article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the busy reader can take the following shortcut : 1
. read theorem [ thm : main ] on page for the main result .
2 . read definition [ def : evmquant ] if its statement is not clear .
3 . look at eqns .
( [ equ : cexdef ] ) and the preceding enumeration for the definition of the counterexample .
let @xmath9 be a smooth @xmath10-dimensional manifold equipped with a poisson structure @xmath0 , making @xmath11 a poisson algebra with bracket @xmath12 . in this paper , we will exclusively deal with the case @xmath13 .
kontsevich @xcite has shown that one can always quantize this algebra , i.e. , find an associative product @xmath2 on @xmath14 $ ] such that for all @xmath15 @xmath16 furthermore , he showed that the set of such star products is , up to equivalence , in one to one correspondence with the set of formal poisson structures on @xmath9 extending @xmath0 .
a _ formal poisson structure _ @xmath6 is a formal bivector field @xmath17\ ] ] satisfying the jacobi identity @xmath18 } } = 0\ ] ] where @xmath19 } } $ ] is the schouten - nijenhuis bracket .
we say that @xmath6 _ extends _ the poisson structure @xmath20 if its @xmath21-component is @xmath0 .
let now @xmath22 be a point and consider the evaluation map @xmath23 it makes @xmath24 into a @xmath11-module , i.e. , for all @xmath15 @xmath25 the question treated in this paper is the following : * main question : * can one quantize the evaluation map @xmath26 ? by this
we mean the following : [ def : evmquant ] let @xmath27 be the kontsevich star product associated to the formal poisson structure @xmath6 on @xmath9 and let @xmath22 be an arbitrary point . a formal map @xmath28\ ] ] will be called _ quantization of @xmath26 _ if the following holds .
* it has the form @xmath29 where the @xmath30 are differential operators evaluated at @xmath31 .
concretely , this means in local coordinates that @xmath32 where the sum is over multiindices and the @xmath33 are constants , vanishing except for finitely many @xmath34 .
* for all @xmath15 @xmath35 let @xmath36 and @xmath22 . if a quantization of @xmath26 exists , then @xmath37 , i.e. , @xmath0 vanishes at @xmath31 .
the @xmath38-component of the equation @xmath39 } } _
\star ) = { { \left[{\rho(f)},{\rho(g)}\right ] } } = 0\ ] ] reads @xmath40 hence @xmath41 .
remark : a similar calculation for a higher dimensional submanifold also yields the higher dimensional coisotropy condition .
from now on we will assume that @xmath41 , or , equivalently , that @xmath42 is coisotropic .
for details on coisotropic submanifolds see @xcite .
the above main question has been answered positively by cattaneo and felder in @xcite , @xcite , provided @xmath0 satisfies certain conditions . adapted to our context
, they proved the following theorem .
[ thm : cf ] for any formal poisson structure @xmath6 on @xmath13 such that @xmath43 for some @xmath22 there exists a linear map @xmath44\ ] ] @xmath45 where the @xmath46 are differential operators evaluated at @xmath31 , that satisfies @xmath47 for all @xmath15 . here
@xmath48 ( the `` anomaly '' ) is a bidifferential operator .
hence , if the `` anomalous '' term on the r.h.s . of ( [ equ : modano ] )
vanishes , one sees that @xmath49 becomes a quantization of @xmath26 .
the precise form of @xmath50 is recalled in section [ sec : cfano ] . when @xmath51 to all orders in @xmath8 , the anomaly is actually at least of order @xmath52 .
furthermore , we will later provide an example for which the @xmath52-term does not vanish . a theorem similar to theorem [ thm : cf ] above holds in the case of higher dimensional coisotropic submanifolds .
there , anomaly terms will also occur in general .
it is still an open question whether the vanishing of these terms is merely a removable technical condition or a fundamental obstruction to quantizability .
our paper gives a partial answer to this question in the simplest possible case . in this section
we review the construction of cattaneo and felder @xcite leading to theorem [ thm : cf ] . we throughout assume familiarity with the construction of kontsevich s star product @xcite .
the map @xmath49 of theorem [ thm : cf ] has the explicit form @xmath53 the sum is over all kontsevich graphs with one type ii , whereas type ii vertices are associated to the functions one intends to multiply . ]
vertex ( associated to @xmath54 ) .
the differential operator @xmath55 is constructed exactly as it is constructed for kontsevich s star product .
the weights @xmath56 are given by the integral formula @xmath57 here the @xmath58 , the _ cattaneo - felder configuration space _
, is ( a compactification of ) the space of all embeddings of the vertex set of @xmath59 into the first quadrant , such that the type ii vertex is mapped into the real axis .
similar to the kontsevich case , the weight form @xmath60 is defined as a product of one - forms , one for each edge in the edge set @xmath61 of @xmath59 . @xmath62 here the edge @xmath63 in the product is understood to point from the vertex @xmath64 that is mapped to @xmath65 to the vertex @xmath66 , that is mapped to @xmath67
. the precise expression for the angle form @xmath68 will never be needed , but we will use the following facts about its boundary behaviour in the appendix : 1 . if @xmath69 lies on the real axis or @xmath70 on the imaginary axis , @xmath71 vanishes .
2 . if @xmath69 and @xmath70 both come close to each other and and a point on the positive real axis , @xmath71 approaches kontsevich s angle form . 3 . if @xmath69 and @xmath70 both come close to each other and a point on the positive imaginary axis , @xmath72 approaches kontsevich s angle form .
i.e. , @xmath71 approaches kontsevich s form after reversal of the edge direction .
the anomaly @xmath50 in theorem [ thm : cf ] can be computed by the formula @xmath73 here the sum is over all _
anomaly graphs_. such a graph is a kontsevich graph , but with a third kind of vertices , which we call _ anomalous _ or _ type iii _ vertices .
an anomaly graph is required to contain at least one such type iii vertex .
these anomalous vertices have exactly 2 outgoing edges .
the weight @xmath56 is computed just as the cattaneo - felder weight , but with the type iii vertices constraint to be mapped to the imaginary axis .
the computation of @xmath55 also remains the same as before , but one has to specify which bivector fields to associate with the new type iii vertices . in local coordinates @xmath74 , @xmath75 , the components of this bivector field will be denoted '' in @xmath76 is not an index , just a label . ]
@xmath77 it is in turn given as a sum of graphs .
@xmath78 here the sum is over all cattaneo - felder graphs with 2 type ii vertices and @xmath55 is again defined as in the cattaneo - felder case before .
however , the weights @xmath79 are computed by the following algorithm : 1 . delete the type ii vertices in @xmath59 and all their adjacent edges . 2 .
reverse the direction of all edges .
compute the kontsevich weight of the resulting graph . with this anomalous vertex , one can construct two kinds of graphs that yield @xmath80-contributions : * the graph with only one vertex , which is anomalous .
it yields the contribution proportional to @xmath81 to the anomaly . * the graphs with one type i and one anomalous vertex as shown in figure [ fig : anogra ] .
together , they yield a symmetric contribution to the anomaly .
we will use the following notation for the parts of @xmath50 of various orders in @xmath8 : @xmath82 here @xmath83 and @xmath84 are antisymmetric .
the contribution to @xmath84 comes from the two graphs in figure [ fig : first_anomally ] , with @xmath85 attached to the vertices of the left graph and a @xmath86 attached to the vertex of the right graph .
note also that if @xmath87 to all orders in @xmath8 , then the contribution of the right graph vanishes and @xmath88 .
it is easily seen that , if the poisson structure @xmath89 is linear , i.e. , if @xmath9 is the dual of a lie algebra , the anomalous vertex @xmath90 vanishes @xcite .
this is because any contributing graph with @xmath10 vertices will contain @xmath91 edges .
but for a graph @xmath59 with different numbers of vertices and edges , @xmath92 by power counting .
hence a contribution will not arise unless @xmath93 .
but the weight of the only possible graph with 2 vertices is 0 by one of kontsevich s lemmatas @xcite , i.e. , a reflection argument .
this also implies that the anomaly can be removed whenever the poisson structure is linearizable .
at least formally , the poisson structure can be linearized whenever the second lie algebra cohomology of the lie algebra defined by the linear order , with values in the symmetric algebra , vanishes .
for example , this is true for semisimple lie algebras .
see @xcite for details . as pointed out to the author by a. s. cattaneo and g. felder ,
the anomaly also vanishes whenever the linear order of @xmath0 does .
this is shown by power counting :
each vertex in an anomaly graph comes with two edges ( derivatives ) and is of at least quadratic order .
but two edges have to be external and do not contribute derivatives , so the result vanishes when evaluated at @xmath31 .
[ sec : cafpoint ] the graphs accounting for the @xmath94 and @xmath52 contributions to the anomaly vertex @xmath90 .
the right graph only contributes if @xmath95 . ]
the author wants to thank his advisor prof .
g. felder for his continuous support and patient explanation of most of the material presented here .
furthermore , the computation of weights in appendix [ sec : nonvanishing ] is mainly due to g. felder and a. s. cattaneo , the author merely filled in some details .
[ thm : main ] there exists a poisson structure @xmath0 on @xmath96 vanishing at @xmath97 , s.t .
no quantization of @xmath7 exists for the kontsevich star product associated to @xmath0 .
furthermore , this also holds for any formal poisson structure @xmath6 extending @xmath0 @xmath98 as long as @xmath87 . placing this theorem into a more general context
, this means that the @xmath11-module structure on a coisotropic submanifold can not always be quantized to a module structure for @xmath99,\star)$ ] , where @xmath100 is the usual kontsevich product .
hence this theorem answers main question 1 negatively .
main question 2 however , is only partially answered . a complete answer to main question 2
we can not give , only some more hints , see @xcite .
without loss of generality we can use the following ansatz for @xmath101 . @xmath102 here @xmath49 is the map of theorem [ thm : cf ] and @xmath103 has the form @xmath104 our goal , and content of the next sections , is to find the lowest order restrictions on @xmath103 coming from the requirement of @xmath101 being a quantum module map , i.e. , eqn .
( [ equ : mod ] ) . concretely ,
the requirement is @xmath105 here we simply inserted ( [ equ : rhoansatz ] ) into ( [ equ : mod ] ) and used ( [ equ : modano ] ) . to order @xmath21
this equation is obviously satisfied . the @xmath38 part of eqn .
( [ equ : mainmod ] ) reads @xmath106 choosing @xmath107 both constant we see that the zeroth derivative part of @xmath108 has to vanish . choosing @xmath54 and @xmath109 both linear the r.h.s . vanishes and
hence the second - derivative contribution to @xmath108 has to vanish . picking @xmath54
quadratic and @xmath109 linear we see that the third - derivative part of @xmath110 must vanish and similarly that all higher derivative parts must vanish as well .
hence @xmath111 for some constants @xmath112 .
we will separately consider the contributions symmetric and antisymmetric in @xmath54 , @xmath109 .
the antisymmetric contribution reads @xmath113 note that if @xmath87 , then @xmath114 . the l.h.s . of ( [ equ : d1precon ] )
is zero if @xmath54 or @xmath109 contains no linear part .
hence it suffices to treat the case where @xmath54 and @xmath109 are both linear .
then the equation becomes @xmath115 where @xmath116 are the components of @xmath0 w.r.t .
the standard coordinates @xmath117 .
the symmetric part yields the constraint @xmath118
picking @xmath107 linear we see that the second derivative part of @xmath119 must be @xmath120 inserting this back into ( [ equ : o2sym ] ) we obtain the same constraint equation for the remaining parts of @xmath119 as we had found for @xmath108 in eqn .
( [ equ : o1 ] ) . by the same logic as there we can hence deduce that @xmath121 for some yet undetermined constants @xmath122 , @xmath75 . here
all derivatives are implicitly understood to be evaluated at zero , e.g. @xmath123 note that the calculations presented so far are valid for any formal poisson structure @xmath98 as long as it vanishes at 0 .
i.e. , the higher order terms do not contribute to the first two orders in @xmath8 of eqn .
( [ equ : mainmod ] ) .
( [ equ : con3 ] ) in terms of graphs .
the black vertices symbolize @xmath0 , the grey ones @xmath124 .
the solid and dashed arrows without a vertex at their tail symbolize @xmath125 and @xmath126 respectively .
the constant coefficients are omitted . ]
we will only need to consider the antisymmetric part and linear @xmath127 and @xmath128 in eqn .
( [ equ : mainmod ] ) . in this case
the @xmath52 part of the equation becomes @xmath129 the first term is the contribution of the @xmath38-term in the formal poisson structure @xmath6 .
it is absent if we consider @xmath89 . to derive the above formula we used the following .
* the r.h.s . of ( [ equ : mainmod ] ) is obviously symmetric in @xmath54 , @xmath109 , hence all contributions come from the l.h.s . * the kontsevich product satisfies @xmath130 } } _
\star=\epsilon \pi^{ij}+\epsilon^2 \pi_1^{ij } + o(\epsilon^3).\ ] ] hence , using ( [ equ : d1con ] ) we obtain @xmath131 } } _
\star ) & = & \epsilon^3 \phi_1(\pi_1^{ij})+\epsilon^3\phi_2(\pi^{ij } ) + o(\epsilon^4)\\ & = & \epsilon^3\left ( d_1^k\partial_k\pi_1^{ij } + \frac{1}{2}d_1^kd_1^l\partial_k\partial_l\pi^{ij } + d_2^k\partial_k\pi^{ij } \right)+o(\epsilon^4)\end{aligned}\ ] ] the constraint ( [ equ : con3 ] ) is displayed graphically in figure [ fig : greqn ] . in this section
we present a @xmath0 such that there are no constants @xmath132 satisfying ( [ equ : d1con ] ) and ( [ equ : con3 ] ) for @xmath133 .
this will prove the first part of theorem [ thm : main ] .
for this , the following data is needed : 1
. some finite dimensional semisimple lie algebra @xmath134 with structure coefficients @xmath135 in some basis @xmath136 .
we denote by @xmath137 its killing form and by @xmath138 its inverse .
2 . some finite dimensional lie algebra @xmath139 such that its second cohomology group @xmath140 .
denote its structure coefficients @xmath141 .
3 . a non - trivial @xmath142 , with coefficients @xmath143 in some basis @xmath144 .
the simplest possible choice would be @xmath145 , @xmath146 as abelian lie algebra and @xmath147 , @xmath148 .
the poisson structure will reside in @xmath149 where @xmath150 .
we use as coordinate functions the above basis @xmath74 and @xmath151 .
then one can define @xmath152 where @xmath153 is the quadratic casimir in @xmath154 .
the above @xmath0 defines a poisson structure on @xmath1 .
denote by @xmath155 , @xmath156 the linear and quadratic parts of @xmath0 respectively .
we need to show that @xmath157 } } = { { \left[{\pi^{(1)}},{\pi^{(1)}}\right ] } } + 2 { { \left[{\pi^{(1)}},{\pi^{(2)}}\right ] } } + { { \left[{\pi^{(2)}},{\pi^{(2)}}\right ] } } = 0\ ] ] where @xmath19 } } $ ] denotes the schouten - nijenhuis bracket .
the linear part of the equation , i.e. , @xmath158 } } = 0 $ ] is satisfied since @xmath159 is a lie algebra .
the cubic part @xmath160 } } = 0 $ ] is trivially satisfied since all vector fields @xmath161 commute with all @xmath74 .
the quadratic part @xmath162 } } = 0 $ ] is equivalent to @xmath163 for all linear @xmath164 . here
@xmath165 are the poisson brackets of the poisson structures @xmath155 and @xmath156 respectively . by trilinearity
, we can separately consider the following cases . * if at least two of the @xmath166 are functions of the @xmath74 s only , the expression trivially vanishes since the set of these functions is closed under @xmath167 , and furthermore @xmath168 . *
if @xmath169 the only contributing term is the second , i.e. , @xmath170 which vanishes by the cocycle property of @xmath143 .
* the remaining case is @xmath171 , leading to ( since @xmath172 ) @xmath173 which vanishes since the casimir element is poisson central .
knowing that @xmath0 defines a poisson structure , we can continue the proof of the main theorem .
this will be done in two lemmata .
the anomaly @xmath174 associated to @xmath0 as in theorem [ thm : cf ] is a nonzero multiple of @xmath175 the anomaly is given by the left graph , say @xmath59 , of figure [ fig : first_anomally ] .
it will be shown in the appendix that its weight @xmath56 is nonzero .
the associated bidifferential operator ( applied to functions @xmath54 , @xmath109 ) is given by @xmath176 here and in the following we adopt the convention that greek indices refer to a basis @xmath177 of @xmath159 , and are summed over @xmath178 if repeated .
in contrast the roman indices @xmath179 , @xmath180 label the basis @xmath74 of @xmath134 only and are summed over @xmath181 if repeated .
similarly , the roman indices @xmath182 , @xmath183 refer to the basis @xmath151 of @xmath139 only and are summed over @xmath184 . for the above @xmath0 and @xmath133 , eqns .
( [ equ : d1con ] ) and ( [ equ : con3 ] ) can not be solved simultaneously .
since @xmath134 is semisimple @xmath185 } } = { \mathfrak{{g}}}$ ] and eqn .
( [ equ : d1con ] ) implies that @xmath186 but then also @xmath187 since @xmath0 contains no part quadratic in the @xmath151 . hence eqn .
( [ equ : con3 ] ) becomes @xmath188 inserting the expression for @xmath0 we obtain from the second equation @xmath189 stating that @xmath143 is a coboundary .
hence , by choice of @xmath143 , this equation can not be solved .
thus the lemma and the first part of theorem [ thm : main ] is proven .
we finally turn to the more general case where @xmath190 , but still @xmath191 .
the construction in this case runs as above , but we make the special choice @xmath192 , where @xmath193 is the ( unique ) non - abelian two dimensional lie algebra .
its cohomology groups are computed in the appendix .
there is , up to normalization , only one non - trivial cocycle we can pick , namely @xmath194 as defined in eqn . ( [ equ : omegadef ] ) in the appendix .
we will call the resulting poisson structure @xmath0 .
the proof of the main theorem [ thm : main ] will then be finished by proving the following lemma . for any formal poisson structure @xmath6 extending
the @xmath0 constructed above , for which @xmath87 , eqns .
( [ equ : d1con ] ) and ( [ equ : con3 ] ) can not be solved simultaneously . by the previous proof
it will be sufficient to show that we can not pick @xmath124 and @xmath195 such that @xmath196 becomes ( the coefficients of ) an exact element of @xmath197 .
the @xmath38 part of the condition ( [ equ : piepsjacobi ] ) that @xmath6 is a poisson structure reads @xmath198 } } = 0.\ ] ] considering only the linear part we have @xmath199 } } = 0.\ ] ] where @xmath155 , @xmath200 are the linear parts of @xmath0 and @xmath124 respectively .
note that we used here that the constant part @xmath201 .
( [ equ : pi1cocycle ] ) means that @xmath200 defines a 2-cocyle of @xmath159 with values in the adjoint module .
equivalently , by projecting on the invariant submodules @xmath134 or @xmath139 , one has two 2-cocycles , with values in the @xmath159-modules @xmath202 and @xmath203 respectively . here
@xmath204 is always understood as equipped with the trivial module structure , and @xmath134 , @xmath139 with the adjoint structures .
the first 2-cocycle is irrelevant to us since it does not occur in ( [ equ : con3cohom ] ) ( since @xmath205 ) .
the second 2-cocycle defines a cohomology class , say @xmath206 $ ] , of @xmath207 eqn .
( [ equ : d1con ] ) means that @xmath208 is a cocycle and defines the class @xmath209\in h^0({\mathfrak{{g}}}\oplus{\mathfrak{{h } } } , { \mathfrak{{h}}}^*).\ ] ] the triviality of ( [ equ : con3cohom ] ) implies that their cup product would have to satisfy @xmath210\cup [ pi_1 ] \neq 0\ ] ] but from the formulas of knneth and whitehead it follows that @xmath211 furthermore , as shown in the appendix , @xmath212 . hence @xmath206=0 $ ] and eqn .
( [ equ : picoho ] ) can not be satisified .
thus the lemma and theorem [ thm : main ] is proven .
the lie algebra @xmath193 is defined as the vector space @xmath214 with the bracket @xmath215 } } = e_2\ ] ] where @xmath216 are the standard basis vectors .
[ lem : khom1 ] @xmath217 all other cohomology groups with values in @xmath204 vanish .
a representative of the equivalence class spanning @xmath218 is @xmath219 it is clear by antisymmetry that @xmath220 for @xmath221 and also that any 2-cochain is a cocycle .
there is only one 2-cochain ( up to a factor ) and it is a coboundary since @xmath222 satisfies @xmath223 } } ) = 1.\ ] ] finally , any 1-cocycle must vanish on @xmath224 } } = { { \mathbb{r}^ { { } } } } e_2 $ ] .
hence it is clear that the map @xmath225 defined in the lemma spans the space of 1-cocycles .
all cohomology groups of @xmath193 with values in @xmath193 vanish , i.e. , @xmath226 there is no central element in @xmath193 , hence @xmath227 .
the cocycle condition for some @xmath228 reads @xmath229 } } ) - { { \left[{l(e_1)},{e_2}\right ] } } - { { \left[{e_1},{l(e_2)}\right ] } } \\ & = & l(e_2)-e_2{l_2(e_2)}-e_2l_1(e_1 ) \\ & = & e_1l_1(e_2)-e_2l_1(e_1 ) \stackrel{!}{= } 0\end{aligned}\ ] ] where @xmath230
. hence @xmath231 . but then @xmath232 } } \ ] ] and hence @xmath225 is exact and @xmath233 .
that @xmath234 for @xmath235 follows as in the proof of the previous lemma .
we now consider the direct sum @xmath192 .
we denote the standard basis by @xmath64, ..
so , e.g. , @xmath237 } } = e_4 $ ] .
knneth s formula and the above lemmata tell us the following : * @xmath238 is spanned by @xmath239 with all other components vanishing .
* @xmath212 .
one still needs to show that the weight of the left graph in figure [ fig : first_anomally ] is nonzero .
we will actually compute the weights of all wheel graphs . instead of defining `` wheel graph ''
, we refer to figure [ fig : sixwheel ] , from which the definition should be clear . to compute the weights , we need the following result interesting in its own right .
let @xmath134 be a lie algebra and equip its dual space with the canonical poisson structure .
let @xmath7 be the evaluation map at zero and @xmath101 its quantization according to cattaneo and felder .
let @xmath240 be the map @xmath241 with @xmath242 the @xmath180-th bernoulli number
. becomes the duflo map when composed with @xmath243 .
] then @xmath244 we want to show recursively that @xmath252 if @xmath253 for @xmath254 . to do this
pick @xmath255 such that @xmath256 and set @xmath257 . exists .
the constants @xmath250 are weights of wheels and do not depend on the lie algebra . ]
then a straightforward calculation proves the claim . a typical anomaly wheel graph whose weight @xmath250 ( here @xmath258 ) is computed in theorem [ thm : wheelweights ] .
to get a cattaneo felder wheel graph as in the proof of theorem [ thm : wheelweights ] one simply removes the two lower edges .
the weight is the same as the kontsevich weight of the graph on the right.,title="fig : " ] a typical anomaly wheel graph whose weight @xmath250 ( here @xmath258 ) is computed in theorem [ thm : wheelweights ] . to get
a cattaneo felder wheel graph as in the proof of theorem [ thm : wheelweights ] one simply removes the two lower edges .
the weight is the same as the kontsevich weight of the graph on the right.,title="fig : " ] [ thm : wheelweights ] the @xmath250 computed in the preceding proof coincide with the weights of the anomaly wheel graphs as depicted in figure [ fig : sixwheel ] , up to possibly signs . in particular ,
the weight of the anomaly graph of figure [ fig : first_anomally ] is nonzero .
pick a cattaneo felder wheel graph @xmath59 .
see figure [ fig : sixwheel ] for an example .
let @xmath260 be the cattaneo felder configuration space as in section [ sec : cafmod ] . to divide out the scale invariance we will fix the central vertex of the wheel to lie on the unit quartercircle @xmath261\}$ ] as depicted in figure [ fig : quartcirclemove ] .
consider the closed form @xmath262 defined on @xmath260 as in section [ sec : cafmod ] , and compute @xmath263 with the help of stokes theorem .
there are several boundary strata contributing to the r.h.s .
they correspond to center- or non - center - vertices approaching the real axis , imaginary axis , or each other .
we divide the strata into the following eight types , each treated separately : 1 .
if all vertices together approach the real axis and each other , the result is 0 by a result of shoikhet @xcite .
if the central vertex approaches the real axis alone , the integral reduces to the integral of @xmath262 over @xmath260 , yielding the cattaneo felder weight @xmath264 .
if any subset of vertices approach the real axis and each other , except the two cases before , the result is zero by property 1 in section [ sec : cafpoint ] .
note that there is always an edge from the collapsing `` cluster '' to the remainder of the graph .
if all vertices together approach the imaginary axis and each other the result is the cattaneo felder anomaly weight @xmath265 by property 3 and the algorithm for computing @xmath265 .
if any proper subset of vertices appraoch the imaginary axis , the result is zero by property 1 .
if more than two vertices come close to each other inside the quadrant , the result is zero by a lemma of kontsevich .
if two non - center vertices come close to each other , the result is zero .
this is because both are linked to the center vertex and hence the boundary integrand will contain a wedge product of at least twice the same form , i.e. , 0
if any non - center vertex approaches the center vertex , the result is zero by similar reasoning as before .
note that automatically another vertex will be connected twice to the `` cluster '' of the two approaching vertices . | let @xmath0 be a poisson structure on @xmath1 vanishing at 0 .
it leads to a kontsevich type star product @xmath2 on @xmath3 $ ] .
we show that 1 .
the evaluation map at 0 @xmath4 can in general not be quantized to a character of @xmath5,\star_\pi)$ ] .
2 . a given poisson structure @xmath0 vanishing at zero can in general not be extended to a formal poisson structure @xmath6 also vanishing at zero , such that @xmath7 can be quantized to a character of @xmath5,\star_{\pi_\epsilon})$ ] .
we do not know whether the second claim remains true if one allows the higher order terms in @xmath8 to attain nonzero values at zero . |
since the beginning of the investigations on the mechanism of particle acceleration at shock fronts , estimating the maximum energy of the accelerated particles has been one of the main goals , in the perspective of assessing the ability of the mechanism to explain the observed energies in cosmic rays .
this issue has played a crucial role especially in the context of the so - called supernova paradigm for the origin of galactic cosmic rays .
in fact , it was understood very soon , in the case of supernova remnants shocks , that acceleration would not be efficient enough and provide sufficiently high energies unless magnetic scattering could be self - generated by the particles themselves through their motion in the upstream plasma .
the mechanism through which this generation of waves could occur was identified in the streaming instability ( @xcite ) and widely discussed later by @xcite . in these latter papers , in particular ,
the maximum energy was estimated for shocks in supernova remnants , already believed to be candidate sources of galactic cosmic rays .
the result of this investigation was however that the maximum energy , even in the presence of self - generated scattering agents , is of the order of @xmath1 gev , quite lower than the energy at which the knee is observed .
today we know that this energy is also lower than the maximum energy observed in the proton component by the kascade experiment ( see @xcite for a comprehensive review of experimental results in the knee region ) that seems to be as high as @xmath2 gev ( see @xcite for a recent review ) .
more recently the problem has been complicated even more by the understanding that particle acceleration can be rather efficient and break the _ test particle _ regime that old calculations were based upon .
this important result was obtained independently within completely different approaches , two - fluid models ( @xcite ) , kinetic models ( @xcite ) and numerical approaches , using both monte carlo and other simulation procedures ( @xcite ) .
the non - linear dynamical reaction of the accelerated particles introduces a positive reaction that results in a flattening of the high energy spectra , which are no longer power laws , thereby enhancing the role of particles at the highest energies .
moreover a gradient in the velocity of the upstream fluid is produced by the pressure of the accelerated particles .
the value of @xmath0 is important to determine the spectrum , and the spectrum determines the amplification of the magnetic field which in turn determines @xmath0 . the formalism of @xcite is not appropriate to describe this type of situation . in this paper
we describe the first calculation of the maximum momentum for cosmic ray modified shocks , including all the relevant non - linear effects .
the description of the acceleration process , with the dynamical reaction of cosmic rays and the magnetic field amplification taken into account , is based on the recent papers by @xcite .
the paper is organized as follows : in [ sec : method ] we describe the method that allows us to compute the acceleration time as a function of the particle momentum ; in [ sec : pmax ] we describe the procedure used to determine @xmath0 in the case of modified shocks and discuss the underlying assumptions ; in [ sec : results ] we present the results we find for the maximum achievable momentum in supernova remnant shocks ; finally a critical discussion of the method adopted and a summary of the main results is provided in [ sec : critical ] .
the full description of the calculations that allow us to compute the spectrum of the accelerated particles , the dynamics of the upstream and downstream plasmas and the strength of the amplified magnetic field due to cosmic rays streaming instability in the stationary regime can be found in the papers by @xcite . in this section
we present the main achievement of our work , which consists of a derivation of the acceleration time @xmath3 ( as a function of the momentum @xmath4 ) for particles accelerated at a shock with an arbitrary level of shock modification .
the expression we derive is a generalization of the well known formulae found in the context of quasi - linear theory ( @xcite ) and reduces to them when the non - linear shock modification is negligible . in general ,
the transport of the accelerated particles is described by the time dependent diffusion - convection equation : @xmath5 + \frac{1}{3 } \frac{du}{dx } p\frac{\partial f}{\partial p } + q(x , p ) , \label{eq : transport}\ ] ] where @xmath6 is the distribution function of the accelerated particles , @xmath7 is the spatial diffusion coefficient and @xmath8 is the injection term , taken in the form @xmath9 , with the assumption that injection takes place only at the shock surface on the downstream side ( although this assumption is not strictly needed ) .
the @xmath10axis goes from upstream infinity ( @xmath11 ) to downstream infinity ( @xmath12 ) and the shock is at @xmath13 .
we will use the index @xmath14 ( @xmath15 ) to refer to the region immediately upstream ( downstream ) of the shock .
the problem of determining the time scale associated with the change of the distribution function around the maximum momentum is impossible to face analytically in all its complexity , as implied by the non - linearity of the system : in general , @xmath16 , @xmath17 and @xmath18 all depend on time , and change on comparable time scales .
moreover , @xmath16 depends in a complicated way on all other quantities ( @xcite ) .
let us assume that we know an approximation of the solution @xmath16 in the form of a solution of the stationary problem for a given value of the maximum momentum @xmath0 , as can be calculated following @xcite .
this value of @xmath0 can be considered as a _ guess _ value to start with . in this case
the velocity profile @xmath19 can be considered as determined in first instance from the guess function @xmath16 as derived by the assumption of quasi - stationarity , using the calculations of @xcite . within this approximate framework
, we can introduce the laplace transform of the distribution function with respect to time : @xmath20 so that the transport equation becomes : @xmath21 + \frac{1}{3 } \frac{du}{dx } p\frac{\partial g}{\partial p } + \frac{1}{s } q(x , p ) , \label{eq : transportlaplace}\ ] ] integration of eq .
[ eq : transportlaplace ] in the vicinity of the shock ( namely between @xmath22 and @xmath23 ) leads to the following boundary condition at the shock : @xmath24_2 - \left[d\frac{\partial g}{\partial x}\right]_1 + \frac{1}{3 } ( u_2-u_1 ) p \frac{\partial g_0(s , p)}{\partial p } + \frac{1}{s } q_0(p ) = 0 , \label{eq : jump}\ ] ] where @xmath25 , @xmath26 and we have used the continuity of the distribution function ( at any time ) at the shock . we can find another condition on the distribution function by integrating the transport equation ( eq . [ eq : transportlaplace ] ) in the upstream section from @xmath11 to @xmath22 , so as to obtain : @xmath27_1 + \frac{1}{3 } p\frac{\partial}{\partial p } \int_{-\infty}^{0 ^ - } dx\ g(s , x , p ) \frac{du(x)}{dx } , \label{eq : jump2}\ ] ] where @xmath28 .
following @xcite we introduce the quantity @xmath29 which allows us to rewrite eq .
[ eq : jump2 ] as @xmath30_1 = s i(s , p ) + g_0 u_p \left ( 1 + \frac{1}{3 } \frac{d\ln u_p}{d \ln p } \right ) - \frac{1}{3 } p \frac{\partial g_0}{\partial p } ( u_1 - u_p ) .
\label{eq : d1}\ ] ] clearly in this equation @xmath31 depends on time through the functions @xmath16 and @xmath17 , which again reflects the non - linearity of the underlying equation .
the agreement was found to be worse when the shock is quasi - linear , namely for small modifications .
the reason for this is in the fact that eq .
[ eq : approx ] satisfies the jump condition in the stationary regime ( @xmath49 ) only when @xmath50 . a better approximation for @xmath42
is in fact of the form , @xmath51 @xmath46 \int_0^x dx'\frac{1}{d(x',p ) } \int_{-\infty}^{x ' } dx '' g(x '' ) \label{eq : approx1}\ ] ] which satisfies the jump condition exactly .
note that eq .
[ eq : approx1 ] differs from eq .
[ eq : approx ] only in that at given momentum the particle distribution is slightly more extended spatially . in the following we adopt eq .
[ eq : approx1 ] as an approximation of the spatial distribution of the accelerated particles . in evaluating the term @xmath52 in eq .
[ eq : g0 ] we retain only the first term of @xmath42 in eq .
[ eq : approx1 ] , since in the end we will be interested in the limit @xmath49 ( large times ) .
this means that eq .
[ eq : g0 ] becomes : @xmath53 where @xmath54 the solution of this equation is easily found to be : @xmath55 @xmath56 or , in a more compact form , @xmath57 , \label{eq : compact}\ ] ] where @xmath58 and @xmath59,\ ] ] with @xmath60 and @xmath61 . at this point
we follow the elegant procedure discussed by @xcite .
we need to calculate the inverse laplace transform @xmath62 we introduce the function @xmath63 one can easily show that the inverse laplace transform of the function @xmath64 is exactly the function @xmath65 as written in eq .
[ eq : compact ] .
it follows that the solution of our problem is in fact @xmath66 . in the approximate framework in which we are ,
the function @xmath67 , as in the linear case discussed by @xcite , represents the probability that a particle starting with energy @xmath68 has reached the momentum @xmath4 in a time @xmath69 ( @xmath67 is normalized to unity ) .
the mean time for the acceleration to momentum @xmath4 is therefore @xmath70 .
one can use the following obvious property of the function @xmath67 @xmath71 , \label{eq : prop}\ ] ] to determine @xmath72 , by deriving eq .
[ eq : prop ] with respect to @xmath73 and calculating it at @xmath74 . from this follows that : @xmath75_{s=0 } = \frac{3 r_{tot}}{u_0 ^ 2}\ \int_{p_{inj}}^p \frac{dp'}{p ' } \left\ { \frac{r_{tot } d_2(p')}{r_{tot}u_p(p')-1 } + \frac{u_0 \lambda(p ' ) } { r_{tot } u_p(p ' ) -1 } \right\}. \label{eq : tacc}\ ] ] for an unmodified shock , with spatially constant fluid velocities upstream and downstream , one has that @xmath76 and the previous formula easily reduces to the well known : @xmath77 , \label{eq : lagage}\ ] ] if the diffusion coefficient is in the form of bohm diffusion . for an arbitrary form of the diffusion coefficient the expression in eq .
[ eq : lagage ] is correct within a factor of order unity .
in general the value of @xmath0 is determined by the competition between the acceleration time and the shortest among the finite age of the source , the time scale for energy losses , and the time scale for escape of particles with momentum @xmath0 from upstream infinity .
here we concentrate on the self - consistent calculation of the maximum momentum in those cases when it is determined by the finite age of the source ( this is expected to be the case for supernova remnants ) .
in such a context each value of @xmath0 therefore reflects a different age @xmath78 of the system . in this sense ,
our previous work ( @xcite ) was based on the implicit assumption that a quasi - stationary regime could be reached on time scales shorter than @xmath78 , but no specific recipe for the calculation of @xmath0 was adopted . in other work on non - linear diffusive particle acceleration at shock waves ,
@xmath0 was either fixed or determined according with a simple recipe based on test - particle theory .
our determination of the maximum momentum proceeds through 2 different steps : 1 ) we first determine the acceleration time @xmath3 ( eq . [ eq : tacc ] ) for a given value of @xmath79 ( at the first step this is just a guess value @xmath80 ) ; 2 ) we compare @xmath81 with the age of the system @xmath78 and determine an updated value of @xmath0 , say @xmath82 .
the procedure is iterated until @xmath83 , where @xmath0 is now the actual maximum momentum for our problem .
it is crucial to keep in mind that due to the non - linearity of the system , changing @xmath0 causes the all system to change ( velocity profile , spectrum of accelerated particles , efficiency of particle acceleration and also the diffusion coefficient , if the scattering is due to self - generated waves ) .
the results obtained by applying this procedure are illustrated in the next section .
following the method outlined in [ sec : method ] we computed the acceleration time as a function of the particle momentum @xmath4 for a system with @xmath84 , under different assumptions on the shock mach number at infinity @xmath85 and on the diffusion properties of the system .
the results of these calculations are summarized in fig .
[ fig : tacc ] . we assumed a velocity of the shock wave of @xmath86 and considered values of @xmath85 ranging from @xmath87 ( unmodified shocks ) to @xmath88 ( already in the asymptotic regime for strongly modified shocks ) .
the injection momentum was taken to be @xmath89 with @xmath90 corresponding to the maxwellian temperature @xmath91 and @xmath92 .
we also related the injection efficiency @xmath93 to the compression ratio at the subshock through the recipe proposed by @xcite and already adopted by @xcite . in the top panel of fig .
[ fig : tacc ] we plot the acceleration time up to a momentum @xmath4 for @xmath94 , namely with self - generated scattering , with @xmath95 the result of quasi - linear theory for scattering induced by self - generated alfvn waves . here
@xmath96 and @xmath97 are the particle velocity and larmor radius ( in the background field @xmath98 ) respectively , and @xmath99 is the dimensionless wave energy density per unit logarithmic band - width ( see e.g. @xcite for details ) . in the bottom panel of fig .
[ fig : tacc ] we show the acceleration time for @xmath100 , with @xmath101 referring to bohm - like diffusion in the amplified magnetic field @xmath102 found by integrating @xmath103 in @xmath104-space : @xmath105 . in all of our calculations
the background magnetic field was taken to be @xmath106 .
the maximum momentum obtained from our iterative procedure is seen as the intersection with the horizontal line @xmath107 years , assumed to be the age of the system .
each of the curves in fig .
[ fig : tacc ] was obtained by following the procedure outlined in [ sec : pmax].the first thing to notice by comparing the two different panels in fig .
[ fig : tacc ] is that the maximum achievable momentum , for a given mach number , is roughly the same ( within a factor of a few ( 2 - 4 ) at most ) for the two diffusion models adopted .
this observation is very important given the fact that the diffusion coefficient is basically unknown and not easily obtained from observations .
still the comparison between the two panels allows to highlight what the main effect of the different diffusion models is : in the case of strongly modified shocks the spectra become very flat at high energies as discussed in detail elsewhere ( e.g. @xcite ) and most of the power of the alfvn waves excited by the streaming cosmic rays is deposited in long wavelength waves ( the ones that are resonant with high energy particles ) , with a relatively smaller fraction going into waves that are able to efficiently scatter the low energy particles .
this is the reason for the behaviour of @xmath3 observed in the upper panel , which may seem surprising at first sight for the flatness of the curves referring to @xmath108 . by comparing the two panels ,
it is clear that things change drastically , with the dependence of @xmath109 on @xmath4 becoming much stronger , if the energy of the waves is redistributed uniformly between different wavelengths , which is what one is effectively assuming by adopting a bohm - like diffusion coefficient in the amplified magnetic field .
finally , let us discuss the behaviour of the maximum achievable momentum as a function of the mach number .
this is qualitatively the same for both models of diffusion : @xmath0 initially increases quickly with the mach number as a result of field amplification in modified shocks and then decreases somewhat , although slowly , when @xmath85 is further increased .
this latter fact is clearly related to the more strongly modified velocity profile that leads to lower and lower values of the flow velocity in the vicinity of the shock and as a consequence increasingly longer residence times upstream .
the dependence of the maximum achievable momentum on the shock mach number is illustrated more clearly in fig .
