xmath38 space .
a plutino s kozai libration behavior could be specified by selecting @xmath358 , calculating all of the associated kozai resonant hamiltonian level curves for that @xmath358 , selecting one of these level curves , selecting a ( @xmath22,@xmath38 ) pair from that level curve trajectory , and then calculating the appropriate @xmath23 based on @xmath22 and @xmath358 .
one could then select a 3:2 libration amplitude , set @xmath13 as above in the non - kozai case , select a random @xmath33 , and calculate @xmath359 to fully specify the orbit and position ( this is the procedure used in @xcite ) .
the above procedure would be the best way to generate kozai plutinos in the three body problem , but the real kozai plutinos do not follow such a nicely defined set of trajectories . in figure
[ af : kz - example ] we show the evolution of one of the ossos kozai plutinos in the 10 myr classification integration superposed over hamiltonian level curves calculated using the approximate hamiltonian from @xcite for two different values of @xmath358 ; the values of @xmath358 are parameterized by @xmath360 , the maximum inclination allowed by conservation of @xmath358 at @xmath361 .
the observed plutino in figure [ af : kz - example ] has @xmath362 , but the figure shows that its evolution in @xmath363 space is not well described by any of the @xmath362 hamiltonian level curves . from visual inspection of the evolution of the five ossos kozai plutinos , we find that all of the kozai plutinos have an average eccentricity over their kozai cycles of @xmath364 despite having @xmath360 ranging from @xmath365 .
the three body hamiltonian predicts that the average @xmath22 should decrease with decreasing @xmath360 , but the real kozai plutinos do not show this and all have average @xmath22 that approximately matches the hamiltonian level curves for @xmath366 ( see the right panel of figure [ af : kz - example ] ) .
the exact path a kozai plutino follows in @xmath363 space does change with changing @xmath360 .
the amplitude of the eccentricity variations in the real kozai plutinos seems to decrease with decreasing @xmath360 , which can be seen in the right panel of figure [ af : kz - example ] where the real @xmath367 object s path is flattened relative to the @xmath366 hamiltonian curves ; a real object with @xmath366 would more closely match the eccentricity amplitude predicted by the level curves .
( left panel ) and @xmath368 ( right panel ) .
the object has @xmath362 , but the poor match in the left panel indicates that the expression for the approximate hamiltonian in the three body problem as given by @xcite is not a good match to the dynamics in the full integrations.,width=595 ] because all of the ossos kozai plutinos appear to share an average eccentricity with the @xmath366 hamiltonian curves , we use this one set of level curves to choose @xmath22 and @xmath38 for all of the kozai plutinos in the survey simulator .
this is the same approach used in g12 because all of their kozai plutinos were well described by those level curves and all had @xmath360 near @xmath369 .
some of the ossos kozai plutinos have significantly lower current orbital inclinations and @xmath360 than the g12 kozai plutinos , so we have altered the way the inclinations of kozai plutinos are calculated within the survey simulator .
after @xmath22 and @xmath38 are chosen from the @xmath366 level curves , we then choose a new value of @xmath360 uniformly in the range @xmath365 and then set the generated kozai plutino s inclination such that @xmath370 .
thus the full procedure we use for generating a kozai plutino is as follows : * @xmath49 is randomly chosen uniformly in the range @xmath349 au * a hamiltonian level curve is randomly chosen from the kozai resonant @xmath366 level curves * a position on that level curve is randomly chosen ( with the simplified assumption that equal time is spent on all parts of the level curve ) , setting @xmath22 and @xmath38 * @xmath360 is chosen randomly from the uniform range @xmath365 * @xmath23 is then calculated from @xmath370 * the object s @xmath76 is chosen randomly from the specified distribution ( see section [ ss:32iaphi ] ) * @xmath13 is given by @xmath350 where @xmath51 is a random number distributed uniformly from @xmath351 * @xmath33 is randomly chosen uniformly in the range @xmath352 * @xmath16 is fully constrained by the above choices and is given by @xmath359 the object s absolute magnitude is then chosen randomly from equation [ eq : spl ] , fully specifying the object s position and brightness . this procedure is not entirely self - consistent in the use of analytical models of the kozai population , but as shown in figure [ af : kz - example ] , the real evolution of the kozai plutinos matches the hamiltonian level curves in a qualitative rather than quantitative sense .
this procedure yields synthetic detected kozai plutinos that are an acceptable match to the real detections . in future work
the procedure will be re - evaluated to see if a more accurate representation of kozai phase space is required .
the procedure for choosing an orbit for a 5:2 object is as follows : * @xmath49 is randomly chosen uniformly in the range @xmath371 au * @xmath22 is chosen randomly from equation [ eq : e ] * @xmath23 is chosen randomly from equation [ eq : i ] * the object s @xmath76 is chosen randomly from the specified distribution ( see section [ s:52 ] ) * @xmath13 is then given by @xmath350 where @xmath51 is a random number distributed uniformly from @xmath351 * @xmath33 is randomly chosen uniformly in the range @xmath352 * @xmath16 is randomly chosen uniformly in the range @xmath352 * @xmath38 is fully constrained by the above choices and is given by @xmath372 the object s absolute magnitude is then chosen randomly from equation [ eq : spl ] , fully specifying the object s position and brightness . to generate an orbit for a 2:1 resonant object ,
we first decide if it is a symmetric or asymmetric librator .
if it is a symmetric librator , we then choose the orbit as follows * @xmath49 is randomly chosen uniformly in the range @xmath373 au * @xmath22 is randomly chosen uniformly from @xmath248 * @xmath23 is chosen randomly from equation [ eq : i ] * @xmath76 is randomly chosen uniformly from @xmath374 * @xmath13 is then given by @xmath350 where @xmath51 is a random number distributed uniformly from @xmath351 * @xmath33 is randomly chosen uniformly in the range @xmath352 * @xmath16 is randomly chosen uniformly in the range @xmath352 * @xmath38 is fully constrained by the above choices and is given by @xmath375 for asymmetric librators , we choose the orbit as follows : * @xmath49 is randomly chosen uniformly in the range @xmath373 au * @xmath23 is chosen randomly from equation [ eq : i ] * @xmath22 is randomly chosen uniformly from @xmath250 * the center of libration , @xmath376 , for that eccentricity is calculated from figure 4 in @xcite * the maximum @xmath76 , for that eccentricity is also calculated from figure 4 in @xcite * @xmath76 is randomly chosen uniformly from @xmath251 * the object is chosen to be in either the leading or trailing asymmetric island and @xmath376 is adjusted appropriately * @xmath13 is then given by @xmath377 where @xmath51 is a random number distributed uniformly from @xmath351 * @xmath33 is randomly chosen uniformly in the range @xmath352 * @xmath16 is randomly chosen uniformly in the range @xmath352 * @xmath38 is fully constrained by the above choices and is given by @xmath375 the object s absolute magnitude is then chosen randomly from equation [ eq : spl ] , fully specifying the object s position and brightness .
the above ranges in @xmath22 and @xmath76 for the symmetric and asymmetric librators are chosen based on numerical modeling and stability analysis of the 2:1 resonance by @xcite and @xcite .
in order to compare population estimates for the resonant populations from the des @xcite , cfeps ( g12 ) , and ossos @xcite , we have to know how to compare the @xmath43 magnitudes of the resonant objects in the surveys different filters .
magnitudes for the des objects are given in the @xmath378 band in @xcite , whereas the @xmath43 magnitudes of the cfeps detections are in @xmath379 band @xcite and the ossos detections are in @xmath380 band . as discussed in section [ ss:32hd ] , the average color for comparing cfeps and ossos resonant objects is @xmath277 based on cfeps objects that were observed in both @xmath380 and @xmath379 band @xcite .
fortunately , some objects seen in the des were serendipitously present in the cfeps fields , allowing a direct estimation of the average color is between the g and vr photometric systems ( light curves , while present , should average out ) .
we compared the list of cfeps resonant detections and des resonant detections and found all instances of overlapping detections where both @xmath297 and @xmath381 are measured .
these objects are listed in table [ at : colors ] .
we find that these resonant objects have colors in the range @xmath382 with an average color of @xmath293 .
this differs from the assumption of @xmath294 in @xcite ; although @xmath291 falls inside the lower extremity of the range , typical colors are much larger .
as such , a comparison of population estimates requires a larger shift between the surveys .
l l l 5:2 & l4j06pd & 2002 gp32 & 7.03 & 6.99 & 0.04 + 5:2 & l4h02pd & 2004 eg96 & 8.36 & 8.01 & 0.35 + 7:4 & k02o03 & 2000 op67 & 8.13 & 7.27 & 0.86 + 2:1 & l4v06 & 2004 vk78 & 8.5 & 8.16 & 0.34 + 7:3 & l5c19pd & 2002 cz248 & 8.5 & 8.27 & 0.23 + 3:2 & l5i06pd & 306792 ( 2001 kq77 ) & 7.48 & 7.2 & 0.28 + 3:2 & l4v13 & 2002 vv130 & 7.6 & 7.51 & 0.09 + 12:5 & l5c12 & 119878 ( 2002 cy224 ) & 6.69 & 6.1 & 0.59 + 3:1 & l4v08 & 307463 ( 2004 vu130 ) & 6.95 & 6.1 & 0.85 + [ at : colors ] , a. , levison , h. f. , & gomes , r. 2008 , the dynamical structure of the kuiper belt and its primordial origin , ed .
barucci , m. a. , boehnhardt , h. , cruikshank , d. p. , morbidelli , a. , & dotson , r. , 275292 | the first two observational sky blocks " of the outer solar system origins survey ( ossos ) have significantly increased the number of well - characterized observed trans - neptunian objects ( tnos ) in neptune s mean motion resonances .
we describe the 31 securely resonant tnos detected by ossos so far , and we use them to independently verify the resonant population models from the canada - france ecliptic plane survey ( cfeps ; * ? ? ?
* ) , with which we find broad agreement .
we confirm that the 5:2 resonance is more populated than models of the outer solar system s dynamical history predict ; our minimum population estimate shows that the high eccentricity ( @xmath0 ) portion of the resonance is at least as populous as the 2:1 and possibly as populated as the 3:2 resonance .
one ossos block was well - suited to detecting objects trapped at low libration amplitudes in neptune s 3:2 resonance , a population of interest in testing the origins of resonant tnos .
we detected three 3:2 objects with libration amplitudes below the cutoff modeled by cfeps ; ossos thus offers new constraints on this distribution .
the ossos detections confirm that the 2:1 resonance has a dynamically colder inclination distribution than either the 3:2 or 5:2 resonances .
using the combined ossos and cfeps 2:1 detections , we constrain the fraction of 2:1 objects in the symmetric mode of libration to be 0.20.85 ; we also constrain the fraction of leading vs. trailing asymmetric librators , which has been theoretically predicted to vary depending on neptune s migration history , to be 0.050.8 .
future ossos blocks will improve these constraints . |
since the measurement of the large @xmath6 at reno , daya bay , and double chooz , the precise measurement of neutrino mass hierarchy , the sign of @xmath7 , has become the focus in neutrino physics@xcite .
it had been believed that the neutrino mass hierarchy can be determined through long - baseline experiments , mainly using accelerator neutrino beams .
recently , the capability of a reactor neutrino experiment at an intermediate baseline to distinguish normal or inverted hierarchy was reported . for an intermediate - baseline neutrino experiment ,
many approaches have been proposed ; they can be categorized into the @xmath8 analysis methods , which are discussed in refs.@xcite , and the fourier - transform methods@xcite .
the @xmath8 analysis methods based on the newly adopted bayesian approach utilizes all the available information from experiments , and it is straightforward to incorporate the uncertainties in order to evaluate the sensitivity , providing robust and complementary results in the fourier - transform methods@xcite .
although the @xmath8 analysis methods are attractive and interesting , the fourier - transform methods are more intuitive .
the prominent merit of the fourier - transform methods are that the mass hierarchy can be extracted without precise knowledge of the reactor antineutrino spectrum , the absolute value of the large @xmath1 , and the energy scale of a detector .
the fourier - transform methods were introduced to enhance and visualize the structures of mass hierarchy in the frequency spectrum , as first discussed in ref .
@xcite . in principle
, the mass hierarchy can be determined through precise measurements of @xmath1 and @xmath2 .
as @xmath9 is very small and is only @xmath10 of @xmath1 , we have to measure @xmath1 and @xmath2 with a precision much better than @xmath11 .
however , @xmath1 and @xmath2 have been measured in many experiments with only @xmath12 precision@xcite .
the intermediate baseline based on reactor neutrino experiments has been explored using the precise measurement of distortions of the energy spectrum with negligible matter effect . learned _
proposed a new method to distinguish normal and inverse hierarchy after a fourier transform of the @xmath3 spectrum of reactor neutrinos@xcite .
they pointed out that the fourier power spectrum has a small but not negligible shoulder next to the main peak , and its relative position could be used to extract the mass hierarchy while a non - zero @xmath13 is considered . in this paper , we analyze the sensitivity of medium - baseline reactor antineutrino experiments to the neutrino mass hierarchy for a baseline range of 3060 km and overall energy resolution , @xmath14 , in the range of 0 to @xmath15 with the fourier - transform method . the optimal baseline length is estimated based on the expected probability of determination .
in a nuclear reactor , antineutrinos are mainly produced via the @xmath16-decay of the fission products of the four types of radioactive isotopes , @xmath17 , and @xmath18 , in the fuel .
the antineutrino flux having energy @xmath19 in mev with thermal power @xmath20 in @xmath21 is given as @xmath22 where @xmath23 and @xmath24 are the relative fission contribution and the energy released per fission of isotope @xmath25 , respectively .
further , @xmath26 is the number of antineutrinos produced per fission and is obtained as follows@xcite : @xmath27 the antineutrino flux is modulated by neutrino oscillation .
the antineutrino survival probability @xmath28 is expressed as @xmath29 the oscillation phase @xmath30 is defined as @xmath31 with a baseline @xmath32 . as @xmath33 and @xmath34
appear simultaneously in eq .
( 3 ) , the effects of mass hierarchy on @xmath28 are hardly recognized . by using the relation between the squared mass differences , @xmath35 we rearrange eq .
( 3 ) to eliminate the @xmath34 term as follows : @xmath36 the plus ( minus ) sign in the fifth term on the right - hand side of eq . ( 6 ) corresponds to the normal ( inverted ) mass hierarchy or nh ( ih ) in short . in ongoing reactor experiments
, we assume that protons will be used as targets to detect electron antineutrinos via the inverse - beta - decay ( ibd ) , which produces a neutron and positron .
the antineutrino distribution observed with a detector having @xmath37 free protons can be expressed for an exposure time @xmath38 as follows : @xmath39 where @xmath40 is the cross section of the ibd process and @xmath32 is the baseline length .
we use the distribution of the expected antineutrino events from the above expression . for the ibd cross section ,
we use the following expression from vogel and beacom s work@xcite : @xmath41 in order to study the sensitivity of the mass hierarchy , we use fourier - transform method together with monte carlo simulations to compare the simulated ibd energy spectrum with the expected spectrum in both the nh and ih cases . taking into account the detector response , the reactor electron antineutrino @xmath0 @xmath3 spectrum becomes @xmath42 where @xmath43 is the actual @xmath0 energy , @xmath44 is the observed @xmath0 energy with the detector response , @xmath45 is the energy resolution , and @xmath46 describes the detector response function including effects such as the energy resolution and energy scale . in this study , we take the normalized gaussian function as the response function : @xmath47 as the neutrino energy is usually measured using scintillators , the energy is typically proportional to the number of photoelectrons , and the error is dominated by the photoelectron statistics .
therefore , the neutrino energy resolution is proportional to @xmath48 . in general
, the detector energy resolution is parameterized into two parts : @xmath49 the first term represents the uncertainty from statistical fluctuation , and the second term originates from the systematic uncertainty . in this study , @xmath50 is assumed for simplicity .
before the measurement of the surprisingly large @xmath51 , it had been known that at the oscillation maximum of @xmath52 , which corresponds to a baseline of approximately 58 km , the sensitivity to the mass hierarchy is maximized at @xmath53 .
as @xmath51 is no longer small , the sensitivity to mass hierarchy needs to be explored as a function of the baseline , @xmath32 , and the detector energy resolution , @xmath54 . in this study , each monte carlo experiment generates a set of 500,000 @xmath0 events by sampling @xmath55 with input parameters , @xmath32 and @xmath54 .
the default oscillation parameters are taken from ref .
@xcite and listed in table [ parameter ] , together with the explored ranges of baseline and energy resolution .
.[parameter]default values of neutrino oscillation parameters and the explored ranges of other input parameters .
[ cols="^,^,^,^",options="header " , ]
we have studied the experimental requirements to determine neutrino mass hierarchy using fourier sine and cosine transform of the reactor neutrino @xmath3 spectrum at @xmath4 .
the parameters @xmath56 and @xmath57 were defined to extract features of the fourier sine and cosine spectra , and the mass hierarchy could be obtained from the determination probability , @xmath58 based on bayes theorem .
since the effect of varying @xmath1 over its uncertainty has little effect on the fst and fct spectra at @xmath4 , the @xmath58 is less dependent on the uncertainty of @xmath1 than the value of @xmath6 . as defined in eq .
( 16 ) and eq .
( 17 ) , the probability @xmath58 is closely related to the different features of each mass hierarchy .
each value of @xmath58 indicates the probability that an experiment will be found inside its correct @xmath59 or @xmath60 region on the @xmath56 and @xmath61 planes .
these different features from different neutrino mass hierarchies suggest that the analysis method described in this paper , which is a simple and straightforward approach , can be used to determine the neutrino mass hierarchy by using the determination probability @xmath58 based on the fourier sine and cosine transform of the @xmath3 spectrum .
this work was supported by the national research foundation of korea ( nrf-2013r1a1a2011108 ) and also supported , in part , by the nrf grant no .
2009 - 0083526 .
00 j. k. ahn _
et al . , _ ( reno collaboration ) phys .
lett . * 108 * , 191802 ( 2012 ) .
f. p. an _ et al . , _
( daya bay collaboration ) phys .
lett . * 108 * , 171803 ( 2012 ) .
m. apollonio _
et al . _ ,
phys . rev . * d61 * , 012001 ( 2000 ) .
ghoshal and s. t. petcov , jhep 1103 ( 2011 ) 058 [ arxiv:1011.1646 ] .
x. qian , d. a. dwyer , r. d. mckeown , p. vogel , w. wang and c. zhang , arxiv:1208.1551 ( 2012 ) .
p. ghoshal and s. t. petcov , arxiv:1208.6473 ( 2012 ) .
li , j. cao , y. wang and l. zhan , arxiv:1303.6733 ( 2013 ) .
yoshitaro takaesu , arxiv:1304.5306 ( 2013 ) .
e. ciuffoli , j. evslin and x. zhang , arxiv:1208.1991 ( 2012 ) .
e. ciuffoli , j. evslin and x. zhang , arxiv:1209.2227 ( 2012 ) .
x. qian , a. tan , w. wang , j. j. ling , r. d. mckeown and c. zhang , arxiv:1210.3651 ( 2012 )
. j. learned , s. t. dye , s. pakvasa and r. c. svoboda , phys .
rev . * d78 * , 071302 ( 2008 ) .
j. beringer _
et al . , _ phys . rev . * d86 * , 010001 ( 2012 ) .
p. vogel and e. engel , phys .
* d39 * , 3378 ( 1989 ) .
p. vogel and j. f. beacom , phys .
* d60 * , 053003 ( 1999 ) .
p. huber and t. schwetz , phys . rev . *
d40 * , 053011 ( 2004 ) . | this article reports the optimized experimental requirements to determine neutrino mass hierarchy using electron antineutrinos ( @xmath0 ) generated in a nuclear reactor .
the features of the neutrino mass hierarchy can be extracted from the @xmath1 and @xmath2 oscillations by applying the fourier sine and cosine transform to the @xmath3 spectrum . to determine the neutrino mass hierarchy above 90% probability ,
the requirements on the energy resolution as a function of the baseline are studied at @xmath4 .
if the energy resolution of the neutrino detector is less than @xmath5 and the determination probability obtained from bayes theorem is above 90% , the detector needs to be located around 4853 km from the reactor(s ) to measure the energy spectrum of @xmath0 .
these results will be helpful for setting up an experiment to determine the neutrino mass hierarchy , which is an important problem in neutrino physics .
pacs numbers : : 13.15.+g , 14.60.pq , 14.60.lm , 29.85.+c keywords : : neutrino mass hierarchy , reactor experiment , fourier transform |
the study of mixing in the @xmath40 and @xmath41 sectors has provided a wealth of information to guide the form and content of the standard model . in the framework of the standard model , mixing in the charm meson sector
is predicted to be small @xcite , making this an excellent place to search for non - standard model effects .
similarly , measurable _ cp _ violation ( cpv ) phenomena in strange @xcite and beauty @xcite mesons are the impetus for many current and future experiments @xcite .
the standard model predictions for cpv for charm mesons are of the order of @xmath42 @xcite , with one recent conjecture of nearly @xmath43 @xcite .
observation of cpv in charm mesons exceeding the percent level would be strong evidence for non - standard model processes .
a @xmath0 can evolve into a @xmath44 through `` ordinary '' on - shell intermediate states , or through off - shell intermediate states such as as those that might be present due to new physics .
we denote the amplitude through the former ( latter ) states by @xmath45 @xmath46 , in units of @xmath47 @xcite . the standard model contributions to @xmath48
are suppressed to @xmath49 and the glashow - illiopolous - maiani @xcite cancelation could further suppress @xmath50 down to @xmath51 .
many non - standard model processes could lead to @xmath52 .
contributions to @xmath48 at this level could result from the presence of new particles with masses as high as 100 1000 tev @xcite .
signatures of new physics include @xmath53 and _ cp _ violating interference between @xmath48 and @xmath54 or between @xmath48 and a direct decay amplitude .
wrong sign ( ws ) processes , such as @xmath6 , can proceed directly through doubly cabbibo - suppressed decay ( dcsd ) or through mixing and subsequent cabbibo favored decay ( cfd ) . both dcsd and mixing followed by cfd
can contribute to the time integrated ws rate , @xmath55 and the inclusive _ cp _ asymmetry @xmath56 , where @xmath57 , @xmath58 is the charge conjugated quantity , and @xmath59 is a wrong - sign final state , such as @xmath60 .
the different contributions to @xmath61 and @xmath62 can be separated by studying the proper decay time dependence of ws final states , as we have done in @xmath63 @xcite , and has been done by focus in @xcite .
the differential ws rate relative to the right - sign ( rs ) process , in time units of the mean @xmath0 lifetime , @xmath64 @xcite , is @xmath65 e^{-t}$ ] @xcite , where @xmath66 and @xmath67 , where @xmath68 is the relative rate of dcsd and @xmath69 is the strong phase between the dcsd and cfd amplitudes .
there are theoretical arguments that @xmath69 should be small @xcite , although this should not be taken for granted .
the coefficient of the term quadratic in @xmath70 is proportional to the relative rate of mixing , @xmath71 .
all of the analyses discussed herein , unless otherwise stated , use the same data set and reconstruction techniques described below .
the data set was accumulated between february 1996 and february 1999 and corresponds to @xmath72 of @xmath73 collision data at @xmath74 gev provided by the cornell electron storage ring ( cesr ) .
the data were recorded by the cleo ii detector @xcite upgraded with the installation of a silicon vertex detector ( svx ) @xcite and by changing the drift chamber gas from an argon - ethane mixture to a helium - propane mixture @xcite .
the upgraded configuration is referred to as cleo ii.v .
the monte carlo simulation of the cleo ii.v detector was based upon geant @xcite , and simulated events were processed in the same manner as the data .
the @xmath0 candidates are reconstructed through the decay sequence @xmath75 @xcite .
the charge of the slow pion ( @xmath76 ) tags the flavor of the @xmath0 candidate at production .
the charged daughters of the @xmath0 are required to leave hits in the svx and these tracks are constrained to come from a common vertex in three dimensions .
the trajectory of the @xmath0 is projected back to its intersection with the cesr luminous region to obtain the @xmath0 production point .
the @xmath76 is refit with the requirement that it come from the @xmath0 production point , and the confidence level of the @xmath77 of this refit is used to reject background . the energy release in the @xmath78 decay , @xmath79 , obtained from the above technique
is observed to have a width of @xmath80 kev @xcite , which is a combination of the intrinsic width and our resolution , where @xmath81 and @xmath82 are the reconstructed masses of the @xmath0 and @xmath83 candidates respectively , and @xmath84 is the charged pion mass .
the reconstruction technique discussed above has also been used by cleo to measure the @xmath85 intrinsic width , @xmath86 kev ( preliminary ) @xcite . in the mixing analyses described below
, the distribution of candidates in the @xmath87 vs @xmath81 plane are fit to determine both right - sign and wrong - sign yields .
we calculate @xmath70 using only the vertical component of the @xmath0 candidate flight distance .
this is effective because the vertical extent of the cesr luminous region has @xmath88
m @xcite .
the resolution on the @xmath0 decay point ( @xmath89 , @xmath90 , @xmath91 ) is typically @xmath92 m in each dimension .
we measure the centroid of the cesr luminous region ( @xmath93 , @xmath94 , @xmath95 ) using @xmath96 ( @xmath97 ) events from sets of data with integrated luminosity of several pb@xmath98 , obtaining a resolution on the centroid of less than @xmath99 m .
we express @xmath70 as @xmath100 , where @xmath101 is the vertical component of the total momentum of the @xmath0 candidate .
the error in @xmath70 , @xmath102 , is typically @xmath103 ( in @xmath0 lifetimes ) , although when the @xmath0 direction is near the horizontal plane @xmath102 can be large .
_ cp _ violation in charm meson decay is expected to be small in the standard model , which makes charm meson decay a good place to look for non - standard model effects .
cabbibo suppressed charm meson decays have all the necessary ingredients for _ cp _ violation
multiple paths to the same final state and a weak phase .
however , in order to get sizable _ cp _ violation , the final state interactions need to contribute non - trivial phase shifts between the amplitudes .
large final state interactions are a likely reason why the prediction for the ratio of branching ratios of @xmath104 yields a value roughly half of the observed value @xcite , hence these may provide a good hunting ground for _ cp _ violation .
previous searches for mixing - induced @xcite or direct @xcite _ cp _ violation in the neutral charm meson system have set limits of @xmath105 or a few percent , respectively .
we present results of searches for direct _ cp _ violation in neutral charm meson decay to pairs of light pseudo - scalar mesons : @xmath106 , @xmath107 , @xmath108 , @xmath109 and @xmath110 .
the asymmetry we want to measure , @xmath113 can be obtained from the asymmetry @xmath114 the slow pion and @xmath0 are produced by the _
cp_-conserving strong decay of the @xmath83 , so the slow pion serves as an unbiased flavor tag of the @xmath0 .
the decay asymmetry can be obtained from the apparent production asymmetry shown above because the production of @xmath115 is _ cp_-conserving .
the asymmetry result is obtained by fitting the energy release ( @xmath87 ) spectrum of the @xmath75 events .
the @xmath0 mass spectra are fit as a check .
the background - subtracted @xmath87 spectrum is fit with a signal shape obtained from @xmath9 data and a background shape determined using monte carlo .
figure [ fig : hill ] shows our fits to both @xmath116 and @xmath117 distributions found in data , for both signs of the slow pion flavor tag .
the parameters of the slow pion dominate the @xmath87 distribution , so all modes have the same shape .
we do the fits in bins of @xmath0 momentum to eliminate any biases due to differences in the @xmath0 momentum spectra between the data and the mc .
the preliminary results are @xmath118 and @xmath119 .
we use many different variations of the fit shapes , both empirical and analytical , to assess the systematic uncertainties due to the fitting procedure ( 0.69% ) .
we also consider biases due to the detector material ( 0.07% ) , the reconstruction software ( 0.48% ) , and forward
backward acceptance variations ( @xmath120 pairs are not produced symmetrically in the forward / backward directions in @xmath73 collisions at @xmath121 gev , and the collision point was not centered exactly in the middle of the detector ) ( 0.014% ) .
the measured asymmetries are consistent with zero , and no _ cp _ violation is seen .
these results compared to previous measurements @xcite are shown in figures [ fig : hillkk ] and [ fig : hillpp ] .
this analysis @xcite differs from the other analyses presented in this paper in some of its reconstruction techniques and in the data set used .
the @xmath125 and @xmath108 final states do not provide sufficiently precise directional information about their parent @xmath0 to use the intersection of the @xmath0 projection and the cesr luminous region to refit the slow pion as described in the general experimental technique section .
the @xmath110 final state is treated the same for consistency .
this analysis uses the data from both the cleo ii and cleo ii.v configurations of the detector , totaling 13.7 @xmath126 of @xmath73 collision data at @xmath121 gev .
the @xmath127 and @xmath128 candidates are constructed using only good quality tracks and showers . the tracks ( showers )
whose combined invariant mass is close to the @xmath127 ( @xmath128 ) mass are kinematically constrained to the @xmath127 ( @xmath128 ) mass , improving the @xmath0 mass resolution .
the tracks used to form @xmath127 candidates are required to satisfy criteria designed to reduce background from @xmath129 decays and combinatorics .
candidate events with reconstructed @xmath0 masses close to the known @xmath0 mass are selected to determine the asymmetry , @xmath130 / \left [ \gamma ( d^0 \rightarrow f ) + \gamma(\overline{d^0 } \rightarrow f)\right]$ ] .
the @xmath87 distributions of the candidates in the three decay modes are shown in figures [ fig : k0pi0 ] , [ fig : pi0pi0 ] , and [ fig : k0k0 ] .
a prominent peak indicative of @xmath75 decay is observed in all three distributions .
the total number of @xmath0 and @xmath44 candidates for a given final state is determined as follows .
we fit the @xmath87 distribution outside of the signal region and interpolate the fit under the signal peak to determine the background in the signal region .
we subtract the background in the signal region from the total number of events there to determine the total number of signal events .
the background shape is approximated as a non - relativistic threshold function with first and second order relativistic corrections @xmath131 .
after background subtraction , we obtain @xmath132 @xmath133 candidates , @xmath134 @xmath125 candidates , and @xmath135 @xmath110 candidates .
the difference in the number of @xmath0 and @xmath44 to a given final state is determined by taking the difference of the number of events in the signal region , and the asymmetry is obtained by dividing by the number of candidates determined above .
this method of determining the asymmetry implicitly assumes that the background is symmetric .
we have searched for any sources of false asymmetries : from the @xmath76 finding ( 0.19% ) , from the fitting ( 0.5% ) , and from the backgrounds ( 0.35% @xmath108 , 0% @xmath125 and 12% @xmath110 ) .
we find no significant biases , but apply the measured corrections and add their uncertainties to the total .
we obtain the results @xmath136 , @xmath137 and @xmath138 where the uncertainties contain the combined statistical and systematic uncertainties .
all systematic uncertainties , except for the 0.5% uncertainty assigned for possible bias in the fitting method , are determined from data and would be reduced in future higher luminosity samples .
all measured asymmetries are consistent with zero and no indication of significant _ cp _ violation is observed .
this measurement of @xmath139 is a significant improvement over previous results , and the other two asymmetries reported are first measurements .
@xmath140 candidates are reconstructed using the selection criteria described in section [ genmethod ] , with additional requirements specific to this analysis . in particular ,
@xmath141 candidates with momenta greater than 340 m@xmath142v/@xmath143 are reconstructed from pairs of photons detected in the csi crystal calorimeter .
backgrounds are reduced by requiring specific ionization of the pion and kaon candidates to be consistent with their respective hypotheses .
the rs mode was recently studied by cleo@xcite and found to have a rich dalitz structure consisting of @xmath144 , @xmath145 , @xmath146 , @xmath147 , @xmath148 , @xmath149 , and @xmath150 resonances and non - resonant contributions .
recent theoretical predictions based on u - spin symmetry arguments@xcite suggest that the wrong sign ( ws ) channel will have a different resonant substructure than the right sign ( rs ) .
this analysis allows for different average ws and rs efficiencies resulting from different ws and rs kinematic distributions in the calculation of the ws rate : @xmath151 the ratio of yields in eq .
( [ eq : rws ] ) is measured by performing a maximum likelihood fit to the two - dimensional distribution in @xmath152 and @xmath87 . the signal distribution in these variables
is taken from the rs data .
the backgrounds are broken down into three categories : 1 ) rs @xmath153 decay combined with an uncorrelated @xmath154 , 2 ) combinations from @xmath155 , @xmath156 , and @xmath157 , and 3 ) combinations from charm particle decays other than the rs or ws signal modes .
the background distributions are determined using a large monte carlo sample , which corresponds to approximately eight times the integrated luminosity of the data sample .
the @xmath87@xmath158 fit yields a ws signal of @xmath159 events and a ratio @xmath160 .
projections of the data and fit results in slices through the signal region in each variable are shown in fig .
[ fig : alexmass ] .
the statistical significance of this signal is found to be 4.9 standard deviations . the average efficiency ratio in eq .
( [ eq : rws ] ) is determined using a fit to the dalitz plot variables @xmath161 and @xmath162 in the ws data . in this fit ,
the amplitudes and phases are initialized to the rs values , and those corresponding to the @xmath163 and @xmath164 resonances are floated relative to the dominant @xmath165 and other minor contributions . combining the square of the fitted amplitude function with a parameterization of the efficiency determined using a large non - resonant monte carlo sample , we measure an average efficiency ratio of @xmath166 .
studies are under way to examine the extent and the significance of this surprising similarity between the rs and ws dalitz plots .
the dominant systematic errors in this analysis come from the uncertainty in the monte carlo background distributions ( 14% ) , uncertainty in the amplitudes and phases that are fixed in the fit ( 8% ) , and uncertainty in the background dalitz plot distributions ( 3% ) .
several powerful checks of the @xmath87@xmath158 and dalitz plot fits are performed in order to verify the validity of these results .
fits using specific background regions of the @xmath87@xmath158 plane test the sensitivity to the monte carlo background distributions .
the ws dalitz plot is also fit using hypotheses which include one of the @xmath163 , @xmath164 , or @xmath165 resonances , in order to provide an upper limit on the error due to this fit .
only the @xmath164 hypothesis leads to an efficiency ratio that differs significantly from one , but this hypothesis is strongly disfavored by the data .
we measure the wrong sign rate to be @xmath167 this result is consistent with the cleo ii.v@xcite and focus @xcite @xmath168 measurements .
this measurement of @xmath169 can be used to obtain limits on @xmath170 as a function of @xmath171 within the limits on @xmath172 set by cleo , as shown in figure [ fig : alexrws ] .
work is in progress to use the lifetime distribution of this sample to yield independent limits on @xmath172 , @xmath173 , and @xmath170 .
in the limit of no @xmath174 violation in the neutral @xmath176 system we can write the time dependent rate for @xmath177 , where @xmath59 is a cp eigenstate , as @xmath178 where @xmath179 is the average @xmath176 width , @xmath180 is the @xmath174 eigenvalue for @xmath59 , and @xmath181 where @xmath182 is the width difference between the physical eigenstates of the neutral @xmath176 . in the limit
that _ cp _ is conserved in charm decays , @xmath183 , where @xmath54 is the mixing parameter defined in section [ sec : intro ] . with the @xmath174 asymmetry for @xmath184 and @xmath185 consistent with zero as discussed in section [ sec : kkpp ] , this limit is well motivated and we will simply use @xmath54 for the rest of this section .
we can then express @xmath54 as @xmath186 where @xmath187 is the lifetime of a @xmath174 neutral state , such as @xmath188 , and @xmath189 is the lifetime of a @xmath174 even state , such as @xmath184 and @xmath185 .
thus to measure @xmath54 we simply take the ratio of the lifetimes of @xmath190 to @xmath191 and @xmath185 . since the final states are very similar , our backgrounds are small and cross - feed among the final states is negligible many of the sources of uncertainty cancel in the ratio .
a similar analysis has recently been published by focus comparing the @xmath188 and @xmath184 final states@xcite .
the left - most plots in figures [ fig : kpimass ] , [ fig : kkmass ] , and [ fig : pipimass ] show the candidate @xmath0 masses for the three samples selected as described in section [ genmethod ] . selecting events in a narrow region around the @xmath176 mass
, we fit their proper time distributions with an unbinned maximum likelihood fit . the proper time is measured as described in section [ genmethod ] .
the signal likelihood function is an exponential convolved with three resolution gaussians .
the width of the primary gaussian is due to the propagation of errors from the track fit to the flight distance and momentum for each candidate .
the second and third gaussians represent candidates that have been mismeasured by the addition of spurious tracking hits or due to hard non - gaussian scatters in the material of the detector .
the first of these has its width determined in the fit to the copious @xmath188 sample while the second is fixed to a large value , 8 ps .
the relative contribution of the two mismeasured signal resolutions is determined using the @xmath188 data sample and fixed for the @xmath184 and @xmath185 samples . according to our simulation ,
the fraction of the second gaussian is about 4% of the well measured signal and the third gaussian is less than 0.1% of the signal .
the probability for a candidate to be signal is determined by its measured mass , and is based on the fit to the mass distributions .
background is considered to have contributions with both zero and non - zero lifetimes .
all parameters that describe the background are allowed to float in the fits except for the width of the widest gaussian which is fixed to 8 ps .
the right - most plots in figures [ fig : kpimass ] , [ fig : kkmass ] , and [ fig : pipimass ] show the lifetime fits to the three samples and the results are summarized in table [ tab : lifes ] .
.summary of the lifetime fits .
the parameters are those described in the text .
note that we have constrained the candidates to a @xmath0 mass of 1.86514 gev , the monte carlo corrected weighted average of the data .
the signal lifetime is highly dependent on the @xmath0 mass used in the constraint , while the lifetime difference is not .
this technique yields the smallest uncertainty in @xmath54 , but it not optimal for measuring the absolute @xmath0 lifetime and was not used in [ 34 ] . [
cols="^,^,^,^",options="header " , ] [ tab : lifes ] note that the background in all the samples has a large component that has a lifetime consistent with the @xmath0 lifetime .
this agrees with the prediction of our simulation that the background with lifetime is dominated by misreconstructed fragments of charm decays .
we calculate @xmath54 separately for the @xmath184 and @xmath185 samples .
systematic uncertainties are dominated by the statistical uncertainty in a monte carlo study used to determine small corrections , consistent with zero , that are applied to the measured result to account for differences between measured and generated values of the lifetimes ( @xmath192 ) .
additional significant systematic uncertainties come from variations in description of the background ( @xmath193 ) , uncertainties in our model of the proper time resolution ( @xmath194 ) , and details of the fit procedure ( @xmath194 ) , where the listed values are the contribution to the average result .
our preliminary results are @xmath195 and @xmath196 we form a weighted average of the two to get @xmath197 this result is consistent with zero .
it is also consistent with our previous result from wrong sign @xmath188 @xcite , the focus results from wrong sign @xmath188 @xcite , and the focus results from @xmath188 versus @xmath184 @xcite .
this result is shown along with some of the previous measurements in figure [ fig : newxy ] .
we present preliminary results of several analyses searching for the effects of cp violation and mixing in the decay of @xmath0 mesons .
we find no evidence of cp asymmetry in five different two - body decay modes of the @xmath0 to pairs of light pseudo - scalar mesons : @xmath198 , @xmath199 , @xmath3 , @xmath4 and @xmath5 .
we present the first measurement of the rate of wrong - sign @xmath6 decay : @xmath7 .
finally , we describe a measurement of the mixing parameter @xmath200 by searching for a lifetime difference between the cp neutral @xmath9 and the cp even @xmath10 and @xmath11 final states . under the assumption
that cp is conserved we find @xmath201 .
we gratefully acknowledge the effort of the cesr staff in providing us with excellent luminosity and running conditions .
m. selen thanks the pff program of the nsf and the research corporation , and a.h .
mahmood thanks the texas advanced research program .
this work was supported by the national science foundation , the u.s .
department of energy , and the natural sciences and engineering research council of canada .
nelson , hep - ex/9908021 .
ktev collaboration , a. alavi - harati _ et al .
lett . * 83 * , 22 ( 1999 ) .
na48 collaboration , v. fanti _
b * 465 * , 335 ( 1999 ) .
babar collaboration , b. aubert _ et al .
_ , a study of time- dependent _ cp_-asymmetries in @xmath202 and @xmath203 decays " , babar - conf-00/01 , slac - pub-8640 , hep - ex/0008048 .
belle collaboration , h. aihara , a measurement of _ cp _ violation in @xmath204 meson decays with belle " , to be published in the proceedings of the 30th international conference on high - energy physics ( ichep 2000 ) , osaka , japan , 27 jul-2 aug 2000 , hep - ex/0010008 .
cdf collaboration , f. abe _
et al . _ ,
81 * , 5513 ( 1998 ) ; t. affolder _ et al . _
( cdf collaboration ) , phys .
d * 61 * , 072005 ( 2000 ) .
_ , ags experiment proposal 926 ( 1996 ) .
kami collaboration , e. cheu _
_ , an expression of interest to detect and measure the direct _ cp _ violating decay @xmath205 and other rare decays at fermilab using the main injector " , 22 september 1997 , hep - ex/9709026 . proposal for an experiment to measure mixing , _ cp _ violation and rare decays in charm and beauty particle decays at the fermilab collider -btev " , 15 may 2000 , http://www-btev.fnal.gov/ public_documents / btev_proposal/. lhcb collaboration , technical proposal a large hadron collider beauty experiment for precision measurements of _ cp _ violation and rare decays " , printed at cern , geneva , switzerland , isbn 92 - 9083 - 123 - 5
. f. bucella _
et al . _ ,
phys . rev .
d * 51 * , 3478 ( 1995 ) .
bigi , flavor dynamics - central mysteries of the standard model " , to be published in the proceedings of the 30th international conference on high - energy physics ( ichep 2000 ) , osaka , japan , 27 jul - 2 aug 2000 , hep - ex/0009021 .
lee , r. oehme , and c.n .
yang , phys . rev . * 106 * , 340 ( 1957 ) ; a. pais and s.b .
treiman , phys .
d * 12 * , 2744 ( 1975 ) .
our @xmath48 and @xmath54 are in terms of mixing amplitudes , while pais and treiman use the eigenvalues of the mixing hamiltonian ; both definitions agree in the limit of _ cp _ conservation . in our convention
, @xmath206 implies the real intermediate states with _
@xmath207 , such as @xmath106 , make the larger contribution to @xmath54 .
glashow , j. illiopolous , and l. maiani , phys .
1285 ( 1970 ) . m. leurer , y. nir , and n. seiberg , nucl .
* b 420 * , 468 ( 1994 ) ; n. arkani - hamed _ et al .
_ , hep - ph/9909326 .
cleo collaboration , r. godang _
et al . _ ,
* 84 * , 5038 ( 2000 ) .
@xmath208 , 95% cl @xmath209 and @xmath210 .
focus collaboration , j.m .
et al . _ , hep - ex/0012048 .
c. caso _ et al . _
( particle data group ) , eur .
j. * c 3 * , 1 ( 1998 ) .
s.b . treiman and r.g .
sachs , phys . rev . * 103 * , 1545 ( 1956 ) .
l. wolfenstein , phys . rev . lett . * 75 * , 2460 ( 1995 ) ; t.e .
browder and s. pakvasa , phys .
b * 383 * , 475 ( 1996 ) ; a.f .
falk , y. nir , and a.a .
petrov , j. high energy phys .
* 9912 * , 019 ( 1999 ) .
y. kubota _ et al .
methods phys .
a * 320 * , 66 ( 1992 ) .
. hill , nucl .
methods phys .
a * 418 * , 32 ( 1998 ) .
d. peterson , nucl .
* b * ( proc . suppl . )
* 54b * , 31 ( 1997 ) .
r. brun _ et al .
_ , geant3 users guide , cern dd / ee/84 - 1 .
charge conjugation is implied throughout , except where the charge conjugate states are explicitly shown , such as in an asymmetry definition .
this result is for the @xmath63 mode .
other modes have similar widths since the uncertainty on the slow pion dominates the width of the @xmath87 distribution .
first measurement of @xmath211 , cleo - conf 01 - 02 .
d. cinabro _ et al .
e * 57 * , 1193 ( 1998 ) .
cleo collaboration , j. bartelt _ et al .
d * 52 * , 4860 ( 1995 ) .
focus collaboration , j.m link _ et al .
b * 491 * , 232 ( 2000 ) ; erratum - ibid . *
495 * , 443 ( 200 ) ; e791 collaboration , e.m .
et al . _ ,
b * 421 * , 405 ( 1998 ) ; e687 collaboration , p.l . frabetti _
et al . _ ,
d * 50 * , 2953 ( 1994 ) ; e691 collaboration , j.c .
et al . _ ,
d * 44 * , 3371 ( 1991 ) .
cleo collaboration , g. bonvicini _
, to be published in phys .
, clns 00/1708 , hep - ex/0012054 .
cleo collaboration , s. kopp _ et al .
_ , hep - ex/0011065 , submitted to phys .
d , november 2000 .
m. gronau , j. l. rosner , hep - ph/0010237 ( 2000 ) , submitted to phys .
b. cleo collaboration , g. bonvicini _
et al . _ , phys .
lett . * 82 * , 4586 ( 1999 ) . focus collaboration , j.m .
et al . _ ,
phys . lett .
b * 485 * , 62 ( 2000 ) . | we present preliminary results of several analyses searching for the effects of _ cp _ violation and mixing in the decay of @xmath0 mesons .
we find no evidence of _ cp _ asymmetry in five different two - body decay modes of the @xmath0 to pairs of light pseudo - scalar mesons : @xmath1 , @xmath2 , @xmath3 , @xmath4 and @xmath5 .
we present the first measurement of the rate of wrong - sign @xmath6 decay : @xmath7 .
finally , we describe a measurement of the mixing parameter @xmath8 by searching for a lifetime difference between the _ cp _ neutral @xmath9 final state and the _ cp _ even @xmath10 and @xmath11 final states . under the assumption
that _ cp _ is conserved we find @xmath12 . 6.5 in 9.0 in -0.50 in 0.00 in 0.00 in d. cronin - hennessy,@xmath13 a.l .
lyon,@xmath13 e. h. thorndike,@xmath13 t. e. coan,@xmath14 v. fadeyev,@xmath14 y. s. gao,@xmath14 y. maravin,@xmath14 i. narsky,@xmath14 r. stroynowski,@xmath14 j. ye,@xmath14 t. wlodek,@xmath14 m. artuso,@xmath15 c. boulahouache,@xmath15 k. bukin,@xmath15 e. dambasuren,@xmath15 g. majumder,@xmath15 r. mountain,@xmath15 s. schuh,@xmath15 t. skwarnicki,@xmath15 s. stone,@xmath15 j.c .
wang,@xmath15 a. wolf,@xmath15 j. wu,@xmath15 s. kopp,@xmath16 m. kostin,@xmath16 a. h. mahmood,@xmath17 s. e. csorna,@xmath18 i. danko,@xmath18 k. w. mclean,@xmath18 z. xu,@xmath18 r. godang,@xmath19 g. bonvicini,@xmath20 d. cinabro,@xmath20 m. dubrovin,@xmath20 s. mcgee,@xmath20 g. j. zhou,@xmath20 a. bornheim,@xmath21 e. lipeles,@xmath21 s. p. pappas,@xmath21 a. shapiro,@xmath21 w. m. sun,@xmath21 a. j. weinstein,@xmath21 d. e. jaffe,@xmath22 r. mahapatra,@xmath22 g. masek,@xmath22 h. p. paar,@xmath22 d. m. asner,@xmath23 a. eppich,@xmath23 t. s. hill,@xmath23 r. j. morrison,@xmath23 r. a. briere,@xmath24 g. p. chen,@xmath24 t. ferguson,@xmath24 h. vogel,@xmath24 a. gritsan,@xmath25 j. p.
alexander,@xmath26 r. baker,@xmath26 c. bebek,@xmath26 b. e. berger,@xmath26 k. berkelman,@xmath26 f. blanc,@xmath26 v. boisvert,@xmath26 d. g. cassel,@xmath26 p. s. drell,@xmath26 j. e. duboscq,@xmath26 k. m. ecklund,@xmath26 r. ehrlich,@xmath26 p. gaidarev,@xmath26 l. gibbons,@xmath26 b. gittelman,@xmath26 s. w. gray,@xmath26 d. l. hartill,@xmath26 b. k. heltsley,@xmath26 p. i. hopman,@xmath26 l. hsu,@xmath26 c. d. jones,@xmath26 j. kandaswamy,@xmath26 d. l. kreinick,@xmath26 m. lohner,@xmath26 a. magerkurth,@xmath26 t. o. meyer,@xmath26 n. b. mistry,@xmath26 e.
nordberg,@xmath26 m. palmer,@xmath26 j. r. patterson,@xmath26 d. peterson,@xmath26 d. riley,@xmath26 a. romano,@xmath26 h. schwarthoff,@xmath26 j. g. thayer,@xmath26 d. urner,@xmath26 b. valant - spaight,@xmath26 g. viehhauser,@xmath26 a. warburton,@xmath26 p. avery,@xmath27 c. prescott,@xmath27 a. i. rubiera,@xmath27 h. stoeck,@xmath27 j. yelton,@xmath27 g. brandenburg,@xmath28 a. ershov,@xmath28 d. y .- j .
kim,@xmath28 r. wilson,@xmath28 t. bergfeld,@xmath29 b. i. eisenstein,@xmath29 j. ernst,@xmath29 g. e. gladding,@xmath29 g. d. gollin,@xmath29 r. m. hans,@xmath29 e. johnson,@xmath29 i. karliner,@xmath29 m. a. marsh,@xmath29 c. plager,@xmath29 c. sedlack,@xmath29 m. selen,@xmath29 j. j. thaler,@xmath29 j. williams,@xmath29 k. w. edwards,@xmath30 r. janicek,@xmath31 p. m. patel,@xmath31 a. j. sadoff,@xmath32 r. ammar,@xmath33 a. bean,@xmath33 d. besson,@xmath33 x. zhao,@xmath33 s. anderson,@xmath34 v. v. frolov,@xmath34 y. kubota,@xmath34 s. j. lee,@xmath34 j. j. oneill,@xmath34 r. poling,@xmath34 a. smith,@xmath34 c. j. stepaniak,@xmath34 j. urheim,@xmath34 s. ahmed,@xmath35 m. s. alam,@xmath35 s. b. athar,@xmath35 l. jian,@xmath35 l. ling,@xmath35 m. saleem,@xmath35 s. timm,@xmath35 f. wappler,@xmath35 a. anastassov,@xmath36 e. eckhart,@xmath36 k. k. gan,@xmath36 c. gwon,@xmath36 t. hart,@xmath36 k. honscheid,@xmath36 d. hufnagel,@xmath36 h. kagan,@xmath36 r. kass,@xmath36 t. k. pedlar,@xmath36 j. b. thayer,@xmath36 e. von toerne,@xmath36 m. m. zoeller,@xmath36 s. j. richichi,@xmath37 h. severini,@xmath37 p. skubic,@xmath37
a. undrus,@xmath37 v. savinov,@xmath38 s. chen,@xmath39 j. fast,@xmath39 j. w. hinson,@xmath39 j. lee,@xmath39 d. h. miller,@xmath39 e. i. shibata,@xmath39 i. p.
j. shipsey,@xmath39 and v. pavlunin@xmath39 @xmath13university of rochester , rochester , new york 14627 + @xmath14southern methodist university , dallas , texas 75275 + @xmath15syracuse university , syracuse , new york 13244 + @xmath16university of texas , austin , texas 78712 + @xmath17university of texas - pan american , edinburg , texas 78539 + @xmath18vanderbilt university , nashville , tennessee 37235 + @xmath19virginia polytechnic institute and state university , blacksburg , virginia 24061 + @xmath20wayne state university , detroit , michigan 48202 + @xmath21california institute of technology , pasadena , california 91125 + @xmath22university of california , san diego , la jolla , california 92093 + @xmath23university of california , santa barbara , california 93106 + @xmath24carnegie mellon university , pittsburgh , pennsylvania 15213 + @xmath25university of colorado , boulder , colorado 80309 - 0390 + @xmath26cornell university , ithaca , new york 14853 + @xmath27university of florida , gainesville , florida 32611 + @xmath28harvard university , cambridge , massachusetts 02138 + @xmath29university of illinois , urbana - champaign , illinois 61801 + @xmath30carleton university , ottawa , ontario , canada k1s
5b6 + and the institute of particle physics , canada + @xmath31mcgill university , montral , qubec , canada h3a 2t8 + and the institute of particle physics , canada + @xmath32ithaca college , ithaca , new york 14850 + @xmath33university of kansas , lawrence , kansas 66045 + @xmath34university of minnesota , minneapolis , minnesota 55455 + @xmath35state university of new york at albany , albany , new york 12222 + @xmath36ohio state university , columbus , ohio 43210 + @xmath37university of oklahoma , norman , oklahoma 73019 + @xmath38university of pittsburgh , pittsburgh , pennsylvania 15260 + @xmath39purdue university , west lafayette , indiana 47907 |
over the last two decades , the long lasting idea of globular clusters hosting single , simple stellar populations has changed dramatically . the classical paradigm of gcs being an excellent example of
a simple stellar population , defined as a coeval and initially chemically homogeneous assembly of stars , has been challenged by observational evidence .
the presence of chemical inhomogeneities , in most cases limited to the light elements ( the chemical pairs c - n , o - na , and mg - al anti - correlated with each other ) , have been known for decades and recognized to be the signature of high - temperature h - burning .
this was initially framed within a stellar evolutionary scenario ( see , e.g. , * ? ? ?
* and references therein ) given that gc abundance work based on high - quality data was limited to bright , evolved giants .
it was only at the turn of the century that the availability of high - resolution spectrographs mounted on 8 m class telescopes made it possible to carry out studies on the compositions of stars down to the main - sequence , which revealed light elements variations analogous to those found among giants @xcite . given that the atmospheres of warm main - sequence stars in large part retain the composition of the gas from which they were formed , the unavoidable conclusion was that the abundance inhomogeneities are of primordial origin .
the most extensive spectroscopic survey of gcs undertaken so far @xcite revealed that these inhomogeneities are ubiquitous in galactic gcs , though they do not appear to occur in other star formation environments .
however , the extent of the inhomogeneity varies from cluster to cluster , and appears to correlate strongly with the present - day total mass of the gcs , and also with metallicity .
the improvement in available instrumentation and techniques has also led to the discovery of a much higher degree of complexity of gc color - magnitude diagrams .
in fact , while some clusters seem to photometrically comply with the simple stellar population paradigm , a growing number of them are found to be characterized by multiple main sequences and/or subgiant and/or giant branches , ( * ? ? ?
* e.g. , ) , which have been associated to variations in the content of he and cno , as well as age spread @xcite .
this observational evidence led to a general scenario where gcs host multiple stellar populations .
these are often assumed to be associated with different stellar generations : the ejecta of the slightly older stars , probably mixed with varying amounts of gas from the original star forming cloud , creates subsequent younger generation of stars ( see e.g. , * ? ? ?
* ; * ? ? ?
* ) , though alternative scenarios are also being considered ( see , e.g. , * ? ? ?
it is believed that only a fraction of the first - generation of stars can contribute to the internal enrichment .
the difference in ages among the stellar generations is actually relatively small for the majority of the clusters ( with a few exceptions such as e.g. , @xmath2 cen or m22 ) and is confined to a couple of hundreds myr .
the details of this formation scenario are still far from being understood .
the origin of the polluting material remains to be established and it has obvious bearings on the timescales for the formation of the cluster itself and its mass budget . the observed wide star - to - star variations in c , n , o , na , and al found in each galactic gcs , coupled with the uniformity in fe and ca ( apart from a few notable exceptions ) provide quite stringent constraints and argue against anything but a minor contribution from supernovae @xcite .
proposed candidate polluters include intermediate mass stars in their asymptotic giant branch ( agb ) phase @xcite , fast rotating massive stars losing mass during their main sequence phase @xcite , novae @xcite and massive binaries @xcite .
these potential contributions obviously operate on different time scales and require a different amount of stellar mass in the first - generation .
all of the candidates proposed so far fall short of reproducing the full variety of observations .
advances in the theoretical modeling of star formation and evolution are likely needed to improve our understanding of these issues , including the spanning of a larger range of the parameter space ( e.g. , mass , metallicity , and mass loss ) . however , from an observational point of view , the increase in the high - quality abundance work available for gcs , both in the sheer number of stars and clusters , as well as in terms of chemical species considered , is paramount , as it creates a more complete picture of the phenomena involved . the apache point
observatory galactic evolution experiment ( apogee * ? ? ?
* ) is a three year , near - infrared ( 15,090 to 16,990 ; * ? ? ?
* ) , high - resolution spectroscopic survey of about 100,000 red giant stars included as part of the 3rd sloan digital sky survey ( sdss - iii * ? ? ?
* ) . with a nominal resolving power of 22,500
, apogee is deriving abundances of up to 15 elements for nearly 100,000 stars , although fewer elements are generally detected in weak - lined metal - poor stars .
apogee is in a unique position among the various galactic spectroscopic surveys such as gaia - eso , @xcite , rave @xcite , and galah @xcite , as it uses the sloan 2.5 m telescope at apache point observatory @xcite , and thereby has access to the northern hemisphere .
apogee observes a large sample of northern globular clusters , something that makes it possible to analyse these clusters in a homogeneous way , which has not been done before for these objects .
the study of gcs with apogee plays an important role not just because it has access to many northern gcs .
its high - resolution near - ir spectra allow the simultaneous determination of many elemental abundances generally not available in optical spectroscopic work of gc stars .
c and n , which are elements heavily affected by the pollution phenomenon in gcs , are often not included in studies of metal poor stars because the strongest features ( ch and cn ) lie in the near - uv , far from the optical lines of na , mg and al , and thus multiple detectors or setups are required to obtain both sets .
in addition , because these studies usually focus on fairly red stars longer exposure times are required to acquire sufficient snr to analyse the near - uv features .
the spectra used in this paper are publicly available as part of the tenth data release ( dr10 , * ? ? ?
* ) of sdss - iii .
the initial set of stars selected were the same used by @xcite to check the accuracy and precision of apogee parameters published in dr10 .
however , instead of using the automatic aspcap pipeline , we will make use of photometry and theoretical isochrones to constrain the effective temperature ( t@xmath3 ) and surface gravity @xmath4 , and use an independent semi - automated method for elemental abundance determination for 10 northern gcs .
some of these clusters are well studied , such as m3 , m13 , m92 , m15 , while others have been poorly studied ( ngc 5466 ) , or been only recently discussed in the literature , such as m2 @xcite .
table 1 lists the globular clusters apogee observed in its first year , along with the adopted [ fe / h ] , e(@xmath5 ) , and ages from the literature .
targets were selected as cluster members if 1 ) there is published abundance information on the star as a cluster member , 2 ) the star is a radial velocity member , or 3 ) if it has a probability @xmath650% of being a cluster member based on their proper motion adopted from the literature . after this initial selection , we checked the position of stars in the t@xmath7 diagram based on apogee observations and deleted those that were not red giant branch ( rgb ) stars .
stars that have metallicity 0.3 dex ( typically 3 @xmath8 scatter ) larger or smaller than the cluster average also need to be deleted , but this last step resulted in no rejections .
the cluster target selection process is described in more detail in @xcite .
the final sample consists of 428 stars from 10 globular clusters .
high s / n spectra are essential to determine abundances from atomic and molecular features , thus all selected targets have at least s / n=70 as determined by @xcite .
llrrc ngc 7078 & m15 & 23 & [email protected] & 0.10 + ngc 6341 & m92 & 47 & [email protected] & 0.02 + ngc 5024 & m53 & 16 & [email protected] & 0.02 + ngc 5466 & & 8 & [email protected] & 0.00 + ngc 6205 & m13 & 81 & [email protected] & 0.02 + ngc 7089 & m2 & 18 & [email protected] & 0.06 + ngc 5272 & m3 & 59 & [email protected] & 0.01 + ngc 5904 & m5 & 122 & [email protected] & 0.03 + ngc 6171 & m107 & 42 & [email protected] & 0.33 + ngc 6838 & m71 & 12 & [email protected] & 0.25 + lrrrrrrrrrrrrr 2m21301565 + 1208229 & m15 & -104.5 & 4836 & 1.56 & -2.12 & & & & 0.16 & -0.06 & 0.35 & 0.19 & + 2m21301606 + 1213342 & m15 & -108.3 & 4870 & 1.64 & -2.31 & & & & 0.10 & 0.57 & 0.46 & 0.53 & + 2m21304412 + 1211226 & m15 & -102.7 & 4715 & 1.28 & -2.12 & & & & -0.45 & 0.63 & 0.60 & 0.35 & + 2m21290843 + 1209118 & m15 & -106.0 & 4607 & 1.03 & -2.07 & & & & -0.11 & 0.75 & 0.41 & & + 2m21294979 + 1211058 & m15 & -107.6 & 4375 & 0.56 & -2.31 & -0.44 & 0.95 & 0.44 & 0.17 & 0.64 & 0.44 & 0.06 & + abundances presented in this paper are defined for each individual element x heavier than helium as @xmath10 = \log_{10}(n_x / n_h)_{\rm star } - \log_{10}(n_x / n_h)_\odot\ ] ] where @xmath11 and @xmath12 are respectively the number of atoms of element x and hydrogen , per unit volume in the stellar photosphere .
to derive abundances from stellar spectra , we first have to estimate four main atmospheric parameters : t@xmath3 , @xmath4 , microturbulent velocity , and overall metallicity ( [ fe / h ] ) . in the following sub - sections we present our methodology for determining these parameters and our reasons for not using the values available for each star from the apogee stellar parameters and chemical abundances pipeline ( aspcap ; * ? ? ?
* ) in dr10 . in order to evaluate the accuracy and precision of the aspcap parameters ,
@xcite carried out a careful comparison with literature values using 559 stars in 20 open and globular clusters .
these clusters were chosen to cover most of the parameter range of stars apogee is expected to observe .
@xcite provided a detailed explanation of the accuracy and precision of these parameters , and also derived empirical calibrations for t@xmath3 , @xmath4 , and [ m / h ] using literature data . in the sections below we
will briefly review these calibrations along with their limitations .
we adopt a photometric effective temperatures calculated from the @xmath13 colors using the equations of @xcite .
their calibration was chosen because of its proximity of only 30@xmath1440 k to the absolute temperature scale .
de - reddened @xmath13 were calculated the same way as by @xcite , from @xmath15 , listed in table 1 for each cluster , using @xmath16 .
the aspcap dr10 effective temperatures were compared with photometric ones using calibrations by @xcite based on 2mass @xmath13 colors @xcite . @xcite found that small systematic differences , in the range of 100@xmath14200k , are present between the aspcap and photometric temperatures .
the aspcap dr10 t@xmath3 were also compared to literature spectroscopic temperatures , and the average of these differences were found to be negligible .
the corrected aspcap dr10 temperatures were calculated between 3500 k and 5500 k using a calibration relation derived from the comparison with the @xcite scale .
aspcap dr10 raw temperatures above 5000 k in metal - poor stars showed significant , 300@xmath14500 k offsets compared to photometry , and are thus believed to be not accurate enough for abundance analysis .
this issue is mostly limited to metal poor stars , and does not affect stars at [ fe / h]@xmath171 , where the vast majority of apogee targets are .
the adoption of a purely photometric temperature scale enables us to be somewhat independent of aspcap ( while still using the same spectra ) , which gives important comparison data for future aspcap validation . besides providing an independent comparison dataset for apogee , the photometric temperatures allowed us to include stars in the sample that are hotter than 5000 k. because of these reasons , the final results presented in this paper are based on the photometric temperatures , and we only use the aspcap dr10 raw temperatures to estimate our errors related to the atmospheric parameters . the apogee dr10 release contains metallicities derived by aspcap for all stars analysed , thus providing an alternative scale to manually derived metallicities .
that metallicity , [ m / h ] , tracks all metals relative to the sun , and gives the overall metallicity of the stars because it was derived by fitting the entire wavelength region covered by the apogee spectrograph .
this is different from most literature publications that use fe lines to track metallicity in a stellar atmosphere .
we use [ fe / h ] in this paper whenever we refer to metallicity presented here , because we use fe i lines to measure it . for the most part
, one can treat values of [ m / h ] as if they were [ fe / h ] .
when aspcap metallicity was compared with individual values from high - resolution observations from the literature , a difference of 0.1 dex is found below [ m / h]=@xmath141 , and this discrepancy increased with decreasing metallicity reaching [email protected] dex around [ m / h]=@xmath142 @xcite .
the calibrated dr10 metallicities map well onto [ fe / h ] , because the calibration process uses [ fe / h ] values from the optical . because an alternative metallicity based on only iron lines did not exist for apogee in dr10 , and because of these small zero point offsets , we decided to derive our own [ fe / h ] based on fe i lines found in the h band , instead of using the published apogee dr10 [ m / h ] values .
we assumed that all stars have the literature cluster mean metallicity before starting the calculations , after which small wavelength windows around the fe lines listed in table 3 were used to revise the individual star metallicities . in this study
we adopt gravities from stellar evolution calculations .
following @xcite , we derive gravities for our sample using isochrones from the padova group @xcite .
the cluster metallicities collected from the literature used in the isochrones are listed in table 1 , while the ages of all clusters were chosen to be 10 gyr .
the final set of temperatures and gravities corresponding to them from the isochrones are listed in table 2 .
the aspcap dr10 surface gravities were compared to both isochrones and kepler @xcite asteroseismic targets observed by apogee @xcite to estimate their accuracy and precision .
an average difference of 0.3 dex was found at solar metallicity in both cases , but this increased to almost 1 dex for very metal poor stars .
the asteroseismic gravities are believed to have errors in the range of [email protected] dex , thus far superior to spectroscopic measurements , but they are only available for metal rich stars with [ m / h]@xmath181.0 . the discrepancy between spectroscopic and asteroseismic surface gravities is a topic of ongoing investigation ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . the final calibration from @xcite combined surface gravities derived from isochrones below [ m / h]@xmath191.0 with the asteroseismic dataset for high metallicities .
values after the calibration still show some small ( @xmath200.1 dex ) offsets and large scatter around the isochrones for the globular clusters .
thus , in order to minimize errors in abundances related to the uncertainty in the aspcap corrected spectroscopic gravities , we decided to use the pure isochrone gravities in this study .
a relation between microturbulent velocity and surface gravity , @xmath21 , was used for aspcap analyses in dr10 @xcite . for consistency with aspcap , and for easier future comparisons with apogee results
, we adopted that equation in this work .
the program called _ autosynth _ was developed especially for this project to simplify the large amount of synthesis required for abundance determination .
the program takes atmospheric parameters as input and carries out spectral synthesis to derive elemental abundances .
the core of _ autosynth _ is moog2013 @xcite , which does the spectrum synthesis , while _ autosynth _ compares the synthetic spectrum with the observed spectrum and determines the best abundances with @xmath22 minimization in wavelength windows specified by the user .
the program can read the moog formatted atlas and marcs model atmospheres , and can also convert the original atlas and marcs formats into a moog compatible format . the line list adopted for this study includes both atomic and molecular species .
it is an updated version of what was used for the dr10 results , version m201312160900 ( also used for dr12 , apogee s next public data release ; * ? ? ?