[ fig : pmax ] , where we plot @xmath0 as a function of @xmath85 for four different models of diffusion .
the plot refers again to a system with @xmath110 and all other values of the relevant parameters are the same as above , including the background magnetic field strength and the recipe for @xmath68 .
the solid curve refers to @xmath111 as described above , the dashed curve is for @xmath112 , the dotted curve is for @xmath113 ( namely the diffusion coefficient equals the bohm diffusion calculated in the amplified magnetic field at the shock surface , without spatial variation ) and finally the dot - dashed curve refers to bohm diffusion in the background magnetic field .
the first aspect that clearly appears from the figure is the enhancement of @xmath0 that arises from magnetic field amplification . for high mach number shocks the magnetic field strength at the shock
is amplified by a factor of order 25 and this implies an enhancement by a factor 10 - 30 ( depending on the diffusion model adopted ) of the maximum achievable energy ( comparison of the solid , dashed or dotted curves with the dot - dashed one ) . on the other hand
, @xmath0 does not strongly depend on the detailed diffusion model adopted , in terms of the turbulence spectrum , with the difference between bohm diffusion and resonant scattering with self - generated alfvn turbulence ( comparison between the solid and dashed curves ) being a factor of 2 at most .
a few remarks are worth being made .
note that the values of the parameters adopted throughout our calculations are reminiscent of those appropriate to the case of young supernova remnants ( snrs ) .
this is because we wished to address here the issue of the maximum energy that can be achieved in this class of sources , as candidate sources of galactic cosmic rays .
moreover there seems to be now solid evidence for amplified magnetic fields in snrs ( @xcite ) , so that there is observational support for the idea that cosmic rays might be efficiently accelerated at snrs shocks .
the shock modification is a natural consequence of the effectiveness of these shocks as cosmic ray accelerators .
the maximum energy that the accelerated particles can achieve in the context of modified shocks is always less than one would naively guess by using a @xcite type of estimate with the upstream velocity set equal to @xmath114 and the magnetic field evaluated at the shock surface .
this is due to the combination of two effects .
on one side , the net effect of the shock modification is to slow down the upstream fluid ( precursor ) , so that the effective velocity _ felt _ by the particles of given momentum is always lower than @xmath114 , thereby making the acceleration process slower .
this is illustrated effectively through the dotted line in fig .
[ fig : pmax ] , where the diffusion coefficient is taken as bohm - like calculated with the magnetic field at the shock , assumed constant in the precursor . in this case
the value of @xmath0 is always lower than @xmath115 gev / c for mach number @xmath116 , while in the absence of shock modification one would get @xmath117 gev / c .
this factor @xmath118 is solely due to the existence of a precursor .
there is however a second ( less important ) effect that causes a further decrease of @xmath0 : the spatial distribution of the accelerated particles in the precursor also determines a spatial profile of the strength of the magnetic field associated with the self - generated alfvn waves , so that the effective diffusion coefficient is also larger than @xmath119 .
this effect is visible when comparing the dashed line with the dotted line in fig .
[ fig : pmax ] . the combination of the two effects leads to an overall decrease of @xmath0 by a factor @xmath120 compared with the naive expectation _ a la _ @xcite , more evident for higher values of the mach number .
in this section we summarize the reasons for the need of the type of calculations presented here and we discuss some shortcomings of the proposed mathematical approach . to our knowledge
this represents the first attempt to determine the maximum momentum of the particles accelerated at a cosmic ray modified shock , where the nonlinear dynamical reaction of the accelerated particles changes the shocked fluid , the particles self - generate their own scattering centers , and the modified system in turn determines the spectral shape and energy content of the accelerated particles distribution .
the creation of a shock precursor in the upstream region and the amplification of the background magnetic field through streaming instability are the main reasons to expect that the maximum achievable momentum may be appreciably different from the one calculated in linear theory for the same value of the shock mach number .
the precursor , due to the gradual slowing down of the upstream plasma would by itself be expected to slightly decrease the maximum momentum : this effect can in fact be seen in fig .
[ fig : pmax ] , where the dash - dotted line shows the maximum momentum as a function of the mach number when the background field is not amplified . the reduced velocity in the upstream plasma slows down the process of return to the shock for the particles in the upstream section of the fluid . in the presence of magnetic field amplification
the maximum momentum is expected to increase .
this trend continues up to mach numbers for which the pressure in the form of accelerated particles can not increase any more ( saturation ) , so that the self - generated magnetic field also saturates . on the other hand the compression factor ( shock modification ) keeps increasing , so that the corresponding @xmath0 in fig .
[ fig : pmax ] shows a slightly decreasing trend ( roughly the same that one obtains in the presence of shock modification but without field amplification , dash - dotted line ) .
one should be careful not to mistake this trend for a temporal trend .
the behaviour of the maximum momentum is to be interpreted only as a trend as a function of the fluid mach number at upstream infinity . in a more realistic situation the mach number
can also change with time , which might introduce additional effects , not taken into account here . on the other hand ,
if the conclusions of @xcite remain valid for this more complex scenario , one should expect that most of the acceleration of cosmic rays occurs between the end of the free expansion phase and the beginning of the sedov phase .
this is the phase we concentrated our attention upon , so that the assumption of temporally constant mach number adopted here is justified .
the most important equation derived in this work is eq .
[ eq : tacc ] for the acceleration time in the presence of shock modification .
it is important to realize that although the calculations shown here have been carried out by adopting the theory of @xcite for the evaluation of the modification and field amplification , eq .
[ eq : tacc ] can be applied in a very general way .
in fact all the quantities involved in the evaluation of eq .
[ eq : tacc ] can be determined using very different approaches , for instance monte carlo simulations , while eq .
[ eq : tacc ] can be adopted to estimate the _ instantaneous _ value of @xmath0 . moreover , since eq .
[ eq : tacc ] directly contains the diffusion coefficients and the dynamical quantities ( e.g. fluid velocity ) , it can be applied independently of the specific mechanism of wave production . if the non - resonant wave generation mechanism recently proposed by @xcite were to be responsible for the wave production and particle scattering , this would change the mathematical form of the diffusion coefficient , and possibly the dynamical reaction of the plasma ( if the amplified field reaches a dynamically important value ) , but eq .
[ eq : tacc ] could still be applied unchanged .
the most relevant shortcoming of the calculation illustrated here is in the fact that we used the assumption of quasi - stationarity in the upstream fluid ( not downstream ) .
this assumption is fundamental to allow us to apply the method of laplace transform , which can be used only for linear differential equations . at the same time
quasi - stationarity is also required in order to use our previous calculations ( @xcite ) for the shock modification and field amplification .
these are actually two different aspects of the same problem of this approach , which however can not be avoided if an analytical attempt has to be made to calculate the maximum momentum in these complex situations .
it should however be stressed that the shortcomings just discussed also apply to the approach of @xcite : their calculations are only valid for linear shocks ( @xmath121 ) so that the velocity and density profiles at the shock are spatially constant and given .
on the other hand the diffusion coefficient , which is self - generated in their approach , is time - dependent because the particle distribution function is time - dependent . despite this
, the authors fix the diffusion coefficient ( and the distribution function of the accelerated particles ) and proceed by applying the method of the laplace transform . in this sense
our approach should be considered as the extension of the work of @xcite to the case in which the dynamical reaction of the accelerated particles is not neglected .
the common finding of several independent approaches to particle acceleration at shocks is that shock modification manifests itself in a prominent way , as a consequence of efficient acceleration , therefore our extension seems particularly important at this time .
\1 ) the maximum momentum of accelerated particles can be as high as @xmath115 gev / c for protons , when all effects are taken into account .
the approach of @xcite would give @xmath117 gev / c when the magnetic field is assumed spatially constant to its maximum value at the shock surface . the same approach for @xmath122 would reproduce the well known value @xmath123 gev / c for @xmath124 and @xmath125 .
\2 ) the net effect of the shock modification is to create a precursor in both the fluid velocity and magnetic field strength .
the former decreases while moving towards the shock while the latter increases .
both phenomena contribute to lowering @xmath0 with respect to the value @xmath126 gev / c .
\4 ) the expression provided in eq .
[ eq : tacc ] is the only expression presently available that allows one to calculate the acceleration time for cosmic ray modified shock waves .
this expression is very general and can be applied independently of the specific treatment of the shock modification and reduces to the well known formula for weakly modified shocks . | particle acceleration at non - relativistic shocks can be very efficient , leading to the appearance of non - linear effects due to the dynamical reaction of the accelerated particles on the shock structure and to the non - linear amplification of the magnetic field in the shock vicinity .
the value of the maximum momentum @xmath0 in these circumstances can not be estimated using the classical results obtained within the framework of test particle approaches .
we provide here the first attempt at estimating @xmath0 in the cosmic ray modified regime , taking into account the non - linear effects mentioned above .
[ firstpage ] acceleration of particles - shock waves |
optoelectronic oscillators ( oeos ) combine a nonlinear modulation of laser light with optical storage to generate ultra - pure microwaves for lightwave telecommunication and radar applications @xcite
. their principal specificity is their extremely low phase noise , which can be as low as @xmath0 dbrad@xmath1/hz at @xmath2 khz from a @xmath2 ghz carrier . despite some interesting preliminary investigations ,
the theoretical determination of phase noise in oeos is still a partially unsolved problem .
the qualitative features of this phase noise spectrum can be recovered using some heuristical guidelines or rough approximations , but however , a rigorous theoretical background is still lacking .
there are several reasons which can explain that absence of theoretical background .
a first reason is that before refs .
@xcite , there was no time domain model to describe such systems , so that stochastic analysis could not be used to perform the phase noise study .
moreover , unlike most of oscillators , the oeo is a delay - line oscillator , and very few had been done to study the effect phase noise on time - delay induced limit - cycles .
finally , the oeo is subjected to multiple noise sources , which are sometimes non - white , like the flicker ( also referred to as @xmath3 " ) noise which is predominant around the microwave carrier .
the objective of this work is to propose a theoretical study where all these features are taken into account .
the plan of the article is the following . in section [ phasediffusion ] ,
we present the phase diffusion approach in autonomous systems .
it is a brief review where the fundamental concepts of phase diffusion are recalled , and where some important earlier contributions are highlighted .
then , we derive in section [ phasesdde ] a stochastic delay - differential equation for the phase noise study .
we show that for our purpose , the global interaction of noise with the system can be decomposed into two contributions , namely an additive and a multiplicative noise contribution .
section [ belowth ] is devoted to the study of the noise spectrum below threshold .
it will appear that the spectrum below threshold will not only be important to validate the stochastic model , but also that it enables an accurate calibration of additive noise . in section [ aboveth ] , we address the problem of phase noise when there is a microwave output using fourier analysis , and we show that it is possible to have an accurate image of the phase noise spectrum in all frequency ranges . the last section concludes the article .
for an ideal ( noise - free ) oscillator , the fourier spectrum is a collection of dirac peaks , standing for the fundamental frequency and its harmonics .
the effect of amplitude white noise is to add a flat background , while the peaks do keep their zero linewidth ; it is the effect of phase noise to widen the linewidth of these peaks .
some pioneering papers on the topic of phase noise in autonomous oscillators using stochastic calculus had been published forty years ago @xcite . in particular , it was demonstrated that a general framework to study the problem of phase noise in a self - sustained oscillator could be built using some minimalist assumptions .
the first point is that a strong nonlinearity is an essential necessity in oscillators , in the sense that nonlinearity can not be regarded as small because it controls the operating level of the oscillator .
the second important point is that the phase is only _ neutrally _ stable , so that quasilinear methods which assume that fluctuations from some operating point are small ( linearization techniques ) can not be applied _ directly_. the phase is neutrally stable as a consequence of the phase - invariance of autonomous oscillators . in other words , limit - cycles are stable against amplitude perturbations , while there is no mechanism able to stabilize the phase to a given value : hence , phase perturbations are undamped , but they do not diverge exponentially , though . in a noise free oscillator ,
the stroboscopic " state point on the limit - cycle is immobile , but in the presence of noise , it moves randomly along the limit - cycle : in other words , the phase of the oscillator undergoes a _ diffusion _ process , in all points similar to a one - dimensional brownian motion . in the most simple case ,
the random fluctuations of the phase are referred to as a _ wiener process _ , obeying an equation of the kind @xmath4 , where @xmath5 is a gaussian white noise with autocorrelation @xmath6 , while @xmath7 is a parameter referred to as the _ diffusion constant_. it can be demonstrated that the phase variance diverges linearly as @xmath8 , and the single - side band phase noise spectrum ( in dbc / hz ) explicitly reads @xmath9 $ ] , so that @xmath7 is the unique parameter characterizing all the statistical and spectral features of phase fluctuations . experimental set - up.,width=226 ] on the base of earlier works by lax @xcite and krtner @xcite , demir , mehrota and roychowdhury have proposed few years ago a unifying theory of phase noise in self - sustained oscillators subjected to white noise sources @xcite .
their approach , which had later been extended by demir to the case colored noise sources @xcite , relies on stochastic calculus .
the principal point of their contribution was the introduction of a decomposition of phase and amplitude noise through a projection onto the periodic time - varying eigenvectors ( the so called _ floquet eigenvectors _ ; also see ref .
@xcite ) , and they proved that it provides the correct solution to the problem .
demir _ et al _ have shown that if the sources of noise are gaussian and white , the phase noise around the fundamental peak ( and its harmonics ) has a lorentzian lineshape , and therefore is fully determined by an effective " diffusion constant @xmath10 which is the unique parameter needed for the phase noise determination . however
, if the demir _ et al _ theory has the great and essential advantage of mathematical rigorousness , its principal drawback is that exactitude is obtained at the expense of simplicity : the calculation of @xmath10 is very complex , as it requires an accurate determination of all the time - varying eigenvectors related to the autonomous flow . in general this task can only be performed numerically using quite complicated algorithms , and this lack of flexibility explains why this method is scarcely used in the phase noise studies available in the literature .
the key challenge for the study of phase noise in oeo would be provide an accurate description of the phase noise spectrum , while avoiding the determination of floquet eigenvectors , which is an extremely complicated task in delayed system .
the oeo under study is organized in a single - loop architecture as depicted in fig .
[ setup ] .
the oscillation loop consists of : @xmath11 a wideband integrated optics linbo@xmath12 mach - zehnder ( mz ) modulator , seeded by a continuous - wave semiconductor laser of optical power @xmath13 ; the modulator is characterized by a half - wave voltage @xmath14 v. @xmath15 a thermalized @xmath16 km fiber performing a time delay of @xmath17 @xmath18 on the microwave signal carried by the optical beam ; the corresponding free spectral range is @xmath19 khz . @xmath20 a fast photodiode with a conversion factor @xmath21 .
@xmath22 a narrow band microwave radio - frequency ( rf ) filter , of central frequency @xmath23 ghz , and @xmath24 db bandwidth of @xmath25 mhz ; @xmath26 a microwave amplifier with gain @xmath27 . @xmath28 a variable attenuator , in order to scan the gain .
@xmath29 all optical and electrical losses are gathered in a single attenuation factor @xmath30 .
the dynamics of the microwave oscillation can therefore be described in terms of the dimensionless variable @xmath31 whose dynamics obeys @xcite @xmath32 \ , , \label{oeo_original}\end{aligned}\ ] ] where @xmath33 is the normalized loop gain , @xmath34 is the mach - zehnder offset phase , while @xmath35 and @xmath36 are the characteristic timescale parameters of the bandpass filter .
since we are interested by single - mode microwave oscillations , the solution of eq .
( [ oeo_original ] ) can be expressed under the form @xmath37 where @xmath38 $ ] is the slowly varying amplitude of the microwave @xmath39 .
we can significantly simplify the right - hand side term of eq .
( [ oeo_original ] ) because the cosine of a sinusoidal function of frequency @xmath40 can be fourier - expanded in harmonics of @xmath40 . in other words , since @xmath39 is nearly sinusoidal around @xmath40 , then the fourier spectrum of @xmath41 $ ] will be sharply distributed around the harmonics of @xmath40 using the relationship @xmath42/2 $ ] and the jacobi - anger expansion @xmath43 where @xmath44 is the @xmath45-th order bessel function of the first kind .
hence , since the filter of the feedback loop is narrowly resonant around @xmath40 , it can be demonstrated that discarding all the spectral components of the signal except the fundamental is an excellent approximation , so that eq .
( [ oeo_original ] ) can be rewritten as @xmath46 \ , \cos [ \omega_0 ( t - t)+ \psi(t - t ) ] \ , .
\label{oeo_j1}\end{aligned}\ ] ] in order to include noise effects in this equation , we will consider two main noise contributions in this system .
the first contribution is an _ additive noise _ , corresponding to random environmental and internal fluctuations which are uncorrelated from the eventual existence of a microwave signal .
the effect of this noise can be accounted for by addition as a langevin forcing term , to be added in the right - hand side of eq .
( [ oeo_j1 ] ) .
this additive noise can be assumed to be spectrally white , and since we are interested by its intensity around the carrier frequency @xmath40 , it can be explicitly written as @xmath47 where @xmath48 is a complex gaussian white noise , whose correlation is @xmath49 , so that the corresponding power density spectrum is @xmath50 .
the second contribution is a _ multiplicative noise _ due to a noisy loop gain .
effectively , the normalized gain parameter explicitly reads @xmath51 \ , .
\label{gamma_explicit}\end{aligned}\ ] ] if all the parameters of the system are noisy ( i.e. , we replace @xmath30 by @xmath52 , @xmath21 by @xmath53 , etc . ) , then the gain @xmath54 may be replaced in eq .
( [ oeo_j1 ] ) by @xmath55 , where the @xmath56 is the overall gain fluctuation .
we therefore introduce the dimensionless multiplicative noise @xmath57 which is in fact the relative gain fluctuation . in the oeo configuration
, we have @xmath58 .
this noise is in general spectrally complex , as it is the sum of noise contributions which are very different ( noise from the photodetector , from the amplifier , etc . ) . in agreement with the usual noise spectrum of amplifiers and photodetectors
, we will here consider that this multiplicative noise is flicker ( i.e. , varies as @xmath3 ) near the carrier , and white above a certain knee - value .
we therefore assume the following empirical noise power density @xmath59 where @xmath60 is the low corner frequency of the flicker noise , while @xmath61 is the high corner frequency .
more precisely , we consider that the noise is white below @xmath62 and above @xmath63 , while it remains flicker in between .
typically , we may consider @xmath64 hz and @xmath65 khz , so that the flicker noise is extended over a frequency span of more than 4 orders of magnitude . + to avoid the integral term of eq .
( [ oeo_j1 ] ) which is complicated to manage analytically , it is mathematically convenient to use the intermediate integral variable @xmath66 which is also nearly sinusoidal with a zero mean value . using eqs .
( [ oeo_j1 ] ) , ( [ decomp_noises ] ) and ( [ multiplicative_noise ] ) , it can be shown that the slowly - varying amplitude @xmath67 obeys the stochastic equation @xmath68\ , \ ,
\left [ \frac{1}{2}e^{i \omega_0 ( t - t)}e^{i\psi_t } + { \rm c.c . } \right ] \nonumber \\ & & \times { \rm j_1}[2 |\dot{{\cal b}}_t+i \omega_0 { \cal b}_t|]+ 2 \delta \omega \left [ \frac{1}{2}\zeta_a(t)e^{i \omega_0 t } + { \rm c.c . } \right ] \ , , \nonumber
\\ \label{eq_calb_stoch_1}\end{aligned}\ ] ] where c.c .
stands for the complex conjugate of the preceding term .
we can assume @xmath69 and @xmath70 ; the relationship @xmath71 therefore gives @xmath72 , so that we can finally derive from eq .
( [ eq_calb_stoch_1 ] ) the following stochastic equation for the slowly varying envelope @xmath73 @xmath74 \ , { \rm jc_1}[2 |{\cal a}_t| ] { \cal a}_t \nonumber \\ & & + \mu e^{i\vartheta } \zeta_a(t)\ , , \label{eqt_a_stoch}\end{aligned}\ ] ] where @xmath75 is the first order _ bessel cardinal _ function of the first kind .
the phase condition has been set to @xmath76 , so that the dynamics of interest is restricted to the case @xmath77 .
the key parameters of this equation are @xmath78\ , , \label{def_param_sdde}\end{aligned}\ ] ] where @xmath79 is the quality factor of the rf filter . since @xmath80
, we may simply consider that @xmath81 and @xmath82 .
the complex term @xmath83 is a kind of `` filter operator '' , which can be simply equated to the half - bandwidth @xmath84 when the @xmath85-factor of the filter is sufficiently high , as it was done in ref .
it is also noteworthy that in the complex amplitude equation ( [ eqt_a_stoch ] ) , the initial multiplicative noise remains a _
variable , while the additive noise becomes _
complex_. we had recently shown , in agreement with the experiment , that the oeo has three fundamental regimes @xcite . for @xmath86 , the system does not oscillate and the trivial fixed point is stable ; for @xmath87 , the system sustains a pure microwave oscillation , with a constant amplitude and frequency ; and at last , for @xmath88 , the system enters into a regime where the amplitude of the microwave is unstable , and turns to be nonlinearly modulated .
we can consider that this phenomenology is still correct as long as @xmath80 .
with the aid of the stochastic delay - differential equation ruling the dynamics of @xmath89 , we may now derive analytically the power spectrum density of the oscillator , below and above threshold .
however , it should be stressed that in _ all _ cases , stochastic variables should be manipulated with respect to the rules of stochastic calculus when an integral / differential transformation is applied to them .
in general , no interest is paid to the study of the noise power density spectrum below threshold in oeos .
this lack of interest can be explained by the fact that there is no oscillation in this regime , and the system randomly fluctuates around the trivial equilibrium .
however , as we will further see , this regime is particularly interesting because it enables to understand how the noise interacts with the system . from the stability theory of delay - differential equations with complex coefficients , the deterministic solution of eq .
( [ eqt_a_stoch ] ) below threshold is the trivial fixed point @xmath91 .
after linearization around this solution , eq .
( [ eqt_a_stoch ] ) can simply be rewritten as @xmath92 where we have used @xmath93 .
this equation indicates that the multiplicative noise has no significative influence below threshold , because the product @xmath94 is a second - order term .
therefore , the noise power below threshold is _ essentially _ determined by additive noise .
equation ( [ a_bel_th ] ) is linear with constant coefficients : hence , the power density spectrum can directly be obtained as @xmath95 one can determine the total output power below threshold due to the white noise fluctuations in the system through the formula @xmath96 where @xmath97 is the output impedance ( in our case , @xmath98 @xmath99 ) .
the dimensionless power @xmath100 can not be calculated analytically for @xmath101 : it can nevertheless be determined either by numerical simulation of eq .
( [ a_bel_th ] ) , or through a numerical computation of the integral @xmath102 , where @xmath103 is given by eq .
( [ a_bel_th_four_2 ] ) .
however , in the open - loop configuration ( @xmath104 ) , the noisy output power can be analytically determined as @xmath105 through the use the fourier integral , or using fundamental results from stochastic calculus since eq .
( [ a_bel_th ] ) degenerates to the well - known orstein - uhlenbeck equation .
therefore , knowing the bandwidth @xmath106 of the rf filter and the half - wave voltage @xmath107 of the mz interferometer , an open - loop measurement of the output power can directly give an experimental a value for the white noise power density @xmath108 through eq .
( [ noise_pow_gam0 ] ) . in our system
, we have experimentally measured @xmath109 nw ( or @xmath110 dbm ) , which corresponds to @xmath111 rad@xmath1/hz . this value for the power
can also be obtained by other means [ see appendix a ] . the curve displaying the power variation as a function of the normalized gain under threshold is shown in fig .
[ noise_power_profile ] , and there is an excellent agreement between the experimental data and our analytical formula of eq .
( [ noise_pow_true ] ) .
it may be interesting to note that the noise power apparently diverges at @xmath112 .
in fact , one should not forget that this result is obtained using eq .
( [ a_bel_th ] ) , which is only valid for @xmath113 . when @xmath114 , the amplitude of @xmath89 increases and the higher order terms of the bessel cardinal function are not negligible anymore , so that the eqs .
( [ a_bel_th ] ) and ( [ a_bel_th_four_2 ] ) are no more valid .
hence , divergence of the noise power is prevented by the nonlinear terms of eq .
( [ eqt_a_stoch ] ) which become predominant in a very narrow range just below the threshold .
a noteworthy study on this topic of noisy oscillators near threshold is ref .
@xcite .
it is also noteworthy that for @xmath104 , the noise spectrum follows the spectral shape of the rf filter .
however , when @xmath54 is increased ( still below threshold ) , a first qualitative difference emerges , since the spectrum still follows the spectral shape of the filter , but its fine structure is composed by a collection of peaks which are the signature of microwave ring - cavity modes , as it can be seen in figs .
[ spec_bt_num ] and [ spec_bt_exp ] .
variation of the rf noise output power @xmath115 as a function of the normalized gain , under threshold .
the solid line is the theoretical prediction of eq .
( [ noise_pow_true ] ) with @xmath116 rad@xmath1/hz , and the symbols represent the experimentally measured data .
the gain was varied through attenuation in the electric branch of the loop.,width=302 ]
above threshold , the amplitude of the microwave obeys the nonlinear algebraic equation @xmath118=1/(2 \gamma)$ ] .
linearizing eq .
( [ eqt_a_stoch ] ) around this solution yields the following equation @xmath119 \ , { \cal a}_t + \mu e^{i\vartheta } \zeta_a(t)\ , .
\label{a_above_th}\end{aligned}\ ] ] we should now look for an equation for the phase @xmath120 in order to find its power density spectrum @xmath121 . using the it rules of stochastic calculus [ see appendix b ] , we derive the following time - domain equation for the phase dynamics @xmath122 where @xmath123 is a _ real _ gaussian white noise of correlation
@xmath124 ( same variance as @xmath125 ) .
we can use eq .
( [ eq_psi_above_th ] ) to obtain the fourier spectrum @xmath126 of the phase @xmath127 , and then its power density spectrum following @xmath128 } \right|^2 \ , .
\label{eq_psi_ab_th_four}\end{aligned}\ ] ] note that here , the influence of gain on phase noise spectrum is not explicit anymore : it is implicitly contained in @xmath129 .
figure [ spec_at ] displays the phase noise spectrum explicitly expressed by eq .
( [ eq_psi_ab_th_four ] ) , and we can now analyze how the spectrum behaves according to the various frequency ranges .
theoretical power density spectrum @xmath130 of the microwave noise signal below threshold , with @xmath131 rad@xmath1/hz and @xmath132 , using eq .
( [ a_bel_th_four_2 ] ) .
the semi - logarithmic scale is adopted because it enables to see at the same time the fine structure of regularly spaced ring - cavity peaks , and the global variation shaped by the rf filter bandwidth .
this spectrum is divided into two areas : a quasi - flat area within bandwidth , and a @xmath133 db / dec decrease outside the bandwidth .
, width=302 ] an experimental power density spectra of the microwave noise below threshold .
the spectrum has been scaled to its maximum ( reference at @xmath134 db ) .
( a ) spectrum in a @xmath135 mhz window , showing how the noisy power density is profiled by the rf filter .
( b ) zoom - in with near the @xmath2 ghz central frequency in a @xmath136 khz window , showing the noisy ring - cavity peaks.,title="fig:",width=302 ] an experimental power density spectra of the microwave noise below threshold .
the spectrum has been scaled to its maximum ( reference at @xmath134 db ) .
( a ) spectrum in a @xmath135 mhz window , showing how the noisy power density is profiled by the rf filter .
( b ) zoom - in with near the @xmath2 ghz central frequency in a @xmath136 khz window , showing the noisy ring - cavity peaks.,title="fig:",width=302 ] here , we consider the spectrum for frequencies which are relatively close to the carrier ( with @xmath138 , however ) .
qualitatively , this corresponds to the frequencies that are much smaller than the high corner value @xmath61 of the multiplicative flicker noise . in this region ,
flicker noise is stronger than white noise , so that @xmath139 . on the other hand , we can also consider that @xmath140 .
therefore , taking into account the fact that @xmath141 , eq .
( [ eq_psi_ab_th_four ] ) can be simplified into @xmath142 some remarks can be made at this stage .
first , the lesson effect is here very explicit : the phase noise spectrum decreases as @xmath143 due to the @xmath144 flicker noise @xcite .
secondly , the phase noise is inversely proportional to @xmath145 .
thirdly , the phase noise is practically independent of the microwave amplitude @xmath129 , as long as multiplicative noise is stronger than additive noise near the carrier .
finally , the phase noise decreases as @xmath146 , therefore justifying the need for very long delay - lines to reduce phase noise close to the carrier .
this dependence was also recovered analytically by yao and maleki , using another theoretical approach @xcite .
hence , in first approximation there are three ways to reduce phase noise close to the carrier : reduce the power @xmath147 of the flicker noise , increase the delay @xmath148 or increase the @xmath85 factor of the rf filter . the frequencies of concern are here those which are well within the bandwidth , but not too close from the carrier .
this range typically lies between @xmath150 khz and few mhz , and contains the parasite ring - cavity peaks .
it is also an area where the multiplicative and additive noises are both white [ in the sense that in that range , @xmath151 and @xmath152 are both constant ] .
the local minima of the spectrum in that area are obtained for @xmath153 , so that the floor of the phase noise after the flicker decrease is @xmath154 when @xmath85 is sufficiently high .
this level is @xmath155 db below the additive white noise power density scaled to the power of the microwave .
the recipe for a low phase noise floor is then quite simple , and also quite conventional : low additive noise @xmath108 , and high power @xmath156 for the microwave signal .
the bandwidth does not play any role in this case , as long as the multiplicative noise is not too strong .
the spurious peaks are localized around integer multiples of the round - trip frequency @xmath157 khz .
more precisely , a fourth order taylor expansion of the denominator of eq .
( [ eq_psi_ab_th_four ] ) shows that around these resonance frequencies , the phase noise can be expressed as @xmath158 by finding the minima of this taylor - expanded denominator , it can be shown that the spurious peaks are in fact frequency - shifted according to @xmath159 then , their height relatively to the phase noise floor can also be calculated as @xmath160 ^ 4 = 120 \ , { \rm{db } } - 40 \log n \ , .
\label{eq_psi_peaks}\end{aligned}\ ] ] it appears that the level of the spurious peaks increases with the rf bandwidth and with the delay : therefore , a large delay may lead to a lower phase noise near the carrier [ see eq .
( [ eq_psi_flicker ] ) ] , but it also leads to a higher level for the spurious peaks , so that an optimal trade - off has to be found .
it is also noteworthy that this level is _ independent _ of the power densities @xmath108 or @xmath147 . in our case , the height of the first spurious peak relatively to the floor is theoretically equal to @xmath161 db , in excellent agreement with the experimental results of fig .
[ spec_at_exp ] , where a height of @xmath162 db has been measured .
it can also be shown from eq .
( [ eq_psi_expand ] ) that the -@xmath163 db bandwidth of the spurious peaks is @xmath164 an extremely small value which is experimentally confirmed with the results of fig . [ spec_at_exp]b . in comparison
, these spurious peaks typically have a linewidth higher than @xmath165 khz below threshold ( see fig . [ spec_bt_num ] ) , but their linewidths sharply narrow as @xmath114 . a theoretical phase noise spectrum above threshold in a @xmath136 mhz window , with @xmath166 rad@xmath1/hz , @xmath167 rad@xmath1/hz , @xmath168 khz and @xmath169 hz .
the dimensionless amplitude of the microwave oscillation is @xmath170 , corresponding to a power of @xmath171 dbm.,width=302 ] @xmath172 experimental phase noise spectrum in a @xmath173 khz window , showing a noise floor around @xmath174 dbrad@xmath1/hz for a microwave power of @xmath175 dbm .
@xmath176 enlargement of the spectrum around the first spurious peak at the frequency @xmath177 khz .
the maximum of this spurious peak is at @xmath178 dbrad@xmath1/hz [ height of the peak : @xmath179 db ] , and its @xmath24 db bandwidth is around @xmath180 mhz .
all these experimental data are in excellent agreement with the theory .
note that the height of the peak in fig .
@xmath172 is not indicative because of insufficient resolution .
also note that the peak at @xmath150 hz is a parasite peak originating from the electric mains supply.,title="fig:",width=302 ] @xmath172 experimental phase noise spectrum in a @xmath173 khz window , showing a noise floor around @xmath174 dbrad@xmath1/hz for a microwave power of @xmath175 dbm .
@xmath176 enlargement of the spectrum around the first spurious peak at the frequency @xmath177 khz .
the maximum of this spurious peak is at @xmath178 dbrad@xmath1/hz [ height of the peak : @xmath179 db ] , and its @xmath24 db bandwidth is around @xmath180 mhz .
all these experimental data are in excellent agreement with the theory .
note that the height of the peak in fig .
@xmath172 is not indicative because of insufficient resolution .
also note that the peak at @xmath150 hz is a parasite peak originating from the electric mains supply.,title="fig:",width=302 ] here , the term @xmath182 $ ] progressively becomes negligible as @xmath183 is increasing , so that the ring - cavity peaks excited by white noise become strongly damped ( for being outside the rf bandwidth ) . in this case
, the phase noise decays as @xmath184 however , the phase noise does not decrease monotonically as @xmath185 up to infinity : in fact , for @xmath186 , there is a second phase noise floor induced by the coupling between phase fluctuations and amplitude fluctuations ( second - order effect , see ref .
this article has presented a theoretical study of phase noise in oeos .
our approach has consisted in a langevin formalism , that is , in adding noise sources to a core deterministic model for the microwave dynamics .
we have found a excellent agreement between the main predictions of the model and the experimental results .
there is also an good agreement between this theory and the results that are known from the literature , or from our earlier works .
the main advantage of this approach is that it enables within the same framework to understand the behavior of the system under and above threshold , as the same model continuously accounts for all the observed features independently of the value of the gain .
however , we have not taken into account in this first model the noise generated by the filter ( noisy @xmath18 and @xmath187 ) , and the delay time ( noisy @xmath148 ) .
fluctuations associated to these parameters may induce interesting stochastic features , that will be adressed in future work .
another line of investigation is to achieve a better spectral and statistical fitting of the multiplicative noise @xmath188 , which is an essential variable for the determination of phase noise spectra .
future work will also emphasize on phase noise reduction methods , such as optical filtering , multiple - loop architectures , or quadratic crossed nonlinearities @xcite .
in the open - loop configuration , the total output power can also be obtained using some quantum electronics formulas .
effectively , the output power can be explicitly expressed as @xmath190\,g \delta f \ , , \label{app_1}\end{aligned}\ ] ] where @xmath27 is the total gain of our two cascaded amplifiers ( @xmath191 and @xmath192 db at @xmath2 ghz ) , @xmath193 db is the noise figure of the first amplifier , @xmath194 k is the room temperature , @xmath195 is the boltzmann constant , @xmath196 is the electron charge , @xmath197 ma is the photodiode current , @xmath198 @xmath99 is the equivalent load impedance for the photodiode , and @xmath199 mhz is the bandwidth of the rf filter .
the formula gives @xmath200 nw , while we have measured @xmath201 nw . the combination of eqs .
( [ noise_pow_gam0 ] ) and ( [ app_1 ] ) also gives a method to determine @xmath108 directly from the specifications of the various optoelectronic components used in the oscillation loop .
we use it chain rules to derive the stochastic differential equation for the phase .
we first rewrite eq .
( [ a_above_th ] ) under the differential form @xmath202 { \cal a}_t dt + \mu e^{i\vartheta } \ , d{\cal w}_a \ , , \label{app_9}\end{aligned}\ ] ] where @xmath203 is a differential wiener process .
note that @xmath204 and @xmath205 .
the fact that @xmath206 explains why the differential terms of second order should be taken into account in stochastic calculus , so that usual differentiation and chain rules do not generally apply . when considering the second order one may consider @xmath207 , and discard higher order terms since @xmath208 \ll dt$ ] for @xmath209 .
we set @xmath210 , where @xmath211 is an auxiliary variable . at order @xmath212
, we have @xmath213 ^ 2 \nonumber \\ & = & - \mu e^{i\vartheta } \ , dt + \mu e^{i\vartheta } [ 1 + \eta_m ( t ) ] \ , { \frac{{\cal a}_t}{{\cal a } } } \,dt \nonumber \\ & & + \frac{\mu e^{i\vartheta } } { { { \cal a } } } \ , d{\cal w}_a \ , , \label{app_4}\end{aligned}\ ] ] where we have considered @xmath214 . assuming second order fluctuations for the amplitude ( that is , @xmath215 ) , we are led to @xmath216\,\sin[\vartheta + \psi_t -\psi ] \,dt \nonumber \\ & & + \frac{\mu}{|{\cal a}_0| } \ , dw_{a,\psi } \ , , \label{app_11}\end{aligned}\ ] ] where @xmath217 is a real gaussian white noise with zero mean and variance @xmath218 .
since @xmath219 , we have @xmath220 \simeq \vartheta - ( \psi -\psi_t)$ ] so that finally , @xmath221 this result is also the one we may have recovered through the usual rules of differential calculus ( however , note that it is not so for the equation ruling the power variable @xmath222 ) . also note that this equation is valid only as long as the approximation of neglecting @xmath223 in eq .
( [ eq_calb_stoch_1 ] ) is valid .
it is possible to gain a different physical insight into the phase noise problem in oeos , using an alternative methodology related to the conventional theory of feedback oscillators .
we hereafter briefly sketch the main lines of this heuristical approach .
the oscillator consists of an amplifier of gain @xmath224 ( constant ) and of a feedback path of transfer function @xmath225 in closed loop .
the function @xmath225 selects the oscillation frequency , while the gain @xmath224 compensates for the feedback loss .
this general model is independent of the nature of the amplifier and of the frequency selector .
we assume that the barkhausen condition @xmath226 for stationary oscillation is verified at the carrier frequency @xmath227 by through a gain - control mechanism . under this hypothesis
, the phase noise is modeled by the scheme shown in fig .
[ scheme_barkhausen ] , in which all signals are _ the phases of the oscillator loop _ @xcite .
the main reason for describing the oscillator in this way is that we get rid of the non - linearity , pushing it in the loop - gain stabilization .
the ideal amplifier repeats " the phase of the input , for it has a gain of @xmath165 ( exact ) in the phase - noise model .
the real amplifier introduces the random phase @xmath228 in the loop . in this representation ,
the phase noise is always additive noise , regardless of the physical mechanism involved .
this eliminates the mathematical difficulty inherent in the parametric nature of flicker noise and of the noise originated from the environment fluctuations .
the feedback path is described by the transfer function @xmath230 of the phase perturbation . in the case of the delay - line oscillator ,
the feedback path is a delay line of delay @xmath148 followed by a selector filter .
the latter is necessary , otherwise the oscillator would oscillate at any frequency multiple of @xmath231 , with no preference . implementing the selector as a bandpass filter ( a resonator ) of group delay @xmath232 ,
the phase - perturbation response of the feedback path is @xmath233 we assume that all the phase perturbations in the loop are collected in the random function @xmath228 , regardless of the physical origin ( amplifier , photodetector , optical fiber , etc . ) . denoting with @xmath234 the oscillator output phase ,
the oscillator is described by the phase - perturbation transfer function @xmath235 . by inspection on fig .
[ scheme_barkhausen ] , and using the basic equations of feedback , the oscillator transfer function reads @xmath236 and the oscillator phase noise spectrum would be given by @xmath237 .
x. s. yao and l. maleki , optoelectronic oscillator for photonic systems " , _
ieee j. of quantum electron .
_ * 32 * , pp . 11411149 ( 1996 ) .
y. kouomou chembo , l. larger , h. tavernier , r. bendoula , e. rubiola and p. colet `` dynamic instabilities of microwaves generated with optoelectronic oscillators '' , _ opt .
_ * 32 * , 25712573 ( 2007 ) .
a. demir , a. mehrotra , and j. roychowdhury , phase noise in oscillators : a unifying theory and numerical methods for characterization " , _ ieee trans .
theory and appl .