* ; * ? ? ?
* ) , and includes atomic and molecular species .
the molecular line list is a compilation of literature sources including transitions of co , oh , cn , c@xmath23 , h@xmath23 , and sih .
all molecular data are adopted without change with the exception of a few obvious typographical corrections .
the atomic line list was compiled from a number of literature sources and includes theoretical , astrophysical and laboratory oscillator strength values .
these literature line positions , oscillator strengths , and damping values were allowed to vary in order to fit to the solar spectrum and the spectrum of arcturus , thus generating a tuned astrophysical line list .
the solution is weighted such that the solar solution has twice the weight as the arcturus solution to properly consider the fact that the abundance ratios in arcturus are more poorly understood than those of the sun .
the code used for this process was based on the lte spectral synthesis code moog @xcite but adapted to our unique needs . for lines with laboratory oscillator strengths , we did not allow the astrophysical gf value to vary beyond twice the error quoted by the source . a more detailed description of this process and the line list can be found in @xcite .
the choice for the local continuum set can greatly affect the derived abundances , thus we needed a reliable automated way to determine the continuum placement .
this was done with a separate @xmath22 minimization from the one that was used for the abundance determination .
continuum normalized observation points around 1 are multiplied by a factor between 0.7 and 1.1 with 0.001 steps for each synthesis emulating slightly different choices for the location of the local continuum .
multiplication is necessary because it preserves the original spectrum .
the @xmath22 near the continuum is calculated and compared to the continuum of the observation and minimized separately for every abundance step .
the @xmath22 calculation for the abundances determination happens between certain flux ranges ( usually between 0.3 and 1.1 ) using the continuum placement determined in the previous step . the individual abundances were determined using the 1d local thermodynamic equilibrium ( lte ) model atmospheres calculated with atlas9 @xcite .
the model atmospheres were generated using solar reference abundances from @xcite , the same way as the main apogee model atmosphere database was generated @xcite . because the overall metallicity of these clusters were well known from the literature , initially we calculated atmospheres using the average literature metallicity for each cluster adopting the photometric effective temperatures and isochrone gravities .
these initial model atmospheres were later revised to have consistency with the synthesis .
the windows used to derive the individual abundances were determined based on the analysis of fts stars in the h@xmath14band using the apogee line list by @xcite . in the case of fe we measured [ fe / m ] , relative to the literature cluster metallicity for each line .
the abundance of na is very important in discussing the spread of o in gcs , and two na lines are available in the apogee spectral band .
however , these two na lines are weak even at solar metallicities .
we carried out a number of tests attempting to derive na abundances , but we found that the two na lines become very weak around [ fe / h][email protected] , and non - detectable below about @xmath24 , thus we were not able to determine na abundances for any of our targets .
the list of wavelength regions used in our analysis , and the solar reference values for each element is listed in table 3 .
figures [ fig : spectra1 ] and [ fig : spectra2 ] show examples of observed fe , mg , al , oh , co , and cn line profiles and their fitted synthesis for one star from m71 and m13 . the wavelength regions shown in these figures are only a fraction of what has been used from table 3 .
cn lines spread over most of the h@xmath14band , hence it is important to calculate the cno abundances before the atomic ones .
it is also important to use self consistent model atmospheres , because stars in globular clusters exhibit low carbon and high @xmath1content , which significantly alters the structure of the atmosphere compared to a solar scaled one @xcite .
taking into account all this , we developed the following procedure to produce the final abundances for each star : lrc fe & 7.45 & 15210 - 15213.5 ; 15397 - 15401 ; 15651 - 15654 + & & 15966 - 15973 ; 16044 - 16048 ; 16156 - 16160 + & & 16168 - 16171 + c & 8.39 & 15572 - 15606 ; 15772 - 15791 ; 15980 - 16037 + & & 16172 - 16248 ; 16617 - 16677 ; 16839 - 16870 +
n & 7.78 & 15240 - 15417 + o & 8.66 & 15267 - 15272 ; 15281 - 15288 ; 15372 - 15380 + & & 15386 - 15390 ; 15394 - 15397 ; 15404 - 15414 + & & 15499 - 15502 ; 15508 - 15511 ; 15539 - 15542 + & & 15561 - 15566 ; 15569 - 15574 ; 15887 - 15904 + & & 16188 - 16198 ; 16207 - 16213 ; 16233 - 16237 + & & 16244 - 16247 ; 16251 - 16261 ; 16300
- 16305 + & & 16314 - 16319 ; 16707 - 16714 ; 16718 - 16720 + & & 16731 - 16735 ; 16888 - 16892 ; 16898 - 16912 + mg & 7.53 & 15741 - 15757 ; 15767 - 15773 + al & 6.37 & 16720 - 16727 ; 16751 - 16759 ; 16765 - 16770 + si & 7.51 & 15962 - 15966 ; 16062 - 16066 ; 16097 - 16101 + & & 16218 - 16223 ; 16683 - 16687 ; 16830 - 16834 + ca & 6.31 & 16139 - 16143 ; 16153.5 - 16164 + ti & 4.90 & 15546.5 - 15549.5 ; 15718 - 15721.5 + 1 . a model atmosphere is generated using literature cluster average metallicities , the photometric temperature and an isochrone gravity . because all of our targets are rgb stars , we choose [ c / fe ] @xmath25 , [ o / fe ] @xmath26 , and [ n / fe ] @xmath27 dex for this initial model .
individual fe i lines are fit with _ autosynth _ , and an average [ fe / h ] is calculated for each star .
3 . a new model atmosphere is calculated using this new [ fe / h ] value , but still using the starting cno abundances .
4 . we set the abundances of c , n and o before the remaining elements , because they can have a significant effect on the atmospheric structure in cool stars .
since molecular features generally disappear from metal - poor spectra above 4500 k , we divide our stars into two temperature groups . for the stars cooler than 4500 k ,
we first determine [ o / fe ] using oh lines , then create a new model atmosphere with [ @xmath28/fe ] equal to [ o / fe ] .
we then determine c and o abundances from co lines , then recreate the model atmosphere again with these new [ c / fe ] and [ o / fe ] abundances .
finally , we derive n abundance using cn lines . for stars hotter than 4500 k
, we leave the c , n , and o abundances at their inital values .
the abundances of the remaining elements ( mg , al , si , ca and ti ) are determined with _ autosynth _ , using the stellar parameters , metallicities , and c , n and o abundances previously determined . for each element
, we average together the abundance results from the different wavelength regions to obtain final values .
although the size of each region is different , we did not find it necessary to use weights based on their ranges or line strengths , because that approach did not produce abundances significantly different from a straightforward average .
data reduction errors or missing data affected some of these regions , resulting in erroneous fits , and because of this we carefully examined each fit by eye .
these wavelength regions were not included when constructing the final average abundances .
the final abundance values are listed in table 2 .
lccccccccc @xmath29t@xmath3 & 0.08 & 0.12 & 0.16 & 0.18 & 0.04 & 0.07 & 0.05 & 0.06 & 0.15 + = 150k & & & & & & & & & + @xmath29t@xmath3 & 0.05 & 0.08 & 0.11 & 0.12 & 0.03 & 0.05 & 0.03 & 0.04 & 0.10 + = 100k & & & & & & & & & + m15 & 0.11 & 0.21 & 0.30 & 0.06 & 0.08 & 0.16 & 0.13 & 0.19 & 0.26 + m92 & 0.10 & 0.18 & 0.30 & 0.05 & 0.07 & 0.12 & 0.12 & 0.22 & 0.17 + m53 & 0.06 & 0.16 & 0.12 & 0.05 & 0.06 & 0.14 & 0.11 & 0.11 & 0.18 + n5466 & 0.06 & 0.18 & 0.15 & 0.03 & 0.07 & 0.03 & 0.07 & 0.14 & 0.17 + m13 & 0.05 & 0.12 & 0.06 & 0.05 & 0.05 & 0.12 & 0.07 & 0.06 & 0.12 + m2 & 0.04 & 0.11 & 0.06 & 0.04 & 0.07 & 0.10 & 0.09 & 0.05 & 0.09 + m3 & 0.04 & 0.12 & 0.08 & 0.05 & 0.03 & 0.10 & 0.08 & 0.05 & 0.12 + m5 & 0.04 & 0.08 & 0.05 & 0.05 & 0.04 & 0.06 & 0.07 & 0.04 & 0.09 + m107 & 0.04 & 0.04 & 0.03 & 0.05 & 0.04 & 0.08 & 0.07 & 0.04 & 0.10 + m71 & 0.02 & 0.03 & 0.04 & 0.04 & 0.04 & 0.05 & 0.07 & 0.06 & 0.07 + m15 & 0.12 & 0.22 & 0.32 & 0.13 & 0.09 & 0.17 & 0.13 & 0.19 & 0.28 + m92 & 0.11 & 0.20 & 0.32 & 0.13 & 0.09 & 0.13 & 0.12 & 0.22 & 0.20 + m53 & 0.08 & 0.18 & 0.16 & 0.13 & 0.07 & 0.15 & 0.11 & 0.12 & 0.21 + n5466 & 0.08 & 0.20 & 0.19 & 0.12 & 0.08 & 0.06 & 0.08 & 0.15 & 0.20 + m13 & 0.07 & 0.14 & 0.13 & 0.13 & 0.06 & 0.13 & 0.08 & 0.07 & 0.16 + m2 & 0.06 & 0.14 & 0.13 & 0.13 & 0.08 & 0.11 & 0.09 & 0.06 & 0.13 + m3 & 0.06 & 0.14 & 0.14 & 0.13 & 0.04 & 0.11 & 0.09 & 0.06 & 0.16 + m5 & 0.06 & 0.11 & 0.12 & 0.13 & 0.05 & 0.08 & 0.08 & 0.06 & 0.13 + m107 & 0.06 & 0.09 & 0.11 & 0.13 & 0.05 & 0.09 & 0.08 & 0.06 & 0.14 + m71 & 0.05 & 0.09 & 0.12 & 0.13 & 0.05 & 0.07 & 0.08 & 0.07 & 0.12 + the uncertainty in the atmospheric parameters strongly affects the final abundances derived from some of the spectral features we consider . to test the sensitivity of abundances due to changes in the atmospheric parameters we used the results from the aspcap raw temperature scale .
the same exact steps described in the previous section were followed , but instead of adopting the photometric temperature scale we adopt the aspcap dr10 raw temperature scale , which results in new surface gravities and microturbulent velocities .
this way we could track systematics uncertainties sensitive to these parameters as well .
errors give the standard deviation around the mean of the differences .
, width=336 ] lrrrrrrrrrrrrll 2m21290843 + 1209118 & m15 & & & -2.37 & & & & & & 0.36 & & 0.31 & b5 & h + 2m21293353 + 1204552 & m15 & & & -2.37 & & & & & & & 0.13 & & k22 & h + 2m21293871 + 1211530 & m15 & & & -2.37 & & & & & & & 0.14 & & k47 & h + 2m21294979 + 1211058 & m15 & & & -2.40 & & & & & & 0.53 & 0.28 & & k144 & h + 2m21294979 + 1211058 & m15 & & & -2.27 & & & 0.42 & & & & 0.38 & & k144 & i + the differences in abundances as a function of photometric temperatures are demonstrated in figure [ fig : errors ] .
the top left panel displays the differences in the measured abundances by using aspcap and photometric temperature , while the rest of the panels are assigned to each element .
the color scale in all panels represents @xmath29 t@xmath3 .
we defined the estimated errors associated with the atmospheric parameters based on the standard deviation around the mean differences between the two temperature scales .
the calculated standard deviation in of the differences in temperatures is 146 k ( which we round to 150 k ) .
this standard deviation corresponds to the sum of the uncertainty in the photometric temperature and the aspcap temperature in quadrature . in order to estimate the uncertainty of just the photometric temperature component
, we carried out calculations of temperatures with varied j@xmath14ks colors , reddenings and metallicities for m107 .
m107 was chosen because it is the cluster with the highest reddening in our sample .
the uncertainty of 2mass photometry is usually between [email protected] magnitude for these stars , and by using 0.03 as baseline , we estimate an uncertainty of 0.05 magnitude for the j@xmath14ks color . changing j@xmath14ks by 0.05 magnitudes typically produces a change of 80 k in the photometric temperature .
the reddening of m107 was changed by 0.03 , simulating a 10% uncertainty in reddening , and this produced a difference of about 40 k in the photometric temperature . a change of 0.1 dex in metallicity results in 1 k or less uncertainty in temperature
; thus the uncertainties in metallicity can be neglected . by adding the uncertainty from photometry and reddening in quadrature
, we estimate the uncertainty of the photometric temperature to be about 100 k. the top panel of table 4 lists uncertainties associated with both 150 and 100 k changes in temperature . in our methodology
a 100 k change in temperature also introduces an 0.3 dex systematic difference in surface gravity , and 0.1 km / s in microturbulent velocity , so by coupling these two parameters to t@xmath3 , our systematic uncertainties include the investigation of abundance sensitivity to these parameters as well .
besides the uncertainty coming from the adopted atmospheric parameters , the uncertainty of the fit was also calculated using the @xmath8 of the residuals between the best fit synthesis and the observation .
these calculations estimate random uncertainties .
this @xmath8 of the fit is calculated within the windows used in the @xmath22 calculation .
for determining the uncertainty of the fit , we multiply the observed spectra by 1+@xmath8 and 1-@xmath8 which simulates two spectra slightly different from the original spectrum .
then , the fit of each line is repeated while keeping all other parameters unchanged .
the average differences between these two new fits and the original best - fit spectrum is the defined uncertainty associated with the fit itself .
this uncertainty estimate is mainly sensitive to variations of noise in the spectrum in the defined windows for the @xmath22 fit .
if its value is close to 0 , then the @xmath30 spectrum and @xmath31 spectrum give very similar , or the same abundances .
this is expected when working at high s / n and with a well defined continuum . while this uncertainty estimation method is reliable in most cases , it has its limitations for very noisy spectra , very weak lines , or when abundances are near upper limits . uncertainties are usually overestimated for noisy spectra , while they are underestimated for weak lines and upper limits .
thus , we decided to use one uncertainty from _ autosynth _ per element per cluster by simply averaging together all uncertainties reported by the program .
this resulted in one uncertainty estimate for every element in each cluster .
_ autosynth _ uncertainties are listed in the middle section of table 4 .
the final estimated uncertainties were calculated by adding together in quadrature the uncertainties for 100 k difference in temperature ( also 0.3 dex systematic difference in @xmath4 , and 0.1 km / s in microturbulent velocity ) , and the average _ autosynth _ values that estimate random uncertainties .
these final estimations for each element per cluster are given in the bottom panel of table 4 .
we take [ x / fe ] abundance values for c , n , o , mg , al , si , ca and ti from high - resolution spectroscopic studies in the literature as a point of comparison for our abundance determinations . we use the same literature sources as @xcite , and added more recently published papers listed in table 5 .
the derivation of stellar parameters @xmath32 and @xmath4 are described in detail in section 2 ; and they were also compared to the literature .
our goal for these comparisons is not to cross - calibrate our new abundance determinations with the literature ; rather , we are looking for cases where our abundances are systematically different from the literature , or particular clusters or elements for which our homogeneously observed and analysed data set can clarify conflicts in the literature .
cross - identification between the globular cluster stars in the dr10 apogee data set and the literature was performed using the simbad online service , based on 2mass identifiers and ( @xmath28 , @xmath33 ) coordinates . because there is a large and heterogeneous literature on chemical abundances in globular cluster stars
, we are providing our cross - identifications as a resource for the community .
table 5 lists 2mass i d and position , and alternate stellar identifiers from literature abundance studies for all of the stars considered in this study .
lrrrrrrrrrr m15 & -2.28 & -0.41 & 0.95 & 0.54 & 7.09 & 0.11 & 0.34 & 0.44 & 0.16 & 0.15 + m92 & -2.23 & -0.41 & 0.93 & 0.58 & 7.19 & 0.14 & 0.42 & 0.45 & 0.10 & 0.09 + m53 & -1.95 & -0.50 & 1.06 & 0.56 & 7.49 & 0.11 & 0.37 & 0.41 & 0.23 & 0.28 + n5466 & -1.82 & -0.56 & 0.84 & 0.63 & 7.60 & 0.14 & -0.24 & 0.29 & 0.04 & 0.29 + m13 & -1.50 & -0.53 & 0.89 & 0.28 & 7.69 & 0.13 & 0.61 & 0.40 & 0.26 & 0.20 + m2 & -1.49 & -0.48 & 0.90 & 0.41 & 7.76 & 0.26 & 0.45 & 0.35 & 0.24 & 0.27 + m3 & -1.40 & -0.46 & 0.69 & 0.40 & 7.84 & 0.15 & 0.21 & 0.30 & 0.12 & 0.11 + m5 & -1.24 & -0.46 & 0.76 & 0.27 & 7.85 & 0.23 & 0.36 & 0.34 & 0.20 & 0.26 + m107 & -1.01 & -0.21 & 0.69 & 0.33 & 8.15 & 0.24 & 0.47 & 0.48 & 0.15 & 0.21 + m71 & -0.68 & -0.10 & 0.91 & 0.51 & 8.65 & 0.38 & 0.51 & 0.39 & 0.21 & 0.42 + m15 & 0.10 & 0.13 & 0.35 & 0.19 & 0.14 & 0.24 & 0.52 & 0.16 & 0.25 & 0.08 + m92 & 0.10 & 0.11 & 0.23 & 0.19 & 0.13 & 0.23 & 0.48 & 0.12 & 0.17 & 0.20 + m53 & 0.07 & 0.16 & 0.21 & 0.06 & 0.09 & 0.08 & 0.51 & 0.05 & 0.17 & 0.13 + n5466 & 0.08 & 0.01 & 0.10 & 0.04 & 0.08 & 0.06 & 0.35 & 0.09 & 0.25 & 0.14 + m13 & 0.07 & 0.07 & 0.17 & 0.17 & 0.15 & 0.15 & 0.53 & 0.09 & 0.14 & 0.14 + m2 & 0.08 & 0.05 & 0.15 & 0.21 & 0.23 & 0.07 & 0.51 & 0.06 & 0.12 & 0.08 + m3 & 0.08 & 0.08 & 0.13 & 0.23 & 0.15 & 0.06 & 0.43 & 0.04 & 0.08 & 0.13 + m5 & 0.08 & 0.13 & 0.17 & 0.27 & 0.19 & 0.09 & 0.34 & 0.07 & 0.11 & 0.12 + m107 & 0.06 & 0.09 & 0.27 & 0.15 & 0.15 & 0.10 & 0.15 & 0.05 & 0.21 & 0.16 + m71 & 0.07 & 0.13 & 0.32 & 0.09 & 0.13 & 0.05 & 0.08 & 0.05 & 0.10 & 0.09 + figure [ fig : litcomp1 ] shows comparisons of our stellar parameters and abundances against the literature values .
different globular clusters are represented by different colored points . in general ,
we find a systematic offset of @xmath34 k between our photometric @xmath35 values and the spectroscopic effective temperatures from the literature , with a reasonably small scatter .
this is similar to the typical difference between spectroscopic and photometric temperatures reported by @xcite . because of the degeneracies in deriving stellar parameters ,
the slightly higher temperatures in the literature are accompanied by a systematic offset of @xmath36 dex in @xmath4 .
there are a few systematic differences between our abundance results and those in the literature .
these can mainly be traced back to a change in the solar abundance scale as derived by @xcite .
as one example , the [ fe / h ] metallicities we derive are typically @xmath37 dex higher than those from the literature , which is quite similar in magnitude to the change in a(fe)@xmath38 from @xmath39 @xcite to @xmath40 @xcite . also , the solar abundance of oxygen was revised significantly , from a(o)@xmath38=8.93 @xcite to 8.66 @xcite .
since we use the more recent solar abundance values from @xcite whereas our earlier literature sources do not , we expect the @xmath41 dex offset visible in figure [ fig : litcomp1 ] . the systematic difference in [ n / fe ] abundance is also likely to be due to the updated solar abundances .
carbon is the only element studied in this paper for which this explanation does not hold : although the solar abundance was revised down , from a(c)@xmath38=8.56 @xcite to @xmath42 @xcite , our [ c / fe ] values are typically lower than the literature values .
unfortunately , literature values are available for only a subset of our stars , which makes it difficult to verify precisely any systematic behavior . in the other elements we consider in the present study ,
there is generally good agreement between our results and those from the literature , which is encouraging .
however , we find significant differences in al abundances in stars with below - solar metallicities .
this is mainly driven by stars in m3 , where the [ al / fe ] abundances from @xcite are larger than in the other literature sources .
we also see significant scatter in [ ca / fe ] and [ ti / fe ] at low abundances ; this is likely caused by our abundance determination method having difficulty with weak lines . the [ mg / fe ] versus [ al / fe ] relations in m3 , m13 and m5 provide useful examples of how our new data set compares with the literature . in m3 ,
our abundance determinations show a clear bimodality in the mg - al distribution .
the study of @xcite found a similar bimodality , but @xcite found a smooth distribution .
in m13 we find an extended anticorrelation between [ mg / fe ] and [ al / fe ] , and the literature abundances lie within it .
we have seven stars in common with @xcite , and they span the full anticorrelation , while the three and two stars , respectively , that we have in common with @xcite and @xcite are consistent with our abundance results but happen to inhabit small regions of abundance space .
m5 is a similar case , where our [ al / fe ] values span a range of @xmath140.3 dex to @xmath431.1 dex , while the @xcite al abundances are limited to between @xmath140.2 and @xmath430.7 dex . as in m13 ,
our sample is larger than the literature sample , and covers the full rgb , while the literature sample does not fully span the parameter space .
although there are well - known abundance patterns within globular clusters , large homogeneous studies that include a wide range of abundances are rare . since we can determine abundances of most of the light elements for the stars in our sample , we examine the behavior of the known abundance patterns across a wide range in cluster metallicity , and search for unexpected variations in the @xmath1elements . in this section
we focus on the range of abundance within each cluster , to separate real abundance variations , bimodalities and trends from possible measurement errors .
table 6 lists the average and standard deviation for each elemental abundance in each cluster . in figures [ fig : femgal]-[fig : sicati ] we show all nine derived abundances as a function of effective temperature . from figure
[ fig : femgal ] we conclude that we measure constant fe abundances in all of the clusters .
mg and al abundances show a large range of values in some clusters ( as discussed in section 2 , and further in section 6 ) , but no significant trends with t@xmath3 except in m107 and m71 .
this trend is very weak in m71 , and data are consistent with showing no trend within the uncertainties .
however , in m107 the trend is stronger .
we currently do not fully understand where these small correlations come from .
we suspect that this is a result of a combination of effects .
one such effect may be the use of model atmospheres that assume local thermodynamic equilibrium ( lte ) , but that non - lte effects @xcite act to make the strong mg lines in these metal - rich stars give the appearance of higher abundance in the cooler stars .
other possible effects are small systematic errors from estimating t@xmath3 and @xmath4 culminating during the synthesis .
abundances are also sensitive to the microturbulent velosity , so if the @xmath44v@xmath45 expression used is not so accurate in this metallicity range , or the surface gravity is badly estimated , this propagates into systematically off abundances through v@xmath45
. the minimum [ c / fe ] ( figure [ fig : cno ] ) that can be measured from the co lines strongly depends on temperature .
because rgb stars in gcs generally have low carbon abundances , we can only set upper limits for a number of our stars .
our [ c / fe ] values are an average of the derived abundances from five co windows . as a result determining
the upper limit is more challenging because co lines in certain windows disappear faster with rising temperature than in others .
we carefully checked every co window fit and selected upper limits if the co band head was not visible in more than three windows by looking at the flatness of the @xmath22 fit around the minimum value .
because the derived abundance of n from the cn lines is anticorrelated with the value of [ c / fe ] used , all stars with upper limits in [ c / fe ] have also lower limits in [ n / fe ] . these upper limits for [ c / fe ] and lower
limits of [ n / fe ] are identified with open triangles in all figures .
we see clear correlations between [ c / fe ] and t@xmath3 in m13 , m3 , m2 , and m5 when omitting upper limits .
we interpret these as a sign of `` deep mixing '' , a nonconvective mixing process that causes steady depletion of surface carbon abundance and an enhancement in nitrogen abundance in all low - mass rgb stars @xcite .
we derive only upper limits on [ c / fe ] for m15 , m92 , m53 and ngc 5466 because the co bands are quite weak at such low metallicity . with only upper limits ,
it is not possible to identify abundance trends in these clusters .
however , previous studies @xcite have found signs of the same deep mixing process in action in those clusters . the only cluster in which we see a sign of nitrogen enrichment with declining temperature is m13 .
possible trends in nitrogen are not necessarily a result of astrophysics , because in our methodology small systematic errors in temperature can result in systematic errors in n abundance because deriving [ n / fe ] is challenging from the cn lines .
other than the relatively large errors of n , one also has to be careful with possible correlations of c and n with temperature , because they can be generated on the rgb as part of deep mixing and also in a previous stellar evolution event that are responsible for the pollution . in figure
[ fig : sicati ] there is a weak trend , on par with the error , visible in the si abundance in m13 , but not in the other clusters .
while ca does not show any correlation with temperature , it does show temperature - dependent scatter in most of the gcs .
the three ca lines used in our analysis are generally weak , and get significantly weaker above 4700 k , which leads to higher errors related to increased sensitivity to the uncertainty in the continuum placement .
titanium , unexpectedly , shows a decline with decreasing temperature in m2 , m5 and ( marginally ) in m107 in figure [ fig : sicati ] .
we suspect that inaccuracies in our analysis of the hotter star spectra are driving this apparent trend , since the s / n is lower for those stars than for the cooler , brighter giants . because these trends in mg ( figure [ fig : femgal ] ) , si and ti ( figure [ fig : sicati ] ) only show up in a handful of clusters
, we believe that they are a result of difficulties in data analysis for certain lines in certain stars and not any systematic mishandling in , e.g. , @xmath46 estimation .
we choose to fit these various observed trends with a linear equation , regardless of their origin , in order to explore the scatter around the trend .
we fit lines to [ mg / fe ] in m107 and m71 , [ c / fe ] in m13 , m2 , m3 and m5 , and [ ti / fe ] in m2 and m5 .
we are only using these fits to determine the internal scatter , which we define as the standard deviation about the fitted line . throughout the rest of this paper
we use the abundance values directly , except when discussing the cn anticorrelations in section 6.4 .
an internal abundance scatter significantly larger than the estimated errors suggests an astrophysical origin .
in fact , significant scatter in n , o , mg and al is well documented in literature @xcite .
the large star - to - star variations in these elements are the result of the cno , ne - na , and mg - al cycles , which play an important role in the nuclear fusion processes at the early stages of rgb ( cno ) and agb ( ne - na , mg - al ) evolution .
figure [ fig : rms ] shows the internal scatter calculated as the standard deviation around the mean values ( in some cases around a fitted linear equation , as explained earlier ) , and the final combined estimated uncertainties from table 4 , as a function of cluster average metallicity .
we note that the small number of stars analysed in ngc 5466 limits our ability to make a detailed analysis of c , n , and o in this cluster : because we were only able to measure the abundance of these elements in the three stars that have temperatures below @xmath47k , the scatter is unrealistically higher than the calculated uncertainty .
the star - to - star scatter in [ fe / h ] is quite similar to our measurement uncertainties , indicating that no significant metallicity variations in the clusters are detected in this study .
recently , @xcite have discovered three distinctive groups with different iron abundances in m2 .
the first group has a dominant peak at [ fe / h]=-1.7 , while the second and third have smaller peaks at 1.5 and 1.0 , but the membership for the latter group is not conclusive .
we see no such behaviour in m2 .
the discrepancy between these two studies is most likely the result of a selection effect .
@xcite selected stars that belong to a second rgb @xcite within the cluster , while our sample selection were based upon previous observations @xcite and all of our stars belong to the main rgb of this cluster .
the @xmath1elements si and ti also show constant distribution , but the scatter in [ ca / fe ] is larger than our measurement uncertainties for most clusters in our sample .
we attribute this to an increasing inaccuracy in [ ca / fe ] with rising temperature , as discussed in section 5.1 , and not a real range in calcium abundance .
little to no range in iron and @xmath1elements is what we expect for normal globular clusters , although the clusters in which these abundances do vary are an intriguing puzzle ( see , e.g. , * ? ? ?
m107 stands out as having the largest discrepancy between the scatter of [ ca / fe ] and its estimated uncertainty .
however , the large internal scatter is not related to astrophysics , but to spectral contamination from telluric oh lines . due to the cluster s radial velocity ,
two of the three ca lines lie on top of atmospheric oh lines . because the telluric correction is not perfect ,
the ca lines are compromised in a way that is not included in our uncertainty estimation .
we do see scatter above the level of the uncertainties in some of the light - element abundances . in the case of carbon , we find that the scatter around the overall carbon depletion trend for m13 , m2 , m3 and m5 ( as shown by dashed lines in figure [ fig : cno ] ) is fairly small , although our measurement uncertainties get quite large at low metallicity .
we do not see clear carbon depletion in m107 or m71 ; however , as discussed in @xcite , the rate of carbon depletion due to deep mixing is lower in higher - metallicity stars . the range in [ n / fe ] abundance ( figure [ fig : rms ] , second panel from the top , in the left column ) in m71 , m107 , m5 , m3 , m13 , and m2 is clearly larger than the estimated uncertainties . however , the difference between the scatter and the uncertainty decreases with decreasing cluster metallicity , and below [ fe / h][email protected] ( ngc 5466 , m53 , m92 , and m15 ) our derived [ n / fe ] values are all lower limits .
the decreasing discrepancy is a result of larger uncertainties in more metal poor clusters , and the fact that the minimum n abundance of a cluster is slightly increasing as the average metallicity decreases ( figure [ fig : cno ] ) , while the maximum n is constant above [ m / h]=-1.7 .
the uncertainties in m15 and m92 are significantly larger than in other clusters , and our measurements are not precise enough to judge the amount of n enrichment in these clusters . fitting the cn lines in our defined cn window is also challenging .
this is because they are usually blended with other molecular lines , and because n abundance strongly depends on the derived c abundance , which is an upper limit in many cases . a more detailed analysis of the c - n anticorrelation and n bimodality can be found in section 6.4 .
lcrrrrrrrrrrr m15 & 1 & 10 & 0.43 & -2.31 & -0.49 & 0.78 & 0.64 & 0.20 & -0.19 & 0.31 & 0.08 & 0.11 + & 2 & 13 & 0.57 & -2.26 & -0.33 & 1.11 & 0.43 & 0.04 & 0.75 & 0.53 & 0.24 & 0.17 + m92 & 1 & 14 & 0.30 & -2.23 & -0.42 & 0.89 & 0.67 & 0.33 & -0.23 & 0.41 & 0.08 & 0.09 + & 2 & 33 & 0.70 & -2.23 & -0.38 & 1.03 & 0.40 & 0.05 & 0.70 & 0.47 & 0.11 & 0.09 + m53 & 1 & 8 & 0.50 & -1.96 & -0.43 & 0.91 & 0.59 & 0.38 & -0.11 & 0.42 & 0.23 & 0.24 + & 2 & 8 & 0.50 & -1.95 & -0.54 & 1.13 & 0.55 & 0.25 & 0.84 & 0.41 & 0.23 & 0.31 + n5466 & 1 & 6 & & -1.83 & -0.56 & 0.84 & 0.64 & 0.15 & -0.42 & 0.28 & 0.14 & 0.29 + & 2 & 2 & & -1.79 & -0.56 & 0.86 & 0.62 & 0.11 & 0.31 & 0.30 & -0.16 & + m13 & 1 & 32 & 0.40 & -1.50 & -0.44 & 0.83 & 0.56 & 0.21 & 0.04 & 0.38 & 0.24 & 0.18 + & 2 & 49 & 0.60 & -1.49 & -0.58 & 0.92 & 0.12 & 0.09 & 0.99 & 0.40 & 0.28 & 0.21 + m2 & 1 & 7 & 0.39 & -1.46 & -0.45 & 0.75 & 0.56 & 0.30 & -0.10 & 0.35 & 0.19 & 0.26 + & 2 & 11 & 0.61 & -1.50 & -0.49 & 1.01 & 0.30 & 0.24 & 0.79 & 0.35 & 0.27 & 0.27 + m3 & 1 & 39 & 0.66 & -1.40 & -0.42 & 0.59 & 0.53 & 0.18 & -0.08 & 0.30 & 0.13 & 0.11 + & 2 & 20 & 0.34 & -1.38 & -0.54 & 0.86 & 0.16 & 0.10 & 0.79 & 0.29 & 0.10 & 0.12 + m5 & 1 & 60 & 0.49 & -1.24 & -0.41 & 0.62 & 0.41 & 0.24 & 0.06 & 0.34 & 0.20 & 0.24 + & 2 & 62 & 0.51 & -1.24 & -0.50 & 0.87 & 0.16 & 0.21 & 0.64 & 0.35 & 0.20 & 0.28 + m107 & 1 & 6 & & -1.04 & -0.16 & 0.47 & 0.38 & 0.36 & 0.51 & 0.50 & 0.21 & 0.12 + & 2 & 6 & & -1.00 & -0.25 & 0.92 & 0.29 & 0.37 & 0.53 & 0.45 & 0.22 & 0.16 + m71 & 1 & 3 & & -0.72 & 0.00 & 0.57 & 0.53 & 0.46 & 0.44 & 0.41 & 0.24 & 0.38 + & 2 & 4 & & -0.69 & -0.17 & 1.16 & 0.49 & 0.44 & 0.53 & 0.38 & 0.21 & 0.45 + our uncertainties of the [ o / fe ] abundance ( figure [ fig : rms ] ) are fairly constant as a function of cluster metallicity .
this is because the error is dominated by a strong sensitivity to the temperature used in the model atmosphere , and the internal scatter varies between clusters .
unfortunately , the na lines available in the apogee spectra are too weak to investigate the o - na anticorrelation in our sample , so we can only limit our discussion on the scatter of [ o / fe ] .
we use the later as an indication of the strength of o - na anticorrelation in a cluster . in m71 and m107 the o scatter is comparable to the uncertainties , suggesting that the o - na anticorrelation is not as extended as in the rest of the clusters .
a clear spread is visible in m5 , m3 , m2 , m13 , m15 and m92 , while in m53 the o abundances are constant and the scatter is significantly smaller than what is expected from the uncertainties .
the effects of the mg - al cycle are very noticeable for all clusters except m71 and m107 ( figure [ fig : rms ] , second panels from the bottom ) .
the scatter in mg abundance in those clusters is close to the estimated uncertainties after taking the linear correlations with t@xmath3 into account ( figure [ fig : femgal ] ) . in other clusters
the scatter of [ mg / fe ] also closely follows the uncertainties , except for m15 and m92 where the most mg poor stars can be found .
we see no al enrichment in m107 and m71 ; however , as the average metallicity decreases , the amount of internal scatter rapidly starts to deviate from the error , and we see a large spread in al in all other gcs .
a more detailed analysis of the mg - al anticorrelations and al populations can be found in section 6.2 and 6.3 .
based on the examination of the spread of n and al abundances , we conclude that all clusters possess multiple stellar populations either based on n or al , or both .
we examine these populations in more detail in the next section .
the derived @xmath1element abundances are also fairly constant in all globular clusters presented here , similar to what is reported in the literature @xcite .
most light elements show star - to - star variations in all gcs .
these large variations are generally interpreted as the result of chemical feedback from an earlier generation of stars @xcite , rather than inhomogeneities in the original stellar cloud these stars formed from .
thus , the current scenario of gc evolution generally assumes that more than one population of stars were formed in each cluster .
the first generation of gc stars formed from gas that had been enriched by supernovae in the very early universe , while the second formed by combining gas from the original star - forming cloud with ejecta from the first - generation stars .
only the fraction of first - generation stars contribute to the pollution , and the time scale of the formation of these second generation stars depend on the nature of polluters ; it is a couple of hundred myr in the case of intermediate - mass agb stars , but it is only a few myr for fast rotating massive stars and massive binaries . these first - generation pollutors are thought to have been fast rotating massive stars @xcite , or intermediate mass ( m@xmath48 3 m@xmath38 ) agb stars @xcite , or massive binaries @xcite .
our data reveal the expected signatures of pollution from material enriched from the hot hydrogen burning cycles such as the cno , ne
na , or mg
al cycles in all globular clusters in our sample . in this section
we explore the various correlations between these elements , and we discuss individual star formation , and/or pollution events suggested by separate groups found in the c - n and mg - al anticorrelations . to separate the various populations under study , we follow an approach similar to that of @xcite , who used @xmath49-means clustering @xcite to identify multiple populations in @xmath2 cen . here
, we use the _ extreme - deconvolution _ ( xd ) method of @xcite to identify population groups and assign membership .
this method fits the distribution of a vector quantity , here the elemental abundances , as a sum of @xmath49 gaussian populations , whose amplitudes , centers and covariance matrices are left entirely free .
the algorithm can be applied to noisy or incomplete data , thus making the best possible use of all available data . in the current application ,
xd s main advantage is the latter , since we are not able to measure abundances for all nine elements in all stars in our sample . similar to the @xmath49-means ,
the number @xmath49 of populations to fit is an input to the algorithm ; xd itself does not determine the optimal number of components to fit the distribution .
briefly , the xd algorithm works by optimizing the likelihood of the gaussian mixture model of the data .
the optimization proceeds by an iterative procedure consisting of repeated _ expectation _ ( e ) and _ maximization _ ( m ) steps . in the e step , each datum is probabilistically assigned ( a ) membership in each population and ( b ) error - free values of each noisy or missing elemental abundance ; in the m step , each gaussian s parameters are updated using its members mean abundances and covariance calculated from the error - free values obtained in the e step .
these steps are repeated until the likelihood stops increasing to within a small tolerance .
the algorithm is proven to increase the likelihood in each em - step .
we analyse each globular cluster in our sample using this algorithm , with the [ mg / fe],[al / fe ] , [ si / fe ] , [ ca / fe ] , [ ti / fe ] abundances as well as only [ mg / fe ] and [ al / fe ] .
we find that the two separate analyses provide nearly identical results , indicating that si , ca , and ti do not drive the populations . for m71 and m107 we used [ n / fe ] instead of [ al / fe ] , because al does not show any spread in these two clusters .
we run xd using uncertainties given by the @xmath50final combined errors@xmath51 section in table 4 , and also without any uncertainties , and found that the two different runs produce identical results .
we include missing abundance measurements by using a large uncertainty for these measurements .
group membership for each star is determined using the best - fit gaussian mixture by calculating the posterior probability for each star to be a member of each population based on its elemental abundances and their uncertainties ; stars are then assigned to the population for which this probability is the largest .
these membership assignments are typically unambiguous ( probabilities @xmath52 in most cases ) , with only a few cases for which the maximum probability is below 90 % .
the abundance averages of populations are listed in table 7 . in the next few subsections
we discuss the results from the population fitting in more detail for each individual cluster .
lrrrrrr m15 & 0.16 & 0.037 & 4.3 & 0.22 & 0.055 & 4.0 + m92 & 0.28 & 0.029 & 9.7 & 0.06 & 0.038 & 1.6 + m53 & 0.13 & 0.035 & 3.7 & 0.01 & 0.055 & @xmath201 + m13 & 0.12 & 0.014 & 8.6 & 0.02 & 0.018 & 1.1 + m2 & 0.06 & 0.038 & 1.6 & 0.00 & 0.043 & @xmath201 + m3 & 0.08 & 0.011 & 7.3 & 0.01 & 0.024 & @xmath201 + m5 & 0.03 & 0.008 & 3.8 & 0.01 & 0.014 & @xmath201 + in order for the mg - al cycle to operate , high temperatures above 70 million kelvin are required @xcite
. because current cluster main - sequence stars are unable to reach these temperatures , the high [ al / fe ] abundances we see in some globular cluster stars imply that a previous generation of higher - mass or evolved stars must have contributed to their chemical composition .
figure [ fig : almganti ] shows [ mg / fe ] versus [ al / fe ] for all ten clusters in our study .
the two xd populations are plotted in red and blue , and a representative error bar is included in the top right corner of each panel . as noted previously , stars in m107 and m71
do not show a strong mg - al anticorrelation , and their xd population assignments are based on n abundances .
the light blue points in the panels for m107 and m71 are stars that are warmer than 4500 k , and thus have no cno measurements .
the extended distribution of mg and al abundances apparent in figure [ fig : almganti ] is typical for globular clusters , and it shows the influence of the mg - al fusion cycle , which converts mg into al .
however , there is a variety in the structure of this relationship that has not been thoroughly explored before : in some clusters ( m92 , m53 , ngc 5466 , m2 and m3 ) , there are two distinct abundance groups with a gap , while in other clusters ( m15 , m13 and m5 ) there is no gap .
this result strongly suggests that there is diversity in the process of stellar chemical feedback and star formation in globular clusters , which may relate to the larger environment in which they formed . to date , the largest homogeneous study of mg and al abundances was carried out by @xcite .
we have several clusters in common , m15 , m5 , m107 and m71 which enables a direct comparison . m107 and m71 do not show any anticorrelation in either of the studies .
@xcite only have upper limits for al and mg in m15 , while we were able to make direct measurements , but even with upper limits the two studies show similar anticorrelations with largest spreads in both al and mg from the whole sample of clusters .
the largest difference between the two studies can be seen in case of m5 , where our values span a range of @xmath140.3 dex to + 1.1 dex , while the al abundances of @xcite are limited to between @xmath140.2 and 0.7 dex .
the difference may be explained by a simple selection effect , as @xcite observed 13 stars in m5 , while our sample size is 122 and covers almost the full extent of the rgb .
we discuss the extent of anticorrelation through differences in the average of the mg abundance between the populations , divided by the estimated uncertainty in this average mg difference .
we use the final combined errors from table 4 .
table 8 lists the errors of the average differences with values of how much @xmath8 detection is each difference .
we find that in m15 ( 4.3@xmath8 ) , m92 ( 9.7@xmath8 ) , m13 ( 8.6@xmath8 ) , and m3 ( 7.3@xmath8 ) the differences are statistically significant and the mg - al anticorrelation exists .
two clusters , m53 ( 3.7@xmath8 ) and m5 ( 3.8@xmath8 ) , also show large @xmath8 detections , however , we would like to use a more conservative approach to the detection of anticorrelations out of a concern that our errors may be underestimated , given the degeneracies between stellar parameters and abundances in our analysis .
m2 stands out as having no statistically significant mg - al anticorrelation .
the summed abundance a(mg+al ) is expected to be constant as a function of t@xmath3 when material is completely processed through the mg - al cycle , and that is what our results show in figure [ fig : mgaladd ] .
the only exception is m107 , but the slight correlation is due mg correlating with temperature ( see figure [ fig : femgal ] ) , and this trend was not removed here .
al is expected to correlate with elements enhanced by proton - capture reactions ( n , na ; figure [ fig : sicor ] ) and anti - correlate with those depleted in h - burning at high temperature ( o , mg ; figure [ fig : sicor ] ) .
the al - o anticorrelations and mg - o correlations can be clearly seen in our data in figure [ fig : sicor ] for all clusters except m53 , m107 and m71 .
the small number of stars observed in ngc 5466 does not allow us to investigate the correlations in that cluster in detail . however , the slight correlation of al with si that can be seen in m15 in fig .
[ fig : sicor ] is the evidence of @xmath0si leaking from the mg - al cycle , which is an intriguing result .
si participating in the light - element abundance pattern was first reported in ngc 6752 by @xcite . since then , @xcite have also found si enhancement correlated with al in ngc 2808 .
those authors interpretation was that it is only in low - metallilcity clusters , where the agb stars burn slightly hotter , or in high - mass clusters , where the chemical enrichment is more efficient , that a si enhancement will be observed in second - generation gc stars .
the difference in si abundance between the two xd - identified populations , listed in in table 8 , is significant relative to the error on that measurement .
the si - al correlation in m15 is also accompanied by a si - mg anticorrelation ( figure [ fig : sicor ] ) , which is further evidence of @xmath0si being produced by hot bottom burning in agb stars @xcite .
the spread of al abundances ( figure [ fig : rms ] ) also increases significantly below [ fe / h][email protected] , thus we can conclude that high mass , low metallicity agb polluters that are able to process material at high , 60@xmath1470 million k , play an increasingly larger role in more metal - poor cluster .
this behaviour of al abundances was previously observed by @xcite , but the larger range of al values presented in this paper make this correlation clearer .
theoretical agb nucleosynthesis modeling indeed predict this behavior @xcite .
the high - mass agb stars reach higher temperatures at the bottom of the convective envelope ; i.e. , stronger hbb and advanced ( mg - al ) nucleosynthesis occurs with decreasing metallicity .
we are seeing very advanced mg - al nucleosynthesis in the most metal poor clusters such as m15 and m92 , while the most metal - rich clusters like m107 and m71 do not show this , as theoretically expected , if the high - mass agb stars are the polluters .
also , this is corroborated by the al - o anticorrelation ; the most al - rich stars are o - poor showing the effects of very strong hot bottom burning ( hbb ) , because hbb proceeds completely and destroys o. an other related issue is that hbb is activated for lower masses with decreasing metallicity .
high - mass and very low - metallicity agb stars don not exists in gcs today due to their extremely short lifetimes , but this trend seems to be confirmed both theoretically @xcite and observationally at least for solar metallicities down to that of the magellanic clouds . from the observational point of view , we know that high - mass agb stars in the small magellanic cloud ( [ fe / h][email protected] ) activate hbb for progenitor masses ( m @xmath63 m@xmath38 ) lower than their solar metallicity counterparts ( m @xmath6 4 m@xmath38 ) @xcite .
thus , at the lowest metallicities in m15 and m92 we also would expect more hbb agb stars ( i.e. , with several degrees of hbb and mg - al nucleosynthesis ) , because the minimum stellar mass to activate the hbb process ( and advanced mg - al nucleosynthesis ) decreases with decreasing metallicity .
thus , we conclude that more polluted material would be present at the lowest metallicities . because of these reasons
, we believe that our results support and add some evidence to the high - mass agbs as gcs polluters . .
a clear correlation is visible in m13 , m2 , m3 , and m5 , for more discussion see section 6.4 . ,
width=336 ] carbon , nitrogen and sometimes oxygen are influenced by two independent processes in gc giants : a primordial anticorrelation , with the same first - generation sources as the o - na and mg - al patterns ; and also stellar evolution , driven by circulation between the hydrogen - burning shell and the surface @xcite . as discussed in the previous section ,
we see large star - to - star variations in n abundance in all clusters ; however , our inability to measure c abundances below [ fe / h]=-1.7 dex leads us to only having upper limits , and therefore we restrict the discussion of c - n anticorrelations to the more metal - rich clusters .
figure [ fig : cnanti ] shows [ n / fe ] versus [ c / fe ] for all clusters .
deep mixing has to be taken into account when discussing the cn anticorrelations , so in m13 , m2 , m3 , and m5 , where we see deep mixing clearly , the correlation of [ c / fe ] with t@xmath3 was removed . because deep mixing is not visible in m107 and m71 , we used the original values .
a clear anticorrelation can not be seen in any of the clusters .
while m107 and m71 show weak anticorrelations , neither of these is statistically significant .
the anticorrelation itself may be obscured by the relatively large errors for [ c / fe ] and [ n / fe ] . as mentioned in section 5.1 , we see correlations of [ c / fe ] with temperature in m13 , m3 , m2 , and m5 , but these are not accompanied by increasing [ n / fe ] at the same time .
we believe that this is due to our inability to measure c and n accurately : most of our abundance determinations are upper limits or close to them . if the anticorrelations exist , they must span ranges smaller than our uncertainties , which are always higher than 0.14 dex in [ c / fe ] and 0.12 dex in [ n / fe ] .
it is interesting to note that in m3 , m107 and m71 there is a clear cn weak and cn strong group of stars , while in all more metal poor clusters , n abundances fill out a broad distribution .
we have c , n , and o abundances available for a large number of stars , thus we are able to investigate the c+n+o content in each cluster in detail .
studies from the literature showed that the c+n+o content in globular clusters are fairly constant to within 0.3 dex @xcite , except for n1851 where @xcite find a spread of 0.57 dex . according to @xcite the large spread in c+n+o in n1851
is probably attributed to larger than usual pollution from lower mass agb stars than in other clusters , but these results were questioned by @xcite as they did not see spread of c+n+o larger than 0.3 dex .
we also see near constant c+n+o in our sample , which is consistent with the material in these stars having undergone cno cycling in the first - generation of stars in the rgb phase . in figure
[ fig : cnoadd ] we show a(c+n+o ) as a function of t@xmath3 .
the spread is significantly larger than what is reported in literature , between 0.4 and 0.6 dex for m13 , m2 , m3 , and m5 .
the spread in these clusters is at the level of what has been found in n1851 , but we think this is mostly associated with large uncertainties of [ n / fe ] measurements .
we do not find clear separation in the amount of a(c+n+o ) between the two populations found in any of the clusters .
as previously mentioned , there are currently two leading models to explain the nature of the polluters : the first assumes that intermediate - mass agb stars that are thermally pulsing and undergoing hot bottom burning expel material to the intra - cluster medium by strong mass loss @xcite , while the second assumes that the pollution comes from hot , fast rotating stars @xcite .
according to the first theory , intermediate - mass hbb - agb stars could produce or not produce large c+n+o variations ( see , e.g. , * ? ? ? * and references therein ) .
this is because the c+n+o predictions ( which basically depend on the number of third dredge - up episodes ) for these stars are extremely dependent on the theoretical modelling , the convective model and the mass loss prescription used ( see , e.g. , * ? ? ?
* ; * ? ? ?
a ( c+n+o)-al correlation is not clearly visible in the lower panel of figure 10 . according to the first theory , agb stars may explain a large spread in a(c+n+o ) because they are expected to produce a substantial increase in the c+n+o abundance as they enhance na and al and deplete o and mg @xcite .
this should result in a ( c+n+o)-al correlation , which is not clearly visible in the lower panel of figure [ fig : cnoadd ] .
weak correlations in m107 and m71 , and weak anticorrelations in m13 and m3 are visible , but they are probably the result of large uncertainties in c and n. more accurate measurements of [ n / fe ] will help to clarify the possible ( c+n+o)-al correlation .
we investigated the abundances of nine elements for 428 stars in ten globular clusters using apogee dr10 spectra .
a homogeneous analysis of these gcs has not been accomplished previously , something that apogee is uniquely able to do because it can observe all the bright gcs in the northern hemisphere .
a semi automated code called _ autosynth _ was developed to provide abundances independent of those derived by aspcap .
our main goal was to examine the stellar populations in each cluster using a homogeneous dataset . based on our abundances ,
we find the following : \1 . from the examination of the star - to - star scatter in @xmath1element abundance , only o and mg show an abundance scatter larger than our measurement errors .
however , this is expected , because o and mg are part of the well - known globular cluster light - element abundance anticorrelations .
population analysis using [ n / fe ] for m107 and m71 and
[ al / fe ] in the rest of the clusters confirms that each cluster can be divided into two discrete populations , though in m107 and m71 that division is clearest in the n - al plane , whereas in the other clusters it is best made in the mg - al plane .
the anticorrelation of mg and al is clearly shown in m15 , m92 , m13 and m3 , and to a lesser extent in m53 and m5 .
interestingly , m2 , which is also a metal poor cluster , does not have a mg - al anticorrelation .
the increased number of stars in our sample compared to the literature enables us to discover more al - rich stars , making the anticorrelation clearer in this dataset than in previous studies .
the spread of al abundances increases with decreasing cluster metallicity .
this is in agreement with theoretical agb nucleosynthesis predictions by @xcite .
this suggests that high - mass hbb agb stars , which are able to enrich al while destroying o , are more common polluters in metal - poor clusters than in metal - rich ones .
a correlation between al and si in m15 indicates particularly high - temperature hydrogen burning in the stars that contributed the abundance pattern of the second - generation stars .
this is consistent with si enhancement in globular cluster stars as discussed in the literature , where it appears to be confined to low - metallicity and high - mass clusters .
besides accessing northern gcs , apogee is also unique among other spectroscopic surveys , because it is capable of measuring abundances of cno .
the total abundance of c+n+o is found to be constant in each cluster , within our errors .
our data set do not show an unambiguous correlation between a(c+n+o ) and [ al / h ] . in principle , this finding is not against the idea of intermediate - mass hbb - agb stars as the primary source of chemical self - enrichment in globular clusters .
however , more precise measurements of c and n abundances will clarify this issue , and we expect that aspcap will be able to provide these in the future .
freeman , k. c. 2012 , in asp conf .
458 , galactic archaeology : near - field cosmology and the formation of the milky way , ed .
w. aoki , m. ishigaki , t. suda , t. tsujimoto , & n. arimoto ( san francisco , ca : asp ) , 393 | we investigate the light - element behavior of red giant stars in northern globular clusters ( gcs ) observed by the sdss - iii apache point observatory galactic evolution experiment ( apogee ) .
we derive abundances of nine elements ( fe , c , n , o , mg , al , si , ca , and ti ) for 428 red giant stars in 10 globular clusters . the intrinsic abundance range relative to measurement errors is examined , and the well - known c - n and mg - al anticorrelations are explored using an extreme - deconvolution code for the first time in a consistent way .
we find that mg and al drive the population membership in most clusters , except in m107 and m71 , the two most metal - rich clusters in our study , where the grouping is most sensitive to n. we also find a diversity in the abundance distributions , with some clusters exhibiting clear abundance bimodalities ( for example m3 and m53 ) while others show extended distributions .
the spread of al abundances increases significantly as cluster average metallicity decreases as previously found by other works , which we take as evidence that low metallicity , intermediate mass agb polluters were more common in the more metal poor clusters .
the statistically significant correlation of [ al / fe ] with [ si / fe ] in m15 suggests that @xmath0si leakage has occurred in this cluster .
we also present c , n and o abundances for stars cooler than 4500 k and examine the behavior of a(c+n+o ) in each cluster as a function of temperature and [ al / fe ] .
the scatter of a(c+n+o ) is close to its estimated uncertainty in all clusters and independent on stellar temperature .
a(c+n+o ) exhibits small correlations and anticorrelations with [ al / fe ] in m3 and m13 , but we can not be certain about these relations given the size of our abundance uncertainties .
star - to - star variations of @xmath1elements ( si , ca , ti ) abundances are comparable to our estimated errors in all clusters . |
the study of quantum graphs has become popular in a number of fields in quantum mechanics ranging from molecular physics and the physics of disordered systems to quantum chaology and quantum computation ( see _ e.g. _ @xcite ) .
quantum graphs serve as computationally inexpensive models with the ability to mimic a variety of features also present in more realistic quantum systems .
for example , the now 20 years old conjecture by bohigas , giannoni and schmidt ( bgs ) @xcite , stating that the spectral statistics of quantum systems whose classical limit is chaotic follow those of random hermitian or unitary matrices in the semiclassical limit is well reproduced on quantum graphs @xcite . in this paper , we will address two fundamental , but seemingly disconnected questions related to quantum graphs , namely , we will look at * _ a ways to introduce symmetries on connected quantum graphs _ and investigate * _ the degree of complexity or randomness necessary on a quantum graph to fall within the universal regime of random matrix statistics .
_ the first point has hardly been addressed in the context of quantum graphs .
symmetries on quantum graphs play an important role in studies on quantum random walks considered recently in the context of quantum computation ( see _ e.g. _ @xcite ) .
speed up of mixing - parameters of quantum random walks over classical random walks found on certain graphs is indeed related to interference effects due to symmetries in the quantum propagation .
we will suggest a method for imposing a large class of symmetries on certain types of graphs which has potential applications in the design of effective quantum random walks .
it is furthermore expected that symmetries on graphs will have a profound influence on the statistical properties of spectra of quantum graphs .
the existence of discrete symmetries and associated good quantum numbers on connected quantum graphs is expected to lead to deviations from random matrix results .
the second point addresses the range of validity of the bgs - conjecture .
it is widely believed that the spectra of unitary propagators on quantum graphs follow random matrix statistics if the correlation exponents of an underlying stochastic dynamics are bound away from zero in the limit of large graphs sizes and the length of the arcs of the graph are incommensurate @xcite .
we will argue here that the last condition can be considerably relaxed and that , in context of regular graphs , the existence or absence of random matrix statistics is related to the commutativity properties of certain sets of _ connectivity matrices _ to be defined in detail later . similar results for the spectra of laplacians of regular graphs have been reported in @xcite .
a related discussion of spectra of adjacency matrices of cayley graphs of certain groups can be found in @xcite . + we start by briefly reviewing the notion of a quantum graph . a quantum graph is given by an underlying graph @xmath0 and a set of local scattering matrices at the vertices as well as a set of arc lengths .
a ( finite ) _ directed graph _ or _ digraph _ consists of a finite set of _ vertices _ and a set of ordered pairs of vertices called _
arcs_. we denote by @xmath1 and @xmath2 the set of vertices and the set of arcs of the digraph @xmath0 , respectively . given an ordering of the vertices ,
the _ adjacency matrix _ of a digraph @xmath0 on @xmath3 vertices , denoted by @xmath4 , is the @xmath5 @xmath6-matrix where the @xmath7-th element is defined by @xmath8{ll}% $ 1 $ & if $ ( i , j)\in e^{g},$\\ $ 0 $ & otherwise .
\end{tabular } \right.\ ] ] an _ undirected graph _
( for short , _ graph _ ) is a digraph whose adjacency matrix is symmetric .
the _ line digraph _ of a digraph @xmath0 , denoted by @xmath9 , is defined as follows , ( see _ e.g. _ @xcite ) : @xmath10 and , given @xmath11 , the ordered pair @xmath12 if and only if @xmath13 . a _ quantum graph _ associated with a digraph @xmath0 on @xmath3 vertices may then be defined in terms of a set of unitary vertex scattering matrices @xmath14 on vertices @xmath15 and a set of arc - lengths @xmath16 defined for every arc @xmath17 .
waves propagate freely along the directed arcs , transitions between incoming and outgoing waves at a given vertex @xmath18 are described by the scattering matrix @xmath14 .
the two sets specify a unitary propagator of dimension @xmath19 defining transitions between arcs @xmath20 which has the form @xcite @xmath21{ccc } $ s^{g}=d\ , v$ & with & $ d_{(i , j)(i^{\prime},j^{\prime})}= \delta _ { i , i^{\prime}}\ , \delta _ { j , j^{\prime}}\ , e^{\mathrm{i}kl_{(i , j)}},$% \end{tabular}\ ] ] where @xmath22 is the wave number and @xmath21{ccc } $ v_{(i , j)(i^{\prime},j^{\prime})}=a_{(i , j)(i^{\prime},j^{\prime})}^{lg}% \sigma_{ij^{\prime}}^{(j)}$ & with & $ a_{(i , j)(i^{\prime},j^{\prime})}% ^{lg}=\delta_{j , i^{\prime}}$ $ .$% \end{tabular}\ ] ] the local scattering matrices @xmath23 depend on the boundary conditions and local potentials at the vertex @xmath24 which we do not want to specify here any further . for our purpose , we may regard the @xmath23 s as arbitrary unitaries .
let @xmath25 and @xmath26 be the number of incoming and outgoing arcs of a vertex @xmath24 , respectively .
a sufficient and necessary condition for a digraph @xmath0 to be quantisable in the way above is then , that for every vertex @xmath27 , @xmath28 @xcite .
this means in particular that if @xmath0 is an undirected graph then it is quantisable .
the `` classical '' dynamics corresponding to a quantum graph defined by a unitary propagator @xmath29 is given by a stochastic process with transition matrix @xmath30 @xmath31 note that both the quantum mechanics as well as the associated stochastic dynamics relate to transitions between arcs and is thus defined on the line digraph of @xmath0 .
the paper is organised as follows . in section [ sec : reg - graph ] , we will introduce the notion of regular quantum graphs and discuss a factorisation of the propagator in terms of connectivity matrices for a special class of such graphs . in section [ sec : sym - reg ] , we relate the existence or absence of symmetries on a connected regular quantum graph to properties of the connectivity matrices .
we discuss some specific examples for completely connected graphs including statistical properties of the spectra in section [ sec : examples ] . in section [ sec : rmt ] we show numerically that by inscribing a _ single _ ( @xmath32 unitary matrix into a large regular quantum graph one still obtains random matrix statistics despite huge degeneracy in the set of arc lengths and scattering matrices for a generic choice of connectivity matrices .
we will implement symmetries on quantum graphs for which the wave dynamics at the vertices of the digraph is `` locally indistinguishable '' when going from one vertex to the next .
we will restrict ourselves to wave dynamics on @xmath33-regular digraphs . recall that a digraph @xmath0 is said to be @xmath33 _ -regular _ if , for every vertex @xmath34 , @xmath35 and thus @xmath36 . extending the concept of local indistinguishableness to quantum graphs
, we will consider quantum graphs on @xmath33-regular digraphs with local @xmath37 scattering matrices @xmath23 and set of outgoing arc lengths @xmath16 being identical at every vertex @xmath24 up to permutations of the incoming or outgoing channels .
that is , there are @xmath37 unitary matrices @xmath38 and @xmath39 with @xmath40 and local permutation matrices @xmath41 , such that @xmath21{ccc}% $ \sigma^{(i)}=p^{(i)}\sigma q^{(i)}$ & and & $ d^{(i)}(k)=p^{(i)}d(k)\left ( p^{(i)}\right ) ^{-1}$ $ .$ \end{tabular } \ \ \ \\ ] ] combining the local matrices @xmath38 and @xmath39 to a single matrix @xmath42 , we obtain @xmath43 we call a quantum graph with these properties a _ regular quantum graph_. the matrix @xmath44 is called the _ coin _ in the context of quantum random walks on graphs @xcite .
we denote by @xmath45 the @xmath5 matrix with all elements being equal to 1 and @xmath46 is the identity matrix .
the following observation will be useful in what follows , ( see also @xcite ) : [ proadjlg ] let @xmath4 be the adjacency matrix of a @xmath33-regular digraph @xmath0 .
the adjacency matrix of @xmath9 has up to reordering the arcs the form @xmath47 where the matrices @xmath48 of dimension @xmath3 have entries 0 or 1 and represent discrete functions on the vertex set , that is @xmath49 and in addition @xmath50 * remark : * a given matrix @xmath51 assigns to every vertex @xmath18 a specific arc @xmath52 , see fig .
[ fig : graph ] . note that , for @xmath53 , the choice of matrices @xmath48 is not unique and that the @xmath48 s do not need to be invertible .
different sets of @xmath48 s fulfilling the conditions in the proposition give rise to adjacency matrices of the line digraph which are equivalent up to permutations in the ordering of the arcs in ( [ arc - ord ] ) .
+ condition [ cond - rho ] ensures that @xmath54 for every @xmath24 and @xmath18 ; writing out eqn .
( [ adj - lg ] ) , we obtain @xmath55{cccc } \rho_{1 } & 0 & \cdots & 0\\ 0 & \rho_{2 } & \cdots & 0\\ \vdots & \vdots & \ddots & \vdots\\ 0 & 0 & \cdots & \rho_{d}% \end{array } \right ] , \quad \mbox{and thus } \quad \left ( \bigoplus_{i=1}^{d}\rho_{i}\right ) \cdot \left(j_{d}\otimes i_{n}\right ) = \left [ \begin{array } [ c]{cccc}% \rho_{1 } & \rho_{1 } & \cdots & \rho_{1}\\ \rho_{2 } & \rho_{2 } & \cdots & \rho_{2}\\ \vdots & \vdots & \ddots & \vdots\\ \rho_{d } & \rho_{d } & \cdots & \rho_{d}% \end{array } \right ] \ , .\ ] ] the choice of matrices @xmath51 fixes now a certain ordering of the arcs ; ordering the arcs according to @xmath56 one deduces that non - zero matrix elements of @xmath57 as defined in ( [ adj - lg ] ) refer to transitions @xmath58 for every @xmath59 and every @xmath60 , which are exactly the allowed transition in the line digraph of @xmath0 .
we pointed out in the introduction that the wave propagation on a quantum graph actually lives on the line digraph of the underlying digraph . generalising ( [ adj - lg ] ) to describe unitary propagation on digraphs
, we write @xmath61{cccc } c_{11}\,\rho_{1 } & c_{12}\,\rho_{1 } & \cdots & c_{1d}\,\rho_{1}\\ c_{21}\,\rho_{2 } & c_{22}\,\rho_{2 } & \cdots & c_{2d}\,\rho_{2}\\ \vdots & \vdots & \ddots & \vdots\\ c_{d1}\,\rho_{d } & c_{d2}\,\rho_{d } & \cdots & c_{dd}\,\rho_{d}% \end{array } \right ] , \label{qu - lg}%\ ] ] with @xmath44 being the unitary @xmath37 coin and the matrices @xmath48 fulfil the condition ( [ cond - rho ] ) .
we have to add the additional constraints here that the @xmath48 s are invertible , that is , that they are permutation matrices .
the condition is necessary to ensure that @xmath29 is unitary .
we will refer to the permutations @xmath48 as the _ connectivity matrices _ in what follows , see [ fig : graph ] .
@xmath29 satisfies all the properties of a regular quantum graph as defined above .
the matrix @xmath44 is in particular the coin from which the local scattering matrices @xmath62 at vertices @xmath18 can be deduced .
one obtains @xmath63 the connectivity matrices @xmath64 and @xmath65 specify thus the pair of arcs related through the transition @xmath66 at a given vertex @xmath18 .
* remark : * in contrast to ( [ adj - lg ] ) where different decomposition of @xmath4 of the form ( [ cond - rho ] ) lead to equivalent adjacency matrices ( up to reordering the arcs ) , this is no longer the case for ( [ qu - lg ] ) .
different sets of connectivity matrices lead here to different regular quantum graphs which may have very different spectral properties as will be discussed in the next section .