_ * 47 * , pp . 655674 ( 2000 ) .
g. j. coram , a simple 2-d oscillator to determine the correct decomposition of perturbations into amplitude and phase noise " , _ ieee trans .
i , fund . theory and appl .
_ , * 48 * , pp . 896898 ( 2001 ) .
e. rubiola , _ phase noise and frequency stability in oscillators _ , cambridge university press , november 2008 ( in press ) .
isbn 978 - 0521 - 88677 - 2 . | we introduce a stochastic model for the determination of phase noise in optoelectronic oscillators . after a short overview of the main results for the phase diffusion approach in autonomous oscillators ,
an extension is proposed for the case of optoelectronic oscillators where the microwave is a limit - cycle originated from a bifurcation induced by nonlinearity and time - delay .
this langevin approach based on stochastic calculus is also successfully confronted with experimental measurements .
[ [ keywords ] ] keywords : + + + + + + + + + optoelectronic oscillators , phase noise , microwaves , semiconductor lasers , stochastic analysis . |
our knowledge of the galaxy halo environment is slowing being painted by the study of the interstellar medium in both emission and absorption . understanding the distribution and kinematics of extra planer and extended halo gas ,
compared to that of the host galaxy , can provide constraints on formation and evolutionary models .
diffuse ionized gas ( digs ) , usually observed in and emission , has been studied in edge on disk galaxies .
@xcite found filaments in m82 that flow away from the plane of the disk .
@xcite showed that digs extend out to 13 kpc above the plane of ngc 5775 .
the gas is decreasing in rotational velocity with height above the plane and may have no rotation above several kpcs . in 21cm emission
, @xcite finds gas extending out to at least 5 kpc above the galaxy plane of ngc 891 that rotates 25 to 100 km s@xmath1 more slowly than the disk gas .
models are consistent with a halo that is lagging behind the disk rotation .
fraternali ( @xcite , 2002 , 2004 ) also found halo gas with slower rotation than the disk gas in several galaxies .
this so called `` anomalous gas '' displays a radial inward flow toward the center of the galaxy .
they may be observing the infalling stage of galactic fountains .
they also detect `` forbidden gas '' that moves contrary to disk rotation and does not fit well in the classical fountain picture .
vertical motions and `` holes '' in the distribution have been seen in m31 ( @xcite ) and in the dwarf irregular hoii ( @xcite ) .
these holes might be produced by the expansion of large bubbles around stellar associations via strong outflowing winds and/or supernovae .
it is clear that the disk
halo interface is dynamically and kinematically complex .
the study of halo gas in emission has several observational challenges .
halo gas is intrinsically faint which limits studies to the local universe and the present epoch .
furthermore , only the highest column density regions of inner halos can be studied .
an additional complication is the distinction between halo and disk emission for various galaxy orientations . on the other hand ,
observing halo gas in absorption allows one to probe both the inner and outer halos to much lower column densities at all redshifts .
absorption selected galaxies are chosen by gas cross section only and not by _ a priori _ knowledge of galaxy morphology , orientation , and surface brightness .
absorption studies primarily use the @xmath2 doublet as a tracer of halo gas .
since the early 1990s , the association of absorption with normal , bright , field galaxies has been well established ( e.g. , bergeron & boiss ' e 1991 ; @xcite ) . and
are orientated such that the quasar is down .
displaying the galaxies in this fashion allows one to see the variety of galaxy orientations with respect to the quasar line of sight.,width=518 ] equivalent width .
( bottom right ) asymmetries in galaxy morphologies normalized to impact parameter plotted against @xmath3 equivalent width.,width=508 ]
we have examined the detailed connections between halo gas and galaxies using hires and uves quasar spectra to study the absorption kinematics and _ hst_/wfpc2 images of the quasar fields to measure the host galaxy properties .
all galaxies in our sample are spectroscopically confirmed to have the same redshift as the absorption . in figure [ ggkfig : mosaic ] , we present 36 of 38 galaxies in our current sample having redshifts between 0.3 @xmath4 1 .
each galaxy `` postage stamp '' is @xmath5 with the quasar oriented downward .
the range of impact parameters are @xmath6 kpc .
note that the galaxies exhibit a wide range of orientations with respect to the quasar line of sight .
also , there are a wide variety of galaxy morphologies .
some of the galaxies appear slightly perturbed and/or have bright regions .
others have minor satellites or major companions .
three of the absorbers are associated with double galaxies that could be in the process merging or being harassed .
we modeled the morphology and orientations of the galaxies using gim2d ( ( * ? ? ? * simard 2002 ) ) .
the models comprise a smooth disk and bulge component from which we quantify disk inclinations , disk position angles , disk scale lengths , bulge position angles , bulge effective radii , bulge to total ratios , galaxy half light radii , and morphological asymmetries . an example model and the model residual
are shown in the upper panels of figure [ ggkfig : asym ] .
we used voigt profile fitting to model the hires and vlt spectra ( @xcite ; @xcite ) .
the fits to the spectra provide the number of clouds in an absorption system and their associated column densities , doppler parameters , and velocities . directly from the spectra we also measured the equivalent width , velocity spread , and velocity asymmetry of the absorbing gas . to examine halo geometry ,
we tested for correlations between galaxy orientation and absorption properties for a subsample having a rest frame equivalent width , @xmath7 , less than 1.4 .
this cutoff limits our subsample to `` classical '' systems and is designed to eliminate dlas and wind driven systems ( see @xcite ; @xcite ) .
the orientation of a galaxy is the combined projection of the galaxy s inclination , @xmath8 , and the position angle , @xmath9 .
the position angle is defined as the primary angle between the galaxy major axis and the quasar line of sight .
we find there are no statistically significant correlations between galaxy orientation parameters and absorption properties . in particular ,
if the gas is distributed in a disk geometry then @xmath10 should correlate with @xmath7 .
as shown in the bottom
left panel of figure [ ggkfig : asym ] , the distribution is consistent with being random .
the gim2d model residual image of the example galaxy , shown at the top
right panel of figure [ ggkfig : asym ] , displays a spiral barred structure along with an extend tidal tail on one side . without modeling the galaxy
these underlying features would go unnoticed . to quantify these asymmetries we used two different methods .
one method computes the asymmetries on the residual image ( @xcite ) and the other is computed directly for the science image ( abraham 1994 ; 1996 ) .
we find a @xmath0 correlation using the abraham method and a @xmath11 correlation using the ( * ? ? ?
* schade ) method between @xmath7 and galaxy morphological asymmetry normalized by the impact parameter .
the correlation is shown in the bottom
right panel of figure [ ggkfig : asym ] , where the dashed line is a maximum likelihood fit . in the near future
, we will be conducting a multivariate analysis to further explore the nature of correlations between galaxy and absorption properties ( e.g. , @xcite ) . and @xmath12 .
the projected distance @xmath13 is 38.8 @xmath14 kpc .
( right ) galaxy rotation curve ( top ) as a function of slit position with the kinematics of the observed @xmath3 ( middle ) and @xmath15 absorption profiles ( bottom ) from the hires spectrum of the quasar.,width=508 ]
the lack of correlations with galaxy orientation implies that gaseous halos are not necessarily disk like or spherical or that the distribution of gas could be patchy and have a less than unity covering factor ( also see churchill 2005 _ this volume _ ) .
the 3.2 @xmath16 correlation suggests that a galaxy with strong perturbations could produce a similar absorption strength at a large galactocentric distance as a galaxy with mild perturbations at a smaller galactocentric distance .
it is possible that strong star formation , winds , or tidal stripping could lead to the development of both morphological perturbations in the galaxy and an increase in the halo gas cross section .
galaxies that are more symmetric may not produce as significant absorption since there is a lacking or inadequate mechanism for repopulating the halo gas .
this would explain why some bright galaxies close to the quasar produce very weak to no absorption ( see @xcite ) . to help understand these results and to see if there may be sample selection effects
, we will employ more advanced models ( i.e. , lagging halos , fountains , infall ) .
furthermore , a full census of absorbers and non
absorbers is central to a deeper interpretation of our results .
in order to further analyze the galaxy halo kinematic connection it is necessary that we obtain spectra of the galaxies .
we can then place the halo absorption features in the same velocity frame of the galaxy .
the velocity field between the galaxy and the halo will tell us how they are kinematically related so that we can differentiate between competing halo models .
a prime example of such a study was by @xcite , who presented the kinematic properties of halo gas for a sample of five edge on galaxies . in figure [ ggkfig : kinematics ] (
left panel ) we present the _ hst _
image of the q1038 + 064 field with an absorber at @xmath170.4415 .
long slit spectroscopy was obtained to get the kinematics of the gas in the disk ( upper right panel ) .
a hires spectrum of the quasar was obtained for kinematic information of the halo gas ( lower right panels ) .
the halo gas kinematics is probed at 38.8 @xmath14 kpc away from the disk .
absorption between @xmath12250 and @xmath12150 km s@xmath1 shows halo gas kinematics similar to that in the disk .
additional absorption between @xmath18 and @xmath19 km s@xmath1 is suggestive of slower rotation with respect to the disk , similar to that mentioned by fraternali ( 2001 , 2002 , 2004 ) .
thus , in this galaxy probing out to distances far beyond the plane , one finds that the halo gas is aware of the the kinematics of the disk . in four of their five cases , @xcite found disk like halo kinematics consistent with a rotating and/or lagging halo .
another study of and edge on galaxy ( @xcite ) found absorption with velocities inconsistent with the galaxy rotation .
this `` forbidden '' gas may actually be caused by superbubbles ( see @xcite ) .
over all , only a half dozen galaxies have been studied in great detail by comparing the kinematics of the absorbing gas and the galaxy kinematics . however , these galaxies were all edge on disk galaxies .
we are carrying out an in depth study , similar to that mentioned above , for our sample of 38 galaxies having a wide range of morphologies and line of sight orientations . here , we have reported an orientation and morphology analysis , in absence of galaxy spectra .
our results hint that there are some relationships between the halo and the host galaxy when looking at the distance normalized galaxy perturbations . obtaining
the galaxy spectra will allow us to perform a full halo model analysis which is much need in order to understand the nature of galactic halos .
we would like to acknowledge a small grant from the iau and sigma xi scientific research society .
michael murphy , michael rauch , wal sargent and alice shapley all made important contributions to the new work presented herein . | we studied a sample of 38 intermediate redshift absorption selected galaxies using ( 1 ) keck / hires and vlt / uves quasar spectra to measure the halo gas kinematics from absorption profiles and ( 2 ) _
hst_/wfpc2 images to study the absorbing galaxy morphologies .
we have searched for correlations between quantified gas absorption properties , and host galaxy impact parameters , inclinations , position angles , and quantified morphological parameters .
we report a @xmath0 correlation between asymmetric perturbations in the host galaxy morphology and the absorption equivalent width .
we suggest that this correlation may indicate a connection between past merging and/or interaction events in absorption selected galaxies and the velocity dispersion and quantity of gas surrounding these galaxies . |
the nuclei above @xmath1fe are produced dominantly by the slow ( @xmath2 ) and rapid ( @xmath0 ) neutron - capture ( @xmath3-capture ) processes .
the main @xmath2-process occurs in low - mass asymptotic giant branch stars and mainly produces the nuclei above @xmath4sr while the weak @xmath2-process occurs in massive stars and mainly produces the nuclei up to @xmath4sr ( e.g. , kppeler , beer , & wisshak 1989 ) .
it is considered here that type ii supernovae ( sne ii ) resulting from evolution of massive stars are the major sources for the @xmath0-process ( e.g. , qian 2000 ) . as
massive stars have much shorter lifetimes than low - mass stars , sne ii made the dominant contributions to the abundances of @xmath3-capture elements in metal - poor ( mp ) stars that were formed in the early galaxy .
detailed abundance patterns covering many @xmath3-capture elements have been obtained for a number of ultra - metal - poor ( ump ) stars with [ fe / h ] @xmath5 ( e.g. , sneden et al .
2000 ; westin et al .
2000 ; hill et al . 2001 ) .
this letter discusses the implications of these abundance patterns for @xmath0-process nucleosynthesis in sne ii . starting with a large abundance ratio of neutrons relative to the seed nuclei ( i.e. , a large neutron - to - seed ratio ) , the @xmath0-process produces a distribution of neutron - rich progenitor nuclei far from stability through the interplay of rapid neutron capture , photo - disintegration , and @xmath6-decay . depending on the neutron - to - seed ratio
, the @xmath0-process may produce progenitor nuclei that subsequently undergo fission on a short timescale .
this limits the range of nuclei involved in the @xmath0-process and may result in a cyclic flow between the terminating nuclei and their fission products ( i.e. , fission cycling ) .
the role of fission in @xmath0-process nucleosynthesis was discussed by the very first papers that proposed the @xmath0-process ( burbidge et al .
1957 ; cameron 1957 ) .
fission cycling was discussed in detail by seeger fowler , & clayton ( 1965 ) and was included in many @xmath0-process studies ( e.g. , rauscher et al .
1994 ; freiburghaus , rosswog , & thielemann 1999 ; cameron 2001 ) .
fission was also invoked to explain certain features of the observed @xmath0-process abundance patterns ( @xmath0-patterns ) .
for example , cameron ( 1957 ) attributed the small peak in the rare earth region ( with mass numbers @xmath7 ) of the solar @xmath0-pattern to fission of the progenitor nuclei with @xmath8 .
an @xmath0-process scenario with fission but no fission cycling is considered here to account for the observed abundance patterns in ump stars ( 2 ) .
it is proposed that fission may be induced by neutrino reactions with the progenitor nuclei after the @xmath0-process freezes out ( i.e. , exhausts all the initial neutrons ) in sne ii .
the conditions under which this scenario could be realized and some possible features of the resulting @xmath0-pattern are discussed ( 3 ) .
further studies to develop this scenario are emphasized ( 4 ) .
extensive observations showed that the abundances of the heavy @xmath3-capture elements from ba up to pt in ump stars exhibit remarkable regularity and follow the solar @xmath0-pattern rather closely ( e.g. , sneden et al . 2000 ; westin et al .
however , as shown in figure 1a , the observed abundances of the light @xmath3-capture elements rh , pd , ag , and cd in the ump star cs 22892 - 052 ( sneden et al . 2000 ) are too low relative to the solar @xmath0-pattern ( arlandini et al .
1999 ) that is translated to fit the region above ba ( @xmath9 ) .
a possible explanation for this is that different production mechanisms may be responsible for the @xmath3-capture elements below and above ba in ump stars .
for example , fission cycling tends to produce a rather robust @xmath0-pattern at @xmath10 ( e.g. , freiburghaus et al .
so the regular abundance pattern for ba and above in ump stars could be attributed to an @xmath0-process with a neutron - to - seed ratio that is sufficiently high for fission cycling to occur
. then the light @xmath3-capture elements in ump stars could be attributed to some other @xmath0-process with a neutron - to - seed ratio that is appropriate for producing these elements only or to the weak @xmath2-process that might operate at some unusually high efficiency in the earliest generations of massive stars .
however , the attribution of the abundances of ba and above in ump stars to an @xmath0-process with fission cycling may be in conflict with the data on another ump star cs 31082 - 001 . as can be seen from figure 1b , the observed abundances in cs 31082 - 001 ( cayrel et al . 2001 ; hill et al .
2001 ) and cs 22892 - 052 ( sneden et al .
2001 ) follow approximately the same pattern except that the differences in the abundances of os and the radioactive elements th and u between the two stars appear to be significantly larger than those for the other elements ( the differences in the abundances of ir and pb may also be significant but the uncertainties are large ) . as can be best seen by considering the abundance ratio th / eu , the data on cs 31082 - 001 and cs 22892 - 052 do not support a robust @xmath0-pattern at @xmath10 that is expected from fission cycling . in view of the extremely low [ fe / h ] values of @xmath5 for the two stars ,
it is reasonable to assume that both stars were formed during the first few gyr after the onset of galactic @xmath0-process nucleosynthesis .
so both stars should have ages of @xmath11 gyr .
under the assumption of a fixed @xmath0-process yield ratio of th relative to eu , the difference @xmath12 in the ages of the two stars is @xmath13 where @xmath14 gyr is the lifetime of @xmath15th .
the data @xmath16 , @xmath17 for cs 31082 - 001 ( cayrel et al .
2001 ; hill et al .
2001 ) and @xmath18 , @xmath19 for cs 22892 - 052 ( sneden et al . 2001 ) , where @xmath20 , give @xmath21 gyr .
this is totally unreasonable .
thus , the assumption of a fixed @xmath0-process yield ratio of th relative to eu must be invalid and the @xmath0-pattern at @xmath10 is not robust . while it may be possible to accommodate the data on both cs 22892 - 052 and cs 31082 - 001 by having variable amounts of fission cycling in the @xmath0-process , an alternative scenario with fission but no fission cycling
is considered here to explain the observed abundance patterns of @xmath3-capture elements in these two and possibly other ump stars .
it is proposed that an @xmath0-process produces a freeze - out distribution of progenitor nuclei covering @xmath22 with a peak at @xmath23 ( corresponding to the magic neutron number @xmath24 ) .
the required neutron - to - seed ratio is high , but it is assumed that no fission cycling occurs during the @xmath0-process .
this may be consistent with some recent studies of fission barriers for extremely neutron - rich nuclei ( mamdouh et al .
2001 ) . as the progenitor nuclei @xmath6-decay towards stability after freeze - out ,
all of those with @xmath25 eventually undergo spontaneous fission ( e.g. , cameron 2001 ) . due to the strong influence of the closed proton and neutron shells at @xmath26sn ,
the fission of @xmath25 is expected to produce one fragment at @xmath27 and the other at @xmath28 although the detailed fission yields may be more complicated ( e.g. , hulet et al .
. some of the progenitor nuclei with @xmath29 would also undergo spontaneous fission during decay towards stability , thereby producing one fragment again at @xmath27 and the other at @xmath30 .
the possibility of fission for @xmath29 may be greatly enhanced by reactions with the neutrinos emitted in sne ii as these nuclei could be highly excited by such reactions .
neutrino reactions may even induce fission of the progenitor nuclei with @xmath31 .
experiments using energetic particles to induce fission of the stable or long - lived nuclei in this mass range showed that the fission mode is dominantly symmetric with no preference for a fission fragment at @xmath27 and the mass ratio of the two fission fragments is @xmath321.2 ( e.g. , britt et al .
1963 ; see mller et al .
2001 for recent theoretical interpretation ) .
so the neutrino - induced fission of @xmath31 is expected to produce fragments at @xmath33 .
thus , in the @xmath0-process scenario proposed here , the fission of the progenitor nuclei with @xmath22 during decay towards stability after freeze - out would produce nuclei with @xmath34 ( see table 1 ) .
in view of the mass range of the progenitor nuclei and their fission products , it is tempting to consider the proposed @xmath0-process scenario as the dominant source for all the @xmath3-capture elements observed in ump stars .
the ratio of the fission yields at @xmath34 relative to the surviving abundances at @xmath35 depends on the number of neutrino reactions experienced by each progenitor nucleus after freeze - out .
variation of this number among individual sne ii may explain why the observed abundances of os and above ( @xmath36 ) relative to those of the @xmath3-capture elements below os are significantly different in cs 22892 - 052 and cs 31082 - 001 .
the conditions under which the proposed scenario could be realized are discussed below .
in the @xmath0-process scenario proposed here , all the initial neutrons are exhausted when the heaviest progenitor nuclei with @xmath37 are produced . except for the progenitor nuclei perhaps with @xmath38 that are mostly produced when the @xmath0-process is running out of neutrons , the freeze - out abundances of the other progenitor nuclei are expected to be inversely proportional to their @xmath6-decay rates .
the @xmath6-decay rates for the progenitor nuclei with @xmath39 ( @xmath24 ) are @xmath11 times smaller than those for @xmath40 ( mller et al .
1997 ) . with this crude guidance to the freeze - out abundances ( see table 1 ) ,
the ratio of the fission yields at @xmath34 relative to the surviving abundances at @xmath35 is @xmath41 if a fraction @xmath42 of the progenitor nuclei with @xmath35 and all of those with @xmath25 undergo fission during decay towards stability after freeze - out . by assigning the @xmath3-capture elements below os to the fission products and considering pb as the dominant decay product of the surviving nuclei with @xmath43
, it can be estimated from figure 1 that @xmath44 is required to account for the gross features of the observed abundance pattern in cs 22892 - 052 and @xmath45 for cs 31082 - 001 . without excitation by energetic particles ,
the probability of fission increases with @xmath46 , where @xmath47 is the atomic number of the nucleus .
all of the progenitor nuclei with @xmath25 are expected to undergo spontaneous fission eventually as @xmath47 increases during their decay towards stability ( e.g. , cameron 2001 ) .
in contrast , as the stable or long - lived nuclei with @xmath31 do not undergo any significant spontaneous fission , neither would their neutron - rich progenitors in the absence of high excitation
. the probability of spontaneous fission during decay towards stability may not be significant even for the progenitor nuclei with @xmath29 ( e.g. , meyer et al . 1989 ) .
however , the progenitor nuclei and their daughters can be highly excited by reactions with the neutrinos emitted in sne ii .
the typical excitation energy is @xmath48 mev for @xmath49 capture and @xmath50 mev for reactions with @xmath51 , @xmath52 , @xmath53 , and @xmath54 ( e.g. , qian et al .
thus , neutrino reactions may be crucial in inducing the fission of @xmath5540% of the progenitor nuclei with @xmath35 during decay towards stability after freeze - out .
note that neutrino - induced fission should be unimportant during the @xmath0-process as the rates for @xmath49 capture must be severely limited in order to produce the abundance peak at @xmath23 ( fuller & meyer 1995 ) .
an essential question is then whether there would be a sufficient number of neutrino reactions with the progenitor nuclei after the @xmath0-process freezes out in sne ii .
the gravitational binding energy of the neutron star formed in sne ii is radiated in neutrinos over a period of @xmath55 s ( e.g. , woosley et al .
this is the upper limit on the duration of the @xmath0-process so that the progenitor nuclei can have a significant number of neutrino reactions after freeze - out .
the duration of the @xmath0-process is dominantly controlled by the @xmath6-decay lifetimes of the progenitor nuclei . for an @xmath0-process path in a low - density environment where electrons are nondegenerate , the typical @xmath6-decay lifetimes are @xmath56 s for @xmath11 progenitor nuclei with closed neutron shells and @xmath57 s for those without ( mller et al .
it would then take @xmath58 s to reach the progenitor nuclei with @xmath37 .
so all the progenitor nuclei could have a significant number of neutrino reactions after freeze - out if the @xmath0-process scenario proposed here is restricted to occur in a low - density environment such as the neutrino - driven wind from the neutron star ( e.g. , woosley et al .
1994 ) instead of a high - density environment such as an accretion disk ( cameron 2001 ) .
another concern is the probability of fission following a neutrino reaction .
a neutron - rich nucleus far from stability can de - excite by emitting neutrons instead of undergoing fission .
increase in @xmath47 would help neutrino - induced fission compete with neutron emission .
based on the @xmath6-decay lifetimes estimated by mller et al .
( 1997 ) , @xmath47 could increase by @xmath59 units within @xmath11 s for the progenitor nuclei with @xmath35 .
thus , the crucial role of neutrino - induced fission in the @xmath0-process scenario proposed here relies on the assumption that the probability of fission following a neutrino reaction is or becomes significant during the first @xmath59 steps in the @xmath6-decay chains of the progenitor nuclei with @xmath35 .
this assumption should be checked by detailed calculations in the future .
if a typical probability of fission is @xmath60 ( i.e. , comparable to that of neutron emission ) , a fraction @xmath61\sim 20$]40% of the progenitor nuclei would undergo fission after a total number @xmath621.6 of neutrino reactions per nucleus .
the values of @xmath63 to give @xmath6440% may be reduced somewhat by increasing @xmath3 to @xmath65 .
a similar level of neutrino reactions would completely account for the production of @xmath66187 by neutrino - induced neutron emission from the progenitor nuclei in the peak at @xmath23 after freeze - out ( qian et al .
1997 ; haxton et al .
1997 ) .
given that the proposed @xmath0-process scenario with fission could be realized , it is interesting to note some features of the corresponding yield pattern that are not so evident from the data for cs 22892 - 052 and cs 31082 - 001 .
as proposed , the freeze - out distribution of progenitor nuclei covers @xmath22 with a peak at @xmath23 .
the fission of @xmath5540% of the progenitor nuclei with @xmath35 and all of those with @xmath25 would produce enhanced fission yields at @xmath67 and @xmath27 .
the enhancement at @xmath67 is expected from fission of the progenitor nuclei in the peak at @xmath23 , while that at @xmath27 from the common fission fragment at this position for all the progenitor nuclei with @xmath68 .
thus , the typical yield pattern from the proposed @xmath0-process scenario has a peak at @xmath23 and enhanced yields at @xmath67 and @xmath27 .
this yield pattern could be modified by capturing the neutrons released from fission .
for example , the positions of the peak and the enhanced fission yields could be shifted to somewhat higher @xmath69 .
such shifts in the yield pattern due to residual neutron capture were found in @xmath0-process scenarios with fission cycling ( rauscher et al .
1994 ; freiburghaus et al .
1999 ) . in any case
, the @xmath0-process scenario proposed here could not account for the peak at @xmath70 in the solar @xmath0-pattern and therefore could provide one of the diverse sources for the @xmath0-process that are required by the model of wasserburg , busso , & gallino ( 1996 ) to explain the meteoritic data on @xmath71i and @xmath72hf .
the yield pattern resulting from this scenario might also be similar to the @xmath0-pattern that was derived by qian & wasserburg ( 2001 ) for the high - frequency kind of sne ii based on the model of wasserburg et al .
( 1996 ) and observations of mp stars
. there may be a hint for the shift of the enhanced fission yields at @xmath27 to higher @xmath69 from studies of the ba isotopic composition in the ump star hd 140283 .
the three @xmath0-process isotopes of ba are @xmath73ba , @xmath74ba , and @xmath75ba .
the odd-@xmath69 isotopes of an element have finite nuclear magnetic moments that lead to hyperfine structure of the atomic spectra .
magain ( 1995 ) performed a detailed analysis of the ba spectra for hd 140283 and concluded that the fraction of the odd-@xmath69 ba isotopes is @xmath76 , corresponding to abundance ratios @xmath75ba/@xmath73ba and @xmath75ba/@xmath74ba of @xmath77 . by capturing the neutrons released from fission ,
the enhanced fission yields at @xmath27 could be shifted to produce this steep rise of the yield at @xmath78 .
it is also possible that the ba in hd 140283 may not be of pure @xmath0-process origin . in any case , the ba isotopic composition in ump stars has important implications for @xmath3-capture processes in the early galaxy and deserves further observational studies .
clearly , much further work is required to develop the @xmath0-process scenario with fission as proposed here . on the observational side , extensive studies of the abundance patterns of @xmath3-capture elements
should be carried out for many more ump stars to identify regularities and variations .
the paucity of data on the light @xmath3-capture elements below ba should be remedied . as pb is the dominant decay product of the progenitor nuclei with @xmath43 , accurate determination of its abundance should be made .
detailed analyses of ba isotopes are also valuable as discussed in 3 . on the theoretical side , the foremost issue to investigate is the probability of neutrino - induced fission for extremely neutron - rich nuclei and the corresponding fission yields .
numerical calculations with an extensive network including neutrino reactions are needed to make a detailed comparison between the theoretical yields and the observed abundance patterns .
the effects of the energy release from fission on the thermodynamic and hydrodynamic evolution of the @xmath0-process environment are also interesting to study .
last but certainly not the least , attempts should be made to understand how the conditions for the proposed @xmath0-process scenario are realized in a low - density environment such as the neutrino - driven wind in sne ii .
a crucial quantity characterizing the @xmath0-process conditions is the neutron - to - seed ratio @xmath79 . by mass conservation , the average mass number @xmath80 for the freeze - out distribution of progenitor nuclei is @xmath81 , where @xmath82 is the mass number of the seed nuclei . with @xmath83
obtained in the neutrino - driven wind ( e.g. , woolsey et al .
1994 ) , @xmath84 is required to produce @xmath85 in the proposed @xmath0-process scenario ( see table 1 ) .
in general , @xmath79 is determined by the electron fraction @xmath86 , the entropy @xmath87 , and the dynamic timescale @xmath88 of the wind . while it remains unclear how the suitable combinations of @xmath86 , @xmath87 , and @xmath88 can be obtained in the wind to give @xmath84 , several possibilities have been suggested .
one is to consider the general relativistic effects of a more massive or compact neutron star . as first shown by qian & woosley ( 1996 ) and
later confirmed by other studies ( cardall & fuller 1997 ; otsuki et al .
2000 ; thompson , burrows , & meyer 2001 ) , such effects increase @xmath87 and decrease @xmath88 , both favoring higher @xmath79 .
another possibility is to consider the effects of neutrino flavor mixing .
the @xmath86 in the wind is mostly determined by the competition between @xmath89 and @xmath90 , and therefore is extremely sensitive to the differences between the luminosities and energy spectra of @xmath91 and @xmath49 ( qian et al .
1993 ) . as @xmath52 and @xmath54
have much higher average energy than @xmath91 and @xmath49 , mixing between @xmath92 and @xmath91 could significantly reduce @xmath86 ( e.g. , qian & fuller 1995 ) , again favoring higher @xmath79 .
it is plausible that by including both the effects of general relativity and neutrino flavor mixing , @xmath84 could be obtained in the wind .
these issues should be examined by future studies .
discussions with roberto gallino , peter mller , and jerry wasserburg are gratefully acknowledged .
i thank the referee , al cameron , for many valuable criticisms .
this work was supported in part by the department of energy under grants de - fg02 - 87er40328 and de - fg02 - 00er41149 . | an @xmath0-process scenario with fission but no fission cycling is considered to account for the observed abundance patterns of neutron - capture elements in ultra - metal - poor stars .
it is proposed that neutrino reactions play a crucial role in inducing the fission of the progenitor nuclei after the @xmath0-process freezes out in type ii supernovae . to facilitate neutrino - induced fission ,
the proposed @xmath0-process scenario is restricted to occur in a low - density environment such as the neutrino - driven wind from the neutron star .
further studies to develop this scenario are emphasized . |
we consider space - time block codes ( stbcs ) for an @xmath2 transmit antenna , @xmath3 receive antenna , quasi - static mimo channel ( @xmath4 mimo system ) with rayleigh flat fading .
the system can be modeled as @xmath5 where @xmath6 is the codeword matrix transmitted over @xmath7 channel uses , @xmath8 is the received matrix , @xmath9 is the channel matrix and the matrix @xmath10 is the additive noise at the receiver .
the entries of @xmath9 and @xmath10 are i.i.d .
zero mean , unit variance , circularly symmetric complex gaussian random variables .
( stbc ) an stbc @xmath11 encoding @xmath12 real independent information symbols , denoted by @xmath13 , @xmath14 , is a set of complex matrices given by @xmath15^t \in \mathcal{a } \right\},\ ] ] where the @xmath16 complex matrices , which are called _ linear dispersion _ or _ weight matrices _ , are linearly independent over the real field @xmath17 @xcite , @xcite , and the finite set is called the _ signal set_. ( code rate ) the rate of an stbc is the average number of information symbols transmitted in each channel use . for the stbc given by ,
the code rate is real symbols per channel use , or complex symbols per channel use ( cspcu ) .
the linear independence of the weight matrices in the definition of an stbc implies that @xmath18 . _ throughout this paper , unless otherwise specified , the code rate is taken to be in terms of complex symbols per channel use . _ generally , the signal set @xmath19 is chosen in such a way that the stbc @xmath11 has full - diversity and large coding gain . in most cases @xmath19 is chosen to be a subset of @xmath20 , where @xmath21 is a full - ranked matrix .
one such instance is when the symbols are partitioned into multiple encoding groups , and each group of symbols is encoded independently of other groups using a lattice constellation , such as in clifford unitary weight designs @xcite ( in which case @xmath22 is an orthogonal matrix ) , quasi - orthogonal stbcs @xcite and coordinate interleaved orthogonal designs @xcite .
there are also instances where the real symbols are encoded independently using regular pam constellations of possibly different minimum distances @xcite , in which case @xmath22 is a diagonal matrix with positive entries . for a complex matrix @xmath23 ,
let its real and imaginary components be denoted by @xmath24 and @xmath25 , respectively .
let @xmath26 denote the complex vector obtained by stacking the columns of @xmath23 one below the other and @xmath27^t.\ ] ] now , the system model given by can be expressed as @xmath28 where , and the _ equivalent channel matrix _
@xmath29 is given by @xmath30 .\ ] ] consider the vector of transformed information symbols which takes values from , where @xmath31 denotes the ring of integers .
the components of @xmath32 take finite integer values , i.e. , @xmath33 , with @xmath34 for some finite @xmath35 .
hence one can use a sphere decoder @xcite to decode @xmath32 and then obtain the ml estimate of the information vector @xmath36 .
the ml decoder output is given by @xmath37 where @xmath38 denotes the frobenius norm of a matrix .
it is claimed in @xcite without proof that is a sufficient condition for the system of equations defined by to be _ not underdetermined _ ,
i.e. , for @xmath39 in section [ sec2 ] , we show that the claim made in @xcite is true only for .
this observation is the gateway to the new results presented from section [ sec2 ] onwards . for a system
where , the sphere decoder complexity , averaged over noise and channel realizations , is _ independent of the constellation size @xmath35 _ and is roughly polynomial in the dimension of the sphere decoding search @xcite , @xcite , @xcite . however , if the rank of @xmath40 is less than @xmath12 , the average sphere decoding complexity is no more independent of the constellation size .
when @xmath41 , the conventional sphere decoder needs to be modified as follows @xcite : the @xmath0 matrix resulting from the @xmath42-decomposition of @xmath43 has the form , where @xmath44 is a @xmath45 upper triangular full - rank matrix .
there is a corresponding partition of @xmath32 as @xmath46^t$ ] .
if @xmath35 is the size of the regular pam constellation used , then for each of the @xmath47 values of @xmath48 , the conditionally optimal estimate of @xmath49 can be found by first removing the interference from @xmath48 , and then using a sphere decoder with the upper triangular matrix @xmath44 to obtain an estimate of @xmath49 .
then , from the resulting @xmath47 estimates of @xmath32 , the optimal vector is chosen .
this observation leads to the following lemma .
[ singular_thm ] for an stbc whose equivalent channel matrix @xmath40 is such that @xmath50 , the sphere decoding complexity as a function of @xmath35 is of the order of @xmath47 , i.e. , @xmath51 . in the rest of the paper , by sphere decoding complexity we mean the average sphere decoding complexity and the focus is on _ the dependence of the complexity on the constellation size @xmath35 and not on the dimension of the sphere decoding search_. we now introduce the notion of _ singularity _ of an stbc which is a direct indicator of its sphere decoding complexity . [ singular ] let @xmath11 be an stbc with rate @xmath52 and let .
we say that @xmath11 is _ singular _ for @xmath3 receive antennas if with probability @xmath1 .
otherwise , @xmath11 is said to be _ non - singular _ for @xmath3 receive antennas . to illustrate the effect of the rank of @xmath40 and singularity of an stbc on the sphere decoding complexity as a function of the constellation size ,
consider the following example .
[ ex : fgd ] in @xcite , a rate @xmath53 code for @xmath54 antennas with @xmath55 real symbols was presented .
define @xmath56 and @xmath57 , and for any two matrices @xmath23 , @xmath58 , denote their kronecker product by .
then , the @xmath59 weight matrices of the stbc in @xcite are as follows : @xmath60 let the number of receive antennas be @xmath61 .
then , and the equivalent channel matrix @xmath40 is of size . now consider the following randomly generated channel matrix @xmath62 0.3457 + 0.2299i & 0.2078 - 0.0723i & -0.7558 - 0.6116i \\
0.7316 - 0.5338i & -0.5567 - 0.1707i & -0.5724 - 0.0212i \\ 0.5140 + 0.9689i & 0.6282 + 0.2257i & -2.0819 - 0.1166i \\ -0.2146 - 1.2102i & -0.8111 + 0.2212i & 1.0171 + 0.4439i \end{bmatrix}.\ ] ] the resulting @xmath40 matrix has rank only @xmath63 and the structure of the upper triangular matrix @xmath0 of size obtained upon its @xmath42-decomposition is shown in at the top of next page .
the non - zero entries of @xmath0 are denoted by ` @xmath64 ' .
it is clear that the first @xmath63 columns of @xmath0 are linearly independent and the last column lies in the span of the first @xmath63 columns . removing the last @xmath65 rows of @xmath0 , which are all zeros , we get the real matrix @xmath66 which is used by the sphere decoder to find the ml estimate of the information vector . in this case , from lemma [ singular_thm ] , the complexity of sphere decoding this stbc for this particular channel realization @xmath9 is of the order of . in section [ sec3a ] , by establishing that with probability @xmath1 , we show that this stbc is singular for @xmath61 .
@xmath67 ' '' '' the contributions and organization of this paper are as follows . *
we introduce the notion of singularity of an stbc which is a direct indicator of its sphere decoding complexity .
* we show that contrary to the claim made in @xcite , is not a sufficient condition for an stbc to be non - singular for @xmath3 receive antennas ( theorem [ th : main_theorem ] , section [ sec2 ] ) . we show that the case of singular stbcs arises only for asymmetric mimo systems with @xmath68 ( proposition [ pr : m_less_than_n ] , section [ sec2 ] ) . *
we show that all known families of high rate multigroup ml decodable - group or multigroup ml decodable if its symbols can be partitioned into @xmath69 groups and each group of symbols can be ml decoded independent of others . ]
codes @xcite , @xcite , @xcite , @xcite and @xcite are non - singular for certain values of @xmath3 ( section [ sec3 ] , see table [ tbl : comparison ] for a summary of results ) . * we derive the sphere decoding complexity of almost all known high - rate multigroup ml decodable codes and show that in each case the sphere decoding complexity is a decreasing function of the number of receive antennas @xmath3 ( section [ sec3 ] , see table [ tbl : comparison ] ) .
we show that even when an stbc is singular , multigroup ml decodability helps reduce the sphere decoding complexity .
the reduction in complexity is from @xmath70 to @xmath71 , where @xmath69 is the number of ml decoding groups ( section [ sec3 ] ) .
some related open problems are discussed in section [ sec4 ] .
* notations : * throughout the paper , matrices ( vectors ) are denoted in bold , uppercase ( lowercase ) letters .
for a complex matrix @xmath23 , its transpose , conjugate and conjugate - transpose are denoted by @xmath72 , @xmath73 and @xmath74 , respectively . for a square matrix @xmath23
, @xmath75 denotes its determinant .
the @xmath76 identity matrix is denoted by @xmath77 and @xmath78 is the null matrix of appropriate dimension .
unless used as a subscript or to denote indices , @xmath79 . for square matrices
@xmath81 , @xmath82 denotes the square , block - diagonal matrix with @xmath83 on the diagonal , in that order . the field of complex numbers is denoted by @xmath84 .
we present a few results which we will use in the following sections to derive the rank of @xmath40 for specific stbcs .
the following result shows that if any stbc is singular for @xmath3 receive antennas , then @xmath85 .
thus , the rank - deficiency problem arises only in asymmetric mimo with .
[ pr : m_less_than_n ] if @xmath86 , every @xmath16 stbc is non - singular for @xmath3 receive antennas .
since @xmath87 is an isomorphism from the @xmath17-vector space @xmath88 to @xmath89 , it is enough to show that @xmath90 are linearly independent with probability @xmath1 .
suppose @xmath9 is full - ranked , i.e. , @xmath91 , then there exists a matrix @xmath92 such that @xmath93 .
if @xmath94 , it would mean that @xmath95 .
since @xmath96 are the weight matrices of an stbc , they are linearly independent and hence @xmath97 , @xmath98 .
thus @xmath99 if @xmath9 is full - ranked .
since @xmath9 is full - ranked with probability @xmath1 @xcite , we have shown that any stbc is non - singular for @xmath3 receive antennas if @xmath86 .
let denote the @xmath17-linear subspace of @xmath100 spanned by the matrices .
[ pr : rank_independent_basis ] let @xmath101 be @xmath16 complex matrices such that and let @xmath102 be any matrix
. then the column spaces of the matrices @xmath103 \textrm { and } \label{eq : ga_of_h}\\ & \textbf{g}_\textbf{b}(\textbf{h } ) = [ \widetilde{vec}(\textbf{b}_1\textbf{h})~\widetilde{vec}(\textbf{b}_2\textbf{h } ) ~ \cdots ~ \widetilde{vec}(\textbf{b}_k\textbf{h } ) ] \label{eq : gb_of_h}\end{aligned}\ ] ] are identical .
in particular , .
let @xmath104 be any vector in the column space of @xmath105 .
then , @xmath106 , for some choice of real numbers @xmath107 , @xmath98 .
since each of the , every @xmath96 can be written as some real linear combination of @xmath101 .
it follows that every @xmath108 can be written as some real linear combination of .
hence @xmath104 belongs to the column space of @xmath109 .
similarly we can show that every vector in the column space of @xmath109 belongs to the column space of @xmath105 also .
this completes the proof .
the following result shows that if every weight matrix of a given stbc is multiplied on the left by a constant invertible matrix , then the rank of the equivalent channel matrix is unchanged .
[ pr : rank_independent_c_mult ] let @xmath110 be any full - rank matrix , @xmath111 , and @xmath9 be any @xmath4 complex matrix .
then we have , where @xmath105 and @xmath109 are as defined in and .
it suffices to show that the subspaces and have the same dimension .
the proof is complete if we show that the vector space homomorphism @xmath112 from the former subspace to the latter , that sends @xmath113 to @xmath114 is a one to one map . if @xmath115 , then . since @xmath116 is invertible
, this means that @xmath117 .
this completes the proof .
it is shown in @xcite that for any @xmath118 , there exists an explicitly constructable set of @xmath119 matrices belonging to @xmath120 that are unitary , hermitian and linearly independent over @xmath17 .
this set of matrices forms a basis for the space of @xmath121 hermitian matrices .
denote by @xmath122 any stbc obtained by using these @xmath119 matrices as weight matrices . for positive integers @xmath123 and @xmath124
, define the function @xmath125 where .
we now state the main result of this paper .
[ th : main_theorem ] the equivalent channel matrix of the stbc @xmath122 for @xmath3 receive antennas has rank @xmath126 with probability @xmath1 .
see appendix a. the rate of @xmath122 is @xmath127 and the rank of the equivalent channel is less than whenever .
thus , this stbc is singular for all .
[ ex : x_3 ] consider the stbc @xmath128 used in an asymmetric mimo system with receive antennas . in this case , and the equivalent channel matrix @xmath40 is of size . from theorem
[ th : main_theorem ] , we know that the rank of @xmath40 is equal to with probability @xmath1 .
hence , this stbc is singular for @xmath129 receive antennas .
the @xmath130 weight matrices of the stbc @xmath128 are as follows @xmath131 @xmath132 @xmath133 the structure of the upper triangular matrix @xmath134 obtained from the * qr * decomposition of @xmath40 when @xmath9 equals the following randomly generated matrix @xmath135 -0.5688 - 0.8117i & -0.1723 + 1.8282i\\ 0.4926 + 0.0742i & 0.1525 - 0.4716i\\ 0.5905 + 0.5107i & -0.8244 + 0.1325i \end{bmatrix},\ ] ] is given by @xmath136 the sphere decoder uses the matrix @xmath137 obtained from @xmath134 by deleting its last @xmath138 rows which are all zero .
hence , for this particular channel realization @xmath9 , the sphere decoding complexity is of the order of .
the remaining part of this section is concerned with multigroup ml decodable codes .
suppose the code obtained from an stbc @xmath11 with a signal set @xmath19 is @xmath69-group ml decodable for some .
the information symbols @xmath139 can be partitioned into @xmath69 vectors @xmath140 of length @xmath141 respectively such that each symbol vector can be ml decoded independently of other symbol vectors .
there is a corresponding partition of the channel matrix into @xmath69 submatrices @xmath142 , such that @xmath143 in theorem 2 of @xcite it is shown that for any @xmath144 and any channel realization @xmath9 , every column of @xmath145 is orthogonal to every column of @xmath146 . as a direct consequence of this , we have the following proposition .
[ pr : sum_of_ranks_multigroup ] for any @xmath69-group ml decodable stbc and any channel realization @xmath9 , .
since the column spaces of @xmath145 , @xmath147 are orthogonal to each other , the dimension of the column space of @xmath40 is equal to the sum of the dimensions of the column spaces of @xmath145 , @xmath147 .
in this section , we show that all known families of high - rate multigroup ml decodable codes are singular for certain number of receive antennas . using the properties of the rank of the equivalent channel matrix derived in the previous section ,
we now derive the sphere decoding complexities of these known multigroup ml decodable stbcs .
table [ tbl : comparison ] summarizes the results of this section .
the table lists the sphere decoding complexity and the minimum number of receive antennas for non - singularity of @xmath122 , the codes in @xcite , @xcite , @xcite , and the codes in @xcite corresponding to even number of ml decoding groups .
@xmath3 & order of sphere + & antennas & & & ( cspcu ) & for & decoding complexity + & @xmath2 & @xmath7 & @xmath69 & @xmath137 & non - singularity & exponent of @xmath148 + @xmath122 ( theorem [ th : main_theorem ] ) & @xmath149 & @xmath2 & @xmath1 & @xmath127 & @xmath2 & @xmath150 + ren et .
@xcite & @xmath138 & @xmath138 & @xmath129 & @xmath151 & @xmath138 & @xmath152 + ren et al .
@xcite&@xmath149 & @xmath153 & @xmath129 & @xmath154 & @xmath2 & @xmath155 + srinath et al .
@xcite&@xmath156 , @xmath157 & @xmath2 & @xmath129 & @xmath158 & @xmath127 & @xmath159 + & @xmath160 , @xmath161&@xmath162&@xmath163&@xmath164 & @xmath165 & @xmath166 + & @xmath167 , @xmath161 & @xmath168 & @xmath163 & @xmath169 & @xmath170 & @xmath171 + [ tbl : comparison ] the size of the real constellation used is denoted by @xmath35 . for any real number @xmath64
, @xmath172 is defined as @xmath173 .
@xcite contains codes for all and not just even values of @xmath69 . ' '' '' in @xcite , a 2-group decodable stbc for 4 transmit antennas with @xmath174 was constructed . for this code , with the notations as used in , @xmath175 , @xmath176 , @xmath177^t$ ] , @xmath178 and @xmath179 $ ] .
the weight matrix corresponding to @xmath180 is @xmath181 .
since any two weight matrices from different groups are hurwitz - radon orthogonal , i.e. , satisfy , all the matrices in the second group must be skew - hermitian , i.e. , @xmath182 , @xmath183 . as a result
, we can express these weight matrices as @xmath184 , @xmath183 , with @xmath185 being hermitian matrices .
it is clear that for any arbitrary channel realization @xmath9 , @xmath186 $ ] which is non - zero with probability @xmath1 .
hence , @xmath187 with probability 1 . from proposition
[ pr : rank_independent_c_mult ] and theorem [ th : main_theorem ] , @xmath188 with probability @xmath1 .
so , 1 . for @xmath61 , @xmath189 .
since @xmath190 and @xmath191 can be decoded independently of each other , from proposition [ singular ] , the sphere decoding complexity of the first group is independent of @xmath35 while that of the second group is @xmath192 .
consequently , the sphere decoding complexity of the code in @xcite is @xmath193 for 3 receive antennas .
2 . for @xmath194 ,
@xmath195 and the stbc in @xcite is non - singular for 4 or more receive antennas .
hence , its sphere decoding complexity is independent of @xmath35 .
a family of @xmath129-group ml decodable stbcs was constructed in @xcite for @xmath196 , @xmath197 antennas with rate @xmath198 cspcu .
this family includes the rate @xmath199 code of @xcite for @xmath200 as a special case .
the number of symbols in the stbc is . in the rest of this subsection
we show that the sphere decoding complexity is @xmath201 which is polynomial in @xmath35 , and so is large for all , where @xmath202 is the smallest integer greater than or equal to @xmath64 .
note that the sphere decoding complexity is a decreasing function of the number of receive antennas @xmath3 .
the stbcs constructed in @xcite have a block diagonal structure .
the weight matrices for @xmath196 antennas are of the form @xmath203 , where @xmath204 are unitary and .
noting that the stbc is 2-group decodable , denote the set of weight matrices belonging to the first and the second group by @xmath205 and @xmath206 , respectively .
for all the matrices in @xmath205 , @xmath207 is constant , say @xmath208 , and for all the matrices in @xmath206 , @xmath209 is constant , say @xmath210 .
each group contains real symbols .
we will now derive the rank of the submatrix @xmath211 of @xmath40 that corresponds to @xmath205 .
let @xmath212 .
consider the set @xmath213 where @xmath214 .
since @xmath116 is unitary , from proposition [ pr : rank_independent_c_mult ] , the rank of @xmath211 equals the rank of @xmath215 , the equivalent channel matrix corresponding to @xmath216 .
since the multiplication of all the weight matrices of an stbc by a unitary matrix does not affect its multigroup ml decodability , for any matrix @xmath58 belonging to @xmath206 and any , we have @xmath217 where @xmath218 and @xmath219 are block diagonal and are of the form @xmath220 for some unitary matrices @xmath221 and @xmath222 . since @xmath218 and @xmath219 satisfy and are block diagonal , from lemma 1 of @xcite , .
thus , for , @xmath221 is a unitary , hermitian matrix .
next we use proposition [ pr : rank_independent_basis ] to find the rank of @xmath215 .
note that is same as the span of @xmath223 { \bf 0 } & { \bf 0 } \\
{ \bf 0 } & \textbf{d}_{n^2 + 1 } \end{bmatrix}.\ ] ] since @xmath224 are linearly independent , @xmath225 are also linearly independent .
further , the first matrix in is linearly independent of the remaining matrices and hence the dimension of the span of the remaining matrices in is @xmath226 . without loss of generality ,
let us assume that @xmath227 are linearly independent , thus @xmath228 is the space of all @xmath76 hermitian matrices .
then , equals the space spanned by @xmath229 { \bf 0 } & { \bf 0 } \\ { \bf 0 } & \textbf{d}_{n^2 } \end{bmatrix}.\ ] ] from proposition [ pr : rank_independent_basis ] ,
it is enough if we concentrate on the stbc whose weight matrices are given by .
let the channel matrix be partitioned as , where .
we need to compute the dimension of the space spanned by the weight matrices multiplied on the right by @xmath9 which is @xmath230{\bf 0 } \\
\textbf{d}_{n^2}\textbf{h}_2 \end{bmatrix } \right\rangle.\ ] ] with probability @xmath1 , @xmath231 is non - zero and hence the first matrix is linearly independent of the remaining matrices . from theorem
[ th : main_theorem ] , the dimension of the span of the remaining @xmath226 matrices is with probability @xmath1 .
thus @xmath232 equals @xmath233 with probability @xmath1 .
a similar result can also proved for the second ml decoding group , i.e. , for @xmath234 . from proposition
[ pr : sum_of_ranks_multigroup ] , which equals @xmath235 comparing this with , we see that the stbc given in @xcite is non - singular only if .
hence the code is singular for all .
now , the two groups of symbols can be ml decoded independently of each other , and the number of symbols in each group is @xmath236 , with @xmath237 .
thus , the sphere decoding complexity of the stbc is @xmath238 instead of @xmath70 .
hence , _ multigroup ml decodability reduces the sphere decoding complexity even if the stbc is singular_. in @xcite , @xmath129-group ml decodable codes for all @xmath239 and all even @xmath240 were constructed with rate .
the number of symbols per group is . in this subsection , we show that the codes of this family are singular for all receive antennas and that their sphere decoding complexity is @xmath241 . using proposition [ pr : m_less_than_n ] ,
it is clear that these codes are non - singular if and only if .
the structure and derivation of the sphere decoding complexity of these codes is similar to that of the codes from @xcite , which was discussed in section [ sec3b ] .
the weight matrices of the stbcs in @xcite are of the form @xmath242 , where @xmath243 .
for all the matrices in the first group , @xmath244 for some constant matrix @xmath208 , and for all the matrices in the second group , @xmath245 for some constant matrix @xmath210 .
the stbcs constructed in @xcite are such that for each @xmath246 , the @xmath247 submatrices of any two weight matrices belonging to different groups are hurwitz - radon orthogonal .
we consider the case where @xmath248 , @xmath246 are semi - unitary i.e. , .
we derive the sphere decoding complexity only for the first group . using a similar argument
the complexity for the second group can be derived , and it is same as that of the first group .
since @xmath210 is semi - unitary , there exists a unitary matrix @xmath249 such that .
consider the new stbc @xmath250 obtained by multiplying all the weight matrices of the original stbc on the left by .
then , the lower submatrix of all the weight matrices of the second group are of the form @xmath251 .
since the lower submatrix of every matrix in the first group is hurwitz - radon orthogonal to @xmath251 , the weight matrices in the first group of @xmath250 have the following structure : @xmath252 \textbf{v}_1 \\
\textbf{b } \\
\textbf{e } \end{bmatrix}$ ] , where , @xmath58 is an @xmath121 hermitian matrix and .
let be any basis for the space of @xmath121 hermitian matrices over @xmath17 , @xmath253 and be @xmath254 @xmath255 which is a standard basis for the space of complex matrices over @xmath17 .
clearly the space spanned by the weight matrices of the first group of @xmath250 is a subspace of the space spanned by the following @xmath236 linearly independent matrices : @xmath256 from dimension count , it is clear that the matrices in form a basis for the space spanned by the weight matrices of the first group of @xmath250 . for any non - zero channel realization , the subspaces , and
are such that their pairwise intersections contain only the all zero matrix .
thus , the rank of the equivalent channel matrix of the first group of @xmath250 is @xmath257 with probability @xmath1 , and from theorem [ th : main_theorem ] , .
it is straightforward to show that @xmath258 with probability @xmath1 . from proposition [ pr :
m_less_than_n ] , we know that every stbc is non - singular for . we thus consider only the case .
then from proposition [ pr : rank_independent_c_mult ] , the rank of the equivalent channel matrix of the first ml decoding group of the stbc given in @xcite is @xmath259 compare this with the number of symbols in the first group .
thus , for any @xmath85 , @xmath260 with probability @xmath1
. using a similar argument it can be shown that the rank of the equivalent channel matrix of the second group also equals with probability @xmath1 . since
, the two groups can be ml decoded independently of each other , the complexity of sphere decoding the stbc proposed in @xcite is @xmath261 for any @xmath85 and the stbc is singular for all . consider the stbc from @xcite for and .
the rate of this code is and the number of symbols per decoding group is . from the above discussion
it is clear that this stbc is singular for .
the rank of the equivalent channel matrix of each ml decoding group equals @xmath63 with probability @xmath1 and hence the sphere decoding complexity is @xmath193 .
it is interesting to compare this code with the code from @xcite which we have discussed in section [ sec3a ] and example [ ex : fgd ] .
both codes have the same parameters @xmath2 , @xmath137 and both have a sphere decoding complexity that is linear in the constellation size @xmath35 .
however , the code in @xcite is fast - group - decodable and @xmath262 levels can be removed from the sphere decoding search tree of the second decoding group .
hence , after conditioning on the value taken by one of the real symbols ( to account for the reduced rank of the equivalent channel matrix ) , the code in @xcite uses a @xmath263-dimensional sphere decoder to decode the second group of @xmath63 symbols . for decoding each ml decoding group , the code from @xcite , uses a @xmath63 dimensional search tree after conditioning on the value of one of the real symbols .
thus , the sphere decoding complexity of the code from @xcite is less than that of @xcite . in @xcite , delay optimal @xmath69-group ml decodable codes
were constructed for all , , , with rate . in this subsection
we show that the sphere decoding complexity of the codes with even @xmath69 is of the order of @xmath264 and that the codes are non - singular only for .
also in @xcite , non - delay optimal codes with @xmath265 , @xmath266 , were constructed .
we show that the sphere decoding complexity of these codes for even values of @xmath69 is of the order of @xmath264 .
simulation results show that the stbcs in @xcite for odd values of @xmath69 are also singular for certain number of receive antennas .
the number of symbols per group is and let .
we will now derive the rank of the equivalent channel matrix @xmath211 of the stbc corresponding to the first group .
the weight matrices of the first group have a block diagonal structure @xmath267 , where each .
the first block @xmath268 is one of the @xmath226 matrices of the form @xmath269 , where is hermitian , and the remaining @xmath270 matrices @xmath221 , are of the form for some set of @xmath69 unitary matrices .
let us multiply all the weight matrices of the first group on the right by .
clearly , the new set of weight matrices also have a block diagonal structure where @xmath268 is one of the @xmath226 matrices of the form , where @xmath222 is hermitian and the remaining @xmath270 blocks are of the form .
it is straightforward to show that there exists a permutation matrix @xmath271 such that for any .
consider the matrices , @xmath272 , where .
let @xmath273 be any basis for the space of hermitian matrices . from dimension count and the structure of the @xmath274 matrices
, @xmath275 equals the span of the matrices @xmath276 which given by @xmath277 , , @xmath278 , , @xmath279 .
since @xmath280 is unitary , the statistics of @xmath9 and @xmath281 are same . along with propositions [ pr : rank_independent_basis ] and [ pr : rank_independent_c_mult ] , it is thus clear that @xmath232 has the same statistics as the rank of the equivalent channel matrix @xmath282 of the stbc whose weight matrices are @xmath276 .
let the channel realization be , where , @xmath147 .
we are interested in the dimension of which is the span of @xmath283 thus , with probability @xmath1 , @xmath284 equals the sum of @xmath270 and the dimension of the span of the first @xmath226 matrices in .
let us rewrite @xmath231 as @xmath285 , where @xmath286 for @xmath287 .
thus , the dimension of the span of first @xmath226 matrices in is same as that of @xmath288 this in turn , is equal to the dimension of the span of the following matrices @xmath289 from theorem [ th : main_theorem ] , this dimension equals @xmath290 with probability @xmath1 .
hence , @xmath291 with probability @xmath1 .
compare this with the number of symbols per group .
thus , the delay optimal stbcs of @xcite for even values of @xmath69 are singular whenever . similar results on the rank of the equivalent channel matrix can be proved for other ml decoding groups also .
thus , the sphere decoding complexity of these codes is of the order of @xmath264 .
this stbc is singular only for . for , and equal values of @xmath2 , the delay - optimal codes of @xcite and the codes of @xcite have equal rate and the same order of sphere decoding complexity .
consider the @xmath292 , @xmath293 delay - optimal code of @xcite .
the underlying stbc has rate @xmath294 and there are @xmath295 symbols per ml decoding group . from the ongoing discussion , for receive antennas with probability @xmath1 for @xmath246 .
hence , this stbc is singular for and the sphere decoding complexity is @xmath193 .
let us denote the weight matrices of @xmath128 given in example [ ex : x_3 ] by .
then , the @xmath296 weight matrices of the @xmath293 , @xmath292 code of @xcite are given by @xmath297 now consider the @xmath40 matrix corresponding to the randomly generated channel realization @xmath298 -0.0583 + 1.2105i & 0.0708 + 0.6795i\\ -1.3669 - 0.1373i & -0.3850 + 0.0877i\\
-0.3104 - 1.5120i & 0.2146 + 1.0159i\\ -1.2690 - 0.5937i & -0.4245 - 1.3866i\\ 0.5942 + 0.9578i & 0.3465 - 0.1398i\\ -0.6279 - 0.7581i & 0.5228 - 0.8541i \end{bmatrix}.\ ] ] the rank of @xmath40 is only @xmath299 and the structure of the @xmath300 upper triangular matrix @xmath301 obtained from the * qr * decomposition is given in at the top of the next page .
the matrix @xmath301 is of the form @xmath252 \textbf{r}_1 & { \bf 0}_{9 \times 10 } \\
{ \bf 0}_{1 \times 10 } & { \bf 0}_{1 \times 10 } \\
{ \bf 0}_{9 \times 10 } & \textbf{r}_2 \\ { \bf 0}_{5 \times 10 } & { \bf 0}_{5 \times 10 } \end{bmatrix}$ ] , where . the all zero submatrix at the upper right corner of @xmath301 is due to the @xmath129-group ml decodability property of the code .
the two sphere decoders corresponding to the two ml decoding groups use the matrices @xmath302 and @xmath303 respectively .
clearly the rank of @xmath302 and @xmath303 is @xmath130 , and hence the sphere decoding complexity is @xmath193 .
@xmath304 ' '' '' the non delay - optimal code for has rate and number of symbols per group .
the weight matrices of the first group are of the form @xmath305 , where each .
the first block @xmath268 is one of the @xmath226 matrices of the form @xmath269 , where is hermitian , and the remaining @xmath270 matrices @xmath221 , are of the form for some set of @xmath69 unitary matrices . using an argument similar to the one used with delay - optimal codes
, it can be shown that the sphere decoding complexity of the non delay optimal codes is of the order of @xmath264 and that the codes are non - singular only for . for @xmath292 and equal
values of @xmath2 and @xmath7 , the non - delay optimal codes of @xcite and the codes of @xcite have the same rate and sphere decoding complexity .
in this paper we have introduced the notion of singularity of stbcs and showed that all known families of high rate multigroup ml decodable codes are singular for certain number of receive antennas .
the following facts which were not known before have been shown .
* though the @xmath54 , @xmath306 code of @xcite and the @xmath54 , @xmath307 code of @xcite have identical rate of @xmath151 cspcu , the sphere decoding complexity of the code from @xcite is less than that of the code from @xcite .
* for , and equal values of @xmath2 , the delay - optimal codes of @xcite and the codes of @xcite have equal rate and the same order of sphere decoding complexity . * for equal values of @xmath2 and @xmath7 , the codes in @xcite and the non - delay optimal codes of @xcite have identical rate and sphere decoding complexities .
the results and ideas presented in this paper have brought to light the following important open problems . * is there an algebraic criterion that ensures that a code is non singular ?
for example , is every code with non vanishing determinant also non - singular for all ?
* do there exist high rate multigroup ml decodable codes that are non - singular for arbitrary values of @xmath3 ?
* do there exist singular high rate multigroup ml decodable stbcs with lower sphere decoding complexity than that of the known codes ?
this work was supported partly by the drdo - iisc program on advanced research in mathematical engineering through a research grant , and partly by the inae chair professorship grant to b. s. rajan
proof of theorem [ th : main_theorem ] from proposition [ pr : m_less_than_n ] it is clear that the theorem is true for .
thus , we will only consider the case . before giving the proof of theorem [ th : main_theorem ] we present two results which are used in the proof .
let @xmath308 be the @xmath2 columns of the matrix @xmath309 .
we first prove the result for .
let the channel realization , where and .
now , consider the matrix .
we have , which is non - zero w.p.1 @xcite .
thus , w.p.1 , the columns of @xmath311 are linearly independent over @xmath84 , and since @xmath312 is the first column of @xmath311 , this means that w.p.1 , @xmath312 does not belong to the column space of the matrix @xmath9 .
now consider any .
let @xmath271 be an permutation matrix such that @xmath313 .
since @xmath271 is full - ranked , @xmath310 belongs to the column space of @xmath9 if and only if @xmath312 belongs to the column space of @xmath314 .
since @xmath271 is unitary , the distribution of @xmath9 and @xmath314 are one and the same , and hence the probability that @xmath312 belongs to the column space of @xmath314 is @xmath315 .
thus , with probability @xmath1 , @xmath310 does not belong to the column space of @xmath9 . for a given channel realization @xmath9 ,
since @xmath9 is of rank @xmath3 w.p.1 , the dimension of @xmath316 over @xmath17 is equal to @xmath317 w.p.1 .
let and for , let be the vector space homomorphism that sends the vector @xmath318 to the real number @xmath319 .
we are interested in the dimension of the subspace of @xmath316 which is composed of vectors whose @xmath320 component is purely real , i.e. , in the dimension of @xmath321 . for any given @xmath9 , and
hence @xmath324 is either @xmath315 or @xmath1 .
suppose , @xmath325 , then , there is no vector @xmath318 in @xmath316 such that @xmath326 is non - zero because if such a @xmath318 exists , the vector @xmath327 belongs to @xmath316 and the imaginary part of its @xmath320 component is @xmath328 , and thus @xmath329 , which is a contradiction . since , for all the vectors in @xmath316 , the @xmath320 component is @xmath315 , we have @xmath330=\textbf{0}\}.\ ] ] thus , the dimension of the column space of @xmath9 and @xmath331 $ ] are the same .
this means that @xmath310 belongs to the column space of @xmath9 . from proposition
[ pr : appendix_first ] , @xmath310 belongs to the column space of @xmath9 w.p.@xmath315 and hence @xmath325 w.p.@xmath315 .
thus , @xmath329 w.p.1 . from rank - nullity theorem , w.p.1 .
_ proof of theorem [ th : main_theorem ] : _ let the weight matrices of the stbc @xmath122 be and let the space of hermitian matrices over @xmath17 be given by
. for a given channel realization @xmath9 , let be the @xmath17-vector space homomorphism that sends the matrix @xmath23 to @xmath332 . clearly , @xmath333 is equal to the dimension of the subspace @xmath334 over @xmath17 .
since @xmath334 is isomorphic to @xmath335 as vector spaces , we have @xmath336 thus , it is enough to show that w.p.1 . let and let denote the @xmath2 rows of @xmath23 .
then , @xmath337 satisfies @xmath338 , and since the @xmath23 is hermitian , the first component of @xmath337 is purely real . from proposition [ pr :
appendix_second ] , whose dimension is @xmath339 w.p.1 . given a choice of @xmath337 , since @xmath23 is hermitian , the first component of @xmath340 equals the conjugate of the second component of @xmath337 , and the second component of @xmath340 is purely real . as a result of these restrictions and since , i.e. , @xmath341 , @xmath340 belongs to a coset of the subspace of @xmath316 whose dimension is .
similarly , given a choice for , where , the first @xmath342 components of @xmath343 are fixed and the imaginary part of the @xmath344 component is zero . hence , @xmath343 belongs to a coset of a subspace of @xmath316 with dimension .
thus , the dimension of @xmath345 equals @xmath346 with probability @xmath1 .
this completes the proof .
+ @xmath347 k. p. srinath and b. s. rajan , `` low ml - decoding complexity , large coding gain , full - rate , full - diversity stbcs for 2x2 and 4x2 mimo systems , '' _ ieee j. sel .
topics signal process . _ ,
3 , no . 6 , pp .
916 - 927 , dec . 2009 | in the landmark paper @xcite by hassibi and hochwald , it is claimed without proof that the upper triangular matrix @xmath0 encountered during the sphere decoding of any linear dispersion code is full - ranked whenever the rate of the code is less than the minimum of the number of transmit and receive antennas . in this paper , we show that this claim is true only when the number of receive antennas is at least as much as the number of transmit antennas . we also show that all known families of high rate ( rate greater than @xmath1 complex symbol per channel use ) multigroup ml decodable codes have rank - deficient @xmath0 matrix even when the criterion on rate is satisfied , and that this rank - deficiency problem arises only in asymmetric mimo with number of receive antennas less than the number of transmit antennas . unlike the codes with full - rank @xmath0 matrix , the average sphere decoding complexity of the stbcs whose @xmath0 matrix is rank - deficient is polynomial in the constellation size .
we derive the sphere decoding complexity of most of the known high rate multigroup ml decodable codes and show that for each code , the complexity is a decreasing function of the number of receive antennas . |
the solar atmosphere extends from the photosphere and chromosphere through the transition region into the corona . in the photosphere and lower chromosphere , where the fraunhofer absorption lines are formed
, the plasma is usually dominating the magnetic field , which is frozen - in .
models of the upper part of the convection zone and the photosphere have to account not only for the interaction of the plasma and the magnetic field ( magneto - convection ) , but also for the radiative transfer ( e.g. * ? ? ?
higher up in the atmosphere the plasma becomes optically thin and the radiative transfer reduces to a radiative loss function , however , other processes such as heat conduction become of importance . what is common for the description of ( almost ) all interesting phenomena on the sun is the interaction of the magnetic field and the plasma
for many models one assumes the magnetic field only to provide a flow channel , e.g. in penumbral filaments or coronal loops . despite the great success of many of these
one - dimensional models one has to account properly for the complex magnetic structure to understand the true nature of the solar atmosphere , and this implies to advance to more complex 3d models .
traditionally the 3d magnetic structure of the outer atmosphere has been described by extrapolations of the magnetic field , from simple potential field models to very complex and computationally intensive force - free extrapolations ( e.g. * ? ? ?
such models are of vital importance to understand the large scale structure , but are also very helpful when , e.g. , investigating how the magnetic field is rearranging itself . during the dynamic phases of the reconnection process @xcite
, however , the field certainly becomes non - force - free , and thus one also needs a model combining the magnetic field and the plasma , i.e. , a mhd model . due to the advancement of computer power
such 3d mhd models of structures of the solar upper atmosphere became available during recent years .
of course , these models can not describe individual structures , as e.g. a coronal loop , in such great detail as 1d models , but they can properly account for the interaction and coexistence of different structures , which is of great importance when investigating a highly structured object such as the solar atmosphere . also on global scales
the interaction of magnetic fields and the plasma is of great importance .
for example , in their global models @xcite investigate the large scale evolution of the ( outer ) corona and its connectivity to heliosphere ( see also the review of * ? ? ? * ) .
they show that always different footpoints are connected to field lines going into the heliosphere , which is due to on - going reconnection .
the dynamic reconfiguration processes can not be properly handled by a sequence of magnetic field extrapolations .
opening up of loops , reconnection of open regions to form loops or interchange reconnection between open and closed regions does happen all the time .
such scenarios have been sketched in the past , of course , but only now global mhd models could show that they indeed operate .
the present paper concentrates on the upper atmosphere of the sun , i.e. effects within the photosphere and chromosphere as well as on global scales will not be discussed .
the aim of this paper is to discuss coronal models for some structures , namely the closed field structures such as ( moderately ) active regions , and open structures like coronal holes . for illustrative purposes the structure of the low corona
is sketched in fig.[f : sketch ] .
it shows the co - existence of closed magnetic structures of various sizes : small loops with lengths below 5 mm connecting magnetic patches within the chromospheric network , low - lying loops with length below 20 mm spanning across network cells and finally larger loops connecting larger magnetic patches over large distances .
the base regions of these large loops might be funnel - type , just like the magnetically open coronal funnels .
the latter should be most prominent in coronal hole , but also might be present in the quiet predominantly magnetically closed corona .
the paper will start with an overview of how to actually measure coronal magnetic fields ( sect.[s : obs.mag ] ) , even though we know only little about this currently . in the following sect.[s : loops ] it will be outlined how observations ( and models ) require to advance from 1d to 3d models and why one really needs 3d models ( sect.[s : why ] ) .
the main part of the paper , sect.[s : box ] , will discuss 3d coronal box models and the vacuum ultraviolet ( vuv ) spectra synthesized from them , which allows a detailed comparison to observations .
open structures , especially the origin and acceleration of the fast wind in coronal holes will be the subject of sect.[s : open ] , before sect.[s : conclusions ] concludes the paper .
before discussing modeling of the solar atmosphere including the magnetic field , the attention should be drawn on how to actually observe the magnetic field in the corona .
there are various techniques providing the coronal magnetic field .
direct coronagraphic measurements of the stokes vector e.g. in the infrared lines of fe at 10747 and 10798 applying the zeeman effect show magnetic fields in bright loops being of the order of 20 gauss @xcite .
such measurements can be applied only for regions seen above the limb , of course , and are most valuable for the large - scale structure of the corona .
smaller areas , especially low in the corona , are not accessible through this technique , because they are occulted by the design of the coronagraph .
radio observations allow to investigate the coronal magnetic field also on the disk , e.g. through the zeeman effect @xcite .
radio interferometers such as vla provide a remarkable spatial resolution of 12arcsec , and experiments for the future fasr facility showed that the magnetic field strength could be retrieved in a reliable way @xcite .
the problems here are the detailed interpretation that the emission at a given frequency does originate from a complex volume in the corona , and that these observations are of value mainly for active regions .
another valuable tool are infrared observations in the he ( 10830 ) triplet using zeeman and hanle effect giving magnetic fields for emerging flux regions @xcite .
however , he is not a coronal line and can thus only indicate magnetic fields in just newly emerging flux regions , which still carry cool plasma .
extrapolation techniques based on photospheric and/or chromospheric measurements are also a very valuable tool to investigate coronal magnetic fields . as a rule of thumb one might trust these extrapolation on scales typically comparable to or larger than a fraction of the chromospheric network , i.e. , some 510 mm . on smaller scales still the expansion of the magnetic field into the chromosphere , i.e. the equilibrium of magnetic and gas pressure ,
can be expected to dominate .
however , whenever interesting events occur , such extrapolations , no matter how fancy , can not be trusted completely , as then the assumption of a force free state is most likely violated . while in such events the overall field structure on the larger scales mentioned above might not change dramatically , the force - free violation might be more severe on smaller scales in a transient manner
. this will be certainly true on scales on which explosive events do occur ( i.e. some mm ; * ? ? ?
* ) as well as on scales not yet resolvable , e.g. in the case of nanoflares . thus such extrapolations are important and helpful to understand the overall structure and long - term evolution , but real measurements of the coronal magnetic field with a spatial resolution of 1 mm and better are desperately needed in order to pin - point the relevant processes during phases when the heating mechanism is showing itself through a dynamic non - force - free event , as in flares , explosive events or nanoflares . while the above mainly concerns closed field regions , the same is also true for magnetically open region as found in coronal holes .
following the idea of e.g. @xcite magnetic reconnection of open field with bipolar network flux is the energy source for the heating of the open corona and the wind acceleration .
it would be of high interest to detect the non - force - free magnetic reconfiguration predicted by this furnace model through the direct measurement of the magnetic field .
the ideal tool to investigate the magnetic field in the corona is to use lines which are formed in the transition region and corona , i.e.to use the same vuv emission which is currently widely used to investigate the corona with imagers such as trace or eit or spectrometers such as sumer or cds .
of course , because of the complicated spatial structure of the corona and the high variability , the interpretation of the polarized light in the vuv will not be possible in a simple and straightforward way , and is thus confronting us with the same problems as radio observations .
the coronal models , which become more and more elaborate and realistic , will be of pivotal importance for a proper interpretation of vuv and radio spectro - polarimetric data .
the importance of uv and vuv spectro - polarimetry from space has been emphasized by @xcite , who suggest to use the lyman series of hydrogen for an investigation of the hanle effect .
another approach is to measure the stokes vector in strong lines such as the c doublet at 1548 and 1550 , which is formed in the transition region .
this is suggested by @xcite , who are currently building an instrument for a series of rocket flights , the first probably in spring 2007 for a test of the system ( solar ultraviolet magnetograph investigation , sumi ) .
because of the limited count rate during the rocket flight , they expect to get a signal only in stokes v , which would be in itself a great step forward , if they succeed . in the future we have to aim at a combination of spectro - polarimetry in the radio and the vuv . in the radio
such techniques already exist , but they have to be refined , also with respect to spatial resolution .
for vuv spectro - polarimetry we will certainly have to wait some time until space instrumentation will be available , but this is the instrument to aim at when investigating the magnetic structure of the corona .
the most conspicuous basic building block of the corona is the coronal loop .
observations show that more or less semi - circular shaped loops are the most prominent structure in the corona , and they are found in active regions , after flares , in quiescent regions in the network , etc . therefore most 1d models for the solar corona dealt with loop models , one of the most well - known being the rtv models @xcite . even though it is not clear what the `` micro - structure '' of an observed loop would be , i.e. if it consists of many individual strands or if it is a bundle of parallel field lines @xcite , many loop models exist solving the mass , momentum and energy balance along a loop - shaped 1d structure , some now with adaptive mesh refinement and even a self - consistent treatment of the ionization and radiation ( e.g. * ? ? ?
* ; * ? ? ?
such loop models allow a very detailed description of the thermal and dynamic properties within the loop ( or strand ) modeling complex non - linear processes , e.g. catastrophic cooling @xcite , which is also found in observations @xcite .
1d models for open structures will be discussed in sect.[s : onedwind ] .
the main disadvantage of these 1d loop models is that they ( usually ) are not able to treat the heating process in a physical way .. ] since the early rtv work almost all 1d models assume a parameterized form of the distribution of coronal heating , e.g. exponentially decaying with height , i.e. not really incorporating a physical heating mechanism into the model .
however some models made an attempt to include a distribution of heat into the 1d models as found from other studies . in the mid 1970ies the skylab observations showed that the chromospheric network is expanding to higher temperatures @xcite , which lead to assume a funnel - type structure of the transition region and low corona @xcite .
however , it became clear later that these funnels _ alone _ are not a good representation of the corona , even though they are an important ingredient ( cf .
fig.[f : sketch ] ) .
for example in the magnetically closed corona the emission measure @xmath1 , i.e. a measure of the emissivity of the corona at a given temperature , could not be reproduced by the funnel models . because of the ineffective heat conduction at low temperatures the 1d models give an extremely thin transition region , i.e. a very steep temperature gradient , and thus in the model there is not enough material at low temperatures to account for the observed increase of the emission measure towards lower temperatures below some @xmath2k .
@xcite proposed that loops at various temperatures could account for that increase and through this introduced a complicated magnetic structure of the low corona , sometimes also called `` magnetic junkyard '' ( cf .
fig.[f : sketch ] ) .
later it was proposed that signatures in the spectra of vuv emission lines support this scenario @xcite .
furthermore there is ample observational evidence from vuv line emission maps acquired with sumer that low - lying loops with lengths below 1020 mm do exist ( e.g. * ? ? ?