* remark : * note that not all regular quantum graphs can be written in the form ( [ qu - lg ] ) .
any pair of permutation matrices @xmath67 and @xmath68 leaving the adjacency matrix of a line graph @xmath57 of a @xmath33-regular graph @xmath0 invariant , that is @xmath69 , transform an associate propagator of a regular graph , @xmath70 , into a propagator of a @xmath33-regular , albeit different , quantum graph @xmath71 .
if @xmath29 is of the form ( [ qu - lg ] ) , one easily finds permutations @xmath67 and @xmath68 such that @xmath72 is not of this form .
so far we have considered general regular digraphs . in the special case where the underlying graph is undirected it is natural to consider associated time - reversal symmetric regular quantum graphs ; that is , regular quantum graphs for which for every ( wave)-paths there exists an equivalent time - reversed paths undergoing the same transitions .
a time - reversal symmetric unitary propagator of the form ( [ qu - lg ] ) for an undirected regular graph can be constructed by choosing symmetric coin and connectivity matrices , that is , @xmath21{ccccc } $ c = c^{\intercal}$ & and & $ \rho_{i}=\rho_{i}^{\intercal},$ & & for every $ i=1,\ldots , d$ $ .$ \end{tabular } \\ ] ] note , that the symmetry conditions for the connectivity matrices severely limit the choice of possible graphs and decompositions .
we note first that if a regular quantum graph can be written in the form ( [ qu - lg ] ) and there exists an invertible @xmath5 matrix @xmath73 such that @xmath74{ccc}% $ \lbrack\pi,\rho_{i}]=0 $ & & for every $ i=1,\ldots d.$% \end{tabular } \
\label{comm}%\ ] ] then @xmath74{ccc}% $ \lbrack p , s^{g}]=0 $ & with & $ p=\left ( i_{d}\otimes\pi\right ) $ % \end{tabular } \ \label{comm1}%\ ] ] independently of the choice of the coin @xmath44 .
the result follows immediately from @xmath75=0\text { } .\ ] ] it is obvious that the condition ( [ comm ] ) implies @xmath76=0\text { } $ ] .
the above property enables us to study certain symmetries of quantum graphs in terms of the symmetries of the connectivity matrices only .
given a @xmath33-regular digraph @xmath0 we can in general find many sets of connectivity matrices which sum up to @xmath4 and which may have very different symmetry properties . or if one is interested in quantum graphs with specific symmetries one may start from a set of connectivity matrices @xmath48 in order to construct quantum graphs with desired properties .
we consider various scenarios here and give some specific examples in the next section . in the special case
when all connectivity matrices commute , every @xmath51 acts as a symmetry @xmath78 .
the spectrum of @xmath29 can then be decomposed into @xmath3 sub - spectra of dimension @xmath33 .
[ prop1 ] let @xmath29 be of the form @xmath79 where the @xmath5 connectivity matrices @xmath48 fulfil @xmath80=0 $ ] , @xmath81 is the adjacency matrix of a d - regular digraph and @xmath44 is a @xmath37 unitary matrix .
let @xmath82 be the @xmath5 unitary matrix simultaneously diagonalising the @xmath48 s , that is , @xmath83 and @xmath84 is the @xmath85-th eigenphase of the connectivity matrix @xmath48 where the order is determined by the transformation @xmath82 .
the spectrum of @xmath29 , @xmath86 , is then @xmath87 where @xmath88 define @xmath89 and note that @xmath90=0.\ ] ] thus @xmath91 there exist permutation matrices @xmath67 such that @xmath92 and thus @xmath93 is indeed block - diagonal of the form stated in the proposition . * remark : * note that the decomposition is independent of the coin @xmath44 .
it can be shown @xcite that a set of commuting connectivity matrices of a connected graph @xmath0 always form a subset of the regular ( permutation ) representation of an abelian group and the underlying symmetry of the corresponding quantum graph is given by that group .
( the commutativity of the @xmath48 s does in fact imply that @xmath0 is a cayley digraph of an abelian group ; it must therefore have the form of a discretised torus ) .
the sub - spectra obtained from @xmath94 may then be characterised in terms of the eigenbasis of the generators of the abelian group represented by the connectivity matrices .
let @xmath95 be the generators of such an abelian group , @xmath21{cccc } with $ a_{i}^{n_{i}}=id$ & and $ \prod_{i=1}^{r}n_{i}=n,$ & where & $ n_{i } \geq2.$ \end{tabular}\ ] ] the eigenbasis of the connectivity matrices may then be written in dirac notation as @xmath96 with @xmath97 and the sub - spectra obtained from ( [ sub - ab ] ) are characterised by a set of @xmath98 `` quantum numbers '' @xmath99 .
such a regular quantum graph is thus a discretized version of a quantum systems whose underlying classical dynamics has @xmath98-integrals of motion in involution .
some additional degrees of freedom are represented by the coin @xmath44 which may or may not be related to classical chaotic dynamics depending on the properties of @xmath44 and the group .
next we consider the case that a symmetry @xmath73 exists with @xmath100=0 $ ] for all @xmath27 , but @xmath101\neq0 $ ] for some @xmath102 .
that implies that @xmath73 has degenerate eigenvalues ; @xmath73 could for example represent a @xmath103 symmetry of the graph , that is , @xmath104 with eigenvalues @xmath105 only .
let us assume that @xmath73 has @xmath106 distinct eigenvalues @xmath107 , @xmath108 , each with multiplicity @xmath109 with @xmath110 let @xmath82 be a unitary matrix diagonalising @xmath73 in the form @xmath111 @xmath82 then brings @xmath48 into block - diagonal form , that is , @xmath112 with @xmath113 .
the spectrum of @xmath70 is now decomposed in the following way : [ prop2]let @xmath29 be of the form ( [ qu - lg ] ) and the matrices @xmath48 , @xmath73 have the properties as described above ; @xmath114 is then of the form @xmath115 with @xmath74{ccc}% $ s_{m}^{g}=\left ( \bigoplus_{i=1}^{d}\tilde{\rho}_{i}^{(m)}\right ) \cdot\left ( c\otimes i_{n_{m}}\right ) $ & where & $ m=1,\ldots , r$ $ .$% \end{tabular } \
\label{sub - pi}%\ ] ] the proof goes along the line of the proof of proposition [ prop1 ] .
the decomposition is again independent of the coin @xmath44 , but the sub - spectra are now of dimension @xmath116 there is a trivial symmetry independent of the particular choice of the @xmath48 s related to the fact that every permutation matrix has an eigenvalue 1 with corresponding eigenvector @xmath117 .
the symmetry @xmath73 in question has the form @xmath118 having two distinct eigenvalues @xmath119 and the eigenvalue @xmath120 has geometric multiplicity @xmath121 . as a consequence any @xmath29 can be block - diagonalised containing @xmath44 as an @xmath37 block , and thus @xmath122
when constructing particular examples , it is useful to start with the completely symmetric graph , namely that of a _ fully connected graph_. this graph , also called the _ complete graph _ , has adjacency matrix @xmath45 . as @xmath124=0 $ ] for every permutation matrix @xmath67 of size @xmath3 , we may indeed construct regular quantum graphs of degree @xmath125 with whatever finite symmetry we want .
in addition , we can make use of the fact that if @xmath126 is a finite group of order @xmath3 and the @xmath127 permutation matrices @xmath48 form a regular representation of @xmath126 then @xmath128 we can thus implement the group properties of any finite group on a regular quantum graph by choosing the regular representations of that group as the connectivity matrices . in what follows
we will consider various decompositions of @xmath45 and see how they effect spectral properties like level statistics .
the simplest abelian group is the cyclic group @xmath129 .
the regular representations @xmath48 are of the form @xmath21{ccccc } $ ( \rho_{j})_{kl}=\delta_{k,(l+j)\operatorname{mod}n}$ & with eigenvalues & $ \chi^j_{m}=e^{2\pi\mathrm{i } \frac{jm}{n}},$ & where & $ j , m=1,\ldots , n\,.$% \end{tabular } \\ ] ] here , @xmath130 and @xmath131 . in order to construct
regular quantum graphs with circular symmetry independent of the coin @xmath44 we use the regular representation of @xmath129 as connectivity matrices . the spectrum of the quantum graph
can then be decomposed into the sub - spectra given by @xmath132 the eigenvalues are characterised in terms of two quantum numbers , an ` angular momentum ' quantum number @xmath85 and a second quantum number @xmath98 , say , counting the eigenvalues in each @xmath85 manifold .
if the spectra for different @xmath85 are uncorrelated , one expects poisson statistics of the total spectrum in the limit @xmath133 .
figure [ fig : stat]a ) shows spectral properties of @xmath29 with @xmath134 , that is , @xmath135 .
we plot here the nearest neighbour spacing ( nns ) distribution @xmath136 and the form factor @xmath137 , the fourier transform of the two - point correlation function .
the coin is of the form ( [ coin ] ) where the local scattering matrix @xmath38 is taken randomly from a @xmath138 - ensemble , but then fixed , and the arc lengths entering the diagonal matrix @xmath139 are chosen independently and identically distributed in @xmath140 $ ] , but then fixed .
the average is taken over the wavelength @xmath22 @xcite . ]
both approaches are equivalent under very general conditions @xcite . ]
the numerical results shown in fig . [ fig : stat]a ) indeed suggest poisson - statistics apart form deviations in the form factor on scales @xmath141 due to the ` chaotic nature ' of the coin . and
nearest neighbour spacing distribution @xmath136 for ( a ) @xmath51 s are the regular representation of the cyclic group @xmath142 ; ( b ) @xmath51 s represent the symmetric group @xmath143 ; ( c ) a generic set @xmath51 s without non - trivial symmetries .
the dashed curve in ( b ) labelled red .
poisson corresponds to a distribution of degenerate levels being poisson distributed otherwise.,height=718 ] next , we consider a specific example of a non - abelian group , namely the symmetric group @xmath144 with @xmath134 elements ; we will discuss spectral properties of general groups elsewhere @xcite .
the regular representation of @xmath144 can be decomposed in terms of its irreducible representations ( for short _ irreps _ ) ; each @xmath48 contains each @xmath33-dimensional irrep exactly @xmath33 times .
the group @xmath144 has 2 one - dimensional , 1 two - dimensional and 2 three - dimensional irreps , such that @xmath145 denote the irreps of the group element @xmath146 as @xmath147 with @xmath148 and the index @xmath149 counting different irreps of the same dimension ; there exists then a transformation @xmath82 such that @xmath150 the connectivity matrices @xmath48 are thus of the form as discussed in section [ sec : sec - part ] .
note that the sub - matrices @xmath151 related to @xmath33-dimensional irreps occur now @xmath33 times in the decomposition .
we thus have 5 independent sub - spectra , 2 of dimension 24 , 1 of dimension 48 and 2 of dimension 72 of which the latter are of multiplicity two and three , respectively .
the huge degeneracy in the spectra can clearly be seen in the spectral statistics ; it is manifest in the peak a @xmath152 in @xmath136 ( see figure [ fig : stat]b ) and leads to @xmath21{ccc } $ k(\tau)=(2\cdot3^{3}+1\cdot2^{3}+2\cdot1^{3})=8/3 $ & for & $ \tau>3/24\,.$% \end{tabular } \ \\ ] ] the spectra appear to be uncorrelated otherwise ; note however , that the spectrum for each sub - matrix @xmath151 alone are correlated following @xmath138 statistics , which manifests itself in the deviations from purely poisson behaviour in @xmath136 ( cf . dashed curve ) as well as in the behaviour of the form factor for @xmath153 which is dominated by the sub - spectra of the three dimensional irreps .
the overwhelming number of decompositions of the form ( [ cond - jd ] ) will of course have no common symmetry apart from the trivial symmetry discussed in section [ sec : sec - part ] .
even though no further analytical results can be given in this case , a numerical study may reveal interesting insights into the range of validity of the universal rmt - regime .
figure [ fig : stat]c ) shows the level statistics of a regular quantum graph obtained from a fully connected graph for a generic choice of connectivity matrices .
one indeed finds good agreement with random matrix theory for the cue - ensemble .
deviations in the formfactor for small @xmath154 can be attributed to the fact that the spectrum of @xmath44 is contained in the full spectrum . after removing this separable part of the spectrum as done for the nns in figure [ fig : stat]c )
there is good agreement with random matrix results .
it is worth keeping in mind , that this is a highly non - random matrix ; we are dealing here with the spectrum of the @xmath155 unitary matrix @xmath70 which has only @xmath156 non - zero elements of which only @xmath157 are independent .
in particular , the arc lengths in the graph are not incommensurate , the @xmath157 arcs in the graph share indeed only @xmath3 different lengths among them .
still , universality is obtained .
the origin of the complexity in this type of quantum graphs is here clearly not due to the randomness in the choice of the matrix elements but due to _ the lack of a common symmetry _ in the set of connectivity matrices .
the results in the last section suggest that spectral statistics of regular quantum graphs can to a large extent be controlled by properties of the permutation matrices @xmath51 independently of the coin @xmath44 .
it is thus natural to ask whether we can reduce the dimension of the coin to its smallest possible value , namely @xmath158 , by considering large 2 - regular quantum graphs and still obtain random matrix correlations .
we can only expect random matrix statistics on a regular quantum graph if the corresponding quantum graph with randomly chosen arc - lengths falls into the random matrix category .
we therefore need to consider 2-regular graphs leading to fast ( classical ) mixing and not for example diffusive networks like ring graphs @xcite exhibiting 1d - anderson localisation .
the d - regular digraphs with the fastest mixing rates are so - called @xmath33-ary _ de bruijn graphs _ of order @xmath22 , @xmath159 , being the @xmath160st line graph generation of a complete graph of size @xmath33 .
that is @xmath161 and @xmath162 is iteratively defined as @xmath163 @xcite .
de bruijn graphs have size @xmath164 and are equipped with a complete symbolic dynamics of order @xmath33 .
they play an important role in coding theory and parallel algorithms @xcite .
numerical evidence suggests that quantum graphs based on de bruijn digraphs with incommensurate arc lengths follow random matrix statistics in the limit of large graphs sizes even for @xmath165 @xcite . in fig .
[ fig : debrstat ] we show results for a regular quantum graph based on a binary de bruijn graph @xmath166 with k = 9 and quantum propagator @xmath167 where @xmath168 are permutation matrices with @xmath169 and @xmath44 is a @xmath170 unitary matrix . the statistics in fig .
[ fig : debrstat ] are obtained by averaging over the space of @xmath171 unitaries with respect to the haar measure .
the connectivity matrices of dimension @xmath172 have been chosen randomly . to avoid accidental symmetries , different sets of connectivity matrices
have been produced and the statistics of the corresponding ensemble averages combined .
the spectral statistics of these regular quantum graphs agrees again very well with cue statistics .
recall that the unitary matrix @xmath173 of size @xmath174 = 1024 has @xmath175 nonzero matrix elements which do , however , take on only 4 different ( complex ) values ! we have thus constructed extremely non - random matrices which still show universal random matrix statistics and have thereby shown that the bgs - conjecture is valid far beyond regimes previously thought to be included in the conjecture . and nearest neighbour spacing distribution @xmath136 for binary de bruijn graph of order @xmath176 with @xmath177.,height=188 ] similar numerical results were also found for the spectra of the laplacian of regular graphs @xcite .
a laplacian on a @xmath33 - regular undirected , loop - less graph @xmath0 is @xmath178 and is thus a symmetric matrix whose non - zero matrix elements take on the values 1 or @xmath33 only . when averaging over sets of @xmath33 - regular graphs agreement with goe statistics was found .
this underlines once more that the origin of universality in spectral statistics lies not in the randomness of the matrix elements .
we introduce a decomposition of certain regular quantum graphs which separates the quantum propagator on a graph into a topological part containing the connectivity matrices and a trivial part containing the quantum scattering information at the vertices .
this allows one to implement global symmetries on the graph by choosing the connectivity matrices according to desired symmetry properties .
we demonstrate that the complexity in the quantum spectrum ( which may be seen to take on its maximal value when the statistics coincides with rmt ) can here be linked to the amount of complexity contained in the set of permutation matrices building up the quantum graph .
we present examples , where for a given graph and a fixed coin matrix , we were able to construct anything from poisson to rmt - statistics just by changing the set of connectivity matrices . by doing so ,
we leave the local properties of the graph invariant , but change the way in which incoming and outgoing channels between vertices are connected and thus the global structure of the wave dynamics .
we take this concept to its extreme by demonstrating numerically that unitary matrices representing 2-regular quantum graphs whose non - zero matrix elements take on only four different values still follow cue statistics for de bruijn graphs .
we believe that our results open up new perspectives in understanding universality in spectral statistics .
it transforms the question from a continuous into an essentially discrete problem focusing on the way local scattering processes are connected and condensing the parameter space to an absolute minimum , ( namely four - dimensional ) . +
* acknowledgement : * we would like to thank andreas winter for interesting discussions and stephen creagh and jens marklof for reading the manuscript and for valuable comments .
the main part of the work was carried out during a _ royal society industrial fellowship _ by one of us ( gt ) who would like to thank the royal society and hewlett - packard , bristol , for financial support .
g. tanner , _ j. phys . a : math .
gen . _ * 34 * 8485 - 8500 ( 2001 ) . nlin.cd/0104014 .
a. terras , _ amer . math .
monthly _ * 109 * , 121 - 139 ( 2003 ) ; a. terras , _ fourier analysis on finite groups and applications _ , ( cambridge university press , cambridge , 1999 ) . | we introduce the concept of regular quantum graphs and construct connected quantum graphs with discrete symmetries .
the method is based on a decomposition of the quantum propagator in terms of permutation matrices which control the way incoming and outgoing channels at vertex scattering processes are connected .
symmetry properties of the quantum graph as well as its spectral statistics depend on the particular choice of permutation matrices , also called connectivity matrices , and can now be easily controlled .
the method may find applications in the study of quantum random walks networks and may also prove to be useful in analysing universality in spectral statistics . |
quantum entanglement of gaussian states constitutes a fundamental resource in continuous variable quantum information processing and communication @xcite . for the class of gaussian states there exist necessary and sufficient criteria of entanglement @xcite and quantitative entanglement measures @xcite .
the quantum information processing tasks are difficult to implement , due to the fact that any realistic quantum system is not isolated and it always interacts with its environment . quantum coherence and entanglement of quantum systems
are inevitably influenced during their interaction with the external environment . as a result of the irreversible and uncontrollable phenomenon of quantum decoherence ,
the purity and entanglement of quantum states are in most cases degraded .
therefore in order to describe realistically quantum information processes it is necessary to take decoherence and dissipation into consideration .
decoherence and dynamics of quantum entanglement in continuous variable open systems have been intensively studied in the last years @xcite .
when two systems are immersed in an environment , then , in addition to and at the same time with the quantum decoherence phenomenon , the environment can also generate a quantum entanglement of the two systems and therefore an additional mechanism to correlate them @xcite . in this paper
we study , in the framework of the theory of open systems based on completely positive quantum dynamical semigroups , the dynamics of the continuous variable entanglement of two identical harmonic oscillators coupled to a common thermal environment .
we are interested in discussing the correlation effect of the environment , therefore we assume that the two oscillators are uncoupled , i.e. they do not interact directly .
the initial state of the subsystem is taken of gaussian form and the evolution under the quantum dynamical semigroup assures the preservation in time of the gaussian form of the state . in section 2
we write the markovian master equation in the heisenberg representation for two uncoupled harmonic oscillators interacting with a general environment and the evolution equation for the covariance matrix of the considered subsystem . by using the peres - simon criterion for separability of two - mode gaussian states
@xcite , we investigate in section 3 the dynamics of entanglement for this system . for certain values of the environment temperature
, the state keeps for all times its initial type : separable or entangled . for
other values of the temperature , entanglement generation , entanglement sudden death or a repeated collapse and revival of entanglement take place . in recent papers , in a different model , paz and roncaglia @xcite studied numerically the exact entanglement behaviour of two identical oscillators in an infinite bath of other oscillators by using the exact master equation for quantum brownian motion and showed that the entanglement can undergo three qualitatively different dynamical phases : sudden death , sudden death and revival , and no sudden death of entanglement .
finally , we determine the asymptotic gaussian maximally entangled mixed states ( gmems ) and the corresponding asymptotic maximal logarithmic negativity , which characterizes the degree of entanglement of these states . a summary is given in section 4 .
we study the dynamics of the subsystem composed of two identical non - interacting oscillators in weak interaction with a thermal environment . in the axiomatic formalism based on completely positive quantum dynamical semigroups , the irreversible time evolution of an open system
is described by the following general quantum markovian master equation for an operator @xmath0 in the heisenberg representation ( @xmath1 denotes hermitian conjugation ) @xcite : @xmath2+{1\over 2\hbar}\sum_j(2v_j^{\dagger}av_j - v_j^{\dagger } v_ja - av_j^{\dagger}v_j).\label{masteq}\end{aligned}\ ] ] here , @xmath3 denotes the hamiltonian of the open system and the operators @xmath4 defined on the hilbert space of @xmath5 represent the interaction of the open system with the environment .
we are interested in the set of gaussian states , therefore we introduce such quantum dynamical semigroups that preserve this set during time evolution of the system . consequently @xmath3 is taken to be a polynomial of second degree in the coordinates @xmath6 and momenta @xmath7 of the oscillators and @xmath8 are taken polynomials of first degree in these canonical observables . then in the linear space spanned by coordinates and momenta there exist only four linearly independent operators @xmath9 @xcite : @xmath10 where @xmath11 are complex coefficients .
the hamiltonian @xmath3 of the two uncoupled identical harmonic oscillators of mass @xmath12 and frequency @xmath13 is given by @xmath14 the fact that the evolution is given by a dynamical semigroup implies the positivity of the following matrix formed by the scalar products of the four vectors @xmath15 whose entries are the components @xmath16 respectively : @xmath17 we take this matrix of the following form , where all diffusion coefficients @xmath18 ... and dissipation constant @xmath19 are real quantities ( we put from now on @xmath20 ) : @xmath21 it follows that the principal minors of this matrix are positive or zero . from the cauchy - schwarz inequality
the following relations hold for the coefficients defined in ( [ coef ] ) : @xmath22 a two - mode gaussian state is completely characterized by its first and second moments of canonical variables .
we therefore introduce the following @xmath23 bimodal covariance matrix , which entirely specifies a two - mode gaussian state ( all first moments have been set to zero by means of local unitary operations which do not affect the entanglement ) : @xmath24 the problem of solving the master equation for the operators in heisenberg representation can be transformed into a problem of solving first - order in time , coupled linear differential equations for the covariance matrix elements .
namely , from ( [ masteq ] ) we obtain the following system of equations for the quantum correlations of the canonical observables , written in matrix form @xcite ( @xmath25 denotes a transposed matrix ) : @xmath26 where @xmath27 @xmath28 the time - dependent solution of eq .
( [ vareq ] ) is given by @xcite @xmath29 m^{\rm t}(t)+\sigma(\infty),\label{covart}\end{aligned}\ ] ] where the matrix @xmath30 has to fulfill the condition @xmath31 in order that this limit exists , @xmath32 must only have eigenvalues with negative real parts .
the values at infinity are obtained from the equation @xmath33
the characterization of the separability of continuous variable states using second - order moments of quadrature operators was given in refs .
a two - mode gaussian state is separable if and only if the partial transpose of its density matrix is non - negative [ necessary and sufficient positive partial transpose ( ppt ) criterion ] .
a two - mode gaussian state is entirely specified by its covariance matrix ( [ covar ] ) , which is a real , symmetric and positive matrix with the block structure @xmath34 where @xmath0 , @xmath35 and @xmath36 are @xmath37 hermitian matrices . @xmath0 and @xmath35 denote the symmetric covariance matrices for the individual one - mode states , while the matrix @xmath36 contains the cross - correlations between modes .
simon @xcite derived a ppt criterion for bipartite gaussian continuous variable states : the necessary and sufficient criterion for separability is @xmath38 where @xmath39- \frac{1}{4}(\det a+\det b ) \label{sim1}\end{aligned}\ ] ] and @xmath40 is the @xmath37 symplectic matrix @xmath41 since the two oscillators are identical , it is natural to consider environments for which @xmath42 then both unimodal covariance matrices are equal , @xmath43 and the entanglement matrix @xmath36 is symmetric . in order to describe the dynamics of entanglement , we use the ppt criterion @xcite according to which a state is entangled if and only if the operation of partial transposition does not preserve its positivity .
concretely , we have to analyze the time evolution of the simon function @xmath44 ( [ sim1 ] ) . for a thermal environment characterized by the temperature @xmath45 we consider such diffusion coefficients , for which @xmath46 @xmath47 this corresponds to the case when the asymptotic state is a gibbs state @xcite .
we consider two cases , according to the type of the initial gaussian state : 1 ) separable and 2 ) entangled .
\1 ) to illustrate a possible generation of the entanglement , we represent in figure 1 the dependence of function @xmath44 on time @xmath48 and temperature @xmath49 for a separable initial gaussian mixed state .
we notice that , according to peres - simon criterion , for relatively small values of the temperature @xmath45 the initial separable state @xmath50 becomes entangled shortly after the initial moment of time @xmath51 . for relatively large values of @xmath45 @xmath44 is strict positive and the state remains separable for all times .
depending on the environment temperature , there are three situations in the case of a generated entanglement @xcite : a ) entanglement may persist forever , including the asymptotic final state ; b ) there exist repeated collapse and revival of entanglement ; c ) the entanglement is created only for a short time , then it disappears and the state becomes again separable .
the entanglement of the two modes can be generated from an initial separable state during the interaction with the environment only for certain values of mixed diffusion coefficient @xmath52 and dissipation constant @xmath53 \2 ) an example of the evolution of an entangled initial state is illustrated in figure 2 , where we represent the dependence of function @xmath44 on time @xmath48 and temperature @xmath49 for an entangled initial gaussian mixed state .
for relatively small values of @xmath45 the initial entangled state may remain entangled for all times @xcite .
for relatively large values of temperature @xmath45 at some finite moment of time , @xmath44 takes non - negative values and therefore the state becomes separable .
this is the so - called phenomenon of entanglement sudden death .
this phenomenon is in contrast to the loss of quantum coherence , which is usually gradual [ 12 ] .
depending on the values of the temperature , it is also possible to have a repeated collapse and revival of the entanglement . on time
@xmath48 and environment temperature @xmath54 for @xmath55 @xmath56 and for a separable initial gaussian mixed state with initial correlations @xmath57 we take @xmath58 , title="fig : " ] , title="fig : " ] for gaussian states , the measures of entanglement of bipartite systems are based on some invariants constructed from the elements of the covariance matrix @xcite .
heisenberg uncertainty principle can be expressed as a constraint on the global symplectic invariants @xmath59 and @xmath60 @xcite : @xmath61 and the symplectic eigenvalues @xmath62 which form the symplectic spectrum of covariance matrix @xmath63 are determined by @xmath64 in terms of @xmath65 relation ( [ heis ] ) takes the form @xmath66 . in order to quantify the degrees of entanglement of the infinite - dimensional bipartite system states of the two oscillators
it is suitable to use the logarithmic negativity . for a gaussian density operator ,
the logarithmic negativity is completely defined by the symplectic spectrum of the partial transpose of the covariance matrix .
it is given by @xmath67 where @xmath68 is the smallest of the two symplectic eigenvalues of the partial transpose @xmath69 of the 2-mode covariance matrix @xmath70 @xmath71 here @xmath72 is given by @xmath73 a state is separable if and only if @xmath74 and logarithmic negativity quantifies the violation of inequality ( [ seigen ] ) . in our model , the logarithmic negativity
is calculated as @xmath75 , \end{aligned}\ ] ] where @xmath76 ^ 2-\det\sigma(t)}\right)^{1/2}.\end{aligned}\ ] ] it determines the strength of entanglement for @xmath77 and if @xmath78 then the state is separable .
as expected , the logarithmic negativity has a behaviour similar to that one of the simon function in what concerns the characteristics of the state of being separable or entangled @xcite . from ( [ covarinf ] ) ,
( [ envcoe1 ] ) and ( [ envcoe2 ] ) we obtain the following elements of the asymptotic matrices @xmath79 @xmath80 and of the entanglement matrix @xmath81 @xmath82 @xmath83 @xmath84 the mixedness of a quantum state @xmath85 is characterized by its purity @xmath86 . for a two - mode gaussian state
the purity is given by @xmath87 .
the marginal purities @xmath88 ( @xmath89 ) of the reduced states in mode @xmath90 are given by @xmath91 and @xmath92 the global and marginal purities range from @xmath93 to @xmath94 , and they fulfill the constraint @xmath95 as a direct consequence of heisenberg uncertainty relations . in ref .
@xcite the following upper and lower bounds on the invariant @xmath96 have been obtained in terms of global and marginal purities : @xmath97 the invariant @xmath96 has a direct physical interpretation @xcite : at given global and marginal purities , @xmath96 determines the amount of entanglement of the state and the smallest symplectic eigenvalue @xmath68 of the partially transposed state is strictly monotone in @xmath98 consequently , the entanglement of a gaussian state with fixed global purity @xmath99 and marginal purities @xmath100 is strictly increasing with decreasing @xmath96 .
according to double inequality ( [ ineqd ] ) , giving lower and upper bounds on @xmath101 there exist both maximally and minimally entangled gaussian states .
we analyze the existence of the entanglement in the asymptotic regime in the symmetric situation @xmath102 first we consider the particular case @xmath103 in the limit of long times , we obtain the following quantities : @xmath104 where we used the notations : @xmath105 then we obtain @xmath106 these values saturate the lower bound in inequalities ( [ ineqd ] ) and this situation entails a maximal entanglement .
consequently , the corresponding states are gaussian maximally entangled mixed states ( gmems ) .
they are thermal squeezed states with the squeezing parameter @xmath107 given by @xmath108 in the pure case ( @xmath109 ) , these states are equivalent to two - mode squeezed vacua with the squeezing parameter determined only by the temperature of the thermal environment : @xmath110 according to ref .
@xcite , these states are separable in the range @xmath111 then , for a given temperature @xmath45 we obtain that the asymptotic final state is separable for the following range of positive values of the mixed diffusion coefficient @xmath112 : @xmath113 we remind that , according to inequalities ( [ coefineq ] ) , the coefficients have to fulfill also the constraint @xmath114 for a given temperature of the environment and for this range of mixed diffusion coefficients , no entanglement can occur for these states . outside this region ,
i.e. for a temperature and diffusion coefficient satisfying @xmath115 they are gaussian maximally entangled mixed states ( gmems ) .
the asymptotic logarithmic negativity has the form @xmath116 outside the separable region , this is the maximum possible value of the logarithmic negativity , attained by gmems .
it depends only on the mixed diffusion coefficient , dissipation constant and temperature , and does not depend on the initial gaussian state . when @xmath117 but @xmath118 then @xmath119 and the asymptotic final state is always separable .
if both these diffusion coefficients are non - zero , @xmath120 @xmath118 then the lower bound in inequalities ( [ ineqd ] ) is not anymore saturated and , consequently , the corresponding states are gaussian non - maximally entangled mixed states .
we notice that the asymptotic persistence of entanglement is the result of the competition between thermal decoherence , determined by the uni - modal diffusion coefficients ( [ envcoe1 ] ) and the statistical coupling of the two modes , due to the presence of position - position and momentum - momentum diffusion coefficients , which are chosen of equal relevance in our model ( @xmath121 ) and position - momentum diffusion coefficients @xmath122
in the framework of the theory of open quantum systems based on completely positive quantum dynamical semigroups , we investigated the markovian dynamics of the quantum entanglement for a subsystem composed of two noninteracting modes embedded in a thermal environment . by using the peres - simon necessary and sufficient criterion for separability of two - mode gaussian states ,
we have described the evolution of entanglement in terms of the covariance matrix for gaussian input states .
the dynamics of the quantum entanglement is sensitive to the initial states and the parameters characterizing the environment ( diffusion and dissipation coefficients and temperature ) .
for some values of these parameters , the state keeps for all times its initial type : separable or entangled . in other cases , entanglement generation or entanglement suppression ( entanglement sudden death ) take place or even one can notice repeated collapse and revival of entanglement .
we have also shown that , independent of the type of the initial state - separable or entangled , for certain values of temperature , the initial state evolves asymptotically to an equilibrium state which is entangled , while for other values of temperature the asymptotic state is separable . for a given temperature , we calculated the range of mixed diffusion coefficients which determine the existence of asymptotic gaussian maximally entangled mixed states ( gmems ) .
we obtained also the expression of the maximal logarithmic negativity , which characterizes the degree of entanglement of these states .
the author acknowledges the financial support received within the project cncsis - idei 497/2009 and project pn 09 37 01 02/2009 .
the author thanks victor dodonov and salomon mizrahi for invitation and kind hospitality . | in the framework of the theory of open systems based on completely positive quantum dynamical semigroups , we give a description of the dynamics of entanglement for a system consisting of two uncoupled harmonic oscillators interacting with a thermal environment . using peres - simon necessary and sufficient criterion for separability of two - mode gaussian states , we describe the evolution of entanglement in terms of the covariance matrix for a gaussian input state .
for some values of the temperature of environment , the state keeps for all times its initial type : separable or entangled . in other cases , entanglement generation , entanglement sudden death or a repeated collapse and revival of entanglement take place .
we determine the asymptotic gaussian maximally entangled mixed states ( gmems ) and their corresponding asymptotic maximal logarithmic negativity . |
( left scale , closed symbols ) and spin pumping voltage @xmath51 ( right scale , open symbols ) determined from experiments as a function of pt thickness @xmath112 for constant microwave driving field @xmath113 at @xmath10 .
( b ) calculated magnon - electron temperature difference @xmath18 ( left scale , closed symbols ) between yig and pt at the yig / pt interface and measured spin seebeck voltage @xmath47 ( right scale , open symbols ) at constant laser power @xmath114 impinging on the sample . ] figure [ fig : si_v](a ) shows the precession cone angle @xmath13 as determined from the line width of the ferromagnetic resonance spectra and the spin pumping voltage @xmath51 determined from experiment for constant microwave driving field @xmath66 at @xmath10 . at constant @xmath115 , @xmath13 is determined mainly by the properties of the ferromagnet and thus there is no systematic evolution of @xmath13 with @xmath112 .
generally however , it can be observed that larger @xmath13 results in larger @xmath51 as already demonstrated in @xcite .
furthermore , it can be seen that due to electrical short circuiting , the inverse spin hall effect conversion of the spin current to @xmath51 tends to be more efficient at small @xmath112 .
figure [ fig : si_v](b ) depicts the simulated magnon - electron temperature difference @xmath18 between yig and pt at the yig / pt interface . for details of the calculations that rely on a finite kapitza interface resistance
refer to ref .
@xcite . a systematic evolution of @xmath18 as a function of @xmath112 is observed .
the reduced @xmath18 at small @xmath112 is due to the incomplete absorption of the laser light .
the scatter in the simulated @xmath18 reflects the influence of spin backflow on @xmath18 which hence depends on the normal metal resistivity @xcite .
the observed spin seebeck voltage @xmath47 is reduced for large @xmath112 due to electrical short circuiting .
all y@xmath7fe@xmath8o@xmath9 ( yig ) thin films were grown ( 111 ) oriented on either gd@xmath7ga@xmath8o@xmath9 ( ggg ) or y@xmath7al@xmath8o@xmath9 ( yag ) substrates by pulsed laser deposition and covered in situ with au , cu and/or pt by electron beam evaporation .
all samples were diced in two parts .
one part was patterned into a hall bar geometry for laser - induced longitudinal spin seebeck and spin hall magnetoresistance measurements while the unpatterned part was used for the spin pumping experiments .
table [ table : bisamples ] lists all investigated bilayer samples and table [ table : trisamples ] the trilayer samples .
lc sample & @xmath37 ( n@xmath116 m ) + ggg / yig(50)/pt(7 ) & 409.4 + ggg / yig(54)/pt(7 ) & 406.5 + ggg / yig(53)/pt(2.5 ) & 719 + ggg / yig(65)/pt(6.6 ) & 582.6 + ggg / yig(46)/pt(3.5 ) & 306.6 + ggg / yig(69)/pt(2.7 ) & 453.6 + ggg / yig(58)/pt(2.2 ) & 761.7 + ggg / yig(57)/pt(1.3 ) & 1089.9 + ggg / yig(61)/pt(11.1 ) & 334.5 + ggg / yig(52)/pt(16.9 ) & 339.2 + ggg / yig(53)/pt(8.5 ) & 348.3 + yag / yig(59)/pt(6.8 ) & 487.7 + yag / yig(64)/pt(3 ) & 622.2 + yag / yig(61)/pt(19.5 ) & 361.3 + yag / yig(63)/pt(6.5 ) & 412 + yag / yig(60)/pt(9.7 ) & 429 + yag / yig(60)/pt(12.8 ) & 434.9 + yag / yig(50)/pt(3 ) & 513 + lc sample & @xmath37 ( n@xmath116 m ) + ggg / yig(31)/cu(8.8)/pt(7.3 ) & 410 + ggg / yig(20)/au(7)/pt(7 ) & 400 + ggg / yig(20)/cu(9)/pt(7 ) & 400 + yag / yig(55)/au(9.2)/pt(9 ) & 370 + yag / yig(45)/au(9.4)/pt(2.9 ) & 860 + the trilayer spin pumping and spin seebeck data are evaluated by using @xmath117 instead of @xmath82 in eqs .
( 6 ) and ( 8) in the main text to account for the additional electrical shortcircuit due to x. for the spin seebeck effect , we calculate @xmath18 as the magnon - electron temperature difference between the yig and the adjacent normal metal for the trilayer samples . for the smr trilayer data , we again assume that only pt contributes to the smr , while the second metallic layer acts as a mere electrical short - circuit .
we extract the charge current in pt @xmath118 by assuming that pt and x layers act as parallel resistors for the charge current .
we then derive the smr in pt @xmath119 by assuming there is no change in @xmath120 due to the spin hall magnetoresistance .
we use these expression to replace @xmath105 in eq .
( 4 ) and @xmath106 in eq . ( 9 ) in the main text to evaluate @xmath97 and @xmath43 for the trilayer samples .
h. nakayama , m. althammer , y .- t .
chen , k. uchida , y. kajiwara , d. kikuchi , t. ohtani , s. geprgs , m. opel , s. takahashi , r. gross , g. e. w. bauer , s. t. b. goennenwein , and e. saitoh , phys .
lett . * 110 * , 206601 ( 2013 ) .
m. althammer , s. meyer , h. nakayama , m. schreier , s. altmannshofer , m. weiler , h. huebl , s. geprgs , m. opel , r. gross , d. meier , c. klewe , t. kuschel , j .-
schmalhorst , g. reiss , l. shen , a. gupta , y .- t .
chen , g. e. w. bauer , e. saitoh , and s. t. b. goennenwein , phys .
b * 87 * , 224401 ( 2013 ) .
f. d. czeschka , l. dreher , m. s. brandt , m. weiler , m. althammer , i .-
imort , g. reiss , a. thomas , w. schoch , w. limmer , h. huebl , r. gross , and s. t. b. goennenwein , phys . rev . lett .
* 107 * , 046601 ( 2011 ) . | we perform a quantitative , comparative study of the spin pumping , spin seebeck and spin hall magnetoresistance effects , all detected via the inverse spin hall effect in a series of over 20 yttrium iron garnet / pt samples .
our experimental results fully support present , exclusively spin current - based , theoretical models using a single set of plausible parameters for spin mixing conductance , spin hall angle and spin diffusion length .
our findings establish the purely spintronic nature of the aforementioned effects and provide a quantitative description in particular of the spin seebeck effect .
pure spin currents present a new paradigm in spintronics @xcite and spin caloritronics @xcite .
in particular , spin currents are the origin of spin pumping @xcite , the spin seebeck effect @xcite and the spin hall magnetoresistance ( smr ) @xcite .
taken alone , all these effects have been extensively studied , both experimentally @xcite and theoretically @xcite . from a theoretical point of view , all these effects are governed by the generation of a current of angular momentum via a non - equilibrium process .
the flow of this spin current across a ferromagnet / normal metal interface can then be detected .
the relevant interface property that determines the spin current transport thereby is the spin mixing conductance .
nevertheless , there has been an ongoing debate regarding the physical origin of the measurement data acquired in spin seebeck and smr experiments due to possible contamination with anomalous nernst effect @xcite or anisotropic magnetoresistance @xcite caused by static proximity polarization of the normal metal @xcite . to settle this issue ,
a rigorous check of the consistency of the spin - current based physical models across all three effects is needed . if possible contamination effects are absent , according to the spin mixing conductance concept @xcite , there should exist a generalized ohm s law between the interfacial spin current and the energy associated with the corresponding non - equilibrium process .
this relation should invariably hold for the spin pumping , spin seebeck and spin hall magnetoresistance effects , as they are all based on the generation and detection of interfacial , non - equilibrium spin currents .
we here put forward heuristic arguments that are strongly supported by experimental evidence for a scaling law that links all aforementioned spin(calori)tronic effects on a fundamental level and allows to trace back their origin to pure spin currents . in the ferromagnet ( yig in our experiments )
is excited either resonantly ( a ) or thermally ( b ) .
the @xmath0 precession around @xmath1 is damped via the emission of a spin current @xmath2 with polarization @xmath3 into the normal metal ( pt in our experiments ) .
( c ) the spin hall magnetoresistance is due to the torque exerted on @xmath0 by an appropriately polarized @xmath2 which yields a change in the reflected spin current @xmath4 .
the interconversion between @xmath2 ( @xmath4 ) and the charge currents @xmath5 ( @xmath6 ) are due to the ( inverse ) spin hall effect in the normal metal.,title="fig : " ] + we carried out a systematic set of spin pumping , spin seebeck and smr experiments on y@xmath7fe@xmath8o@xmath9 ( yig ) / pt thin film bilayers . in our spin
pumping experiments [ schematically depicted in fig .
[ fig : schematics](a ) ] , we place yig / pt bilayers in a microwave cavity operated at @xmath10 to resonantly excite magnetization dynamics .
the emission of a spin current density @xmath2 across the bilayer interface into the pt provides a damping channel for the non - equilibrium excitations of the magnetization @xmath0 .
it has been established that the magnitude of the dc spin current density is given by @xcite @xmath11 where @xmath12 is the frequency of the microwave , @xmath13 is the cone angle which the precessing magnetization @xmath0 encloses with the effective magnetic field @xmath1 , @xmath14 is the planck constant , @xmath15 is a factor to correct for elliptical precession of @xmath0 @xcite and @xmath16is the spin mixing conductance per unit of interface area and the conductance quantum @xmath17 . as shown in fig .
[ fig : schematics](b ) , thermal excitations of @xmath0 also give rise to a spin current .
this is the so - called spin seebeck effect @xcite .
given a temperature difference @xmath18 between the electrons in the normal metal and the magnons in the ferromagnet , a dc spin current density @xcite @xmath19 is generated .
we investigate the longitudinal spin seebeck effect @xcite , where the temperature gradient is applied across the ferromagnetic insulator / normal metal interface . in eq . , @xmath20 is the gyromagnetic ratio with the effective g - factor @xmath21 and the bohr magneton @xmath22 , @xmath23 is the saturation magnetization and @xmath24 is the magnetic coherence volume given by @xcite @xmath25 with the riemann zeta function @xmath26 , the spin wave stiffness @xmath27 and @xmath28 for our room temperature experiments . as depicted in fig .
[ fig : schematics](c ) , the application of a dc charge current density @xmath5 furthermore allows to inject a dc spin current density direction vector @xmath29 into the yig via the spin hall effect in pt @xcite . here , @xmath30 is the spin hall angle of pt and @xmath3 is the spin current polarization .
if the magnetization @xmath0 of the ferromagnet is oriented perpendicular to @xmath3 , @xmath2 can exert a torque on @xmath0 by being absorbed at the interface .
when @xmath3 is parallel to @xmath0 , the spin current is reflected at the interface , causing a spin current @xmath4 . due to the inverse spin hall effect ,
@xmath4 again generates a charge current density @xmath31 that effectively changes the electrical resistance of the pt film .
the net spin current density @xmath32 for @xmath33 is given by @xcite @xmath34 where @xmath35 is the elementary charge , @xmath36 is the spin diffusion length in pt and @xmath37 and @xmath38 are pt resistivity and thickness , respectively .
we furthermore introduced the correction factor @xcite @xmath39^{-1}\;.\ ] ] as suggested by eqs .
, and , one should thus observe a scaling @xmath40 with an appropriate energy @xmath41 , @xmath42 and @xmath43 that generates the spin pumping , spin seebeck and smr effects , respectively .
note that , due to the inclusion of spin backflow via the correction factor @xmath44 , the spin mixing conductance enters the linear response expression in a non - linear fashion by defining an effective excitation energy .
to experimentally test the scaling between @xmath45 and @xmath46 , we performed a series of spin pumping , spin seebeck and spin hall magnetoresistance measurements on more than 20 samples .
most of these samples consisted of thin film yig / pt bilayers with varying pt thickness . additionally , we used yig / x / pt trilayers in which x was a normal metal ( au or cu ) . a complete list of samples , details of their preparation and relevant material parameters can be found in the appendix .
is extracted in ferromagnetic resonance as indicated .
( b ) data from a spin seebeck experiment performed using a piece of the same sample .
a laser beam is used to generate the thermal perturbation ( see text ) and @xmath47 is obtained by taking half of the voltage difference observed between positive and negative saturation magnetic fields in the geometry sketched to the right .
( c ) dc magnetoresistance measurements are used to extract @xmath48 as the change in bilayer resistance upon rotating @xmath0 from parallel ( 0@xmath49 ) to perpendicular ( 90@xmath49 ) to @xmath5 .
, title="fig : " ] + in spin pumping experiments with electrical spin current detection via the inverse spin hall effect @xcite , it is possible to determine @xmath50from the recorded dc voltage @xmath51 in ferromagnetic resonance ( fmr ) as @xcite @xmath52 with the sample length @xmath53 and the open - circuit spin hall conversion efficiency @xcite @xmath54 with @xmath55 .
the factor @xmath44 is again included to correct for spin backflow is not a backflow correction but is only obtained in a method similar to the backflow correction for spin pumping ] .
note that eq . , is strictly valid only for @xmath56 with the spin flip time @xmath57 in pt @xcite .
@xmath44 is thus exact for the dc smr and a good approximation for our spin pumping data ( @xmath58 ) .
a typical experimental @xmath59 trace recorded for a yig ( 20 nm)/ pt ( 7 nm ) sample at a fixed microwave frequency @xmath10 while sweeping the external magnetic field @xmath60 is shown together with a schematic of the sample in fig .
[ fig : exp](a ) .
we observe a resonant lorentzian line shape of @xmath59 as a function of the external magnetic field @xcite . within experimental error , @xmath61 far away from fmr and @xmath51 is the dc voltage recorded at the resonance magnetic field as indicated in fig .
[ fig : exp](a ) . because yig is a ferrimagnetic insulator and we took great care to position the sample in a node of the microwave electric field , @xmath51 is not contaminated with rectification voltages @xcite .
this is supported by the purely symmetric lorentzian resonance line shapes observed .
the sample length @xmath53 ranged from @xmath62 to @xmath63 in the different samples investigated .
we determined @xmath37 from four point resistance measurements and @xmath38 from x - ray reflectometry .
we now turn to the evaluation of @xmath64 [ cf .
eq . ] .
to this end , we extract @xmath65 @xcite from the fwhm line width of the @xmath51 traces , where @xmath66 is the circular microwave magnetic field that was determined from paramagnetic resonance calibration .
we find @xmath67 in the different samples investigated .
we calculate the ellipticity correction factor @xmath68 as detailed in refs .
@xcite using a saturation magnetization @xmath69 and an effective g factor @xmath70 @xcite .
we are now able to evaluate @xmath50and @xmath41 as a function of the three parameters @xmath71 , @xmath30 and @xmath36 .
we discuss below that with a single set of these parameters we can quantitatively describe the spin pumping , spin seebeck and smr data in the context of the spin mixing conductance concept . in a different set of experiments , using parts of the same samples patterned into hall bar mesas by optical lithography and subsequent ar - ion etching , we determined the dc voltage @xmath47 due to the laser - heating induced spin seebeck effect @xcite .
a laser beam of adjustable power ( @xmath72 ) impinges on the main hall bar ( length @xmath73 and width @xmath74 ) which is oriented perpendicular to the external , in - plane magnetic field .
the laser beam is dominantly absorbed in the pt layer and yields a temperature difference @xmath18 between the magnons in yig and the electrons in the pt at the yig / pt interface .
we use a numerical model incorporating a thermal contact resistance between the yig and the normal metal layers to compute the magnon , phonon and electron temperature profiles in our samples as a function of layer composition and laser power @xcite .
we find @xmath75 .
the spin current @xmath76is detected via the inverse spin hall voltage @xmath77 along the main hall bar .
a typical @xmath77 curve is shown in fig .
[ fig : exp](b ) as a function of the external magnetic field .
the depicted hysteretic @xmath77 vs. @xmath60 loop is consistent with our previous experiments @xcite .
the spin current density is extracted from experiment by @xmath78 where @xmath53 from eq .
is now replaced by @xmath79 with the laser spot radius @xmath80 and the hall bar width @xmath74 .
this stems from lateral integration over the gaussian laser spot profile to account for the fact that the sample is heated only locally as demonstrated in ref .
@xcite and is valid as long as @xmath81 , which is the case for all investigated samples .
we use the identical values for @xmath82 and @xmath44 as for the spin pumping experiments to calculate @xmath83 with eq . .
to quantify @xmath84 , we use the coherence volume @xmath85 which we obtain from eq . by using @xmath86
consistent with theory and a broad range of experiments @xcite . in yet further , independent experiments with the same set of hall - bar samples we measured the smr as the change @xmath87 in bilayer resistance
@xmath88 when rotating @xmath33 ( @xmath89 ) to @xmath90 ( @xmath91 ) .
typical @xmath92 traces for @xmath33 ( 0@xmath49 ) and @xmath93 are shown in figure [ fig : exp](c ) .
one can observe the magnetization switching at the coercive magnetic fields that agree with those extracted from our spin seebeck experiments on identical samples .
as detailed in ref .
@xcite , the change in @xmath88 with @xmath60 can be traced back to the smr from its characteristic dependence on @xmath0 orientation . from the smr data , we extract the spin current density @xcite @xmath94 from the experimentally determined @xmath48 .
the charge current densities in our experiments were @xmath95 .
as a function of the non - equilibrium energy @xmath46 for all investigated samples as determined in spin pumping ( circles ) , spin seebeck ( triangles ) and smr ( squares ) measurements . the solid line is the proportionality constant identified in the text as @xmath96 .
open symbols correspond to yig / au / pt trilayer samples and half - filled symbols to yig / cu / pt trilayer samples .
( b ) the spin mixing conductance as a function of total normal metal thickness [ same symbols as in ( a)].,title="fig : " ] + we plot @xmath97 from eq . as a function of @xmath98 in fig .
[ fig : scaling](a ) ( squares ) for all samples .
in identical fashion , fig .
[ fig : scaling](a ) depicts @xmath99 as a function of @xmath41 ( circles ) and @xmath100 as a function of @xmath42 ( up triangles ) .
we use a single set of parameters , @xmath101 , @xmath102 and @xmath103 for all samples .
these parameters are identical to those extracted from an analysis of the pt thickness dependence of the smr @xcite .
we acquired data points for smr and spin seebeck effect on various samples as a function of charge current density or laser power , respectively .
we furthermore include data recorded using yig / au / pt ( open symbols ) and yig / cu / pt ( half - filled symbols ) trilayer samples ( symbol shape identifies spin pumping , spin seebeck or smr data ) . to evaluate the trilayer data , we assume vanishing @xmath30 and @xmath104 for au and cu . we thus modify @xmath82 for spin pumping and spin seebeck effect as well as @xmath105 and @xmath106 for the smr as detailed in the appendix .
altogether our experimental data span five orders of magnitude in @xmath45 and @xmath46 . in this entire range , we observe that all experimental data points fall on ( or close to ) one line in the plot .
as predicted by theory , the constant of proportionality is found to be @xmath96 .
this has several implications .
first , and most importantly , fig .
[ fig : scaling](a ) is strong evidence for the spin mixing conductance concept , i.e. , that spin pumping , spin seebeck and smr effects indeed arise from pure spin currents physics .
spurious effects due to static proximity polarization in pt @xcite can be excluded based on fig .
[ fig : scaling](a ) , because if the measured spin seebeck effect and spin hall magnetoresistance data were indeed consequence of a static proximity polarization , the obtained data should not fit quantitatively in the spin current injection picture parameterized by the mixing conductance .
our data for both these effects however are consistent with those extracted from spin pumping experiments , showing that all three effects have one and the same microscopic origin , namely spin current flow .
last but not least , our data enable a quantitative understanding of the spin seebeck effect which has remained elusive so far @xcite .
the scaling of our data shows that a spin hall angle @xmath102 and spin diffusion length @xmath103 are reasonable material parameters for our pt thin films .
these parameters are in accordance with the product @xmath107 extracted in previous studies @xcite and more recent findings @xcite . in our work , owing to the different functional dependence of the three effects on these parameters , a more reliable extraction of @xmath36 and @xmath30 becomes possible .
figure [ fig : scaling](b ) shows @xmath108 as a function of the total normal metal thickness @xmath109 ( x is au or cu ) for all bilayer and trilayer samples .
the symbol definitions are identical to that in fig .
[ fig : scaling](a ) .
the solid line depicts @xmath101 and the shaded region corresponds to @xmath110 . the majority of our data points lie within the shaded region , so @xmath16is constant within @xmath111 for all our samples and regardless of the experimental method used to extract it .
there is no discernible trend in the @xmath45 to @xmath46 ratio as a function of pt ( or x / pt ) thickness .
this suggests that eqs .
, and are sufficiently accurate in the entire thickness range investigated . the unsystematic scatter in @xmath16 in fig .
[ fig : scaling](b ) can be accounted for as being experimental errors and varying interface properties between the different samples .
our trilayer samples exhibit a @xmath16similar or slightly lower than that of our yig / pt bilayers and previous findings for the yig / au interface @xcite . in summary
, we have experimentally demonstrated that spin pumping , spin seebeck and smr all share the same purely spintronic origin and thus experimentally validated the spin mixing conductance concept .
spurious contributions due to proximity ferromagnetism in pt can be ruled out , thereby supporting existing models for smr and spin seebeck effect .
a relevant set of parameters for a ferromagnetic insulator / normal metal bilayer ( or according trilayer ) obtained from rather straightforward smr experiments may be used to predict results for spin seebeck or spin pumping experiments on the same samples . financial support from the dfg via spp 1538 spin caloric transport `` , project no .
go 944/4 - 1 , the dutch fom foundation , ec project
macalo '' , the national natural science foundation of china ( no . 11004036 , no .
91121002 ) and the german excellence initiative via the nanosystems initiative munich " ( nim ) is gratefully acknowledged . |
30 chakravarty s. and schmid a. 1986 _ phys . rep . _ * 140 * 193 .
beenakker c.w.j .
1997 _ rev .
phys . _ * 69 * 731 .
golubev d.s . and zaikin a.d .
2001 _ phys .
lett . _ * 86 * 4887 .
bagrets d.a . and nazarov yu.v
. 2005 _ phys .
lett . _ * 94 * 056801 .
golubev d.s . and zaikin a.d . 2004 _ phys .
_ b * 69 * 075318 .
brouwer p.w .
, lamacraft a. , and flensberg k. 2005 _ phys .
_ b * 72 * 075316 .
schn g. and zaikin a.d .
1990 _ phys . rep .
_ * 198 * 237 .
taniguchi n. , simons b.d . , and altshuler b.l .
1996 _ phys .
_ b * 53 * 7618 .
golubev d.s . and
. 2006 _ phys .
_ b * 74 * 245329 .
ingold g.l . and nazarov yu.v
. 1992 _ single charge tunneling _ , ( plenum press , new york ) _ nato asi series _
b * 294 * , p. 21 .
golubev d.s . and
zaikin a.d .
1998 _ phys .
lett . _ * 81 * 1074 ; 1999 _ phys .
b * 59 * 9195 .
golubev d.s . and zaikin a.d . 2000 _ quantum physics at mesoscopic scale _ ( edp sciences , 2000 ) , p. 491 ; cond - mat/9907497 .
pivin d.p .
_ 1999 _ phys .
lett . _ * 82 * 4687 .
huibers a.g .
_ 1999 _ phys .
lett . _ * 83 * 5090 .
hackens b. _ et al .
_ 2005 _ phys .
lett . _ * 94 * 146802 .
our results derived for a double dot system can also at least qualitatively apply to a single dot provided fluctuating voltage is not strictly uniform inside the cavity .
somewhat more accurate treatment of the coulomb blockade correction amounts to expanding the exact result only in @xmath82 .
then one again arrives at eq .
( [ wb ] ) with @xmath169 .
one can also demonstrate that at @xmath170 the coulomb blockade correction saturates . for @xmath171
this correction remains unimportant down to exponentially small energies below which weak coulomb blockade turns into strong at any non - zero @xmath14 .
aleiner i.l .
, altshuler b.l . , and gershenson m.e .
1999 _ waves random media _ * 9 * 201 ; golubev d.s . and zaikin a.d .
2000 _ phys .
b _ * 62 * 14061 ; golubev d.s . , zaikin a.d . , and schn g. 2002 _ j. low temp . phys . _
* 126 * 1355 ; aleiner i.l .
, altshuler b.l . , and vavilov m.g
. 2002 _ j. low temp .
* 126 * 1377 ; von delft j. cond - mat/0510563 ; golubev d.s . and zaikin a.d . 2003 _ j. low temp
. phys . _ * 132 * 11 ; cond - mat/0512411 ; saminadayar l. _ et al .
_ 2007 _ physica e _ * 40 * 12 .
golubev d.s . and
zaikin a.d .
2007 _ physica e _ * 40 * 32 . | we non - perturbatively analyze the effect of electron - electron interactions on weak localization ( wl ) in relatively short metallic conductors with a tunnel barrier .
we demonstrate that the main effect of interactions is electron dephasing which persists down to @xmath0 and yields suppression of wl correction to conductance below its non - interacting value .
our results may account for recent observations of low temperature saturation of the electron decoherence time in quantum dots .
electrons propagating in a disordered conductor get scattered and interfere .
this quantum interference is possible only as long as the electron wave functions remain coherent . in any realistic situation , however , interactions between electrons and with other degrees of freedom may limit phase coherence and , hence , reduce electrons ability to interfere . the interplay between scattering , quantum coherence and interactions
yields a rich variety of non - trivial effects and significantly impacts electron transport in disordered conductors .
the so - called weak localization ( wl ) correction to the conductance of a disordered system @xmath1 is most sensitive to electron coherence and is known to arise from interference of pairs of time - reversed electron paths @xcite . in a system of two scatterers separated by a cavity ( quantum dot ) and in the absence of interactions this correction can be directly evaluated @xcite .
the effect of electron - electron interactions can be described in terms of fluctuating voltages . provided the voltage drops only across the barriers and not inside the cavity electron - electron interactions yield energy dependent logarithmic renormalization of the dot channel transmissions @xcite but
_ do not _ cause any dephasing @xcite .
the latter result can easily be understood if one observes that the voltage - dependent random phase acquired by the electron wave function @xmath2 along any path turns out to be the same as that for its time - reversed counterpart .
hence , in the product @xmath3 these random phases cancel each other exactly and quantum coherence of electrons remains preserved .
it is important , however , that this cancellation occurs only in the case of two scatterers , whereas in a system of three or more scatterers the situation is entirely different .
consider , e.g. , a system of two quantum dots depicted in fig . 1 and again assume that fluctuating voltages are concentrated at the barriers .
the phase factor accumulated along the path ( see fig .
1 ) which crosses the central barrier twice ( at times @xmath4 and @xmath5 ) and returns to the initial point ( at a time @xmath6 ) is @xmath7}$ ] , where @xmath8 is the fluctuating voltage across the central barrier .
similarly , the phase factor picked up along the time - reversed path reads @xmath9}$ ] .
hence , the overall phase factor acquired by the product @xmath3 for a pair of time - reversed paths is @xmath10 , where @xmath11 .
averaging over phase fluctuations , which for simplicity are assumed gaussian , we obtain @xmath12 where we defined the phase correlation function @xmath13 should this function grow with time the electron phase coherence decays and @xmath1 gets suppressed below its non - interacting value . via an ohmic shunt resistor @xmath14.,width=321 ] the above arguments are not specific to the system of three barriers and after proper generalization can be applied to virtually any disordered conductor . at the same time , these arguments are not yet sufficient to quantitatively describe the decoherence effect of electron - electron interactions for two important reasons : ( i ) fluctuating voltages are treated as external ( classical ) fields rather than quantum fields produced internally by fluctuating electrons and ( ii ) fermi statistics is not yet accounted for . below we will cure both these problems and non - perturbatively evaluate wl correction @xmath1 for a metallic system with a tunnel barrier and ( at least ) two more scatterers in the presence of electron - electron interactions which turn out to reduce phase coherence of electrons at any temperature down to @xmath0 .
we will consider a system with a tunnel barrier with dimensionless conductance @xmath15 , which separates two sufficiently short disordered metallic conductors with thouless energies @xmath16 and dimensionless conductances @xmath17 .
this system is described by the hamiltonian @xmath18 where @xmath19 is the hamiltonian of the left ( right ) lead , @xmath20 is the single electron hamiltonian in the left ( right ) lead , and @xmath21 $ ] is the tunnel hamiltonian . here
@xmath22 is the phase operator , which is related to the voltage drop across the junction @xmath23 , and the @xmath24-integration runs over the junction area .
finally , @xmath25 is the quadratic hamiltonian of electromagnetic fields , the precise form of which depends on the circuit configuration and will not be specified here .
following the standard procedure we integrate out fermionic degrees of freedom and arrive at the effective action @xmath26 $ ] , where @xmath27 is the green - keldysh function for our system . expanding the action in powers of the tunnel hamiltonian we obtain @xmath28 , where the term @xmath29 describes the action of the left and right conductors , @xmath30 is ambegaokar - eckern - schn action @xcite and @xmath31 here @xmath32 , @xmath33 is the phase variable on the forward ( backward ) branch of the keldysh contour , @xmath34 are @xmath35 matrix green - keldysh functions in the left and right conductors ( @xmath36 ) and we use the convention @xmath37 , @xmath38 .
we assume that @xmath39 have the equilibrium form @xmath40 , where @xmath41 are retarded and advanced green functions , @xmath42 @xmath43 @xmath44 is the fermi function and @xmath45
. our next step amounts to averaging the products of retarded and advanced propagators in the action ( [ s2 ] ) over disorder in each conductor separately .
we have ( see , e.g. , @xcite ) @xmath46 where @xmath47 , @xmath48 and @xmath49 are respectively the density of states , the diffuson and the cooperon in the left conductor , @xmath50 , @xmath51 and @xmath52 are respectively the fermi wave vector and elastic mean free path .
the same averaging procedure applies to the right conductor .
finally we assume that the transmission amplitude @xmath53 is random , quickly oscillating real function .
averaging over these oscillations yields @xmath54 , where @xmath55 is the local conductance of the barrier .
after all these steps eq .
( [ s2 ] ) reduces to a sum of different terms . here
we will select only the terms responsible for weak localization which involve the product of two cooperons @xmath56 and @xmath57 . collecting all such contributions we obtain @xmath58 \nonumber\\ & & \times\ , \left [ h(\tau_2-t_3)e^{i\frac{\varphi^-(t_3)}{2}}f(t_1+t_3-t_4-\tau_1 ) \right . \nonumber\\ & & \left .
-\ , f(\tau_2-t_3 ) e^{-i\frac{\varphi^-(t_3)}{2}}h(t_1+t_3-t_4-\tau_1 ) \right ] \nonumber\\ & & \times\ , \left [ e^{-i\frac{\varphi^-(t_4)}{2}}f(-t_1+t_2+t_4-\tau_2 ) + \ , e^{i\frac{\varphi^-(t_4)}{2}}h(-t_1+t_2+t_4-\tau_2)\right ] \nonumber\\ & & + \,\big\ { l\leftrightarrow r ,
\varphi^\pm\to -\varphi^\pm\big\}. \label{swl}\end{aligned}\ ] ] here we defined `` classical '' @xmath59 and `` quantum '' @xmath60 phases and introduced @xmath61 and @xmath62 . the action ( [ swl ] ) fully accounts for the effects of electron - electron interactions on wl via the fluctuating phases @xmath63 . in order to find the wl correction to
the current across the central barrier @xmath64 we make use of the following general formula @xmath65}{\delta \varphi^-}\ , e^{is_{l , r}+is^{(1)}_{\rm t}}. \label{i}\ ] ] in the limit @xmath66 this integral remains gaussian in @xmath63 at all relevant energies and can easily be performed .
the effective expansion parameter in this case is @xmath67 . combining eqs .
( [ swl ] ) and ( [ i ] ) and introducing the average voltage at the barrier @xmath68 we find @xmath69 -\,\big\ { v\to -v \big\}. \label{iwl1}\end{aligned}\ ] ] here @xmath70 and @xmath71 ( @xmath72 ) are the fourier transforms of respectively the cooperons @xmath73 and the functions @xmath74{\cal q}_{j}(t_1,t_2,t_3 ) , \label{calp}\ ] ] where @xmath75 and @xmath76 coincides with the phase correlation function ( [ f ] ) .
the terms @xmath77 read @xmath78 } \nonumber\\ & & \times\ , \big\{2e^{i\left[k(|t_1+t_2+t_3|)+k(t_1+t_3)+k(t_1+t_2)\right ] } -\,e^{i\left[k(t_1+t_2+t_3)+k(|t_1+t_3|)+k(|t_1+t_2|)\right ] } \big\ } , \nonumber\\ & & { \cal q}_2=e^{i\left[k(|t_1+t_2+t_3|)-k(t_2)-k(|t_3|)\right ] } e^{i\left[k(t_1+t_3)-k(|t_1+t_2|)-k(t_3-t_2)\right ] } , \nonumber\\ & & { \cal q}_3= e^{i\left[k(t_1+t_2+t_3)-k(|t_2|)-k(|t_3|)\right ] } e^{-i\left[k(t_1+t_3)+k(t_1+t_2)-k(|t_3-t_2|)\right ] } , \label{pj}\end{aligned}\ ] ] where @xmath79 \right\rangle$ ] is the response function . eqs .
( [ iwl1])-([pj ] ) represent the central result of our paper .
they fully determine wl correction to the current in our system .
the non - interacting result is reproduced by the first two lines of eq .
( [ iwl1 ] ) before the square brackets , while the terms in the square brackets exactly account for the effect of interactions .
the same result follows from the non - linear @xmath80-model approach @xcite .
our result demonstrates that the whole effect of electron - electron interactions is encoded in two different correlators of fluctuating phases @xmath81 and @xmath82 .
these correlation functions are well familiar from the so - called @xmath83-theory @xcite .
they read @xmath84\frac{1-\cos\omega t}{\omega^2 } , \label{fff}\\ k(t)&=&e^2\int\frac{d\omega}{2\pi}\,{\rm re}\big[z(\omega)\big ] \frac{\sin\omega t}{\omega } , \label{kkk}\end{aligned}\ ] ] where @xmath85 is an effective impedance `` seen '' by the central barrier .
both functions ( [ fff ] ) and ( [ kkk ] ) are purely real and , hence , @xmath86 . at times
@xmath87 ( an effective @xmath88-time @xmath89 will be defined later ) we obtain @xmath90 and @xmath91 , where @xmath92 , @xmath93 and @xmath94 is the euler constant .
we observe that while @xmath81 grows with time at any temperature including @xmath0 , the function @xmath82 always remains small in the limit @xmath95 considered here .
hence , the combination ( [ calf ] ) should be fully kept in the exponent of ( [ calp ] ) while the correlator @xmath82 can be safely ignored in the leading order in @xmath96 .
then all @xmath97 , the fermi function @xmath44 drops out from the result and we get @xmath98 , where @xmath99 identifying @xmath100 and @xmath101 we observe that the exponent in the third line of eq .
( [ gwlf ] ) exactly coincides with the expression ( [ phase ] ) derived from simple considerations involving electrons propagating along time - reversed paths in an external fluctuating field .
thus , in the leading order in @xmath96 the wl correction @xmath1 is affected by electron - electron interactions via dephasing produced _ only _ by the `` classical '' component @xmath102 of the fluctuating field which mediates such interactions .
fluctuations of the `` quantum '' field @xmath103 turn out to be irrelevant for dephasing and may only cause a ( weak ) coulomb blockade correction to be considered below .
it is worthwhile to point out that a similar conclusion was previously reached for spatially extended disordered conductors within a different approach @xcite .
we also note that a close relation between the results @xcite and the @xmath83-theory @xcite was already demonstrated earlier @xcite .
our present results make this relation even more transparent .
our further calculation is concentrated on a system of two ( identical ) dots depicted in fig .
1 . for simplicity
the outer barriers are supposed to be open , @xmath104 and @xmath105 .
then the cooperons take a simple form @xmath106 , where @xmath107 and @xmath108 are respectively the dot volume and dwell time .
we also define the effective impedance seen by the central tunnel junction @xmath109 with the real part @xmath110 , \label{imped}\end{aligned}\ ] ] where @xmath111 , @xmath112 , @xmath113 and @xmath114 , @xmath115 and @xmath116 are the capacitances of respectively left ( right ) barriers , the central junction and the gate electrode . substituting the cooperons @xmath117 and the correlator @xmath81 ( [ fff ] ) , ( [ imped ] ) into eq .
( [ gwlf ] ) we observe that contribution of @xmath118 in eq .
( [ imped ] ) drops out . performing the time integrals we arrive at the final expression for the wl correction @xmath119 in the presence of electron - electron interactions : @xmath120 where @xmath121 is the dot mean level spacing .
this result is plotted in fig .
2a demonstrating that interactions suppress @xmath119 below its non - interacting value @xcite @xmath122 . [
cols="^,^ " , ] let us define @xmath123 and consider the limit of metallic dots @xmath124 . at @xmath125
the wl correction saturates to @xmath126 whereas at @xmath127 and for @xmath128 we find @xmath129 let us phenomenologically define the electron decoherence time @xmath130 by taking the cooperons in the form @xmath131 which yields @xcite @xmath132 . resolving this equation for @xmath130
we obtain @xmath133 which yields @xmath134 for @xmath125 and @xmath135 .
eqs .
( [ t0])-([tphi ] ) , although not directly applicable to a single quantum dot , account for key features of the dependence @xmath136 ( fig .
2b ) observed in various ucf experiments @xcite with quantum dots @xcite . at higher temperatures we find @xmath137 with non - universal @xmath138-dependent power @xmath139 , while at lower @xmath140 the electron decoherence time @xmath141 saturates to a constant in agreement with the observations @xcite .
it was pointed out @xcite that the available experimental values of @xmath142 scale as @xmath143 for a variety of dot sizes and dwell times @xmath108 varying by @xmath144 decades .
our result ( [ t0],[tphi ] ) should be consistent with this scaling provided at low @xmath145 the right - hand side of eq .
( [ tphi ] ) remains of order one .
note that the phenomenological definition of @xmath130 is identical to that used before in ref .
@xcite where we also demonstrated that for an arbitrary array of quantum dots our expression for the weak localization correction determines the system magnetoconductance if we substitute @xmath146 , where @xmath147 is the electron dephasing time due to the external magnetic field @xmath148 .
thus , our definition of @xmath130 is fully consistent with the standard procedure of extracting the electron dephasing time from the magnetoconductance curves .
furthermore , it is straightforward to demonstrate @xcite that , e.g. , in the case of quasi-1d arrays of quantum dots our definition for @xmath130 just yields the standard result for the magnetoconductance of a diffusive wire , cf .
eq .
( 60 ) of ref .
@xcite .
we also would like to emphasize that there exists no contradiction between the definition of @xmath130 adopted here and the fact that no dephasing occurs for electon paths confined within a single quantum dot , as discussed in the beginning of our manuscript .
as it was demonstrated , electron dephasing occurs as soon as time - reversed paths cross the central barrier twice and return to the initial point inside the dot ( see fig .
1 ) . in the presence of fluctuating electromagnetic potentials ( dropping across the central barrier ) the forward path and
its time - reversed counterpart pick up different random phases . after averaging over both fluctuating fields and electron paths
one arrives at a decaying in time contribution to the cooperons which is just captured by our phenomenological definition .
of course , other definitions of @xmath130 can also be employed .
however , our basic conclusion about non - vanishing electron dephasing by electron - electron interactions down to @xmath149 will not be sensitive to any particular definition of @xmath130 , since this conclusion is based on the result ( [ gwlf ] ) demonstrating the interaction - induced suppression of the wl correction to conductance ( as well as of the magnetoconductance , cf .
ref .
@xcite ) at any temperature including @xmath0 .
the basic physics behind this result is exactly the same as that already elucidated by the well known @xmath83-theory @xcite : tunneling electrons can exchange energies with an effective electromagnetic environment .
this process results in broadening of the distribution function for such electrons even at @xmath0 which inevitably yields electron dephasing . finally , it is instructive to establish the relation to the ordinary perturbation theory in the interaction which is reproduced by formally expanding our exact result ( [ iwl1 ] ) to the first order in @xmath85 .
we obtain @xmath150 where @xmath151 , \label{wlf}\end{aligned}\ ] ] @xmath152 \nonumber\\ & & + \ , i\,{\rm im}\ , z(\omega)\ ; c_l(\omega)c_r(\omega ) \big\}. \label{wlk}\end{aligned}\ ] ] here we defined the function @xmath153 and fourier transformed cooperons @xmath154 .
the two terms @xmath155 and @xmath156 are linear in respectively @xmath81 and @xmath82 .
exactly the same results ( [ pert])-([wlk ] ) are reproduced from the first order diagrammatic perturbation theory in the interaction . in order to observe the equivalence of the two approaches one should keep in mind that @xmath81 is proportional to the keldysh component of the photon green function , while @xmath82 is proportional to the retarded photon green function .
one should also remember that the photon green function in our model is coordinate independent in both quantum dots .
one can actually demonstrate that the terms @xmath157 @xmath158 come from the so - called `` self - energy '' diagrams , while the terms @xmath159 emerge from the `` vertex '' diagrams .
the term @xmath156 represents the coulomb blockade correction to @xmath160 and is entirely different from the dephasing term @xmath155 .
in contrast to the latter , the term @xmath156 is non - linear in @xmath68 describing the standard coulomb offset at large @xmath68 and turning into @xmath161 for @xmath162 in the linear in @xmath68 regime .
thus , the coulomb blockade correction remains small @xcite in the metallic limit @xmath163 .
we also note that @xmath156 involves the combination @xmath164 which enters _ only _ in the first order in the interaction . as in the case of
spatially extended conductors , at @xmath0 some terms contained in @xmath165 partially cancel similar contributions to @xmath166 .
this cancellation , however , remains incomplete and , as demonstrated by our exact result , by no means implies the absence of electron dephasing at @xmath167 .
more information on the debates on low temperature decoherence by electron - electron interactions can be obtained , e.g. , from refs . @xcite and further references therein . without going into details , we would only like to emphasize that our present manuscript does not make any use of the techniques introduced in our previous works on decoherence in disordered conductors and ,
hence , is formally independent on those . in summary ,
we have non - perturbatively treated the effect of electron - electron interactions on weak localization in relatively short metallic conductors .
the most significant effect of interactions is electron decoherence which persists down to @xmath0 and in agreement with experiments @xcite yields saturation of @xmath141 at @xmath140 .
the physics behind this effect is exactly the same as that discussed , e.g. , within the well known @xmath83-theory @xcite .
it is also worth pointing out that very recently @xcite we generalized our present approach to arbitrary arrays of quantum dots and derived the expression for @xmath168 which describes both weakly and strongly disordered conductors and quantitatively explains numerous experimental data available to date . in the case of weakly disordered conductors our results
@xcite match with those derived previously @xcite by means of a different technique .
this work was supported in part by the eu framework programme nmp4-ct-2003 - 505457 ultra-1d `` experimental and theoretical investigation of electron transport in ultra - narrow 1-dimensional nanostructures '' and by rfbr grant 06 - 02 - 17459 . |
imagine the following situation .
you are a manufacturer of antennas . in order to save power
, your antennas should communicate along a wedge - shape area , that is , an angular and practically infinite section of certain angle @xmath1 whose apex is the antenna .
the smaller the angle is the better it is in terms of power saving .
you are supposed to build many copies of these antennas to be used in various different communication networks . you know nothing about the future positioning of the antennas and you want them to be generic in the sense that they will fit to any possible finite set of locations .
when installed , each antenna may be directed to an arbitrary direction that -directional antennas and their ( connected ) communication graph.,width=264 ] will stay fixed forever .
therefore , you wish to find the minimum @xmath10 so that no matter what finite set @xmath11 of locations of the antennas is given , one can always install the antennas and direct them so that they can communicate with each other .
this is to say that the _ communication graph _ of the antennas should be a connected graph .
the communication graph is the graph whose vertex set is the set @xmath11 and two vertices ( antennas ) are connected by an edge if the corresponding two antennas can directly communicate with each other , that is , each is within the transmission - reception wedge of the other ( see figure [ fig : antennas ] for an illustration ) .
this problem was formulated by carmi , katz , lotker , and rosn @xcite , who also found the optimal @xmath1 which is @xmath2 ( a different model of a _ directed _ communication graph of directional antennas of bounded transmission range was studied in @xcite ) . in this paper
we provide a much simpler and more elegant proof of this result .
we also improve on the result in @xcite by obtaining a much simpler connected communication graph for @xmath2 whose diameter is at most @xmath12 .
our graph in fact consists of a path of length @xmath13 while every other vertex is connected by an edge to one of the three vertices of the path . in order to state our result and bring the proof we now formalize some of the notions above
. given two rays @xmath14 and @xmath15 with a common apex in the plane , we denote by @xmath16 the closed convex part of the plane bounded by @xmath14 and @xmath15 . for three noncollinear points in the plane @xmath17
we denote by @xmath18 the wedge @xmath19 , whose apex is the point @xmath20 .
@xmath18 is a wedge of angle @xmath21 .
let @xmath22 be @xmath23 wedges with pairwise distinct apexes .
the _ wedge - graph _ of @xmath22 is by definition the graph whose vertices correspond to the apexes @xmath24 of @xmath25 , respectively , where two apexes @xmath26 and @xmath27 are joined by an edge iff @xmath28 and @xmath29 . using this terminology
we wish to prove the following theorem whose first part was proved by carmi _
et al . _
@xcite .
[ theorem : main ] let @xmath11 be a set of @xmath23 points in general position in the plane and let @xmath30 be the number of vertices of the convex hull of @xmath11 .
one can always find in @xmath31-time @xmath23 wedges of angle @xmath32 whose apexes are the @xmath23 points of @xmath11 such that the wedge - graph with respect to these wedges is connected .
moreover , we can find wedges so that the wedge graph consists of a path of length @xmath13 and each of the other vertices in the graph is connected by an edge to one of the three vertices of the path .
the angle @xmath32 in theorem [ theorem : main ] is best possible , as shown in @xcite . indeed , for any @xmath33 one can not create a connected communication graph for a set of @xmath1-directional antennas that are located at the vertices of an equilateral triangle and on one of its edges .
we note that the result in theorem [ theorem : main ] is optimal in the sense that it is not always possible to find an assignment of wedges to the points so that the wedge graph consists of less than three vertices the union of neighbors of which is the entire set of vertices of the graph . to see this consider a set of points evenly distributed on a circle .
notice that if each wedge is of angle @xmath34 , then in any wedge graph each vertex is a neighbor of at most one third of the vertices .
our main tool in proving theorem [ theorem : main ] is a basic geometric lemma that we call the `` ice - cream lemma '' .
suppose that we put one scoop of ice - cream in a very large 2-dimensional cone , such that the ice - cream touches each side of the cone at a single point .
the distances from these points to the apex of the cone are not necessarily equal .
however , we show that there is always a way of putting the ice - cream in the cone such that they are equal . more formally , we prove : [ lemma : ice ] let @xmath4 be a compact convex set in the plane and fix @xmath5 .
there exist a point @xmath6 in the plane and two rays , @xmath14 and @xmath15 , emanating from @xmath6 and touching @xmath4 at two _ single points _
@xmath7 and @xmath8 , respectively , that satisfy @xmath9 and the angle bounded by @xmath35 and @xmath14 is @xmath1 .
see figure [ fig : ice - cream ] for an illustration .
the requirement in lemma [ lemma : ice ] that the two rays touch @xmath4 at _ single _ points will be crucial in our proof of theorem [ theorem : main ] . at a first glance
the statement in lemma [ lemma : ice ] probably looks very intuitive and it may seem that it should follow directly from a simple mean - value - theorem argument . however , this is not quite the case .
although the proof ( we will give two different proofs ) indeed uses a continuity argument it is not the most trivial one .
the reader is encouraged to spend few minutes trying to come up with a simple argument just to get the feeling of lemma [ lemma : ice ] before continuing further . because the proof of lemma [ lemma : ice ] is completely independent of theorem [ theorem : main ] and of its proof ,
we first show , in the next section , how to prove theorem [ theorem : main ] using lemma [ lemma : ice ] .
we postpone the two proofs of the ice - cream lemma to the last section .
let @xmath11 be a set of @xmath23 points in general position in the plane .
we denote the convex hull of @xmath11 by @xmath36 , and recall that it can be computed in @xmath37 time @xcite .
call two vertices of @xmath36 a _ good pair _ if there are a point @xmath6 and two rays @xmath38 emanating from it creating an angle of @xmath32 such that @xmath39 , @xmath9 , and @xmath40 .
note that given @xmath36 and two vertices of it @xmath41 , we can check in constant time whether @xmath41 is a good pair .
indeed there are exactly two points that form an equilateral triangle with @xmath41 , and for each of these two possible locations of @xmath6 , we only need to check whether the neighbors of @xmath7 and @xmath8 in @xmath36 lie in @xmath42 .
lemma [ lemma : ice ] guarantees that @xmath36 has a good pair .
next we describe an efficient way to find such a good pair .
suppose that @xmath41 is a good pair .
observe that there are two other rays that form an angle of measure @xmath43 , contain @xmath36 in their wedge , and such that one of them contains an edge of @xmath36 that is adjacent to @xmath7 or @xmath8 and the other ray contains the other point .
( these rays can be obtained by continuously rotating the ray through @xmath7 while forcing the ray through @xmath8 to form a @xmath43 angle with it , until an edge of @xmath36 is hit .
note that the distances from @xmath7 and @xmath8 to the apex of the new wedge are no longer equal . ) therefore , to find a good pair @xmath41 , it is enough to find for every edge @xmath44 of @xmath36 the ( at most four ) points @xmath8 such that @xmath8 is a vertex of @xmath36 and a line through @xmath8 that forms @xmath43 angle with the line through @xmath44 supports @xmath36 . for every such point , we can check in constant time whether @xmath41 or @xmath45 is a good pair . finding the points @xmath8 for an edge @xmath44 as above
can be done in @xmath46-time by a binary search on the vertices of @xmath36 .
thus , a good pair @xmath47 and the corresponding @xmath15 , @xmath14 , and @xmath6 can be found in @xmath31-time .
let @xmath48 be a line creating an angle of @xmath32 with both @xmath14 and @xmath15 such that @xmath11 is contained in the region bounded by @xmath49 and @xmath48 and there is a point @xmath50 on @xmath48 .
( note that @xmath51 can be found in @xmath46-time . )
let @xmath52 denote the intersection points of @xmath48 with @xmath14 and @xmath15 , respectively ( see figure [ fig : wedges ] ) .
let @xmath53 be such that @xmath54 is equilateral .
let @xmath55 be such that @xmath56 is equilateral .
* case 1 : * @xmath57 and @xmath58 . in this case
let @xmath60 be a wedge of angle @xmath32 with apex @xmath51 containing both @xmath7 and @xmath8 .
let @xmath61 be the wedge @xmath62 and let @xmath63 be the wedge @xmath64 .
see figure [ fig : case1 ] .
observe that the wedge - graph that corresponds to @xmath65 is connected ( @xmath51 is connected by edges to both @xmath7 and @xmath8 ) .
* case 2 : * without loss of generality @xmath66 . in this case
let @xmath67 , let @xmath68 , and let @xmath69 , where @xmath70 is such that @xmath71 is equilateral .
see figure [ fig : case2 ] .
again we have that @xmath7 is connected by edges to both @xmath8 and @xmath51 in the wedge - graph that corresponds to @xmath72 . finally , observe that in both cases the wedge - graph contains a 2-path on the vertices @xmath73 , and @xmath74 contains @xmath75 and hence the entire set of points @xmath11 .
we can now easily find for each point @xmath76 in constant time a wedge of angle @xmath32 and apex @xmath77 such that in the wedge - graph that corresponds to the set of all these edges , each such @xmath77 will be connected to one of @xmath73 .
depth1pt it is left to prove our main tool , lemma [ lemma : ice ] .
this is done in the next section .
we will give two different proofs for lemma [ lemma : ice ] .
* proof i. * in this proof we will assume that the set @xmath4 is _ strictly _ convex , that is , we assume that the boundary of @xmath4 does not contain a straight line segment .
we assume this in order to simplify the proof , however this assumption is not critical and can be avoided .
we bring this proof mainly for its independent interest ( see claim [ claim : per ] below ) .
the second proof of lemma [ lemma : ice ] is shorter and applies for general @xmath4 . for an angle @xmath78
we denote by @xmath79 a ( possibly translated ) copy of @xmath4 rotated in an angle of @xmath80 .
let @xmath14 and @xmath15 be two rays emanating from the origin @xmath6 and creating an angle @xmath1 .
observe that for every @xmath80 there exists a unique translation of @xmath79 that is contained in @xmath16 and both @xmath14 and @xmath15 touch @xmath79 at a _ single _ point ( here we use the fact that @xmath4 is _ strictly _ convex ) .
for every @xmath78 we denote by @xmath81 and @xmath82 the two touching points on @xmath14 and on @xmath15 , respectively .
let @xmath83 denote the distance between @xmath82 and the origin ( see figure [ fig : proof1 ] ) .
consider the boundary @xmath87 with the positive ( counterclockwise ) orientation .
note that since @xmath4 is convex , at each point @xmath88 there is a unique supporting line pointing forward ( here after positive tangent ) , and a unique supporting line pointing backwards ( the two tangent lines coincide iff @xmath87 is smooth at @xmath35 ) .
let @xmath89 be a unit speed curve traveling around @xmath87 .
we define a function @xmath90 in the following way : if @xmath91 belongs to the part of the boundary of @xmath79 between @xmath81 and @xmath82 that is visible from @xmath6 , then we set @xmath90 to be equal to the length of the orthogonal projection of the unit positive tangent at @xmath91 on the @xmath92-axis ( see figure [ fig : proof1 ] ) .
otherwise we set @xmath93 .
the simple but important observation here is that for every @xmath78 the expression @xmath94 is equal to the @xmath92-coordinate of @xmath82 .
this , in turn , is equal by definition to @xmath95 . to see this observation
take a small portion of the boundary of @xmath96 of length @xmath97 that is visible from @xmath6 .
its orthogonal projection on the @xmath92-axis has ( by definition ) length @xmath98 .
notice that the orthogonal projection of the entire part of the boundary of @xmath96 that is visible from @xmath6 on the @xmath92-axis ( whose length , therefore , equals to this integral ) is precisely all the points on the @xmath92-axis with smaller @xmath92-coordinate than that of @xmath82 . to see this
observe that @xmath91 is visible from @xmath6 through the rotation of @xmath4 precisely from where it lies on the ray @xmath15 until it lies on the ray @xmath14 . through this period the angle which the positive tangent at @xmath91 creates with the @xmath86-axis varies from @xmath102 to @xmath103
. combining ( [ eq:1 ] ) and ( [ eq:2 ] ) we conclude : analogously to @xmath83 we define @xmath106 to be the distance from @xmath81 to the origin @xmath6 .
lemma [ lemma : ice ] is equivalent to saying that there is a @xmath80 for which @xmath107 . by a similar argument or by applying the result of claim [ claim : per ] to a reflection of @xmath4
, we deduce that @xmath108 .
in particular @xmath109 .
because @xmath110 and @xmath111 are continuous we may now conclude the following : * proof ii .
* consider the point @xmath6 such that the two tangents of @xmath4 through @xmath6 create an angle of @xmath1 and such that the area of the convex hull of @xmath114 is maximum . by a simple compactness argument such @xmath6 exists .
we will show that the point @xmath6 satisfies the requirements of the lemma .
let @xmath14 and @xmath15 be the two rays emanating from @xmath6 and tangent to @xmath4 .
let @xmath7 and @xmath115 be the end points of the ( possibly degenerate ) line segment @xmath116 and assume @xmath117 .
similarly , let @xmath8 and @xmath118 be the two ( possible equal ) points such that the intersection of @xmath15 and @xmath4 is the line segment connecting @xmath8 and @xmath118 and assume @xmath119 .
assume to the contrary that @xmath125 .
without loss of generality assume that @xmath4 lies to the left of @xmath14 and to the right of @xmath15 ( see figure [ fig : proof2 ] ) .
let @xmath126 be very small positive number and let @xmath127 be the directed line supporting @xmath4 having @xmath4 to its left that is obtained from @xmath128 by rotating it counterclockwise at angle @xmath129 .
let @xmath130 be the directed line supporting @xmath4 having @xmath4 to its right that is obtained from @xmath131 by rotating it counterclockwise at angle @xmath129 ( see figure [ fig : proof2 ] ) .
let @xmath132 be the intersection point of @xmath127 and @xmath133 .
note that @xmath134 and @xmath130 create an angle of @xmath1 .
we claim that the area of the convex hull of @xmath135 is greater than the area of the convex hull of @xmath114 . indeed , up to lower order terms the difference between the two equals @xmath136 .
this contradicts the choice of the point @xmath6 .
depth1pt lemma [ lemma : ice ] clearly holds for non - convex ( but compact ) sets @xmath4 as well , since we can apply it on the convex hull of @xmath4 and observe that if a line supports the convex hull and intersects it in a single point , then this point must belong to @xmath4 . from both proofs of lemma [ lemma : ice ] it follows , and is rather intuitive as well , that one can always find at least two points @xmath6 that satisfy the requirements of the lemma . in the first proof notice that both functions @xmath110 and @xmath111 are periodic and therefore if they have the same integral over @xmath137 $ ]
, they must agree in at least _ two _ distinct points , as they are continuous .
in the second proof one can choose a point @xmath6 that minimizes the area of the convex hull of @xmath6 and @xmath4 and obtain a different solution .
i. caragiannis , c. kaklamanis , e. kranakis , d. krizanc and a. wiese , communication in wireless networks with directional antennas , _ proc .
20th symp . on parallelism in algorithms and architectures _ , 344351 , 2008
m. damian and r.y .
flatland , spanning properties of graphs induced by directional antennas , _ electronic proc . 20th fall workshop on computational geometry _ , stony brook university , stony brook , ny , 2010 .
s. dobrev , e. kranakis , d. krizanc , j. opatrny , o. ponce , and l. stacho , strong connectivity in sensor networks with given number of directional antennae of bounded angle , _ proc .
4th int .
conf . on combinatorial optimization and applications _ , 7286 , 2010 | what is the minimum angle @xmath0 such that given any set of @xmath1-directional antennas ( that is , antennas each of which can communicate along a wedge of angle @xmath1 ) , one can always assign a direction to each antenna such that the resulting communication graph is connected ? here
two antennas are connected by an edge if and only if each lies in the wedge assigned to the other .
this problem was recently presented by carmi , katz , lotker , and rosn @xcite who also found the minimum such @xmath1 namely @xmath2 . in this paper
we give a simple proof of this result .
moreover , we obtain a much stronger and optimal result ( see theorem [ theorem : main ] ) saying in particular that one can chose the directions of the antennas so that the communication graph has diameter @xmath3 . our main tool is a surprisingly basic geometric lemma that is of independent interest .
we show that for every compact convex set @xmath4 in the plane and every @xmath5 , there exist a point @xmath6 and two supporting lines to @xmath4 passing through @xmath6 and touching @xmath4 at two _ single points _ @xmath7 and @xmath8 , respectively , such that @xmath9 and the angle between the two lines is @xmath1 . |
let @xmath0 be a compact 3-manifold such that @xmath1 has no 2-sphere components .
a heegaard splitting of @xmath0 is a pair @xmath2 , where @xmath3 and @xmath4 are compression bodies such that @xmath5@xmath6 , and @xmath7@xmath8 .
@xmath8 is called a heegaard surface of @xmath0 .
the splitting is often denoted as @xmath9@xmath10@xmath4 or @xmath11 .
it is known that any compact 3-manifold has a heegaard splitting .
let @xmath12 be a heegaard splitting of @xmath0 .
@xmath13 is said to be reducible if there exist essential disks @xmath14 and @xmath15 with @xmath16 .
otherwise , it is said to be irreducible .
w. haken[h ] showed that any heegaard splitting of a reducible , compact 3-manifold is reducible , kneser[kn ] and milnor[m ] showed that any orientable , compact 3-manifold is a connected sum of @xmath17 irreducible 3-manifolds @xmath18 where @xmath19 are unique up to isotopy .
thus any heegaard splitting of @xmath0 is the connected sums of the heegaard splittings of @xmath17 factors @xmath18 defined as following : now let @xmath20 be a reducible heegaard splitting .
then @xmath21 is a 2-sphere .
suppose that @xmath22 cuts @xmath0 into @xmath23 and @xmath24 .
then @xmath25 separates @xmath3 into @xmath26 and @xmath27 , @xmath28 separates @xmath4 into @xmath29 and @xmath30 .
we may assume that @xmath31@xmath32 and @xmath33@xmath34 .
let @xmath35 and @xmath36 where @xmath37 and @xmath38 are two 3-balls .
then @xmath0 is the connected sum of @xmath39 and @xmath40 , denoted by @xmath41 .
let @xmath42@xmath43 , and @xmath44@xmath45
. then @xmath46 and @xmath47 are two compression bodies such that @xmath48 and @xmath49 .
hence @xmath50 is a heegaard splitting of @xmath39 and @xmath51 is a heegaard splitting of @xmath40 . in this case , @xmath52 is called the connected sum of @xmath53 and @xmath54 .
a heegaard splitting @xmath55 is said to be stabilized if there are two properly embedded disks @xmath56 and @xmath57 such that @xmath58 intersects @xmath59 in one point ; otherwise , it is said to be unstabilized .
some important results on stabilizations of heegaard splittings have been given in [ rs ] , [ s ] , [ st ] and [ w ] .
an interesting problem on stabilizations of heegaard splittings offered by c. gordon is the following : * _ gordon s conjecture . _ * the connected sum of two heegaard splittings is stabilized if and only if one of the two factors is stabilized .
( see problem 3.91 in [ ki ] . ) in this paper , we shall give a proof to gordon s conjecture .
the main result is the following theorem : * _ theorem 1 . _ * the connected sum of two heegaard splittings is stabilized if and only if one of the two factors is stabilized .
* comments on theorem 1 .
* \(1 ) by haken s lemma , the connected sum of the minimal heegaard splittings of @xmath39 and @xmath40 is unstabilized ; but there are many manifolds which have unstabilized heegaard splittings of distinct genera .
there are examples , given by a. casson and c. gordon[cg1 ] , independently by t. kobayashi[ko ] , which have irreducible heegaard splittings of arbitrarily high genera .
now let @xmath60 be such a manifold , and @xmath61 be any compact 3-manifold .
then , by theorem 1 , @xmath62 has unstabilized heegaard splittings of arbitrarily high genus .
\(2 ) david bachman[b ] announced that if @xmath39 and @xmath40 are closed and irreducible then gordon s conjecture is true . in this case ,
@xmath53 and @xmath54 are irreducible .
\(3 ) two applications of theorem 1 have been given in [ qm ] : \(a ) a heegaard splitting @xmath63@xmath4@xmath64@xmath3 is said to be boundary reducible if there is an essential disk @xmath65 of @xmath0 which intersects @xmath66 in an essential simple curve in @xmath66 .
a. casson and c. gordon proved that any heegaard splitting of a boundary reducible 3-manifold is boundary reducible in [ cg2 ] . in [ qm ] , we shall prove that any heegaard splitting @xmath67 of a boundary reducible , irreducible 3-manifold is obtained by doing boundary connected sums and self - boundary connected sums from heegaard splittings of @xmath17 3-manifolds @xmath18 , where @xmath68 is either boundary irreducible or a solid torus .
furthermore , @xmath67 is unstabilized if and only if one of the factors is unstabilized .
this result can be taken as the disk version of gordon s conjecture .
\(b ) suppose that @xmath60 and @xmath61 are two compact 3-manifolds with boundary .
let @xmath69 be an incompressible annulus in @xmath70 , and @xmath71 .
let @xmath72@xmath4@xmath73@xmath3@xmath74 be an unstabilized heegaard splitting of @xmath68 .
then @xmath0 has a natural heegaard splitting @xmath4@xmath75@xmath3 induced by @xmath4@xmath76@xmath77@xmath78 and @xmath4@xmath79@xmath3@xmath80 such that @xmath81@xmath4@xmath82@xmath4@xmath83@xmath4@xmath84 . without loss of generality , we may assume that @xmath85@xmath77@xmath74 .
we denote by @xmath86 the manifold obtained by attaching a 2-handle to @xmath68 along @xmath69 , @xmath9@xmath87 the manifold obtained by attaching a 2-handle to @xmath3@xmath74 along @xmath69 .
then @xmath88@xmath89@xmath73@xmath3@xmath87 is a heegaard splitting of @xmath86 . in [ qm ]
, we shall prove that if @xmath90@xmath4@xmath76@xmath3@xmath91 and @xmath92@xmath4@xmath79@xmath3@xmath93 are unstabilized , then @xmath4@xmath75@xmath3 is unstabilized .
0.5 mm we shall use a basic tool in 3-manifold theory , called band sums of disks , to prove theorem 1 . the argument in this paper
is self - contained .
we shall give an outline of the proof of theorem 1 in chapter 2 .
in this chapter , we shall introduce the ideas of the proof of theorem 1 . before doing this
, we first give some notations and related basic observations . * _ definition 2.1.1 . _ * ( 1 ) let @xmath94 $ ] .
\(2 ) suppose that @xmath95 is an arc in a surface @xmath8 , @xmath96 be a neighborhood of @xmath95 in @xmath8 such that @xmath97 .
for a sub - arc @xmath98 of @xmath95 , we denote @xmath99 by @xmath100 . in this case ,
@xmath101 and @xmath96 have the same width .
0.5 mm let @xmath102 be a subset of @xmath103 possibly not in a natural order .
this means that it is possible @xmath104 when @xmath105 . *
_ definition 2.1.2 .
_ * suppose that @xmath8 is a closed surface , @xmath106 are pairwise disjoint disks in @xmath8 , and @xmath107 are pairwise disjoint arcs in @xmath8 satisfying the following conditions : \(1 ) @xmath108 .
\(2 ) if we denote by @xmath109 the arc obtained by pushing @xmath110 off @xmath111 in @xmath112 , then each component of @xmath113 is a tree when we take @xmath114 as a fat vertex and @xmath109 as an edge .
+ then we say @xmath115 is an abstract tree.0.5 mm let @xmath115 be an abstract tree in a closed surface , and @xmath116 be a regular neighborhood of @xmath117 in @xmath8 satisfying the conditions : ( @xmath118 ) if @xmath119 , then either @xmath120 or each component of @xmath121 is an arc @xmath122 which is a core of @xmath123 , and each component of @xmath124 is @xmath125 .
( @xmath126 ) each component of @xmath127 is @xmath128 where @xmath129 is a properly embedded arc in @xmath130 .
( @xmath131 ) for @xmath132 , @xmath133 , for @xmath134 , @xmath135.0.5 mm * _ lemma 2.1.3 . _ * if conditions ( @xmath118 ) , ( @xmath126 ) and ( @xmath131 ) are satisfied , then @xmath136 is a compact surface . * _ proof . _
* let @xmath137 .
then , by condition ( @xmath126 ) , each component of @xmath138 is @xmath139 where @xmath140 is a properly embedded arc in @xmath111 .
hence @xmath98 is compact and @xmath141 .
we denote by @xmath142 the components of @xmath143 .
then @xmath144 is an arc with @xmath145 .
hence @xmath146 is a surface . by conditions ( @xmath118),(@xmath126 ) and ( @xmath131 ) and
induction on @xmath102 , we can prove lemma 2.1.3.q.e.d . * _ definition 2.1.4 . _
* let @xmath147 be an abstract tree in a closed surface . if conditions ( @xmath118 ) , ( @xmath126 ) and ( @xmath131 ) are satisfied , then @xmath148 is called a surface generated by @xmath115 .
figure 1 figure 2 * _ example 1 : _ * ( 1 ) .
suppose that @xmath149 are three disks in a surface @xmath8 and @xmath150 are two arcs in @xmath8 as in figure 1(a ) .
then @xmath151 is an abstract tree .
suppose that @xmath152 and @xmath153 are as in figure 1(b ) .
then @xmath154 is not an abstract tree . in this case , if @xmath155 is an arc such that @xmath156 and @xmath157 is disjoint from @xmath158 , then @xmath159 is not a tree.0.5 mm ( 3 ) .
figure 2 is a surface generated by an abstract tree @xmath115 as in figure 1(a ) . in this case , @xmath160 .
note that the surface generated by a fixed abstract tree is not unique .
now we consider a kind of special arcs in a surface generated by an abstract tree . *
_ definition 2.1.5 .
_ * let @xmath66 be a surface generated by an abstract tree @xmath115 .
if @xmath95 is a properly embedded arc in @xmath66 such that \1 ) for @xmath132 , each component of @xmath161 is a core of @xmath123 which is contained in @xmath162 , \2 ) each component of @xmath163 is a core of @xmath164 for some @xmath165 , \3 ) for each @xmath165 , there is at most one component of @xmath163 which is a core of @xmath164 .
+ then @xmath95 is said to be regular in @xmath66.0.5 mm * _ example 2 : _ * the arc @xmath95 in figure 2 is regular in @xmath166 . in this case , @xmath95 intersects @xmath167 in two cores of @xmath168 , @xmath169 in one core of @xmath170 , @xmath171 in two components , one of which is a core of @xmath172 and the other is a core of @xmath168 .
@xmath173 contains three components @xmath174 with @xmath175.0.5 mm * _ lemma 2.1.6 .
_ * let @xmath95 be a regular arc in a surface @xmath66 generated by an abstract tree @xmath147 . then @xmath176 satisfying the following conditions : \1 ) @xmath177 , @xmath178 is a properly embedded arc in @xmath179 which is disjoint from @xmath180 .
\2 ) @xmath181 , and @xmath182 is a core of @xmath183 .
\3 ) for @xmath184 , @xmath185 and for @xmath186 , @xmath187 , @xmath188 , @xmath189 .
\4 ) for @xmath190 , @xmath191 , for @xmath192 , @xmath193 .
\5 ) for each @xmath194 , @xmath195 contains at most one component . *
_ * since @xmath95 is regular in @xmath66 , by definition 2.1.5(1 ) and ( 2 ) , @xmath196 is disjoint from @xmath197 , and each component of @xmath198 is a core of @xmath164 for some @xmath165 . by definition 2.1.4 ,
@xmath199 , each component of @xmath200 is an properly embedded arc in @xmath114 for some @xmath201 .
hence @xmath202 , and ( 1 ) , ( 2 ) , ( 3 ) holds . by definition 2.1.5(3 ) , @xmath203 for @xmath204 . now if @xmath205 for @xmath206 , then @xmath115 is not an abstract tree , a contradiction . hence ( 4 ) holds . ( 5 ) follows from ( 1 ) and ( 4 ) .
q.e.d.0.5 mm * _ remark 2.1.7 . _ * the properties of regular arcs in lemma 2.1.6 are important in the proof of theorem 1 .
lemmas 3.1.2 and 3.3.1 follow from lemma 2.1.6 and they take roles in the inductive proof of theorem 1 .
0.5 mm * _ lemma 2.1.8 .
_ * let @xmath136 be a surface generated by an abstract tree .
then @xmath207 is also a surface generated by an abstract tree where @xmath208 is any integer . * _ proof . _
* since @xmath209 is an abstract tree , @xmath210 is also an abstract tree .
obviously , conditions ( @xmath118 ) , ( @xmath126 ) and ( @xmath131 ) are satisfied .
hence @xmath211 is a surface generated by @xmath210.q.e.d .
* it is easy to see that if one of @xmath212 and @xmath213 is stabilized then @xmath214 is stabilized . so in order to obtain a contradiction
, we may assume that each of @xmath53 and @xmath215 is unstabilized and @xmath11 is stabilized .
* * assumption ( * ) . *
( 1 ) let @xmath216 be a pair of stabilized disks such that @xmath217 , @xmath218 . now @xmath58 intersects @xmath59 in only one point @xmath219 . recalling the two disks @xmath25 and @xmath220
defined in chapter 1 .
we may assume that each component of @xmath221 is an arc in both @xmath58 and @xmath25 , each component of @xmath222 is an arc in both @xmath59 and @xmath220 .
it is easy to see that if @xmath223 or @xmath224 , then one of @xmath53 and @xmath213 is stabilized .
so we may assume that @xmath225 and @xmath226 .
* _ assumption ( * ) .
_ * ( 2 ) @xmath227 .
\(3 ) @xmath228 and @xmath229.1.5 mm now consider @xmath221 . by assumption ,
each component @xmath230 of @xmath221 separates @xmath58 into two disks @xmath231 and @xmath232 such that @xmath233 .
now we denote by @xmath234 the disk in @xmath232 which is bounded by @xmath230 , with some components in @xmath221 and some arcs in @xmath235 , such that @xmath236 is disjoint from @xmath25 as in figure 3 .
then @xmath234 is a properly embedded disk in @xmath237 or @xmath27 .
figure 3 * _ definition 2.2.1 .
_ * a component @xmath230 in @xmath238 is labeled a symbol @xmath239 where @xmath240 if @xmath241 and @xmath242 if @xmath243.0.5 mm similarly , each component @xmath230 of @xmath222 separates @xmath59 into two disks @xmath244 and @xmath245 such that @xmath246 .
now we denote by @xmath247 the disk in @xmath245 which is bounded by @xmath230 with some components in @xmath222 and some arcs in @xmath248 such that @xmath249 is disjoint from @xmath220 .
then @xmath247 is a properly embedded disk in @xmath29 or @xmath250 defined in section 1.0.5 mm * _ definition 2.2.2 . _
* a component @xmath230 in @xmath251 is labeled a symbol @xmath239 where @xmath240 if @xmath252 and @xmath242 if @xmath253.0.5 mm by assumption ( * ) , we number the components of @xmath221 , @xmath254 , and the components of @xmath222 , @xmath255 , so that if @xmath256 and @xmath257 , then @xmath258 , @xmath259 .
now we denote by @xmath260 the disk @xmath261 , @xmath262 the disk @xmath263 for @xmath264 and @xmath265.0.5 mm * _ definition 2.2.3 . _
* for each @xmath266 in @xmath238 and each @xmath267 in @xmath222 , let @xmath268 and @xmath269 the following two lemmas are immediately from definitions . *
_ lemma 2.2.4 . _ *
1 ) if @xmath270 , then @xmath271 , and \2 ) if @xmath272 , then @xmath273 . * _ proof . _
* we need only to prove ( 1 ) . if @xmath270 , then @xmath274 . by definitions ,
0.5 mm * _ lemma 2.2.5 .
_ * ( 1 ) @xmath275 if and only if @xmath276 for each @xmath270 .
\(2 ) if @xmath277 , then @xmath278 .
\(3 ) @xmath279 if and only if @xmath280 for each @xmath281 .
\(4 ) if @xmath282 , then @xmath283.0.5 mm * _ proof . _ * this lemma is immediately from definitions and simple observations.q.e.d.0.5 mm since @xmath284 , each component of @xmath285 is either the disk @xmath260 for some @xmath264 with @xmath275 or a disk containing @xmath219 , denoted by @xmath286 . by definition of @xmath260 , @xmath286 is disjoint from @xmath260 for @xmath275 . now
each component of @xmath287 is the disk @xmath260 for some @xmath264 with @xmath288 .
similarly , each component of @xmath289 is either the disk @xmath262 for some @xmath290 with @xmath279 or a disk containing @xmath219 , denoted by @xmath291 , each component of @xmath292 is the disk @xmath262 for some @xmath290 with @xmath293.0.5 mm * _ definition 2.2.6 .
_ * let @xmath294 and @xmath295.0.5 mm * _ lemma 2.2.7 .
_ * 1 ) if @xmath296 , then @xmath276 .
\2 ) if @xmath297 , then @xmath280 .
\3 ) @xmath298 , @xmath299 . * _ proof .
_ * by assumption , @xmath300 .
hence @xmath301 , @xmath302 .
if @xmath296 , then @xmath303 . if @xmath297 , then @xmath304
now @xmath305 separates @xmath58 into @xmath306 and @xmath307 such that @xmath308 .
suppose , otherwise , @xmath309 . by definition ,
@xmath305 is disjoint from @xmath310 .
this means that there is an integer @xmath311 such that @xmath266 separates @xmath305 and the point @xmath219 in @xmath306 . by definition ,
hence @xmath313 , a contradiction . q.e.d.0.5
mm recalling the definitions of @xmath53 and @xmath314 in section 1 . now
@xmath262 and @xmath291 are properly embedded in @xmath29 or @xmath30 for @xmath290 .
it is easy to see that for each @xmath208 , there is an arc @xmath315 in @xmath25 such that @xmath316 bounds a disk in @xmath317 , denoted by @xmath318 , and a disk in @xmath319 , denoted by @xmath320 as in figure 4 .
note that for each @xmath321 , @xmath322 and @xmath323 .
thus if @xmath279 , then @xmath324 is a properly embedded disk in @xmath46 , and if @xmath293 , then @xmath325 is a properly embedded disk in @xmath47 . specially , @xmath326 is a properly embedded disk in @xmath46 .
figure 4 figure 5 * _ lemma 2.2.8 . _ * @xmath327 and @xmath328 are two sets of pairwise disjoint arcs properly embedded in @xmath25 such that \(1 ) @xmath266 intersects @xmath315 in at most one point , and \(2 ) @xmath329 for each @xmath330 and @xmath290 . * _ proof . _
* since @xmath331 , @xmath332 and @xmath333 , by assumption ( * ) , @xmath266 is properly embedded in @xmath25 and @xmath267 is properly embedded in @xmath220 such that @xmath266 is disjoint from @xmath267 .
note that @xmath334 . hence either the two end points of @xmath267 lie in the same component of @xmath335 or the two end points of @xmath267 lie in the distinct components of @xmath336 as in figure 5 .
hence the lemma holds.q.e.d . *
_ lemma 2.2.9 . _
* there are four sets of pairwise disjoint properly embedded disks @xmath337 in @xmath26 , @xmath338 in @xmath27 , @xmath339 in @xmath46 , and @xmath340 in @xmath47 satisfying the following conditions : \(1 ) @xmath341 , @xmath342 , @xmath343 , @xmath344 .
\(2 ) if @xmath345 , then @xmath346 , @xmath347 , @xmath348 , @xmath349 .
\3 ) if @xmath350 , then @xmath346 . * _ proof . _
* ( 1 ) follows from the construction of @xmath315 . since @xmath351 , @xmath352 for each @xmath264 and @xmath290 . by the constructions of @xmath353 , @xmath354 , @xmath315 , ( 2 ) and ( 3 ) holds.q.e.d .
* the idea of the proof of theorem 1 * now let @xmath3@xmath355@xmath356 where @xmath357 and @xmath358 are regular neighborhoods of @xmath260 and @xmath286 in @xmath26 , @xmath3@xmath359@xmath360 where @xmath357 is a regular neighborhood of @xmath260 in @xmath27
. then @xmath3@xmath361@xmath362 and @xmath9@xmath363@xmath364 are two compression bodies .
let @xmath89@xmath365@xmath366 , where @xmath367 and @xmath368 are regular neighborhoods of @xmath353 and @xmath354 in @xmath46 , @xmath89@xmath369@xmath370 where @xmath367 is a regular neighborhood of @xmath353 in @xmath47 . now @xmath59 and
@xmath58 defined in section 2.2 are a pair of stabilized disks of the connected sum of @xmath3@xmath371@xmath89@xmath372 and @xmath3@xmath373@xmath4@xmath374 .
if theorem 1 is true , then one of @xmath3@xmath371@xmath89@xmath372 and @xmath3@xmath373@xmath4@xmath374 , say @xmath3@xmath371@xmath4@xmath372 , is stabilized .
this means that there are two essential disks @xmath375@xmath372 and @xmath376@xmath372 such that @xmath377 intersects @xmath378 in only one point .
hence @xmath377 is obtained by doing band sums from @xmath353 with @xmath279 and @xmath354 , and @xmath378 is obtained by doing band sums from @xmath260 with @xmath275 and @xmath286 .
we do want to do this .
recalling @xmath379 , @xmath380 which are defined in section 2.2 . * _ definition 2.3.1 . _
* for @xmath381 and @xmath382 .
let @xmath383 , @xmath384 , @xmath385 and @xmath386.0.5 mm by definition , if @xmath387 , then @xmath388 , @xmath389 .
furthermore , @xmath390 . by definitions ,
hence @xmath383 , @xmath392 .
similarly , @xmath393 .
0.5 mm * to prove theorem 1 , we only need to prove the following propositions:*0.5 mm * _ proposition 1 . _ * for each @xmath381 , there are a surface @xmath394 in @xmath395 , a subset of @xmath103 , say @xmath396 , and two sets of pairwise disjoint arcs @xmath397 and @xmath398 in @xmath394 satisfying the following conditions : \(1 ) @xmath399 is a surface generated by an abstract tree @xmath400 such that \(i ) for each @xmath401 with @xmath402 , @xmath403 ; \(ii ) for each @xmath401 with @xmath404 , @xmath405 is disjoint from @xmath406 .
\(2 ) for @xmath407 , @xmath408 is regular in @xmath409 .
\(3 ) for @xmath407 and @xmath410 , if @xmath411 , then either @xmath412 with @xmath413 or @xmath408 is disjoint from @xmath414 .
\(4 ) for each @xmath415 , @xmath416 is a properly embedded arc in @xmath394 lying in @xmath417 for some @xmath194 .
furthermore , for each @xmath418 , @xmath419 intersects @xmath416 in at most one point .
\(5 ) for each @xmath415 , @xmath418 and @xmath410 , @xmath420 , @xmath421.0.5 mm * _ definition 2.3.2 . _
* if @xmath422 intersects @xmath416 in one point , then we say @xmath423 and @xmath424.0.5 mm * _ proposition 2 . _
* for each @xmath425 and @xmath426 .
\(1 ) if @xmath427 , then @xmath428 .
\(2 ) if @xmath429 , then @xmath430 .
\(3 ) @xmath431 and @xmath432 .
+ where @xmath433 is as in proposition 1 .
0.5 mm * _ proposition 3 . _
* for each @xmath381 , there are two sets of pairwise disjoint disks @xmath434 properly embedded in @xmath26 and @xmath435 properly embedded in @xmath436 satisfying the following conditions : \(1 ) @xmath437 , @xmath438 , @xmath439 , @xmath440 .
\(2 ) @xmath441 , @xmath442 , @xmath443 , @xmath444 where @xmath445 in lemma 2.2.9 . 0.5 mm recalling the set @xmath446 in proposition 1.0.5 mm * _ proposition 4 . _ * for each @xmath425 , there are a surface @xmath447 in @xmath448 , and two sets of pairwise disjoint arcs @xmath449 and @xmath450 in @xmath447 satisfying the following conditions : \(1 ) @xmath451 is a surface generated by an abstract tree @xmath452 such that \(i ) for each @xmath401 , if @xmath404 , then @xmath453 ; \(ii ) if @xmath402 , @xmath454 is disjoint from @xmath455 .
\(2 ) for @xmath407 , @xmath456 is regular in @xmath457 .
\(3 ) for @xmath407 and @xmath410 , if @xmath411 , then either @xmath412 with @xmath458 or @xmath456 is disjoint from @xmath459 .
\(4 ) for each @xmath415 , @xmath460 is a properly embedded arc in @xmath447 lying in @xmath461 for some @xmath194 .
furthermore , for each @xmath418 , @xmath462 intersects @xmath460 in at most one point .
\(5 ) for each @xmath415 , @xmath418 and @xmath410 , @xmath463 , @xmath464 .
0.5 mm * _ definition 2.3.3 . _
* if @xmath465 intersects @xmath460 in one point , then we say @xmath466 and @xmath467 . 0.5 mm * _ proposition 5 . _ * @xmath468 , @xmath469 . for each @xmath425 and @xmath470 .
\(1 ) if @xmath471 , then @xmath472 .
\(2 ) if @xmath473 , then @xmath474.0.5 mm * _ proposition 6 . _ * for each @xmath381 , there are two sets of pairwise disjoint disks @xmath475 properly embedded in @xmath476 and @xmath477 properly embedded in @xmath47 such that \(1 ) @xmath478 , @xmath479 ; \(2 ) @xmath480 . 0.5 mm * now we prove theorem 1 under the assumption that propositions 1 - 6 are true . * * _ the proof of theorem 1 . _ * suppose that @xmath481 . by the definition of @xmath482 ,
@xmath483 . by proposition 3(2 ) , @xmath484 . by proposition 3(1 ) , @xmath485 , @xmath486 .
hence @xmath487 and @xmath488 .
this means that @xmath53 is stabilized.q.e.d . *
* though theorem 1 follows immediately from proposition 3 . but propositions 1 , 2 , 4 , 5 and 6 are necessary in the inductive proof of proposition 3 .
see section 2.5 we organize the proofs of propositions 1 - 6 as follows : in section 2.5 , we shall prove propositions 1 - 6 for @xmath489 .
furthermore , we shall introduce the ideas of the inductive proofs of propositions 1 - 6 . in chapter 3 , we shall study properties of @xmath490 and @xmath491 under the assumptions that propositions 1 - 6 are true for @xmath492 . in chapter 4
, we shall prove that propositions 4 - 6 are true for @xmath493 under the assumptions : @xmath494 and propositions 1 - 6 are true for @xmath492 . in chapter 5 , we shall prove that propositions 4 - 6 are true for @xmath493 under the assumptions : @xmath495 and propositions 1 - 6 are true for @xmath492 . in chapter 6
, we shall finish the proofs of propositions 1 - 6 .
* _ lemma 2.5.1 . _ * propositions 1 - 3 are true for @xmath489 . * _ proof .
_ * by the definition , @xmath25 is a disk in @xmath496 .
hence @xmath25 is a surface generated by an trivial abstract tree @xmath497 .
now let @xmath498 @xmath499 for @xmath264 and @xmath290 . by lemma 2.2.8
, @xmath500 intersects @xmath501 in at most one point and @xmath502 .
specially , let @xmath503 .
thus proposition 1 holds .
we denote by @xmath504 the disk @xmath353 for @xmath279 , @xmath505 the disk @xmath354 , @xmath506 the disk @xmath260 for @xmath275 , @xmath507 the disk @xmath286 .
then , by lemma 2.2.9 , we obtain two sets of pairwise disjoint properly embedded disks @xmath508 in @xmath26 and @xmath509 in @xmath46 . by definition 2.3.1 , @xmath510 , @xmath511 , @xmath512 , @xmath513 . by lemma 2.2.9 , proposition 3 holds.q.e.d .
* _ lemma 2.5.2 . _ * propositions 4 - 6 are true for @xmath489 . * _ proof .
* by the definitions of @xmath47 and @xmath476 , @xmath25 is a disk in @xmath514 .
hence @xmath25 is a surface generated by an trivial abstract tree @xmath497 in @xmath515 .
now let @xmath516 @xmath517 . by lemma 2.2.8
, @xmath518 intersects @xmath519 in at most one point and @xmath520 .
it is easy to see that @xmath521 and @xmath469 .
hence propositions 4 and 5 holds .
we denote by @xmath504 the disk @xmath353 for @xmath293 , @xmath506 the disk @xmath260 for @xmath288 .
then , by lemma 2.2.9 , we obtain two sets of pairwise disjoint properly embedded disks @xmath522 in @xmath27 and @xmath523 in @xmath524 .
note that @xmath525 . by definition 2.3.1 , @xmath510 , @xmath526 . by lemma 2.2.9 ,
proposition 6 holds.q.e.d . * _ the ideas of the proofs of propositions 1 - 6 .
_ * by lemmas 2.5.1 and 2.5.2 , propositions 1 - 6 are true for @xmath489 .
now we may assume that propositions 1 - 6 are true for @xmath492 .
we only need to prove propositions 1 - 6 are true for @xmath493 .
the inductive proofs depend on @xmath527 .
@xmath494 . by proposition 1
, @xmath528 is a properly embedded arc in @xmath529 .
let @xmath530 . by definition 2.3.2 and proposition 1 , either @xmath531 or @xmath532 .
now let @xmath533 . by proposition 3 , there is a properly embedded disk @xmath534 in @xmath26 such that @xmath535 . by proposition 1
, @xmath536 is an arc in @xmath529 if @xmath537 .
we can prove that @xmath538 intersects @xmath536 in only one point , and @xmath539 .
let @xmath540 be a regular neighborhood of @xmath541 in @xmath542 , and @xmath543 be a regular neighborhood of @xmath536 in @xmath529 .
in fact , @xmath544 , @xmath545 for @xmath546 , and @xmath547(@xmath548 ) is the band sum of @xmath549 with some copies of @xmath541 along @xmath536 .
specially , @xmath550 and @xmath551 for @xmath546 with @xmath279 ; @xmath552 is the connected sum of @xmath286 with some copies of @xmath534 , and @xmath553 is the connected sum of @xmath554 with some copies of @xmath534 for @xmath548 with @xmath275 .
note that propositions 1 - 6 are true for @xmath492 . by propositions 2 , 4 and lemmas 2.5.1 , 2.5.2 , we can prove that @xmath555 and @xmath556 .
since @xmath539 , by proposition 6 , @xmath557 is a properly embedded disk in @xmath47 such that @xmath558 . we can also prove that @xmath559 intersects @xmath538 in only one point if @xmath537 .
let @xmath560 be a regular neighborhood of @xmath561 in @xmath562 , and @xmath563 be a regular neighborhood of @xmath538 in @xmath564 .
in fact , @xmath565 , @xmath566 for @xmath548 , and @xmath567 is the connected sum of @xmath568 with some copies of @xmath569 along @xmath538 for @xmath546 .
specially , @xmath570 for @xmath548 with @xmath288 , and @xmath571 is the connected sum of @xmath572 with some copies of @xmath557 for @xmath573 with @xmath293 .
@xmath495 .
now let @xmath574 . by proposition 4 , we can prove that @xmath575 .
thus we have an alternating proof with the one of case 1 .
so propositions 4 - 6 are necessary to show @xmath575 when @xmath495.q.e.d .
by the argument in section 2.5 , propositions 1 - 6 holds for @xmath489 . in this chapter , we shall give some properties of @xmath577 and @xmath578 under the assumptions that propositions 1 - 6 hold for @xmath492 .
we first consider @xmath578 . *
_ lemma 3.1.1 .
_ * suppose that @xmath580 , and @xmath581 for each @xmath132 with @xmath458 .
then @xmath568 is regular in @xmath564 . * _ proof . _
* now by proposition 4(3 ) , if @xmath411 , then @xmath568 is disjoint from @xmath582 . by proposition 4(2 ) , the lemma holds.q.e.d . * _ lemma 3.1.2 . _ * for each @xmath583 , @xmath584 satisfying the following conditions : \(1 ) @xmath585 , and @xmath586 is a core of @xmath587 for @xmath588 .
\(2 ) @xmath589 , @xmath590 is a properly embedded arc in @xmath591 which is disjoint from @xmath592 for @xmath593 .
\(3 ) for @xmath594 , @xmath595 , @xmath596 .
\(4 ) @xmath597 , @xmath598 , @xmath599 , @xmath600 .
\(5 ) for each @xmath194 , @xmath601 intersects @xmath602 in at most one component , denoted by @xmath603 . furthermore ,
if @xmath604 , then @xmath605 ; if @xmath606 for each @xmath607 , then @xmath608 .
* _ proof . _ * by proposition 4(2 ) , @xmath568 is regular in @xmath609 .
hence the lemma follows from lemma 2.1.6 .
q.e.d.0.5 mm * _ remark .
_ * it is possible that @xmath610 . in this case ,
hence @xmath568 is disjoint from @xmath582 for @xmath612.0.5 mm now by proposition 4(4 ) , @xmath590 intersects @xmath579 in at most one point . * _ definition 3.1.3 .
_ * we say @xmath613 if @xmath590 intersects @xmath579 in one point .
* _ lemma 3.1.4 . _ * ( 1 ) if @xmath614 , then @xmath615 .
\(2 ) @xmath616 . * proof .
* by proposition 4(4 ) , if @xmath590 intersects @xmath579 in one point , then @xmath617 . by lemma 3.1.2(3 ) , ( 1 ) holds . by lemma 3.1.2(5 ) , each component of @xmath618 is @xmath590 for some @xmath619 . hence ( 2 ) holds .
0.5 mm * _ lemma 3.1.5 . _ * @xmath555 and @xmath556 . * _ proof . _
* by lemmas 2.5.1 and 2.5.2 , @xmath620 .
now by propositions 2 and 5 , the lemma holds .
0.5 mm * _ lemma 3.1.6 . _
* ( 1 ) each component of @xmath621 is either @xmath603 or a core of @xmath622 for some @xmath132 and @xmath623 .
furthermore , @xmath624 , @xmath458 .
\(2 ) if @xmath625 , then either @xmath626 or @xmath627 for some @xmath273 . * _ proof . _ * ( 1 ) by lemma 3.1.2 , @xmath584 such that @xmath628 and @xmath629 . now
if one component @xmath630 of @xmath631 is not @xmath603 , then , by lemma 3.1.2(2 ) and ( 4 ) , @xmath632 for some @xmath633 where @xmath586 is a core of @xmath587 . since @xmath634
. by proposition 4(1 ) ,
@xmath635 and @xmath636 .
hence @xmath637 for @xmath638 .
by proposition 4(2 ) , @xmath639 is a core of @xmath640 .
note that @xmath641 . by induction , ( 1 ) holds .
\(2 ) let @xmath642 be a point in @xmath643 .
then , by ( 1 ) , either @xmath644 or @xmath645 where @xmath630 is a core of @xmath640 for some @xmath132 and @xmath646 . by proposition 4(5 ) , @xmath579 intersects @xmath630 in one point if and only if it intersects @xmath647 in one point .
by induction , ( 2 ) holds .
0.5 mm 0.5 mm * for each @xmath649 and @xmath132 , either @xmath650 or @xmath581 .
we may assume that @xmath651 .
in this case , if @xmath652 for each @xmath132 , then @xmath653 . * * _ definition 3.2.1 . _ * ( 1 ) let @xmath654 if @xmath651 and @xmath655 , \(2 ) let @xmath656 if @xmath652 for each @xmath132 or @xmath657 with @xmath658 . definition 3.2.1 means that @xmath659 is either a disk or an arc . recalling the equality : @xmath584 for each @xmath649 in lemma 3.1.2 , and @xmath603 in lemma 3.1.2(5).0.5 mm * _ lemma 3.2.2 . _ * if @xmath660 , then @xmath661 is disjoint from @xmath662 for each @xmath663 . * _ proof . _ * without loss of generality , we may assume that @xmath651 , @xmath664 and @xmath665 .
assume that @xmath666 .
since @xmath667 , @xmath668 is disjoint from @xmath669 .
assume now that @xmath670 .
if @xmath671 , then @xmath672 . by proposition 4(1 ) , @xmath564 is generated by an abstract tree . by definition 2.1.4 ,
each component of @xmath673 is @xmath674 where @xmath675 is a core of @xmath676 . since @xmath603 and @xmath677 are properly embedded in @xmath602 , @xmath678 .
hence @xmath679 . by lemma 2.2.4 ,
@xmath680 . by proposition 4(3 ) , @xmath681 .
note that @xmath682 , contradicting lemma 3.1.2(2).q.e.d.0.5 mm * _ lemma 3.2.3 .
_ * if @xmath683 , then either @xmath684 or @xmath650 with @xmath458 .
* _ proof .
_ * suppose that @xmath685 for @xmath686 and @xmath581 . if @xmath687 , then , by proposition 4(3 ) , @xmath568 is disjoint from @xmath582 .
suppose that @xmath612 .
then , by lemma 3.1.2 , @xmath590 is disjoint from @xmath688 for each @xmath593 . by proposition 4(2 ) , each component of @xmath689 is a core of @xmath690 . since @xmath564 is generated by an abstract tree , @xmath691 for each @xmath692 . since @xmath693 , each component of @xmath689 lies in @xmath694
( see the equality of @xmath568 in lemma 3.1.2 . )
thus each component of @xmath695 is a core of @xmath582 . by definition 2.1.4 , @xmath696 .
0.5 mm * _ lemma 3.2.4 .
_ * if @xmath697 for some @xmath132 and some @xmath663 , then each component of @xmath698 is either @xmath699 for some @xmath646 or @xmath700 where @xmath701 is a core of @xmath702 for some @xmath703 and @xmath651 .
furthermore , @xmath704 , @xmath705 . *
_ * now by proposition 4(1 ) , @xmath458 and if @xmath706 is a component of @xmath707 , then @xmath708 for some @xmath646 .
hence @xmath709 .
if @xmath710 . then , by lemma 3.1.2 , @xmath711 for some @xmath712 . by induction , the lemma holds.q.e.d . * _ corollary 3.2.5 . _ * if @xmath713 , then @xmath568 and @xmath582 are disjoint from @xmath538 for each @xmath714 and @xmath132 .
the corollary follows from lemma 3.1.6 and lemma 3.2.4 and proposition 4(5 ) . *
_ lemma 3.2.6 .
* ( 1 ) suppose that @xmath715 and @xmath651 .
then @xmath716 is disjoint from @xmath582 .
\(2 ) suppose that @xmath717 .
then @xmath718 is disjoint from @xmath582 .
* _ proof .
_ * ( 1 ) suppose that @xmath658 .
then , by proposition 4(3 ) and definition 3.2.1 , @xmath719 is disjoint from @xmath720 . now suppose that @xmath655 . by proposition 4(1 ) and definition 2.1.4 , @xmath721 is @xmath722 where @xmath723 is a core of @xmath724 . by lemma 3.2.3 , @xmath716 is disjoint from @xmath582 .
\(2 ) suppose that @xmath658 .
then , by lemma 3.1.2 and definition 3.2.1 , @xmath725 is disjoint from @xmath582 . now suppose that @xmath655 . by proposition 4(1 ) and definition 2.1.4 , each component of @xmath726 is @xmath727 , where @xmath728 is a core of @xmath724 . since @xmath729 is disjoint from @xmath582 , ( 2 ) holds
. q.e.d . 0.5 mm * _ definition 3.3.0 . _
* let @xmath574.0.5 mm by proposition 4(4 ) , @xmath731 lies in @xmath602 for some @xmath663 . without loss of generality , we may assume that @xmath732 . recalling @xmath733 defined in lemma 3.1.2(5 ) .
now if @xmath734 , then , by proposition 4(4 ) and lemma 3.1.2 , @xmath735 intersects @xmath538 in one point and @xmath736 .
now we rearrange all the elements in @xmath737 as @xmath738 according to the order of @xmath739 lying in @xmath538 as in figure 6 .
figure 6 * _ lemma 3.3.1 . _ * for each @xmath740 , @xmath741 satisfying the following conditions : \(1 ) for @xmath742 , @xmath743 is a core of @xmath744 for some @xmath745 .
\(2 ) for @xmath742 and @xmath746 is a properly embedded arc in @xmath747 which is disjoint from @xmath748 .
\(3 ) @xmath749 is a properly embedded arc in @xmath750 .
\(4 ) for @xmath751 , @xmath752 , @xmath753 , and @xmath754 .
\(5 ) for @xmath742 , @xmath755 , @xmath756 , @xmath757 , @xmath758 , @xmath759 , @xmath760 . * _ proof . _
* since @xmath740 , @xmath761 , by lemma 3.1.2 , @xmath762 .
now we denote by @xmath763 the integer @xmath764 , @xmath765 the integer @xmath766 , @xmath746 the arc @xmath767 , @xmath743 the arc @xmath768 for @xmath769 .
let @xmath770 and @xmath771 .
in particular , we denote by @xmath749 the arc @xmath772 . by lemma 3.1.2 , the lemma holds.q.e.d.0.5 mm * _ remark 3.3.2 .
* 1 ) in order to simplify the formulation , we shall write as @xmath773 .
in the formulation , `` @xmath774 '' means `` the union index from @xmath775 to @xmath776 , then 0 , then @xmath777 , and then to @xmath778 '' , but `` @xmath779 '' means `` the union index from @xmath775 to @xmath776 , then @xmath777 , and then to @xmath778 '' .
\2 ) if @xmath780 , then @xmath781 . *
_ lemma 3.3.3 . _
* ( 1 ) @xmath782 where @xmath714 , @xmath740 , and @xmath783 is defined in definition 3.2.1 .
\(2 ) @xmath784 for @xmath132 . *
_ * ( 1 ) now each component of @xmath785 is a point . by assumption ,
@xmath786 . if @xmath787 , then @xmath788 . by lemma 3.1.6 ,
each component of @xmath789 is either @xmath735 or a core of @xmath790 for some @xmath132 and @xmath623 .
suppose that @xmath791 is a component of @xmath792 such that @xmath793 .
if @xmath794 , then @xmath734 . if @xmath795 , then @xmath796 .
\(2 ) follows from lemma 3.2.4 and proposition 4(5).q.e.d.0.5 mm figure 7 * _ lemma 3.3.4 . _
* if @xmath797 , then there are two integers @xmath798 and @xmath799 satisfying the following conditions : \(1 ) @xmath800 .
\(2 ) if @xmath801 , then @xmath802 and @xmath803 .
\(3 ) @xmath804 .
\(4 ) @xmath805 .
\(5 ) if @xmath806 , then @xmath807 .