* ) . however , all these scenarios are not really based on a physical model , but are merely trying to find a plausible way to draw a sketch of the magnetic structure of the low corona . in open field regions such as coronal holes
it seems problematic that the above scenario for the ( predominantly ) closed field regions holds .
there other processes might be more important .
recently @xcite have shown that the ( fast ) solar wind outflow of the plasma along funnel type structures might well result in significant violation of the ionization equilibrium by lifting a sizable fraction of neutral material higher up into hotter regions of the funnel .
this can then produce an increased emission in ly-@xmath3 which agrees well with observation .
however , it still has to be explored to which extent this process will predict the observed emission from the transition region , i.e. of lines formed at a couple of 100000k . while large - scale structures such as a ( polar ) coronal hole or an active region complex seem to be rather stable for days and weeks ,
on smaller scales of below 10 mm the corona is highly dynamic .
when analyzing the spatial magnetic structure on these smaller scales this variability is adding to the complexity of the problem .
for example , when modeling an explosive event , it is usually assumed that the `` background '' magnetic structure remains the same , while only a small perturbation is shuffling around the magnetic field leading to a reconnection event ( e.g. * ? ? ?
when investigating observations with imagers or spectrographs , one sees a very strong variability in time indicating that the `` natural '' state of the the transition region and corona on scales below 10 mm is a dynamic one @xcite .
one could argue that a composite model consisting of a large number of loops and open structures would be sufficient to describe the solar corona .
furthermore one could argue to neglect the magnetically open structures in these models as they provide only a minor contribution to the emission in x - rays or the far vuv , i.e. in the wavelength bands accessible to yohkoh , or the eit or trace coronal channels . as the corona and transition region
are built up by ( loop - like ) magnetic structures which are rooted in the solar photosphere , one could use an extrapolation of the magnetic field to identify the loops and then investigate the resulting structure . for certain problems
this is certainly a good approach .
for example @xcite did a global extrapolation of the observed photospheric magnetic field to define the loops and used a simple scaling law for the flux of mechanical energy @xmath4 into the loop , @xmath5 . here
@xmath6 is the magnetic field at the base of the structure , @xmath7 its half length , @xmath8 $ ] a factor accounting for reduced heating in sunspots , and @xmath0 and @xmath9 are free parameters . each loop was then described in a static 1d model .
the authors then calculated how this constructed multi - loop corona would appear in an x - ray observation ( yohkoh sxt ) and compared it to the real observation . by varying @xmath0 and @xmath9
they found that a heating scaling with @xmath10 and @xmath11 would give the best fit .
even the best fit is not really representing the real corona , which is not surprising as it is based on a very simple model .
nevertheless it can be used to estimate how the energy input into the corona is roughly scaling with the magnetic structures .
most important , this kind of model opens the possibility for more realistic global models of stellar corona , which are not accessible to direct observations . for other purposes one - dimensional models , or such multi - loop models would not be sufficient .
for example @xcite showed that depending on the distribution of internetwork magnetic flux only some 50% of the coronal magnetic field above the quiet sun does come from network magnetic patches .
this implies that the simple idea of a coronal loop being rooted in a single ( or even a few ) patches of strong magnetic flux in the photosphere is not applicable .
furthermore @xcite could show that based on current instrumentation we are not even able to determine where the coronal field is really rooted , on a scale of a fraction of a super - granule ( say 5 mm ) . for only small differences in the distribution of weak internetwork flux elements one gets radically different connections from the photosphere into the corona .
therefore when describing the corona in more detail on a scale corresponding to a fraction of a single active region or the chromospheric network , one has to account for the complex magnetic connectivity , because a simple picture of a coronal structure rooted in a single ( or few ) strong magnetic patch(es ) does not really apply . on top of that the problem is even more complicated , because the distribution of magnetic flux in the photosphere changes constantly on time scales from minutes ( granules , @xmath121 mm ) to days ( super - granules , @xmath1220 mm ) , which is shorter than the lifetime of many loops . loop arcades , e.g. as seen by trace , overall last longer ( many days ) than the nano - flaring processes associated with them , so they appear to be rigid , or solid
. however , the magnetic connectivity within the loop structures might constantly change while the loop is sitting seemingly still , i.e. the many strands of the loop could constantly change their identity .
currently it is not clear if , and if yes to what extent , this affects the appearance of the loop as a whole .
therefore we might ask ourselves the question what sense the concept of a coronal loop as a rigid magnetic loop really makes . in consequence , to describe the structure of the transition region and corona , one has to account for the changing `` boundary conditions '' in the photosphere and include a proper treatment of the 3d structure of the magnetic field .
one of the most important ingredients for a coronal model is a proper energy equation . because of the strong flows observed in the active regions as well as quiet corona ,
the enthalpy flux can play a major role . and only if the energy balance , including furthermore the energy input , heat conduction and radiation are treated well enough , the density of the corona will be correct ( as good as possible ... ) .
this is because basically the energy input sets the pressure of the coronal structure : for larger energy input more energy will have to be radiated in the transition region . when the heating rate increases this consequently leads to ( chromospheric ) evaporation ,
while the coronal temperature changes only little as the heat conduction acts as a thermostat through the strong dependency on temperature ( @xmath13 ) .
thus only a proper energy equation will assure proper densities and consequently proper velocities within the computational domain .
a severe problem is that the energy dissipation processes are occuring on scales much smaller than currently observable or resolvable with the mhd - type models to describe e.g. an active region complex or a coronal loop .
thus in terms of kinetic physics it is questionable is a `` proper energy equation '' in fluid terms can even be defined ( e.g. * ? ? ?
. however , one might ( or have to ) hope that when describing the corona with a resolution of the current mhd models ( some km in 1d , some 100 km in 3d ) , the energy deposition as a function of space and time is comparable to the result a micro - physical model smoothed to the mhd scales would give . also the resistivity in the mhd models is basically set by the ( spatial ) resolution they can achieve . due to the limitations for the magnetic reynolds number this is a common property for all mhd - type models .
however mhd studies with different spatial resolutions and resistivities showed that the energy deposition does not change significantly over a range of scales ( for resolution and resistivity ) accessible to mhd experiments @xcite .
future work will have to show to what extent the mhd approach is a good representation for the real sun , but currently it is the most appropriate ( and only ) tool to describe structures such as active region systems or loop complexes .
the basic test for every coronal model has to be whether it is matching with observations .
the most straight - forward way for such a test is to compare observable quantities such a line intensities or shifts synthesized from the model to real observations , as this eliminates the many problems of inversions of vuv spectral data ( e.g. * ? ? ?
in the following a 3d mhd coronal box forward model will be discussed which compares very well with observations concerning e.g. the emission measure , line shifts or temporal variability .
a 3d mhd model presented by @xcite includes the atmosphere from the photosphere to the lower corona in a 60@xmath1460 mm horizontal times 37 mm vertical box .
it solves the mass , momentum and energy balance , the latter including classical heat conduction @xcite and optically thin radiative losses as piecewise power laws .
the temperature of the chromosphere is kept near a prescribed profile by newtonian cooling .
initially the magnetic field is given by a potential field extrapolation with the lower boundary as observed by mdi / soho for an active region .
further on in the simulation the field evolves self - consistently and becomes non - potential , of course .
the system is driven by horizontal motions at the lower boundary .
the flow field is constructed using a voronoi - tessellation technique @xcite and reproduces the typical pattern of the granular motions of the sun @xcite . by this
also the power spectra of the velocity and the vorticity are reproduced .
the spacing of the ( non - uniform ) grid goes down to 150 km and is as good it can be given the size of the structure modeled here ( 60 mm ) . of course all the shortcomings of mhd models
as discussed above apply here . for the upper boundary condition it is assumed that the magnetic field above the computational domain is potential , which , of course , prevents the development of larger open magnetic structures ( and was not the aim of that study ) .
this has to be kept in mind when discussing the doppler shifts of coronal lines ( sect.[s : diagnostics ] ) and the apparent lack of hot blueshifted material , i.e. a wind .
this procedure leads to a heating of the corona just in the way @xcite suggested .
the field line braiding gives high coronal temperatures of a million k , and a system of hot loops is forming , connecting the magnetic concentrations of the active region .
furthermore the system reaches some sort of quasi - stationary state , with large fluctuations in time and space , or in other words , the system is quite dynamic .
the detailed discussion of the mhd results can be found in @xcite . based on the above mhd model results for the density and temperature @xcite calculated the emissivity at each grid point under the assumption of ionization equilibrium using the atomic data base chianti . assigning a spectral line profile at each grid point with a line width corresponding to the thermal width
, one can integrate the spectra from the 3d box along a line of sight , e.g. the vertical .
this results in 2d maps of spectral profiles , which can be analyzed in the same way as observations , e.g. through maps of line intensities or shifts . for a detailed discussion of the synthesis of the emission lines from the transition region and corona and
the assumptions and limitations see @xcite .
figure [ f : images ] shows in the middle panel the vertical magnetic field at the bottom of the computational box , i.e. in the photosphere . to
the left and right are maps in line intensity and line shift as derived from the spectra computed from the mhd model .
they represent a view from the top onto the box , corresponding to an observation at disk center .
as found with observations the spatial structures in the transition region line c ( 1548 ) at 10@xmath15k are much smaller and finer than in the corona seen in mg ( 625 ) formed at 10@xmath16k . in the following a short overview
will be given how vuv spectra synthesized from a 3d coronal box model can be used to test the model .
it should be stressed that the model results presented in the following show the best overall match to the observations published so far for the differential emission measure , doppler shifts or rms fluctuations . ] for each time step @xcite used the radiances of a number of vuv emission lines synthesized from a mhd model to perform a differential emission measure ( dem ) analysis , i.e. they used the maps of the spectral profiles ( `` synthetic observations '' ) as described in the previous section and integrated the line profiles over the whole map .
thus the line radiances represent the total emission of the given vuv lines integrated over the whole computational domain . as with real observations
one can then use an inversion procedure to derive the differential emission measure ( dem ) , e.g. with the help of the procedures provided with the chianti atomic data package @xcite .
fig.[f : dem ] shows the resulting dem curve as a solid line for a single time step derived from the intensities of a given set of lines , which characterizes the ( average ) small active region simulated in the mhd model . for comparison
the thick dashed line shows the inversion of actually observed intensities of the same lines .
the match of the model with the observations is remarkable , especially , the 3d mhd model used in the @xcite study @xcite is reproducing the observed trend of the emission measure below @xmath17k , while previous ( 1d and 2d ) models failed to predict this increase of dem towards lower temperatures ( cf .
sect.[s : loops ] ) . in the above model
the increase towards the low transition region is caused by numerous low - lying ( intermittent ) cool structures .
this confirms the scenario outlined by @xcite , in which a hierarchy of large and small , hot and cool loops exist in the transition region and corona as already discussed in sect.[s : loops ] .
despite the variability of the low corona , the emission measure ( averaged over the computational domain , viz .
the active region ) changes only little with time . ] since the discovery of the systematic transition region line redshifts by @xcite , the explanation of these persistent doppler shifts has been a challenge for modelers .
the discovery that the lines formed in the ( low ) corona show a net blueshift by @xcite and @xcite using sumer / soho data added a new quality to this challenge .
the trend of the net doppler shift at quiet sun disk center is shown in fig.[f : doppler ] as a thick dashed line ( data compiled from * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in order to derive the average doppler shift corresponding to disk center observations one can investigate the doppler maps of the synthetic spectra with a vertical line - of - sight , i.e. when looking at the computational box from straight above ( cf . ) .
the average line shifts as computed by @xcite for two different time steps 7min apart are plotted in fig.[f : doppler ] as bars and rectangles ( the heights of the bars and rectangles indicating the scatter of the doppler shifts ) .
the line shifts are caused by the dynamic response of the atmosphere to the energy input due to the braiding of magnetic field lines .
the flows are partly induced through energy deposition in and subsequent expansion of the corona , partly from evaporation of chromospheric material due to heating at low heights and the heat flux from the corona into the chromosphere . throughout the transition region ( @xmath18k )
the match of the observed doppler shifts with those of the spectra synthesized from the model is quite good .
it should be stressed here that no fine - tuning was applied , but these doppler shifts follow naturally from the driving of the corona through the footpoint motions of the magnetic field lines . in the low corona ( @xmath19k ) the synthesized spectra
do not show the blueshifts as they are observed @xcite , but this problem might be overcome ( at least partly ) by a more appropriate handling of the upper boundary condition . ]
based on the synthesized maps of line intensities and shifts generated from a line of sight integration along the vertical direction , the rms fluctuations at each spatial pixel within the respective map have been evaluated by @xcite .
figure [ f : rms ] shows their results for a number of vuv lines as a function of line formation temperature .
the diamonds show the ( spatially ) averaged values for the rms fluctuation for each line and the bars represent the scatter of the rms values ( standard deviation ) .
this is exactly the same procedure as used by @xcite to reduce their observational data from cds and sumer on soho .
the trends they found with their observations is over - plotted as thick dashed lines in fig.[f : rms ] .
it has to be emphasized that the spatial resolution of the maps in line intensity and shift derived from the mhd model has been reduced to match the spatial resolution of the soho instruments , as this has an effect on the absolute values of the rms fluctuations ( the trend with temperature , however , does not depend on the spatial resolution ) .
together with the good match between the average doppler shifts and emission measure from observations and the spectra computed from the 3d mhd coronal box model presented above , this is yet another strong indication that the underlying 3d coronal model is a good description for the coronal heating mechanism . as a function of line intensity ( a : c ,
the dashed lines show the median value of plasma@xmath0 in a given interval of intensity .
please note that the 2d histograms are plotted on a log - scale .
following @xcite .
[ f : beta ] ] to investigate the interaction of the magnetic field and the plasma one can discuss plasma@xmath0 , i.e. the ratio of the gas pressure to the magnetic pressure , @xmath20 .
fig.[f : beta ] shows the relation of @xmath0 to the synthesized transition region ( a ; c ) and coronal ( b ; mg ) emission within the computational box as done by @xcite .
the plots show the 2d histograms of the plasma beta as a function of ( normalized ) intensity of the respective line , i.e. the color coding represents the fraction of the volume where one finds the respective combination of plasma-@xmath0 and normalized intensity .
the median relations between plasma-@xmath0 and the normalized intensities are over - plotted as dashed lines .
the transition region and corona are mostly in a low-@xmath0 state , however with noticeable exceptions in the coronal part !
it is clear that in the whole transition region the assumption of a low-@xmath0 plasma is very good ( fig.[f : beta]a ) .
this is different , however , in the corona when relating @xmath0 to the mg emission ( fig.[f : beta]b ) .
a large fraction of the not - low - beta plasma in the corona is at low densities , i.e. at @xmath21 , where the magnetic field is very weak , too .
as these are regions of low emission , one might be inclined to disregard this .
however , even in the volume with a relative intensity larger than the median intensity ( i.e. @xmath22 in fig.[f : beta]b ) contributing @xmath2390% to the total coronal mg emission , some @xmath235% of this bright material has values of @xmath24 @xcite .
these are regions of rather high densities but low magnetic field , and they can be found e.g. in regions along magnetic neutral lines across which the direction of the magnetic field changes and thus strong currents heat the plasma and finally cause the density to increase . in conclusion , in significant patches of the corona , even above active regions , @xmath0 is in fact not much less than unity .
this again emphasizes that for `` good '' coronal model it is not enough to only extrapolate the magnetic field and then solve a lot of 1d loop - like problems along each field line .
instead , one has to account for the interaction of the plasma and the magnetic field in a more complex 3d model .
so far the discussion concentrated on magnetically closed structures , such as active regions or the chromospheric network . in magnetically open regions the plasma is not trapped by the magnetic field , but can escape along magnetic field lines .
this is the basic reason why coronal holes , i.e. the source region of the fast solar wind , appear dark . because a large part of the energy goes into acceleration of the wind
, there is less energy to heat the plasma .
consequently the temperature is a bit lower than in the quiet sun , typically below @xmath25k @xcite .
the effect of the reduced heating rate is much stronger on the pressure , which results in a lower density in the corona above the holes we see less emission from the coronal holes . as for the quiet and active sun structures , which are dominated by loops , the first models for the open corona , which is the same as for the fast wind , were 1d models stretching from the base of the corona into interplanetary space , with an area expansion factor accounting for the ( super ) radial expansion of the magnetic field . as with the loops these models used a parameterized form of the energy input , which was typically concentrated close to the sun . as mentioned earlier , in order to properly understand the mass loss , especially to describe the mass loss rate self - consistently set by the energy input , one has to include a transition region to account for the heat flux back to the sun @xcite . only then one can describe properly the interplay between coronal heating , densities , temperatures and solar wind acceleration , as done first by @xcite . however , as for the 1d loop models ( cf .
sect.[s : loops ] ) , these models did not include a physical heating mechanism .
in contrast , @xcite put forward an idea that small - scale reconnection in the magnetic concentrations of the chromospheric network ( i.e. between open and closed field lines ) leads to high - frequency alfvn waves which propagate upwards into the corona , a concept also called magnetic furnace .
these waves can then resonantly interact with the protons and heavy ions in the open corona and drive the wind as has been described by @xcite and @xcite .
furthermore one can also account for the expansion of the magnetic field directly above the magnetic concentrations in the chromosphere forming large funnels @xcite .
these 1d models including a physical mechanism for the heating and acceleration of the plasma describe a situation where the material is continuously and constantly accelerated from the chromosphere out into the solar wind ( e.g. * ? ? ? * ) . by this
they implicitly assume that the magnetic configuration from the upper chromosphere into the solar wind is rather stable !
this would strongly contradict the scenario outlined for the closed field regions in the previous section , where the magnetic field in the photosphere is constantly shuffled , resulting in a continuous change also higher up .
it has been known for a while that the fast wind is originating from coronal holes , and observations of blueshifts in coronal lines within coronal holes confirmed this @xcite .
even the transition region lines of c showed a higher fraction of blueshifted regions in coronal holes than in the quiet sun @xcite .
the strongest blueshifts are found in the darkest coronal regions @xcite and they seem to be concentrated at intersections of network lanes @xcite .
a comparison of the doppler shift patterns to the photospheric magnetic field showed that these strong blueshifts , interpreted as the base of the outflow , coincide with strong magnetic concentrations in the network @xcite .
these data seemed to support the wind scenario described above , namely that funnel - type structures emerging from strong network magnetic patches define the flow channels of the wind reaching out into the interplanetary medium .
one can investigate the magnetic connectivity from the photosphere to the corona in coronal holes and compare this to the doppler shift signals of transition region and ( low ) coronal lines .
this shows that the blueshifts in coronal lines coincide with the magnetic funnels , but that at those locations no systematic blueshifts can be found in transition region lines within the instrumental uncertainties @xcite .
based on this observation @xcite argued that the plasma of the solar wind outflow has to be injected somewhere between the formation heights of the transition region and coronal lines they inspected ( c and ne ) , i.e. somewhere around 10 mm . if this is correct , it changes dramatically our view of the formation of the wind at its very base , because it would challenge the old paradigm that there is a continuous outflow of the wind from the chromosphere into the heliosphere .
however , because of the higher densities in the transition region the expected outflow speed of the wind should be rather low ( @xmath261km / s ) at the level where one can expect c to form .
this is just at the resolution limit of current instrumentation , e.g. sumer / soho can determine doppler shifts down to some 12km / s @xcite .
therefore it remains unclear , whether there is a continuous outflow from the chromosphere , with a small not yet detectable velocity at c heights , or if the velocity in c is really zero and the wind is injected between the levels of c and ne as suggested by @xcite . ] to investigate the scenario proposed by @xcite a numerical experiment was conducted by @xcite .
they start with a magnetic configuration given at the lower boundary from observations and then drive the system by whirl flows imposed on the lower boundary . through this
they bring together open and closed magnetic field structures which results in reconnection .
their model contains a plasma
neutral gas coupling , i.e. it includes effects beyond mhd . for the reconnection they used a switch - on - resistivity , i.e. the resistivity is non - zero only where the current - carrier - velocity ( @xmath27 ) is above a certain threshold . by this their treatment of the reconnection is more detailed than the 3d box models for the closed corona discussed in sect.[s : threed ] , however , they used a simplified energy equation , not accounting for heat conduction or energy losses through optically thin radiation ( cf .
* ) . the magnetic configuration ( at the initial state )
is shown in fig.[f : open ] . as a consequence of driving the magnetic field through the whirls at the bottom
, the reconnection between the open and closed regions injects energy and plasma into the open funnels and through this drives a wind out of the funnels .
roughly , the plasma below @xmath2k falls down , plasma above @xmath2k leaves the computational box through the top .
the downdrafts are mostly concentrated above the magnetic concentrations at the bottom boundary .
thus this simulation supports the general picture suggested by @xcite .
however , further modeling with a proper inclusion of an energy equation and also an investigation of the vuv emission to be expected from this model is needed before a final conclusion can be drawn .
at the moment we have to conclude that even in the very quiet coronal holes , the assumption of a stationary magnetic field configuration simply channeling the solar wind outflow is not justified . to really understand the origin of the solar wind we have to account for the complex interaction between different magnetic field structures and also between the magnetic field and the plasma
a realistic model of the ( upper ) solar atmosphere has to account for the complexity of its magnetic structure . together with the constant rearrangement of the magnetic field at photospheric levels this leads to an ongoing energy release in the low parts of the transition region and corona and through this drives the highly dynamic upper atmosphere .
the current box models for magnetically closed regions as discussed in sect.[s : box ] show a qualitatively good match to observed quantities , such as emission measures , doppler shifts of rms - variability .
this presents ample evidence that the heating through field line braiding is a prime candidate to heat the corona .
of course , other means of energy input into the corona might also reproduce the average quantities of e.g. emission measure , doppler shift or line widths as presented in this paper .
it remains to be seen if other models based on different heating processes will give the same or different results , once they are pushed to produce observables as now done for the field line braiding model . in the future
more efforts have to be undertaken not only to increase the spatial resolution of the simulations , but also to actually include a physical mechanism for the dissipation of magnetic energy during the reconnection events .
the box models for magnetically open region , i.e. for the onset of the fast wind ( sect.[s : open ] ) , have not yet been analyzed in such detail concerning observable properties , but they support some recent interpretation of observational data suggesting that plasma is injected in funnel - type structures to form the fast solar wind . together with the very successful global coronal models , which have only briefly been touched upon in this paper , now complex three - dimensional models become available accounting properly for the plasma and the magnetic field as well as their interaction .
we might hope that the coming years will bring further advances of these models and their diagnostic capabilities .
thus we will improve our understanding of the structure , dynamics and heating of the upper atmospheres of the sun and solar - like stars .
peter , h. , gudiksen , b. , nordlund , . , 2005 .
coronal heating through braiding of magnetic field lines : synthesized coronal euv emission and magnetic structure . in : chromospheric and coronal magnetic fields .
esa sp-596 .
peter , h. , gudiksen , b. , nordlund , . , 2006 .
forward modeling of the corona of the sun and solar - like stars : from a three - dimensional magnetohydrodynamic model to synthetic extreme - ultraviolet spectra .
apj 638 , 10861100 .
trujillo bueno , j. , landi deglinnocenti , e. , casini , r. , martnez pillet , v. , 2005 .
the scientific case for quantum spectropolarimetry from space . in : chromospheric and coronal magnetic fields .
esa sp596 . | the atmosphere of the sun is highly structured and dynamic in nature . from the photosphere and chromosphere into the transition region and the corona plasma-@xmath0 changes from above to below one , i.e.while in the lower atmosphere the energy density of the plasma dominates , in the upper atmosphere the magnetic field plays the governing role one might speak of a `` magnetic transition '' .
therefore the dynamics of the overshooting convection in the photosphere , the granulation , is shuffling the magnetic field around in the photosphere .
this leads not only to a ( re-)structuring of the magnetic field in the upper atmosphere , but induces also the dynamic reaction of the coronal plasma e.g. due to reconnection events .
therefore the ( complex ) structure and the interaction of various magnetic patches is crucial to understand the structure , dynamics and heating of coronal plasma as well as its acceleration into the solar wind .
the present article will emphasize the need for three - dimensional modeling accounting for the complexity of the solar atmosphere to understand these processes .
some advances on 3d modeling of the upper solar atmosphere in magnetically closed as well as open regions will be presented together with diagnostic tools to compare these models to observations .
this highlights the recent success of these models which in many respects closely match the observations .
sun : atmosphere , magnetic field |
with massive new data sets shrinking the statistical error bars on cosmological quantities , it is becoming increasingly important to avoid inaccuracies in their modeling and analysis .
for example , one of the less interesting aspects of modern studies of large scale structure is having to deal with complex angular masks .
this humble but essential task is rendered more time - consuming by the fact that angular masks may require updating as a survey progresses .
the purpose of this paper is to describe a scheme intended to remove much of the drudgery and scope for inadvertent error or unnecessary approximation involved in defining and using angular masks .
the scheme is implemented in a publically available software package mangle , which can be obtained , along with complete documentation , from http://casa.colorado.edu/~ajsh / mangle/. the present paper is not a software manual : for that , visit the aforesaid website .
rather , the purpose of this paper is to set forward the philosophy and to detail the algorithms upon which the software is built .
angular masks of galaxy survey have grown progressively more complicated through the years .
the first complete redshift surveys of galaxies , the first center for astrophysics redshift survey ( cfa1 ; @xcite ) , and the first southern sky redshift survey ( ssrs1 ; @xcite ) , had rather simple angular boundaries , defined by simple cuts in declination and in galactic latitude .
the level of complexity increased with the _ iras _ redshift surveys .
the first of these , the _ iras _ 2 jy redshift survey @xcite , in addition to a cut in galactic latitude , excluded 1465 lunes of high cirrus or contamination by local group galaxies , each lune being an approximately @xmath2 square with boundaries of constant ecliptic latitude and longitude .
the angular masks of subsequent _ iras _ surveys followed a similar theme , leading up to the pscz survey @xcite , whose high - latitude mask ( the one most commonly used in large scale structure studies ) consisted of the whole sky less 11,477 @xmath2 ecliptic lunes .
the automatic plate measuring ( apm ) survey ( @xcite , b , 1996 ) and associated surveys such as the apm - stromlo survey @xcite , consisted of a union of a couple of hundred photographic plates , cut to @xmath3 , each drilled with a smrgsbord of holes to avoid bright stars , satellite trails , plate defects , and the like .
the edges of the excluded holes were straight lines on photographic plates , but unlike the _ iras _ surveys , the holes were not necessarily rectangles , their boundaries were not necessarily lines of constant latitude and longitude , and different holes could overlap .
= 6.5 in the anglo - australian telescope 2 degree field survey ( 2df ; @xcite ; @xcite ) is a redshift survey of galaxies from the apm survey , and thus inherits the holey angular mask of the apm survey .
superimposed on the apm backdrop , the 2df angular mask consists of several hundred overlapping @xmath4 diameter circular fields .
the various overlaps of the circular fields have , at least in early releases of the data , various degrees of completeness . the sloan digital sky survey ( sdss ; @xcite ) has an angular mask comparable in complexity to that of the 2df survey .
it consists of several stripes from the parent photometric survey , peppered with holes masked out for a variety of reasons .
superimposed on the stripes are circular fields from the redshift survey .
recently , the sdss team used the mangle scheme described in the present paper as part of the business of computing the 3d galaxy power spectrum @xcite . with both the 2df and sdss data going public , it has seemed sensible to publish the scheme so that others can use it too .
the scheme described in the present paper began life in the delightful atmosphere of an aspen center for physics workshop in 1985 .
the mathematics of harmonization ( [ harmonize ] ) and other aspects of the computation of angular integrals are written up in an appendix to @xcite .
the methods described therein were first applied by @xcite , and have been used regularly by him since that time .
the idea of adapting the methods to deal with angular masks in a rather general way , and in particular the concept of balkanization , is new to the present paper .
the mangle software has been applied to the 2df 100k survey by @xcite , and to the sdss by @xcite .
the figures in this paper were prepared from files generated by the mangle software .
figure [ twoqzfig ] shows a zoom of a small piece of the northern angular mask of the 2df qso redshift survey ( 2qz ) 10k release @xcite .
the angular mask of this survey is defined by files ( downloadable from http://www.2dfquasar.org/spec_cat/masks.html ) giving the boundaries of : ( 1 ) @xmath3 ukst plates , ( 2 ) holes in ukst plates , and ( 3 ) @xmath4 fields .
these boundaries are illustrated in the top left panel of figure [ twoqzfig ] .
the 2qz team provide the completeness of the angular mask in the @xmath5 pixelized form illustrated in the top right panel of figure [ twoqzfig ] .
the 2qz mask is typical of the way that angular masks are defined in modern galaxy surveys .
[ cols= " < , < " , ] motivated by common practice , an angular mask is defined in the present paper to be an arbitrary union of arbitrarily weighted angular regions bounded by arbitrary numbers of edges .
the restrictions on the mask are 1 . that each edge must be part of some circle on the sphere ( but not necessarily a great circle ) , and 2 . that the weight within each subregion of the mask must be constant
this definition of an angular mask by no means covers all theoretical possibilities , but it does reflect the actual practices of makers of galaxy surveys .
the broad utility of spherical polygons to delineate angular regions is widely appreciated ; for instance , they play an integral part in the sdss database .
the definition implies that an angular mask is a union of arbitrarily weighted non - overlapping polygons .
a polygon is defined to be the intersection of an arbitrary number of caps , where a cap is defined to be a spherical disk , a region on the unit sphere above some line of constant latitude with respect to some arbitrary polar axis . for reference ,
table [ deftable ] collects definitions of mask , polygon , cap , and certain other terms used in this paper .
the bottom left panel of figure [ twoqzfig ] shows the 2qz mask ` balkanized ' ( see [ balkanize ] below ) into non - overlapping polygons .
the bottom right panel of figure [ twoqzfig ] shows the mask reconstructed from spherical harmonics up to @xmath6 ( see [ harmonize ] below ) .
the information specifying a mask ( its angular boundaries and completeness ) is collected in files which we refer to as ` polygon files ' .
typically a command in the mangle suite of software will : 1 . read in one or more polygon files , possibly in different formats ; 2 .
do something to or with the polygons ; 3 .
write an output file , possibly a polygon file , or files .
the strategy adopted in the mangle software is to permit the most flexible possible input format for polygon files , the idea being to be able to read the files provided by the makers of a galaxy survey as far as possible in their original form , or perhaps mildly edited .
mangle reads and writes several different formats of polygon files : * circle ; * vertices ; * edges ; * rectangle ; * polygon . for convenience , there are five additional formats that provide useful information about polygons , but that can only be written , not read , because the information they provide is too limited , or ambiguous , to specify polygons completely .
the five output only formats are : area ; graphics ; i d ; midpoint ; weight .
an abbreviated description of each format appears below ; see http://casa.colorado.edu/~ajsh/mangle/ for full details .
the * circle * format is able to describe polygons in all generality .
a circle is defined by the azimuth @xmath7 and elevation @xmath8 of its north polar axis , and by the angular radius , the polar angle @xmath9 , of the circle .
each circle defines a cap .
a polygon is an intersection of caps , and a line of the form @xmath10 containing @xmath11 angles defines a polygon with @xmath12 caps .
the * vertices * format specifies polygons by a sequence of vertices , assumed to be joined by great circles .
the general form of a line specifying a polygon in vertices format is @xmath13 which defines a polygon with @xmath12 caps . in vertices format ,
a line with @xmath14 angles defines a polygon with n caps .
the * edges * format is a souped - up version of the vertices format .
whereas the vertices format joins each pair of vertices with a great circle , the edges format uses an additional point ( or additional points ) between each pair of vertices to define the shape of the circle joining the vertices .
although the edges format retains more information about a polygon than the vertices format , in general it does not retain all information about a polygon .
a rectangle is a special kind of 4-cap polygon bounded by lines of constant azimuth and elevation .
the * rectangle * format is offered not only because some masks are defined this way ( for example , the _ iras _ masks ) , but also because the symmetry of rectangles permits accelerated computation of their spherical harmonics ( [ harmonize ] ) .
a line in rectangle format looks like @xmath15 with precisely 4 angles .
the * polygon * format is the default output format for polygon files . besides the circle format , it is the only other format that is able to describe polygons in all generality without loss of information .
it stores each cap , not as three angles as in the circle format , but rather as a unit vector along the north pole of the cap , together with a quantity @xmath16 , which is equal both to the area of the cap divided by @xmath17 , and to half the square of the 3-dimensional distance between the north pole and the cap boundary .
it seems doubtful that one would want to create an original mask file in polygon format , since it is a bit peculiar , but it is the format used internally by the mangle software , and it specifies polygons in the manner expected for many years past by the fortran backend .
the advantage of the format is that some computational operations are simpler and faster if the cap axis is stored as a unit vector rather than as an azimuth and elevation . of the purely output formats ,
one of the most useful is the * graphics * format , which is useful for making plots of polygons .
the mangle software does not incorporate any plotting software : it is assumed that you have your own favourite plotting package .
the graphics format is similar to edges format , but is generally more economical .
whereas in edges format there is a specified number of points per edge , in graphics format there is a specified number of points per @xmath17 of azimuthal angle along each edge .
thus in graphics format curvier edges get more points than straighter edges .
the graphics format is implemented only as output , not as input , because of ambiguity in the interpretation of the format .
another useful output format is the * midpoint * format , which returns a list containing the angular position of a point inside each polygon of a mask .
this can be helpful in assigning weights to the polygons of a mask , if you have your own software that returns a weight given an angular position .
[ midpoint ] for more about how midpoints of polygons are computed .
the * area * , * i d * , and * weight * output formats give lists of , respectively , the areas , identity numbers , and weights ( completenesses ) of the polygons of a mask .
one of the basic tasks that the mangle software does is resolve a mask into a set of non - overlapping polygons .
this takes place in a sequence of four steps , elaborated in the subsections following . 1 .
snap ; 2 .
balkanize ; 3 . weight ; 4 .
unify . resolving a mask into non - overlapping polygons
greatly simplifies the logic of dealing with a mask , since it allows subsequent processing ( generation of random catalogues , computation of spherical harmonics , plotting , etc . ) to proceed without recourse to the intricate hierarchy of overlapping geometric entities and the associated complicated series of inclusion and exclusion rules that tend to characterize a survey mask .
each individual polygon is by definition an intersection of caps .
geometrically , this implies that a polygon is convex : the interior angles at the vertices of a polygon are all less than @xmath18 .
the requirement that a polygon be an intersection of caps greatly simplifies the logic , since it means that a point lies inside a polygon if and only if it lies inside each of the caps of the polygon .
the first thing that must be done on all the original polygon files of a mask is to ` snap ' them .
this process identifies almost coincident cap boundaries and snaps them together .
the problem is that the positions of the intersections of two almost but not exactly coincident circles ( cap boundaries ) on the unit sphere may be subject to significant numerical uncertainty . to avoid numerical problems ,
such circles must be made exactly coincident .
you might think that that near - but - not - exactly - coincident circles would hardly ever happen , but in practice they occur often , because a mask designer tries to make two polygons abut , but imprecision or numerical roundoff defeats an exact abutment .
the snap process adjusts the edges of each polygon , but it leaves the number and order of polygons the same as in the input file(s ) .
edges that appear later in the input file(s ) are snapped to earlier edges .
the snap process offers four tunable tolerances : * axis tolerance .
are the axes of two caps within this angular tolerance of each other , either parallel or anti - parallel ?
if so , change the axis of the second cap to equal that of the first cap . * latitude tolerance . if the axes of two caps coincide , are their latitude boundaries within this tolerance of each other ?
if so , change the latitude of the second cap to equal that of the first cap .
the two caps may lie either on the same or on opposite sides of the latitude boundary . * edge tolerance , and edge to length tolerance .
are the two endpoints and midpoint of an edge closer to a cap boundary than the lesser of ( a ) the edge tolerance , and ( b ) the edge to length tolerance times the length of the edge ?
in addition , does at least one of the two endpoints or midpoint of the edge lie inside all other caps of the polygon that owns the cap boundary ?
if so , change the edge to align with the cap boundary .
the purpose of the first two of these tolerances , the axis tolerance and the latitude tolerance , is obvious .
the remaining two tolerances , the edge tolerance and the edge to length tolerance , are necessary because it is possible for two edges , if they are short enough , to almost coincide even though the axes and latitudes of their corresponding caps differ significantly . by default , the three angular tolerances ( axis , latitude , and edge ) are all two arcseconds , which is probably sufficient for typical large scale structure masks . the tolerances can be tightened considerably before numerical problems begin to occur , so it is fine to tighten the tolerance for a mask whose edges are more precisely defined . the default edge to length tolerance @xmath19 , should be fine in virtually all cases .
the snap process accomplishes its work in two stages : * snap axes and latitudes of pairs of caps together , passing repeatedly through all pairs of caps until no more caps are snapped . *
snap edges of polygons to other edges , again passing repeatedly through all pairs of caps until no more caps are snapped . as a finishing touch ,
snap prunes each of the snapped polygons in order to eliminate superfluous caps , those whose removal leaves the area of the polygon unchanged . the process of resolving a mask into disjoint polygons we dub ` balkanization ' , since it fragments an input set of possibly overlapping polygons into many non - overlapping connected polygons .
the process involves two successive stages : 1 .
fragment the polygons into non - overlapping polygons , some of which may be disconnected .
identify disconnected polygons and subdivide them into connected parts .
= 1.8 in the algorithm for the first stage of balkanization is simple and pretty : * is the intersection of two polygons neither empty nor equal to the first polygon ?
if so , find a circle , a cap boundary , of the second polygon that divides the first polygon , and split the first polygon into two along that circle .
* iterate .
notice that only one of the two parts of the split polygon overlaps the second polygon , and that only the overlapping part needs iterating . for any pair of polygons ,
iteration ceases when the overlapping part lies entirely inside the second polygon .
the final overlapping part is equal to the intersection of the original first polygon with the second polygon .
all other fragments of the first polygon lie outside the second polygon .
figure [ balkanizefig ] illustrates an example of the first stage of balkanization for two overlapping polygons a and b. first , a is split against b , which takes two iterations of the above cycle .
then , b is split against a. again , this takes two iterations of the above cycle .
the final system consists of 5 non - overlapping polygons .
note that splitting the system shown in panel ( a ) of figure [ balkanizefig ] into its three connected parts ( the part of a that does not intersect b , the part of b that does not intersect a , and the intersection ab of a and b ) would not constitute a successful balkanization , since two of these regions are not convex and hence not polygons .
one might ask , why not stop at panel ( c ) in figure [ balkanizefig ] ?
do not the three polygons there already form a satisfactory set of non - overlapping polygons ?
the answer is that the intersection polygon ab may well have a weight different from those of the non - overlapping parts of the parent a and b polygons ( this is typically true for example in the 2df and sdss surveys ) . to deal with this eventuality
, balkanization must continue to completion , as illustrated in panel ( e ) .
the question of whether two polygons overlap is determined by computing the area of the intersection of the polygons .
the area is proportional to the monopole harmonic , computed as described in [ harmonize ] .
the intersection of two polygons is itself a polygon , consisting of the intersection of the two sets of caps defining the polygons .
stage 1 of the balkanization procedure yields polygons that can contain two or more connected parts , as illustrated in figure [ discpolfig ] .
stage 2 attempts to subdivide such disconnected polygons into connected parts by computing the connected boundaries of the polygon , and lassoing ( see [ lasso ] ) each connected boundary with an extra circle .
figure [ eyefig ] illustrates a polygon that has two distinct connected boundaries by virtue of being not simply - connected rather than not connected .
a region is said to be * simply - connected * if , according to the usual mathematical definition , it is connected and any closed curve within it can be continuously shrunk to a single point . loosely speaking ,
this means that a simply - connected polygon has no holes .
because it is connected , the polygon of figure [ eyefig ] need not be split .
the strategy to deal with non - simply - connected polygons is based on the following theorem , proven in the appendix : * a connected part of a polygon is simply - connected if and only if all the boundaries of the connected part belong to a single group*. a group is defined here as follows : two circles are friends , belonging to the same group , if they intersect ( anywhere , not necessarily inside the polygon ) , and friends of friends are friends . according to this definition of group , the circles on a single connected boundary necessarily all belong to the same group . however , the circles on two distinct connected boundaries may or may not belong to the same group .
this theorem implies that it is necessary to lasso only those boundaries of a polygon that belong to the same group . in figure [ discpolfig ] , for example , the two boundaries of the polygon belong to the same group of three intersecting circles , so these two boundaries must be lassoed , partitioning the polygon into two parts . in figure [ eyefig ] , on the other hand , the two boundaries belong to two separate groups , and need not be lassoed . figure [ discpolsfig ] illustrates a more complicated polygon , similar to the polygon of figure [ discpolfig ] but pierced with two circular holes .
the polygon contains four boundaries belonging to three groups .
the two original boundaries inherited from figure [ discpolfig ] belong to the same intersecting group of circles , but the additional two holes form two separate groups . here
only the two boundaries belonging to the same group need lassoing .