* _ proof . _
* since @xmath808 , @xmath809 .
suppose that @xmath810 and @xmath811 are two integers such that @xmath803 for @xmath801 .
since @xmath812 is an abstract tree , @xmath802 for @xmath813 .
now @xmath800 .
assume now that @xmath814 and @xmath815 , and @xmath816 for some @xmath817 or @xmath818 and some @xmath819 or @xmath820 .
since @xmath821 , @xmath822 is not an abstract tree even if @xmath823 , contradicting proposition 4(1 ) .
hence ( 1 ) , ( 2 ) , ( 3 ) and ( 4 ) hold .
suppose that @xmath824 .
now @xmath803 and @xmath825 , @xmath802 .
hence @xmath826 . by proposition 4(4 ) and ( 5 ) , @xmath827 is disjoint from @xmath828 .
hence @xmath829 intersects @xmath746 in one point if and only if @xmath829 intersects @xmath830 in one point as in figure 7 . hence ( 5 ) holds .
q.e.d .
figure 8(a ) 0.5 mm recalling the equality @xmath773 in remark 3.3.2 , and the two integers @xmath810 and @xmath811 in lemma 3.3.4 . in this case
, @xmath746 is an arc in @xmath747 and @xmath743 is a core of @xmath744 for @xmath769 .
in particular , @xmath749 is an arc in @xmath750 . by lemma 3.3.4 , @xmath803 and @xmath802 for @xmath831 . * _ definition 3.4.1 . _ * for each @xmath832 .
\(1 ) let @xmath833 be the disk in @xmath747 which is bounded by @xmath746 and @xmath830 with two arcs in @xmath834 .
\(2 ) let @xmath835 be the disk in @xmath836 which is bounded by @xmath743 and @xmath837 with two arcs in @xmath838 as in figure 8(a ) .
\(3 ) let @xmath839 be the disk in @xmath750 which is bounded by @xmath749 and @xmath840 with two arcs in @xmath841 as in figure 8(a ) .
\(4 ) let @xmath842 .
now we rewrite @xmath843 as @xmath844 .
now suppose that @xmath845 . in this section
, we shall prove that @xmath846 , and @xmath843 is a disk in @xmath564 .
0.5 mm * _ lemma 3.4.2 . _ * let @xmath714 .
if @xmath847 for some @xmath848 , then @xmath849 . furthermore , if @xmath850 , then @xmath851 or @xmath852 .
* _ proof .
* assume that @xmath853 for some @xmath854 .
by lemma 3.1.6 , each component of @xmath855 is either @xmath856 or a copy of @xmath857 for some @xmath646 and @xmath132 with @xmath612 .
there are two cases : case 1 .
one component @xmath630 of @xmath855 is @xmath856 . now if @xmath858 or @xmath859 , then , by lemma 3.3.2 , either @xmath860 or @xmath861 .
suppose that @xmath862 . by lemma 3.2.2 , @xmath863 is disjoint from @xmath864 . since @xmath854 , @xmath865 and @xmath866 .
now @xmath867 .
now we claim that @xmath868 and @xmath869 .
suppose that @xmath870 , then @xmath871 . by proposition 4(1 ) , @xmath872 . by proposition 4(3 ) , @xmath873 . by proposition 4(2 ) , @xmath568 is regular in @xmath874 .
this means that @xmath875 intersects @xmath876 in at least two cores of @xmath876 , contradicting lemma 3.1.2(3 ) .
now @xmath877 , @xmath878 and @xmath869
. hence @xmath568 intersects @xmath879 in at least a core of @xmath879 . by lemma 3.1.2 , @xmath880 as in figure 8(b ) . by induction , @xmath881 , and @xmath882 . contradicting the order of @xmath883 we rearrange .
one component @xmath630 of @xmath884 is a copy of @xmath857 . by the argument in case 1 , @xmath885 or @xmath859 and @xmath886 .
assume now that @xmath887 .
since @xmath888 and @xmath889 . then , by proposition 4(3 ) , @xmath890.q.e.d
. figure 8(b ) * _ lemma 3.4.3 . _
* if @xmath891 for some @xmath848 , then each component of @xmath892 lies in one of @xmath893 and @xmath894 .
furthermore , if one component of @xmath895 lies in @xmath896 , then @xmath897 . * _ proof .
_ * suppose that @xmath898 . by lemma 3.2.4 , each component of @xmath899
is contained in @xmath900 for some @xmath901 and @xmath664 .
furthermore , @xmath902 . if @xmath903 , then , by lemma 3.2.2 and definition 3.2.1 , @xmath896 and @xmath904 are disjoint from @xmath900 . in this case , if @xmath905 , then @xmath906 . by the proof of lemma 3.4.2 , this is impossible . q.e.d . * _ lemma 3.4.4 . _ * if @xmath907 , then @xmath879 is disjoint from @xmath908 . *
_ * by lemma 3.3.1 , @xmath909 .
we may assume that @xmath910 .
if @xmath911 is disjoint from @xmath836 , then @xmath912 is disjoint from @xmath913 .
suppose now that @xmath914 .
now by proposition 4(1 ) and definition 2.1.4 , each component of @xmath915 is a disk in @xmath836 .
if one component of @xmath916 , say @xmath917 , is not disjoint from @xmath879 , then either one of the two arcs @xmath918 and @xmath919 lies in @xmath879 as in figure 9(a ) or one of @xmath743 and @xmath920 lies in @xmath917 as in figure 9(b ) . here
@xmath921 is defined in definition 3.4.1 .
we first suppose that one of @xmath918 and @xmath922 , say @xmath918 , lies in @xmath879 . by definition 2.1.4 ,
hence @xmath924 . by lemma 3.2.4
, there is @xmath208 such that a copy of @xmath925 is contained in @xmath926 , and @xmath927 , contradicting lemma 3.4.2 .
assume that one of @xmath743 and @xmath920 , say @xmath743 , lies in @xmath917 as in figure 9(b ) .
then @xmath928 .
note that @xmath929 . by lemma 3.1.2(2 ) , this is impossible .
q.e.d.0.5 mm figure 9 * _ lemma 3.4.5 .
_ * for each @xmath930 and @xmath931 , @xmath932 is disjoint from @xmath933 .
furthermore , if @xmath934 , then @xmath879 is disjoint from @xmath933 . * _ proof . _ * by lemma 3.2.4 , each component of @xmath935 is contained in @xmath936 , where @xmath937 and @xmath650 .
hence @xmath938 .
since @xmath939 , by lemma 3.3.1 , @xmath940 and @xmath941 are disjoint from @xmath942 .
thus if @xmath943 , then one component of @xmath935 lies in @xmath933 .
hence @xmath944 lies in @xmath933 for some @xmath945 . by lemma 3.4.2
, this is impossible .
note that @xmath946 and @xmath947 .
if @xmath948 , then , by lemma 2.1.4 , @xmath949 .
q.e.d.0.5 mm * _ lemma 3.4.6 .
_ * ( 1 ) if @xmath950 , then @xmath951 .
furthermore , if @xmath952 , then @xmath890 or @xmath953 .
\(2 ) if @xmath954 for @xmath955 and @xmath956 , then @xmath957 . furthermore , @xmath958 or @xmath953 . * _ proof_*. ( 1 ) by lemma 3.3.1 and definition 3.4.1 , @xmath959 .
if @xmath950 , then @xmath960 ; otherwise , @xmath961 . by proposition 4(3 ) , @xmath962 . by proposition 4(2 ) and definition 2.1.5 , each component of @xmath963 is a core of @xmath836 which lies in @xmath729 .
thus @xmath964 . by lemma 3.4.2 ,
( 1 ) holds .
\(2 ) now suppose that @xmath965 and @xmath956 .
then each component of @xmath966 is @xmath967 where @xmath630 is a core of @xmath836 . by definition 2.1.4 ,
hence @xmath969 . by lemma 3.4.3 ,
@xmath970 . furthermore ,
if @xmath971 , then @xmath958 . in this case
, each component of @xmath972 is @xmath973 when @xmath974 .
hence ( 2 ) holds .
q.e.d.0.5 mm * _ lemma 3.4.7 .
_ * ( 1 ) suppose that @xmath845
. then @xmath975 and @xmath976 .
\(2 ) @xmath843 is a disk in @xmath564 .
\(3 ) for @xmath977 , let @xmath978 where @xmath979 and @xmath980 .
\(i ) if @xmath981 , then @xmath982 . furthermore , if @xmath983 for each @xmath984 , then @xmath985 for some @xmath984 , \(ii ) if @xmath954 for @xmath801 and @xmath986 , then @xmath987 .
furthermore , @xmath988 for some @xmath984 . * _ proof .
_ * by lemma 3.3.1 , if @xmath594 , then @xmath989 .
hence @xmath990 if @xmath594 . by lemmas 3.4.4 and 3.4.5 , @xmath991 is a disk in @xmath564 .
see in definition 3.4.1 .
we first prove that @xmath992 and @xmath993 .
suppose , otherwise , that @xmath994 . then @xmath995 and @xmath996 for @xmath997 .
furthermore , @xmath998 for @xmath997 .
now let @xmath999 , @xmath1000
. then @xmath1001 and @xmath1002 are two disks .
now by lemma 3.2.2 and lemma 3.4.2 , @xmath1003 is disjoint from @xmath1004 , otherwise , @xmath1005 .
hence @xmath1006 .
since @xmath998 , by lemma 3.4.3 , @xmath1007 is disjoint from @xmath1008 .
similarly , @xmath879 is disjoint from @xmath1009 . by the proof of lemma 3.4.4 ,
@xmath879 is disjoint from @xmath1010 for @xmath594 .
note that @xmath1011 ; otherwise , @xmath1012 separates @xmath1013 in @xmath836 , and @xmath1014 separates @xmath1015 and @xmath1016 in @xmath1017 , but by the proof of lemma 3.4.2 , this is impossible .
now @xmath1018 is a disk in @xmath564 .
hence @xmath1019 separates @xmath1020 and @xmath1021 in @xmath1022 . since @xmath564 is generated by @xmath1023 .
so if @xmath1024 for some @xmath194 , then @xmath1025 is disjoint from @xmath602 . now
if @xmath1026 , then @xmath1027 .
this means that @xmath1028 , @xmath1029 , a contradiction . similarly , @xmath1030 .
let @xmath1031 and @xmath1032 .
let @xmath1033 , @xmath1034
. then @xmath1035 and @xmath1036 .
by the above argument , @xmath1037 is a disk in @xmath564 . by lemma 3.4.2 ,
lemma 3.4.3 and lemma 3.4.6 , ( 3 ) and ( 4 ) hold . by induction , @xmath845
. then @xmath975 and @xmath976 .
furthermore , @xmath1038 is a disk as in figure 10(a ) .
( 3 ) is immediately from lemmas 3.4.2 , 3.4.3 and 3.4.6 .
q.e.d .
figure 10(a ) figure 10(b ) * _ lemma 3.4.8 . _ * @xmath1039 is a disk . * _ proof . _ * by lemma 3.4.7 , @xmath1040 if @xmath1041 or @xmath1042 .
now suppose that @xmath1043 .
then @xmath1044 . since @xmath1045 , @xmath1046 for @xmath1047 and @xmath1048 . now if @xmath1049 , then @xmath1050 is disjoint from @xmath1004 .
otherwise , one of @xmath830 and @xmath1051 , say @xmath830 , separates @xmath746 and @xmath1052 in @xmath747 , but by the proof of lemma 3.4.2 , @xmath840 separates @xmath749 and @xmath1053 in @xmath750 , a contradiction .
similarly , if @xmath1054 , then @xmath1055 is disjoint from @xmath879 . if @xmath1056 , then @xmath1057 .
since @xmath1058 for each @xmath1059 and @xmath1060 , by the proof of lemma 3.4.4 , @xmath1061 is disjoint from @xmath1062 . by lemma 3.4.3 , @xmath1061 is disjoint from @xmath1063 .
now @xmath1039 is a disk as in figure 10(b).q.e.d . for each @xmath740 , recalling the equality : @xmath741 . * _ definition 3.5.1 . _ * for @xmath1048 , \(1 ) let @xmath1066 , \(2 ) let @xmath1067.0.5 mm * _ lemma 3.5.2 .
_ * ( 1 ) @xmath1068 and @xmath1069 .
\(2 ) if @xmath1041 , then @xmath1070 .
\(3 ) if @xmath1042 , then @xmath1070 . * _ proof . _
* the lemma follows from definition 3.5.1 and lemma 3.4.7.q.e.d.0.5 mm * _ lemma 3.5.3 . _ * ( 1 ) suppose that @xmath1071 or @xmath1072 for each @xmath132 . if @xmath1073 , then @xmath1074 .
\(2 ) suppose that @xmath1075 . if @xmath1076 , then either @xmath1077 or @xmath897 .
* _ proof .
_ * by definition 3.5.1 , @xmath1066 . by lemma 3.3.4 , @xmath1078 , @xmath1079 for @xmath1080 .
\(1 ) suppose that @xmath1071 or @xmath1081 for each @xmath132 . in this case , by definition 3.2.1 , @xmath1082 . if @xmath1083 , then , by proposition 4(3 ) , @xmath1064 is disjoint from @xmath582 .
so @xmath1084 . by lemma 3.3.1
, @xmath1085 is disjoint from @xmath582 .
note that @xmath1078 for @xmath1086 .
now by the proof of lemma 3.2.3 , @xmath1087 is disjoint from @xmath582 for @xmath1088 .
\(2 ) suppose that @xmath1075 , @xmath1089 and @xmath1090 . if @xmath1084 , then , by ( 1 ) , @xmath1087 is disjoint from @xmath582 .
now suppose that @xmath1091 . since @xmath1092 , by lemma 2.2.5 , @xmath1081 . by proposition 4(3 ) , @xmath1068 is disjoint from @xmath582 . since @xmath1093 , each component of @xmath972 is @xmath1094 where @xmath1095 .
now if @xmath1090 , then @xmath1096 , a contradiction.q.e.d.0.5 mm * _ lemma 3.5.4 .
_ * if @xmath1097 and @xmath1098 , then @xmath1099 is disjoint from @xmath1100 . * _ proof . _
* there are three cases : case 1 .
@xmath1101 for each @xmath132 with @xmath458 .
now by definition 3.2.1 , @xmath1102 and @xmath1103 . since @xmath1104 , the lemma holds .
@xmath1072 for each @xmath1105 with @xmath458 , and @xmath1106 for some @xmath1107 with @xmath1108 . since @xmath1109 , @xmath1110 . since @xmath1078 for @xmath1086 .
hence @xmath1111 for @xmath1060 , and @xmath1112 . by the definition , @xmath1066 and @xmath1113 .
now if @xmath1114 , then , by proposition 4(3 ) , @xmath1064 is disjoint from @xmath1115 .
suppose that @xmath1116 . by proposition 4(2 ) , either @xmath1117 or each component of @xmath1118 is a core of @xmath1119 .
hence either @xmath1120 or each component of @xmath1121 is a core of @xmath1100 .
by lemma 3.2.2 , @xmath1122 is disjoint from @xmath1123 .
now if @xmath1124 , then @xmath1125 . by lemma 3.5.3(1 ) , this is impossible .
@xmath1075 , @xmath1108 . now if @xmath1126 , then the lemma holds .
suppose that @xmath1127 . by proposition 4(1 ) , each component of @xmath1128 is @xmath1129 where @xmath1130 is a core of @xmath1131 . by ( 2 ) ,
@xmath1132 is disjoint from @xmath1099 .
hence the lemma holds .
q.e.d.0.5 mm * _ lemma 3.5.5 .
_ * for each @xmath1048 , there are an arc @xmath1133 in @xmath1134 and an arc @xmath1135 in @xmath1136 such that \(1 ) @xmath1137 and @xmath1138 , @xmath1139 and @xmath1140 , \(2 ) @xmath1141 bounds a disk @xmath1142 in @xmath564 .
\(3 ) @xmath1143 , @xmath1144 .
\(4 ) @xmath1145 and @xmath1146 . furthermore ,
if @xmath1147 , then @xmath1148 for some @xmath984 .
* _ proof . _
* we may assume that @xmath1149 .
by lemma 3.4.7 , for @xmath1150 , @xmath1151 . by lemma 3.4.7(3 ) and definition 3.5.1 , @xmath1152 . by lemma 3.4.7 and lemma 3.5.4 ,
@xmath1153 for @xmath1154 .
now let @xmath1133 be in @xmath1155 such that @xmath1156 and @xmath1157 as in figure 11(a ) .
in fact , @xmath1133 can be obtained by pushing @xmath1158 slightly where @xmath1159 as in figure 11(a ) .
then @xmath1133 is disjoint from @xmath1160 .
similarly , let @xmath1135 be in @xmath1161 such that @xmath1162 and @xmath1163 as in figure 11(a ) .
hence @xmath1164 bounds a disk @xmath1165 in @xmath564 .
figure 11(a ) suppose that @xmath1166
. then , by lemma 3.1.6 and lemma 3.2.2 , one component of @xmath1167 , say @xmath630 , lies in @xmath1168 .
if @xmath1169 is disjoint from @xmath1170 , then @xmath1133 can be isotoped to be disjoint from @xmath630 .
if one end point of @xmath630 lies in @xmath1170 , then , by the proof of lemma 3.4.1 , ( 2 ) holds . in this case , either @xmath1171 or @xmath985 for some @xmath1172 .
\(3 ) follows from lemma 3.2.4 and the proof of lemma 3.4.3 .
q.e.d .
figure 11(b ) * _ lemma 3.5.6 . _ * @xmath1173 is a disk .
* _ proof . _ * by lemma 3.4.8 , @xmath1174 is a disk . by lemma 3.5.4 ,
lemma 3.4.7 , @xmath1175 for @xmath1176 as in figure 10(b ) .
we may assume that @xmath1177 .
then , by lemma 3.4.7 , @xmath1178 for each @xmath1059 . then @xmath1179 .
furthermore , @xmath1180 for each @xmath1059 . by the proof of lemma 3.5.5
, @xmath1173 is a disk as in figure 11(b).q.e.d . * _ lemma 3.5.7 . _ * ( 1 ) for each @xmath1048 , @xmath1181 .
\(2 ) if @xmath1182 , then @xmath1183 is disjoint from @xmath1165 . * _ proof .
_ * by lemma 3.5.4 , @xmath1184 . since @xmath1045 for each @xmath1048 .
hence , by proposition 4(3 ) , @xmath1185 is disjoint from @xmath1099 for @xmath1048 . by lemma 3.5.5(3 ) and lemma 3.4.7 , @xmath1186 is disjoint from @xmath1165 .
suppose that @xmath1182 .
then @xmath1187 for each @xmath1059 .
hence @xmath1183 is disjoint from @xmath1188 . by lemma 3.5.5(3 ) and lemma 3.4.7 , ( 2 ) holds.q.e.d .
figure 12 * _ lemma 3.5.8 . _ * ( 1 ) if @xmath1189 , @xmath579 intersects @xmath1190 in one point if and only if @xmath579 intersects @xmath1191 in one point .
\(2 ) if @xmath1192 , then @xmath579 intersects @xmath1133 in a one point if and only if @xmath579 intersects @xmath1193 in one point , @xmath579 intersects @xmath1135 in one point if and only if @xmath579 intersects @xmath1194 in one point . * _ proof . _ * ( 1 ) since @xmath1189 , @xmath1195 and @xmath1196 . hence @xmath1197 and @xmath1198 as in figure 12(a ) . by lemma 3.1.2 , @xmath1199 . by proposition 4(4),(5 ) , @xmath1200 is disjoint from @xmath1201 , and @xmath579 is properly embedded in @xmath602 . since @xmath1202 , @xmath538 intersects each of @xmath1203 and @xmath749 in one point , we can moved @xmath538 so that @xmath1190 is disjoint from @xmath538 . similarly ,
if @xmath579 intersects each of @xmath1203 and @xmath749 in one point , then @xmath579 is disjoint from @xmath1190 .
hence ( 1 ) holds .
\(2 ) by lemma 3.3.1 , @xmath1204 and @xmath1205 are properly embedded arcs in @xmath1206 and @xmath1207 and @xmath1208 are properly embedded arcs in @xmath1209 . since @xmath1210 , by lemma 3.3.1 , @xmath1211 . by proposition 4(5 ) , @xmath1201 is disjoint from @xmath582 and @xmath568 .
hence @xmath579 is disjoint from @xmath1212 .
note that @xmath1213 and @xmath1214 as in figure 12(b ) . by the argument in ( 1 ) , ( 2 ) holds .
q.e.d . *
_ remark 3.5.9 .
_ * since propositions 1 - 6 hold for @xmath1215 . by lemma 3.1.5 , @xmath1216 for @xmath492 . by propositions 1 and 2 , all the arguments in sections 3.1 - 3.5 are true when we take place of @xmath1217 with @xmath1218 , @xmath458 with @xmath413 in lemma 3.1.6 and lemma 3.2.3 , @xmath705 with @xmath1219 in lemma 3.2.4 .
assume now that propositions 1 - 6 hold for each @xmath1220 now we only need to prove that propositions 1 - 6 hold for @xmath493 .
there are two cases : \1 .
@xmath494 .
@xmath495 . in this chapter
, we shall prove that propositions 4 - 6 hold for case 1 .
hence , in this chapter , we assume that @xmath494 .
we first construct @xmath1221 from @xmath1222 , then we shall prove propositions 4 - 6 for this case . since propositions 1 - 6 hold for @xmath492 .
so @xmath1223 , @xmath1224 , and @xmath1225 are well defined for @xmath492 .
recall that @xmath1226 defined in definition 3.3.0 . *
_ definition 4.1.1 .
_ * let @xmath533 . in particular ,
if @xmath531 , then let @xmath1227 . without loss of generality ,
we assume that @xmath786 .
recalling the argument in section 3.3 , we can number all the elements in @xmath737 as @xmath1228 according to the order of @xmath739 lying in @xmath538 . in some time
, we shall use @xmath1229 to take place of @xmath1226 . by lemma 3.3.1 and remark 3.3.2 , for each @xmath1230 , @xmath1231.0.5 mm * _ lemma 4.1.2 .
_ * suppose that @xmath1232 .
\(1 ) in @xmath564 , @xmath538 intersects @xmath1185 in one point . furthermore , if @xmath1182 , then @xmath1183 is disjoint from @xmath538 ; if @xmath1233 , then @xmath720 is disjoint from @xmath538 .
\(2 ) in @xmath529 , @xmath528 intersects @xmath1234 in one point .
furthermore , if @xmath1182 , then @xmath528 is disjoint from @xmath1235 ; if @xmath1233 , then @xmath1236 is disjoint from @xmath528 . * _ proof . _ * by assumptions , propositions 1 - 6 hold for @xmath1237 .
\(1 ) by proposition 4(4 ) , @xmath538 is a properly embedded arc in @xmath564 lying in @xmath750 . since @xmath1238 , @xmath1203 intersects @xmath538 in one point . by proposition 4(5 ) and lemma 3.1.2 , @xmath1239 .
suppose , otherwise , @xmath1185 intersects @xmath538 in at least two points .
then , by lemma 3.1.6 , @xmath796 for some @xmath1182 , contradicting the minimality of @xmath1229 .
similarly , if @xmath1182 , then @xmath538 is disjoint from @xmath1183 . by lemma 3.2.4 , if @xmath1233 , then @xmath582 is disjoint from @xmath538 .
\(2 ) follows from remark 3.5.9 and propositions 1 and 2 .
q.e.d.0.5 mm * _ lemma 4.1.3 .
_ * ( 1 ) if @xmath1232 , then @xmath1240 .
\(2 ) @xmath1241 is a properly embedded disk in @xmath47 such that @xmath1242 intersects @xmath538 in one point which lies in @xmath1243 .
* _ proof . _ * ( 1 ) now consider @xmath529 . by assumptions , propositions 1 - 3
hold for @xmath1237 .
suppose , otherwise , @xmath1244 . by assumption ,
@xmath494 . by proposition 3
, @xmath534 is a properly embedded disk in @xmath26 and @xmath1241 is a properly embedded disk in @xmath46 . by definition 2.3.1 and lemma 2.2.4 , @xmath1245 . by lemma 3.1.5 , @xmath530 .
note that @xmath1246 . by proposition 3 , @xmath1247 .
if @xmath1248 , then , by lemma 2.2.4 , @xmath1182 . by lemma 4.1.2(2 ) , @xmath1235 is disjoint from @xmath528 . by also lemma 4.1.2(2 ) ,
@xmath534 intersects @xmath1241 in only one point .
hence @xmath1249 is stabilized , a contradiction .
\(2 ) by ( 1 ) and proposition 6 , @xmath1241 is a properly embedded disk in @xmath47 . by proposition 6 , @xmath1250 . since @xmath1251 , @xmath1182 . by lemma 4.1.2
, @xmath1252 intersects @xmath538 in one point lying in @xmath1243 .
q.e.d.0.5 mm * _ lemma 4.1.4 .
_ * suppose that @xmath1232 .
\(1 ) @xmath1253 for each @xmath1254 with @xmath458 .
furthermore , if @xmath1248 , then @xmath1255 for each @xmath132 .
\(2 ) @xmath1185 is properly embedded in @xmath564 , @xmath1183 is properly embedded in @xmath564 for @xmath1251 . * _ proof . _ * by lemma 4.1.3 , @xmath1240 .
now if @xmath1256 for some @xmath132 , then , by lemma 2.2.5 , @xmath413 . by lemma 3.1.1
, @xmath1185 is regular in @xmath564 . by definition 2.1.5
, @xmath1185 is properly embedded in @xmath564 .
suppose that @xmath1248 . by definition 2.3.3 ,
hence @xmath1258 . by lemma 2.2.5 , @xmath1259 for each @xmath132 . by lemma 3.1.1 and definition 2.1.5
, @xmath1183 is properly embedded in @xmath564.q.e.d .
0.5 mm * _ lemma 4.1.5 . _
* suppose that @xmath1232 .
then there is a regular neighborhood of @xmath1242 in @xmath448 , say @xmath1260 , satisfying the following conditions : \(1 ) @xmath1261 , @xmath1262 .
\(2 ) if @xmath1263 , then @xmath1264 is disjoint from @xmath582 , if @xmath1251 and @xmath709 , then @xmath1265 is disjoint from @xmath582 .
\(3 ) if @xmath1266 , then each component of @xmath1267 is @xmath1268 where @xmath1269 is a core of @xmath582 , if @xmath1248 and @xmath1270 , then each component of @xmath1271 is @xmath1268 where @xmath1272 is a core of @xmath582 .
\(4 ) if @xmath1273 and @xmath1274 , then @xmath568 is disjoint from @xmath1260 .
\(5 ) for each @xmath1275 , @xmath1276 is disjoint from @xmath1260 .
\(6 ) if @xmath1266 , then @xmath1277 is disjoint from @xmath582 for @xmath1278 .
\(7 ) for each @xmath548 with @xmath288 , @xmath1279 is disjoint from @xmath1280 , for each @xmath546 with @xmath293 , @xmath1281 is disjoint from @xmath1282 .
\(8 ) @xmath1283 . * _ proof . _ * now @xmath1284 . by lemma 3.1.1 and lemma 4.1.4(1 ) , @xmath1285 are regular in @xmath564 . by definition 2.1.5 and lemma 2.1.6 , @xmath1286 are disjoint from @xmath582 for each @xmath132 . hence ( 1 ) follows from proposition 6(1 ) .
( 2 ) follows from proposition 4(3 ) and lemma 4.1.4(1 ) .
( 3 ) follows from proposition 4(2 ) and definition 2.1.5 . by proposition 4
, @xmath1287 is a set of pairwise disjoint arcs in @xmath564 . since @xmath1273 and @xmath1274 , @xmath1288 . by ( 1 ) , ( 4 ) holds . since @xmath1289 , by proposition 4(5 ) , @xmath1290 is disjoint from @xmath1242 for each @xmath1275 . hence ( 5 ) holds .
( 6 ) follows from lemma 3.1.2 and lemma 3.3.1 .
( 7 ) follows from proposition 6 .
( 8) follows from lemma 3.5.7 .
q.e.d.0.5 mm in this section , we shall prove propositions 4 - 6 for the speical case : @xmath493 and @xmath713 .
* _ the proof of proposition 4 .
_ * suppose that @xmath713 .
now by corollary 3.2.5 , for each @xmath649 and @xmath132 , @xmath568 and @xmath582 are disjoint from @xmath1291 where @xmath1291 is a regular neighborhood of @xmath538 in @xmath750 .
obviously , @xmath1292 is disjoint from @xmath1291 for @xmath548 .
now let @xmath1293 .
we denote by @xmath1294 the disk @xmath602 for @xmath1295 , @xmath1296 the two components of @xmath1297 .
let @xmath1298 , @xmath1299 for @xmath649 , @xmath1300 for @xmath1301 .
specially , let @xmath531 , @xmath1302 .
now if @xmath1303 , then @xmath1304 lies in one of @xmath1305 and @xmath1306 .
similarly , if @xmath1307 , then @xmath735 lies in one of @xmath1305 and @xmath1306 . since @xmath531 , @xmath1308 .
hence proposition 4 holds.q.e.d.0.5 mm * _ the proof of proposition 5 .
_ * by the above argument , @xmath1309 , @xmath1298 for @xmath548 .
note that @xmath1310 .
hence proposition 5 holds.q.e.d.1.5 mm * _ the proof of proposition 6 .
_ * since @xmath531 , @xmath1227 . by definition 3.2.1 , @xmath1311
now we denote by @xmath571 the disk @xmath1312 for @xmath649 with @xmath293
. then @xmath1313 . by definition 2.3.1 ,
we denote by @xmath553 the disk @xmath1315 for @xmath548 with @xmath288 .
since @xmath538 is disjoint from @xmath1316 and @xmath567 .
hence @xmath1317 , and @xmath1318.q.e.d . by the argument in section 4.2 , in the following argument
, we shall assume that @xmath537 .
in this section , we shall construct @xmath1304 for @xmath548 , @xmath567 for @xmath1319 , @xmath1320 for @xmath1321 , and @xmath1316 from @xmath1322 .
let @xmath1260 be a regular neighborhood of @xmath1242 in @xmath514 in lemma 4.1.5 .
let @xmath1291 be a regular neighborhood of @xmath538 in @xmath750 . by lemma 4.1.3
, @xmath1252 intersects @xmath1291 in an arc @xmath1323 , and @xmath1260 intersects @xmath1291 in @xmath96 .
( see figure 13 . )
figure 13 * _ definition 4.3.1 . _
* let @xmath1324.0.5 mm * _ lemma 4.3.2 . _ *
@xmath1316 is a compact surface in @xmath514 such that @xmath538 is disjoint from @xmath1316 . * _ proof . _ * by lemma 4.1.4 , @xmath1185 and @xmath1183 for @xmath1248 are properly embedded in @xmath564 . by lemma 4.1.5(1 ) and lemma 4.1.3 , the lemma holds.q.e.d.0.5 mm * _ definition 4.3.3 . _
* let @xmath1325 for @xmath1295 .
let @xmath1305 be the component of @xmath1297 containing @xmath1326 , @xmath1327 be the component of @xmath1328 containing @xmath1329 .
0.5 mm * _ definition 4.3.4 . _
* let @xmath1298 for @xmath548.0.5 mm * _ construction ( * ) _ * by lemma 3.3.3 , in @xmath564 , @xmath1330 where @xmath1331 is defined in definition 3.2.1 . by lemma 4.1.4 ,
note that @xmath1246 .
hence , @xmath1333 . by definition 3.5.1 , @xmath1334 .
suppose that @xmath1048 . if @xmath1335 for @xmath1336 , then @xmath1337 . if @xmath1081 for each @xmath132 , then @xmath1338 . by lemma 4.1.5(2 ) and ( 4 ) , @xmath1260 is disjoint from @xmath1339 for @xmath1340 . by lemma 4.1.5(8 ) , @xmath1341 .
now @xmath1342 separates @xmath1343 into two disks @xmath1344 and @xmath1345 .
we may assume that \(0 ) @xmath1346 $ ] for @xmath1149 , and @xmath1347 $ ] for @xmath1348 .
now @xmath1349 is as in figure 14 .
figure 14 figure 15 now let @xmath1350 $ ] be a sub - interval of @xmath1351 $ ] such that \(1 ) if @xmath1041 , then @xmath1352 .
\(2 ) if @xmath1353 , then @xmath1354 .
now @xmath1355 is as in figure 15 .
let @xmath1356 be the center point of @xmath1357 , and @xmath1358 be an arc connecting @xmath1359 to @xmath1360 in @xmath1361 such that \(3 ) @xmath1362 , where @xmath1133 is as in lemma 3.5.5 and @xmath1363 is an arc in @xmath1364 as in figure 16 . by definition 3.5.1 , @xmath1365 .
similarly , @xmath1366 be an arc connecting @xmath1367 to @xmath1368 in @xmath1369 such that \(4 ) @xmath1370 and @xmath1371 is an arc in @xmath1372 .
figure 16 figure 17 by definition 3.5.1 , @xmath1373 . for @xmath1048 , let @xmath1374 as in figure 17
. then @xmath1375 is an arc in @xmath1316 by lemma 3.5.8 .
now let @xmath1376 be a neighborhood of @xmath1375 in @xmath1316 satisfying the following conditions : \(5 ) @xmath1377 .
\(6 ) if @xmath1075 , then @xmath1378 ; if @xmath1071 or @xmath1379 , then @xmath1380 .
if @xmath1381 with @xmath1075 , then there is a homeomorphism @xmath1382 from @xmath1099 to @xmath1376 such that @xmath1382 is an identifying map on @xmath1383 .
if @xmath1071 or @xmath1379 , then there is a homeomorphism @xmath1382 from @xmath1188 to @xmath1384 . in this case ,
( see definition 3.2.1.)q.e.d.(construction ( * ) ) * _ lemma 4.3.5 .
_ * ( 1 ) if @xmath1386 , then @xmath1387 .
\(2 ) @xmath1388 is an injective map from @xmath1389 to @xmath1390 .
\(3 ) if @xmath546 , and @xmath1391 , then @xmath1392 . furthermore , either @xmath1171 or @xmath985 for some @xmath1393 .
\(4 ) @xmath1394 . if @xmath1395 , then @xmath1396 , and @xmath1148 for some @xmath1393 .
\(5 ) @xmath1397 . * _ proof .
_ * ( 1 ) by construction ( * ) ( 5 ) and ( 6 ) , we only need to prove that @xmath1398 . without loss of generality ,
we assume that @xmath1041 . by lemma 3.4.7 , @xmath1040 .
suppose that @xmath1399 . since @xmath1079 for @xmath1060 @xmath1400 for @xmath1401 . by lemma 3.3.4 , @xmath1402 .
@xmath1403 .
suppose that @xmath1404 .
since @xmath1064 separates @xmath1405 and @xmath1132 in @xmath1406 .
now by the choice of @xmath1357 , ( 1 ) holds .
\(2 ) by lemma 3.5.4 , @xmath1407 for @xmath1098 . by ( 1 ) ,
hence ( 2 ) holds .
( 3 , 4 ) by construction ( * ) ( 3 ) and ( 4 ) , @xmath1409 and @xmath1410 is an arc in @xmath1411 . by lemma 4.1.5(2 ) , ( 4 ) and ( 6 ) , @xmath568 and @xmath582 are disjoint from @xmath1412 for @xmath1413 and @xmath1301 . now by lemma 3.5.5 , ( 3 ) and ( 4 ) holds .
\(5 ) since @xmath1414 , @xmath1415 . by lemma 3.5.8 , @xmath1416 is disjoint from @xmath1291 . hence ( 5 ) holds .
q.e.d . * _ definition 4.3.6 . _
* let @xmath1417 for @xmath546 .
0.5 mm * _ lemma 4.3.7 .
_ * @xmath1418 is a set of pairwise disjoint arcs properly embedded in @xmath1316 which lies in one of @xmath1294 . * _ proof . _ * since @xmath548 , @xmath1298 is disjoint from @xmath1291 .
hence @xmath1419 lies in @xmath1294 for some @xmath194 by proposition 4(4 ) .
furthermore , if @xmath1420 , then @xmath1419 lies in one of @xmath1305 and @xmath1306 . by proposition 4(4 ) and lemma 4.1.5(5 ) , @xmath1419 is properly embedded in @xmath1316 .
q.e.d.0.5 mm * _ lemma 4.3.8 .
_ * ( 1 ) @xmath1421 is a set of pairwise disjoint arcs in @xmath1316 such that @xmath1422 .
\(2 ) if @xmath1423 and @xmath1424 for each @xmath1059 , then @xmath1425 .
\(3 ) if @xmath1426 for some @xmath1176 , then @xmath1427 .
\(4 ) if @xmath1428 , then @xmath1429 . * _ proof . _
* ( 1 ) by proposition 4(2 ) and ( 3 ) , each component @xmath1430 is a core of @xmath1131 if @xmath1431 .
hence each component of @xmath1432 is a core of @xmath1099 even if @xmath858 . by lemma 4.3.5 ,
@xmath1433 and @xmath1434 . by lemma 3.3.3 and lemma 4.1.4 , @xmath1435 . by definition 3.5.1 ,
@xmath1334 . by lemma 4.3.5(5 ) , @xmath567 is an arc in @xmath1316 . by proposition 4 ,
@xmath1436 for @xmath667 . since @xmath1437 is injective , @xmath1438 .
\(2 ) if @xmath889 for each @xmath1048 and @xmath1439 , then , by proposition 4(3 ) , @xmath568 is disjoint from @xmath1440 . hence ( 2 ) holds .
\(3 ) suppose that @xmath1426 for some @xmath1176 . then @xmath1441 .
furthermore , @xmath1442 . by the argument in ( 2 ) , ( 3 ) holds .
\(4 ) if @xmath1428 , then @xmath1443 for each @xmath1048 . by proposition 4(3 ) , @xmath1444 for each @xmath1059 . by ( 2 ) , ( 4 ) holds .
q.e.d.0.5 mm in fact , if @xmath1445 , then @xmath567 is obtained by doing band sums with copies of @xmath1242 to @xmath568 .
see lemma 4.6.3.0.5 mm * _ definition 4.3.9 .
_ * ( 1 ) let @xmath1446 for @xmath132 with @xmath413 .
\(2 ) suppose that @xmath646 for some @xmath1301 with @xmath458 .
if @xmath1447 , then let @xmath1448 .
if @xmath1449 , then let @xmath1450 .
\(3 ) suppose that @xmath132 with @xmath458 .
let @xmath1451 , @xmath1452 .
specially , let @xmath1453 , @xmath1454 .
0.5 mm * _ lemma 4.3.10 . _
* ( 1 ) for each @xmath1321 , @xmath1320 is a disk in @xmath1316 such that @xmath1455 for @xmath132 and @xmath1456 .
\(2 ) suppose that @xmath646 for some @xmath1457 with @xmath458 . if @xmath1458 for each @xmath1059 , then @xmath1459 .
\(3 ) suppose that @xmath1108 .
then @xmath1460 . *
_ * suppose that @xmath413 . by proposition 4(1 ) and ( 5 ) , @xmath582 is disjoint from @xmath1461 . hence @xmath1462 is a disk .
suppose that @xmath458 .
there are four cases : case 1 .
@xmath1233 .
now if @xmath1463 , then @xmath1464 . by lemma 4.1.2 and lemma 4.3.7 , @xmath1465 is disjoint from @xmath1291 . by definition 4.3.9 , @xmath1462 is a disk .
case 2 . by the construction
, @xmath1466 is a disk in @xmath1316 .
@xmath1447 and @xmath1467 for each @xmath1059 .
now @xmath1468 . since @xmath1080 , @xmath1078 .
by lemma 3.5.3 , if @xmath1090 , then , @xmath1469 . by proposition 4(1 ) , @xmath1470 . by definition 2.1.4 ,
each component of @xmath1471 is @xmath1472 where @xmath630 is a core of @xmath1099 . by lemma 3.2.6 , if @xmath623 and @xmath1473 , then @xmath1474 is disjoint from @xmath1188 . hence ( 2 ) holds . by lemma 4.1.5(2 ) , @xmath582 is disjoint from @xmath1475 . by lemma 4.3.5 , if @xmath1476 , then @xmath1477 . since @xmath1437 is injective , @xmath1320 is a disk in @xmath1316 .
@xmath1478 for some @xmath1479 .
now @xmath1480 . by definition 4.3.9 and the argument in case 3
, @xmath1320 is a disk in @xmath1316 and ( 3 ) holds
. q.e.d .
0.5 mm by the construction , @xmath1316 is a surface in @xmath1481 . in this section
, we shall prove that @xmath1482 is an abstract tree and @xmath1316 is a surface generated by @xmath1482 , which satisfies proposition 4(1).0.5 mm * _ lemma 4.4.1 .
_ * ( 1 ) @xmath1483 .
\(2 ) for @xmath1484 , @xmath1485 . * _ proof . _ * by the construction , @xmath1486 , where @xmath1487 . since @xmath1488 , @xmath1489 . hence ( 1 ) follows from the construction . since @xmath564 is generated by @xmath1490 .
hence , by lemma 4.3.10 , @xmath1485 for @xmath692 . by lemma 4.3.10 and lemma 4.1.5(1 ) , @xmath1491 for @xmath132
. q.e.d.0.5 mm * lemma 4.4.2 . *
if @xmath1492 with @xmath402 , then @xmath1493 is disjoint from @xmath1494 . * _ proof .
_ * since @xmath402 , by lemma 4.1.3 , @xmath1495 . by proposition 4(1 ) and ( 5 ) , @xmath582 is disjoint from @xmath1496 . by definition 4.3.9 , @xmath1446 .
suppose that @xmath1497 . by definition 4.3.9 ,
@xmath1300 for @xmath1498 .
note that @xmath1499 and @xmath1325 for @xmath1295 . in this case , @xmath1500 .
hence @xmath1501 is disjoint from @xmath1494 .
suppose that @xmath1502 . by proposition 4(1 ) , @xmath1503 is disjoint from @xmath1504 . by lemma 4.1.5(1 ) and ( 2 ) , @xmath1505 is disjoint from @xmath1506 . by definition 4.3.9 ,
hence @xmath1508 is disjoint from @xmath1494 .
q.e.d.0.5 mm * lemma 4.4.3 .
* suppose that @xmath1492 with @xmath404
. then @xmath1509 . *
_ * by assumption , proposition 4 holds for @xmath1237 .
hence @xmath1510 for @xmath165 with @xmath404 .
since @xmath404 , by lemma 2.2.5 and lemma 4.1.3 , @xmath1511 . by definition 2.3.1 , @xmath1512 .
now there are four cases : case 1 .
@xmath1513 . in this case ,
@xmath1514 . by definition 4.3.9 , @xmath1515 and @xmath1446 for @xmath1516 . by lemma 4.3.8 and lemma 2.2.4 , @xmath1517 for @xmath1518 or @xmath646 .
hence @xmath1519 .
by lemma 4.1.2 , @xmath582 is disjoint from @xmath1291 . by definition 4.3.9 ,
the lemma holds .
@xmath1520 . by definition 4.3.9 ,
@xmath1446 for @xmath1498 . and @xmath1521 .
if @xmath1248 , then @xmath1182 . by lemma 4.3.8
, @xmath1517 is disjoint from @xmath1291 . by lemma 4.1.5(1 ) and ( 2 ) , @xmath1522 .
hence the lemma holds .
@xmath1523 and @xmath1524 for each @xmath1059 . in this case , @xmath1525 . by lemma 4.1.5(1 ) and ( 2 ) , @xmath1505 is disjoint from @xmath1506 . by proposition 4(1 ) , @xmath1526 is disjoint from @xmath1527 . by definition 4.3.9 , for each @xmath1301 , @xmath1528 . hence @xmath1529 is disjoint from @xmath1530 . since @xmath1502 , by lemma 4.3.10(2 ) , @xmath1531 .
hence the lemma holds .
@xmath1532 . by lemma 4.3.5 and construction ( * ) , @xmath1533 . by lemma 4.1.5 and definition 4.3.9 , @xmath1534 . by the argument in case 3 and lemma 4.3.10(3 ) , the lemma holds q.e.d.0.5 mm * _ lemma 4.4.4 .
_ * if @xmath1535 , then each component of @xmath1536 is an arc @xmath1537 which is a core of @xmath1538 , and each component of @xmath1539 is @xmath1540 . *
_ * recalling the assumption that proposition 4 holds for @xmath1237 .
there are five cases : case 1 .
@xmath1541 . now @xmath1300 and @xmath1542 .
since @xmath564 is a surface generated by the abstract tree @xmath1543 , the lemma holds .
( see definition 2.1.4 . )
@xmath1544 .
now @xmath1545 .
by lemma 4.4.3 , @xmath1546 . by lemma 4.3.8 , @xmath1517 . by lemma 4.1.5(1 ) , ( 2),(3 ) and lemma 4.4.1(2 ) , the lemma holds .
@xmath1547 . by definition 4.3.9 ,
@xmath1548 $ ] . by lemma 4.1.5
, each component of @xmath1549 is @xmath1550 where @xmath630 is a core of @xmath724 . by lemma 4.1.5(6 ) , @xmath1277 is disjoint from @xmath1551 .
note that @xmath1552
. hence each component of @xmath1553 is either in @xmath1554 or in @xmath1555 .
since @xmath1556 , by lemma 4.3.5(4 ) , @xmath1557 is disjoint from @xmath582 . since @xmath1558 , by lemma 3.2.6 , @xmath1559 is also disjoint from @xmath582 .
hence each component of @xmath1560 is either in @xmath1561 or in @xmath1562 . since @xmath1502 , @xmath1563 by lemma 3.3.1 . hence if @xmath1564 , then , by lemma 3.5.3 , @xmath1565 . by definition 2.1.4 and proposition 4(1 ) , each component of @xmath1566 is @xmath1567 where @xmath630 is a core of @xmath1099 .
hence each component of @xmath1568 is either in @xmath1188 or in @xmath1569 . by lemma 4.3.10(2 ) , ( 3 ) and lemma 4.4.3 , @xmath1570 where @xmath1571 and @xmath1572 . by the construction , @xmath1573 where @xmath1574 is a core of @xmath1376 . by the above argument
, the lemma holds .
@xmath1575 . by lemma 4.1.5
, @xmath1505 is disjoint from @xmath1454 . by lemma 4.3.10(2 ) , ( 3 ) and lemma 4.4.3 , @xmath1576 .
note that each component of @xmath1577 is @xmath1578 . by definition 4.3.9
, @xmath1579 intersects @xmath1580 in @xmath1581 . since @xmath1069 , by definition 4.3.9 , @xmath1582 . by lemma 4.1.5(8 ) and construction(*)(3),(4 ) , @xmath1583 is disjoint from @xmath1466 . by the construction , @xmath1376 intersects @xmath1466 in @xmath1584 . since @xmath1585 , by proposition 4 for @xmath1237 , the lemma holds .
@xmath1586 . by lemma 4.1.5 ,
@xmath1587 are disjoint from @xmath1588 . by lemma 4.3.5 ,
@xmath1434 . by definition 4.3.10(2 ) , @xmath1589 where @xmath1590 and @xmath1591 . since @xmath1437 is injective , the lemma holds.q.e.d .
* _ lemma 4.4.5 . _ * @xmath1316 is a surface generated by an abstract tree @xmath1592 . * _ proof . _ * by lemmas 4.4.1 - 4.4.4 , we only need to prove that @xmath1593 is an abstract tree . by lemma 4.3.10 ,
@xmath1594 for @xmath132 and @xmath1595 .
let @xmath1596 be a set of pairwise disjoint arcs obtained by pushing @xmath1597 off @xmath1598 in @xmath1599@xmath1600 . by proposition 4(1 ) for @xmath1237 , each component of @xmath1601 is an tree . by definition 4.3.3 ,
@xmath1602 for @xmath1603 , and @xmath1604 .
hence each component of @xmath1605 is also an tree .
note that @xmath1606 . by lemma 3.3.1 ,
since @xmath1185 intersects @xmath538 in one point lying in @xmath1608 , and @xmath1609 , @xmath1610 contains at least two components .
furthermore , @xmath1305 and @xmath1611 lies in one component of @xmath1610 , @xmath1612 and @xmath1613 lies in another component of @xmath1614 as in figure 18 .
since @xmath1615 and @xmath1616 , so each component of @xmath1617 is also an tree . by definition 2.1.2
, @xmath1618 is an abstract tree . q.e.d .
figure 18 * _ lemma 4.5.1 . _
* suppose that @xmath1619 and @xmath1492 .
if @xmath1620 , then either @xmath567 is disjoint from @xmath1621 or @xmath1622 and @xmath404 . * _ proof . _ * by assumption , proposition 4 holds for @xmath1237 .
suppose that @xmath1620 and @xmath1623 .
then , by proposition 4(3 ) , @xmath568 is disjoint from @xmath1505 for @xmath165 .
now there are two cases : case 1 .
@xmath1624 .
suppose that @xmath1625 . since @xmath1626 , by lemma 4.1.5 , lemma 4.3.7 , @xmath1299 is disjoint from @xmath1627 .
suppose @xmath165 .
since @xmath1428 , by lemma 4.3.5(3 ) , @xmath1299 is disjoint from @xmath1628 .
hence @xmath567 is disjoint from @xmath1376 . by definition 4.3.9 ,
hence @xmath567 is disjoint from @xmath1621 .
@xmath1630 . since @xmath1631 , @xmath1502 . by lemma 4.1.5 ,
@xmath1632 are disjoint from @xmath1279 .
if @xmath1439 , then , by lemma 3.5.3 , @xmath568 is disjoint from @xmath1188 except @xmath858 .
now there are two sub - cases : \(1 ) @xmath889 for each @xmath740 . by lemma 4.3.7(2 ) , @xmath1633 . by construction(*)(3 ) , ( 4 ) and lemma 4.3.5(3 ) , @xmath1634 is disjoint from @xmath1628 . by lemma 4.3.10 and lemma 4.4.3 , @xmath1635 . by lemma 4.3.5(4 ) , @xmath1636 is disjoint from @xmath1628 . by lemma 4.3.10(2 ) , @xmath1637 where @xmath1638 and @xmath1639 .
since @xmath1437 is injective , by proposition 4(3 ) for @xmath1237 , the lemma holds .
\(2 ) @xmath1640 for some @xmath1641 .
now @xmath1642 .
by lemma 4.3.10 and lemma 4.4.3 , @xmath1643 . by assumption ,
hence , by the same argument in ( 1 ) , the lemma holds.q.e.d.0.5 mm * _ lemma 4.5.2 .
_ * if @xmath1645 , then each component of @xmath1646 is a core of @xmath1647 . * _ proof . _ * by proposition 4(2 ) for @xmath1237 and the same argument as that in the proof of lemma 4.4.4 , the lemma holds.q.e.d.0.5 mm * _ lemma 4.5.3 .
_ * for each @xmath1648 , @xmath1649 satisfying the following conditions : \(1 ) @xmath629 is as in lemma 3.1.2 , and @xmath1650 is a core of @xmath1651 for @xmath1652 .
\(2 ) @xmath619 is as in lemma 3.1.2 .
furthermore , if @xmath1653 , then @xmath735 is either properly embedded in @xmath1305 or @xmath1306 which is disjoint from @xmath1654 , if @xmath1655 , then @xmath590 is properly embedded in @xmath1656 which is disjoint from @xmath1657 for @xmath607 .
\(3 ) @xmath567 is regular in @xmath1658 . *
_ * by lemma 3.1.2 , in @xmath564 , @xmath584 such that @xmath628 and @xmath590 is disjoint from @xmath592 .
since @xmath1310 , @xmath603 is disjoint from @xmath1659 even if @xmath1660 .
hence @xmath735 lies in one of @xmath1305 and @xmath1306 .
suppose that @xmath612 .
there are two cases : \(1 ) @xmath1428 .
now @xmath1661 .
by lemma 4.3.7 and definition 4.3.9 , @xmath1662 and @xmath1299 . furthermore , @xmath1446 for @xmath612 . in this case
, @xmath586 is also a core of @xmath1651 . by lemma 3.1.2
, the lemma holds .
\(2 ) @xmath1445 . by lemma 4.1.5
, @xmath1663 is disjoint from @xmath1664 .
now we claim that @xmath1665 is disjoint from @xmath1666 . by construction(*)(3 ) and ( 4 ) , if @xmath1667 , then @xmath1668 for some @xmath1669 . by assumption , @xmath983 . by lemma 4.3.5 , @xmath985 , contradicting lemma 3.1.2 . if @xmath612 , then , by lemma 3.1.2 , @xmath590 is disjoint from @xmath1670 . by definition 4.3.9 ,
hence @xmath590 is disjoint from @xmath1493 .
now @xmath1649 where @xmath1672 is a core of @xmath1673 .
now each component of @xmath1674 is @xmath1650 for @xmath1652 . by lemma 4.5.1 and 4.5.2 , ( 1 ) and
( 2 ) holds . by lemma 4.3.7 , @xmath1422 . by lemma 4.3.10 , @xmath1675 for @xmath132
hence @xmath1676 is disjoint from @xmath1320 . since @xmath1445 , @xmath1274 , by lemma 4.1.5(4 ) , @xmath1677 is disjoint from @xmath1678 . by lemma 4.3.7(2 ) and definition 4.3.9 , @xmath568 is properly embedded in @xmath1658 . by lemma 3.1.2 ,
@xmath1679 and @xmath1680 for @xmath594 . by lemma 4.5.2
, @xmath567 is regular in @xmath1658.q.e.d . * _ lemma 4.5.4 . _ * if @xmath1681 , then @xmath1682 is disjoint from @xmath1320 . * _ proof . _
* suppose that @xmath1681 .
there are three cases : case 1 .
@xmath1683 . by definition 4.3.9 , @xmath1446 . by lemma 4.3.5(4 ) , the lemma holds .
@xmath1684 . by lemma 3.5.2 and definition 4.3.9 , @xmath1685 .
hence @xmath1582 . by lemma 4.1.5(8 ) and construction(*)(3 ) and ( 4 ) , the lemma holds .
case 3 . @xmath1686 . by lemma 4.3.5(4 ) , @xmath1687 .
note that @xmath1688 if @xmath1689 . since @xmath1681 , @xmath1690 if @xmath1691 . by lemma 4.3.10(2 ) and ( 3 ) and lemma 4.4.3 , @xmath1692 .
note that @xmath1693 . by lemma 4.3.5 , @xmath1408 .
hence the lemma holds.q.e.d.0.5 mm * _ lemma 4.5.5 . _
* suppose that @xmath1694and @xmath1048 .
then \(1 ) each component of @xmath1695 is a core of @xmath1320 for @xmath1696 .
\(2 ) each component of @xmath1697 lies in @xmath1294 for @xmath1698 .
\(3 ) @xmath1699 is regular in @xmath1700 . *
_ * since @xmath1701 . by lemma 4.3.7(3 ) , @xmath1702 where @xmath1703 . by definition 3.5.1 , @xmath1704 . by lemma 3.3.1 ,
@xmath1705 . by the argument in lemma 4.5.3 ,
@xmath1706 satisfying the following conditions : \(1 ) @xmath746 is an arc in @xmath1707 such that @xmath1085 is disjoint from @xmath1320 for @xmath1681 .
\(2 ) @xmath1708 is a core of @xmath1709 . by the construction , @xmath1710 .
note that @xmath1711 and @xmath1712 .
let @xmath1713 .
note that @xmath1714 for @xmath1047 . by definition 4.3.9 ,
@xmath1715 . by lemma 3.3.1 , @xmath1716 . by lemma 3.5.1 ,
hence @xmath1718 intersects @xmath1320 in a core of @xmath1320 for @xmath1719 , and @xmath1707 in an arc @xmath1720 for latexmath:[$f\in \bigl\{f_{i,0 } \
particular , by definition 4.3.9 , @xmath1718 intersects @xmath1466 in a core of @xmath1722 . by lemma 4.5.4
, @xmath1416 is disjoint from @xmath1493 for @xmath1681 .
now we claim that @xmath1720 is disjoint from @xmath1493 for @xmath1681 .
there are two cases : case 1 .
@xmath1723 .
suppose that @xmath1233 . by definition 4.3.9 ,
@xmath1300 . by lemma 4.1.5(6 ) , @xmath1724 is disjoint from @xmath1320 . since @xmath1725 , @xmath1726 is disjoint from @xmath1720 .
@xmath1447 . since @xmath1681 , @xmath1690 .
since @xmath1727 , by lemma 3.5.3 , @xmath582 is disjoint from @xmath1562 . by lemma 4.3.10 ,
note that @xmath1693 . by lemma 4.3.5 , @xmath1408 .
hence @xmath1320 is disjoint from @xmath1729 .
thus @xmath1720 is disjoint from @xmath1493 . by lemma 3.3.4(3 ) , @xmath1730 . by lemma s 4.5.1 and 4.5.2
, @xmath567 is regular in @xmath1731.q.e.d.0.5 mm in this section , we shall finish the proofs of propositions 4 - 6 for the case : @xmath493 and @xmath494 .
* _ lemma 4.6.1 . _ * @xmath1732 is disjoint from @xmath567 and @xmath1733 . * _ proof . _ * by definition 4.3.3 , @xmath1298 . by lemma 4.1.5(5 ) , @xmath1732 is disjoint from @xmath1260 . by lemma 3.5.8 , @xmath1734 is disjoint from @xmath1376 . by proposition 4(5 ) , @xmath1732 is disjoint from @xmath568 and @xmath582 . by definition 4.3.9 , @xmath1735 and @xmath1736 .
hence the lemma holds.q.e.d . * _ the proof of proposition 4 . _ * by lemmas 4.4.1 - 4.4.5 , proposition 4(1 ) holds . by lemma 4.5.3(3 ) and lemma 4.5.5(3 ) , @xmath567 is regular in @xmath1737 .
hence proposition 4(2 ) holds .
proposition 4(3 ) follows from lemma 4.5.1 . if @xmath1648 , then , by proposition 4(4 ) , @xmath1298 intersects @xmath590 in at most one point . by lemma 4.5.3 and lemma 4.3.7
, @xmath1738 intersects @xmath1419 in at most one point .
suppose that @xmath858 . by lemma 3.3.4(4 ) , @xmath805 .
suppose fist that @xmath1739 , then , by lemma 3.5.8 , @xmath1740 @xmath1741 .
note that @xmath1742 . by lemma 3.3.1 , @xmath1743 .
suppose now that @xmath1744 .
then , by lemma 3.3.1 , @xmath1745 . by lemma 4.5.5 , @xmath1746 . by lemma 4.3.7 , there is at most one component @xmath1747 of @xmath1748 such that @xmath1749 for each @xmath548 . by lemma 4.1.5(6 ) , construction(*)(3 ) , ( 4 ) and lemma 3.5.8 , @xmath1298 intersects @xmath1358 in at most one point . furthermore , @xmath1298 intersects @xmath1358 in one point if and only if @xmath1750 in one point , where @xmath1133 is as in lemma 3.5.5 .
similarly , @xmath1298 intersects @xmath1366 in one point if and only if @xmath1751 in one point . by lemma 4.1.5(6 ) , @xmath1419 intersects @xmath1720 in one point if and only if @xmath1292 intersects @xmath1752 in one point . by proposition 4(4 ) and lemma 3.3.1 , @xmath1298 intersects @xmath746 in at most one point .
hence proposition 4(4 ) holds .
proposition 4(5 ) follows from lemma 4.6.1.q.e.d.1.5 mm * _ the proof of proposition 5 .
_ * suppose that @xmath1310 . by the argument in the proof of lemma 4.5.3 and lemma 3.1.4 , @xmath1753 .
suppose that @xmath1754 . by the argument in the proofs of lemma 4.5.5 and proposition 4 for @xmath493 , @xmath1755 .
now there are two cases : case 1 .
@xmath1189 . by lemma 3.5.8(1 ) , @xmath1756 . by lemma 3.3.1 ,
@xmath1757 . by proposition 4(4 ) , @xmath1758 for @xmath769 .
hence , @xmath1759 .
@xmath1744 . by lemma 3.3.4 , @xmath1760 for @xmath1761 . by lemma 3.5.8(2 ) , @xmath1762 . similarly , @xmath1763 . hence @xmath1764 .
+ q.e.d.0.5 mm * _ lemma 4.6.3 . _ * @xmath1765 is obtained by doing band sums with some copies of @xmath1242 to @xmath568 . * _ proof . _ * since each component of @xmath1432 is a core of @xmath1099 , say @xmath1766 . by definition 4.3.6 , @xmath1767 intersects @xmath1768 in @xmath1769 , @xmath1770 in @xmath1771 , @xmath1554 in @xmath1772 .
now by lemma 3.5.5 , @xmath1773 bounds a disk in @xmath1316 .
hence , the lemma holds.q.e.d . * _ the proof of proposition 6 . _
* suppose that @xmath548 and @xmath288 .
then , by proposition 6 , there is a properly embedded disk @xmath1315 in @xmath27 such that @xmath1774 .
now we denote by @xmath553 the disk @xmath1315 . by lemma 4.1.5(7 ) , @xmath1260 is disjoint from @xmath1775 for each @xmath548 with @xmath288 . since @xmath1776 , @xmath1777 . in this case , if @xmath1778 , then @xmath1314 , if not , then @xmath1779 .
now suppose that @xmath546 and @xmath293 . by lemma 4.1.5(7 ) , @xmath1260 is disjoint from @xmath1282 . by proposition 6 , @xmath1780 .
now let @xmath1781 . then @xmath1782 . since @xmath1240 , by lemma 2.2.4 and lemma 4.1.3 , @xmath1783 .
hence @xmath1311 . by lemma 4.6.3
, @xmath1784 is obtained by doing band sums with copies of @xmath1242 to @xmath1282 .
hence @xmath1784 bounds a disk in @xmath47 , denoted by @xmath571 . now by proposition 6 , @xmath1785 for @xmath548 with @xmath288 and @xmath649 with @xmath293 .
now there are two case : \(1 ) @xmath1786 . now @xmath1779 .
hence @xmath1787 .
\(2 ) @xmath1778 .
now @xmath1314 . by proposition 6(2 )
, @xmath538 is disjoint from @xmath1788 .
obviously , @xmath538 is disjoint from @xmath567 and @xmath1789 .
note that @xmath1790 .
hence @xmath1791 .
q.e.d 0.5 mm
in this section , we shall propositions 4 - 6 for the case : @xmath493 and @xmath495 . in the following argument
, we assume @xmath495 . recalling @xmath574 , and @xmath1792 .
* _ lemma 5.1.1 . _ * suppose that @xmath1793
. then \(1 ) @xmath1794 such that @xmath1795 is disjoint from @xmath564 .
\(2 ) @xmath541 intersects @xmath1185 in one point @xmath1796 .
\(3 ) if @xmath1797 , then @xmath568 is disjoint from @xmath541 ; if @xmath1683 , then @xmath582 is disjoint from @xmath541 .
\(4 ) @xmath575 . * _ proof . _ * ( 1 ) since @xmath495 , by proposition 6 , @xmath1794 and @xmath1798 .
if @xmath1799 , then , by lemma 2.2.4 , @xmath1800 . by definition 2.3.1 ,
@xmath1801 . thus ( 1 ) holds .
\(2 ) by ( 1 ) , @xmath1802 . by the proof of lemma 4.1.2 , @xmath538 intersects @xmath1185 in one point lying in @xmath1803 .
\(3 ) since @xmath1804 . hence @xmath1805 and @xmath1806 . by the minimality of @xmath1226 in @xmath1807 ,
( 3 ) follows from the proof of lemma 4.1.2 .
\(4 ) suppose that @xmath539 . then , by proposition 6 , @xmath1808 . by ( 1 ) , @xmath1809 . since @xmath1810 , by lemma 2.2.4 , @xmath1811
. by ( 2 ) and ( 3 ) , @xmath534 intersects @xmath557 in one point
. hence @xmath215 is stabilized , a contradiction .
q.e.d.0.5 mm * _ lemma 5.1.2 . _
* suppose that @xmath1232 .
if @xmath1812 for some @xmath1813 with @xmath1814 , then there is a neighborhood of @xmath1185 in @xmath564 , say @xmath1815 $ ] , which satisfies the following conditions : \(1 ) @xmath1816 $ ] .
\(2 ) if @xmath1817 , then each component of @xmath1815\cap e^{l}_{\gamma}\times i$ ] is @xmath1818\subset e^{l}_{\gamma}\times ( 0,1)$ ] where @xmath1269 is a core of @xmath690 .
\(3 ) if @xmath1819 , then @xmath1820 $ ] is disjoint @xmath582 .
\(4 ) if @xmath1821 , then @xmath582 is disjoint from @xmath1822\cup [ -2.5,-1])$ ] .
\(5 ) if @xmath1823 , then @xmath568 is disjoint from @xmath1815 $ ] , if @xmath1824 , then @xmath568 is disjoint from @xmath1822\cup [ -2.5,-1])$ ] .
\(6 ) @xmath1825\cup [ 1.5,2.5]$ ] is disjoint from @xmath582 for each @xmath132 and @xmath1047 , @xmath1826 $ ] is disjoint from @xmath582 for @xmath1233 .
\(7 ) @xmath1815 $ ] is disjoint from @xmath1099 for @xmath1048 , and @xmath1827\cap d^{*}_{0,\alpha}=a^{0}_{\alpha}\times [ -2.5,2.5]\cap d^{*}_{0,\alpha}$ ] .
\(8 ) for each @xmath1275 , @xmath1276 is disjoint from @xmath1815 $ ] .
* _ proof .
_ * ( 1 ) is trivial .
\(2 ) since @xmath1233 , by proposition 4(2 ) and definition 2.1.5 , each component of @xmath1828 is a core of @xmath582 , say @xmath630 , which is a core of @xmath724 .
note that @xmath1829 . by proposition 4(1 ) and definition 2.1.4 , each component of @xmath1830 is @xmath1831 .
\(3 ) by lemma 3.2.3 and lemma 3.2.6 , @xmath1832 is disjoint from @xmath582 for @xmath1833 .
\(4 ) if @xmath1834 , then , by proposition 4(1 ) , each component of @xmath1835 .
\(5 ) if @xmath1823 , then , by proposition 4(3 ) , @xmath568 is disjoint from @xmath1832 . by proposition 4(2 ) and definition 2.1.5 , if @xmath1824 , then each component of @xmath1836 is @xmath1837 .
\(6 ) if @xmath1263 , then either @xmath1838 or @xmath1839 is disjoint from @xmath582 by lemma 3.2.6 . if @xmath1266 , then , by lemma 3.3.1 and definition 2.1.4 , @xmath1840 is disjoint from @xmath582 . if @xmath1841 , then each component of @xmath1842 .
\(7 ) by lemma 3.2.6 and lemma 3.5.7 , @xmath1832 is disjoint from @xmath1188 for @xmath1048 .
\(8 ) by proposition 4(5 ) , @xmath1276 is disjoint from @xmath1843 .
now @xmath1844 $ ] can be obtained by making @xmath1839 wide slightly .
hence lemma 5.1.2 holds .