= 2.5 in figure [ fractalfig ] illustrates a yet more complicated multiply - connected polygon .
the mangle software balkanizes this polygon correctly into seven polygons , a stringent test of the algorithms .
a corollary of the theorem proven in the appendix is that the polygon formed by the intersection of the caps bounded by the circles of a single group must be a union of simply - connected parts . for example
, the two parts of the polygon in figure [ discpolfig ] must be simply - connected which evidently they are because the circles of the polygon all belong to the same group .
in the course of the proof in the appendix , it is shown that if the boundary of a polygon falls into two ( or more ) groups , then the circles of a second group must lie entirely inside exactly one of the simply - connected parts of the polygon bounded by the first group .
for example , each of the two circles bounding the two holes in the polygon of figure [ discpolsfig ] must lie entirely inside exactly one of the two simply - connected parts of the original polygon from figure [ discpolfig ] , which again is evidently true .
it follows from the statement of the previous paragraph that in lassoing the connected boundaries of a group , it is necessary to consider only the boundaries belong to the same group : any boundary belonging to another group can be ignored , because it must lie entirely inside one of the simply - connected parts bounded by the first group .
thus each lassoing circle is required to enclose fully its connected boundary , while excluding fully all other connected boundaries belonging to the same group ; there is no constraint on the lasso from boundaries belonging to other groups .
if a group of circles of a polygon defines a single boundary , then that boundary needs no lassoing , but stage 2 balkanization nevertheless attempts to lasso the boundary if the number of caps of the group exceeds the number of vertices .
for example , in its original configuration the polygon shown in figure [ eggboxfig ] has a large number of caps .
none of the caps can be discarded , since each excludes a small piece of the sky .
here it is advantageous to lasso the polygon with an extra circle , allowing most of the original caps to be discarded as superfluous .
a lasso that lassos the lone boundary of a group is discarded if the lasso completely encloses all the circles of the group to which the boundary belongs . for
then either the lasso completely encloses the polygon , in which case it is superfluous , or else the lasso lies completely inside the simply - connected region bounded by the lone boundary , in which case the lasso , if kept , would divide the simply - connected region in two , which would be incorrect . if on the other hand the lasso of a lone boundary of a group intersects at least one of the circles of the group , then the lasso must completely enclose the simply - connected region bounded by the lone boundary , as in figure [ eggboxfig ] ; the lasso can not lie inside the simply - connected region because it is being assumed that the lasso intersects a circle of the group , whereas no such circle can exist within the simply - connected region .
stage 2 balkanization may need more than one pass to succeed .
figure [ chainfig ] shows an example of a polygon , bounded by one large cap punctuated by fifteen small caps , that contains four parts bounded by four boundaries all belonging to the same single group .
the top and bottom boundaries can be lassoed successfully with single circles , but the middle two boundaries can not : any circle that encloses either of the middle boundaries necessarily intersects another boundary somewhere . here
stage 2 balkanization succeeds by submitting the polygon to two passes . in the first pass ,
the polygon is split into three polygons , consisting of the top and bottom connected parts , plus a third polygon containing the two middle parts . in the second
pass , the third polygon is split into two , completing the partitioning of the original polygon into its four parts . in certain convoluted cases , such as the polygon shown in figure [ vfig ]
, it can be impossible to lasso any of the connected boundaries of the polygon with a circle that wholly encloses a connected boundary while wholly excluding all other connected boundaries in the same group .
stage 2 balkanization gives up attempting to lasso a boundary after a certain maximum number of attempts , but it keeps a record of the best - attempt lasso , the one that encloses as much as possible of a boundary while wholly excluding all other boundaries in the same group .
stage 2 balkanization proceeds to split the polygon into two parts with the best - attempt lasso , and then submits the two parts to a further pass .
the polygon of figure [ vfig ] , for example , contains two parts bounded by two connected boundaries neither of which can be lassoed with a circle that completely encloses the boundary while completely excluding the other boundary .
finding no satisfactory lasso , stage 2 balkanization splits the polygon into two polygons with a best - attempt lasso , shown as a dashed line in figure [ vfig ] .
the two polygons are then submitted to further passes of stage 2 balkanization , which in this case succeeds with one pass .
the upshot is that the original polygon is balkanized into four polygons .
= 2.5 in it is conceivable that the algorithm of the above paragraph could continue for ever , continually splitting a polygon into two and continually failing to lasso successfully all the boundaries of the split polygons .
however , polygons that defy lassoing have to be filamentary in character ( long , thin , and windy ) , such as that shown in figure [ wfig ] , and splitting such a polygon in two generally makes it less filamentary , like putting spaghetti in a blender . in the case of the polygon of figure [ wfig ] ,
stage 2 balkanization forcibly splits the polygon 8 times , eventually balkanizing the 13-part polygon into 34 disjoint parts .
suffice to say that we know of no polygon that fails the algorithm , and it is possible that no such polygon exists . if the reader finds one , please tell us about it . in practice , the mangle software bails out if a polygon has to be split forcibly in two more than a certain maximum number ( 100 ) of times . even in this last gasp case , the set of polygons output by balkanization still constitutes a valid set of non - overlapping polygons that completely tile the mask .
the only problem is that the ` failed ' polygons , those which could not be partitioned completely , may contain two or more disjoint parts with different weights .
to finish , balkanization prunes each of the balkanized polygons in order to eliminate superfluous caps , those whose removal leaves the area of the polygon unchanged .
for example , pruning discards the many superfluous caps of the polygon of figure [ eggboxfig ] .
the caps are tested in order , with any new lassoing cap being tested last , so that the many superfluous caps are discarded , and the lassoing cap is kept .
the algorithm to lasso a connected boundary of a polygon is to pick a point , initially taken to be the barycentre of the centres of the edges of the connected boundary ( or , if the connected boundary consists of a single circle , the centre of that circle ) , and find the circle centred on that point which most tightly encloses the boundary .
the lassoing circle is enlarged slightly if possible , as a precaution against numerical problems that might potentially occur if the lasso just touched an edge or vertex of the boundary .
a lasso that lassos , i.e. that encloses completely , one connected boundary of a polygon , is required to exclude completely all other connected boundaries belonging to the same group .
a lasso attempt can sometimes fail if a polygon has two or more connected boundaries .
a lasso attempt fails if the angular distance from the centre of the lasso to the farthest point @xmath20 on the to - be - lassoed connected boundary is greater than the angular distance from the centre of the lasso to the nearest point @xmath21 on all other connected boundaries belonging to the same group .
if a lasso attempt fails , then the centre point of the lasso is shifted over the unit sphere along the vector direction from @xmath21 to @xmath20 , by an amount that puts the centre point just slightly closer to @xmath20 than @xmath21 .
the lasso is then reattempted .
the process of shifting the centre point and retrying a lasso is repeated until either the lasso succeeds , or until a certain maximum number of attempts has been made .
multiple intersections occur where 3 or more circles ( cap boundaries ) intersect at a single point .
multiple intersections pose a potential source of numerical problems , because the topology around multiple intersections may vary depending on numerics , as illustrated in figure [ multfig ] .
circles kiss if they just touch . again , kissing circles pose a potential source of numerical problems , because whether two circles kiss may vary depending on numerics , as illustrated in figure [ kissfig ] .
mangle is equipped to deal with both multiply - intersecting and kissing circles , and should cope in almost all cases , although it is possible to fool manglewith a sufficiently complicated polygon , for example a polygon whose vertices have a fractal distribution of separations .
the strategy is as follows .
circles are considered to be multiply intersecting , crossing at a single vertex , if the intersections are closer than a certain tolerance angle .
similarly , circles are considered to kiss , touching at a single vertex , if their kissing distance is closer than the tolerance angle .
the algorithm is friends - of - friends : two vertices closer than the tolerance are friends , and a friend of a friend is a friend . the position of each vertex @xmath22 of a polygon , where edge @xmath23 intersects ( or kisses ) edge @xmath24 , is computed two different ways , first as the intersection of edge @xmath23 with edge @xmath24 , then as the intersection of edge @xmath24 with edge @xmath23 . for each of the two ways of computing it ,
the intersection is tested against all other circles , to determine whether the intersection is or is not a vertex of the polygon , that is , whether the intersection lies on the edge of the polygon , or outside the polygon . for consistency
, the test should give the same result in both computations : the intersection should be a vertex in both cases , or it should not be a vertex in both cases . if an inconsistency is detected , then the tolerance angle is doubled ( or set to a tiny number , if the tolerance is zero ) , and the computation is repeated for the inconsistent polygon , until consistency is achieved . by default ,
the initial tolerance angle for multiple intersections and kissings is @xmath25 arcseconds . in figure
[ multfig ] , the intersection of the two diagonals is a vertex of the polygon in the left panel , but is not a vertex in the right panel , because it lies outside the polygon . in the case of an exact multiple intersection , as in the middle panel of figure [ multfig ] , the intersection @xmath22 is considered to be a vertex of the polygon only if @xmath23 and @xmath24 are both edges of the polygon .
thus the intersection of the two diagonals in the middle panel of figure [ multfig ] is a vertex , because both diagonals are edges , but the intersection of the horizontal with either diagonal is not a vertex , because the horizontal is not an edge of the polygon . if @xmath23 is an edge , and it intersects multiply with a bunch of other circles , then the adjacent edge @xmath24 is formed by the circle which ` bends most tightly ' around the polygon , that is , the circle whose interior angle at the vertex is the smallest , or , if two circles subtend the same interior angle ( within the tolerance angle ) , then the circle whose polar angle @xmath9 is the smallest . in figure [ kissfig ] , the two circles @xmath23 and @xmath24 intersect at no vertices in the left panel , and at two vertices in the right panel . in the case of an exact kiss , as in the middle panel of figure [ kissfig ] , the kissing point @xmath22 is considered to be a vertex only if @xmath23 and @xmath24 are both edges of the polygon .
thus the kissing point @xmath22 is _ not _ a vertex of the the upper and lower polygons , the two disks , but it is a vertex of the middle polygon , the pointy one .
two polygons that just touch or kiss at a single isolated point ( or at a set of isolated points ) are considered to be disconnected from each other .
thus for example the top and bottom polygons in the middle panel of figure [ kissfig ] are considered to be disconnected from each other ; and similarly the left and right polygons in the middle panel of figure [ kissfig ] are considered to be disconnected from each other . in practice , consistency of the topology of the distribution of vertices around a polygon is checked by means of a 64-bit check number . if the intersection @xmath22 of edge @xmath23 with edge @xmath24 ( two edges can intersect at two separate points , so @xmath22 is an ordered pair , going from edge @xmath23 to edge @xmath24 right - handedly around the boundary of the polygon ) is determined to be a vertex of the polygon , then a 64-bit pseudo - random integer is added to the check number , and if the same intersection @xmath22 of edge @xmath24 with edge @xmath23 is determined to be a vertex of the polygon , then the same 64-bit pseudo - random integer is subtracted from the check number . for consistency , the check number should be zero for the entire polygon .
it is conceivable , with probability 1 in @xmath26 , or less than 1 in 10 billion billion , that the check number could evaluate to zero accidentally , but this seems small enough not to worry about , especially since inconsistency should be a rare occurrence in the first place .
figure [ flowerfig ] illustrates a mask designed to be as ` difficult ' as possible : it contains many multiply intersecting and nearly multiply - intersecting circles , and many kissing and nearly kissing circles , including several simultaneously multiply - intersecting and multiply - kissing circles .
the mangle software copes with this , a non - trivial accomplishment .
the sum of the areas of the 332 polygons of the balkanized mask of figure [ flowerfig ] differ from the area , @xmath27 , of the overall bounding rectangle by @xmath28 , which is definitely satisfactory .
given the algorithms , one could expect the numerical uncertainty in the area of a single polygon to be no better than machine precision times @xmath29 , which on the machine used for this computation was about @xmath30 .
when balkanizing , does the order of the polygons in the input polygon files matter ?
the answer is yes , if the input polygons overlap , and if the overlapping polygons carry different weights .
as described in the following subsection [ weight ] , if two polygons overlap , then the weight of the polygon that appears later in the input file(s ) overrides the weight of the earlier polygon .
if all polygons have the same weight ( say 1 ) , then the order of the input polygon files does not really matter .
however , it may lead to a slightly smaller eventual polygon file ( after unifying , see [ unify ] ) if large , coarse polygons are put first , and small , finely detailed polygons are put last .
each connected polygon of a mask may have a different weight . in galaxy surveys ,
the ` weight ' attached to a polygon is the completeness of the survey in that polygon .
these weights must be supplied by the user .
if no weights are supplied , then the weight defaults to 1 .
if the input polygons of a mask overlap , then the policy adopted by the mangle software is to allow the weights of later polygons in polygon files to override the weights of earlier polygons .
thus for example , to drill empty holes in a region , one would put the polygons of the parent region first ( with weight 1 , perhaps ) , and follow them with polygons specifying the holes ( with weight 0 ) .
there are three ways to apply weights to polygons .
the first way is simply to edit the polygon file or files specifying the mask . attached to each polygon in a polygon file
is a line that includes a number for the weight ; one simply edits that number .
the second way is to specify weights in a file .
the mangle software contains a facility to read in these weights , and to apply them successively to the polygons of a polygon file .
suppose that you have your own software that returns a weight given an angular position in a mask .
the mangle software includes a utility ( [ midpoint ] ) to create a file giving the angular position of a point inside each polygon of a mask .
this file of angular positions becomes input to your own software , which should create a file of weights , which in turn can be fed back to mangle .
the third way to apply weights to polygons is to write a subroutine ( in either fortran or c ) that returns a weight given an angular position , and compile it into mangle .
the mangle software includes some template examples of how to do this . in our experience ,
method two is the method of choice , except in cases that are simple enough that method one suffices .
the set of non - overlapping polygons that emerges from balkanizing and weighting may be more complicated than necessary .
the mangle software includes a facility for simplifying the polygons of a mask , which we call unification .
unification is not strictly necessary , but it tidies things up , and it can save subsequent operations , such as harmonization ( see [ harmonize ] ) , a lot of computer time .
= 1.8 in unification eliminates polygons with zero weight , and does its best to merge polygons with the same weight .
the algorithm is to pass repeatedly through a set of polygons , merging a pair of polygons wherever the pair can be merged into a single polygon by eliminating a single abutting edge .
figure [ unifyfig ] illustrates an example of the unification procedure .
unification does not necessarily accomplish the most efficient unification , nor , as illustrated in figure [ nounifyfig ] , is unification necessarily exhaustive .
the mangle software contains several utilities that do various things with a mask .
one of the most important of these is a utility to take the spherical harmonic transform of a mask , a process we call harmonization . in particular , the area of a mask is proportional to the zeroth harmonic .
computation of the area of a polygon is basic to several of the mangle algorithms .
for example , whether two polygons intersect is determined by whether the area of their intersection is non - zero .
the method for computing the spherical harmonics of a mask consisting of a union of polygons is described in the appendix of @xcite .
the algorithm is recursive and stable , able to compute harmonics to machine precision to arbitrarily high order , limited only by computer power and patience .
the recursion , as implemented in the mangle software , recovers correctly from underflow , which can occur at large harmonic number @xmath31 .
while the recursive algorithm by itself is fast , there is a numerical penalty to be paid for allowing the polygons of a mask to have arbitrary shape : the computation time for harmonics up to @xmath31 increases as @xmath32 , a pretty steep penalty when @xmath31 is large .
the computation time is proportional to the number of edges of the polygons of the mask , and on a 750mhz pentium iii it takes 3 cpu minutes per edge to compute harmonics up to @xmath6 .
thus the method is slow compared to fast algorithms specially designed for regular pixelizations , such as healpix @xcite .
the most time - consuming part of the computation is rotating the harmonics of an edge from its natural frame of reference into the final frame of reference : it is this rotation that takes @xmath32 time . the rotation is unnecessary if the edge is a line of constant latitude in the final reference frame , and the computation goes faster , as @xmath33 , in this case .
another acceleration is possible if two edges are related by a rotation about the polar axis of the final frame .
although computing the harmonics @xmath34 of a single edge @xmath35 still takes @xmath32 time , the harmonics @xmath36 of a second edge @xmath37 rotated right - handedly by azimuthal angle @xmath38 from edge @xmath35 are @xmath39 , which is fast to compute . in practice , the mangle software currently implements the latter acceleration only in the special case where ( some of ) the polygons of a mask are rectangles , polygons bounded by lines of constant azimuth and elevation .
the acceleration applies only if at least two rectangles of the mask have the same minimum and maximum elevation .
two such rectangles need not be adjacent in the polygon file : mangle reorders the computation of polygons so as to take advantage of acceleration where possible . for completeness , we give here an overview of the method detailed by @xcite .
the spherical harmonic coefficients @xmath40 of a mask @xmath41 , a function of angular direction @xmath42 , are defined by @xmath43 @xmath44 where @xmath45 are the usual orthonormal spherical harmonics , and @xmath46 denotes an interval of solid angle about @xmath47 .
the key mathematical trick is to convert the integral ( [ olm ] ) for @xmath40 from an integral over the solid angle of the mask to an integral over its edges .
this is done by introducing the square @xmath48 of the angular momentum operator @xmath49 into the integrand @xmath50 which is valid except for the monopole harmonic @xmath51 , dealt with below .
the hermitian character of the angular momentum operator @xmath52 allows equation ( [ olml ] ) to be rewritten @xmath53 by assumption , the mask is a sum over polygons @xmath54 , and @xmath41 is a constant @xmath55 within each polygon .
it follows that @xmath56 in equation ( [ olmp ] ) is a sum over polygons , with the contribution from each polygon being a vector whose magnitude is @xmath55 times @xmath57 times a dirac delta - function , and whose direction is along the boundary of the polygon , winding right - handedly about the polygon .
thus the integral ( [ olmp ] ) reduces to a sum of integrals over the boundaries @xmath58 of the polygons @xmath59 the boundary @xmath58 of a polygon is a set of edges , so the integral in equation ( [ olmo ] ) becomes a sum of integrals over each edge of each polygon .
thus a harmonic @xmath40 is a sum of contributions @xmath60 from each edge @xmath35 of each polygon @xmath54 @xmath61 hence the analytic problem reduces to that of determining the harmonics @xmath34 of the edge @xmath35 of a polygon .
the problem is well suited to computation , and could easily be parallelized if required ( currently , mangle is not parallelized ) .
stable recursive formulae for computing the harmonics @xmath34 of a polygon edge are given by @xcite .
first , the harmonics of an edge are computed in a special frame of reference where the axis of the edge cap is along the polar direction ( the @xmath62 direction ) of the spherical harmonics .
the harmonics of the edge are then rotated into the actual frame of reference .
the most time consuming part of the computation is the second part , the rotation , with the computational time going as @xmath32 .
the above derivation of the harmonics @xmath40 of the mask fails for the monopole harmonic @xmath51 , for which equation ( [ olml ] ) is invalid .
the monopole harmonic @xmath63 is @xmath64 where @xmath65 is the weighted area of the mask , a weighted sum of the areas @xmath66 of the polygons @xmath54 @xmath67 the general formula for the area @xmath66 of a polygon @xmath54 is @xmath68 where @xmath69 , an integer , is the euler characteristic (= faces minus edges plus vertices of any triangulation ) of the polygon , @xmath70 and @xmath71 are the lengths and polar angles of the edges @xmath35 of the polygon , and @xmath72 are the exterior angles ( @xmath73 @xmath18 minus interior angle ) at the vertices @xmath74 of the polygon .
the euler characteristic @xmath69 of a polygon , a topological quantity , is calculable topologically it equals two plus the number of connected boundaries minus twice the number of groups to which the boundaries belong but in practice it is quicker to compute the euler characteristic as follows .
first , in the trivial case that the polygon is the whole sky , its area is @xmath75 .
second , in the special case that the polygon is a single cap with edge @xmath35 , its area is @xmath76 .
otherwise , the polygon is an intersection of two or more caps .
if the area of any one of the caps is less than or equal to @xmath17 , then the area of the polygon must be less than @xmath17 , so the euler characteristic @xmath69 must take that integral value which makes the area lie in the interval @xmath77 .
this leaves the case where every one of the caps of the polygon has area greater than @xmath17 .
the policy in this case is to introduce an extra cap which splits the polygon into two parts , each of whose areas is less than @xmath17 ; the area of the polygon is then the sum of the areas of the two parts . in practice ,
polygons which are intersections of caps all of whose areas exceed @xmath17 are fairly uncommon , so the slow down involved in splitting these particular polygons is not great .
moreover the splitting is necessary only for the monopole harmonic : the higher order harmonics can be computed from the original polygon without splitting it .
the mangle software contains a ` map ' utility to reconstruct the mask at arbitrary points @xmath42 from spherical harmonics up to a given maximum @xmath78 @xmath79 for example , the bottom right panel of figure [ twoqzfig ] was generated using this utility .
another important feature provided in the suite of mangle software is a utility to compute precisely the angular cross - correlation @xmath0 at given angular separation @xmath9 between given ` data ' points and ` random ' points in the mask .
the angular integral is done analytically and evaluated to machine precision , rather than by monte carlo integration with random points . if the ` data ' points are chosen randomly within the mask , as in [ random ]
, then the cross - correlation becomes equivalent to the angular autocorrelation @xmath1 , for random - random , at given angular separation between pairs of points in the mask .
the advantage of computing the @xmath0 angular integral analytically over the traditional monte carlo method is that it eliminates unnecessary shot noise . as discussed for example by @xcite
, this unnecessary shot noise can adversely affect the performance of estimators of the correlation function at small scales . in the case of the pscz high latitude mask , which balkanizes into 744 polygons , it takes 5 cpu minutes per 1000 ` data ' points to compute @xmath0 at 1000 angular separations with a 1.2ghz pentium iii .
the contribution of a data point at position @xmath42 to the angular correlation @xmath0 at angular separation @xmath9 is a weighted sum over the polygons of the mask of the azimuthal angle @xmath80 subtended within the polygon by a circle of radius @xmath9 centred at @xmath42 .
the correlation @xmath0 at angular separation @xmath9 is an average over the contributions from each data point .
the computation of @xmath0 is done at a specified angular separation or set of separations @xmath9 , not , as is common with monte carlo integration , over a bin of angular separations . for a finite number of data points , and a mask consisting , as in the present paper , of a union of weighted polygons , the correlation @xmath0 is a continuous function of angular separation @xmath9 , except that , if a data point @xmath42 happens to coincide with the axis of an edge of a polygon , then the function will be discontinous at a separation @xmath9 equal to the polar angle of the said edge .
furthermore , the correlation @xmath0 is a differentiable function of separation @xmath9 except at a finite , possibly large , number of discontinuities where @xmath9 equals the separation between a data point @xmath42 and either a vertex of a polygon or a point on the edge of a polygon where the separation from @xmath42 is extremal . although @xmath0 , as a function of separation @xmath9 , is thus typically discontinuous in the derivative , and occasionally discontinuous in itself , in practical galaxy surveys @xmath0 tends to be relatively smooth , especially when the number of data points is large .
thus in the practical case it is usually fine to sample @xmath0 at a suitably large number of angular separations , and to interpolate ( linearly ) on such a table .
the dr utility loops in turn through each data point @xmath42 , so that two points take twice as much cpu time as one point . for each data point , the dr utility attempts to accelerate the computation with respect to the angular separations @xmath9 by first computing the minimum and maximum angles between the point @xmath42 and each polygon in the mask .
the information about the minimum and maximum angles is used to decide whether the circle about @xmath42 lies entirely outside or entirely inside a polygon , in which case the ( unweighted ) angle @xmath80 subtended within the polygon is zero or @xmath17 . in practical cases
the angle subtended is often zero for the great majority of polygons of a mask , especially when the mask is composed of many polygons .
since calculation of the subtended angle can be skipped if the angle is zero , computation can be greatly speeded up .
further acceleration comes from ordering the polygons in increasing order of the minimum angle from the given point @xmath42 to each polygon .
this allows the computation to loop to the next value of @xmath9 as soon as it hits a polygon for which the subtended angle is zero , rather than checking through large numbers of polygons that all have zero subtended angle .
the angular auto - correlation @xmath1 between pairs of points in a mask can be computed in various ways . the traditional method ,
as suggested by the random - random designation , is to count pairs of random points .
however , the traditional method is certainly not the most precise method , and it may not be the most efficient , especially at small scales , if large numbers of random points are needed to reduce the shot noise to a subdominant level .
an alternative method , mentioned above , is to compute @xmath1 using the @xmath0 algorithm with the ` data ' points chosen randomly within the mask .
although this method is subject to some shot noise , the shot noise is liable to be substantially less than that of the traditional method at small scales . at the largest angular separations
@xmath9 , a choice method is to compute @xmath1 from its spherical harmonic expansion , truncated at some suitably large harmonic @xmath78 : @xmath81 where @xmath82 is a legendre polynomial , which is accurate at angular scales @xmath83 .
the exact expression for the autocorrelation @xmath1 at angular separation @xmath9 is @xmath84 for a mask of polygons as considered in this paper , this double integral over solid angles can be transformed into a double integral over the edges of the polygons @xcite @xmath85 where @xmath65 is the weighted area , equation ( [ a ] ) , of the mask , and @xmath86 is the green s function of the scalar product @xmath87 of angular momentum operators : @xmath88 } & ( x_{12 } \geq x ) \\ 0 & ( x_{12 } \leq x ) \ .
\end{array } \right.\ ] ] unfortunately the integral ( [ rrexact ] ) can not be solved analytically , and we have not attempted to implement its numerical solution in the mangle software .
@xcite gives a series expansion of the expression ( [ rrexact ] ) valid at small scales , but the series expansion is liable to break down already at tiny scales in the typically complex masks of modern surveys ( the coefficients of the series expansion change discontinuously wherever the angular separation equals the distance between distinct vertices , or more generally any extremal distance between pairs of edges ) , making the series expansion of limited applicability .
the mangle software contains a utility for generating random points inside a mask .
as indicated above , this can be useful for example in computing the angular auto - correlation @xmath1 of pairs of points in the mask .
the algorithm , which is quite fast , is as follows : 1 . select randomly a polygon in the mask , with probability proportional to the product of the polygon s weight and area .
lasso that polygon with a circle that is intended to be a tight fit , but is not necessarily minimal .
generate a point randomly within the circle , test whether the point lies inside the polygon , and keep the point if it does .
iterate .
a lasso is computed for a polygon as needed , but is then recorded , so that a lasso is computed only once for any polygon .
if the desired number of random points exceeds the number of polygons in the mask , then the computation starts by lassoing every polygon in the mask .
the mangle software contains several other utilities , described below .
one simple but oft - used utility is one that copies a polygon file or files into another polygon file in a different format
see [ format ] for an abbreviated description of the possible formats .
for example , most of the figures in this paper were produced from points generated by copying polygon files into graphics format .
the utility for copying polygon files has some switches to copy polygons with weights only in some interval , or with areas only in some interval .
this makes it easy for example to discard polygons with small weights or small areas . the mangle software contains routines that return the vertices of the polygons of a mask , and to return the positions of points along the edges of a mask .
for example , when a polygon file is copied into graphics format , the copy utility invokes these routines .
a single polygon can have more than one connected boundary , as illustrated in figures [ discpolfig][fractalfig ] and [ chainfig][wfig ] . here
the routines return the vertices and edge points on distinct boundaries as distinct sets .
the routines to determine the distinct connected boundaries of a polygon are used in stage 2 of the balkanization process , [ balkanize ] .
the mangle software contains a routine to find a point inside each polygon of a mask .
the aim is to find a point that is squarely inside the polygon , well away from its edges .
for example , the weights attached to the balkanized set of polygons shown in the lower left panel of figure [ twoqzfig ] were obtained by picking a point inside each polygon , and evaluating the weight at that point from the @xmath5 pixelized map provided by the 2qz team .
the pixelization means that the map is reliable only away from the edges of a polygon , so it is important to pick the point squarely inside the polygon .
the algorithm finds one point for each distinct connected boundary of a polygon .
if the polygon contains non - simply - connected parts , as in figures [ eyefig][fractalfig ] , that means that the algorithm will return more points than there are distinct connected parts of the polygon ; however , having more points than necessary is not a problem .
the algorithm to find a point inside a polygon is mildly paranoid . for each connected boundary of a polygon , the algorithm first determines the barycentre of the midpoints of the edges of the connected boundary .
this barycentre can not be guaranteed to lie inside the polygon , so it is not enough to stop here . instead
, great circles are drawn from the barycentre to each of the midpoints of the edges of the connected boundary . on each of these great circles ,
the midpoint of the segment of the great circle lying within the polygon , with one end of the segment being the midpoint of an edge of the connected boundary , is determined .
each of these segment midpoints inside the polygon is tested to see how far away it is from the nearest vertex or edge of the polygon , including edges and vertices other than on the connected boundary .
the desired point inside the polygon is chosen to be that segment midpoint which is furthest away from any vertex or edge .
the mangle software contains a utility to find inside which polygon or polygons of a polygon file a given point lies .
a point may lie inside zero polygons , or one polygon , or more than one polygon .
the utility takes no short cuts : it tests all points against all polygons .
this can take time if there are large numbers of points and large numbers of polygons .
if the polygon file has been produced by balkanization , then a point would normally lie inside at most one polygon .
however , a point at the edge of two abutting polygons is considered to lie inside both polygons .
the observing strategies of modern galaxy surveys typically produce angular masks with complex boundaries and variable completeness .
the purpose of this paper has been to set forward a scheme that is able to deal accurately and efficiently with such angular masks , and thereby to reduce both the labour and the chance for inadvertent error .
the fundamental idea is to resolve a mask into a union of non - overlapping polygons each of whose edges is part of a circle ( not necessarily a great circle ) on the sphere .
the scheme has been implemented in a suite of software , mangle , downloadable from http://casa.colorado.edu/~ajsh / mangle/. the mangle software includes several utilities for accomplishing common tasks associated with angular masks of galaxy surveys .
this includes generating random catalogues reflecting the angular selection function ( a tool employed in almost all galaxy survey analysis ) , measuring the @xmath0 and @xmath1 angular integrals ( needed for estimating the correlation function ) , and expanding the mask in spherical harmonics ( a key step in various techniques for measuring the power spectrum and redshift space distortions ) .
the scheme was originally motivated by the nature of real angular masks of real galaxy surveys , and the underlying angular routines have been battle - tested over many years .
the full apparatus of the mangle software has been used on the 2df survey by @xcite , and on the sdss survey by @xcite .
ajsh was supported by nasa atp award nag5 - 10763 and by nsf grant ast-0205981 .
mt was supported by nsf grants ast-0071213 & ast-0134999 , nasa grants nag5 - 9194 & nag5 - 11099 , and fellowships from the david and lucile packard foundation and the cottrell foundation .
we thank yongzhong xu for his contributions to an early version of the balkanization software , and michael blanton for suggestions .
colless m. et al .
( the 2dfgrs team ; 29 authors ) , 2001 , mnras , 328 , 1039 ( astro - ph/0106498 ) croom s. m. , smith r. j. , boyle b. j. , shanks t. , loaring n. s. , miller l. , lewis i. j. , 2000 , mnras , 322 , l29 ( 2qz , available at http://www.2dfquasar.org/spec_cat/ ) da costa l. n. , pellegrini p. s. , davis m. , meiksin a. , sargent w. l. w. , tonry j. l. , 1991 , apjs , 75 , 935 gorski k. m. , wandelt b. d. , hansen f. k. , hivon e. , banday a. j. , 1999 , astro - ph/9905275 hamilton a. j. s. , 1993a , apj , 406 , l47 hamilton a. j. s. , 1993b , apj , 417 , 19 huchra , j. , davis , m. , latham , d. and tonry , j. ( 1983 ) apjs , 52 , 89 kerscher m. , szapudi i. , szalay a. , 2000 , apj , 535 , l13 ( astro - ph/9912088 ) lewis i. j. et al .
( 2df team ; 23 authors ) , 2002 , mnras , 333 , 279 ( astro - ph/0202175 ) loveday j. , peterson b. a. , maddox s. j. , efstathiou g. , 1996 , apjs , 107 , 201 maddox s. j. , efstathiou g. , sutherland w. j. , 1990b , mnras , 246 , 433 maddox s. j. , efstathiou g. , sutherland w. j. , 1996 , mnras , 283 , 1227 maddox s. j. , sutherland w. , efstathiou g. , loveday j. , 1990a , mnras , 243 , 692 saunders w. , sutherland w. j. , maddox s. j. , keeble o. , oliver s. j. , rowan - robinson m. , mcmahon r. g. , efstathiou g. p. , tadros h. , white s. d. m. , frenk c. s. , carraminana a. , hawkins m. r. s. , 2000 , mnras , 317 , 55 ( pscz , available at http://www-astro.physics.ox.ac.uk/~wjs/pscz.html ) strauss m. a. , huchra j. p. , davis m. , yahil a. , fisher k. b. , tonry , j. , 1992 , apjs , 83 , 29 tegmark m. , hamilton a. j. s. , xu y. , 2002 , mnras , 335 , 887 ( astro - ph/0111575 ) tegmark m. et al .
( sdss collaboration ) , 2003 , apj , to be submitted york d. g. , et al .
( sdss collaboration , 144 authors ) , 2000 , aj , 120 , 1579
this appendix proves the following theorem , invoked in [ balkanize2 ] ( see table [ deftable ] for a definition of the term * group * ) : first , suppose that a polygon contains a region which is connected but not simply - connected .
it is required to prove that the boundaries of this region belong to at least two distinct groups . by definition of simply - connectedness
, a continuous line can be drawn entirely inside the non - simply - connected region such that one connected boundary of the region lies entirely on one side of the line , and another connected boundary of the region lies entirely on the other side of the line . in the polygon of figure [ eyefig ] , for example
, such a line would circulate around the central boundary while remaining inside the outer boundary .
the continuous line can not intersect any of the circles forming the caps of the polygon , because if the line did intersect a circle , then the line would be inside the polygon on one side of the intersection , and outside the polygon on the other side of the intersection , contradicting the assumption that the line lies entirely inside the polygon . hence the continuous line must partition the circles of the polygon into two non - intersecting groups .
the two connected boundaries on either side of the continuous line must therefore belong to two distinct groups , as was to be demonstrated .
conversely , suppose that a polygon has boundaries that belong to at least two distinct groups .
it is required to prove that the two groups delineate a connected but non - simply - connected region of the polygon .
consider the polygon , call it a , formed by the intersection of all the caps of one group .
the polygon a so formed must consist of one or more simply - connected parts ; for if any connected part of a were not simply - connected , then according the previous paragraph the boundaries of that part would belong to different groups , contradicting the assumption that the caps of a all belong to a single group .
now consider similarly the polygon , call it b , formed by the intersection of all the caps of a second group .
the boundaries of polygon b must lie entirely inside one and only one of the simply - connected parts of a. for certainly b must lie inside at least one part of polygon a , since a must enclose the parent polygon , and the boundary of b lies inside ( at the border of ) the parent polygon .
but the boundary of polygon b can not lie in more than one part of a , because if it did then , because the circles of b are all in the same group and therefore connected to each other , there would be a path lying along the circles of b traversing continuously from one part of a to another , and therefore necessarily intersecting one of the boundaries of polygon a , contradicting the assumption that the circles of a and b belong to distinct groups that nowhere intersect . similarly , the boundaries of polygon a must lie entirely inside one and only one of the simply - connected parts of b. this argument has identified two special boundaries : a boundary of a that entirely encloses b ; and a boundary of b that entirely encloses a. the region enclosed by these two special boundaries delineates a region of the polygon that is connected but not simply - connected . for consider a continuous line which is displaced slightly off the special boundary of a , towards the special boundary of b. by construction , the continuous line lies entirely inside both polygons a and b. the continuous line can not intersect any of the circles of a or b , because if it did then the line would lie inside a ( or b ) on one side of the intersection , and outside a ( or b ) on the other side , contradicting the fact that the line lies entirely inside a and b. the continuous line could possibly intersect circles belonging to a third group of circles of the parent polygon . however , if the continuous line is displaced off the special boundary of a by a sufficiently small amount that it does not encounter any third group , then the continuous line will lie entirely inside the parent polygon .
this continous line forms a line inside the polygon that can not be shrunk continuously to a point , so the polygon must contain a region that is connected but not simply - connected , as was to be demonstrated . | this paper presents a scheme to deal accurately and efficiently with complex angular masks , such as occur typically in galaxy surveys .
an angular mask is taken to be an arbitrary union of arbitrarily weighted angular regions bounded by arbitrary numbers of edges .
the restrictions on the mask are ( i ) that each edge must be part of some circle on the sphere ( but not necessarily a great circle ) , and ( ii ) that the weight within each subregion of the mask must be constant .
the scheme works by resolving a mask into disjoint polygons , convex angular regions bounded by arbitrary numbers of edges .
the polygons may be regarded as the ` pixels ' of a mask , with the feature that the pixels are allowed to take a rather general shape , rather than following some predefined regular pattern . among other things ,
the scheme includes facilities to compute the spherical harmonics of the angular mask , and data - random @xmath0 and random - random @xmath1 angular integrals .
a software package mangle which implements this scheme , along with complete software documentation , is available at http://casa.colorado.edu/~ajsh / mangle/. large - scale structure of universe methods : data analysis |
the class of problems known as assembly line balancing problems ( albps ) concerns the optimization of processes related to the manufacturing of products via assembly lines .
their importance in the industrial world is shown by the fact that much research efforts have been dedicated to many different types of albps during the past 50 - 60 years @xcite .
the specific problem considered in this paper is the so - called simple assembly line balancing problem ( salbp ) @xcite , a well - studied scientific test case .
an assembly line is composed of a set of work stations arranged in a line , and by a transport system which moves the product to be manufactured along the line .
the product is manufactured by executing a given set of tasks .
each of these tasks has a pre - defined processing time . in order to obtain a solution to a given salbp instance ,
all tasks must be assigned to work stations subject to precedence constraints between the tasks . in the context of the salbp
, all work stations are considered to be of equal size .
moreover , the assembly line is assumed to move in constant speed .
this implies a maximum of @xmath0 time units the so - called _
cycle time_for processing the tasks assigned to each work station .
the salbp has been tackled with several objective functions among which the following ones are the most studied ones in the literature : * given a fixed cycle time @xmath0 , the optimization goal consists in minimizing the number of necessary work stations .
this version of the problem is refered to as salbp-1 . *
given a fixed number @xmath1 of work stations , the goal is to minimize the cycle time @xmath0 .
the literature knows this second problem version as salbp-2 .
the feasibility problem salbp - f arises when both a cycle time @xmath0 and a number of work stations @xmath1 is given and the goal is to find a feasible solution respecting @xmath0 and @xmath1 .
in this work we will deal with the salbp-2 version of the problem . for what concerns the comparison between salbp-1 and salbp-2 ,
much of the scientific work has been dedicated to the salbp-1 .
however , also for the salbp-2 exists a considerable body of research papers .
an excellent survey was provided by @xcite .
approaches for the salbp-2 can basically be classified as either _ iterative solution approaches _ or _ direct solution approaches_. iterative approaches tackle the problem by iteratively solving a series of salbp - f problems that are obtained by fixing the cycle time .
this process is started with a cycle time that is set to some calculated upper bound .
this cycle time is then decremented during the iterative process , which stops as soon as no solution for the corresponding salbp - f problem can be found .
in contrast to these indirect approaches , direct approaches intend to solve a given salbp-2 instance directly .
heuristic as well as complete approaches have been devised for the salbp-2 . among the existing complete methods we find iterative approaches such as the ones proposed in @xcite but also direct approaches such as the ones described in @xcite .
moreover , the performance of different integer programming formulation of the salbp-2 has been evaluated in @xcite .
the currently best - performing exact method is salome-2 @xcite .
surprisingly this exact method even outperforms the existing heuristic and metaheuristic approaches for the salbp-2 .
the most successful metaheuristic approach to date is a tabu search method proposed in @xcite .
another tabu search proposal can be found in @xcite .
other metaheuristic approaches include evolutionary algorithms @xcite and simulated annealing @xcite .
finally , a two - phase heuristic based on linear programming can be found in @xcite , whereas a heuristic based on petri nets was proposed in @xcite .
[ [ contribution - of - this - work . ] ] contribution of this work .
+ + + + + + + + + + + + + + + + + + + + + + + + + + subsequently we propose to tackle the salbp-2 by means of an iterative approach based on beam search , which is an incomplete variant of branch & bound .
the resulting iterative beam search algorithm is inspired by one of the current state - of - the - art methods for the salbp-1 , namely beam - aco @xcite .
beam - aco is a hybrid approach that is obtained by combining the metaheuristic ant colony optimization with beam search . in this work
we propose to use the beam search component of beam - aco in an iterative way for obtaining good salbp-2 solutions .
our computational results show indeed that the proposed algorithm is currently a state - of - the - art method for the salbp-2 .
it is able to obtain optimal , respectively best - known , solutions in 283 out of 302 test cases .
moreover , in further 9 cases the algorithm is able to produce new best - known solutions .
[ [ organization - of - the - paper . ] ] organization of the paper .
+ + + + + + + + + + + + + + + + + + + + + + + + + + in section [ sec : salbp-2 ] we present a technical description of the tackled problem .
furthermore , in section [ sec : algo ] the proposed algorithm is described .
finally , in section [ sec : results ] we present a detailed experimental evaluation and in section [ sec : conclusions ] we conclude our work and offer an outlook to future work .
the salbp-2 can technically be described as follows .
an instance @xmath2 consists of three components .