0.5 mm * _ lemma 5.1.3 . _
* suppose @xmath1845 .
if @xmath1253 for each @xmath132 with @xmath458 , there is a neighborhood of @xmath1185 in @xmath564 , say @xmath1815 $ ] , which satisfies the following conditions : \1 ) if @xmath1273 , then @xmath1815 $ ] is disjoint from @xmath568 .
\2 ) if @xmath1846 , then @xmath1815 $ ] is disjoint from @xmath582 .
\3 ) if @xmath1847 , then each component of @xmath1815\cap e^{l}_{\gamma}\times i$ ] is @xmath1818\subset e^{l}_{\gamma}\times ( 0,1)$ ] where @xmath1269 is a core of @xmath690 .
\4 ) @xmath1848 $ ] is disjoint from @xmath688 for each @xmath132 .
\(5 ) @xmath1848 $ ] is disjoint from @xmath1290 for each @xmath1275 .
\(6 ) @xmath1815 $ ] is disjoint from @xmath1099 for @xmath1059 , and @xmath1827\cap d^{*}_{0,\alpha}=a^{0}_{\alpha}\times [ -2.5,2.5]\cap d^{*}_{0,\alpha}$ ] . * _ proof . _ * by lemma 5.1.1(4 ) , @xmath575 . hence if @xmath1849 then @xmath458 . by assumption , @xmath1253 for each @xmath1301 . now ( 1 ) follows from proposition 4 .
( 2 ) follows from lemma 3.2.3 .
( 3 ) follows proposition 4(2 ) and definition 2.1.5 .
( 4 ) follows from ( 2 ) and lemma 3.3.1 .
( 5 ) follows from proposition 4(5 ) .
( 6 ) follows from lemma 3.5.7.q.e.d.0.5 mm * _ lemma 5.1.4 .
_ * suppose that @xmath1845 .
then there is a neighborhood of @xmath541 in @xmath448 , say @xmath1850 , satisfying the following conditions : \(1 ) @xmath1851 .
\(2 ) if @xmath1852 , then @xmath568 and @xmath582 are disjoint from @xmath1291 .
\(3 ) @xmath1853 are disjoint from @xmath1291 for each @xmath649 and @xmath132 .
\(4 ) @xmath1185 intersects @xmath1291 in an arc @xmath1323 , and @xmath1815 $ ] intersects @xmath1291 in @xmath1854 $ ] .
see figure 19 .
\(5 ) for @xmath548 , @xmath1292 is disjoint from @xmath1291 , if @xmath288 , then @xmath1315 is disjoint from @xmath1850 .
\(6 ) for each @xmath649 with @xmath293 , @xmath1788 is disjoint from @xmath1850 . * _ proof . _ * ( 1 ) follows from lemma 5.1.1(1 ) .
( 2 ) follows from lemma 5.1.1(3 ) and ( 4 ) .
( 3 ) follows from proposition 4(5 ) .
( 4 ) follows from lemma 5.1.1(2 ) .
( 5 ) and ( 6 ) follow from proposition 6.q.e.d .
figure 19 * _ lemma 5.1.5 . _ * if @xmath713 , then propositions 4 - 6 hold for the case : @xmath493 and @xmath495 . * _ proof . _
* by corollary 3.2.5 , @xmath1855 are disjoint from @xmath1291 where @xmath1291 is a regular neighborhood of @xmath538 in @xmath750 .
let @xmath1856 .
we denote by @xmath567 the arc @xmath568 for @xmath580 , @xmath1320 the disk @xmath582 for @xmath132 , @xmath1294 the disk @xmath602 for @xmath1295 . in particular , we denote by @xmath1296 the two components of @xmath1297 . by proposition 4 , @xmath1857 for @xmath548 .
now we denote by @xmath1304 the arc @xmath579 .
let @xmath1227 .
for @xmath548 with @xmath288 and @xmath580 , we denote by @xmath553 the disk @xmath554 , @xmath1858 the disk @xmath572 . by the argument in section 4.2 , the lemma holds.q.e.d . by lemma 5.1.5 , in the following argument
, we may assume that @xmath1860 . hence @xmath1232 . in sections 5.2 and 5.3
, we shall construct @xmath1861 from @xmath1862 , where @xmath1863 are just @xmath1864 in propositions 4 - 6 , but @xmath1865 are not @xmath1866 in propositions 4 - 6 . in this section
, we first assume that @xmath1867 for each @xmath132 with @xmath458 .
let @xmath1850 be a neighborhood of @xmath1868 in @xmath514 satisfying lemma 5.1.4 , and @xmath1815 $ ] be a neighborhood of @xmath1185 in @xmath564 satisfying lemma 5.1.3.0.5 mm * _ definition 5.2.1 .
_ * we denote by @xmath1869 the two components of @xmath1297 . figure 20 figure 21 * _ definition 5.2.2 . _ * ( 1 ) let @xmath1870 .
see figure 20 .
\(2 ) let @xmath1316 be the surface obtained by identifying @xmath1871 $ ] to @xmath1872 $ ] and @xmath1873 $ ] to @xmath1874 $ ] in @xmath514@xmath1875 .
see figure 21 .
let @xmath1876 . then @xmath1877 . now by lemma 5.1.4(4 ) , @xmath1878\cup [ 1.5,2])\cup b\times i$ ] is a disk .
we denote by @xmath1879 the two components of @xmath1880 where @xmath95 is as in lemma 5.1.4 .
now there is an arc @xmath1881 in @xmath1882\cup [ 1.8,2])\cup b\times i$ ] connecting @xmath1883 to @xmath1884 . furthermore , @xmath1881 intersects @xmath101 in @xmath1885 .
similarly , there is an arc @xmath1886 in @xmath1887\cup [ 1.8,2])\cup b\times i$ ] connecting @xmath1888 to @xmath1889 .
furthermore , @xmath1886 intersects @xmath101 in @xmath1890 . now @xmath1891 separates @xmath1878\cup [ 1.5,2])\cup b\times i$ ] into three disks @xmath1892 as in figure 21 . without loss of generality , we may assume that @xmath1893 and @xmath1894 as in figure 21 .
note that @xmath1895\cup [ 1.8,2])\cup b\times i$ ] . now there is a homeomorphism @xmath1896 form @xmath1897\subset d_{j_{0}}^{l}\times [ -2,2]$ ] to @xmath1898 such that @xmath1896 is an identifying map on @xmath1899 for @xmath1900 .
let @xmath1901 .
figure 22 by lemma 3.3.1 and remark 3.3.2 , @xmath1902 . * _ definition 5.2.3 . _ * ( 1 ) for each @xmath132 , let @xmath1903 , @xmath1904)\cup h(e^{l}_{\gamma}\times i\cap ( h_{1}\cup h_{2})\times [ -2,2])$ ] .
\(2 ) for @xmath546 , let @xmath1905 .
\(3 ) for @xmath548 , let @xmath1906)\cup h(c_{i}^{l}\cap ( h_{1}\cup h_{2})\times [ -2,2])$ ] .
\(4 ) for @xmath1295 , let @xmath1907)\cup h(e^{l}_{f}\cap ( h_{1}\cup h_{2})\times [ -2,2])$ ] , let @xmath1908)\cup h(e_{0}\cap ( h_{1}\cup h_{2})\times [ -2,2])$ ] , let @xmath1909)\cup h(e_{l+1}\cap ( h_{1}\cup h_{2})\times [ -2,2])$ ] .
\(5 ) let @xmath1910 , and @xmath1911 with @xmath1912=d_{j_{0}}^{l}\times [ -1.5,-1.6]$ ] .
see figure 22 . in this section
, we shall construct @xmath1913 from @xmath1914 for the case : @xmath1812 for some @xmath1813 with @xmath1814 .
let @xmath1850 be a neighborhood of @xmath1868 in @xmath514 satisfying lemma 5.1.4 , and @xmath1844 $ ] be a neighborhood of @xmath1185 in @xmath564 satisfying lemma 5.1.2 .
recalling the disks @xmath1869 , @xmath1915 and the arcs @xmath1916 in section 5.2 .
figure 23 * _ definition 5.3.1 . _ * ( 1 ) let @xmath1917 .
\(2 ) let @xmath1316 be the surface obtained by identifying @xmath1871 $ ] to @xmath1872 $ ] and @xmath1873 $ ] to @xmath1874 $ ] in @xmath514@xmath1875 .
see figure 23 . * _ definition 5.3.2 .
_ * ( 1 ) for each @xmath132 with @xmath1918 , let @xmath1903 , @xmath1919)\cup h(e^{l}_{\gamma}\times i\cap ( h_{1}\cup h_{2})\times [ -2,2])$ ] .
\(2 ) for each @xmath132 with @xmath1920 , let @xmath1903 , @xmath1921 .
\(3 ) let @xmath1910 , and @xmath1911 with @xmath1912=d_{j_{0}}^{l}\times [ -1.5,-1.6]$ ] .
see figure 24 . * _ definition 5.3.3 .
_ * ( 1 ) for @xmath546 , let @xmath1905 .
\(2 ) for @xmath548 , let @xmath1906)\cup h(c_{i}^{l}\cap ( h_{1}\cup h_{2})\times [ -2,2])$ ] .
\(3 ) for @xmath1295 , let @xmath1907)\cup h(e^{l}_{f}\cap ( h_{1}\cup h_{2})\times [ -2,2])$ ] , let @xmath1908)\cup h(e_{0}\cap ( h_{1}\cup h_{2})\times [ -2,2])$ ] , let @xmath1909)\cup h(e_{l+1}\cap ( h_{1}\cup h_{2})\times [ -2,2])$ ] . see figures 23 and 24 .
figure 24 in this section , we shall introduce simple properties of @xmath1922 . * _ lemma 5.4.1 . _
* @xmath1923\cup [ 1.5,2])\cup b\times i\subset e^{l+1}_{0}\cup e^{l+1}_{l+1 }
\cup_{i=\delta(0)}^{\theta(0 ) } e^{l+1}_{f_{i,0}}\cup_{i=\delta(0)}^{\theta(0 ) } e^{'}_{\gamma_{i,0}}\times i\cup e^{'}_{m^{l+1}}\times i$ ] . *
_ * by lemma 3.3.1 , @xmath1902 , where @xmath1752 is an properly embedded arc in @xmath1924 , @xmath1925 is a core of @xmath1926 for @xmath769 , and @xmath1203 is an arc in @xmath750 . by lemma 5.1.4 ,
we denote by @xmath1927 the two components of @xmath1928 .
we may assume that @xmath1929 and @xmath1930 .
then @xmath1931=(\cup_{i=\delta(0)}^{-1 } d_{i,0}\cup h^{*}\cup_{i=\delta(0)}^{-1 } e_{i,0})\times [ -2,2]$ ] and @xmath1932=(h^{**}\cup_{i=1}^{\theta(0 ) } d_{i,0}\cup_{i=1}^{\theta(0 ) } e_{i,0})\times [ -2,2]$ ] .
note that @xmath1923\cup [ 1.5,2])\cup b\times i = b_{1}\cup b_{2}\cup b_{3}$ ] .
now by definitions 5.2.3 , 5.3.2 and 5.3.3 , the lemma holds.q.e.d . *
_ lemma 5.4.2 .
_ * @xmath1933 is a set of pairwise disjoint arcs in @xmath1316 .
furthermore , @xmath1934 is properly embedded in @xmath1935 .
* _ proof . _
* now there are two cases : case 1 .
@xmath1812 for some @xmath1936 with @xmath1814 . by lemma 5.1.2
, @xmath1905 is disjoint from @xmath1937 $ ] . by proposition 4(2 ) , @xmath568 is properly embedded in @xmath874 .
if @xmath1428 , then , by lemma 5.1.4 , @xmath568 is disjoint from @xmath1850 .
hence @xmath568 is also properly embedded in @xmath1938 . if @xmath1630 , then @xmath1939 for each @xmath1940 . by definitions 5.3.2 , 5.3.3 and lemma 5.4.1 , @xmath1941 . by lemma 5.1.2 , if @xmath1942 , then @xmath1824 .
hence the lemma holds .
@xmath1943 or @xmath1944 . by lemma 5.1.3 and the argument in case 1 , the lemma holds .
q.e.d.0.5 mm * _ lemma 5.4.3 .
_ * ( 1 ) @xmath1294 is a disk in @xmath1316 .
\(2 ) @xmath1945 is a set of pairwise disjoint arcs properly embedded in @xmath1316 .
furthermore , @xmath1419 lies in one of @xmath1294 for some @xmath194 .
\(3 ) @xmath1946 for each @xmath1321 , and @xmath1947 for each @xmath1948 .
\(4 ) @xmath1419 is obtained by doing band sums with copies of @xmath541 to @xmath1292 .
* _ proof . _ * ( 1 ) for each @xmath1295 , by lemma 3.1.6 , each component of @xmath1949 is a properly embedded arc @xmath630 in @xmath602 , and each component of @xmath1844 \cap e^{l}_{f}$ ] is @xmath1950 $ ] in @xmath602 .
since @xmath1951 , @xmath1294 is a disk in @xmath1316 . since each component of @xmath1952 is a properly embedded arc @xmath630 in @xmath750 . by lemma 5.1.4(4 ) , each component of @xmath1844 \cap e_{f}$ ] is @xmath1953 $ ] in @xmath602 , where @xmath1954 is a properly embedded arc in @xmath1955 for @xmath1956 .
hence @xmath1296 are two disks in @xmath1316 .
\(2 ) let @xmath548 . by proposition 4(4 ) and lemma 5.1.4(5 ) , either @xmath1292 lies in @xmath602 for some @xmath1295 or @xmath1292 lies in @xmath1955 for @xmath1956 .
we may assume that @xmath1957 .
by lemma 3.1.6 , each component of @xmath1958 is a point @xmath642 .
furthermore , @xmath1959 where @xmath630 is one component of @xmath1949 . by lemma 5.1.2(8 ) and lemma 5.1.3(5 ) , each component of @xmath1844\cap c_{i}^{l}=p\times [ -2,2]\subset intc_{i}^{l}$ ] . by the construction
, @xmath1419 is a properly embedded arc in @xmath1316 which lies in @xmath1294 .
\(3 ) by lemma 5.1.2(8 ) and lemma 5.1.3(5 ) , @xmath1201 is disjoint from @xmath1815 $ ] . by proposition 4(5 ) and the construction , ( 3 ) holds .
\(4 ) since @xmath1960-d_{j_{0}}^{l}\times ( -1.5,1.5)$ ] is a disk . by ( 2 ) , each component of @xmath1961 $ ] is an arc .
hence the homeomorphism @xmath1437 means band sums.q.e.d.0.5 mm * _ lemma 5.4.4 . _
* for each @xmath1321 , @xmath1962 is a disk in @xmath1316 such that \(1 ) @xmath1963 for @xmath132 , \(2 ) @xmath1964 for @xmath1965 .
\(3 ) @xmath1966 and @xmath1967 .
\(4 ) @xmath1968 for @xmath1969 . * _ proof .
_ * there are two cases : case 1 .
@xmath1814 .
suppose that @xmath1970 .
if @xmath1971 , then @xmath1447 . by definition 5.3.2 , @xmath1921 . by lemma 5.1.2(4 ) , @xmath582 is disjoint from @xmath1972\cup [ 1.5,2.5]$ ] . hence @xmath1973 .
if @xmath1974 , then , by lemma 5.1.2(3 ) , @xmath582 is disjoint from @xmath1975 $ ] . by definition 5.3.2 , @xmath1976 .
suppose that @xmath1977 .
by lemma 5.1.2 , each component of @xmath1844\cap e^{l}_{\gamma}\times i$ ] is @xmath1950\subset e^{l}_{\gamma}\times ( 0,1)$ ] . by lemma 5.1.4 , @xmath582 is disjoint from @xmath1978 . by definition 5.3.2 ,
@xmath1979 is a disk in @xmath1316 .
it is easy to see that @xmath1980 is a disk in @xmath1316 . by proposition 4(1 ) and definition 2.1.4 , @xmath1981 , @xmath1982 for @xmath1983 .
hence ( 1 ) follows from the construction , and @xmath1984 for @xmath1985 . by proposition 4(1 ) and lemma 5.1.2 , @xmath1986 $ ] is disjoint from @xmath582 for @xmath132 even if @xmath1838 . by lemma 5.3.2 , @xmath1987)$ ] .
hence @xmath1988 for @xmath132 . thus ( 2 ) holds . by lemma 3.3.1 ,
hence @xmath1989 and @xmath1990 . by definitions 5.3.1 , 5.3.2 and 5.3.3 , ( 3 )
see figure 24 .
@xmath1943 or @xmath1944 . by lemma 5.1.3 and the argument in case 1 , the lemma holds for this case.q.e.d.0.5 mm * _ lemma 5.5.1 .
_ * @xmath1992 . * _ proof .
_ * by lemma 5.4.1 , @xmath1993\cup[1.5,2]))\cup b\times i\subset \cup_{f } e^{l+1}_{f}\cup_{i=\delta(0)}^{\theta(0 ) } e^{'}_{\gamma_{i,0}}\times i\cup e^{'}_{m^{l+1}}\times i$ ] . by proposition 4(1 ) , the lemma holds .
q.e.d.0.5 mm * _ lemma 5.5.2 .
_ * if @xmath402 , then @xmath1994 is disjoint from @xmath1995 . * _ proof . _ * by proposition 4(1 ) , if @xmath165 with @xmath402 , then @xmath1503 is disjoint from @xmath1996 .
there are two cases : case 1 .
@xmath1814 .
there are three sub - cases : case 1.1 .
@xmath1997 . by lemma 5.4.3 , @xmath1998 . by proposition 4(1 ) , @xmath1503 is disjoint from @xmath1999 . by lemma 5.1.4(1 ) , proposition 4(5 ) , @xmath1503 is disjoint from @xmath1850 . by lemma 5.1.2(3 ) and ( 4 ) , @xmath2000\cup[1.5,2]$ ] is disjoint from @xmath1505 . by definitions 5.3.2 and 5.3.3 , @xmath2001\cup[1.5,2]$ ] . hence @xmath1994 is disjoint from @xmath2002 .
@xmath1625 . now if @xmath1498 , then , by lemma 5.1.4(2 ) , @xmath582 is disjoint from @xmath1850 . by definitions 5.3.2 and 5.3.3 , @xmath2003)\cup b_{1}\cup b_{2}$ ] ,
@xmath1980 where @xmath1898 is as in section 5.2 .
hence the lemma holds .
now if @xmath1516 , then @xmath2005 . by lemma 5.1.4 ,
@xmath582 is disjoint from @xmath1978 . by proposition 4(1 ) , @xmath1503 is disjoint from @xmath2006 . by lemma 5.1.2 ,
each component of @xmath1844\cap e^{l}_{\lambda}\times i$ ] is @xmath1950\subset e^{l}_{\lambda}\times ( 0,1)$ ] where @xmath630 is a core of @xmath2007 . hence @xmath2008 $ ] is disjoint from @xmath2006 . by definitions 5.3.2 and 5.3.3
, the lemma holds .
@xmath1943 or @xmath1944 . by lemma 5.1.3 , definition 5.2.3 and the argument in case 1 , the lemma holds.q.e.d.0.5 mm * _ lemma 5.5.3 .
_ * if @xmath404 , then @xmath2009 .
* _ proof .
_ * suppose that @xmath1492 and @xmath404 . by lemma 5.1.4(4 ) , @xmath2010 . by proposition 4(1 ) , @xmath2011 .
now there are two cases : case 1 .
there are four sub - cases : case 1.1 .
@xmath2012 . in this case ,
@xmath2013 . by definition 2.3.1 ,
@xmath2014 . since @xmath2015 , by lemma 2.2.5 , @xmath2016 . by lemma 5.4.4 , @xmath1968 . by lemmas 5.1.2 ,
@xmath2017\cup [ 1,2]$ ] is disjoint from @xmath1505 . by proposition 4(1 ) , @xmath2018)=\cup_{r\in i(w_{\lambda},l ) } ( d_{r}^{l}\times i)_{\lambda}$ ] . by proposition 4(5 ) and lemma 5.1.4(1 ) , @xmath1505 is disjoint from @xmath2019 . by lemma 5.4.1 , the lemma holds .
@xmath2020 . in this case , @xmath2013 . by definition 2.3.1 , @xmath2014 .
now by lemma 5.1.2(2 ) , ( 3 ) and definitions 5.3.2 , 5.3.3 , @xmath1839 is disjoint from @xmath2021 . by the argument in case 1.1 , the lemma holds .
@xmath2022 . in this case ,
@xmath2023 . by lemma 5.1.4 and lemma 5.4.4
, @xmath2024 is disjoint from @xmath1978 . by lemma 5.1.2(3 ) , @xmath2025 is disjoint from @xmath1844 $ ] . by definition 5.3.2 and 5.3.3 , @xmath2026 .
hence the lemma holds .
@xmath2027 . in this case , @xmath2023 .
now we denote by @xmath2028 the surface @xmath2029 .
now @xmath2030)\cap ( p_{\lambda}-p_{\lambda}\cap d_{j_{0}}^{l}\times [ -2,2]))\cup ( h(e^{l}_{\lambda}\times i\cap d_{j_{0}}^{l}\times [ -2,2])\cap h(p_{\lambda}\cap d_{j_{0}}^{l}\times [ -2,2]))$ ] . hence the lemma holds .
@xmath1943 or @xmath1944 . by lemma 5.1.3 and definition 5.2.3 , and
the argument in case 1 , the lemma holds.q.e.d.0.5 mm * _ lemma 5.5.4 .
_ * if @xmath2031 , then each component of @xmath2032 is @xmath2033 where @xmath2034 is a core of @xmath2035 . * _ proof . _ * by proposition 4(1 ) and definition 2.1.4 , if @xmath2036 , then each component of @xmath2037 is @xmath2038 where @xmath2039 is a core of @xmath724 . without loss of generality , we may assume that @xmath1814 .
now there are six cases : case 1 .
@xmath2040 and @xmath2041 . by lemma 5.4.3 ,
@xmath1998 and @xmath1968 .
hence the lemma holds .
@xmath2042 .
since @xmath575 , by lemma 5.4.4(2 ) and lemma 5.5.2 , @xmath2043 is disjoint from @xmath1979 .
@xmath2044 . by lemma 3.2.4 , and lemma 5.1.4(1 ) , each component of @xmath2045 is @xmath2046 where @xmath2047 .
furthermore , by lemma 5.1.2(3 ) and ( 4 ) , @xmath2046 lies either in @xmath2048 $ ] or in @xmath1839 . if @xmath2046 lies in @xmath1839 .
then @xmath2049 is disjoint from @xmath2043 .
if @xmath2046 lies in @xmath2050 $ ] , then , by definitions 5.3.3 , @xmath2046 intersects @xmath2043 in @xmath2051 , where @xmath2052 is a core of @xmath2053 . hence the lemma holds .
@xmath2054 . by lemma 5.4.4(4 ) , @xmath2055 . by lemma 5.1.2(3 ) , @xmath1505 is disjoint from @xmath1844 $ ] . by lemma 5.1.4 , @xmath2056 . by lemma 3.2.4 , each component of @xmath2057 is @xmath2046 where @xmath2039 . by lemma 5.1.4 , @xmath582 is disjoint from @xmath1850 . by definition 5.3.2 , @xmath2058 , where @xmath2059 and @xmath2060)$ ] . hence the lemma holds .
. now by lemma 5.1.2 and lemma 5.4.4 , @xmath2024 is disjoint from @xmath2062 $ ] even if @xmath2020 . by definition 5.3.2 , @xmath2063 , where @xmath2064)\cap e^{l}_{\gamma}\times i$ ] and @xmath2065 . by lemma 3.3.1 ,
since @xmath1233 , by lemma 5.1.2(6 ) , @xmath2067 $ ] and @xmath2068 $ ] are disjoint from @xmath582 .
hence each component of @xmath2037 is either in @xmath2069 $ ] or in @xmath2070 $ ] . by the argument in case 4 , the lemma holds .
@xmath2071 . by definition 5.3.2 , @xmath2072 where @xmath2073 and @xmath2074 .
since @xmath1437 is a homeomorphism , the lemma holds .
q.e.d.1.5 mm * _ lemma 5.5.5 . _
* @xmath2075 is a abstract tree , and @xmath1316 is generated by @xmath2076 . * _ proof . _ * by lemma 5.4.4 and lemmas 5.5.1 - 5.5.4 , we only need to prove that @xmath2077 is a abstract tree . by lemma 5.4.4 , @xmath2078 and @xmath2079 . by the argument in lemma 4.4.5 , the lemma holds.q.e.d .
* _ lemma 5.6.1 . _ * suppose that @xmath2080 , @xmath1321 and @xmath2081 .
if @xmath411 , then @xmath1934 is disjoint from @xmath1962 . * _ proof . _
* suppose that @xmath411 and @xmath2082 . by proposition 4(3 ) , @xmath568 is disjoint from @xmath582 for @xmath132 . without loss of generality , we may assume that @xmath1814 .
now there are three cases : case 1 .
@xmath2083 , @xmath2084 .
now @xmath2085 . by definition 5.3.2 and lemma 5.4.4 , @xmath1968 and @xmath1905
. hence @xmath1934 is disjoint from @xmath582 .
@xmath2086 . since @xmath2087 , @xmath2088 . by lemma 5.1.2(5 ) , @xmath1905 is disjoint from @xmath1975 $ ] . by lemma 5.1.4(2 ) , @xmath1934 is disjoint from @xmath1868 .
hence @xmath1934 is disjoin from @xmath2089\cup \partial v_{l+1}^{l}\times i$ ] .
@xmath2090 . now by lemma 5.1.2(5 ) and lemma 5.1.4(2 ) , @xmath568 is disjoint from @xmath1850 and @xmath1975 $ ] . by definition 5.3.2 ,
@xmath2091\cup \partial v_{l+1}^{l}\times i$ ] .
hence @xmath1905 is disjoint from @xmath1962 .
q.e.d.0.5 mm * _ lemma 5.6.2 . _
* suppose that @xmath2080 , @xmath1321 . if @xmath717 , then each component of @xmath2092 is a core @xmath2093 of @xmath2094 .
* _ proof .
_ * suppose that @xmath717 .
by proposition 4(2 ) , each component of @xmath2095 is a core @xmath2096 of @xmath724 for @xmath132 . without loss of generality
, we may assume that @xmath1814 .
now there are four cases : case 1 .
@xmath1970 . by definition 5.3.2 and lemma 5.4.4 , @xmath1905 and @xmath1968 .
hence the lemma holds .
@xmath2097 . by definition 5.3.2 , @xmath1980 . by lemma 5.1.4 , @xmath2098 . by lemma 5.1.2 and lemma 3.1.6
, each component of @xmath2099 is an arc @xmath2096 which lies either in @xmath2070 $ ] or in @xmath2100 $ ] .
if @xmath630 lies in @xmath2070 $ ] , then @xmath630 is disjoint from @xmath2043 .
if @xmath630 lies in @xmath2101 $ ] , then @xmath630 intersects @xmath2043 in a core of @xmath2102 .
hence the lemma holds .
@xmath2103 .
now @xmath2087 .
by lemma 5.1.4(2 ) , @xmath1905 is disjoint from @xmath1850 . by lemma 5.1.2(5 ) , @xmath1934 is disjoint from @xmath1844 $ ] . by definition 5.3.2 , @xmath2091\cup\partial v_{l+1}^{l}\times i$ ] .
hence @xmath2104 , and the lemma holds .
@xmath2105 .
now by lemma 5.1.2(5 ) , @xmath1905 is disjoint from @xmath2017\cup [ 1,2]$ ] . by lemma 5.1.2(6 ) , @xmath2106 $ ] and @xmath2107 $ ] are disjoint from @xmath582 .
hence each component of @xmath2095 is either in @xmath2070 $ ] or disjoint from @xmath1975 $ ] . by lemma 3.1.6 , each component of @xmath2108 is @xmath2096 with @xmath2109 and @xmath2110 . by lemma 5.1.4
, @xmath582 is disjoint from @xmath1978 . now @xmath2111 , where @xmath2112)\cap e^{l}_{\gamma}\times i$ ] and @xmath2113)$ ] .
hence the lemma holds .
q.e.d.0.5 mm * _ lemma 5.6.3 .
_ * suppose that @xmath1648 .
then \(1 ) @xmath1934 is regular in @xmath1938 .
\(2 ) for each @xmath548 , @xmath2114 intersects @xmath1419 in at most one point . furthermore , @xmath2114 intersects @xmath1419 in one point if and only if @xmath2115 intersects @xmath1292 in one point . * _ proof . _ * without loss of generality , we may assume that @xmath1814 . by lemma 3.1.2 , in @xmath564 , @xmath2116 , where @xmath590 is a properly embedded arc in @xmath591 which is disjoint from @xmath592 and @xmath586 is a core of @xmath587 for some @xmath585 .
furthermore , @xmath2117 and @xmath595 for @xmath594 . by lemma 5.1.2
, @xmath568 is disjoint from @xmath2118\cup [ 1,2]$ ] .
hence @xmath586 is also a core of @xmath2119 . since @xmath1310 , @xmath590 is disjoint from @xmath1291 and @xmath1850 even if @xmath1653 .
furthermore , if @xmath1653 , then @xmath735 lies in in one of @xmath2120 and @xmath2121 . see definition 5.2.1 . by definition 5.3.2 ,
@xmath2122\cup[1,2]\cup \partial v_{l+1}^{l}\times i$ ] . hence if @xmath612 , then @xmath590 is disjoint from @xmath2123 . now
each component of @xmath2124 is @xmath586 for @xmath1652 . by lemmas 5.6.1 and 5.6.2 , ( 1 ) holds .
now @xmath2125 . since @xmath2126)\cup h(c_{i}^{l}\cap d_{j_{0}}^{l}\times [ -2,2])$ ] .
since @xmath1273 , @xmath590 is disjoint from @xmath1844 $ ] .
hence ( 2 ) holds .
q.e.d.0.5 mm * _ lemma 5.6.4 .
_ * suppose that @xmath2127 and @xmath1048 .
then \(1 ) @xmath2128 .
\(2 ) @xmath746 is disjoint from @xmath2129 .
\(3 ) @xmath746 intersects @xmath1304 in one point if and only if @xmath746 intersects @xmath579 in one point for @xmath2130 or @xmath2131 .
\(4 ) for @xmath1681 , each component of @xmath2132 is contained either in @xmath2133 or in @xmath1087 . *
_ * by definition 3.5.1 , @xmath2134 .
hence @xmath1334 , @xmath746 is disjoint from @xmath1291 and @xmath1850 for @xmath2135 or @xmath2136 .
now by the argument in lemma 5.6.3 , ( 1 ) , ( 2 ) and ( 3 ) hold .
suppose that @xmath2137
. then @xmath2138 .
hence @xmath2139 is disjoint from @xmath1291 and @xmath1850 . by the proof of lemma 5.6.3 ,
@xmath2140 is disjoint from @xmath1962 for @xmath1681 .
now we assume that @xmath1189 . since @xmath1048 , by proposition 4(3 ) and lemma 3.3.1 , @xmath746 is disjoint from @xmath1815 $ ] .
now @xmath2141 where @xmath2142 is an arc in @xmath2143 , @xmath2144 is an arc in @xmath2145 $ ] ( by lemma 5.1.2(7 ) ) , @xmath2146 is an arc in @xmath2147 . by definition 5.2.3 and definitions 5.3.2 and 5.3.3 , @xmath2148 $ ] .
hence @xmath2149 are disjoint from @xmath1962 for @xmath1681 . thus ( 4 ) holds .
q.e.d . in this section
, we shall prove the following lemma : * _ lemma 5.7.1 .
_ * there are two sets of pairwise disjoint disks @xmath2150 properly embedded in @xmath27 and @xmath2151 properly embedded in @xmath47 such that \(1 ) @xmath2152 , @xmath2153 ; \(2 ) @xmath2154 . * _ proof . _
* suppose that @xmath548 and @xmath288 .
then , by proposition 6 , @xmath1315 is a properly embedded disk in @xmath27 such that @xmath2155 . by lemma 5.1.4 , @xmath1315 is disjoint from @xmath1850 . by assumption ,
@xmath495 . by lemma 2.2.4 , @xmath1786 . by definition 2.3.1 , @xmath1779 .
now let @xmath2156 . by lemma 5.4.3(4 ) , @xmath1304 and @xmath2157 are obtained by doing band sums with copies @xmath541 to @xmath1292 and @xmath829 .
hence @xmath2158 bounds a disk in @xmath27 , denoted by @xmath553 . since @xmath2159 .
hence @xmath2160 .
suppose now that @xmath2080 and @xmath293 .
then , by proposition 6 , @xmath1312 is a properly embedded disk in @xmath564 such that @xmath2161 . by lemma 5.1.4 , @xmath1788 is disjoint from @xmath1850 .
we denote by @xmath2162 the disk @xmath1312 .
there are two cases : case 1 .
see definition 3.2.1 . since @xmath546 and @xmath1813 , by lemma 2.2.5 , @xmath2164 . by definition 2.3.1 ,
@xmath1311 . since @xmath1905 and @xmath2165 , @xmath2166 . by proposition 6 , @xmath2167 is disjoint from @xmath1788 .
hence @xmath2168 .
hence ( 2 ) holds .
@xmath1944 . in this case , @xmath2169 .
now by proposition 6 , @xmath1185 is disjoint from @xmath2170 . by definition 5.2.1 , @xmath1185 is disjoint from @xmath1316 . by definition 5.2.3 , @xmath1185 is disjoint from @xmath1419 .
hence @xmath2171 and @xmath2172.q.e.d . 0.5 mm in this section , we shall first construct @xmath2173 from @xmath1934 and @xmath1962 for @xmath2080 and @xmath1321 .
then we shall prove propositions 4 - 6 for the case : @xmath493 and @xmath495 . * _ construction ( * * ) .
_ * since @xmath2174 for @xmath1048 , @xmath2175 if @xmath1689 . by lemma 5.1.2(7 ) and lemma 5.1.3(6 ) , @xmath1820 $ ] is disjoint from @xmath1188 . by definitions 3.2.1 and lemma 5.4.4 , @xmath2176 , @xmath2177 if @xmath1075 , and @xmath2178 if @xmath1071 or @xmath1379 . without loss of generality
, we may assume that , in @xmath564 , @xmath1815\cap d^{*}_{0,\alpha}\subset d_{j_{0}}^{l}\times [ 0,2.5]$ ] for @xmath1149 , and @xmath1815\cap d^{*}_{0,\alpha}\subset d_{j_{0}}^{l}\times [ -2.5,0]$ ] for @xmath1348 .
see lemma 3.5.5 . by definition 5.2.2(2 ) and definition 5.3.1(2 ) , in @xmath1316 , @xmath2179 $ ] is a disk such that @xmath2180=\partial d_{j_{0}}^{l}\times [ -2,-1.5]$ ] , and @xmath2181 $ ] is a disk such that @xmath2182=\partial d_{j_{0}}^{l}\times [ 2,2.5]$ ] . for @xmath1149 ,
let @xmath2183 be a simple closed curve in @xmath2181 $ ] satisfying the following conditions : \(1 ) @xmath2183 intersects @xmath2184 $ ] in a core @xmath2185 of @xmath2184 $ ] where @xmath2186 , @xmath2183 intersects @xmath2043 in a core of @xmath2053 lying in @xmath2187 , say @xmath2188 .
\(2 ) if @xmath1041 , then @xmath2189 .
\(3 ) if @xmath1386 , then @xmath2190 . see figure 25 . for @xmath1348 ,
let @xmath2183 be a simple closed curve in @xmath2191 $ ] satisfying the following conditions : \(4 ) @xmath2183 intersects @xmath2191 $ ] in a core @xmath2185 of @xmath2191 $ ] where @xmath2192 , @xmath2183 intersects @xmath2043 in a core of @xmath2053 lying in @xmath2193 , say @xmath2188 .
\(5 ) if @xmath1042 , then @xmath2194 .
\(6 ) if @xmath1386 , then @xmath2190 . see figure 25 .
figure 25 now @xmath2183 bounds a disk in @xmath1972\subset f^{l+1}$ ] for @xmath1348 , and @xmath2183 bounds a disk in @xmath2181\subset f^{l+1}$ ] for @xmath1149 .
note that @xmath1066 , and @xmath2195 . since @xmath1064 and @xmath2196 is disjoint from @xmath1975 $ ] . by definition 5.2.3 ,
definitions 5.3.2 and 5.3.3 , @xmath2197 and @xmath2198 .
in particular , if @xmath1189 . by lemma 3.5.8 , definition 5.2.3 and definition 5.3.3 , we may assume that @xmath2199 and @xmath2200 .
figure 26 now let @xmath1358 be an arc in @xmath2201 connecting @xmath2202 to , @xmath2203 such that \(7 ) @xmath2204=a_{0,\alpha}-d_{j_{0}}^{l}\times [ -2.5,2.5]$ ] and @xmath2205 $ ] is an arc lying in @xmath2206\cup [ 2,2.5]$ ] .
now let @xmath1366 be an arc in @xmath2207 connecting @xmath1367 to , @xmath2208 such that \(8 ) @xmath2209=b_{0,\alpha}-d_{j_{0}}^{l}\times [ -2.5,2.5]$ ] and @xmath2210 $ ] is an arc lying in @xmath2211\cup [ 2,2.5]$ ] .
+ where @xmath2212 are as in lemma 3.5.5 .
see figure 26 . by lemma 5.1.4(1 ) and the proof of lemma 4.3.5 , @xmath1416 is disjoint from @xmath1291 and @xmath1850 . furthermore , @xmath1398 for @xmath2213 , @xmath1098 .
now let @xmath2214 , and @xmath1376 be a neighborhood of @xmath1375 in @xmath1316 satisfying the following conditions : \(9 ) @xmath2215 if @xmath1075 .
\(10 ) @xmath2216 if @xmath1071 or @xmath1379 .
\(11 ) @xmath2217 .
\(12 ) for @xmath1386 , @xmath2218 . q.e.d.(construction ( * * ) ) * _ lemma 5.8.1 . _ * ( 1 ) if @xmath2219 , then @xmath2220 .
furthermore , either @xmath1171 , or @xmath985 for some @xmath2221 .
\(2 ) @xmath2222 . if @xmath2223 , then @xmath2224 .
furthermore , @xmath2225 for some @xmath2221 . * _ proof .
_ * by the proof of lemma 4.3.5 and lemma 5.1.4(1 ) , @xmath2226 is disjoint from @xmath1850 . by construction(**)(7 ) and ( 8) , @xmath2227 is disjoint from @xmath1975 $ ] .
hence @xmath2227 is disjoint from @xmath2228\cup\partial v_{l+1}^{l}\times i$ ] .
suppose @xmath546 and @xmath132 . by definitions 5.2.3 , 5.3.2 and 5.3.3 , @xmath2229 and @xmath2230 . by lemma 4.3.5 , the lemma holds.q.e.d.0.5 mm * _ lemma 5.8.2 .
_ * @xmath2231 satisfying the following conditions : \(1 ) @xmath2232 is disjoint from @xmath1962 for @xmath1321 .
\(2 ) @xmath2233 is a core of @xmath2094 .
\(3 ) @xmath2188 is a core of @xmath2102 .
+ where @xmath2188 is as in construction(**)(1 ) . * _ proof . _ * ( 1 ) by definition 3.5.1 , @xmath2234 for each @xmath1048 .
hence @xmath2235 for @xmath2236 or @xmath2237 .
hence @xmath1848 $ ] is disjoint from @xmath732 . since @xmath2238\cup [ 2,2.5]$ ] , by definitions 5.2.3 , 5.3.2 and lemmas 5.1.2(8 ) , 5.1.3(4 ) , @xmath2239 is disjoint from @xmath1962 for each @xmath2240 .
\(2 ) by definitions 5.2.3 , 5.3.2 , @xmath2233 is a core of @xmath2094 .
\(3 ) follows from construction ( * * ) , definitions 5.2.3 and 5.3.2 .
q.e.d.0.5 mm * _ lemma 5.8.3 .
_ * ( 1 ) if @xmath2241 , then @xmath2242 is disjoint from @xmath1962 .
\(2 ) @xmath2243 .
\(3 ) if @xmath2244 , then each component of @xmath2245 is a core of @xmath2094 .
* _ proof . _ * ( 1 ) suppose that @xmath2241 .
since @xmath2246 , by lemma 5.1.2(2 ) and ( 4 ) , lemma 5.1.3(2 ) , @xmath582 is disjoint from @xmath2242 . by lemma 5.4.4 ,
hence ( 1 ) holds .
\(2 ) by construction(**)(1 ) , @xmath2247 . since @xmath2234 , by lemma 5.1.4(4 ) , @xmath2248 is disjoint from @xmath1291 and @xmath1978 .
furthermore , @xmath2249 , and @xmath2089\cup \partial v_{l+1}^{l}\times i$ ] . hence ( 2 ) holds .
\(3 ) suppose that @xmath2250
. then @xmath1683 . by ( 2 )
, @xmath2248 is disjoint from @xmath1291 . by lemma 5.5.2 , @xmath2251 is disjoint from @xmath1962 . by definition 5.2.3 and definition 5.3.3 , @xmath2252 . by lemma 5.1.2 and lemma 5.1.3 , ( 3 ) holds.q.e.d.0.5 mm * _ definition 5.8.4 . _ * ( 1 ) if @xmath1075 , then let @xmath1382 be a homeomorphism from @xmath1188 to @xmath1376 such that @xmath1382 is an identifying map on @xmath2253 .
\(2 ) if @xmath413 or @xmath1379 , let @xmath1382 be a homeomorphism from @xmath2254 to @xmath1375 such that @xmath1382 is an identifying map on @xmath2255.0.5 mm * _ definition 5.8.5 .
_ * ( 1 ) for @xmath546 , let @xmath2256 .
\(2 ) for @xmath1321 , if @xmath2257 , let @xmath2258 ; if @xmath2259 , let @xmath2260.0.5 mm * _ lemma 5.8.6 . _
* @xmath567 is isotopic to @xmath1934 in @xmath2261 . * _ proof . _ * by lemma 3.5.5 , @xmath2262 bounds a disk in @xmath1316 .
see figure 26 . by construction ( *
* ) , @xmath2183 bounds a disk in @xmath1316 .
hence @xmath1064 is isotopic to @xmath1375 . by lemma 5.6.1 and lemma 5.6.2
, each component of @xmath2263 is a core of @xmath1188 even if @xmath858 , say @xmath2264 . by definitions 5.8.4 and 5.8.5
, @xmath2265 is a core of @xmath1376 .
hence @xmath567 is isotopic to @xmath1934 .
. now we prove propositions 4 - 6 for the case : @xmath493 and @xmath495 . * _ the proofs of propositions 4 - 6 .
_ * by lemma 5.4.4 , lemmas 5.5.1 - 5.5.5 , @xmath1316 is generated by the abstract tree @xmath2266 satisfying the following conditions : \(1 ) if @xmath402 , then @xmath1994 is disjoint from @xmath2002 .
\(2 ) if @xmath404 , then @xmath2267 . by lemma 5.6.1 ,
if @xmath411 and @xmath2081 , then @xmath1934 is disjoint from @xmath1962 . by lemma 5.6.2 , each component of @xmath2092 is a core of @xmath2094 for @xmath2268 and @xmath717 . by lemma 5.6.3
, @xmath1934 is regular in @xmath1938 for @xmath1310 . by lemma 5.4.3
, @xmath1419 is an arc properly embedded in @xmath1316 which lies in one of @xmath1294 for some @xmath194 ; @xmath1946 for each @xmath1321 , and @xmath1947 for each @xmath1948 .
now if we take place of @xmath2269 with @xmath2270 , then proposition 4 holds except that \(i ) @xmath2271 is regular in @xmath2272 ; \(ii ) @xmath2273 intersects @xmath1304 in at most one point . +
where @xmath740 .
now by lemmas 5.6.4 , 5.8.1 , 5.8.2 , 5.8.3 and the argument in chapter 4 , proposition 4 holds for the case : @xmath493 and @xmath495 . by lemmas 5.6.3 , 5.6.4 , 5.8.2 and the proof of proposition 5 in section 4.6 , proposition 5 holds for the case : @xmath493 and @xmath495 .
proposition 6 follows from lemmas 5.8.6 , 5.7.1 and the proof of proposition 6 in section 4.6.q.e.d.0.5 mm
now we prove propositions 1 - 3 for @xmath493 under the assumptions that propositions 1 - 6 hold for @xmath492 .
then we finish the proofs of propositions 1 - 6 .
* acknowledgement : * the author is grateful to professor shicheng wang for many helpful discussions with the author on this problem and some comments on this paper .
the author is grateful to professors c. gordon , t. kobayashi , fengchun lei and ying - qing wu for some helpful discussions on this problem .
the author thank doctor jiming ma for careful reading of this paper . | cameron gordon([problem 3.91 in [ ki ] ) conjectured that a connected sum of two heegaard splittings is stabilized if and only if one of the two factors is stabilized . in this paper
, we shall prove this conjecture . * keywords * : abstract tree , band sum , connected sum , stabilization .
ams classification : 57m25 |
a variety of applications share the core task of selecting a subset of columns from a short , wide matrix @xmath3 with @xmath4 rows and @xmath5 columns
. the criteria for selecting these columns typically aim at preserving information about the span of @xmath3 while generating a well - conditioned submatrix .
classical and recent examples include experimental design , where we select observations or experiments @xcite ; preconditioning for solving linear systems and constructing low - stretch spanning trees ( here @xmath3 is a version of the node - edge incidence matrix and we select edges in a graph ) @xcite ; matrix approximation @xcite ; feature selection in @xmath0-means clustering @xcite ; sensor selection @xcite and graph signal processing @xcite . in this work ,
we study an elegant randomized approach that holds promise for all of these applications .
this approach relies on sampling columns of @xmath3 according to a probability distribution defined over its submatrices : the probability of selecting a set @xmath6 of @xmath0 columns from @xmath3 , with @xmath7 , is @xmath8 where @xmath9 is the submatrix consisting of the selected columns .
this distribution is reminiscent of _ volume sampling _
, where @xmath10 columns are selected with probability proportional to the determinant @xmath11 of a @xmath12 matrix , i.e. , the squared volume of the parallelepiped spanned by the selected columns .
( volume sampling does _ not _ apply to @xmath13 as the involved determinants vanish . ) in contrast , @xmath14 uses the determinant of an @xmath15 matrix and uses the volume spanned by the _ rows _ formed by the selected columns .
hence we refer to @xmath14-sampling as _ dual volume sampling ( dvs)_. [ [ contributions . ] ] contributions .
+ + + + + + + + + + + + + + despite the ostensible similarity between volume sampling and dvs , and despite the many practical implications of dvs outlined below , efficient algorithms for dvs are not known and were raised as open questions in @xcite . in this work ,
we make two key contributions : =2em we develop polynomial - time randomized sampling algorithms and their derandomization for dvs .
surprisingly , our proofs require only elementary ( but involved ) matrix manipulations .
we establish that @xmath14 is a _ strongly rayleigh _
measure @xcite , a remarkable property that captures a specific form of negative dependence .
our proof relies on the theory of real - stable polynomials , and the ensuing result implies a provably fast - mixing , practical mcmc sampler .
moreover , this result implies concentration properties for dual volume sampling .
the selection of @xmath16 columns from a short and wide matrix has many applications .
our algorithms for dvs hence have several implications and connections ; we note a few below .
* experimental design .
* the theory of optimal experiment design explores several criteria for selecting the set of columns ( experiments ) @xmath6 .
popular choices are @xmath17 here , @xmath18 denotes the moore - penrose pseudoinverse of @xmath3 , and @xmath6 the minimization ranges over all @xmath6 such that @xmath9 has full row rank @xmath4 .
a - optimal design , for instance , is statistically optimal for linear regression @xcite .
finding an optimal solution for these design problems is np - hard ; and most discrete algorithms use local search @xcite .
3.1 ) show that dual volume sampling yields an approximation guarantee for both a- and e - optimal design : if @xmath6 is sampled from @xmath14 , then @xmath19 \le {
m - n + 1\over k - n+1}\|a^\dagger\|_f^2;\quad \mathbb{e}\left[\|a_s^\dagger\|_2 ^ 2\right ] \le \left(1 + { n(m - k)\over k - n+1}\right ) \|a^\dagger\|_2 ^ 2.\end{aligned}\ ] ] @xcite provide a polynomial time sampling algorithm only for the case @xmath20 .
our sampling algorithms achieve the bound in expectation , and the derandomization in section [ sec : derand ] achieves the bound deterministically .
@xcite recently ( in parallel ) achieved approximation bounds for a - optimality via a different algorithm combining a convex relaxation and greedy method .
other methods for experimental design includes leverage score sampling @xcite and predictive length sampling @xcite . * low - stretch spanning trees and applications . * objectives [ eq : opt ] also arise in the construction of low - stretch spanning trees , which have important applications in graph sparsification , preconditioning and solving symmetric diagonally dominant ( sdd ) linear systems @xcite , among others @xcite . in the node - edge incidence matrix @xmath21 of an undirected graph @xmath22 with @xmath4 nodes and @xmath23 edges ,
the column corresponding to edge @xmath24 is @xmath25 .
let @xmath26 be the svd of @xmath27 with @xmath28 .
the stretch of a spanning tree @xmath29 in @xmath22 is then given by @xmath30 @xcite . in those applications ,
we hence search for a set of edges with low stretch .
* network controllability . *
the problem of sampling @xmath31 columns in a matrix also arise in network controllability problems .
for example , in @xcite , the authors consider selecting control nodes @xmath6 ( under certain constraints ) over time in complex networks to control a linear time - invariant network . after transforming the problem into a column subset selection problem from a short and wide controllability matrix ,
the objective becomes essentially an e - optimal design problem , for which the authors use greedy heuristics .
we focus on a specific , widely applicable setup of column subset selection : _ picking columns from a short and wide matrix _ ( to optimize certain criteria ) .
this setup was analyzed by @xcite where they minimize a norm of the pseudoinverse of the induced submatrix .
formally , let @xmath32 with @xmath5 be a full row - rank matrix , and @xmath0 a sampling parameter with @xmath2 .
let @xmath33 denote the rank of @xmath3 , thus @xmath34 .
let @xmath35 denote the submatrix of @xmath3 with rows and columns indexed by @xmath36 $ ] and @xmath37 $ ] , respectively .
for brevity we use @xmath9 to denote columns of @xmath3 indexed by @xmath37 $ ] .
the objective is to find a subset @xmath38 $ ] , such that @xmath39 @xmath40 denotes the frobenius norm or spectral norm , respectively , and @xmath18 denotes the moore - penrose pseudoinverse of @xmath3 .
the initial motivation of @xcite for using this objective was to find low - stretch spanning trees in an undirected graph @xcite , a problem that reduces to with @xmath41 .
however , problem arises much more widely . for example
, it is known that a solution to is the statistically optimal solution for experimental design in linear regression @xcite .
additional applications include low - rank matrix approximation @xcite , feature selection in @xmath0-means clustering @xcite and sensor selection @xcite , among many others .
given its many applications , finding efficient methods to solve is valuable .
a promising approach advanced in @xcite is a sampling algorithm based on determinants of submatrices .
although @xcite refer to their approach as `` volume sampling , '' it is different from the original volume sampling method of @xcite therefore , to avoid confusion , we will refer to the method as _ dual volume sampling ( dvs ) _
( see [ sec : setup ] ) .
it has been shown that dvs solves well .
specifically , if we sample a subset @xmath6 of cardinality @xmath31 via dvs , then the approximation guarantees hold @xmath19 \le { m - n + 1\over k - n+1}\|a^\dagger\|_f^2;\quad \mathbb{e}\left[\|a_s^\dagger\|_2 ^ 2\right ] \le \left(1 + { n(m - k)\over k - n+1}\right ) \|a^\dagger\|_2 ^ 2.\end{aligned}\ ] ] unfortunately , @xcite provide a polynomial time sampling algorithm only for the case @xmath20 .
no polynomial time algorithm for dvs for the more important setting @xmath13 is currently known .
further , no _
derandomized _ algorithm that satisfies the dvs bounds deterministically is known either .
@xcite raise development of efficient algorithms for both these problems as open problems .
the main focus of this paper is to provide positive answers to these open problems . as a result
, dvs becomes a practical method for column subset selection ( in the harder @xmath42 regime ) .
in particular , we provide two polynomial time sampling methods for dvs : ( i ) an exact algorithm polynomial in @xmath23 , @xmath4 ; and ( ii ) an efficient approximate sampling algorithm via markov chains .
moreover , we also present a derandomized version of our exact sampling algorithm that deterministically samples a subset @xmath6 satisfying .
our exact sampling algorithm uses a perturbation argument to compute the partition function and marginal probabilities .
this computation , if done naively , will take time exponential in @xmath0 .
we refer to the perturbation as the `` _
_ @xmath43-trick _ _ '' .
further , based on computation of marginal probabilities , we propose a derandomization of our first algorithm to obtain a method that samples @xmath6 deterministically . to obtain our fast mixing markov chain sampler
, we appeal to the powerful theory of stable polynomials and _ strongly rayleigh _
measures @xcite .
in particular , we first establish that the measure induced by dual volume sampling is ( homogeneously ) strongly rayleigh .
thereafter , fast mixing follows from the recent result of @xcite , that provides a fast mixing ( for suitably initialized ) markov chain for sampling from strongly rayleigh measures .
* contributions . * in summary ,
the key contributions of this paper are : 1 .
the _ first _ polynomial - time exact dvs algorithm ( section [ sec : dualvol ] ) 2 .
the _ first _ polynomial - time derandomized dvs algorithm that deterministically generates a subset @xmath6 satisfying ( section [ sec : derand ] ) .
3 . a fast - mixing approximate dvs sampler built on markov chains for strongly rayleigh measures ;
this yields a method that runs in time _
linear _ in the dataset size @xmath23 ( section [ sec : mcmc ] ) .
[ [ notation . ] ] notation . + + + + + + + + + from a matrix @xmath44 with @xmath45 columns , we sample a set @xmath38 $ ] of @xmath0 columns ( @xmath7 ) , where @xmath46 : = \{1,2,\ldots , m\}$ ] .
we denote the singular values of @xmath3 by @xmath47 , in decreasing order .
we will assume @xmath3 has full row rank @xmath48 , so @xmath49 .
we also assume that @xmath50 for every @xmath38 $ ] where @xmath51 .
by @xmath52 , we denote the @xmath0-th elementary symmetric polynomial of @xmath3 , i.e. , the @xmath0-th coefficient of the characteristic polynomial @xmath53 . similarly , one can define the distribution @xmath54 induced by _ volume sampling _ , that is , @xmath55 in this case , one samples a subset @xmath38 $ ] such that @xmath56 .
observe that while ostensibly similar , these distributions are quite different : one is for sampling @xmath51 and the other for @xmath56 .
both are related to volumes induced by submatrices @xmath14 is proportional to the square of the volume of the parallelepiped spanned by the _ rows _ of @xmath9 , while @xmath54 is proportional to the volume spanned by the _ columns _ of @xmath9 .
these differences and connections make @xmath14 closely related yet different from @xmath54 , which explains our choice of name `` dual volume sampling . ''
we describe in this section our method to sample from the distribution @xmath14 .
our first method relies on the key insight that , as we show , the marginal probabilities for dvs can be computed in polynomial time . to demonstrate this , we begin with the partition function and then derive marginals .
the partition function has a convenient simple closed form , which follows from the cauchy - binet formula and was also derived in @xcite .
[ lem : partition ] for @xmath32 with @xmath34 and @xmath57 , we have @xmath58}\det(a_s a_s^\top ) = \binom{m - n}{k - n}\det(a a^\top).\ ] ] next , we will need the marginal probability @xmath59 that a given set @xmath60 $ ] is a subset of the random set @xmath6 . in the following theorem
, the set @xmath61\setminus t$ ] denotes the ( set ) complement of @xmath29 , and @xmath62 denotes the orthogonal complement of @xmath63 .
[ thm : marginal ] let @xmath64 $ ] , @xmath65 , and @xmath66 .
let @xmath67 be the singular value decomposition of @xmath68 where @xmath69 , and @xmath70 .
further define the matrices @xmath71 q^\top a_{t_c } \in\mathbb{r}^{r(a_t)\times ( m-|t|)}.\end{aligned}\ ] ] let @xmath72 be the eigenvalue decomposition of @xmath73 where @xmath74 .
moreover , let @xmath75 $ ] and @xmath76 .
then the marginal probability of @xmath29 in dvs is @xmath77 \times \left[\prod_{j=1}^{r(b)}\sigma_j^2(b)\right ] \times \gamma \over z_a}.\end{aligned}\ ] ] we prove theorem [ thm : marginal ] via a perturbation argument that connects dvs to volume sampling .
specifically , observe that for @xmath78 and @xmath79 it holds that @xmath80 carefully letting @xmath81 bridges volumes with `` dual '' volumes .
the technical remainder of the proof further relates this equality to singular values , and exploits properties of characteristic polynomials .
a similar argument yields an alternative proof of lemma [ lem : partition ] .
we show the proofs in detail in appendix [ app : sec : partition ] and [ app : sec : marginal ] respectively . in this section ,
we show how to compute the partition function and marginal distributions for @xmath82 .
we base our analysis on a simple perturbation argument , which we call the _ @xmath43-trick_.
this serves as a bridge between dual volume sampling and volume sampling , based on which we can perform the desired computations in polynomial time .
theorem [ thm : partition ] illustrates the @xmath43-trick , which is subsequently used in theorem [ thm : marginal ] to obtain polytime computation of the marginal distribution .
this theorem is the key building block for our final polynomial time exact dvs algorithm of section [ sec : dualvol ] .
before we dive into partition functions and marginals , we recall two easily verified facts about determinants that will be useful in our analysis : @xmath83 the partition function of @xmath82 , happens to have a pleasant closed - form formula . although this formula is known @xcite , and follows immediately by an application of the cauchy - binet identity , we present an alternative proof based on the @xmath43-trick perturbation for its conceptual value and subsequent use .
[ thm : partition ] given @xmath32 where @xmath34 and @xmath57 , we have @xmath84}\det(a_s a_s^\top ) = \left(\begin{array}{c}m - n\\k - n\end{array}\right ) \det(a a^\top).\end{aligned}\ ] ] first note that for @xmath57 and any @xmath66 , by we have @xmath85 taking limits as @xmath86 on both sides we have @xmath87 let us focus on @xmath88 .
we construct an identity matrix @xmath89 , then we have @xmath90}\right ) .
\end{split}\ ] ] in other words , this value is proportional to the probability of sampling columns from @xmath91}$ ] using volume sampling .
therefore , using the definition of @xmath92 we have @xmath93}&\det(a_s^\top a_s + { \varepsilon}i_k ) = { 1\over { \varepsilon}^{k - n } } e_k ( a^\top a + { \varepsilon}i_m)\\ & = { 1\over { \varepsilon}^{k - n } } e_k ( \operatorname{diag}([(\sigma_1 ^ 2(a ) + { \varepsilon } ) , ( \sigma_2 ^ 2(a ) + { \varepsilon}),\ldots , ( \sigma_n^2(a ) + { \varepsilon } ) , { \varepsilon},\ldots,{\varepsilon}]))\\ & = \binom{m - n}{k - n}\prod\nolimits_{i=1}^n ( \sigma_i^2(a ) + { \varepsilon } ) + o({\varepsilon}).\end{aligned}\ ] ] now taking the limit as @xmath86 we obtain @xmath84}\det(a_s a_s^\top ) & = \lim_{{\varepsilon}\to 0}\binom{m - n}{k - n}\prod\nolimits_{i=1}^n ( \sigma_i^2(a ) + { \varepsilon } ) + o({\varepsilon } ) = \binom{m - n}{k - n}\det(a a^\top).\end{aligned}\ ] ] we can see that in this proof , the @xmath43-trick works as a bridge connecting dual volume sampling and volume sampling , whose partition function is easy to compute .
this perturbation trick proves essential in computing the marginal distribution , as shown below .
[ thm : marginal ] let @xmath64 $ ] , @xmath65 , and let @xmath94\backslash t$ ] . we write @xmath95 as the singular value decomposition of @xmath68 where @xmath96 , @xmath97 be the complement column space of @xmath98 . further we define the two matrices @xmath99 q_t^\top a_{t_c } \in\mathbb{r}^{r(a_t)\times ( m-|t|)}.\end{aligned}\ ] ]
let @xmath100 be the eigenvalue decomposition of @xmath101 where @xmath102 .
moreover , let @xmath103 $ ] and @xmath104 .
then the marginal probability of sampling @xmath29 via dual volume sampling is given by @xmath105 \times \left[\prod_{j=1}^{r(b_{t_c})}\sigma_j^2(b_{t_c})\right ] \times \gamma_t \over \binom{m - n}{k - n}\det(a a^\top)}.\end{aligned}\ ] ] the proof is an application of the @xmath43-trick ; due to its length we defer it to appendix [ app : sec : marginal ] . [ [ complexity . ] ] complexity .
+ + + + + + + + + + + the numerator of @xmath106 in theorem [ thm : marginal ] requires @xmath107 time to compute the first term , @xmath107 to compute the second and @xmath108 to compute the third .
the denominator takes @xmath107 time , amounting in a total time of @xmath108 to compute the marginal probability .
the marginal probabilities derived above directly yield a polynomial - time _ exact _
dvs algorithm . instead of @xmath0-sets
, we sample ordered @xmath0-tuples @xmath109^k$ ] .
we denote the @xmath0-tuple variant of the dvs distribution by @xmath110 : @xmath111 sampling @xmath112 is now straightforward . at the @xmath113th step we sample @xmath114 via @xmath115 ; these probabilities are easily obtained from the marginals in theorem [ thm : marginal ] .
[ coro : cond ] let @xmath116 , and @xmath106 as in thm .
[ thm : marginal ] . then , @xmath117 as a result , it is possible to draw an exact dual volume sample in time @xmath118 .
the full proof may be found in the appendix .
the running time claim follows since the sampling algorithm invokes @xmath119 computations of marginal probabilities , each costing @xmath108 time .
we are now ready to state our polynomial time exact dvs algorithm . instead of focusing on @xmath0-subsets
, the key idea here is to use dvs for @xmath0-tuples , and to decompose the tuple distribution into a product of conditional distributions that allow us to recover the desired @xmath0-subset . a similar strategy
is followed by @xcite .
denote @xmath0-tuple dvs distribution as @xmath110 , and let @xmath109^k$ ] be the sampled tuple .
since each permutation of the @xmath120 is equally likely , it follows that @xmath121 the joint probability of the tuple can be decomposed into a product of conditional probabilities as @xmath122 given these conditional probabilities , sampling @xmath123 is straightforward : in the @xmath113-th iteration , we calculate the conditional probability of @xmath124 given @xmath125 for each @xmath126 $ ] , and then sample @xmath127 with this conditional probability .
the full algorithm is outlined as algorithm [ algo : dualvol ] .
thus , a polynomial - time algorithm for computing conditional distributions directly implies a polynomial - time _
exact _ dual volume sampling algorithm .
the computation of the conditional probability is based on theorem [ thm : marginal ] , and is given in the following corollary : * input : * matrix @xmath32 to sample columns from , @xmath2 the target size * output : * set @xmath6 such that @xmath128 and @xmath129 .
initialize @xmath112 as empty tuple compute the conditional distribution @xmath130 ( corollary [ coro : cond ] ) sample @xmath131 with probability proportional to @xmath132 @xmath133 output @xmath112 as a set @xmath6 .
[ coro : cond ] let @xmath116 , then @xmath134 \times \left[\prod_{j=1}^{r(b_{t_c\backslash\{i\}})}\sigma_j^2(b_{t_c\backslash\{i\}})\right ] \times \gamma_{t\cup\{i\ } } \over ( k - t+1)\left[\prod_{i=1}^{r(a_t)}\sigma_i^2(a_t)\right ] \times \left[\prod_{j=1}^{r(b_{t_c})}\sigma_j^2(b_{t_c})\right ] \times \gamma_t}.\end{aligned}\ ] ] let @xmath116 . by the definition of conditional probability and theorem [ thm : marginal ] we have @xmath135,r\cap t = \emptyset}\det(a_{t\cup\{i\}\cup r}a_{t\cup\{i\}\cup r}^\top)\over{(k - t+1)!\over k!}\sum_{|r| = k - t+1,r\subseteq[n],r\cap t = \emptyset}\det(a_{t\cup r}a_{t\cup r}^\top)}\\ & = { \left[\prod_{i=1}^{r(a_{t\cup\{i\}})}\sigma_i^2(a_{t\cup\{i\}})\right ] \times \left[\prod_{j=1}^{r(b_{t_c\backslash\{i\}})}\sigma_j^2(b_{t_c\backslash\{i\}})\right ] \times \gamma_{t\cup\{i\ } }
\over(k - t+1 ) \left[\prod_{i=1}^{r(a_t)}\sigma_i^2(a_t)\right ] \times \left[\prod_{j=1}^{r(b_{t_c})}\sigma_j^2(b_{t_c})\right ] \times \gamma_t}.\end{aligned}\ ] ] [ [ computational - cost . ] ] computational cost .
+ + + + + + + + + + + + + + + + + + + it takes @xmath108 time to compute the conditional probability for one element given @xmath136 already selected elements .
since this probability has to be computed for @xmath137 elements and there are a total of @xmath0 such iterations , the net running time of algorithm [ algo : dualvol ] is @xmath118 .
[ [ remark ] ] remark + + + + + + a potentially more efficient approximate algorithm could be derived by noting the relations between volume sampling and dvs .
specifically , we add a small perturbation to dvs as in equation [ eq : eps ] to transform it into a volume sampling problem , and apply random projection for more efficient volume sampling as in @xcite .
please refer to appendix [ app : sec : approx ] for more details .
[ [ alternative - random - projections . ] ] alternative : random projections .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + given the connection between dvs and volume sampling via perturbation , one may consider an alternative sampling method that performs volume sampling on @xmath138^\top$ ] .
doing so introduces a small distortion factor of order @xmath139 .
this procedure can be further accelerated by applying volume - preserving random projections @xcite , resulting in an algorithm with running time @xmath140 .
this time is linear in @xmath23 but of high order in @xmath0 , and thus efficient only if @xmath141 . for sdd matrices ,
recent work may yield improvements to .