@xmath3 is a set of @xmath4 tasks .
each task @xmath5 has a pre - defined processing time @xmath6 .
moreover , given is a precedence graph @xmath7 , which is a directed , acyclic graph with @xmath8 as node set .
finally , @xmath1 is the pre - defined number of work stations which are ordered from 1 to @xmath1 .
an arc @xmath9 indicates that @xmath5 must be processed before @xmath10 .
given a task @xmath10 , @xmath11 denotes the set of tasks that must be processed before @xmath12 .
a feasible solution is obtained by assigning each task to exactly one work station such that the precedence constraints between the tasks are satisfied .
the objective function consists in minimizing the so - called cycle time .
the salbp-2 can be expressed in the following way as an integer programming ( ip ) problem . @xmath13
subject to : @xmath14 this ip model makes use of the following variables and constants : @xmath15 is a binary variable which is set to 1 if and only if task @xmath5 is assigned to work station @xmath16 .
the objective function ( 1 ) minimizes the cycle time @xmath17 . , while fixed cycle times are denoted by @xmath0 . ] the constraints ( 2 ) ensure that each task @xmath5 is assigned to a single work station @xmath16 .
constraints ( 3 ) reflect the precedence relationships between the tasks .
more specifically , if task @xmath10 is assigned to a work station @xmath16 , all tasks @xmath18 must be assigned to work stations @xmath19 with @xmath20 .
the constraints ( 4 ) ensure that the sum of the processing times of the tasks assigned to a work station @xmath16 do not exceed the cycle time @xmath21 . note that this model is not necessarily the most efficient ip model for solving the salbp-2 .
an evaluation of several different models can be found in @xcite .
the following solution representation is used for the description of the algorithm as given in section [ sec : algo ] . a solution
@xmath22 is an ordered list @xmath23 of @xmath1 sets of tasks , where @xmath24 denotes the set of tasks that are assigned to the @xmath25-th work station . abusing notation we henceforth call @xmath24 a work station .
note that for a solution @xmath22 to be valid the following conditions must be fulfilled : 1 .
@xmath26 and @xmath27 , that is , each task is assigned to exactly one work station .
2 . for each task
@xmath28 it is must hold that @xmath29 .
this ensures that the precedence constraints between the tasks are not violated .
the reverse problem instance @xmath30 with respect to an original instance @xmath2 is obtained by inverting the direction of all arcs of @xmath31 .
it is well - known from the literature @xcite that tackling the reverse problem instance may lead an exact algorithm faster to an optimal solution , respectively , may provide a better heuristic solution when tackled with the same heuristic as the original problem instance . moreover , a solution @xmath32 to the reverse problem instance @xmath30 can easily be converted into a solution @xmath22 to the original problem instance @xmath2 as follows : @xmath33
as mentioned in the introduction , the basic component of our algorithm for the salbp-2 consists of beam search ( bs ) , which is an incomplete derivative of branch & bound @xcite .
initially bs has especially been used in the context of scheduling problems ( see , for example , @xcite ) . to date only very few applications to other types of problems
exist ( see , for example , @xcite ) . in the following
we briefly describe how one of the standard versions of bs works .
the crucial aspect of bs is the parallel extension of partial solutions in several ways . at all times
, the algorithm keeps a set @xmath34 of at most @xmath35 partial solutions , where @xmath34 is the so - called _ beam _ , and @xmath35 is known as the _ beam width_. at each step , at most @xmath36 feasible extensions of each partial solution in @xmath34 are selected on the basis of greedy information . in general , this selection is done deterministically . at the end of each step
, the algorithm creates a new beam @xmath34 by choosing up to @xmath35 partial solutions from the set of selected feasible extensions . for that purpose ,
bs algorithms determine in the case of minimization a lower bound value for each extension . only the maximally @xmath35
best extensions with respect to these lower bound values are included in @xmath34 .
finally , if any complete solution was generated , the algorithm returns the best of those .
note that the underlying constructive heuristic that defines feasible extensions of partial solutions and the lower bound function for evaluating partial solutions are crucial for the working of bs . in the following
we first present a description of the implementation of the bs component , before we describe the algorithmic scheme in which this bs component is used . the bs component described in this section
see algorithm [ algo : bs ] for the pseudo - code is the main component of the proposed algorithm for the salbp-2 .
the algorithm requires a problem instance @xmath2 , a fixed cycle time @xmath0 , a beam width @xmath35 , and a maximal number of extensions @xmath36 as input .
given a fixed cycle time @xmath0 and @xmath1 ( the number of work stations ) bs tries to find at least one feasible solution .
as mentioned before , the crucial aspect of bs is the extension of partial solutions in several possible ways . at each step
the algorithm extends each partial solution from @xmath34 in a maximum number of ways .
more specifically , given a partial solution with @xmath37 work stations already filled , an extension is generated by assigning a set of so - far unassigned tasks to the next work station @xmath38 such that the given cycle time @xmath0 is not surpassed and the precedence constraints between the tasks are respected ( see lines 1112 of algorithm [ algo : bs ] ) .
the algorithm produces extensions in a ( partially ) probabilistic way rather than in the usual deterministic manner .
each generated extension ( partial solution ) is either stored in set @xmath39 in case it is a complete solution , or in set @xmath40 otherwise ( see lines 1319 of algorithm [ algo : bs ] ) .
however , a partial solution is only stored in set @xmath40 if it uses at most @xmath41 work stations , and if its @xmath42-th work station is different to the @xmath42-th work station of all partial solutions that are already in @xmath34 .
finally , bs creates a new beam @xmath34 by selecting up to @xmath43 solutions from set @xmath40 of further extensible partial solutions ( see line 22 of algorithm [ algo : bs ] ) .
this is done in function selectsolutions(@xmath40,@xmath43 ) on the basis of a lower bound function lb@xmath44 . in the following we describe in detail the extension of partial solutions and the working of function selectsolutions(@xmath40,@xmath43 ) .
an instance @xmath2 , a fixed cycle time @xmath0 , a beam width @xmath35 , and @xmath36 @xmath45 initialization of an empty solution @xmath22
@xmath46 @xmath47 @xmath48 @xmath49 @xmath50 @xmath51 extendpartialsolution(@xmath52 ) @xmath53
@xmath54 @xmath55selectsolutions(@xmath40,@xmath43 ) if @xmath56 the output is @xmath57 , otherwise @xmath58 [ [ extending - partial - solutions . ] ] extending partial solutions . + + + + + + + + + + + + + + + + + + + + + + + + + + + + the generation of an extension of a partial solution @xmath59 with @xmath60 work stations already filled works as follows .
unassigned tasks are iteratively assigned to work station @xmath61 until the sum of their processing times is such that no other task can be added to @xmath61 without exceeding the given cycle time @xmath0 .
this procedure is pseudo - coded in algorithm [ algo : fill - station ] . at each step
, @xmath62 denotes the set of so - far unassigned tasks that may be added to @xmath61 without violating any constraints .
the definition of this set of _ available tasks _ is given in line 3 , respectively 8 , of algorithm [ algo : fill - station ] .
a partial solution @xmath59 , the index @xmath42 of the work station to be filled , and the cycle time @xmath0
@xmath63 @xmath64 @xmath65 @xmath66choosetask(@xmath67 ) @xmath68
@xmath64 @xmath69 filled work station @xmath61 it remains to describe the implementation of function choosetask(@xmath67 ) of algorithm [ algo : fill - station ] .
for that purpose let us first define the following subset of @xmath62 : @xmath70 this definition is such that @xmath71 contains all tasks that _ saturate _ , in terms of processing time , the @xmath42-th work station @xmath38 .
the choice of a task from @xmath62 is made on the basis of greedy information , that is , on the basis of values @xmath72 that are assigned to all tasks @xmath73 by a greedy function .
the first action for choosing a task from @xmath62 consists in flipping a coin for deciding if the choice is made deterministically , or probabilistically . in case of a deterministic choice
, there are two possibilities .
first , if @xmath74 , the best task from @xmath71 is chosen , that is , the task with maximal greedy value among all tasks in @xmath71 .
otherwise , we choose the task with maximal greedy value from @xmath62 . in case of a probabilistic decision , a task from @xmath62 is chosen on the basis of the following probability distribution : @xmath75 for completing the description of function choosetask(@xmath67 ) , we must describe the definition of the greedy values @xmath76 , @xmath77 . in a first step a term @xmath78
is defined as follows : @xmath79 hereby , @xmath80 denotes the set of all tasks that can be reached from @xmath25 in precedence graph @xmath31 via a directed path .
this definition combines two greedy heuristics that are often used in the context of assembly line balancing problems .
the first one concerns the task processing times and the second one concerns the size of @xmath80 .
the influence of both heuristics can be adjusted via the setting of weights @xmath81 and @xmath82 . in order to be more flexible we decided to allow for both weights a value from @xmath83 $ ] .
this means that we consider for each heuristic potentially also its negation .
given the @xmath78-values , the greedy values @xmath76 are then derived as follows : @xmath84 where @xmath85 , respectively @xmath86 , denote the minimum , respectively maximum , values of all @xmath78 .
interestingly , for obtaining well - working greedy values , parameters @xmath81 and @xmath82 have to be chosen in a problem - instance - dependent way .
tuning experiments are present in section [ sec : tuning ] .
[ [ the - lower - bound - function . ] ] the lower bound function .
+ + + + + + + + + + + + + + + + + + + + + + + + + the new beam @xmath34 is at each step chosen from @xmath40 .
this choice is implemented by function selectsolutions(@xmath40,@xmath43 ) of algorithm [ algo : fill - station ] .
first , the solutions in @xmath40 are ranked with respect to increasing lower bound values lb(@xmath87 ) .
then , the min@xmath88 highest ranked partial solutions from @xmath40 are selected .
let us denote by @xmath89 the set of tasks that have not yet been assigned to work stations in partial solution @xmath59 .
then : @xmath90 note that this lower bound is inspired by splitting - based bounds for the one - dimensional bin - backing problem .
an instance @xmath2 @xmath91 determinestartingcycletime ( ) @xmath92 , @xmath93
@xmath94 @xmath95 beamsearch(@xmath96 ) @xmath95 beamsearch(@xmath97 ) @xmath98 * then * @xmath99 * end if * @xmath100 @xmath101 @xmath94 within 5% of time limit * then * @xmath102 , @xmath103 * else * @xmath104 , @xmath105 * end if * @xmath95 beamsearch(@xmath96 ) @xmath95 beamsearch(@xmath97 ) @xmath98 * then * @xmath100 * else * @xmath106 * end if * @xmath99 cycle time @xmath0 the bs component outlined in the previous section is used by an iterative algorithmic scheme that is presented in algorithm [ algo : scheme ] .
henceforth this algorithmic scheme is labelled iterated beam search ( ibs ) .
the first step consists in determining a starting cycle time @xmath0 , which is computed in funcion determinestartingcycletime ( ) of algorithm [ algo : scheme ] as @xmath107 the algorithm works in two phases . in the first phase
( see lines 3 - 11 of algorithm [ algo : scheme ] ) the algorithm tries to quickly find a first cycle time @xmath0 for which a valid solution can be found . for this purpose bs
is applied with the setting @xmath108 and @xmath109 .
note that this setting was chosen after tuning by hand .
moreover , note that the first phase only takes a fraction of a second of computation time .
this holds for all instances considered in section [ sec : results ] .
the second phase of the algorithm iteratively tries to find a valid solution for the next smaller cycle time . in this phase ,
the algorithm disposes over a certain time limit for each considered cycle time .
remember that the working of bs is partially probabilistic .
therefore , bs can repeatedly be applied to the same instance with potentially different outcomes .
the first five percent of the above - mentioned time limit are spent by bs applications that use the setting @xmath102 and @xmath103 .
this is done with the intention of not wasting too much computation time , if not necessary .
however , if bs is not able to solve the given cycle time with this setting , the remaining 95% of the available time are spent by bs applications using the setting @xmath104 and @xmath105 . with this setting bs
is considerably slower .
however , the probability of finding feasible solutions is much higher than with the setting described before .
the second phase of the algorithm ends when the time limit has passed without having found a feasible solution for the considered cycle time .
ibs was implemented in ansi c++ , and gcc 3.4.0 was used for compiling the software .
experimental results were obtained on a pc with an amd64x2 4400 processor and 4 gb of memory . in the following
we first describe the set of benchmark instances that we used for the experimental evaluation .
subsequently we present the tuning process that we conducted and the results of the proposed algorithm .
we used the usual set of 302 benchmark instances from the literature .
they can be obtained together with information about optimal and best - known solutions from a website especially dedicated to all kind of assembly line balancing problems maintained by armin scholl , http://www.assembly-line-balancing.de .
each instance consists of a precedence graph @xmath31 and a given number @xmath1 of work stations .
the benchmark set is composed of two subsets of instances , henceforth called dataset1 and dataset2 .
dataset1 consists of 128 instances based on 9 different precedence graphs with a number of tasks between 29 to 111 .
dataset2 is composed of 174 instances based on 8 different precedence graphs with a number of tasks varying from 53 to 297 . during preliminar experiments we realized that parameters @xmath35 and @xmath36 have a rather low impact on the final results of ibs .
in other words it is easy to find a reasonable setting for these parameters quite quickly .
their setting dynamically changes during a run of the algorithm as specified in section [ sec : algo - scheme ] . on the contrary , parameters @xmath81 and @xmath82 ( see eq .
[ eq : gamma ] ) have a high impact on the algorithms performance .
remember that @xmath81 is the weight of the greedy function concerning the task processing times , while @xmath82 is the weight of the greedy function concerning the number of tasks that have to be processed after the task under consideration .
as mentioned before , for both parameters we allowed values from @xmath83 $ ] . instead of trying to find a good parameter setting for each instance
, we decided for a tuning process based on precedence graphs , that is , we wanted to choose a single setting of @xmath81 and @xmath82 for all instances concerning the same precedence graph .
for that purpose we applied a specific version of ibs for all combinations of @xmath110 to all 302 instances .
this makes a total of 441 different settings for each instance . the specific version of ibs that we applied for the tuning process differs from ibs as outlined in algorithm [ algo : scheme ] in that lines 14 - 24 were replaced by a single , deterministic , application of beam search with @xmath111 and @xmath112
this was done for the purpose of saving computation time . based on the obtained results we chose the settings presented in table [ tab : tuning ] for the different precedence graphs .
it is interesting to note that , apart from a few exceptions , the greedy heuristic based on task processing times does not seem necessary for obtaining good results .
.[tab : tuning ] values of parameters @xmath81 and @xmath82 for the final experiments . [ cols="<,<,<,<,<,<",options="header " , ]
in this work we have proposed an iterative beam search algorithm for the simple assembly line balancing problem with a fixed number of work stations , salbp-2 .
the experimental evalution of the algorithm has shown that it is currently a state - of - the - art method for this problem .
appart from producing optimal , respectively best - known , solutions in 283 out of 302 test cases , our algorithm generated new best - known solutions in further 9 test cases . encouraged by the results for the salbp-1 version of the problem ( as published in @xcite ) and the results obtained in this paper for the salbp-2 we intent to apply similar algorithms based on beam search to other assembly line balancing problems .
this work was supported by grant tin2007 - 66523 ( formalism ) of the spanish government .
moreover , christian blum acknowledges support from the _ ramn y cajal _ program of the spanish ministry of science and innovation .
many thanks go to armin scholl for verifying the new best - known solutions found by the algorithm proposed in this work .
finally , we would also like to express our thanks to cristbal miralles who was involved as a co - author of a similar work for the more general assembly line worker assignment and balancing problem
. 10 h. akeba , m. hifib , and r. mhallah . a beam search algorithm for the circular packing problem .
, 36(5):15131528 , 2009 . e. j. anderson and m. c. ferris .
genetic algorithms for combinatorial optimization : the assembly line balancing problem .
, 6:161173 , 1994 . c. blum . for simple assembly line balancing
, 20(4):618627 , 2008 .
c. blum , m. j. blesa , and m. lpez ibez .
beam search for the longest common subsequence problem .
, 36(12):31783186 , 2009 . c. blum and c. miralles . on solving the assembly line worker assignment and balancing problem via beam search
, 38(1):328339 , 2011 .
the application of a tabu search metaheuristic to the assembly line balancing problem .
, 77:209227 , 1998 .
m. ghirardi and c. n. potts .
makespan minimization for scheduling unrelated parallel machines : a recovering beam search approach .
, 165(2):457467 , 2005 .
s. gosh and r. j. gagnon . a comprehensive literature review and analysis of the design , balancing and scheduling of assembly systems .
, 27:637670 , 1989 .
s. t. hackman , m. j. magazine , and t. s. wee .
fast , effective algorithms for simple assembly line balancing problems . , 37:916924 , 1989 .
a. henrici . a comparison between simulated annealing and tabu search with an example from the production planning . in h.
dyckhoff et al . , editor , _ operations research proceedings _ , pages 498503 .
springer verlag , berlin , germany , 1994 .
o. kilincci .
a petri net - based heuristic for simply assembly line balancing problem of type 2 . , 46:329338 , 2010 .
r. klein and a. scholl . maximizing the production rate in simple assembly line balancing a branch and bound procedure .
, 91:367385 , 1996 .
lee and d. l. woodruff .
beam search for peak alignment of nmr signals .
, 513(2):413416 , 2004 .
a. c. nearchou .
balancing large assembly lines by a new heuristic based on differential evolution .
34:10161029 , 2007 . p. s. ow and t. e. morton .
filtered beam search in scheduling .
, 26:297307 , 1988 . r. pastor , l. ferrer , and a. garca . evaluating optimization models to solve salbp . in o.
gervasi and m. l. gavrilova , editors , _ proceedings of iccsa 2007 international conference on computational science and its applications _ ,
volume 4705 of _ lecture notes in computer science _ , pages 791803 .
springer verlag , berlin , germany , 2007 .
i. sabuncuoglu and m. bayiz .
job shop scheduling with beam search .
, 118:390412 , 1999 .
m. e. salveson .
the assembly line balancing problem .
, 6:1825 , 1955 .
a. scholl . . in h.
dyckhoff et al . , editor , _ operations research proceedings _ , pages 175181 .
springer verlag , berlin , germany , 1994 .
a. scholl . .
physica verlag , heidelberg , germany , 2nd edition edition , 1999 .
a. scholl and c. becker .
state - of - the - art exact and heuristic solution procedures for simple assembly line balancing . , 168(3):666693 , 2006 .
a. scholl and s. voss .
simple assembly line balancing
heuristic approaches .
, 2:217244 , 1996 .
h. f. ugurdag , r. rachamadugu , and c. a. papachristou . designing paced assembly lines with fixed number of stations .
, 102:488501 , 1997 .
j. m. s. valente and r. a. f. s. alves . filtered and recovering beam search algorithms for the early / tardy scheduling problem with no idle time .
, 48(2):363375 , 2005 .
t. watanabe , y. hashimoto , l. nishikawa , and h. tokumaru .
line balancing using a genetic evolution model .
, 3:6976 , 1995 . | the simple assembly line balancing problem ( salbp ) concerns the assignment of tasks with pre - defined processing times to work stations that are arranged in a line .
hereby , precedence constraints between the tasks must be respected .
the optimization goal of the salbp-2 version of the problem concerns the minimization of the so - called cycle time , that is , the time in which the tasks of each work station must be completed . in this work
we propose to tackle this problem with an iterative search method based on beam search .
the proposed algorithm is able to obtain optimal , respectively best - known , solutions in 283 out of 302 test cases .
moreover , for 9 further test cases the algorithm is able to produce new best - known solutions .
these numbers indicate that the proposed iterative beam search algorithm is currently a state - of - the - art method for the salbp-2 . |
beyond the original goal of detecting gravitational waves , modern interferometric sensors ( ifo ) ( such as the future advanced ligo ( aligo ) @xcite , the advanced virgo ( avirgo ) @xcite , the lcgt @xcite , the aei 10 m @xcite interferometer ; and the proposed third generation detectors such as the einstein telescope ( et ) @xcite ) can be viewed as unique pioneering instruments capable of measuring induced displacements below the @xmath2 scale @xcite .
this feature opens up myriads of new possibilities for fundamental science . in searches for gravitational waves ( gw )
the signature of the induced displacement of the ifo s test mass ( tm ) ranges from the basically unknown to the well - predicted but yet undetected .
in contrast to this , one may design a precision experiment in which the local gravitational field around the tm of the interferometric sensor is modulated at a given frequency on a well - controlled manner by a dynamic gravity field generator ( dfg ) .
a dfg consists essentially of a rotating mass of null odd and non - null even moments , from which the quadrupole term dominates . in the newtonian limit ,
its effect on the detector is to add a signal at the even - harmonics of the rotation frequency .
a well - characterized dfg has the potential to provide sub - percent calibration for the gw detectors in phase and amplitude as well as to evaluate current theoretical newtonian noise estimates .
this is the subject of a separate publication @xcite .
it has been shown in the past that devices capable of producing gravity field gradients can be employed , in conjunction with a suitable displacement sensor , for testing newton s @xmath0 law in the laboratory scale . in the first experiment in 1967 forward and
miller @xcite used an orbiter sensor , originally developed for measuring the lunar mass distribution , to test newton s @xmath0 law in the @xmath3 scale .
weber and sinsky @xcite used a gw bar detector as a sensor , acoustically stressing a volume of matter at @xmath4 , measuring an excess in noise in the detector consistent to theory . in the 1980s at the university of tokyo , a series of experiments were carried through to test violations to the inverse square law ( isl ) up to a distance of @xmath5 @xcite . in these studies ,
the coupling between the dynamic gravity field generated by a rotating mass , and the quadrupole moment of a mechanical oscillator antenna was measured as a function of the rotor - antenna separation . with this method
limits on non - newtonian gravity were provided and a measurement of the newtonian constant , g , was found in agreement with previous experiments @xcite . in the 1990s ,
the gravitational - wave group at the university of rome developed and carried out experiments @xcite on the cryogenic gw bar detector , explorer , at cern .
a device , rotating at a frequency close to half of the antenna s resonant frequency was developed , and the resulting dynamic field was measured as a function of the source - sensor separation .
the results were then used to derive upper limits to yukawa - like gravitational potential violations at laboratory scale .
the experimental designs cited above consisted in the use of a single dfg in conjunction with bar type gw detectors . in this article we propose a significantly different design , exploiting the exceptional bandwidth and sensitivity of state - of - the - art interferometric sensors .
we illustrate the method through the sensitivity of advanced and third - generation gravitational wave detectors to emphasize the feasible scientific reach . however , it is expected that due to the technical details of gravitational wave detectors , the dfgs based isl measurements shall require dedicated custom configured interferometry .
nevertheless , the sensitivity measure we derive is a solid baseline .
the concept is based on a null - experiment configuration , where a _ pair _ of well matched and symmetrical dfgs , rotating at the same frequency but @xmath6 out of phase , generates a null - signal in the absence of violations to newton s @xmath0 law . in the presence of violations , and within experimental uncertainties , the effect of the two dfgs would not cancel , and a measure of such deviations would be achievable . + the most common way of interpreting composition independent tests of non - newtonian gravity ( for review
see e.g. @xcite and @xcite ) is through the yukawa formalism , where an additional so - called yukawa term ( @xmath7 ) is added to the classical newtonian gravitational potential ( @xmath8 ) : @xmath9\ ] ] in this expression @xmath10 is the gravitational constant , @xmath11 and @xmath12 are the interacting masses , and @xmath13 denotes the distance between the two point masses .
the yukawa parameters of interest are @xmath14 , denoting the yukawa interaction coupling strength , and @xmath15 , giving the length scale of the coupling . for a yukawa range of @xmath16 , the current limit on
the coupling strength , @xmath14 , is in the order of @xmath17 for both negative , positive , and absolute value of @xmath14 .
( see fig .
[ summary ] . ) .
the concept designed here shows promise to explore deviations from isl significantly below present bounds , maybe even down to the order of @xmath18 if enabled by technology and cost .
in this work we study the application of future interferometric sensors for artificially generated gravity fields by a pair of hypothetical dfgs in a null experiment configuration .
we show that a pair of symmetrical and matched dfgs cancel each other s effect on the tm at twice the rotation frequency in the newtonian limit .
thus we relate the estimated test mass displacement to non - newtonian parametrization of the law of gravitation .
we also consider the contribution from some of the undesired system asymmetries due to production / measurement uncertainties .
we numerically compute the tolerances of a hypothetical experimental setup to measure yukawa - like deviations from the @xmath0 law and we estimate a limit on the coupling strengths , @xmath14-s , that might be achieved with future measurements .
consider a hypothetical dfg , consisting of two point masses , @xmath19 and @xmath20 , separated by a distance of @xmath21 and @xmath22 respectively from the center of rotation , rotating at a frequency of @xmath23 .
a point mass , @xmath12 , representing the tm , is then placed at distance @xmath24 from the dfg s center of rotation .
and @xmath20 are rotating with a frequency of @xmath25 and radii of @xmath21 and @xmath22 from the center of rotation , respectively .
the dfg center of rotation is placed in a distance of @xmath26 from the mass @xmath12 , representing the tm center of mass .
the distance between the dfg masses and the tm are @xmath27 and @xmath28 , respectively , @xmath29 , and only accelerations of the tm along the ifo cavity axis are considered.,title="fig:",width=326 ] + we first calculate the acceleration of @xmath12 induced by the dfg along the axis connecting @xmath12 and the center of rotation of the dfg .
this axis corresponds to the optical axis of the ifo . introducing the dimensionless parameters @xmath30 , and @xmath31 ,
the distance between the dfg s @xmath32-th mass and the point mass @xmath12 , @xmath33 , can be written as @xmath34 where @xmath29 .
the newtonian potential at the position of point mass @xmath12 is then given by @xmath35 the magnitude of the induced acceleration on @xmath12 along the optical axis can be written as @xmath36 here @xmath37 is a geometrical factor @xmath38 as we have shown elsewhere @xcite , for the case of a much smaller lever arm @xmath39 than distance @xmath26 ( @xmath40 ) , considering only the dominant terms in the first few harmonics , the induced displacement of a free mass @xmath12 along the axis connecting it to the dfg s center of rotation can be approximated as @xmath41}\end{aligned}\ ] ] where @xmath42 describes the @xmath43-th multipole moment of the dfg s mass distribution .
we observe that using eq .
( [ displ2 ] ) and in the case of a dfg mass distribution invariant under rotation by pi , all odd moments vanish and the induced displacement is dominated by the quadrupole moment @xmath44 at twice the rotation frequency .
we note , that the solution represented in eq .
( [ displ2 ] ) is valid for a free body , which , in case of a suspended tm is a good approximation for frequencies well above the eigenfrequencies of the suspension ( typically around @xmath45 @xcite ) . in the same way as above
, we calculate the field dynamics arising from a yukawa - like potential perturbation to the classical newtonian field . in view of the fact that the potentials are additive and all operations to calculate the acceleration are linear
, the acceleration of point mass @xmath12 due to the two potential terms will also be additive .
using eq .
( [ pot ] ) , the yukawa perturbation to the newtonian potential at @xmath12 due to the dfg can be expressed as @xmath46 where @xmath47 is given by eq .
( [ hi ] ) . in the same way ,
the magnitude of the tm induced acceleration along the optical axis is @xmath48 where @xmath49 , and @xmath50 is a function of the length scale of the yukawa coupling , @xmath15 : @xmath51 } & \times & \\ \nonumber & & \hskip -1.5 in \times \exp{\bigg{(}-\frac{d}{\lambda } \sqrt{1 + r_\mathrm{i}^2 - 2r_\mathrm{i } \cos \theta}\bigg{)}}\end{aligned}\ ] ] analytical integration of eq .
( [ cla_acc ] ) and eq .
( [ yuk_acc ] ) can not be easily formulated without using some additional level of approximation .
for this reason , we treat the problem numerically , facilitating the treatment of some of the experimental uncertainties .
we have modeled both the newtonian and the non - newtonian dynamics on the tm while taking into account many of the dfg fabrication and measurement procedural uncertainties . in this work
the tm is approximated as a driven , damped pendulum .
let the pendulum s longitudinal eigenfrequency be @xmath52 with quality factor @xmath53 .
casting the differential equations of motion ( eq . ( [ cla_acc ] ) and eq . ( [ yuk_acc ] ) ) in the laplace frequency domain , the transfer function relating the acceleration @xmath54 to the induced displacement @xmath55 is @xmath56 where @xmath57 ( the upper index @xmath58 denotes that the same expression applies to both newtonian and yukawa case ) .
we used monte carlo simulations to model a null - experiment design using dfgs to measure hypothetical violations to the classical inverse square law . to do this , we first numerically computed the induced acceleration due to the newtonian and non - newtonian terms using eq .
( [ cla_acc ] ) and eq .
( [ yuk_acc ] ) .
these results , generated in the time domain , are then mapped in the fourier - domain to generate an acceleration spectrum , which is then filtered through the transfer function presented in eq .
( [ susp ] ) .
the displacement contributions at the different harmonics , for different sets of parameters and uncertainties are then computed and analyzed . to assess the short and long term feasibility of the proposed method , the simulated results are compared to the displacement sensitivity of future most sensitive interferometric gw detectors .
[ senscurves ] . shows the design sensitivities for aligo , avirgo , the et ( for a reference to these , see @xcite ) , lcgt @xcite , and the aei 10 m detector @xcite .
two conceptualized dfgs are placed in - line with one of the ifo s arms at a distance of @xmath59 and @xmath60 from the tm .
the dfgs positions and quadrupole moments are chosen such that when operated at a relative phase of @xmath61 the net displacement at @xmath62 will largely depend on the non - newtonian coupling strength @xmath14.,title="fig:",width=326 ] + in order to mitigate the effect of the ifo s calibration uncertainty , two dfgs rotating at the same frequency ( @xmath25 ) , but out of phase ( @xmath63 ) , are set up to generate a null effect on the tm in the newtonian limit at twice the rotation frequency .
for this reason , the uncertainties of the proposed measurement , whose setup is shown in fig . [ tworotord ] .
, will solely depend on the production and procedural uncertainties of dfgs .
for example , similarly to the one proposed in @xcite as a detector calibration tool , a single hypothetical dfg may consist of an aircraft grade ( 6al/6v/2sn ) titanium disc @xmath64 in diameter and @xmath3 in height .
the disc has two cylindrical slots , @xmath65 apart ( @xmath66 ) , which can hold different materials .
tungsten cylinders , @xmath67 in diameter , can serve as rotating masses , with @xmath68 effective mass each , where @xmath69 and @xmath70 are mass densities of tungsten and titanium respectively , and @xmath71 is the volume of a cylinder . in the ideal case of symmetrical and matched dfgs ,
no effect on the tm is observed at @xmath62 in the absence of @xmath0 violations . in the classical sense , this condition is satisfied if the dfg radii and placement follow the @xmath72 scaling laws , where indices @xmath73 and @xmath74 are associated with the two dfgs , and @xmath75 is the mass ratio ( see fig .
[ tworotord ] . ) .
note , that if we take into account the cylindrical geometry of the tm , the scaling laws presented in eq .
( [ etarelations ] ) will differ by o@xmath76 , where the corresponding @xmath77 and @xmath60 will still be analytically calculable .
in non - newtonian dynamics , using the yukawa formalism , limits on the parameter alpha are given as a function of the yukawa range , lambda .
first we address an ideal case , where the parameters of both dfgs are known exactly . as an example , let s assume the first dfg , with parameters described above , positioned at @xmath78 from the center of mass of the ifo tm .
specifications and position for the second dfg are determined by eq .
( [ etarelations ] ) . on fig .
we plot the interaction range for which the tm displacement at @xmath62 is maximal , @xmath79 , as a function of @xmath80 ( continuous line ) . on the same plot
, the dashed line shows the corresponding tm rms displacement at @xmath81 due to the yukawa term . using the above case and @xmath82 , fig .
gives an interaction range of @xmath83 , corresponding to a maximum displacement of the tm due to the yukawa term with an rms value of @xmath84 .
for the case of the aei 10 m detector with a noise floor of @xmath85 at @xmath86 , and an integration time of @xmath87 day , a limit of @xmath88 can be provided with a signal to noise ratio ( defined as the ratio of the rms signal to the displacement noise spectrum density integrated for a time @xmath89 ) of @xmath90 , given the basic physical limitations described later in this section .
in general terms , for an arbitrary noise floor @xmath91 , and integration time @xmath89 , the signal to noise ratio scales as @xmath92 expressed in @xmath59 units , where @xmath93 denotes the yukawa interaction range , for which the @xmath62 rms displacement due to the yukawa term is maximal . the corresponding rms displacement of the tm ( using @xmath94 , @xmath95 , @xmath96 , and @xmath97 ) , expressed in @xmath98 units , is shown by the dashed curve.,title="fig:",width=326 ] + there are two fundamental classes of effects that limit our capability of measuring an inverse - square - law - violating gravitational force within the experimental concept .
the first one is due to the noise level of the ifo , which effect can be suppressed by using sufficiently long integration times , as discussed in sec .
[ yuk ] . in this section
we discuss some of the effects from the second class , which are due to the finite measuring and manufacturing precision when setting up the null - measurement .
errors due to conceivable machining precision and procedure , uncertainties related to measurement of parameters of the final dfg products as well as their final placement relative to the tm and to each other will all infer an uncertainty in the induced displacement measurement , and propagate an error into the measurement of the non - newtonian parameter @xmath14 . in case of purely symmetric dfgs positioned in a null experiment configuration , and taking into account only the newtonian field component ,
the induced displacement of the tm will be dominated by a @xmath99 term . in the presence of asymmetries ,
the displacement will also have components at the odd harmonics and the classical terms at @xmath62 will not cancel each other completely .
this means that the newtonian peak due to this imperfect cancellation will appear at @xmath100 , and will contaminate a potential yukawa - signal from the very beginning of the measurement . first , we calculate the achievable lower limits on the yukawa strength parameter , @xmath101 , in case of ideal dfgs with no parameter errors .
we chose the parameters of the first dfg to be @xmath102 , @xmath103 , and @xmath104 , and maximized the integration time of the measurement at @xmath105 s ( @xmath106 months ) . because of the slight dependence of the tm response on @xmath80 ( see fig .
[ eta ] . ) , and taking into account manufactural limitations , we set @xmath80 to 2 .
we chose a conservative limit of @xmath107 as the condition for detection .
the optimal dfg frequencies that maximize the snr in terms of @xmath101 for the different detectors ( see fig . [ senscurves ] ) , along with the lowest achievable limits on the strength parameter , @xmath108 , are given in table [ classicerrors ] .
results on @xmath101 as a function of the yukawa scale parameter , @xmath15 , for five different ifos are provided in fig .
note , that these results are also valid for the case , when the measurement setup errors are kept low enough to allow the detector noise levels to be the main limiting factors .
[ classicerrors ] in order to study the feasibility and real - life limitations of null - experiment concept , we numerically studied the effect of uncertainties associated with the system parameters included in the simulations ( rotation frequency ( @xmath25 ) , mass ( @xmath11 ) and radius ( @xmath13 ) of each dfg , as well as the dfg positions relative to the tm and to each other ( @xmath26 and @xmath109 ) ) , thereby mimicking a realistic construction and measurement procedure . to do this , a large number ( @xmath110 ) of hypothetical setups were monte carlo simulated with the dfg parameters normally distributed around a mean .
the mean of the parameter distributions ( e.g. @xmath111 kg , @xmath112 and @xmath113 = 0.2 m ) are chosen such as to maximize the response of the ifo in terms of @xmath101 , while taking into account the technically achievable range of values for each parameter , as well as safety issues .
we again set @xmath80 to @xmath114 for all setups , similarly to the ideal case .
following eq .
( [ etarelations ] ) , @xmath80 in this case determines the means of the distributions corresponding to parameters of the second dfg . the means of dfg operational frequencies were chosen to be equal to the values given in table [ classicerrors ] .
the low optimal dfg frequencies allow the effective masses of dfg fillings for the first dfg ( rotating closer to the ifos vacuum chamber ) to @xmath115 .
the highest corresponding mechanical energy of a dfg ( @xmath116 ; @xmath117 ; @xmath118 ) would still be less than @xmath119 of the energy proposed in a similar experiment described in @xcite .
note , that as an alternative to the dfg geometry we proposed in sec .
[ yuk ] , a rotating mass design similar to the one suggested by @xcite can also be used in the measurement .
naturally , we are seeking to achieve the highest precision measurement currently possible or feasible in the near future , and we let the parameter uncertainties vary to values corresponding to state - of - the - art limits with the investment of plausible amount of work and finances .
our goal is to find , using monte carlo simulations , the limits on @xmath101 , allowed by using a two - dfg configuration with parameter uncertainties .
each parameter associated to hypothetical setups were randomized independently , from a gaussian distribution , using the same variances for corresponding parameters of different setups .
we used a chosen set of variances for the parameter distributions ( discussed in sec .
[ feasibility ] ) that are within the limits of current state - of - the - art machining and measurement technologies , and at the same time , minimize the newtonian peak we get at @xmath120 due to imperfect cancellation .
we generated @xmath121 two - dfg setups , and determined the 95% quantile of the induced newtonian displacement values at @xmath120 ( from now on , denoted by @xmath122 ) .
note that this quantile @xmath122 is completely independent from the detector noise and thus from the measurement integration time .
we then , for each @xmath15 , calculated the @xmath101 value for which the resulting yukawa peak at @xmath120 equals @xmath122 .
the limits on @xmath101 defined this way , versus @xmath15 , are shown in fig .
[ perf ] . via laser interferometric method
proposed here .
the grey shadings show the parameter regions excluded by previous measurements @xcite that were sensitive to negative , positive , and absolute alpha values only .
the thick black line corresponds to the limit achievable with an ongoing experiment , denoted by uw - uci @xcite .
the red curves show the results for various ifos in case of infinite dfg machining precision , using an integration time of @xmath123 sec .
the ifo+pa curve gives the 95% confidence limits for all ifos but lcgt and aei 10 m , when allowing dfg parameter uncertainties given in sec .
[ feasibility ] .
the limiting factor in the measurement is the finite integration time when using lcgt and aei 10 m , and the finite machining precision in case of all the other ifos .
the coloured background areas correspond to different theoretical models presented in @xcite , that predict yukawa parameters within the areas .
our tolerable parameter uncertainties were chosen such as to give the best coverage of the parallelogram corresponding to model `` t '' , while taking into account the feasible parameter adjustment precision.,title="fig:",width=326 ] + using a two - dfg setup with allowed parameter uncertainties given in sec .
[ feasibility ] , the limiting factor in case of the lcgt and aei 10 m ifo is its relatively high noise level at @xmath120 .
however this means , that we can tolerate higher parameter uncertainties compared to the ones given in sec .
[ feasibility ] , when using lcgt and aei 10 m and an integration time of 4 months . in case of the other three ifos ( aligo , avirgo , and et ) the limiting factor is the imperfect cancellation of the newtonian peaks at @xmath120 due to parameter errors , when considering a 4 months long integration time . therefore in order to reach the best possible @xmath101 limits for the chosen set of parameter uncertainties
, we can use a reduced integration time of o(@xmath124 ) , depending on which ifo we choose .
the reach of the allowed exclusion region suggests that the displacement sensitivity of any dedicated ifo for the null - measurement should have comparable displacement sensitivity to advanced interferometric gravitational wave detectors .
custom configured interferometry could possibly have several advantages such as double dfg compatible suspensions , being able to reduce the distance between dfgs and the ifo test mass , and thus to obtain better measurement results .
monte carlo test results presented in sec .