[ thm : approx ] for any @xmath142 and @xmath143 there is an algorithm that , in time @xmath144 , samples a subset from an approximate distribution @xmath145 with @xmath146 and @xmath147.\end{aligned}\ ] ] in summary , we obtain two sampling algorithms : an exact one with running time @xmath118 , and an approximate one with time @xmath140 .
next , we derandomize the above sampling algorithm to _ deterministically _ select a subset that satisfies the bound for the frobenius norm , thereby answering another question in @xcite .
the key insight for derandomization is that conditional expectations can be computed in polynomial time , given the marginals in theorem [ thm : marginal ] : [ cor : conditional ] let @xmath148^{t-1}$ ] be such that the marginal distribution satisfies @xmath149 .
the conditional expectation can be expressed as @xmath150 = \frac{\sum_{j=1}^n p ' ( \{i_1 , \ldots , i_{t-1}\ } \subseteq s | s\sim p(s;a_{[n]\setminus \{j\ } } ) ) } { p ' ( \{i_1 , \ldots , i_{t-1}\ } \subseteq s | s\sim p(s;a ) ) } , \end{aligned}\ ] ] where @xmath151 are the unnormalized marginal distributions , and it can be computed in @xmath152 time .
we show the full derivation in appendix [ app : sec : derand ] .
corollary [ cor : conditional ] enables a greedy derandomization procedure . starting with the empty tuple @xmath153 , in the @xmath154th iteration , we greedily select @xmath155 $ ] and append it to our selection : @xmath156 .
the final set is the non - ordered version @xmath157 of @xmath158 .
theorem [ thm : derandbounds ] shows that this greedy procedure succeeds , and implies a deterministic version of the bound ( [ eq : bounds ] ) .
[ thm : derandbounds ] the greedy derandomization selects a column set @xmath6 satisfying @xmath159 the proof is done by constructing an iterative algorithm where in each iteration , we go through all elements that have not been selected and compute the expectation conditioned on this element included to current set .
then we choose the element with the lowest conditional expectation to actually include into current set .
such inclusion of elements will only decrease the conditional expectation , thus retaining the bound in theorem [ thm : derandbounds ] .
detailed proof is deferred to appendix [ app : sec : greedy ] . * complexity . * at each iteration , greedy selection requires @xmath152 to compute the conditional expectation for @xmath160 items .
thus , the total running time for @xmath0 iterations is @xmath161 .
the approximation bound for the spectral norm is slightly worse than that in , but of the same order if @xmath162 .
[ [ complexity.-1 ] ] complexity .
+ + + + + + + + + + + at each iteration , algorithm [ algo : derand ] takes @xmath152 to compute the conditional expectation for @xmath160 items , thus the total running time for @xmath0 iterations is @xmath161 . also note that the approximation bound for spectral norm is slightly worse than that in , though it will of the same order if @xmath162
next , we investigate dvs more deeply and discover that it possesses a remarkable structural property , namely , the _ strongly rayleigh ( sr ) _
@xcite property .
this property has proved remarkably fruitful in a variety of recent contexts , including recent progress in approximation algorithms @xcite , fast sampling @xcite , graph sparsification @xcite , extensions to the kadison - singer problem @xcite , and certain concentration of measure results @xcite , among others . for dvs
the sr property has two major consequences : it leads to a fast mixing practical mcmc sampler , and it implies results on concentration of measure . [ [ strongly - rayleigh - measures . ] ] strongly rayleigh measures .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + sr measures were introduced in the landmark paper of @xcite , who develop a rich theory of negatively associated measures . in particular , we say that a probability measure @xmath163}\to { \mathbf{r}}_+$ ] is _ negatively associated _ , if @xmath164 , for @xmath165 increasing functions on @xmath166}$ ] with _ disjoint _ support .
this property reflects a `` repelling '' nature of @xmath167 , a property that occurs more broadly across probability , combinatorics , physics , and other fields see @xcite and references therein .
the negative association property turns out to be quite subtle in general ; the class of sr measures captures a strong notion of negative association and provides a framework for analyzing such measures .
specifically , sr measures are defined via their connection to real - stable polynomials @xcite .
a multivariate polynomial @xmath168 $ ] where @xmath169 is called _ real stable _ if all its coefficients are real and @xmath170 whenever @xmath171 for @xmath172 .
a measure is called an _ sr measure _ if its multivariate generating polynomial @xmath173 } \mu(s)\prod_{i\in s}z_i$ ] is real - stable .
notable examples of sr measures are determinantal point processes @xcite , balanced matroids @xcite , bernoullis conditioned on their sum , among others .
it is known ( see ( * ? ? ?
* pg . 523 ) ) that the class of sr measures is exponentially larger than the class of determinantal measures .
theorem [ thm : sr ] establishes the sr property for dvs and is the main result of this section . here and
in the following , we use the notation @xmath174 .
[ thm : sr ] let @xmath175 and @xmath2 .
then the multiaffine polynomial @xmath176 } \det(a_s a_s^\top)\prod_{i\in s}z_i\quad=\quad \sum_{|s|=k , s\subseteq
[ m ] } \det(a_s a_s^\top)z^{s},\ ] ] is real stable .
consequently , @xmath14 is an sr measure .
the proof of theorem [ thm : sr ] relies on key properties of real stable polynomials and sr measures established in @xcite .
essentially , the proof demonstrates that the generating polynomial of @xmath177 can be obtained by applying a few carefully chosen stability preserving operations to a polynomial that we know to be real stable .
stability , although easily destroyed , is closed under several operations noted in the important proposition below .
[ prop : basic ] let @xmath178 be a stable polynomial .
the following properties preserve stability : * substitution * : @xmath179 for @xmath180 ; * differentiation * : @xmath181 for any @xmath38 $ ] ; * diagonalization * : @xmath182 is stable , and hence @xmath183 ; and * inversion * : @xmath184 .
in addition , we need the following two propositions for proving theorem [ thm : sr ] . [ prop : det ] let @xmath185 be hermitian , @xmath186 and @xmath187 ( @xmath188 be hermitian semidefinite matrices .
then , the following polynomial is stable : @xmath189 [ prop : elsym ] for @xmath190 and @xmath191 , we have @xmath192 .
let @xmath193_{i=1}^m)$ ] be an diagonal matrix . using the cauchy - binet identity we have @xmath194 } \det ( ( ay)_{:,t})\det((a^\top)_{t , : } ) = { \sum\nolimits}_{|t|=n , t\subseteq [ m]}\det(a_t^\top a_t)y^t .
\end{split}\ ] ] thus , when @xmath195 , the ( diagonal ) indicator matrix for @xmath6 , we obtain @xmath196 .
consequently , in the summation above only terms with @xmath197 survive , yielding @xmath198 we are now ready to sketch the proof of theorem [ thm : sr ] .
notationally , it is more convenient to prove that the `` complement '' polynomial @xmath199}\det(a_sa_s^\top)z^{s_c}$ ] is stable ; subsequently , an application of prop .
[ prop : basic]-(iv ) yields stability of . using matrix notation @xmath200 , @xmath201 , our starting stable polynomial ( this stability follows from prop . [ prop : det ] ) is @xmath202 which can be expanded as @xmath203}\det(w_s+l_s)z^{s_c } = { \sum\nolimits}_{s\subseteq [ m ] } \left({\sum\nolimits}_{t\subseteq s}w^{s\setminus t}\det(l_{t , t})\right)z^{s_c}.\ ] ] thus , @xmath204 is real stable in @xmath205 variables , indexed below by @xmath6 and @xmath206 where @xmath207 . instead of the form
above , we can sum over @xmath208 $ ] but then have to constrain the support to the case when @xmath209 and @xmath210 .
in other words , we may write ( using iverson - brackets @xmath211 ) @xmath212}{\llbracket s_c\cap r=\emptyset \wedge s_c\cap t=\emptyset \rrbracket}\det(l_{t , t})z^{s_c}w^{r}.\ ] ] next , we truncate polynomial at degree @xmath213 by restricting @xmath214 . by ( * ? ? ?
* corollary 4.18 ) this truncation preserves stability , whence @xmath215 \\ |s_c\cup r|=m - n}}{\llbracket s_c\cap r=\emptyset \rrbracket}\det(l_{s\backslash r , s\backslash r})z^{s_c}w^{r},\ ] ] is also stable . using prop .
[ prop : basic]-(iii ) , setting @xmath216 retains stability ; thus @xmath217 \\
r|=m - n } } { \llbracket s_c\cap r=\emptyset \rrbracket}\det(l_{s\backslash r , s\backslash r})z^{s_c}y^{|r|}\\ & = \sum_{s\subseteq [ m]}\bigl({\sum\nolimits}_{|t|=n , t\subseteq s}\det(l_{t , t})\bigr)y^{|s|-|t|}z^{s_c } = \sum_{s \subseteq [ m ] } e_n(l_{s , s})y^{|s|-n}z^{s_c},\end{aligned}\ ] ] is also stable .
next , differentiating @xmath218 , @xmath219 times with respect to @xmath220 and evaluating at @xmath221 preserves stability ( prop .
[ prop : basic]-(ii ) and ( i ) ) .
in doing so , only terms corresponding to @xmath222 survive , resulting in @xmath223}e_n(l_{s , s})z^{s_c } = ( k - n)!\sum_{|s|=k , s\subseteq[m]}\det(a_s a_s^\top ) z^{s_c},\ ] ] which is just @xmath224 ( up to a constant ) ; here , the last equality follows from prop .
[ prop : elsym ] .
this establishes stability of @xmath224 and hence of @xmath225 .
since @xmath225 is in addition multiaffine , it is the generating polynomial of an sr measure , completing the proof .
the sr property of @xmath14 established in theorem [ thm : sr ] implies a fast mixing markov chain for sampling @xmath6 .
the states for the markov chain are all sets of cardinality @xmath0 .
the chain starts with a randomly - initialized active set @xmath6 , and in each iteration we swap an element @xmath226 with an element @xmath227 with a specific probability determined by the probability of the current and proposed set .
the stationary distribution of this chain is the one induced by dvs , by a simple detailed - balance argument .
the chain is shown in algorithm [ algo : mcmc ] .
* input : * @xmath32 the matrix of interest , @xmath0 the target cardinality * output : * @xmath228 initialize @xmath38 $ ] such that @xmath222 and @xmath229 draw @xmath230 uniformly pick @xmath226 and @xmath231\backslash s$ ] uniformly randomly + @xmath232 @xmath233 with probability @xmath234 the convergence of the markov chain is measured via its mixing time : the _ mixing time _ of the chain indicates the number of iterations @xmath235 that we must perform ( starting from @xmath236 ) before we can consider @xmath237 as an approximately valid sample from @xmath14 .
formally , if @xmath238 is the total variation distance between the distribution of @xmath237 and @xmath14 after @xmath235 steps , then @xmath239 is the _ mixing time _ to sample from a distribution @xmath43-close to @xmath14 in terms of total variation distance .
we say that the chain mixes fast if @xmath240 is polynomial in the problem size .
the fast mixing result for algorithm [ algo : mcmc ] is a corollary of theorem [ thm : sr ] combined with a recent result of @xcite on fast - mixing markov chains for homogeneous sr measures .
theorem [ thm : fastmc ] states this more precisely .
[ thm : fastmc ] the mixing time of markov chain shown in algorithm [ algo : mcmc ] is given by @xmath241 since @xmath14 is @xmath0-homogeneous sr by theorem [ thm : sr ] , the chain constructed for sampling @xmath6 following that in @xcite mixes in @xmath242 time .
[ [ implementation . ] ] implementation .
+ + + + + + + + + + + + + + + to implement algorithm [ algo : mcmc ] we need to compute the transition probabilities @xmath243 .
let @xmath244 and assume @xmath245 .
by the matrix determinant lemma we have the acceptance ratio @xmath246 thus , the transition probabilities can be computed in @xmath247 time .
moreover , one can further accelerate this algorithm by using the quadrature techniques of @xcite to compute lower and upper bounds on this acceptance ratio to determine early acceptance or rejection of the proposed move .
[ [ initialization . ] ] initialization .
+ + + + + + + + + + + + + + + a remaining question is initialization .
since the mixing time involves @xmath248 , we need to start with @xmath236 such that @xmath249 is sufficiently bounded away from @xmath221 .
we show in appendix [ app : sec : init ] that by a simple greedy algorithm , we are able to initialize @xmath6 such that @xmath250 , and the resulting running time for algorithm [ algo : mcmc ] is @xmath251 , which is _ linear _ in the size of data set @xmath23 and is efficient when @xmath0 is not too large . [ [ concentration . ] ] concentration .
+ + + + + + + + + + + + + + @xcite show concentration results for strong rayleigh measures . as a corollary of our theorem [ thm : sr ] together with their results , we directly obtain tail bounds for dvs . [ [ algorithms - for - experimental - design . ] ] algorithms for experimental design .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + widely used , classical algorithms for finding an approximate optimal design include fedorov s exchange algorithm @xcite ( a greedy local search ) and simulated annealing @xcite .
both methods start with a random initial set @xmath6 , and greedily or randomly exchange a column @xmath252 with a column @xmath253 .
apart from very expensive running times , they are known to work well in practice @xcite .
yet so far there is no theoretical analysis , or a principled way of determining when to stop the greedy search . curiously , our mcmc sampler is essentially a randomized version of fedorov s exchange method .
the two methods can be connected by a unified , simulated annealing view , where we define @xmath254 with temperature parameter @xmath255 .
driving @xmath255 to zero essentially recovers fedorov s method , while our results imply fast mixing for @xmath256 , together with approximation guarantees . through this lens
, simulated annealing may be viewed as initializing fedorov s method with the fast - mixing sampler . in practice
, we observe that letting @xmath257 improves the approximation results , which opens interesting questions for future work .
[ thm : fastmc ] the mixing time of markov chain shown in algorithm [ algo : mcmc ] is given by @xmath241 the proof of the fast mixing time bound relies on the theory of strongly rayleigh measures @xcite , for which the following important result was recently shown in @xcite : [ thm : srmix ] for any strongly rayleigh k - homogeneous probability distribution @xmath258}\to \mathbb{r}_+$ ] , @xmath259 $ ] where @xmath260 and @xmath66 , we have @xmath261 . * ( theorem [ thm : fastmc ] ) .
* since @xmath177 is @xmath262-homogeneous strongly rayleigh by theorem [ thm : sr ] , the chain constructed for sampling @xmath263 following that in @xcite mixes in @xmath242 time .
one can first sample @xmath263 and take the complement to get @xmath6 when the chain has mixed well , resulting in algorithm [ algo : mcmc ] .
it remains to prove theorem [ thm : sr ] .
let @xmath264\backslash s$ ] be the complement of @xmath6 .
theorem [ thm : sr ] proves that the distribution @xmath265 , which is closely related to @xmath14 , is strongly rayleigh .
note that @xmath177 samples @xmath266\backslash s$ ] .
we also use the notation @xmath174 .
[ thm : sr ] let @xmath175 and @xmath2 .
then the multiaffine polynomial @xmath176 } \det(a_s a_s^\top)\prod_{i\in s_c}z_i\quad=\quad \sum_{|s|=k , s\subseteq
[ m ] } \det(a_s a_s^\top)z^{s_c},\ ] ] is real stable . in other words
, @xmath177 is a @xmath0-homogeneous strongly rayleigh ( sr ) measure .
the proof of theorem [ thm : sr ] relies on key properties of real stable polynomials and sr measures established in the remarkable work of @xcite .
a polynomial @xmath168 $ ] where @xmath169 is called real stable if all its coefficients are real and @xmath170 whenever @xmath171 for @xmath172
. a measure is called sr measure if its generating polynomial is real stable .
we have the following properties on real stable polynomial that are useful in our proof : [ prop : det ] let @xmath185 be hermitian , @xmath186 and @xmath187 ( @xmath188 be hermitian semidefinite matrices .
then , the following polynomial is stable : @xmath189 stability can be subtle ; although easily destroyed , it can sometimes be preserved , as noted below .
[ prop : basic ] let @xmath178 be a stable polynomial .
the following closure properties hold : * substitution * : @xmath179 for @xmath180 is stable * differentiation * : @xmath181 for any @xmath38 $ ] is stable ; * diagonalization * : @xmath182 is stable , and hence @xmath183 is stable . with propositions [ prop : det ] and [ prop : basic ]
in hand , we are now ready to prove theorem [ thm : sr ] . * ( theorem [ thm : sr ] ) .
* let @xmath193_{i=1}^m)$ ] be an @xmath267 diagonal matrix of indeterminates . using the cauchy - binet identity we obtain @xmath194 } \det ( ( ay)_{:,t})\det((a^\top)_{t , : } ) = { \sum\nolimits}_{|t|=n , t\subseteq [ m]}\det(a_t^\top a_t)y^t . \end{split}\ ] ] thus ,
when @xmath268 is a diagonal matrix with ones for indices in @xmath6 and zeros otherwise , we obtain @xmath196 . consequently , in the summation above only terms with @xmath197 survive , yielding @xmath269 where we introduced @xmath191 .
thus we can rewrite the polynomial as @xmath270 } \det(a_sa_s^\top)z^{s_c } = \sum_{|s|=k , s\subseteq [ m ] } e_n(l_{s , s})z^{s_c}. \end{split}\ ] ] our aim is to prove stability of .
to that end , we introduce the determinantal polynomial @xmath202 which can be expanded as @xmath271}\det(w_s+l_s)z^{s_c } = \sum_{s\subseteq [ m ] } \left({\sum\nolimits}_{t\subseteq s}w^{s\setminus t}\det(l_{t , t})\right)z^{s_c}.\ ] ] this is an sr polynomial in @xmath205 variables , indexed by @xmath6 and @xmath206 , where @xmath207 .
it has support @xmath209 and @xmath210 . in other words
, we have ( using iverson - brackets @xmath211 ) @xmath272}{\llbracket s_c\cap r=\emptyset \wedge s_c\cap t=\emptyset \rrbracket}\det(l_{t , t})z^{s_c}w^{r}.\ ] ] now we truncate this sr polynomial at @xmath213 by restricting @xmath214 . by (
* corollary 4.18 ) this truncation preserves the sr property , whence @xmath215 \\ |s_c\cup
r|=m - n}}{\llbracket s_c\cap r=\emptyset \rrbracket}\det(l_{s\backslash r , s\backslash r})z^{s_c}w^{r}\ ] ] is also stable . using prop .
[ prop : basic]-(iii ) , setting @xmath216 retains stability , yielding @xmath217 \\ |s_c\cup
r|=m - n } } { \llbracket s_c\cap r=\emptyset \rrbracket}\det(l_{s\backslash r , s\backslash r})z^{s_c}y^{|r|}\\ & = \sum_{s\subseteq [ m]}\bigl({\sum\nolimits}_{|t|=n , t\subseteq s}\det(l_{t , t})\bigr)y^{|s|-|t|}z^{s_c } = \sum_{s \subseteq [ m ] } e_n(l_{s , s})y^{|s|-n}z^{s_c},\end{aligned}\ ] ] where the second equality follows since @xmath214 and @xmath273 imply that @xmath274 .
differentiating @xmath218 , @xmath219 times with respect to @xmath220 and evaluating at @xmath221 preserves stability ( prop .
[ prop : basic]-(ii ) and ( i ) ) . in doing so , only terms corresponding to @xmath222
survive , resulting in @xmath223}e_n(l_{s , s})z^{s_c } = ( k - n)!\sum_{|s|=k , s\subseteq[m]}\det(a_s a_s^\top ) z^{s_c},\ ] ] which is just ( [ eq:1 ] ) ( up to a constant ) .
thus , we have shown that @xmath225 is stable .
since it is multiaffine , it is actually an sr polynomial , and its corresponding measure @xmath177 is sr .
[ [ implementation - notes . ] ] implementation notes . + + + + + + + + + + + + + + + + + + + + + to implement algorithm [ algo : mcmc ] we need to compute the transition probabilities @xmath243 .
let @xmath244 and assume @xmath245 .
[ eq:4 ] we have the acceptance ratio @xmath246 thus , the transition probabilities can be computed in @xmath247 time .
moreover , one can further accelerate the algorithm by using the techniques developed in @xcite to compute lower and upper bounds on this quantity to determine early acceptance or rejection of the proposed move .
a remaining question is initialization .
since the mixing time is related to @xmath248 , we need to start with @xmath236 such that @xmath249 is bounded away from @xmath221 . here again the @xmath43-trick proves useful .
set @xmath275 , whereby @xmath276^\top\bigr).\end{aligned}\ ] ] the rhs is a distribution induced by volume sampling .
greedily choosing columns of @xmath3 one by one gives a @xmath277 approximation to the maximum volume submatrix @xcite .
this results in a set @xmath6 such that @xmath278 thus , @xmath250 , and the resulting running time for algorithm [ algo : mcmc ] is @xmath251 , which is _ linear _ in the size of data set @xmath23 and is efficient when @xmath0 is not large . in practice
it is hard to set @xmath43 to be exactly @xmath279 , but a small value will approximately suffice .
we report selection performance of dvs on real regression data ( compact(s ) , abalone and bank32nh ] ) for experimental design .
we use 4,000 samples from each dataset for estimation .
we compare against various baselines , including uniform sampling ( ` unif ` ) , leverage score sampling ( ` lev ` ) @xcite , predictive length sampling ( ` pl ` ) @xcite , the sampling ( ` smpl`)/greedy ( ` greedy ` ) selection methods in @xcite and fedorov s exchange algorithm @xcite .
we run the mcmc sampler for dvs for 10,000 iterations , which empirically yields selections that are sufficiently good .
we measure performances via ( 1 ) the prediction error @xmath280 , and 2 ) running times .
figure [ fig : cpu ] shows the results for these three measures with sample sizes @xmath0 varying from @xmath281 to @xmath282 .
further experiments ( including for the interpolation @xmath283 ) , may be found in the appendix . in terms of prediction error
, dvs performs well and is comparable with ` lev ` .
its strength compared to the greedy and relaxation methods ( ` smpl ` , ` greedy ` , ` fedorov ` ) is running time , leading to good time - error tradeoffs .
these tradeoffs are illustrated in figure [ fig : cpu ] for @xmath284 .
in other experiments ( shown in appendix [ app : sec : exp ] ) we observed that in some cases , the optimization and greedy methods ( ` smpl ` , ` greedy ` , ` fedorov ` ) yield better results than sampling , however with much higher running times . hence , given time - error tradeoffs , dvs may be an interesting alternative in situations where time is a very limited resource and results are needed quickly .
we show full results on compact(s ) , abalone and bank32nh datasets in figure [ app : fig : cpu ] , [ app : fig : abalone ] and [ app : fig : bank ] respectively .
we also run dvs- * , which is @xmath375-generalized dvs algorithm .
we observe that decreasing @xmath255 sometimes helps but sometimes not . in figure
[ app : fig : bank ] we observe that optimization- or greedy - based methods , while taking a huge amount of time to run , perform better than all sampling - based methods , thus for these selection methods , one is not always superior than another .
in this paper , we study the problem of dvs and develop an exact ( randomized ) polynomial time sampling algorithm as well as its derandomization .
thereafter , we study dual volume sampling via the theory of real - stable polynomials and prove that its distribution satisfies the `` strong rayleigh '' property .
this result has remarkable consequences , especially because it implies a provably fast - mixing markov chain sampler that makes dual volume sampling much more attractive to practitioners .
finally , we observe connections to classical , computationally more expensive experimental design methods ( fedorov s method and sa ) ; together with our results here , these could be a first step towards a better theoretical understanding of those methods . in this paper
, we develop polynomial time algorithms for dual volume sampling , a promising method for column subset selection problems .
we propose the first polynomial time exact sampling algorithm and its derandomized alternative , thereby closing two open problems raised in @xcite .
further , we show that a close variant of the dual volume distribution under consideration is strongly rayleigh . as a result
we immediately obtain a fast mixing markov chain for approximate dual volume sampling .
our algorithms thus enhance the applicability of dual volume sampling in its downstream applications like column subset selection and experimental design .
j. borcea and p. brndn .
applications of stable polynomials to mixed determinants : johnson s conjectures , unimodality , and symmetrized fischer products .
_ duke mathematical journal _ , pages 205223 , 2008 .
a. magen and a. zouzias .
near optimal dimensionality reductions that preserve volumes . in _
approximation , randomization and combinatorial optimization .
algorithms and techniques _ , pages 523534 .
springer , 2008 .
the partition function of @xmath82 , happens to have a pleasant closed - form formula . although this formula is known @xcite , and follows immediately by an application of the cauchy - binet identity , we present an alternative proof based on the perturbation argument for its conceptual value and subsequent use . first note that for @xmath57 and any @xmath66 , by we have @xmath85 taking limits as @xmath86 on both sides we have @xmath87 let us focus on @xmath88 .
we construct an identity matrix @xmath89 , then we have @xmath90}\right ) .
\end{split}\ ] ] in other words , this value is proportional to the probability of sampling columns from @xmath91}$ ] using volume sampling .
therefore , using the definition of @xmath92 we have @xmath93}&\det(a_s^\top a_s + { \varepsilon}i_k ) = { 1\over { \varepsilon}^{k - n } } e_k ( a^\top a + { \varepsilon}i_m)\\ & = { 1\over { \varepsilon}^{k - n } } e_k ( \operatorname{diag}([(\sigma_1 ^ 2(a ) + { \varepsilon } ) , ( \sigma_2 ^ 2(a ) + { \varepsilon}),\ldots , ( \sigma_n^2(a ) + { \varepsilon } ) , { \varepsilon},\ldots,{\varepsilon}]))\\ & = \binom{m - n}{k - n}\prod\nolimits_{i=1}^n ( \sigma_i^2(a ) + { \varepsilon } ) + o({\varepsilon}).\end{aligned}\ ] ] now taking the limit as @xmath86 we obtain @xmath84}\det(a_s a_s^\top ) & = \lim_{{\varepsilon}\to 0}\binom{m - n}{k - n}\prod\nolimits_{i=1}^n ( \sigma_i^2(a ) + { \varepsilon } ) + o({\varepsilon } ) = \binom{m - n}{k - n}\det(a a^\top).\end{aligned}\ ] ]
the marginal probability of a set @xmath60 $ ] for dual volume sampling is @xmath285 theorem [ thm : partition ] shows how to compute the denominator , thus our main effort is devoted to the nominator .
we have @xmath286 using the @xmath43-trick we have @xmath287 by decomposing @xmath288 we have @xmath289 now we let @xmath95 be the singular value decomposition of @xmath68 where @xmath96 , @xmath290 and @xmath291 . plugging the decomposition in the equation we obtain @xmath292 q_t^\top a_r\\ & = a_r^\top q_t q_t^\top a_r - { \varepsilon}a_r^\top q_t \left[\begin{array}{cccc } { 1\over \sigma_1 ^ 2(a_t ) + { \varepsilon } } & 0 & \ldots & 0\\ 0 & { 1\over \sigma_2 ^ 2(a_t ) + { \varepsilon } } & \ldots & 0\\ \vdots & \vdots & \ddots & \vdots\\ 0 & 0 & \ldots & { 1\over \sigma_{r(a_t)}^2(a_t)+ { \varepsilon } } \end{array}\right ] q_t^\top a_r.\end{aligned}\ ] ]
thus it follows that @xmath293 q_t^\top a_r + { \varepsilon}i_{|r|}\\ & = b_r^\top b_r + { \varepsilon}c_r^\top c_r + { \varepsilon}i_{|r|},\end{aligned}\ ] ] where @xmath294 is the projection of columns of @xmath295 on the orthogonal space of columns of @xmath68 .
let @xmath97 be the complement column space of @xmath98 , then we have @xmath296 .
moreover , @xmath297
q_t^\top a_r \in\mathbb{r}^{r(a_t)\times |r|}.\end{aligned}\ ] ] we further let @xmath298 and @xmath299 q_t^\top a_{t_c } \in\mathbb{r}^{r(a_t)\times ( m-|t|)}\end{aligned}\ ] ] where @xmath61\backslash t$ ] .
then we have @xmath300}{\left[\begin{array}{c}b_{t_c}\\ \sqrt{{\varepsilon}}u_{t_c } \\
\sqrt{{\varepsilon}}c_{t_c}\end{array}\right]}^\top\right)\end{aligned}\ ] ] where we construct an orthonormal matrix @xmath301 whose columns are basis vectors
. since we are free to chose any orthonormal @xmath302 , we simply let it be @xmath303 .
let @xmath103 $ ] , we have @xmath304}{\left[\begin{array}{c}b_{t_c}\\ \sqrt{{\varepsilon}}u_{t_c }
\\ \sqrt{{\varepsilon}}c_{t_c}\end{array}\right]}^\top\right ) & = \left({\left[\begin{array}{c}b_{t_c}\\ \sqrt{{\varepsilon}}w_{t_c } \end{array}\right]}{\left[\begin{array}{c}b_{t_c}\\ \sqrt{{\varepsilon}}w_{t_c } \end{array}\right]}^\top\right ) \\ & = f_{t_c } \in\mathbb{r}^{(m+n-|t|)\times ( m+n-|t|)}\end{aligned}\ ] ] the properties of characteristic polynomials imply that @xmath305 where @xmath306 $ ] and @xmath307\backslash s_1 $ ] .
further we have @xmath308 hence it follows that @xmath309 \times \\&\sum_{|s| = k-|t|}{\varepsilon}^{k-|t|-|s_1|}\det((b_{t_c})_{s_1}(b_{t_c})_{s_1}^\top ) \times\\ & \quad\quad\quad\quad \det((w_{t_c})_{s_2 } ( w_{t_c})_{s_2}^\top - ( w_{t_c})_{s_2 } ( b_{t_c})_{s_1}^\top((b_{t_c})_{s_1}(b_{t_c})_{s_1}^\top)^{-1}(b_{t_c})_{s_1 } ( w_{t_c})_{s_2}^\top ) \end{aligned}\ ] ] ( since @xmath310 and @xmath311 ) @xmath312 \times \\&\sum_{|s| = k-|t|}{\varepsilon}^{k-|t|-r(b_{t_c})}\det(b_{t_c}b_{t_c}^\top ) \det((w_{t_c})_{s_2 } ( w_{t_c})_{s_2}^\top - ( w_{t_c})_{s_2 } b_{t_c}^\top(b_{t_c}b_{t_c}^\top)^{-1}b_{t_c } ( w_{t_c})_{s_2}^\top ) + o({\varepsilon})\\ & = \left[\prod_{i=1}^{r(a_t)}\sigma_i^2(a_t)\right ] \times \left[\prod_{j=1}^{r(b_{t_c})}\sigma_j^2(b_{t_c})\right ] \sum_{s_2}\det((w_{t_c})_{s_2 } ( w_{t_c})_{s_2}^\top - ( w_{t_c})_{s_2 } b_{t_c}^\top(b_{t_c}b_{t_c}^\top)^{-1}b_{t_c } ( w_{t_c})_{s_2}^\top ) \end{aligned}\ ] ] where @xmath313\backslash [ r(b_{t_c})]$ ] and @xmath314 .
let @xmath315 be the eigenvalue decomposition of @xmath101 where @xmath102 .
further , let @xmath316 be the complement column space of @xmath317 , thus we have @xmath318 \left[\begin{array}{c c } q_{b_{t_c } } & q_{b_{t_c}}^\perp \end{array}\right ] = i_{|t_c| } = i_{n-|t|}\end{aligned}\ ] ] then for any @xmath313\backslash [ r(b_{t_c})]$ ] we have @xmath319 it follows that @xmath320 combining all the above derivations , we obtain that @xmath321 \times \left[\prod_{j=1}^{r(b_{t_c})}\sigma_j^2(b_{t_c})\right ] \times \gamma_t \over \left(\begin{array}{c}n - m\\k - m\end{array}\right ) \det(a a^\top)}.\end{aligned}\ ] ]
it may happen in practice that @xmath322 but @xmath0 is of the same order as @xmath4 .
in such case we can transform the dual volume sampling to slightly distorted volume sampling based on and then take the advantage of determinant - preserving projections to accelerate the sampling procedure . concretely , instead of sampling column subset @xmath6 with probability proportional to @xmath323
, we sample with probability proportional to a distorted value @xmath324 for small @xmath66 . denoting this distorted distribution as @xmath325 ,
we have @xmath326 letting @xmath327 be the minimum singular value , we have @xmath328 we further let @xmath329 when @xmath43 sufficiently small . sampling from @xmath330 will yield @xmath331-approximate dual volume sampling ( in the sense of @xcite and our theorem [ thm : approx ] ) .
we can sample from @xmath330 via _ volume sampling _ with distribution @xmath332})$ ] . with the volume sampling algorithm proposed in @xcite ,
the resulting running time would be @xmath333 .
[ thm : projection ] for any @xmath334 , @xmath335 and @xmath336 , the random gaussian projection of @xmath337 where @xmath338 satisfies @xmath339 for all @xmath340 $ ] and @xmath341 where @xmath342 is the projected matrix .
( corollary [ thm : approx ] ) the idea is to project @xmath91}$ ] to a lower - dimensional space in a way that the values for submatrix determinants are preserved up to a small multiplicative factor .
then we perform volume sampling .
we project columns of @xmath91}$ ] , which is in @xmath346 , to vectors in @xmath347 where @xmath348 so as to achieve a @xmath349 approximation by thm .
[ thm : projection ] .
let @xmath22 be a @xmath350-dimensional i.i.d .
gaussian random matrix , then we have @xmath351 } = g_a a + \sqrt{{\varepsilon}}g_a'\end{aligned}\ ] ] where @xmath352 and @xmath353 are two independent gaussian random matrix .
the projected matrix can be computed in @xmath354 time .
after that , if we use volume sampling algorithm proposed in @xcite the resulting running time would be @xmath355 .
thus the total running time would be @xmath345 .
an interesting observation is that the resulting running time is independent of @xmath356 , which means one can set @xmath43 arbitrarily small so as to make the approximation in the first step as accurate as possible , without affecting the running time .
however , in practice , a very small @xmath43 can result in numerical problems .
in addition , the dimensionality reduction is only efficient if @xmath357 .
we use @xmath358 denote the matrix @xmath359\backslash\{j\},:}$ ] , namely matrix @xmath3 with row @xmath113 deleted .
we have @xmath360\\ & = \sum_{(i_t,\ldots , i_k)\in[m]^{k - t+1 } } \|a_s^\dagger\|_f^2 \overrightarrow{p}(s_1=i_1,\ldots , s_k = i_k ; a \mid s_1=i_1,\ldots , s_{t-1}=i_{t-1})\\ & = \sum_{(i_t,\ldots , i_k)\in[m]^{k - t+1 } } \|a_s^\dagger\|_f^2 { \overrightarrow{p}(s_1=i_1,\ldots , s_k = i_k ; a ) \over \overrightarrow{p}(s_1=i_1,\ldots , s_{t-1}=i_{t-1 } ; a)}\\ & = { \sum_{(i_t,\ldots , i_k)\in[m]^{k - t+1}}\det(a_{\{i_1,\ldots , i_k\ } } a_{\{i_1,\ldots , i_k\}}^\top ) \|a_{\{i_1,\ldots , i_k\}}^\dagger\|_f^2 \over \sum_{(i_t,\ldots , i_k)\in[m]^{k - t+1}}\det(a_{\{i_1,\ldots , i_k\ } } a_{\{i_1,\ldots , i_k\}}^\top ) } \\ & = { \sum_{j=1}^n\sum_{(i_t,\ldots , i_k)\in[m]^{k - t+1 } } \det(a_{\{i_1,\ldots , i_k\}}^j ( a_{\{i_1,\ldots , i_k\}}^j)^\top ) \over \sum_{(i_t,\ldots , i_k)\in[m]^{k - t+1}}\det(a_{\{i_1,\ldots , i_k\ } } a_{\{i_1,\ldots , i_k\}}^\top ) } \\\end{aligned}\ ] ] while the denominator is the ( unnormalized ) marginal distribution @xmath361 , the numerator is the summation of ( unnormalized ) marginal distribution @xmath362 for @xmath363 . by theorem [ thm : marginal ]
we can compute this expectation in @xmath152 time .
* input : * matrix @xmath32 to sample columns from , @xmath364 the target size * output : * set @xmath6 such that @xmath128 with the guarantee @xmath365 initialize @xmath112 as empty tuple compute conditional expectation @xmath366 $ ] with corollary [ cor : conditional ] .
choose @xmath367 @xmath133 output @xmath112 as a set @xmath6 observe that at each iteration @xmath235 , we have @xmath368\\ & = { \sum\nolimits}_{j\notin \overrightarrow{s}}\overrightarrow{p}(t_i = j | t_1=s_1,\ldots , t_{i-1}=s_{i-1})\mathbb{e}\left[\|a_t^\dagger\|_f^2 \mid t_1 = s_1,\ldots , t_{i-1}=s_{i-1 } , t_i = j\right],\end{aligned}\ ] ] and we choose @xmath113 such that @xmath369 $ ] is minimized . since at the beginning we have @xmath370 \le { m - n + 1\over k - n+1}\|a^\dagger\|_f^2 ; \quad t\sim p(t;a),\end{aligned}\ ] ] it follows that the conditional expectation satisfies @xmath371 \le { m - n + 1\over k - n+1}\|a^\dagger\|_f^2.\end{aligned}\ ] ] hence we have @xmath372 \le { m - n + 1\over k - n+1}\|a^\dagger\|_f^2.\end{aligned}\ ] ] further , by using standard bounds relating the operator norm to the frobenius norm , we obtain @xmath373
set @xmath275 , whereby @xmath374^\top\bigr).\end{aligned}\ ] ] the rhs is a distribution induced by volume sampling .
greedily choosing columns of @xmath3 one by one gives a @xmath277 approximation to the maximum volume submatrix @xcite .
this results in a set @xmath6 such that @xmath278 thus , @xmath250 .
note that in practice it is hard to set @xmath43 to be exactly @xmath279 , but a small approximate value suffices . | we study _ dual volume sampling _ , a method for selecting @xmath0 columns from an @xmath1 short and wide matrix ( @xmath2 ) such that the probability of selection is proportional to the _ volume _ spanned by the rows of the induced submatrix .
this method was proposed by @xcite ( @xcite ) , who showed it to be a promising method for column subset selection and its multiple applications .
however , its wider adoption has been hampered by the lack of polynomial time sampling algorithms .
we remove this hindrance by developing an exact ( randomized ) polynomial time sampling algorithm as well as its derandomization .
thereafter , we study dual volume sampling via the theory of real - stable polynomials and prove that its distribution satisfies the `` strong rayleigh '' property .
this result has remarkable consequences , especially because it implies a provably fast - mixing markov chain sampler that makes dual volume sampling much more attractive to practitioners .
this sampler is closely related to classical algorithms for popular experimental design methods that are to date lacking theoretical analysis but are known to empirically work well . |
physical properties of galaxies can often be derived from spectroscopic observations , which allow the detection and characterisation of various emission lines .
however , the complexities related with observation and reduction procedures , make spectroscopic surveys considerably more difficult to pursue than imaging surveys using narrow - band filters .
while such imaging surveys detect only lines whose redshifted wavelength fall inside the filter s band , they have the advantage of detecting spatially extended line emission down to low surface brightness levels .
this imaging technique is , therefore , useful for evaluating the full spatial extent and distribution of the line - emission in nearby galaxies ( kennicutt et al .
2008 ) , as well as to search and identify high - redshift line - emission galaxies in studies where many objects are targeted ( e.g. , thompson , mannucci & beckwith 1996 ; nilsson et al .
2007 ) .
the h@xmath0 line provides a large part of our knowledge of star formation ( sf ) activity in galaxies and is the focus of this paper . observing a galaxy with a filter centred at rest - frame h@xmath0 can
, with reasonable assumptions , provide information on the amount and spatial distribution of the emitted ionising radiation using relatively modest integration times on small telescopes , thus making large galaxy surveys practical .
since the flux measured by narrow - band imaging includes both the spectral line and the continuum emission , it is necessary to measure the continuum contribution . in many cases this
is done by imaging the target in a band significantly wider than the narrow band ( e.g. , waller 1990 ; james et al .
2004 ; kennicutt et al .
2008 ; brosch , spector & zitrin 2011 ) . a scaling factor between the continuum counts of the wide and narrow band images ( wide to narrow continuum ratio - wncr ) is then determined . from its value
the line s counts can be extracted .
this paper discusses the derivation of emission line properties from wide and narrow band imaging , focusing on sources of measurement errors . in section [ sec : derivations_ew ] we derive the relations required to calculate the equivalent width ( ew ) and the flux .
section [ sec : photometry ] discusses methods for obtaining a calibrated measure of the flux . in section
[ sec : err_propagation ] we derive the effect of the wncr estimation error on the measurement errors of the line s equivalent width and flux . in section
[ sec : wncr_measmethod ] we review some commonly used methods to measure wncr .
then , in section [ sec : coloureffects ] we analyse the measurement errors that originate from colour effects when using these methods , and quantify these errors for the case of h@xmath0 measurements in the local universe .
line emission from a galaxy or a region within a galaxy ( target ) can be measured by observing it through a narrow - band filter ( n ) and a wide - band filter ( w ) , where the n band contains the emission - line , and the w band may or may not contain it . in this section
we derive the method to calculate the emission - line s equivalent width ( ew ) and flux ( @xmath1 ) from such observations , and discuss the different assumptions made in this derivation and their implications .
the derivation and discussion are general , and can be applied to h@xmath0 measurements as well as to other emission lines measured in a similar way .
the response of the electro - optical system can be expressed as the integral of the observed flux density over wavelength , weighted by the transmittance functions of the filter and of the atmosphere and by the response function of the rest of the electro - optical system : @xmath2 where : @xmath3 , @xmath4 : : are the measured count rates of the narrow - band ( n ) and wide - band ( w ) filters ( respectively ) in instrumental units - typically adu s@xmath5 ( analog to digital units per second ) ; @xmath6 : : is the spectral flux density of light from the observed target reaching the telescope in absence of an atmosphere ( typically in erg s@xmath5 @xmath7 @xmath8 ) ; @xmath9 , @xmath10 : : are the unitless transmittance functions of the n and w bands , respectively , ranging from 0 ( fully opaque ) to 1 ( fully transparent ) ; @xmath11 , @xmath12 : : are the atmospheric transmittance as function of wavelength , including effects of weather , elevation and airmass of observation , when the n and w bands ( respectively ) were imaged ( unitless , ranging from 0 to 14 ) ; @xmath13 : : is the responsivity as function of wavelength of the rest of the electro - optical system ( i.e. the telescope and sensors , excluding the transmittance effect of the filters ) typically in adu erg@xmath5 @xmath14 ; @xmath15 : : is the wavelength ( typically in @xmath16 ) . :
: this paper considers the case where the measured line is the only significant spectral feature in the n and w bands , i.e. that the target s flux can be divided into line and continuum ( cont ) emission . for the case of h@xmath0
, this means that some adjustments should be made , if the [ nii ] lines flanking it carry significant flux .
the count rate from the observed target can be expressed as the sum of the line and continuum contributions : @xmath17 + @xmath18 where : @xmath19 , @xmath20 : : are respectively the line and continuum contributions to @xmath3 ; @xmath21 , @xmath22 : : are respectively the line and continuum contributions to @xmath4 ; @xmath23 , @xmath24 : : are respectively the line and continuum contributions to @xmath6 .
: : since the width of the emission - line is typically very narrow in comparison to both the n and w bands , its profile can be approximated as a delta function at the centre of the emission line : @xmath25 where : @xmath26 : : is the central wavelength of the emission line ( typically in @xmath16 ) ; @xmath27 : : - the flux ( irradiance ) of the emission line ( typically in erg s@xmath5 @xmath14 ) .
: : using this approximation , the line contribution is simply : @xmath28 the goal of the measurement is to find the emission - line s flux , @xmath1 , and equivalent width , ew .
the line s equivalent width is defined as : @xmath29 while , the flux can be extracted from : @xmath30 the term @xmath31 is obtained by photometric calibrations and is discussed in section [ sec : photometry ] .
the n - band s transmittance , @xmath32 , measured directly or provided by the filter manufacturer , should be corrected to compensate for effects of the telescope s converging beam , since the transmittance curves of interference filters ( the type of filters usually used ) vary with angle of incidence .
we now derive the remaining terms : @xmath19 and @xmath33 . using the two parts of
, we can write @xmath21 in terms of @xmath19 ( based on the @xmath34 function line approximation ) : @xmath35 using this and we can express @xmath22 as : @xmath36 for the derivation of @xmath19 we define the wide to narrow continuum ratio ( wncr ) to be the ratio between the count rate contributed by the continuum in the w band and the count rate contributed by the continuum in the n band : @xmath37 in sections [ sec : err_propagation ] , [ sec : wncr_measmethod ] and [ sec : coloureffects ] we describe the methods most commonly used to measure wncr , and analyse their effect on the accuracy of ew and @xmath1 measurements . from , and we obtain an expression for @xmath19 : @xmath38 extracting @xmath19 yields : @xmath39^{-1 } \end{ieeeeqnarraybox * } \label{cps_line_n_solved}\ ] ] note that this expression includes the atmospheric transmittance at @xmath26 for both bands : @xmath40 and @xmath41 .
these can be estimated from a photometric calibration in the n band , when available .
if the w and n exposures were taken at very similar airmasses and atmospheric conditions , these can be approximated as equal @xmath42 .
typically , the n band is chosen so that @xmath32 is as high as possible ( @xmath26 is close to the n band s peak ) , and is not significantly smaller than @xmath43 . the w band is typically chosen to be significantly wider than the n band , therefore @xmath44 , and also @xmath45 .
this means that the error due to small deviations from the above approximation will typically be minor
. it will decrease the closer @xmath26 is to the n band s peak and the wider the w band is compared to the n band . in order to derive @xmath33 we approximate @xmath24 to be constant within the n band .
this is valid if the n filter is very narrow , and has negligible `` out of band leaks '' ( a requirement which is not trivial for very narrow bands ) .
if @xmath24 is constant within the n band , it can be taken out of the integral for @xmath20 in and be replaced by @xmath33 : @xmath46 using , we get : @xmath47 replacing @xmath33 and @xmath1 in with the expressions from , and using equation yields : @xmath48 if the n band is sufficiently narrow , @xmath13 and @xmath49 can be assumed constant within it .
thus : @xmath50 the equivalent width , ew , can therefore be obtained from and without requiring photometric calibrations .
it is equal to the ratio between the count rates of the line ( @xmath19 ) and the continuum ( @xmath51 ) multiplied by the `` n - band effective width '' @xmath52 .
the `` n - band effective width '' is a measure of the n - band width multiplied by the ratio between the n - band peak transmittance and the n - band transmittance at the emission line s wavelength , @xmath32 .
the ew measurement error contributed by atmospheric effects originates only from @xmath19 . in order to understand how changes in the atmospheric transmittance ( e.g. , clouds ) affect the accuracy of the result
, we write with wncr in the last term expressed through fundamental variables , as defined in and : @xmath53^{-1 } \end{ieeeeqnarraybox*}\ ] ] it can be seen here and in that atmospheric conditions affect ew only through the relative transmissions : @xmath54 and @xmath55 .
`` grey '' wavelength - independent effects do not change these terms , and therefore will not affect the result .
the effect of clouds blocking the field during parts of the exposure will be the same as reducing @xmath11 and/or @xmath12 by an unknown factor .
this effect is virtually independent of wavelength , and therefore does not contribute to the measurement error in ew , whether the w and n exposures were taken in similar airmass and atmospheric conditions or not .
unlike ew , measuring the line s flux , @xmath1 , requires a photometric calibration .
@xmath1 can be obtained from , if @xmath31 is known . a useful method to measure utilises photometric calibration in the n band .
spectrophotometric standard stars ( i.e. stars with known @xmath6 ) should be measured at several airmass ( am ) values .
the measured count rate of the standard stars can be expressed as : @xmath56 where : @xmath57 : : is the count rate of the spectrophotometric standard star imaged through the narrow - band ( n ) filter as function of the airmass ( typically in adu s@xmath5 ) ; @xmath58 : : is the spectral flux density of the spectrophotometric standard star ; @xmath59 : : is the atmospheric transmittance during the photometric calibration as function of wavelength and airmass .
: : if the n band is sufficiently narrow , @xmath13 and @xmath60 can be approximated as constant within the n band .
this yields : @xmath61 from measurements at different airmasses , a fit can be made to the following logarithmic relation : @xmath62 \cong \mbox{am } \cdot \mbox{e}_{\lambda_{line } } + \mbox{m}_0 \label{tatm_fit}\ ] ] where : @xmath63 : : is the extinction coefficient of the atmosphere at @xmath26 ; @xmath64 : : is the zero point of the fit .
: : if the n band image is taken at similar atmospheric conditions as during the calibration , and with am for which the fit still holds , @xmath31 can be simply obtained by : @xmath65 where : @xmath66 : : is the airmass through which the n band image was taken .
: : when photometric calibrations of the n band can not be obtained , one can calibrate the emission - line s flux based on wide - band calibrated images .
the target has to be measured through several calibrated wide bands , to obtain its flux per wavelength , @xmath6 , averaged over each band . from these averaged @xmath6 values a model for the target s continuum flux as function of wavelength , @xmath24
, can be derived .
using with the modelled @xmath33 , and the measured ew , the flux can be simply obtained : @xmath67 when available , wide band magnitudes from the sloan digital sky survey ( sdss - abazajian et al .
2009 ) can be used to model @xmath24 . in some cases a linear fit of the @xmath68 , @xmath69 and @xmath70 magnitudes ( whose average wavelengths are 4686@xmath16 , 6165@xmath16 , and 7481@xmath16 respectively ) is a sufficient approximation for the continuum at h@xmath0 .
the error of this linear approximation can be estimated using the residuals of the fit .
since typically the emission - line s wavelength is within the w band , one might be tempted to approximate @xmath33 to be the average continuum flux measured for w ( when a photometric calibration for band w is available ) .
such an approximation is equivalent to a model in which @xmath24 is constant within the w band , and is therefore highly inaccurate as evident from figure [ f : cont ] .
figure [ f : cont ] shows three examples of normalized @xmath24 curves ( @xmath71myr , @xmath72myr and @xmath73gyr old stellar populations ) . the nave model of constant @xmath24 over the r band is also shown ( dotted horizontal line ) .
using a photometric calibration for w alone ( r in this example ) implies that the dashed line is used instead of the real value . for example , calibrating with r filter photometry alone overestimates @xmath74 of the @xmath71myr old population at @xmath75=0 by 13% .
this systematic calibration error depends on the age of the stellar population as well as on its redshift , and propagates directly through to a systematic measurement error in @xmath1 . of @xmath71 , @xmath72myr and @xmath73gyr old ssps at z=0 , normalized so that the average @xmath24 over the r - band of each stellar population would be 1 .
data adopted from the bruzual - charlot atlas available in ftp://ftp.stsci.edu/cdbs/grid/bc95/templates/ ( also used for the calibration of several _ hubble space telescope _
instruments ) .
the dotted horizontal line shows the model adopted , when calibrating using the r - band alone ( @xmath24 constant over the r band ) .
the dashed vertical line indicates the h@xmath0 wavelength ( @xmath76 ) .
[ f : cont],width=302 ]
in this section we derive the effect of the wncr estimation error ( @xmath77wncr ) on the measurement errors of the equivalent width and flux of the line ( @xmath77ew and @xmath78 ) . using we derive @xmath79 , the measurement error of @xmath19 propagated from @xmath77wncr ( i.e. , @xmath80 if all other sources of error are negligible ) . for reasons discussed above , the right - hand side factor of
typically introduces a small correction to the left - hand side factor . for error propagation purposes this correction can be ignored .
hence : @xmath81 for the purpose of error analysis , we assume that the line s contribution to @xmath4 is minor ( i.e. @xmath82 ) .
this is true in typical conditions , where the w - band width is significantly larger than the ew . using with this assumption
we obtain : @xmath83 from this and , @xmath84 , the measurement error in @xmath1 propagated from @xmath77wncr , can be simply derived as : @xmath85 using and the assumption made in yields : @xmath86 @xmath84 is therefore affected by the relative error in wncr , the n - band effective width , and the underlying continuum flux .
note that @xmath84 does not depend on the line s flux itself ( @xmath1 ) .
+ similarly , @xmath87 , the measurement error in ew propagated from @xmath77wncr , can be derived from , and :
@xmath88 the relative error , obtained by dividing this equation by , is : @xmath89 since the last term in is at least unity , it is evident that the relative error of wncr multiplied by the n - band effective width is a lower limit of the uncertainty in ew : @xmath90 similarly , it is evident from that the relative error of wncr sets a lower limit to the relative error of ew : @xmath91
for well - resolved galaxies , the continuum light can be directly measured from the galaxy itself .
this can be done using the pixel - to - pixel method ( bker et al .
1999 ; rossa & dettmar 2000 ; kennicutt et al . 2008 ; finkelman et al .
2010 ) where @xmath3 is plotted for each pixel as function of @xmath4 . to reduce the data scatter
one may first reject a user - specified percentage of the brightest pixels , suspected to be foreground stars .
galaxies with smoothly varying light distribution , such as elliptical galaxies , can also be fitted with isophotal maps in both the w - band and n - band images . from these
, one can plot the mean @xmath3 vs. mean @xmath4 for pairs of isophotes of the same semidiameter ( e.g. , macchetto et al .
1996 ) .
the resulting relation is expected to show most data points ( presumably lacking line emission ) aligning along a narrow straight line down to the noise level .
the slope of this line yields the wncr estimate
. a smaller number of data points , which trace the h@xmath0 emission , are expected to be above the line , whereas a small amount of scattered data points are also expected below the line .
the latter trace either regions of excess continuum emission , or excess line absorption . note that this wncr estimate is affected by widespread diffuse emission and by underlying stellar h@xmath0 absorption across the galaxy .
a hidden and problematic assumption in applying this method is that of a constant colour .
the fitted regions may show colour variations caused by changes in stellar population across the galaxy and in the line - emission regions .
these may be related to recent or ongoing sf episodes .
note that the derived wncr estimate , which is affected by colour , does not represent the emission line region but is rather an average value across the galaxy .
+ an alternative method to estimate wncr is by calculating the average @xmath4 to @xmath3 ratio of foreground stars ( e.g. , pogge , owen & atwood 1992 ; rand 1996 ; koopman , kenney & young 2001 ; james et al .
this method allows measuring line emission even if the continuum - dominated regions of the target can not be properly sampled .
the main assumption is that the foreground stars represent the mix of stars in the target galaxy , thus matching on average the spectral distribution of its continuum . in order to represent a continuous distribution function of stellar types , as many foreground stars as possible
should be measured .
however , regardless of how numerous , foreground stars represent the local population of the region in the milky way through which the observation is made , rather than the stellar population in the target galaxy .
these two populations may be significantly different .
this method is more useful for target galaxies that are sufficiently distant to allow the n - band to transmit their rest - frame h@xmath0 while blocking the local h@xmath0 ( 6563@xmath16 ) of milky way foreground stars .
if this is not the case , applying this method implies also the assumption that the fraction of foreground stars with strong balmer line absorption or emission ( typically o , b and a stars ) is insignificant .
this assumption typically holds for observations in mid to high galactic latitudes . at these latitudes galactic foreground stars between the observer and a target
tend to be of types later than a due to the age of the halo and the shape of the stellar initial mass function .
+ we note that an alternative set of bands may be used , in which an off - line narrow band replaces the wide w - band ( e.g. almoznino & brosch 1998 ; van zee 2000 ; dale et al .
if the central wavelengths of the off - line and on - line bands are close enough , the wncr uncertainty due to colour effects may be reduced .
however , this comes at the expense of increased random errors due to the lower depth of images obtained with limited telescope time .
an alternative approach utilises two medium to wide off - line bands flanking the h@xmath0 line ( e.g. mackenty et al . 2000 ; lin et al .
2003 ; calzetti et al . 2004 ; ly et al .
this approach reduces colour effects , while having a smaller impact on image depth .
colours of the studied emission - line regions may differ from those of the foreground stars or those of the continuum - dominated galactic regions used for estimating wncr . as a result ,
the estimated wncr may deviate from its value in the studied region .
furthermore , since colours typically vary within a galaxy , different parts of a target galaxy may have different wncr values . in an attempt to improve accuracy
, the value of wncr is sometimes adjusted subjectively to obtain a desirable continuum subtracted image ( rand 1996 ; james et al .
2004 ; kennicutt et al .
2008 ) .
although the uncertainties in estimating wncr are typically of few percent , these can result in a much higher uncertainty in the measured h@xmath0 flux , especially for low ew values ( macchetto et al .
1996 ; hameed & devereux 1999 ; fig . 7 in koopmann
et al . 2001 ) .
for example , for a @xmath92 error in wncr and an n - band effective width of 100@xmath16 , the relative error of ew would be at least 5% ( eq . [ rel_ew_err_limit ] ) , while the absolute error of ew would be at least 5@xmath16 ( eq . [ ew_err_limit ] ) . for low sf rates
the latter can be comparable to the measured ew . in order to estimate the variation of wncr with colour we model observations of different objects with recession velocities up to 30000 km s@xmath5 ( z=0.1 ) .
for each object we simulate observations through a johnson - cousins r filter and a narrow - band filter , 50@xmath16 wide and centred on the redshifted h@xmath0 line . to simulate foreground stars we use various spectra of main sequence ( ms ) and off - ms stars , taken from the bruzual - persson - gunn - stryker atlas ( available at ftp://ftp.stsci.edu/cdbs/grid/bpgs/ ) .
the continua of different parts of target galaxies are represented by redshifted spectra of single stellar population ( ssp ) models at different ages in the range @xmath93-@xmath94yr .
these star - bursts are characterised by a salpeter initial mass function ( imf ) with a lower and upper mass of 0.1 and 125m@xmath95 , respectively .
the data are taken from the bruzual - charlot atlas ( available in ftp://ftp.stsci.edu/cdbs/grid/bc95/templates/ ) . using the spectra of these objects , the filters transmittance curves , and the responsivity of a typical electro - optical system
, we calculate @xmath22 and @xmath20 and from it the theoretical wncr for each object ( equations [ cps_cont_line_detailed ] and [ wncr_def ] ) .
the calculated wncrs for each narrow - band filter are plotted vs. the computed sdss @xmath68-@xmath69 colour of each object ( figure [ f : wncrha ] ) . for each narrow - band filter wncr values
are normalized using the wncr of a theoretical object with constant @xmath96 ( zero colour in the ab magnitude system ) .
linear fits for foreground star points are also shown in each plot . [ cols="^,^ " , ] the error in wncr can be estimated for the different recession velocities of figure [ f : wncrha ] by comparing the wncr value of foreground stars at @xmath68-@xmath69=0.8 with the wncr values of the simulated target galaxy parts ( the redshifted stellar populations ) .
when , for example , measuring a starburst galaxy dominated by a young stellar population ( 10@xmath97yr ) with foreground stars of @xmath68-@xmath69=0.8 , wncr will be underestimated by up to @xmath9830% depending on the galaxy s redshift ( for cz @xmath99 km s@xmath5 ) .
this means underestimating the h@xmath0 ew by a similar fraction ( equation [ rel_ew_err_limit ] ) , and underestimating @xmath1 by @xmath980.3 of the continuum flux within the n - band ( equation [ f_line_err_solved ] ) . for older underlying stellar populations
the error in wncr would be smaller . for 10@xmath100yr the underestimation will reach 8% at 10000 km s@xmath5 , while for 10@xmath101yr wncr will be overestimated by up to @xmath9810% . estimating wncr using galaxy parts believed not to form stars may also be problematic , since it does not take into account the possible colour variations across the galaxy and the expected colour change due to star formation in the h@xmath0 emitting regions .
let us consider a case where a 10@xmath97yr old stellar population is observed within an elliptical galaxy dominated by a 10@xmath101yr old population .
the flux from the young population is expected to overwhelm the flux from the old population , even if it is only a small fraction of the entire star population .
therefore , the measured wncr will correspond to redder objects , i.e. would be underestimated by up to @xmath9840% depending on redshift ( for cz @xmath99 km s@xmath5 ) . these errors in wncr
propagate through @xmath1 to sf measurements .
since the errors depend on colour properties of the galaxies , a bias in the colour properties of a sample may result in a bias in the sf properties concluded by a survey .
this applies also to surveys that measure cosmic sf rate at low redshifts ( low z points in the madau plot ) .
since star forming galaxies tend to be bluer than foreground stars , results may be biased .
as discussed above , such a survey that uses foreground stars and no colour information may result in the underestimation of wncr by tens of percents .
such underestimation , which increases with redshift , will result in the underestimation of the cosmic sf rate .
we identified an important error in estimating extragalactic h@xmath0 emission , originating from commonly used procedures of continuum subtraction .
we showed that when measuring an emission line by observing through a narrow - band ( n ) and a wide - band filter ( w ) , where the n band contains the emission - line , and the w band may or may not contain it : 1 .
the equivalent width , ew , can be obtained without the need for photometric calibrations using and . to improve accuracy :
* the n filter should have a very narrow band and negligible `` out of band leaks '' ; * the exposures of the w and n bands must be taken in very similar atmospheric conditions and air - masses , or with measured atmospheric transmittance at the line s wavelength [ @xmath40 and @xmath41 ] .
an accurate measurement of the line flux , @xmath1 , can be obtained by photometric calibrations in the n - band , if it is sufficiently narrow to assume that the responsivity , @xmath13 , and the atmospheric transmittance , @xmath49 , are constant within it . alternatively ,
if calibrated measurements of the magnitude in several wide - bands are obtained , a fit can be used to estimate the continuum flux under the line , @xmath103 , from which @xmath1 can be obtained .
using calibrations in the w - band alone may lead to systematic errors of over 10% .
+ 3 . an accurate estimation of wncr , the w to n band ratio of count rates contributed by the continuum ( equation [ wncr_def ] ) , is crucial for accurate emission line measurements : * the uncertainty in ew is equal to or greater than the relative error in wncr multiplied by the n - band effective width ( equation [ ew_err_limit ] ) ; * the relative uncertainty in ew is equal to or greater than the relative error in wncr ( equation [ rel_ew_err_limit ] ) ; * the uncertainty in the emission - line flux , @xmath1 , is equal to or greater than the relative error in wncr multiplied by the underlying continuum flux and by the n - band effective width ( equation [ f_line_err_solved ] ) .
4 . the value of wncr may vary significantly with the colour of the measured object . for typical h@xmath0 measurements at redshifts cz @xmath99 km s@xmath5 the dependence on @xmath68-@xmath69 was found to be almost linear . where foreground milky way stars are used for estimating wncr
, colour effects may lead to an underestimation as great as 30% , or in other cases overestimation by as much as 10% .
where regions in the target galaxy itself are used for the estimation , wncr may be underestimated by as much as 40% .
we suggest estimating wncr as a linear function of target colour , rather than as a constant value .
this will reduce errors , and will make the measurement procedure more objective .
the errors described above can bias the results of a study of a class of galaxies by systematically underestimating or overestimating the emission - line flux .
cosmic sf rate measurements at low redshifts may also be biased by this effect .
if not corrected properly , such measurements may underestimate the low z points in the madau plot .
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/. 99 abazajian k. n. et al . , 2009 ,
apjs , 182 , 543 almoznino e. , brosch n. , 1998 , mnras , 298 , 920 bker t. et al . , 1999 ,
apjs , 124 , 95 brosch n. , spector o. , zitrin a. , 2011 , mnras , 415 , 431 calzetti d. , harris j. , gallagher j. s. iii , smith d. a. , conselice c. j. , homeier n. , kewley l. , 2004 , aj , 127 , 1405 .
dale d. a. , barlow r. j. , cohen s. a. , cook d. o. , johnson l. c. , kattner s. m. , moore c. a. , schuster m. d. , staudaher s. m. , 2010 , apj , 712 , 189 finkelman i. , brosch n. , funes j. g. , kniazev a. y. , visnen p. , 2010
, mnras , 407 , 2475 hameed s. , devereux n. , 1999 , aj , 118 , 730 james p.
a. et al . , 2004 , a&a , 414 , 23 kennicutt r. c. jr . , lee j. c. , funes j. g. , sakai s. , akiyama s. , 2008 , apjs , 178 , 247 koopmann r. a. , kenney j. d. p. , young j. , 2001 , apjs , 135 , 125 lin w. et al . , 2003 ,
aj , 126 , 1286 ly c. , lee j. c. , dale d. a. , momcheva i. , salim s. , staudaher s. , moore c. a. , finn r. , 2011 , apj , 726 , 109 macchetto f. , pastoriza m. , caon n. , sparks w. b. , giavalisco m. , bender r. , capaccioli m. , 1996 , a&as , 120 , 463 mackenty j. w. , maz - apellniz j. , pickens c. e. , norman c. a. , walborn n. r. , 2000 , aj , 120 , 3007 nilsson k. k. et al . , 2007 , a&a , 471 , 71 pogge r. w. , owen j. m. , atwood b. , 1992 , apj , 399 , 147 rand r. j. , 1996 , apj , 462 , 712 rossa j. , dettmar r. j. , 2000 , a&a , 359 , 433 thompson d. , mannucci f. , beckwith s. v. w. , 1996 , aj , 112 , 1794 waller w. h. , 1990 , pasp , 102 , 1217 van zee l. , 2000 , aj , 119 , 2757 | we point out an important source of error in measurements of extragalactic h@xmath0 emission and suggest ways to reduce it . the h@xmath0 line , used for estimating star formation rates , is commonly measured by imaging in a narrow band and a wide band , both which include the line .
the image analysis relies on the accurate removal of the underlying continuum .
we discuss in detail the derivation of the emission line s equivalent width and flux for extragalactic extended sources , and the required photometric calibrations .
we describe commonly used continuum - subtraction procedures , and discuss the uncertainties that they introduce .
specifically , we analyse errors introduced by colour effects .
we show that the errors in the measured h@xmath0 equivalent width induced by colour effects can lead to underestimates as large as 40% and overestimates as large as 10% , depending on the underlying galaxy s stellar population and the continuum - subtraction procedure used .
we also show that these errors may lead to biases in results of surveys , and to the underestimation of the cosmic star formation rate at low redshifts ( the low z points in the madau plot ) .
we suggest a method to significantly reduce these errors using a single colour measurement .
[ firstpage ] methods : observational hii regions galaxies : star formation . |
looking for signals of physics beyond the standard model ( sm ) is one of the most important missions of high energy physics .
complementary to direct searches for new physics ( np ) particles in the high energy colliders , the study of @xmath6 physics is of great importance for probing indirect signals of np . in this respect , the @xmath6 factories at slac and kek are doing a commendable job by providing us with a huge amount of data on various @xmath6-meson decays , which are mostly in perfect agreement with the sm predictions .
however , there still exist some unexplained puzzles , such as the unmatched @xmath7 asymmetries in @xmath8 decays @xcite , the abnormally large branching ratios of @xmath9 and @xmath10 decays @xcite , and the large transverse polarization fractions in @xmath11 decays @xcite . confronted with these anomalies , we are forced not only to consider more precise qcd effects , but also to speculate on the existence of possible np scenarios beyond the sm .
it is well - known that the flavor - changing neutral current ( fcnc ) processes arise only from loop effects within the sm , and are therefore very sensitive to various np effects .
since the puzzling @xmath6-meson decay processes mentioned above are all related to the fcnc @xmath12 transitions , these decay channels could be used as effective probes of possible np scenarios .
so , if one kind of np could resolve one of these puzzles , it is necessary to investigate whether the same scenario can also provide a simultaneous resolution to the others .
considering the fact that current theoretical estimations of @xmath13 still suffer from large uncertainties @xcite , and the np scenario with anomalous tensor operators considered in this paper do not contribute to the @xmath14 and @xmath15 decays in the naive factorization ( nf ) approximation , we shall only focus on the @xmath16 and @xmath17 decays .
the recent experimental data on the longitudinal polarization fraction @xmath18 in @xmath19 decay is given as @xmath20 on the other hand , since the two final - state light vector mesons @xmath21 and @xmath22 in this decay mode are flying out fleetly in the rest frame of @xmath6 meson , and the structure of the charged weak interaction current of the sm is left - handed , as well as the fact that high - energy qcd interactions conserve helicity , any spin flip of a fast flying quark will be suppressed by one power of @xmath23 , with @xmath24 the @xmath25 quark mass .
it is therefore expected that , within the sm , both of the final - state hadrons in this decay mode are mainly longitudinally polarized , with @xmath26 while the transverse parts are suppressed by powers of @xmath27 .
obviously , the experimental data eq .
( [ fl - data ] ) deviates significantly from the sm prediction eq .
( [ predifl ] ) , and this polarization anomaly has attracted much interest in searching for possible theoretical explanations both within the sm and in various np models @xcite .
for example , the authors in refs .
@xcite have studied this anomaly and found that the four - quark tensor operators of the form @xmath28 could offer a simple resolution to the observed polarization anomaly within some possible parameter spaces .
since the @xmath29 decays , in analogy with the @xmath11 decays , also involve the @xmath30 transition , it is necessary to investigate the effects of these new types of four - quark tensor operators on the latter . in particular , it is very interesting to see whether these new four - quark tensor operators with the same parameter spaces could also simultaneously account for the measured @xmath31 @xcite , which are much larger than the theoretical predictions within the sm @xcite , and survive bounds from @xmath3 decays .
motivated by these speculations , in this paper , we shall investigate the effects of the following two types of tensor operators on these decay modes ( with @xmath32 the color indices ) @xmath33 and try to find out the allowed parameter spaces characterized by the strengths and phases of these new tensor operators that satisfy all the experimental constraints from these decays .
moreover , since the ( pseudo-)scalar operators @xmath34 can be expressed , through the fierz transformations , as linear combinations of the new tensor operators eq .
( [ eq : tensor ] ) , constraints on these two operators can be derived easily from those on the latter . to further test such a particular np scenario with anomalous tensor operators
, we also give predictions for the branching ratio and the longitudinal polarization fraction of @xmath5 decay , which is also involved the same quark level @xmath35 transition .
all these results could be tested at the fermilab tevatron and the lhc - b experiments .
the paper is organized as follows .
is devoted to the theoretical framework . after a brief entertainment of the qcd factorization formalism ( qcdf ) @xcite
, we discuss the anomalous tensor operator contributions to the @xmath36 and @xmath37 decays . in sec . 3 ,
our numerical analysis and discussions are presented .
contains our conclusions .
appendix a recapitulates the amplitudes for the five decay modes within the sm @xcite .
all the theoretical input parameters relevant to our analysis are summarized in appendix b.
in the sm , the effective hamiltonian responsible for @xmath12 transitions is given as @xcite @xmath38 + { \rm h.c.},\end{aligned}\ ] ] where @xmath39 ( @xmath40 and @xmath41 ) are products of the cabibbo - kobayashi - maskawa ( ckm ) matrix elements @xcite , @xmath42 the wilson coefficients , and @xmath43 the relevant four - quark operators whose explicit forms could be found , for example , in ref .
@xcite .
in addition to @xmath44 , we must employ a factorisation formalism of hadronic dynamics to study the @xmath45 , @xmath46 , @xmath47 , @xmath48 , and @xmath49 decays . to this end
, we take the framework of qcdf @xcite . the factorization formula allows us to calculate systemically the hadronic matrix element of operators in the effective hamiltonian eq .
( [ eq : eff ] ) @xmath50 where @xmath51 is the @xmath52 transition form factor , @xmath53 and @xmath54 are the perturbatively calculable hard kernels , and @xmath55 are the universal nonperturbative light - cone distribution amplitudes ( lcdas ) of the corresponding mesons . in the recent years , the qcdf formalism has been employed extensively to study non - leptonic @xmath6 decays .
for example , all the decay modes considered here have been studied comprehensively within the sm in refs .
we recapitulate the amplitudes for @xmath36 and @xmath37 decays in appendix a. it is noted that , along with its many novel progresses in non - leptonic @xmath6 decays , the framework contains estimates of some power corrections which can not be computed rigorously .
these contributions may be numerically important for realistic @xmath6-meson decays , especially for some penguin - dominated decay modes @xcite .
in fact , there are no reliable methods available at present to _ calculate _ such contributions . to give _ conservative _ theoretical predictions ,
at least , one should leave the associated parameters varying in reasonable regions to show their possible effects . in this work , following closely the treatment made in refs .
@xcite , we will parameterize the end - point divergences associated with these power corrections as @xmath56 in the following numerical calculations , we take the parameter @xmath57 and the phase @xmath58 varying in the range @xmath59 and @xmath60 , respectively . in our calculation , we have neglected possible intrinsic charm content and anomalous gluon couplings related to the meson @xmath61 ,
both of which have been shown to have only marginal effects on the four @xmath16 decays @xcite .
as for the @xmath61@xmath62 mixing effects , we shall adopt the feldmann - kroll - stech ( fks ) scheme @xcite as implemented in ref .
a recent study and comparison of different @xmath61@xmath62 mixing schemes has been given in ref .
@xcite . in the amplitude for @xmath63 decay in eq .
( [ eq : phiksm ] ) , a new power - enhanced electromagnetic penguin contribution to the negative - helicity electroweak penguin coefficient @xmath64 , as first noted by beneke , rohrer and yang @xcite , has also been taken into account in our calculation . in a recent comprehensive study of @xmath65
decays @xcite , it is found that the small @xmath18 could be accommodated within the sm , however , with very large theoretical uncertainties .
since the sm may have difficulties in explaining the large @xmath66 and the measured polarization observables in @xmath11 decays , we shall discuss possible np resolutions to these observed discrepancies .
specifically , we shall investigate whether these discrepancies could be resolved by introducing two anomalous four - quark tensor operators defined by eq .
( [ eq : tensor ] ) .
we write the np effective hamiltonian as @xmath67 + { \rm h.c.},\ ] ] with the tensor operators @xmath68 and @xmath69 defined by eq .
( [ eq : tensor ] ) .
the coefficient @xmath70 describes the relative interaction strength of the tensor operator @xmath71 , and @xmath72 is the np weak phase . in principle , such four - quark tensor operators could be produced in various np scenarios , e.g. , in the minimal supersymmetric standard model ( mssm ) @xcite .
interestingly , the recent study of radiative pion decay @xmath73 at pibeta detector @xcite has found deviations from the sm predictions in the high-@xmath74low-@xmath75 kinematic region , which may indicate the existence of anomalous tensor quark - lepton interactions @xcite . at first , we present the np contributions to the decay amplitudes of @xmath16 and @xmath17 decays due to these new tensor operators .
since their coefficients are unknown parameters , for simplicity , we shall only consider the leading contributions of these tensor operators .
as for the four @xmath29 decays , the relevant feynman diagram due to the tensor operators @xmath76 is shown in fig .
it is easy to realize that the amplitude corresponding to another type of insertion would vanish in the leading order approximation . instead of using the fierz transformation , an easy way to calculate the amplitude in fig .
[ fig1 ] is to use the light - cone projection operator of the meson @xmath61 in momentum space @xcite @xmath77_{\alpha\beta},\ ] ] where @xmath78 , @xmath79 , and @xmath80 are the momentum , leading - twist , and twist-3 lcdas of the meson @xmath61 , respectively . @xmath81 and @xmath82 denote the momenta of the quark and anti - quark in the meson @xmath61 , and are given by @xmath83 with @xmath84 , and @xmath85 the momentum fraction carried by the constituent quark .
the decay constant @xmath86 and the factor @xmath87 in eq .
( [ projector ] ) are defined , respectively , by @xcite @xmath88 where we have used @xmath89 , with @xmath90 and @xmath91 @xcite . after some simple calculations , the np contributions to the decay amplitudes of the four @xmath29 decays due to @xmath92 in eq .
( [ eq : np ] ) can be written as @xmath93 where @xmath94 , and the factor 3 is due to contractions of the involved @xmath95 matrices .
it is interesting to note that the above four decay amplitudes are all proportional to the chirally - enhanced " factor @xmath96 , which has been found to be very important for charmless hadronic @xmath6 decays @xcite .
we now present the np contribution to the decay amplitude of @xmath97 decay . based on the observation that they contribute only to the transverse polarization amplitudes but not to the longitudinal one @xcite
, these anomalous four - quark tensor operators have been proposed to resolve the polarization anomaly observed in @xmath98 decays .
the relevant feynman diagrams are shown in fig .
[ fig2 ] . for fig .
[ fig2 ] ( a ) , we shall use the following matrix elements @xcite @xmath99 for fig .
[ fig2 ] ( b ) , we shall use the light - cone projector operator of the transversely polarized vector meson @xmath21 @xcite @xmath100 where @xmath101 is the leading twist lcda of the meson @xmath21 , and the ellipsis denotes additional parts that have no contributions in our case .
it is easy to find that neither fig .
[ fig2](a ) nor ( b ) contributes to the longitudinal polarization amplitude , and the final decay amplitude can be written as @xmath102\ , \biggl .
\nonumber\\ & & \biggl .
-2 t_{3}^{b\rightarrow k^{\ast}}(m_{\phi}^2)\,\frac{m_{\phi}^2}{m_{b}^{2 } - m_{k^{\ast}}^{2}}\ , ( \varepsilon_{1\perp}^{\ast } \cdot p ) ( \varepsilon_{2}^{\ast}\cdot p ) \biggl\}\,,\end{aligned}\ ] ] with @xmath103 . in the helicity basis ,
the new decay amplitude eq .
( [ phik_np ] ) can be further decomposed into @xmath104\,,\label{hphi2_np}\end{aligned}\ ] ] where @xmath105 is the center - of - mass momentum of final mesons in @xmath106 rest frame .
compared with the sm predictions eqs .
( [ hphi1_sm ] ) and ( [ hphi2_sm ] ) , the new transverse polarization amplitudes @xmath107 are enhanced by a factor of @xmath108 , while the longitudinal part remains unchanged .
it is therefore expected that these new tensor operators might provide a possible resolution to the polarization anomaly observed in @xmath17 decays . from the above discussions , the total decay amplitudes are then given as @xmath109 where @xmath110 denotes the sm results obtained using the qcdf , and @xmath111 the contributions of the particular np scenario with anomalous tensor operators in eq .
( [ eq : np ] ) .
the corresponding branching ratios are @xmath112 is the life time of @xmath6 meson . in the transversity basis @xcite , the decay amplitude for any @xmath113 decay
can also be decomposed into another three quantities @xmath114 , @xmath115 , and @xmath116 , which are related to the helicity amplitudes @xmath117 , @xmath118 , and @xmath119 through @xmath120 in terms of these quantities , we can express the longitudinal polarization fraction as @xmath121 ( the definition of these observables is compatible with that used by the babar and belle collaborations @xcite ) , are potentially very useful for constraining the parameter spaces of np scenario , however , depend on whether the strong phases of the helicity amplitudes could be calculated reliably .
with the theoretical formulas and the input parameters summarized in appendix b for the decay modes of our concerns , we now go to our numerical analysis and discussions . as shown by the np decay amplitudes eqs .
( [ amp1_np])([amp4_np ] ) and ( [ hphi1_np])([hphi2_np ] ) , the allowed regions for the parameters @xmath122 and @xmath72 can be obtained from the four measured @xmath123 , and the ones for @xmath124 from @xmath125 ) and @xmath18 , respectively . generally , we have five branching ratios and one polarization fraction , but only three free parameters : one np weak phase @xmath72 and two effective coefficients @xmath122 and @xmath124 ( or equivalently @xmath126 and @xmath127 ) .
so , it is easy to guess that these decays could severely constrain or rule out the np scenario with two anomalous tensor operators eq .
( [ eq : np ] ) . to make the guess clear
, we shall first find out the allowed regions for the parameters @xmath122 and @xmath72 from the four @xmath128 decays .
then , using the allowed regions for the weak phase @xmath72 , we try to put constraint on the parameter @xmath124 from @xmath129 and @xmath18 .
finally , we can obtain the allowed parameter spaces , if there are , for @xmath126 , @xmath127 , and @xmath72 that satisfy all the experimental data on the decay modes of our concerns . to further test the particular np scenario eq .
( [ eq : np ] ) , we also present our theoretical predictions for @xmath130 decay .
theoretical estimations of the annihilation contributions at present suffer from very large uncertainties , which of course will dilute the requirement of np very much . to show the dilution
, we will give our numerical results for two cases , i.e. , with and without annihilations for comparison .
since the tensor operator contributions to the decay amplitudes of the four @xmath128 decays are all characterized by the parameters @xmath122 and @xmath72 , possible regions for these two np parameters can be obtained from the measured branching ratios of @xmath128 decays .
our main results are shown in tables [ br1 - 4comparison][gtdeltatresult ] and figs .
[ gtdeltat1][gtdeltat2 ] .
the experimental data listed in table [ br1 - 4comparison ] are taken from the heavy flavor averaging group ( hfag ) @xcite .
the sm predictions for the branching ratios of these four decays are presented in the third column of table [ br1 - 4comparison ] , where the theoretical uncertainties are obtained by varying the input parameters within the regions specified in eq .
( [ treat - for - anni ] ) and appendix b. for each decay mode , the first and the second row are evaluated with and without the annihilation contributions , respectively . [ cols="<,^,^,^,^",options="header " , ] from the numerical results presented in tables [ tab : phik ] and [ tab : phiphi ] , we can see that both the branching ratio and the longitudinal polarization fraction of @xmath131 decay is larger than those of the decay @xmath132 , which is due to the relative factor of two in the @xmath133 amplitude eq .
( [ bsphiphi - sm ] ) , as well as the additional contribution from the annihilation coefficient @xmath134 .
in particular , due to an accidental cancelation , the annihilation coefficient @xmath135 ( contributing to both @xmath136 and @xmath137 ) is quite smaller than @xmath138 @xcite .
interestingly , recent calculations made in the perturbative qcd approach also predict a large branching ratio @xmath139 with the longitudinal polarization fraction @xmath140 @xcite .
their result for @xmath141 is smaller than our predictions within the sm .
the predicted results presented in table [ tab : phiphi ] could be tested at the fermilab tevatron and the lhc - b experiments .
in summary , motivated by the observed discrepancies between the experimental data and the sm predictions for the branching ratios of @xmath10 decays and the polarization fractions in @xmath11 decays , we have studied a particular np scenario with anomalous tensor operators @xmath68 and @xmath69 , which has been proposed to resolve the polarization anomalies in the literature @xcite .
after extensive numerical analysis , we have found that the above observed discrepancies could be resolved simultaneously and constraints on the np parameters have been obtained . with both the experimental data and the theoretical input parameters varying within @xmath142 error bars
, we have found the following two solutions : ( i ) both the two tensor operators contribute with the parameter spaces presented in the upper two rows of table [ tab : c15c16result ] ; ( ii ) only @xmath69 contributes with @xmath143 and @xmath72 the same as solution ( i ) .
the above two solutions correspond to the scenario with new operators @xmath144 and @xmath145 added to the sm , with the parameter spaces : ( i ) @xmath146 , @xmath147 and ( ii ) @xmath148 .
their weak phase @xmath149 is the same as @xmath72 . against our early expectations ,
we have found that the solution with two tensor operators are dominated by the color - octet operator @xmath69 .
it would be interesting to investigate whether the available np models could give such effective interactions at the @xmath150 scale . to further test the particular np scenario with two anomalous tensor operators @xmath76 or ( pseudo- ) scalar operators @xmath151
, we have also presented our predictions for the observables in @xmath131 decay , which could be tested more thoroughly at the fermilab tevatron and the lhc - b experiments in the near future .
it should be noted that the strong constraints in table [ tab : c15c16result ] obtained without annihilation contributions may be too optimistic . as has been shown in the table , the uncertainties due to poorly known annihilation contributions could dilute the requirement of np very much .
generally , this caveat could be applied to probe possible np scenarios in exclusive non - leptonic @xmath6 decays .
further theoretical progress is strongly demanded .
the work is supported by national science foundation under contract nos .
10305003 and 10675039 , and the ncet program sponsored by ministry of education , china .
the amplitudes for @xmath16 are recapitulated from ref .
@xcite @xmath152\biggl . \nonumber\\ & & \biggl
. + \sqrt{2 } a_{k^-\eta_s } \big[\delta_{pu}\,\beta_2 + \alpha_3^p + \alpha_4^p - { \frac{1}{2}}\alpha_{3,{\rm ew}}^p -{\frac{1}{2}}\alpha_{4,{\rm ew}}^p + \beta_3^p + \beta_{3,{\rm ew}}^p\big]\biggl . \nonumber\\ & & \biggl .
+ a_{\eta_q k^- } \big [ \delta_{pu}\,(\alpha_1+\beta_2 ) + \alpha_4^p+\alpha_{4,{\rm ew}}^p + \beta_3^p + \beta_{3,{\rm ew}}^p\big]\biggl\}\,,\label{amp1_sm}\end{aligned}\ ] ] @xmath153 \biggl . \nonumber\\ & & \biggl .
+ \sqrt{2 } a_{\bar k^0\eta_s } \big [ \alpha_3^p + \alpha_4^p - { \frac{1}{2}}\alpha_{3,{\rm ew}}^p - { \frac{1}{2}}\alpha_{4,{\rm ew}}^p + \beta_3^p - { \frac{1}{2}}\beta_{3,{\rm ew}}^p \big]\biggl . \nonumber\\ & & \biggl .
+ a_{\eta_q\bar k^0 } \big[\alpha_4^p - { \frac{1}{2}}\alpha_{4,{\rm ew}}^p + \beta_3^p - { \frac{1}{2}}\beta_{3,{\rm ew}}^p\big ] \biggl\}\,,\label{amp2_sm}\end{aligned}\ ] ] @xmath154\biggl .
\nonumber\\ & & \biggl .
+ \sqrt{2 } a_{k^{\ast-}\eta_s } \big[\delta_{pu}\,\beta_2 + \alpha_3^p + \alpha_4^p - { \frac{1}{2}}\alpha_{3,{\rm ew}}^p -{\frac{1}{2}}\alpha_{4,{\rm ew}}^p + \beta_3^p + \beta_{3,{\rm ew}}^p\big]\biggl . \nonumber\\ & & \biggl .
+ a_{\eta_q k^{\ast- } } \big [ \delta_{pu}\,(\alpha_1 + \beta_2)+ \alpha_4^p+\alpha_{4,{\rm ew}}^p + \beta_3^p + \beta_{3,{\rm ew}}^p\big]\biggl\}\,,\label{amp3_sm}\end{aligned}\ ] ] @xmath155 \biggl . \nonumber\\ & & \biggl .
+ \sqrt{2 } a_{\bar k^{\ast0}\eta_s } \big[\alpha_3^p + \alpha_4^p - { \frac{1}{2}}\alpha_{3,{\rm ew}}^p -{\frac{1}{2}}\alpha_{4,{\rm ew}}^p + \beta_3^p - { \frac{1}{2}}\beta_{3,{\rm ew}}^p\big]\biggl .
\nonumber \\ & & \biggl .
+ a_{\eta_q\bar k^{\ast0 } } \big[\alpha_4^p -{\frac{1}{2}}\alpha_{4,{\rm
ew}}^p + \beta_3^p -{\frac{1}{2}}\beta_{3,{\rm ew}}^p\big ] \biggl\}\,,\label{amp4_sm}\end{aligned}\ ] ] where the explicit expressions for the coefficients @xmath156 and @xmath157 could also be found in ref .
we recall that the @xmath158 terms contain one - loop vertex , penguin and hard spectator contributions , whereas the @xmath159 terms are due to the weak annihilation , and the transition form factors are encoded in the factors @xmath160 .
the decay amplitude of @xmath49 mode can be read off from refs .
@xcite @xmath161\,,\ ] ] where the superscript ` @xmath162 ' denotes the helicity of two final - state vector mesons , with @xmath163 corresponding to two outgoing longitudinal , positive , and negative helicity vector mesons , respectively . in the helicity basis ,
the decay amplitude eq .
( [ eq : phiksm ] ) can be further decomposed into the following three helicity amplitudes @xmath164,\label{hphi1_sm}\\
h_{\pm\pm}^{\rm sm}&=&a_{\bar{k}^{\ast0}\phi}^{\pm } \sum_{p = u , c}v_{pb}v_{ps}^{\ast } \big [ \alpha_3^{p,{\pm } } + \alpha_4^{p,{\pm}}- { \frac{1}{2}}\alpha_{3,{\rm ew}}^{p,{\pm } } -{\frac{1}{2}}\alpha_{4,{\rm ew}}^{p,{\pm } } + \beta_3^{p,{\pm } } - { \frac{1}{2}}\beta_{3,{\rm ew}}^{p,{\pm}}\big],\label{hphi2_sm}\end{aligned}\ ] ] with @xcite @xmath165 @xmath166 \,,\\
a^{\pm}_{\bar{k}^{\ast 0}\phi } & = & \frac{i g_f}{\sqrt{2 } } m_{b_d } m_{\phi } f_{\phi}\ , \left[(1+\frac{m_{k^{\ast 0}}}{m_{b_d } } ) a_1^{b\to k^{\ast}}(m_{\phi}^2)\mp ( 1-\frac{m_{k^{\ast 0}}}{m_{b_d } } ) v^{b\to k^{\ast}}(m_{\phi}^2)\right].\end{aligned}\ ] ]
the wilson coefficients @xmath167 have been evaluated reliably to next - to - leading logarithmic order @xcite . their numerical results in the naive dimensional regularization scheme at the scale @xmath168 ( @xmath169 ) are given by @xmath170 the values at the scale @xmath171 , with @xmath172 and @xmath173 , should be used in the calculation of hard - spectator and weak annihilation contributions .
for the ckm matrix elements , we adopt the wolfenstein parameterization @xcite and choose the four parameters @xmath174 , @xmath175 , @xmath176 , and @xmath61 as @xcite @xmath177 with @xmath178 and @xmath179 . as for the quark mass , there are two different classes appearing in our calculation .
one type is the pole quark mass appearing in the evaluation of penguin loop corrections , and denoted by @xmath180 . in this paper , we take @xmath181 the other one is the current quark mass which appears in the factor @xmath182 through the equation of motion for quarks .
this type of quark mass is scale dependent and denoted by @xmath183 .
here we take @xcite @xmath184 where @xmath185 , and the difference between @xmath186 and @xmath187 quark is not distinguished . in this paper , we take the decay constants @xmath191 and the form factors @xcite @xmath192 for the parameters related to the @xmath61-@xmath193 mixing , we choose @xcite @xmath194 the other parameters relevant to the meson @xmath61 can be obtained from the above three ones @xcite . as for the @xmath6 meson wave functions ,
we need only consider the first inverse moment of the leading lcda @xmath196 defined by @xcite @xmath197 where @xmath198 @xcite is the hadronic parameter introduced to parameterize this integral .
h. j. lipkin , arxiv : hep - ph/0608284 ; s. khalil , arxiv : hep - ph/0608157 ; m. gronau and j. l. rosner , phys .
d * 74 * ( 2006 ) 057503 ; s. baek , j. high energy phys .
* 0607 * ( 2006 ) 025 ; r. arnowitt _ et al .
_ , phys .
b * 633 * ( 2006 ) 748 ; h. n. li , s. mishima , and a. i. sanda , phys .
d * 72 * ( 2005 ) 114005 ; x. q. li and y. d. yang , phys . rev .
d * 72 * ( 2005 ) 074007 ; c. s. kim , s. oh , and c. yu , phys . rev .
d * 72 * ( 2005 ) 074005 ; s. khalil , phys . rev .
d * 72 * ( 2005 ) 035007 .
h. j. lipkin , phys .
b * 633 * ( 2006 ) 540 ; a. r. williamson and j. zupan , phys .
d * 74 * ( 2006 ) 014003 ; * 74 * ( 2006 ) 03901(e ) ; j. f. cheng , c. s. huang , and x. h. wu , nucl . phys .
* b701 * ( 2004 ) 54 ; b. dutta _ et al .
_ , eur .
j. c * 37 * ( 2004 ) 273 ; b. dutta , c. s. kim , and s. oh , phys .
* 90 * ( 2003 ) 011801 ; c. w. chiang , m. gronau , and j. l. rosner , phys .
d * 68 * ( 2003 ) 074012 ; a. kundu and t. mitra , phys . rev .
d * 67 * ( 2003 ) 116005 ; e. kou and a. i. sanda , phys .
b * 525 * ( 2002 ) 240 ; d. s. du , d. s. yang , and g. h. zhu , arxiv : hep - ph/9912201 ; d. s. du , c. s. kim , and y. d. yang , phys .
b * 426 * ( 1998 ) 133 . c. h. chen and h. hatanaka , phys .
d * 73 * ( 2006 ) 075003 ; c. s. huang _ et al _ , phys .
d * 73 * ( 2006 ) 034026 ; s. baek _ et al .
_ , phys .
d * 72 * ( 2005 ) 094008 ; y. d. yang , r. m. wang , and g. r. lu , phys .
d * 72 * ( 2005 ) 015009 ; h. n. li and s. mishima , phys .
d * 71 * ( 2005 ) 054025 ; c. h. chen and c. q. geng , phys .
d * 71 * ( 2005 ) 115004 ; h. y. cheng , c. k. chua , and a. soni , phys . rev .
d * 71 * ( 2005 ) 014030 ; h. n. li , phys . lett .
b * 622 * ( 2005 ) 63 ; p. colangelo , f. de fazio , and t. n. pham , phys .
b * 597 * ( 2004 ) 291 ; m. ladisa _ et al .
_ , phys .
d * 70 * ( 2004 ) 114025 ; e. alvarez _ et al _ , phys .
d * 70 * ( 2004 ) 115014 ; c. dariescu _ et al _ , phys . rev .
d * 69 * ( 2004 ) 112003 .
a. datta and d. london , int .
j. mod .
a * 19 * ( 2004 ) 2505 ; d. london , n. sinha , and r. sinha , phys .
d * 69 * ( 2004 ) 114013 ; k. abe , m. satpathy , and h. yamamoto , arxiv : hep - ex/0103002 ; i. dunietz _ et al .
_ , phys .
d * 43 * ( 1991 ) 2193 . | we investigate whether the anomalous tensor operators with the lorentz structure @xmath0 , which could provide a simple resolution to the polarization anomaly observed in @xmath1 decays , could also provide a coherent resolution to the large @xmath2 and survive bounds from @xmath3 decays .
parameter spaces satisfying all these experimental data are obtained , and found to be dominated by the color - octet tensor operator contribution .
constraints for the equivalent solution with @xmath4 operators are also derived and found to be dominated by the color - singlet one . with the constrained parameter spaces , we finally give predictions for @xmath5 decay , which could be tested at the fermilab tevatron and the lhc - b experiments . *
pacs numbers : 13.25.hw , 12.38.bx , 12.15.mm . * |
there has been a recent interest in the links between finsler geometry and the geometry of relativistic spacetimes .
a well - known reason comes from the viewpoint of applicability : finsler metrics are much more general than riemannian ones and , accordingly , if one replaces the lorentz metric of a spacetime by a finslerian counterpart , the possibility to model physical effects is richer .
a more practical reason comes from a purely mathematical correspondence : the geometry of a concrete class of finsler manifolds ( randers spaces ) is closely related to the conformal structure of a class of spacetimes ( standard stationary ones ) .
so , results in one of these two fields can be translated into results on the other one and , sometimes , this can be generalized to more general finsler and lorentz spaces .
the purpose of this paper is to make a brief revision of this topic , emphasizing the applications to physics . after some preliminary definitions and motivations on finsler metrics ( section [ s2 ] ) , we focus on the problem of describing spacetimes by using finsler elements ( section [ s3 ] ) .
we introduce and discuss a notion of lorentz - finsler metric which leads to _ ( conic ) finsler spacetimes_. our definition is quite general and , so , several previous notions in the literature fit in it .
such a finsler spacetime provides a _ cone structure_. this notion has been studied systematically very recently @xcite , and it allows us to introduce classical causality in the framework of finsler spacetimes .
in particular , some of the key results on this topic can be recovered , theorem [ tfs ] .
it is worth pointing out : ( a ) the cone structure of a finsler spacetime generalizes the one provided by the chronological futures of a classical spacetime , and defines implicitly a past cone structure ( but no further assumption on reversibility should be necessary if one liked to include in the definition of conic finsler spacetime the two cones at each point ) , and ( b ) in classical spacetimes , the cone structure is equivalent to the conformal one , but in finsler spacetimes many non - conformally related lorentz finsler metrics will have the same cone structure and , thus , the same causality that is , the information of the lorentz - finsler metric not contained in the cone structure is much richer than in classical spacetimes . in section [ s4 ] , we make a brief survey on the recent developments about the geometric correspondence between standard stationary spacetimes and randers spaces .
we focus in those questions with more clear applications to relativistic spacetimes and , so , topics such as the causal structure of standard stationary spacetimes or the gravitational lensing are emphasized .
finally , the conclusions are summarized in the last section .
let us introduce the very general notion of ( conic ) pseudo - finsler metric in a manifold @xmath0 .
let @xmath1 denote the tangent bundle of @xmath0 , @xmath2 , the natural projection and @xmath3 , an open subset satisfying that it is conic ( that is , if @xmath4 and @xmath5 , then @xmath6 ) , and it projects on all @xmath0 ( i.e. , @xmath7 ) .
we say that a smooth function @xmath8 is a _ ( conic ) pseudo - finsler metric _ if @xmath9 is positive homogeneous of degree 2 , namely , @xmath10 for any @xmath4 and @xmath5 , and the fundamental tensor @xmath11 , defined as the hessian of @xmath12 at every @xmath4 , is nondegenerate . in other words ,
given @xmath4 , the bilinear symmetric form in @xmath13 defined as @xmath14 for any @xmath15 , is non - degenerate . let us remark that classical finsler metrics are a particular case of pseudo - finsler metrics .
more specifically , @xmath9 is a finsler metric if @xmath16 and the fundamental tensor @xmath11 is positive definite for every @xmath4 . in this case , @xmath9 is always positive , since by homogeneity @xmath17 for every @xmath4 and we can consider @xmath18 , which is positive homogeneous of degree one . the function @xmath19 is usually called the finsler metric function , and it is determined by its _ indicatrix _ , i.e. the set of its unit vectors . among the most classical examples of finsler metrics
, we will use randers metrics , which are given by @xmath20 for every @xmath21 , where @xmath22 and @xmath23 and @xmath24 are a riemannian metric and a one - form in @xmath0 respectively .
in fact , the fundamental tensor of @xmath19 is positive definite in @xmath21 if and only if @xmath25 ( see ( * ? ? ?
* corollary 4.17 ) ) ; this restriction is satisfied everywhere if the @xmath23-norm of @xmath24 is smaller than 1 , so defining a randers metric .
these metrics appear naturally in several contexts as in zermelo navigation problem , which aims to describe the trajectories that minimize the time in the presence of a mild wind or current modelled by a vector field @xmath26 .
then , these trajectories are given by geodesics of a randers metric defined as @xmath27 where @xmath21 , @xmath11 is a riemannian metric and @xmath28 in all @xmath0 @xcite .
observe that zermelo metrics are always positive for every @xmath29 and then its fundamental tensor is positive definite . as we will see later , randers metrics also appear naturally associated to stationary spacetimes describing their causal properties ( see section [ s4 ] ) .
other remarkable example is given by matsumoto metric , which describes trajectories minimizing time in the presence of a slope recall that going up is slower than going down .
this metric is defined as @xmath30 for every @xmath21 , and its fundamental tensor is positive definite in @xmath31 if and only if @xmath32 ( * ? ?
* corollary 4.15 ) .
therefore , matsumoto metric is properly _ conic _ if this inequality is not satisfied by some @xmath33 .
randers and matsumoto metrics are particular examples of the class of @xmath34-metrics , which are defined as @xmath35 , being @xmath36 an arbitrary non - negative real function ( see @xcite and references therein ) .
in classical general relativity , a _ spacetime _ is a ( connected , hausdorff ) @xmath37-manifold @xmath0 endowed with a lorentzian metric @xmath11 and a time orientation , i.e. , a continuous choice of one of the two timelike cones at each point , which will be regarded as _ future - directed_. there are some speculative applications of the replacement of @xmath11 by a ( generalization of a ) finsler metric @xmath19 , as modeling possible anisotropies of the spacetime even at an infinitesimal level , or admitting speeds higher than light . in any case
, the possibility to model a general action functional ( homogeneous of degree two , but not necessarily coming from a quadratic form ) justifies the study of the lorentz - finsler approach .
there are different possibilities in order to define a finsler spacetime .
recall that , for a riemannian metric @xmath38 , the unit vectors constitute the indicatrix of a ( standard ) finsler metric ; in particular , each unit sphere @xmath39 is convex ( as the boundary of the unit ball ) at each point .
nevertheless , if @xmath11 is a lorentzian metric , one has to consider the subsets @xmath40 containing , resp .
the spacelike and timelike unit vectors at @xmath41 .
notice that @xmath42 always contains two connected parts , each one _ concave_. for @xmath43 these properties also hold for @xmath44 , but for higher dimensions @xmath44 is connected and , at each point @xmath41 , its second fundamental form ( say , with respect to any auxiliary euclidean product at @xmath45 ) has lorentzian signature . from the physical viewpoint ,
the most important elements of the spacetime are the causal vectors of the metric , since they model the trajectories of massive and massless particles .
observe in particular that the length of a causal curve @xmath46\rightarrow m$ ] in the spacetime is computed as @xmath47 and a fundamental property in the spacetime is that causal geodesics locally maximize this length .
if we want to define a more general way of measuring the length of curves ( but preserving that the length does not depend on the parametrization of the curve and geodesics are local length - maximizers ) , then we must consider a positive one - homogeneous function @xmath48 with fundamental tensor of signature @xmath49 ( see @xcite ) .
this would be enough for timelike curves describing the trajectories of massive particles , but if in addition we want to describe the trajectories of massless particles , we will consider a positive two - homogeneous function @xmath9 , rather than a one - homogeneous function , since for a lorentzian metric @xmath11 , @xmath50 is not smooth when @xmath51 is lightlike . summing up
, we say that a _ ( conic ) finsler spacetime _ is a manifold endowed with a conic pseudo - finsler metric @xmath52 which satisfies the following properties : * each @xmath53 is convex in @xmath45 ( i.e. , the segment in @xmath45 connecting each two vectors @xmath54 is entirely contained in @xmath55 ) , in particular , each @xmath55 must be strictly included in a half - plane of @xmath45 , * @xmath56 has a smooth boundary in @xmath57 , * the extension of @xmath9 as 0 to the closure @xmath58 of @xmath56 is smooth at @xmath59 ( notice that the extension is always continuous at @xmath60 by homogeneity , and this extension can not be smooth there even for a classical finsler metric , except if it comes from a riemannian metric ) , and * both the fundamental tensor @xmath11 in and its extension to the points in @xmath61 have signature @xmath49 . for practical purposes , we will consider always that @xmath9 is extended to @xmath58 and @xmath11 to @xmath62 for all @xmath63 and @xmath64 . in order to make computations related to lightlike geodesics
, one can also assume that @xmath9 is smoothly extended ( in a non - unique way ) on a neighborhood of @xmath65 .
we will call @xmath18 the _ lorentz - finsler metric _ ( defined on @xmath58 ) of the finsler spacetime . observe that our notion of finsler spacetime is based on a `` conic '' element which is essentially present in most of previous literature .
in particular : 1 . in @xcite ,
the function @xmath9 is defined in all @xmath1 and the fundamental tensor is assumed to have signature @xmath49 .
the restriction to one connected component of the subset @xmath66 in every point @xmath63 gives a ( conic ) finsler spacetime as defined here . in @xcite ,
the definition is essentially as in @xcite with the opposite sign of @xmath9 ( and an increasing attention to relax differentiability ) .
the definition in @xcite is somewhat more involved : they introduce an @xmath67-homogeneous function @xmath9 , with @xmath68 endowed with a fundamental tensor of signature one and , then , they associate a homogeneous finsler function @xmath69{l}$ ] .
this allows one to deal with non - differentiability in cases that extend the lightlike vectors in our definition above . in @xcite ,
the author gives a definition analog to ours , but he excludes the conditions on the boundary ( so that one is not worried about lightlike vectors ) .
2 . lorentz - finsler metrics appear in the context of lorentz violation . here , one starts with a background lorentz - minkowski space , but the movement of particles or waves is governed by a lagrangian with a lorentz - finsler behavior ( see @xcite and references therein ) . under natural hypotheses , the following positively homogeneous function is found in @xcite : @xmath70 where @xmath71 is the standard metric of minkowski spacetime @xmath72 , @xmath51 is any timelike vector and @xmath73 are two prescribed vector fields in @xmath72 .
the interplay between @xmath19 and the background metric @xmath71 becomes important for physical applications ( for example , the group velocity of propagating waves may exceed the light speed in @xmath72 ) .
even though @xmath19 can be regarded as a lorentz - finsler metric , in some cases the fundamental tensor of @xmath19 may be degenerate on some timelike directions ( when @xmath74 , it is easy to compute the fundamental tensor of @xmath19 ( * ? ? ?
* proposition 4.17 ) , and this becomes degenerate when @xmath75 ) .
general physical theories of modified gravity ( see @xcite or the very recent articles @xcite and references therein ) yield naturally lorentz - finsler spacetimes as defined above . the exact domain @xmath56 of @xmath9 for general expressions in finsler - cartan gravity ( say , as in ( * ? ? ?
* formula ( 2 ) ) ) may be difficult to compute .
however , one can compute @xmath56 easily in _ general very special relativity _ , which possesses a simple finslerian line element introduced by bogoslovsky ( say , the expression @xmath76 in ( * ? ? ?
* formula ( i.1 ) ) becomes a lorentz - finsler metric with domain @xmath56 equal to the intersection at each point of the future time - cone of @xmath71 and the half - space @xmath77 ) or in randers - type cosmologies as ( * ? ? ?
* formula ( 3.1 ) ) . notice that our definition of finsler spacetime is
a generalization of the structure obtained in a classical spacetime when one considers only its _ future causal cones_. in the case of lorentzian metrics , however , the value of the metric on the causal vectors is enough to determine the metric on all the vectors and , moreover , the lightlike vectors are enough to determine the metric up to a conformal factor .
so , the cone structure for lorentzian metrics is equivalent to the conformal structure .
nevertheless , this does not hold by any means in the case of ( conic ) finsler spacetimes : clearly two lorentz finsler metrics @xmath78 on @xmath0 with the same domain @xmath58 may not be equal up to a multiplicative function .
however , the domain @xmath56 is a _ cone structure _ and , thus , one can reconstruct all the causality theory for finsler spacetimes . let us review this briefly . given a finsler spacetime @xmath79 a tangent vector @xmath21 is called ( future - directed ) timelike if @xmath4 , causal if @xmath80 , lightlike if @xmath81 and null if either @xmath51 is lightlike or the zero vector , and the past directions could be consistently regarded as those in @xmath82 .
it would be natural then to extend @xmath9 to @xmath82 by requiring full homogeneity ( @xmath83 for all @xmath84 ) , but this is not necessary for the issues regarding causality .
however such an extension ( as well as possible further extension of @xmath9 to non - causal vectors ) would be important for the particular physical theory one may be modelling . ] .
a ( piecewise smooth ) curve @xmath85 on @xmath0 is also called timelike , causal etc . depending on the character of its velocity at all the points .
if @xmath86 , we say that @xmath41 lies in the chronological ( resp .
causal ) past of @xmath87 if there exists a ( future - directed ) timelike ( resp .
causal or null ) curve starting at @xmath41 and ending at @xmath87 ; in this case we write @xmath88 ( resp .
@xmath89 ) and we also say that @xmath87 lies in the chronological ( resp .
causal ) future of @xmath41 .
the chronological and causal futures of @xmath41 , as well as its corresponding pasts , are then defined formally as in the case of lorentzian metrics : @xmath90 one says also that a second lorentz finsler metric @xmath91 has cones wider than @xmath19 , denoted @xmath92 when their corresponding domains @xmath93 satisfy @xmath94 . with these definitions
, one can extend directly the _ causal ladder _ of classical spacetimes ( see for example @xcite ) to finsler spacetimes .
we recall some of the steps of this ladder .
a finsler spacetime will be called _ chronological _ ( resp . _
causal _ ) when it does not admit closed timelike ( resp .
causal ) curves , _ stably causal _ when there exists a causal lorentz - lorentz metric @xmath91 with wider cones @xmath92 , and _ globally hyperbolic _ when it is stably causal and the following property holds : @xmath95 is compact for any @xmath86 .
trivially : @xmath96 taking into account the case of spacetimes , one realizes that there are more intermediate levels of the ladder as well as many subtle properties and relations among them to be considered for finsler spacetimes .
however , we focus here just on a pair of them , with deep implications for the global structure . to this aim , we define for a finsler spacetime @xmath79 : * a _ spacelike hypersurface _ is a smooth hypersurface @xmath97 such that no causal vector @xmath51 is tangent to @xmath97 , * a _ ( spacelike ) cauchy hypersurface _ is a spacelike hypersurface @xmath97 such that any causal curve which is inextendible in a continuous way , intersects @xmath97 exactly once , * a _ temporal function _ is a smooth function @xmath98 such that @xmath99 for all ( future - directed ) causal vector @xmath51 ( thus , its levels @xmath100constant are spacelike hypersurfaces ) and , * a _ cauchy temporal function _ is a temporal function such that all its levels are cauchy hypersurfaces . in the case of classical spacetimes , results by geroch @xcite and hawking @xcite obtained at a topological level , plus their improvements to the smooth and metric cases by bernal and snchez @xcite , prove the equivalence between being stably causal ( resp , globally hyperbolic ) and admitting a temporal ( resp .
cauchy temporal ) function . by using different arguments coming from kam theory ,
these results were re - proved and extended to general cone structures by fathi and siconolfi @xcite .
so , as a consequence of the latter we get the following result .
[ tfs ] let @xmath101 be a finsler spacetime .
\(1 ) if @xmath102 is stably causal , then it admits a temporal function ( and , thus , it can be globally foliated by spacelike hypersurfaces ) .
\(2 ) if @xmath102 is globally hyperbolic , then it admits a cauchy hypersurface and a cauchy temporal function .
recall , however , that the finsler spacetime has elements that are not characterized by the cone structure .
for example , one can define a _ finler separation _
@xmath103 $ ] between any @xmath86 by taking the supremum of the lengths of the ( future - directed ) causal curves from @xmath41 to @xmath87 in a close analogous to the _ lorentzian separation _ or _
distance _ for spacetimes with lorentzian metric .
the analogies and differences of @xmath104 with the lorentzian case ( in particular , its interplay with causality ) deserve to be studied further .
one expects that the converse of parts ( 1 ) and ( 2 ) of theorem [ tfs ] , as well as most of classical causality theory , hold . to achieve this , typical lorentzian techniques , as those regarding limit curves or quasi - limits , must be adapted ( recall the lorentzian proofs of the converse in @xcite , ( *
6.6 ) or ( * ? ? ?
* th . 14.38 , cor . 14.39 ) ) . moreover ,
as the results in @xcite work for very general cone structures , theorem [ tfs ] can be applied to a more general class of spacetimes that includes those in ( * ? ? ?
* example 3.3 ) , which can be non - smooth in some timelike directions , and also the cone structure provided by the example ( 2 ) above ( see @xcite ) .
from a classical viewpoint , a correspondence between some purely geometric elements of lorentzian and finslerian manifolds has been developed recently .
such a correspondence provides a precise description of certain objects in ( classical ) spacetimes in terms of an associated finsler space .
the precise correspondence is developed between a particular class of spacetimes , the _ standard complete - conformastationary _ ones , or just _ stationary _ , for short , and a precise class of finsler manifolds , the randers ones ( see ) .
but some consequences and techniques can be extrapolated to general lorentzian and finslerian manifolds .
we start with a trivial observation .
probably the simplest examples of lorentzian manifolds are the products @xmath105 , where @xmath106 and @xmath107 are the natural projections
. the geometric properties of @xmath108 depend on those of @xmath109 .
in particular , each curve @xmath110 parametrized with unit speed yields naturally two lightlike curves @xmath111 ( future - directed with the sign @xmath112 and past - directed with @xmath113 , for the natural time orientation of the spacetime ) , which are geodesics iff @xmath114 is a geodesic in @xmath115 . a less trivial spacetime is obtained if we admit cross terms between the time and space parts ( independent of the time @xmath98 ) .
this can be described by using a 1-form @xmath116 on @xmath0 , namely , we consider the spacetime : @xmath117 now , we introduce the following finslerian _ fermat metrics _ associated to : @xmath118 notice that these are metrics of randers type , and @xmath119 is the reverse finsler metric of @xmath120 , so , we will write simply @xmath19 for the latter and @xmath121 for the former . if we consider curves @xmath122 and @xmath123 in @xmath0 which are unit for @xmath19 and @xmath121 resp . , the curves in the spacetime @xmath124 are again ( future or past directed ) lightlike curves in @xmath125 , and each one is a geodesic up to parametrization iff so is the corresponding original curve @xmath122 or @xmath123 ( see the details in @xcite or @xcite ) .
this suggests the possibility of describing the properties related to the causality and conformal structure in in terms of the geometry of the corresponding randers space @xmath102 .
it is worth pointing out that , as only conformally invariant properties will be taken into account , the class of spacetimes to be considered includes those conformal to .
this class of spacetimes can be characterized intrinsically as those which are distinguishing ( a causality condition less restrictive than strong causality ) and admit a complete timelike conformal vector field .
in fact , the conformal change of the metric @xmath11 by @xmath126 , plus the result in @xcite , allows one to find the expression , which defines a _ normalized standard stationary spacetime_. so , one will find a correspondence between the conformal properties of the elements in this class of metrics @xmath127 and the geometric properties of randers spaces @xmath102 .
this has been carried out at different levels ( see @xcite or @xcite for a review ) , and we will focus here in three of them , with clear physical applications . nicely , fermat metrics allow one to determine the chronological and causal future and past of any point in the stationary spacetime .
for example , if we take @xmath128 then the intersection of @xmath129 with the slice @xmath130 of the spacetime is equal to @xmath131 , where @xmath132 is the open ball of center @xmath41 and radius @xmath133 for the fermat metric @xmath134 . from these considerations ,
one can describe in a precise way the causal structure of the spacetime . concretely
, one has @xcite : [ t1 ] a stationary spacetime is always causally continuous and it is * causally simple ( i.e. , it is causal , and the causal futures and pasts @xmath135 are closed ) if and only if @xmath102 is convex , i.e. any pair @xmath136 in @xmath0 can be connected by means of an @xmath120-geodesic of length equal to the finslerian distance @xmath137 . *
globally hyperbolic if and only if the closed balls for the symmetrized distance of @xmath138 are compact .
* globally hyperbolic with slices @xmath100constant that are cauchy hypersurfaces if and only if @xmath138 is forward and backward complete
. other causal elements which are described naturally with the finslerian elements are the cauchy horizons and developments .
for example , given a subset @xmath139 , its future cauchy horizon @xmath140 is the graph , in @xmath141 of the function which maps each @xmath142 to @xmath143 , i.e. , the @xmath138-distance from the complement of @xmath56 to @xmath144 . the results in stationary spacetimes can also be translated to results in randers metrics , sometimes generalizable to any finsler manifold . among them
, we point out ( see @xcite ) : * the theorem above suggests that the compactness of the closed symmetrized balls of @xmath138 , which is a weaker hypothesis than the commonly used ( forward or backward ) completeness of @xmath138 , can substitute the last hypothesis in many results . and
, in fact , this is the case in classical theorems of finsler geometry , such as myers theorem or the sphere theorem .
* from the known fact that any globally hyperbolic spacetime admits a smooth spacelike cauchy hypersurface , one can deduce that any randers metric @xmath145 with compact symmetrized balls admits a trivial projective change ( @xmath146 ) such that the corresponding new randers metric has _ the same pregeodesics as @xmath145 , and it is forward and backward complete_. this result has been extended by matveev @xcite for any finsler metric . * the results on horizons ( a substantial topic in lorentzian geometry ) in stationary spacetimes , yield directly results on randers spaces .
for example , the translation of a well - known result by beem and krolak @xcite yields the following property for any subset @xmath56 of a randers manifold @xmath147 : _
@xmath63 is a differentiable point of the distance from @xmath41 if and only if it can be crossed by exactly one minimizing segment .
_ this result has been extended to any finslerian manifold by sabau and tanaka @xcite .
assume that in our spacetime , a point @xmath148 represents an event and a line @xmath149 , obtained as an integral curve of @xmath150 , represents the trajectory of a stellar object .
we wonder if there exist lightlike geodesics from @xmath149 to @xmath148 ( i.e. , whether an observer at @xmath148 can see the object @xmath149 ) and , in this case , if there are many of such geodesics ( i.e. , the existence of a lens effect such that @xmath149 is seen in two different directions ) .
this situation is applicable to cosmological models such as friedmann - lemaitre - robertson - walker ones , as they are conformally stationary ( @xmath149 would represent a `` comoving observer '' of the model ) .
the problem becomes more realistic if we choose some ( open ) region @xmath151 which contains @xmath148 and @xmath149 , and search for geodesics contained in this region .
these problems are related to the convexity of @xmath152 for the fermat metric ( in the sense of the last subsection ) which turns to be related to the convexity of its boundary @xmath153 .
there are several different notions for the latter , as the _ local _ and _ infinitesimal convexity _ and , as shown in @xcite , they are equivalent to its _
geometric convexity_. the latter means that , given any pair of points of @xmath152 , any geodesic connecting them and contained in the closure @xmath154 , must be entirely contained in @xmath152 .
the ( geometric ) convexity of @xmath152 turns out equivalent to the _ light ( geometric ) convexity _ of the boundary of @xmath155 ( i.e. , the property of convexity holds when restricted to lightlike geodesics ) and , finally , this is equivalent to the question of existence of connecting geodesics , yielding @xcite : assume that the closed balls of @xmath154 computed for the restriction of the symmetrized distance @xmath156 are compact ( which happens for example , if the intersections with @xmath154 of the closed symmetrized balls in @xmath0 are compact or , simply , if @xmath138 is complete on all @xmath0 ) . then , the following assertions are equivalent : @xmath157 is causally simple ( i.e. @xmath158 is convex , theorem [ t1 ] ) .
@xmath153 is convex for the fermat metric @xmath19 .
@xmath159 is light - convex for the lorentzian metric @xmath160 .
any point @xmath161 and any line @xmath162 , with @xmath163 , can be joined in @xmath164 by means of a future - directed lightlike geodesic @xmath165 $ ] , which minimizes the ( future ) arrival time @xmath166 ( i.e. , such that @xmath167 minimizes the @xmath19-distance in @xmath152 from @xmath41 to @xmath87 ) .
idem to the previous property but replacing `` future '' by `` past '' and `` @xmath19-distance '' by ( the reverse ) `` @xmath168-distance '' .
the previous result solves the question of existence of connecting lightlike geodesics .
the question of multiplicity has some possibilities .
first , the existence of a conjugate point of the lightlike geodesic @xmath169 at @xmath148 .
this is equivalent to the existence of a conjugate point at @xmath41 for its projection @xmath167 ( * ? ? ?
* theorem 13 ) , and it is regarded as trivial .
second , the non - triviality of the topology of @xmath152 may yield a _
topological lensing_. in fact , one can prove that , whenever @xmath152 is not contractible , infinitely many connecting lightlike geodesics ( with diverging arrival times ) will exist .
the results can be also extended to the case of timelike geodesics , prescribing its length ( i.e. , the lifetime of the massive particles represented by such geodesics ) .
the idea relies on a reduction of the problem to the lightlike one , by considering an extra spacelike dimension , see ( * ? ? ?
* section 4.3 ) and ( * ? ? ?
* section 5.2 ) .
the studied _ stationary to randers correspondence _ can be also applied to the study of the causal boundary of the spacetime in terms of finsler elements .
we recall that , in mathematical relativity , the penrose conformal boundary is commonly used , in spite of the fact that it is not an intrinsic construction , and there are problems to ensure its existence or uniqueness .
the causal boundary is a conformally invariant alternative , which is intrinsic and can be constructed systematically in any strongly causal spacetime ( see @xcite for a comprehensive study of this boundary ) .
the computation of the causal boundary and completion of a stationary spacetime has been carried out in full generality in @xcite .
it must be emphasized that this boundary has motivated the definition of a new _ busemann boundary _ in any finslerian manifold .
even more , this has stimulated the study of further properties of the gromov boundary in both , riemannian and finlerian manifolds .
it is also worth mentioning two further topics of current interest .
the first one is the correspondence at the level of curvatures between the weyl tensor of the stationary spacetime and the flag curvature of the randers space , see @xcite .
the second one is the possibility to extend the studied correspondence to the case of spacetimes with a nonvasnishing complete killing vector field , whose causal type may change from timelike to lightlike and spacelike .
such metrics are related to a generalization of the zermelo metrics which includes the possibility of a strong wind ( i.e. , with a speed higher than the one which can be reached by the engine of the ship ) , see @xcite .
as a first goal , we have introduced a very general notion of finsler spacetime which includes many previous ones .
this notion emphasizes the role of the cone structure , showing that many properties of ( classical , relativistic ) spacetimes are potentially transplantable to this setting in particular , this holds for a major lorentzian result such as theorem 1 .
remarkably , in relativistic spacetimes the cone structure is equivalent to the conformal structure and , thus , it is governed by the weyl tensor . however , the metrics @xmath170 of two finsler spacetimes with the same cone structure are not by any means necessarily conformal , and one is lead to deal directly with the cone structure .
this cone structure would remain then at a more basic level than any possible weyl - type tensor and , thus , also at a much basic level than the cartan or chern connections ( of course , such connections will be relevant in order to describe concrete physical effects ) .
the importance on the choice of the connection is stressed because a finsler metric admits several associated connections . among them , one has the classical cartan connection , the berwald connection , the chern - rund connection and the hashiguchi connection ( see @xcite or @xcite ) . thus , for example , two finsler metrics with the same chern connection may have different cartan connections . as a consequence ,
the way to describe physical phenomena by using these connections may be non - equivalent and even controversial @xcite . nevertheless ,
as emphasized here , all the possible physical causal influences will depend exclusively on the cone structure , at least at a non - quantum level . as a second goal we have shown a different type of equivalence between the geometry of any randers space and the conformal geometry of classical standard stationary spacetimes .
this equivalence leads to obtain very precise results which relate : ( a ) the properties of distances on the finslerian side with the causality on the side of the spacetime , ( b ) the properties related to finslerian _
convexity _ wth the properties related to relativistic observability and lensing , ( c ) accurate relations between ideal finslerian boundary and the causal boundary ( and , implicitly then , the conformal one ) of spacetimes , or ( d ) interesting relations between finslerian curvature and weyl curvature @xcite .
this equivalence has suggested new results in both , finslerian and lorentzian geometries .
moreover , a forthcoming work @xcite will show its extendability and applicability to new fields namely , an extension of classical finsler metrics which allows one to model strong winds in zermelo navigation problem is introduced , and its geometrical properties are shown equivalent to the conformal ones of a class of spacetimes wider than standard stationary ones , which allow the possibility of horizons and black holes .
comments by kostelecki , laemmerzahl , perlick , szilasi and the referee are warmly acknowledged .
both authors are partially supported by the grant p09-fqm-4496 ( j. andaluca ) with feder funds .
the first - named author is also partially supported by mineco - feder project mtm2012 - 34037 and fundacin sneca project 04540/germ/06 , spain .
this research is a result of the activity developed within the framework of the programme in support of excellence groups of the regin de murcia , spain , by fundacin sneca , regional agency for science and technology ( regional plan for science and technology 2007 - 2010 ) .
the second - named author acknowledges the support of ihes , bures sur - yvette , france , for a three months stay in 2013 ( which included the period of celebration of the meeting ifwgp ) ; he is also partially supported by mtm201018099 ( micinn - feder ) .
e. caponio , m. .
javaloyes and a. masiello , morse theory of causal geodesics in a stationary spacetime via morse theory of geodesics of a finsler metric , _ ann . inst .
h. poincar anal .
non linaire _ , * 27*(3 ) ( 2010 ) , 857876 . | recent links between finsler geometry and the geometry of spacetimes are briefly revisited , and prospective ideas and results are explained .
special attention is paid to geometric problems with a direct motivation in relativity and other parts of physics . |
history of fractional calculus ( fc ) goes back to seventeenth century , when in 1695 the derivative of order @xmath0 was described by leibnitz . since then , the new theory turned out to be very attractive to mathematicians as well as biologists , chemists , economists , engineers and physicsts .
several books were written on the theories and developments of fc @xcite . in @xcite samko , et al , provides a comprehensive study of the subject .
several different derivatives were introduced : riemann - liouville , hadamard , grunwald - letnikov , riesz and caputo are just a few @xcite . in fractional calculus ,
the fractional derivatives are defined by a fractional integral @xcite .
there are several known forms of the fractional integrals of which two have been studied extensively for their applications @xcite .
the first is the _ riemann - liouville fractional integral _ defined for parameter @xmath1 by , @xmath2 motivated by the _ cauchy integral formula _ @xmath3 well - defined for @xmath4 .
the second is the _ hadamard fractional integral
_ introduced by j. hadamard @xcite , and given by , @xmath5 this is based on the generalization of the integral @xmath6 here we want to present the fractional integration , which generalizes both the riemann - liouville and hadamard fractional integrals into a single form .
new generalization is based on the observation that , for @xmath7 , @xmath8 which suggests the following fractional version @xmath9 where @xmath10 are real numbers .
these integrals are called _ left - sided _ integrals .
similarly we can define _ right - sided _
integrals @xcite . in the subsequent sections , we give conditions for the integration operator @xmath11 to be bounded in @xmath12
. we also establish semigroup property for the generalized fractional integration operator and finally , we give the general definition of the fractional derivatives .
consider the space @xmath13 of those complex - valued lebesgue measurable functions @xmath14 on @xmath15 $ ] for which @xmath16 , where the norm is defined by @xmath17 and for the case @xmath18
@xmath19 \quad ( c \in \textbf{r}).\ ] ] in particular , when @xmath20 the space @xmath21 coincides with the classical @xmath22-space with @xmath23 to derive a formula for the generalized integral , consider for natural @xmath24 and real @xmath25 and @xmath26 , the n - fold integral of the form @xmath27 first notice that - using the dirichlet technique ( see p. 64 of @xcite ) , we have @xmath28 repeating the above step @xmath29 times we arrived at @xmath30 fractional version of ( [ eq : nint ] ) is the following @xmath31 where @xmath32 and @xmath33 are real numbers . when @xmath34 we arrive at the standard riemann - liouville fractional integral , which is used to define both the riemann - liouville and caputo fractional derivatives @xcite .
using lhospital rule , when @xmath35 , we have @xmath36 this is the famous hadamard fractional integral ( [ eq : h ] ) @xcite . hadamard fractional integral has been extensively studied , especially , hadamard - type fractional calculus @xcite , composition and semigroup properties @xcite , mellin transforms @xcite , integration by parts formulae @xcite , and _
g - transform _
representations @xcite are just a few .
in this section we show that the generalized fractional integration operator @xmath37 is well - defined on @xmath12 for @xmath38 .
we have the following theorem [ eq : th1 ] let @xmath39 and let @xmath40 and @xmath41 be such that @xmath42 . then the operator @xmath43 is bounded in @xmath12 and @xmath44 [ eq : thm1 ] where @xmath45 [ eq : con1 ] first consider the case @xmath46 . since @xmath47 , then @xmath48 and we can apply the generalized minkowsky inequality .
we have @xmath49 and hence @xmath50 where @xmath51 thus , theorem [ eq : th1 ] is proved for @xmath52 . for @xmath18 , by ( [ eq : df2 ] ) and ( [ eq : fint ] ) we have @xmath53 after the substitution @xmath54 .
this agrees with ( [ eq : con2 ] ) above .
this completes the proof of the theorem .
note that this proof is similar to the proof of the theorem 2.1 in @xcite .
the following version of the theorem [ eq : th1 ] has been proved in @xcite , for the special case when @xmath35 .
let @xmath55 and let @xmath40 and @xmath56 be such that @xmath42 .
then the operator @xmath57 is bounded in @xmath12 and @xmath58 where @xmath59 for @xmath60 , while @xmath61\ ] ] for @xmath62 , where @xmath63 is the incomplete gamma - function defined by @xmath64 substituting @xmath65 in theorem [ eq : th1 ] and taking ( [ eq : df3 ] ) into account , we deduce the boundedness of the operator @xmath66 in the space @xmath67 .
let @xmath68 and let @xmath40 be such that @xmath69 .
then the operator @xmath57 is bounded in @xmath22 and @xmath70 where @xmath71 [ eq : con3 ] notice that this is corollary 2.2 of @xcite .
we now turn to algebraic properties of the integral operator .
in this section we give semigroup properties of the integral operator [ eq : th2 ] let @xmath72 and let @xmath73 and @xmath74 be such that @xmath42 . then for @xmath75 the semigroup property holds .
that is , @xmath76 [ eq : semi ] using fubinis theorem , for `` sufficiently good '' function @xmath14 , we have @xmath77 the inner integral is evaluated by the change of variable @xmath78 @xmath79 according to the known formulae for the beta function @xcite . substituting ( [ eq : pf2 ] ) into ( [ eq : pf1 ] ) we obtain @xmath80 and thus , ( [ eq : semi ] ) is proved for sufficiently good functions @xmath14 . if @xmath42 then by theorem [ eq : th1 ] the operators @xmath81 and @xmath82 are bunded in @xmath12 , hence the realation ( [ eq : semi ] ) is true for @xmath83 this completes the proof of the theorem [ eq : th2 ] .
we have the following corollary .
let @xmath72 and let @xmath73 be such that @xmath69 .
then for @xmath84 the semigroup property ( [ eq : semi ] ) holds .
we can extend ( [ eq : fint ] ) to an arbitrary complex order @xmath85 and can also define _ right - sided _ generalized fractional integrals .
let @xmath86 \ , ( -\infty < a < b < \infty)$ ] be a finite interval on the real axis @xmath87 .
the generalized fractional integral @xmath88 of order @xmath89 is defined by @xmath90 for @xmath91 and @xmath92 .
this integral is called the _ left - sided _ fractional integral .
similarly we can define the _ right - sided _ fractional integral @xmath93 by @xmath94 for @xmath95 and @xmath92 .
in the next section we will define _ generalized fractional derivatives _ for an arbitrary complex order @xmath89 for riemann - liouville type ( r - g ) and similarly , we can define generalized fractional derivatives for caputo - type ( l - g ) .
the integrals ( [ eq : df1 ] ) and ( [ eq : df2 ] ) allow one to define the corresponding generalized fractional derivatives . here
we only define riemann - liouville type derivatives . for @xmath35
, we will also get the corresponding hadamard - type derivatives as a natural generalization to the standard hadamard fractional derivatives .
the caputo - type derivatives can be defined similarly . when @xmath106 , formulars ( [ eq : df3 ] ) and ( [ eq : df4 ] ) take the following forms , @xmath107 for @xmath91 and @xmath108 for @xmath95 , where @xmath109 , while formulars ( [ eq : df5 ] ) and ( [ eq : df6 ] ) are given by @xmath110 for @xmath91 and @xmath111 for @xmath105 .
let @xmath98 and @xmath116 . the generalised caputo - type derivative of arbitrary order @xmath32 of @xmath117 is defined by @xmath118 for @xmath91 , and @xmath119 for @xmath95 , if the right - hand - sides exist . to demonstrate the use of the new derivative
, we will obtain generalized fractional derivative of the power function for a special case .
for simplicity assume @xmath122 and @xmath123 , and use the following version of the derivative @xmath124 we find the generalized derivative of the function @xmath125 , where @xmath126 .
the formula ( [ eq : simple0 ] ) yields @xmath127 to evaluate the inner integral , we use the substitution @xmath128 to obtain , @xmath129 where @xmath130 is the beta function . combining ( [ eq : simple2 ] ) with ( [ eq : simple1 ] ) , we obtain , @xmath131 for @xmath132 , after using the properties of the beta function @xcite and the relation @xmath133 . when @xmath34 we obtain the riemann - liouville fractional derivative of the power function given by @xcite , @xmath134 to compare results , we plot ( [ eq : gen ] ) for several values of @xmath25 .
we also consider different values of @xmath135 to see the effect on the degree of the power function .
the results are summaries in figure 1 and figure 2 : find an exact formula for the left - sided generalized fractional derivative of the power function , @xmath136 with @xmath137 , that is , evaluate the following integral , @xmath138 for @xmath91 where @xmath85 , @xmath101 and @xmath139 .
according to the figure 1 and figure 2 , we notice that the characteristics of the fractional derivative is highle affected by the value of @xmath25 , thus it provides a new direction for the control applications .
the paper presents a new fractional integration , which generalizes the riemann - liouville and hadamard fractional integrals into a single form , which when a parameter fixed at different values , produces the above integrals as special cases .
conditions are given for such a generalized fractional integration operator to be bounded in an extended lebesgue measurable space .
semigroup property for the above operator is also proved .
we also gave a general definition of the fractional derivatives and gave some examples . in a future paper
, we will derive formulae for the laplace , fourier and mellin transforms for the generalized fractional integral .
we already know that we can deduce hadamard and riemann - liouville integrals for the special cases of @xmath25 .
we want to further investigate the effect on the new parameter @xmath25 .
we will also study the generalize fractional derivatives and their properties .
those results will appear elsewhere .
butzer , p. l. , kilbas , a. a. , and trujillo , j.j .
, _ compositions of hadamard - type fractional integration operators and the semigroup property _ , journal of mathematical analysis and applications , * 269 * , ( 2002 ) , 387 - 400 .
butzer , p. l. , kilbas , a. a. , and trujillo , j.j . , _ fractional calculus in the mellin setting and hadamard - type fractional integrals _ , journal of mathematical analysis and applications , * 269 * , ( 2002 ) , 1 - 27 .
butzer , p. l. , kilbas , a. a. , and trujillo , j.j . , _ melling transform analysis and integration by parts for hadamard - type fractional integrals _ , journal of mathematical analysis and applications , * 270 * , ( 2002 ) , 1 - 15 . | the paper presents a new fractional integration , which generalizes the riemann - liouville and hadamard fractional integrals into a single form , which when a parameter fixed at different values , produces the above integrals as special cases .
conditions are given for such a generalized fractional integration operator to be bounded in an extended lebesgue measurable space .
semigroup property for the above operator is also proved .
finally , we give a general definition of the fractional derivatives and give some examples .
submitted to : applied mathematics and computation |
Subsets and Splits