[ yukawaunc ] were provided by using a chosen set of dfg parameter uncertainties , that gives the best coverage of yukawa parameter predictions given by the `` t '' model ( see @xcite for more details ) , and that are still within the feasible measuring and machining precision .
an optimization of tolerable dfg parameter uncertainties in terms of measuring an isl - violating gravitational force can be given with the fisher matrix method @xcite , however this analysis is beyond the scope of this proof - of - concept paper . in this section
we discuss the dfg parameter uncertainties that correspond to the ifo+pa curve given in fig .
note , that we assumed that the parameter uncertainties are kept at the same levels throughout the whole integration time of the yukawa measurement .
we also discuss the possible methods in measuring and machining that allow us to reach the chosen dfg parameter precision . in the conceptual experimental setup considered the two dfgs are rotating at @xmath125 11 hz , 20 hz , 5 hz , 26 hz , and 5 hz using aligo , avirgo , et , lcgt , and aei 10 m , respectively . in all cases the two - dfgs in a setup have to be in a relative phase of @xmath126 . in our simulations we chose the following measurement uncertainties : * uncertainty in the frequency of the first dfg ( absolute measurement ) : @xmath127 * uncertainty in the frequency of the second dfg , relative to the first dfg ( relative measurement ) : @xmath128 * uncertainty in the initial phase difference between the two dfgs ( absolute measurement ) :
@xmath129 frequencies and the relative phase between the different dfgs can be finetuned to their desired value by calibrating and locking the dfgs rotation signal with an accurate and low - phase - noise clock ( e.g. , good cesium or gps ) , also used for time - stamping the ifo data . a sophisticated dfg frequency ( and phase )
control system can be constructed allowing nanosecond timing precision . in our simulations we used tungsten filling masses of 30 kg for the first , and @xmath130 kg for the second dfg .
the masses had the following permissible uncertainties : * uncertainty in the exact mass of one of the tungsten cylinders in the first dfg ( absolute measurement ) : @xmath131 * uncertainty in the mass of any other dfg cylinders of the two - dfg setup ( relative measurement ) .
the mass of one of the tungsten cylinders built into the second dfg has to be @xmath80 times the mass of the one in the first dfg , where we proposed to use @xmath82 .
this way we can use a relative measurement , and thus , we have to adjust the mass of one of the cylinders in the second dfg to the sum of the masses in the first dfg .
we adjust the mass of the second tungsten cylinder in the second dfg to the mass of the first cylinder in the same dfg .
the mass of the second cylinder in the second dfg can be then finetuned to the mass of the first cylinder within the same dfg . the tolerable uncertainty of these relative measurements and mass adjustments : @xmath132 for the dfg mass measurements
, one can use scales such as the sartorius me415s scale @xcite , that can determine the absolute value of the masses with an uncertainty of @xmath133 kg with 410 g weighing capacity .
using state - of - the - art mass comparators allows us to reduce the relative measurement ( mass - to - mass comparison ) error to @xmath134 kilograms ( e.g @xcite ) . a complete tungsten cylinder can be built from an assembly of several smaller disks , allowing us to reduce uncertainties with precision scales having lower weighing capacity .
these smaller and lighter components of the tungsten fillings are also easier to be manufactured and handled .
a combination of precision absolute scales , mass comparators , and fine mass adjustment through abrasion leads to precise mass standards and also allows manufacturing of matched mass pairs beyond the load limit of the absolute scale .
arm lengths and dfg positions have to be set to definite values ( in our example case @xmath135 , @xmath136 , @xmath103 , and @xmath137 ) to achieve precise cancellation of the newtonian component .
the corresponding parameters have the following uncertainties : * uncertainty in the distance between the first dfg s rotation center and the tm s center of gravity ( @xmath59 absolute measurement ) : @xmath138 * uncertainty in the distance between the two dfgs relative to @xmath59 ( relative measurement ) : @xmath139 * uncertainty in the length of one arm in the first dfg ( @xmath140 absolute measurement ) : @xmath141 * uncertainty of in the equality of arms within one dfg ( @xmath142 ) , and the uncertainty in the length of one arm of the second dfg relative to @xmath143 ( @xmath144 ) ( relative measurements ) : @xmath145 distance - like quantities ( @xmath13-s and @xmath26-s ) can be measured using interferometric methods with accuracy significantly surpassing the wavelength of the laser light .
the exact technique of the measurement greatly depends on the dfg placement , accelerating and controlling configuration , and is beyond the scope of this article .
setting the arm lengths within one dfg to be equal with a precision of @xmath146 m is challenging , but possibly feasible .
the uncertainty of equivalence of arms within one dfg will be in the order of the equivalence of masses within the dfg , and the uncertainty of @xmath147 being zero
. the feasible precision of mass - to - mass comparison can be as low as @xmath148 kg @xcite . in case of a non - zero @xmath149
, the axis of rotation of a spinning dfg will subject a periodic driving force ( that is proportional to @xmath150 ) , causing the support of the axis to vibrate .
the amplitude of the vibration can be measured with high - precision interferometry , and the placement of the center of rotation can be adjusted until the vibration amplitude is zero within the measurement precision .
the adjustment can be carried out using much higher dfg frequencies than the values given in table [ classicerrors ] , allowing an amplification of dfg vibrations in case of a non - zero @xmath149 .
the non - newtonian and the classical gravitational potentials scale differently with @xmath26 , the distance between the tm and the center of rotation of a dfg . using the two dfgs positioned in a null experiment configuration , and measuring the tm displacement spectrum in case of different @xmath26-s , can also help us to determine measurement and setup inaccuracies other than errors of @xmath26 .
it is possible that periodically varying one or more of the tunable parameters ( e.g. , distance or phase ) will allow precise , @xmath151 tuning of the cancellation necessary for the null - measurement .
also , an additional effect can be exploited to estimate the dominating error in the two - dfg setup parameters . by applying the parameter uncertainty tolerances given in this section ,
the tm s induced displacement at @xmath152 due to imperfect cancellation of the newtonian term will be orders of magnitudes above the detector noise level and the yukawa term at @xmath152 , and thus will be detectable . by monitoring this signal at @xmath152 we can give a rough estimation on the dominating error in the parameters of the two - dfg setup .
we investigated the projected contributions of the different parameter uncertainties , relative to each other , to the uncertainty of an @xmath101 measurement . to do this
, we carried out a series of monte carlo simulations , where in each simulation , we set the relative uncertainty of a chosen parameter to @xmath18 , and kept the uncertainties of all the other parameters at zero level .
we simulated @xmath121 two - dfg setups , and calculated the relative error of @xmath101 obtained from the @xmath121 hypothetical measurements of @xmath101 . in order to characterize the relative dominance of the different parameters compared to each other
, we normalized the measured relative errors of @xmath101 with the highest such relative error value we obtained .
this way , our final results were basically independent from the @xmath18 relative uncertainty chosen for each individual setup parameters .
the resulting relative dominances of the different parameter uncertainties in the @xmath101 measurement are given in table [ alpharelativeerrors ] .
our results show that the dominant parameter - for which the allowed uncertainty , @xmath153 , should be kept as low as possible - is the distance between the two dfgs , relative to @xmath154 .
.the relative dominance ( rd ) of parameter uncertainties in terms of their contribution to the uncertainty of @xmath101 in a hypothetical measurement .
we carried out @xmath121 monte carlo trials for each uncertainties , setting the chosen uncertainty to a @xmath18 relative value , and while setting all the other uncertainties to zero .
the relative dominance for each parameter uncertainty was obtained by calculating the relative uncertainty of @xmath101 measured from the @xmath110 trials , and normalizing it by the largest such relative uncertainty value ( thus we get a relative dominance of 1 for the most dominant parameter uncertainty ) . [ cols="^,^,^,^,^,^",options="header " , ] [ alpharelativeerrors ] in this subsection we addressed the fundamental sources of measurement errors and the uncertainty limits for each parameter of interest .
we also pointed out , that uncertainty values for dfg parameters used in our proposed experimental setup are within the limits of current state - of - the - art machining and measurement technologies . for a more realistic application a full dynamic finite element simulation of the mass distributions
must be performed .
the expansion and stress factors of the dfgs under prolonged operating conditions must be modeled and simulated , then subsequently measured and taken into account .
practical limitations due to the complex mechanical and detector dynamics / geometry around the test mass and other second order error sources need to be investigated , via experiments and more sophisticated simulations as they might prevent us from reaching the full potential of dfg device based measurements .
for example , in practice material inhomogeneities can give rise to a substantial uncertainty that could be minimized by using ndt - rated materials @xcite and also reduced by different geometries .
alternative , more simplified dfg geometries different from the one proposed here ( e.g. rotating rods ) could also possibly lead to higher measuring and machining precision , and therefore should be studied carefully .
all these issues will be the scopes of future studies .
significant kinetic energy ( i.e. @xmath155 100 kj ) is stored in the dfg once it rotates , therefore crucial safety considerations must be addressed .
there are two major points of failure management to be concerned with .
( a. ) the vacuum chamber of the dfgs must be made strong enough to withstand the damage of an accidentally disintegrating disk .
this is the standard solution for high speed gyroscopes .
( b. ) for added security , the gap between the inner wall of the vacuum chamber and the outer edge of the rotating disk could be kept relatively _
small_. in the event of an incident where the dfg s material starts to yield or its angular acceleration is uncontrolled the disk will expand radially touching the sidewall and may slowly stop , potentially preempting some of the catastrophic failure modes . adding a heavy ring in contact with the inner lateral wall of the vacuum chamber ( similarly as described in @xcite )
would absorb a huge angular momentum of the dfg and dissipate its rotation energy via the friction against the lateral wall .
these conditions can be met using finite element analysis ( fea ) aided design , in - house destructive testing of sacrificial parts and relying only on the best base materials .
two dfgs in a null experiment setup in conjunction with an interferometric sensor allow for studies of composition independent non - newtonian contribution to the classical gravitational field in the meter scale .
simulation results presented in this paper indicate that by taking advantage of two matched dfgs and the relatively large bandwidth and high sensitivity of state - of - the - art interferometric sensors there is an exciting opportunity to explore @xmath14 below the current limit in the @xmath156 meter yukawa range addressing standing theoretical predictions . for the proof of concept study , we chose a conservative 2 meter distance for our device from the test mass of the interferometric detector and used the sensitivity of advanced and third - generation gravitational wave detectors as the realistic baseline .
we intend to note here , that putting the devices considerably closer to the test mass can yield orders of magnitude better limits in isl violation parameters through the characteristic length scale changes for the yukawa case .
there are many practical details that still need close attention when designing and manufacturing a practical device .
finite element analysis of the dfgs and subsequent experimental studies are necessary to completely understand the stresses the dfg is subjected to .
the dfgs need to be enclosed in separate vacuum chambers seated on seismic attenuation stage(s ) .
prototype design and test will be necessary to balance the disk and test vibration control .
direct gravitational coupling into the test mass suspension and seismic isolation system must be carefully designed , simulated and mitigated .
any structure supporting the test mass must be placed far away ( i.e. , tens of meters ) from the double dfg assembly .
other mostly practical problems , such as safety , can also be solved as was shown in past applications / experiments that have used rapidly rotating instruments ( for examples see references in sec . [ introduction ] ) .
it is also likely that dedicated and custom designed interferometric sensors will be necessary and the direct use of advanced interferometric gravitational wave detectors will not be ideal for technical reasons . in spite of the promising estimates one should not underestimate the technical difficulties imposed by ( 1 ) the required design precision , ( 2 ) the manufacturing of the dfgs and ( 3 ) the unknown , but experimentally possible to study , complex nature of the coupling routes of the dynamic gravity fields of the dfgs into the complicated mechanical structure of the ifo test masses
the authors are grateful for the support of the united states national science foundation under cooperative agreement phy-04 - 57528 and columbia university in the city of new york .
we acknowledge the support of the hungarian national office for research and technology ( nkth ) through the polanyi program ( grant no .
kfkt-2006 - 01 - 0012 ) .
we greatly appreciate the support of ligo collaboration .
we are indebted to many of our colleagues for frequent and fruitful discussion .
in particular we d like to thank m. landry , p. sutton , d. sigg , g. giordano , r. adhikari , v. sandberg , p. shawhan , r. desalvo , h. yamamoto , j. harms , c. matone and z. frei for their support and valuable comments on the manuscript . | we present an experimental opportunity for the future to measure possible violations to newton s @xmath0 law in the @xmath1 range using dynamic gravity field generators ( dfg ) and taking advantage of the exceptional sensitivity of modern interferometric techniques .
the placement of a dfg in proximity to one of the interferometer s suspended test masses generates a change in the local gravitational field that can be measured at a high signal to noise ratio .
the use of multiple dfgs in a null experiment configuration allows to test composition independent non - newtonian gravity significantly beyond the present limits .
advanced and third - generation gravitational - wave detectors are representing the state - of - the - art in interferometric distance measurement today , therefore we illustrate the method through their sensitivity to emphasize the possible scientific reach .
nevertheless , it is expected that due to the technical details of gravitational - wave detectors , dfgs shall likely require dedicated custom configured interferometry
. however , the sensitivity measure we derive is a solid baseline indicating that it is feasible to consider probing orders of magnitude into the pristine parameter well beyond the present experimental limits significantly cutting into the theoretical parameter space . |
the cooperative spontaneous emission of radiation ( superradiance ) from an ensemble was first introduced by dicke in 1954 @xcite , where an atomic ensemble exhibited enhanced coupling to a single electromagnetic mode .
superradiance was initially suggested for sample dimensions @xmath0 much smaller than the wavelength @xmath1 of the resonant transition @xcite .
however , the case in the opposite limit @xmath2 has also attracted extensive research @xcite because in quantum optics the sample is usually large compared to @xmath1 .
in such a large - size sample , the superradiance is difficult to happen according to the standard argument that the dipole - dipole interaction between atoms is too weak to build a macroscopic dipole moment . to have enhancement in the limit of large - size sample with @xmath2 , the quantum coherence and interference
must enter to play a role .
typical example is to employ the quantum coherence to enhance some nonlinear optical processes @xcite .
recently , our group observed an enhanced raman scattering effect by using an atomic spin wave , a coherence prepared between atomic ground - state sublevels @xcite . the subsequent theoretical analysis modeled the experiment where an early spontaneous raman scattering ( srs ) generated the atomic spin wave which causes enhancement of the raman light fields in the second raman scattering @xcite .
the flipped atoms act as seeds to the second raman process which amplifies the light fields all the way through the atomic ensembles . in this paper , based on our previous work @xcite
, we propose a scheme to enhance the raman scattering , termed as correlation - enhanced raman scattering ( cers ) . in the scheme
, a pump field leads to spontaneous emission of the stokes field , accompanying with the generation of atomic spin waves .
then this stokes field used as a seeded signal , with the pump field together , is subsequently input into atomic ensemble to generate a second stokes field .
a cers occurs due to the correlation of the seeding stokes field and atomic spin waves @xcite .
such a light - atom correlation is a new mechanism to enhance raman scattering , which is different from the idea of the so - called super raman scattering initiated by the atom - atom entangled state @xcite .
our article is organized as follows . in sec .
[ ramana ] , the general model involving spatial propagation and light - atom coupling in raman scattering is reviewed .
the correlation of the light field and the atomic spin wave is derived from the light - atom coupling equations .
based on the model , the cers is studied in detail in sec .
[ ramanb ] . in sec .
[ analysis ] , we numerically calculate the intensity of cers based on the correlation of the seeded field and the atomic spin wave . finally , we conclude with a summary of our results .
in this section , we give a brief review of the theoretical model of raman scattering . the raman scattering process in atomic system
is described through a three - level model [ see fig.[fig1](a ) ] coupled to a pump field and a stokes field . for convenience ,
assume the pump field corresponding to a focused beam and the fresnel number @xmath3 ( @xmath4 cross - sectional area , @xmath5 cell length ) is of the order of unity , then only a single transverse spatial mode contributes strongly to emission along the direction of propagation the pump field .
therefore , a simplified one - dimensional model can be enough to describe the raman scattering . in the case of large light detuning
, the atomic excited state can adiabatically be eliminated and one obtains the light - atom coupling equations governing the propagation of the quantized stokes field and the atomic ground - state spin excitations determined by the spin wave @xcite @xmath6 where @xmath7 the spin wave operator , @xmath8 the number of atoms , @xmath9 , @xmath10 the collective atomic operators , where @xmath11 ( @xmath12 ) is the transition operator of the @xmath13th atom between states @xmath14 and @xmath15 and a small and macroscopic volume containing @xmath16 @xmath17 atoms around position @xmath18 , and the commutation relation is @xmath19=l / n\delta(z - z^{\prime})(\delta_{\nu\alpha}% \tilde{\sigma}_{\mu\beta}(z , t)-\delta_{\mu\beta}\tilde{\sigma}_{\alpha\nu } ( z , t))$ ] , @xmath5 the length of the atomic medium .
@xmath20 is a slowly varying envelope operator of the stokes field @xmath21 , and @xmath4 is cross - section area , and its commutation relation @xmath22=l / c\delta(t - t^{\prime}% ) $ ] .@xmath23 the coupling coefficient between spin excitations and stokes field , @xmath24 the rabi frequency of pump field , @xmath25 the atom - field coupling constant , @xmath26 , @xmath27 , and @xmath28 is the optical pumping rate , and @xmath29 is the ac stark shift , and @xmath30 the coherence ( @xmath31 ) decay rate , @xmath32 the decay rates of the excited state @xmath33 to states @xmath34 and @xmath35 ( assuming @xmath36 ) .
@xmath37 describes the population difference between energy levels @xmath34 and @xmath35 . in general , @xmath38 is related to the collective atomic excitation number and the strength of atomic coherence .
here we consider weak excitations , and @xmath38 is approximately determined by eq .
( [ eq3 ] ) . the langevin noise operator @xmath39 has the correlation @xmath40 . for convenience in analysis ,
neglecting the depletion of the pump field by making @xmath41 with @xmath42 being constant and @xmath43 the step function .
similarly , @xmath44 . using the moving coordinates @xmath45 , @xmath46 , the solutions of eqs .
( [ eq1])-([eq3 ] ) are given for the raman scattering @xcite , @xmath47}dt^{\prime\prime},\nonumber\\ & \label{solu3}%\end{aligned}\ ] ] where@xmath48}\frac{\mathcal{w}(t^{\prime\prime})}{c}g_{e}(z^{\prime } , 0,t^{\prime},t^{\prime\prime})]\{ ... \},\nonumber\\ & \mathcal{v}_{s}(z^{\prime},t^{\prime})\{ ... \}=i\int_{0}^{z^{\prime}% } dz^{\prime\prime}\frac{\chi(t^{\prime})}{c}e^{-\gamma(t^{\prime})}% h(z^{\prime},z^{\prime\prime},t^{\prime},0)\{ ... \},\nonumber\end{aligned}\]]@xmath49\{ ... \},\nonumber\\ & \mathcal{v}_{a}(z^{\prime},t^{\prime})\{ ... \}=-i\int_{0}^{t^{\prime}% } dt^{\prime\prime}\mathcal{w}(t^{\prime\prime})\chi^{\ast}(t^{\prime\prime } ) e^{-[\gamma(t^{\prime})-\gamma(t^{\prime\prime})]}\nonumber\\ & \text { \ \ \ \ \ \ \ \ \ \ } \times h(z^{\prime},0,t^{\prime},t^{\prime \prime})\{ ... \},\nonumber\\ & \mathcal{f}_{s}(z^{\prime},t^{\prime})\{ ... \}=i\int_{0}^{t^{\prime}% } dt^{\prime\prime}\int_{0}^{z^{\prime}}dz^{\prime\prime}\frac{\chi(t^{\prime } ) } { c}e^{-[\gamma(t^{\prime})-\gamma(t^{\prime\prime})]}\nonumber\\ & \text { \ \ \ \ \ \ \ \ \ \ } \times h(z^{\prime},z^{\prime\prime},t^{\prime } , t^{\prime\prime})\{ ... \},\nonumber\\ & \mathcal{f}_{a}(z^{\prime},t^{\prime})\{ ... \}=\int_{0}^{t^{\prime}% } dt^{\prime\prime}\int_{0}^{z^{\prime}}dz^{\prime\prime}e^{-[\gamma(t^{\prime } ) -\gamma(t^{\prime\prime})]}\nonumber\\ & \text { \ \ \ \ \ \ \ \ \ \ } \times\lbrack2\delta(z^{\prime}-z^{\prime \prime})+g_{s}(z^{\prime},z^{\prime\prime},t^{\prime},t^{\prime\prime } ) ] \{ ... \},\end{aligned}\ ] ] in which @xmath50\frac{z^{\prime}-z^{\prime\prime}% } { c}}),\nonumber\\ & g_{e}(z^{\prime},z^{\prime\prime},t^{\prime},t^{\prime\prime}% ) = \frac{c(z^{\prime}-z^{\prime\prime})}{q(t^{\prime})-q(t^{\prime\prime}% ) } g_{s}(z^{\prime},z^{\prime\prime},t^{\prime},t^{\prime\prime}),\nonumber\\ & g_{s}(z^{\prime},z^{\prime\prime},t^{\prime},t^{\prime\prime})=\sqrt { \frac{q(t^{\prime})-q(t^{\prime\prime})}{c(z^{\prime}-z^{\prime\prime})}% } \nonumber\\ & \times i_{1}(2\sqrt{[q(t^{\prime})-q(t^{\prime\prime})]\frac{z^{\prime } -z^{\prime\prime}}{c}}).\end{aligned}\ ] ] here @xmath51 , @xmath52 , @xmath53 , and @xmath54 is the modified bessel function of the first kind of order @xmath55 . the integral solutions presented in eqs .
( [ solu1])-([solu2 ] ) indicate that the stokes field @xmath56 contains three parts of contributions , the first one from the input field plus the scattering field , the second from the initial spin excitations , and the third from the atomic fluctuations .
similarly , the generated spin wave @xmath57 contains the contributions from the initial spin excitations , the input field , and the atomic fluctuations .
evidently a light - atom correlation of the stokes field and the generated spin wave , @xmath58 is built through the raman coupling as seen in _ _ eqs .
( [ solu1])-([solu2 ] ) .
is generated by the pump @xmath59 with the initial seeding state ( c ) .
@xmath60 : pump fields ; @xmath61 and @xmath62 : the generated stokes fields and spin waves , respectively ; @xmath63 : the detuning . ]
in the section , we study the role of the correlation of the light field and atomic spin wave in enhancing raman scattering .
the basic idea is as follows .
first we employ a srs process to generate a light field and an atomic spin wave which are correlated with each other .
the srs process is illustrated in figs .
[ fig1](a ) and ( b ) , where an atomic ensemble , initially prepared in the ground state , is pumped by an off - resonance pump field @xmath64 and spontaneously emitted a stokes field @xmath65 . as a result , a coherent spin excitation @xmath66 is built in the atomic ensemble .
based on the analysis given in the above section , the emitted stokes field @xmath65 is correlated with the excited spin wave @xmath67 next , we consider a new raman scattering process , termed as cers , where the stokes field @xmath65 , with its correlated atomic spin wave @xmath66 , is employed as initial seeding as shown in fig . [ fig1](c ) . again
an off - resonance pump field @xmath68 is injected to generate a new stokes emission @xmath69 , as illustrated in fig . [
fig1](d ) .
the experiment to study the cers can simply be setup as in fig .
[ fig2](b ) . an atomic cell , for example
, rubidium vapor cell is used to provide the atomic ensemble .
the generated stokes field @xmath65 and the leaked pump @xmath64 in the srs , experiencing phase shifts @xmath70 and @xmath71 respectively , are injected back into the atomic cell .
the light - atom correlation are used as the initial seeding and the the leaked pump acting as @xmath68 for the cers .
the output intensity of the stokes field @xmath69 is detected for analysis .
generates the stokes field @xmath72 and the atomic spin wave @xmath66 through spontaneous emission , and ( b ) cers : after a delay @xmath73 , the leaked pump @xmath74 , acting as @xmath59 , is again injected into the vapor cell with the srs - generated stokes field @xmath75 .
the srs - generated correlation of the light @xmath72 and the spin wave @xmath66 enhances the intensity of the second raman stokes field @xmath76 .
pbs : polarization beam splitter ; @xmath71 , @xmath77 : phase shifts .
( c ) time sequence for the pump pulses .
@xmath73 : delay time . ] using the relations given in eqs .
( [ solu1])-([solu3 ] ) , we can compare the differences between the srs process and cers process by working out their output intensity in detail . in the case of the srs , the stokes field
is generated from the vacuum and the atomic ground state without light - atom correlation .
hence the initial conditions for the srs are@xmath78 @xmath79 where the subscript @xmath80 denotes the srs process .
the detected intensity of the srs output stokes field at the end of the atomic cell @xmath81 is given by @xmath82 the result reflects the fact that the stokes field is generated from the spontaneous emission .
however , in the case of the cers , one has a completely different initial condition as follows:@xmath83 where the subscript @xmath84 denotes the cers process , @xmath85 indicates the opposite propagation of the two pump fields , and @xmath86 is the pulse duration of the pump @xmath64 .
evidently from eq .
( [ srs ] ) _ , _ the light - atom correlation will play an important role in the cers due to the relations @xmath87 in general , the spin wave @xmath88 decays due to collisional dephasing by a factor @xmath89 with @xmath73 being the delay time . here for convenience , we neglect the slow decay of the spin wave by assuming a short delay time ( @xmath90 ) . versus the dimensionless time @xmath91 : srs case ( dotted line ) , seeded - light only case ( dashed line ) , spin - wave only case ( dotted - dashed line ) , light - atom uncorrelated case ( thin solid line ) , and light - atom correlated case ( thick solid line ) .
the parameters are as follows : @xmath92 , @xmath93 , @xmath94 ghz , @xmath95 hz , @xmath96 , @xmath97 khz , and @xmath98 mhz . ]
the detected intensity of the cers output stokes field can be worked out @xmath99 where @xmath100 the term @xmath101 originates from the same mechanism as in the srs .
the injected seeding field and the initial spin wave contribute the intensity terms @xmath102 and @xmath103 .
the light - atom correlation leads to an interference term @xmath104 , which is sensitive to the phase difference @xmath105 , experienced by the pump and the stokes field in cers .
the cers much depends on the initially - set conditions . a different setting mechanism of initial conditions results in different enhancement . for the case discussed here ,
this can be summarized as follows . when only seeded light ( spin wave ) exists , the corresponding enhancement @xmath102 ( @xmath106 occurs
. then the output intensity is @xmath107 the spin wave ( atomic coherence ) induced enhancement mechanism was studied by both theory and experiment @xcite . when the uncorrelated light field and the spin wave are used as the initial input seeding ,
the enhancement is independently contributed by the light field and the spin wave , respectively . as a result , the output intensity of the stokes field is a simple sum given as @xmath108 the light - atom correlation is a new mechanism which leads to the phase - sensitive interference in the raman scattering .
it is well - known that the constructive interference can greatly enhance the light intensity . in this sense ,
the cers occurs optimally for an appropriate phase difference determined by @xmath109 @xmath110 , where @xmath111 is an overall phase shift offset by the seeded light , initial spin wave and the raman scattering process itself .
in addition , we point out that the light - atom correlation also leads to a maximally reduced raman scattering when the phase difference satisfies the condition @xmath112 @xmath110 .
in this section , we numerically calculate the intensities for different mechanisms of enhancement . for convenience , we define the dimensionless time @xmath113 . according to the different scattering mechanisms , the stokes intensities of different cases are plotted . versus the relative phase @xmath114 .
the other parameters are the same as those of fig .
[ fig3 ] except the dimensionless time at @xmath115 . ] in fig .
[ fig3 ] the intensity determined by the srs mechanism @xmath116 is plotted as the dotted line .
the contributions from the seeded light @xmath102 and the spin wave @xmath103 are shown in the dashed and the dotted - dashed lines , respectively .
the enhanced total intensity without the light - atom correlation @xmath117 is represented by the thin solid line .
the optimal cers output intensity @xmath118 is evidently increased due to the correlation - induced interference for the special phase difference @xmath119 , chosen in the calculation , where the result is displayed as a thick solid line . furthermore , we analyze the phase sensitivity of the cers . in fig .
[ fig4 ] , we plot the intensity @xmath118 as a function of the phase difference @xmath114 .
this figure shows that the intensity is modulated by the relative phase @xmath114 .
the modulation is a reflection of the correlation - induced interference .
the visibility of the interference fringes is given by @xmath120/[(i_{\text{out}}^{\text{cers}})_{\max}+(i_{\text{out}% } ^{\text{cers}})_{\min}]$ ] .
in the cers , the high visibility can be achieved by controlling the seeded light and initial spin wave to realize the case @xmath121 .
a direct application of the phase - sensitive raman scattering proposed here is to realize a nonlinear interferometer @xcite , which can benefit the improvement of phase estimation for optical metrology , imaging , and information processing @xcite .
in conclusion , we have theoretically studied the correlation - induced phase - sensitive raman scattering , which is based on the light - atom correlation through the coupling of the light field with the atomic spin excitations in the raman scattering .
we analyze the phase sensitivity of such a raman scattering process .
an optimally enhanced raman scattering occurs when the accumulated phase difference determined by the pump , the seeded light , and the raman process is appropriate .
similarly , one can also have a reduced raman scattering by tuning the phase difference .
such a correlation - induced phase - sensitive raman scattering process may find applications in a diversity of technological areas such as optical detection , metrology , imaging , precision spectroscopy , and so on .
this work was supported by the national basic research program of china ( 973 program ) under grant no .
2011cb921604 and no . 11234003 ( w.z . ) , the national natural science foundation of china under grant no .
11129402 ( z.y.o . ) , no .
11004058 , no . 11274118 , and supported by innovation program of shanghai municipal education commission 13zz036 ( l.q.c . ) , the national natural science foundation of china under grant no .
11004059 ( c.h.y . ) .
email : @[email protected] | we theoretically propose a method to enhance raman scattering by injecting a seeded light field which is correlated with the initially prepared atomic spin wave .
such a light - atom correlation leads to an interference in the raman scattering .
the interference is sensitive to the relative phase between the seeded light field and initially prepared atomic spin wave . for constructive interference ,
the raman scattering is greatly enhanced .
such an enhanced raman scattering may find applications in quantum information , nonlinear optics and optical metrology due to its simplicity . |
the atmospheres of cool stars are governed by their magnetic fields .
phenomena ranging from flares to coronal mass ejections ( cmes ) to stellar winds are all intimately related to the atmospheric magnetic field .
the best case study we have for a stellar magnetic field is the sun .
the solar magnetic field has been studied in detail over many scales and cycles .
we can not , however , observe the atmospheric ( coronal ) magnetic field directly and so must rely upon theoretical models . as input to these models , information about the magnetic field at the sun s surface ( photosphere ) can be determined through the zeeman effect @xcite .
maps of the line - of - sight magnetic field ( magnetograms ) have been available for many years .
more recently , vector magnetograms of the sun have also been incorporated into coronal models ( e.g. * ? ? ?
* ; * ? ? ?
* ) . on stars other than the sun
, we now have insight into magnetic field structures .
we can use zeeman doppler imaging ( zdi ) to determine stellar magnetograms ( e.g. * ? ? ?
such magnetograms convey only the large - scale structure of a star s magnetic field .
fields on the active region scale and smaller are not observable currently .
much important information about the stellar enviroment , however , can be gained from a knowledge of the large - scale structure of the global magnetic field .
two approaches have been used mainly to study the magnetic coronae of stars ( henceforth ` stars ' refers to stars other than the sun ) .
the first is a potential field extrapolation ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the potential field approach is useful as it gives a unique solution that minimizes the magnetic energy for a given set of boundary conditions .
however , as there is no current ( @xmath1 ) in a potential field , the magnetic field in the corona does not contain twist , which is considered important for the onset of different forms of eruptive behaviour ( e.g. * ? ? ?
* ) . also , the lorentz force completely decouples from the non - magnetic terms in the magnetohydrodynamic ( mhd ) momentum equation .
this decoupling means that certain assumptions need to be made when finding proxies for quantities such as x - ray emission ( e.g. * ? ? ?
the second approach is to perform a full mhd simulation of the global corona ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , where it is possible to capture the large - scale dynamics of the corona .
however , this approach is computationally very expensive and often requires super - computing resources . in this paper
, we introduce a technique that can be thought of as an intermediary step between the potential field model and full mhd simulations , as mentioned above .
we present a magnetohydrostatic ( mhs ) solution which couples the coronal density , pressure and magnetic field together .
perturbations in the hydrostatic density and pressure , due to the magnetic field , can be calculated self - consistently .
as the solution is static , like the potential field model , it is very simple to implement .
the model allows us to calculate , with relative ease , a twisted stellar coronal field where magnetic forces balance hydrostatic forces .
the approach presented in this paper has been applied to the solar corona where large - scale features have been captured successfully @xcite
. our approach will be particularly suited to stellar applications , where magnetograms only convey the large - scale structure of the magnetic field .
this type of model can capture the large - scale field structure ( as evidenced by the solar applications mentioned above ) and improve upon the potential field extrapolations with the inclusion of current density and the coupling of the magnetic field to the background stellar atmosphere . in the following section we give a detailed derivation of the coronal magnetic field and the associated density and pressure perturbations ( correcting typographical errors that appear in previous works ) .
we follow this section with an analysis of how the resulting free parameters affect the geometry and topology of the model coronal magnetic field . after the analysis , an application of the model
is given where a magnetic map of the m - dwarf gj 182 , published by @xcite , is used as input .
the paper ends with a summary .
the mhs equations that describe the balance between the lorentz and hydrostatic forces can be written as @xmath2 @xmath3 @xmath4 here , @xmath5 is the magnetic induction ( referred to as the magnetic field ) , @xmath6 is the curent density , @xmath7 is the plasma presure , @xmath8 is the plasma density and @xmath9 is gravity .
a discussion of appropriate boundary conditions will be left until later . rather than proceeding with equations ( [ momentum ] ) to ( [ divb ] ) in dimensional form
, we write them in dimensionless form by introducing the variables @xmath10 the distance vector is labelled @xmath11 in anticipation of the use of spherical coordinates for stellar applications .
dropping the asterisks , the dimensionless mhs equations are @xmath12 @xmath13 @xmath14 all further expressions will be in dimensionless form unless stated otherwise .
here we have written gravity in terms of its potential @xmath15 , viz .
@xmath16 to make progress , we model the current density as the composition of two distinct terms : @xmath17 this choice of the current density enables us to model the magnetic field through a range of plasma @xmath18 . since equation ( [ current ] ) is linear , we can make analytical progress and take a practical intermediate step between potential field extrapolations and expensive global mhd simulations .
@xmath19 is a constant parameter and the first term on the rhs of equation ( [ current ] ) represents the current density of a linear force - free field . the second term on the rhs represents the current density that produces a lorentz force that can balance the effects of the pressure gradient and gravity .
@xmath20 is a free function which will be specified shortly .
the approach of using this type of term was introduced by @xcite and has been used in several subsequent studies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
@xcite also present a non - potential magnetic field model , in the stellar physics context , with a term similar to the second term on the rhs of equation ( [ current ] ) .
however , they do not consider mhs equilibria .
the combination of the two terms in equation ( [ current ] ) results in components of current density that are both parallel and perpendicular to the magnetic field .
therefore , the total field - aligned current density , if written in the form @xmath21 , has a non - uniform @xmath22 .
although the field in our model can be ( depending on the choice of parameters discussed below ) more complicated than a linear force - free field , our model has the advantage of taking the same input from magnetograms as linear force - free models . as potential field models are a particular case of linear force - free models @xmath23 , potential models can be updated to the mhs model easily .
the following derivation of the model is based on @xcite and @xcite .
an alternative method exploiting the toroidal and poloidal components of the field can also be used @xcite .
we shall utilize the toroidal - poloidal approach later when studying the topology of the field . inserting equation ( [ current ] ) into ampre s law , equation ( [ ampere ] ) , gives , after some manipulation , @xmath24 where @xmath11 is the radial vector .
further , we choose the form of @xmath20 to be @xmath25 where @xmath26 is a free function
. the particular form of @xmath26 will be chosen later to allow for an analytical solution . at this point
it is useful to introduce the angular momentum operator @xmath27 this operator is common in quantum mechanics and a description of its properties can be found in @xcite . by simple manipulation , it can be shown that @xmath28 combining equations ( [ rdot ] ) and ( [ tmp1 ] ) , and making use of equations ( [ fdefine ] ) and ( [ l1 ] ) , gives @xmath29 to proceed , we expand @xmath30 in spherical harmonics , @xmath31 where @xmath32 @xmath33 are constant ( complex ) coefficients .
to determine the @xmath34 , inserting equation ( [ rdotb ] ) into equation ( [ vec_schrod ] ) gives @xmath35g_l^{(j)}=0.\ ] ] where @xmath36 . in deriving equation ( [ g ] ) ,
use has been made of the property @xmath37 to solve equation ( [ g ] ) analytically , a particular form must be chosen for @xmath26 .
notice that equation ( [ g ] ) has the form of a 1d schrdinger equation with @xmath26 playing the role of the potential ( e.g. * ? ? ?
this analogy gives us a starting point for what form @xmath26 can take in order to yield analytical solutions .
several possible analytical solutions are discussed in @xcite .
following @xcite , we adopt @xmath38 where @xmath39 is a free parameter .
this particular form represents the contribution of the non - force - free part of the current density decaying with height , as in the low solar corona ( e.g. * ? ? ?
by considering the following variable transformation : @xmath40 equation ( [ g ] ) transforms into bessel s equation @xmath41f^{(j)}_l = 0.\ ] ] converting back to the original variables , the solutions can be written as @xmath42 where @xmath43 and @xmath44 are bessel functions of the first and second kind respectively . before writing down an expression for the magnetic field
, we require the @xmath33 which are determined using the boundary conditions and an orthogonality constraint .
as mentioned earlier , the solutions of the mhs equations ( [ momentum])-([divb ] ) require boundary conditions . in this problem ,
the required boundary conditions are the radial components of the field at the stellar surface and a source surface . at the stellar surface ,
the radial magnetic field component is provided by a magnetogram .
the modeller is free to select an appropriate form for the source surface and choose the distance of the source surface from the stellar surface . in solar applications
, in - situ satellite observations can constrain the form of the source surface .
no such constraints exist for stellar coronae , although x - ray observations can be used to provide loose constraints ( e.g. * ? ? ?
* ) . for mhs models ,
the effects of outflows ( e.g. stellar winds ) are incorporated into the form of the source surface .
at the stellar surface @xmath45 ( we have taken @xmath46 ) the radial component of the field , @xmath47 , can be written as @xmath48 where @xmath49 note that if the magnetogram comes from observations , it is important to ensure that the magnetic flux is balanced . using the orthogonality condition @xmath50 we find @xmath51 here , the overbar represents the complex conjugate . equation ( [ q1 ] ) is a linear equation for @xmath52 and @xmath53 . applying the same procedure at the source surface , or imposing another constraint , results in another linear equation and this system
can be solved to find the coefficients @xmath54 . to find a closed - form solution for the magnetic field ,
we take the dot product of @xmath11 and equation ( [ ampere_nd ] ) and apply the angular momentum operator , @xmath55 from the definition of @xmath56 in equation ( [ l1 ] ) , equation ( [ lb ] ) contains only derivatives of @xmath57 and @xmath58 .
hence , we can write @xmath59 where @xmath60 .
we require a series expansion of @xmath61 to match the rhs of equation ( [ lb ] ) . following jackson ( 1975 ) and neukirch ( 1995 ) , we find @xmath62 the unknown functions @xmath63 and @xmath64 can be found by two simple considerations . for brevity
, we shall consider a single order for @xmath65 and @xmath66 .
inserting equation ( [ bt ] ) into equation ( [ lb ] ) and making use of equation ( [ l ] ) results in @xmath67 on the application of equation ( [ divb_nd ] ) , and some simple rearrangement , we find @xmath68 the expression for the global coronal magnetic field is then given by @xmath69\end{aligned}\ ] ] the spherical components of equation ( [ globalb ] ) are listed in appendix [ appendix ] .
we now consider what forms the pressure and density take in the mhs model .
inserting the model current density , given by equation ( [ current ] ) , into the force balance equation ( [ mom_nd ] ) gives , after some rearrangement , @xmath70 for a non - trivial current density , we require @xmath71 and @xmath11 to be linearly independent . hence , equation ( [ insertion ] ) implies that @xmath72 . considering the component in the direction @xmath71
, we obtain @xmath73 integrating equation ( [ pf ] ) yields @xmath74 @xmath75 is the background pressure and is found from hydrostatic balance . the second term in equation ( [ pressure ] ) is due to the non - force - free component in the current density . by considering the @xmath11-component of equation ( [ insertion ] )
, the density is found to be @xmath76 here , @xmath77 is the background hydrostatic density profile .
in this section , we analyze the effects of the free parameters , @xmath19 and @xmath39 , on the geometry and topology of the coronal magnetic field .
these parameters determine the form of the two model current density terms given in equation ( [ current ] ) .
the main assumption of the mhs model is the form of the current density in equation ( [ current ] ) .
the first term that contributes to the current density , @xmath78 , represents the linear force - free part of the current that is parallel to the magnetic field .
this term is well known for adding twist to the magnetic field .
analogous to vorticity in a sheared flow , current density can manifest itself through a sheared magnetic field . to demonstrate this effect , we shall consider a dipolar field with very simple boundary conditions .
these are @xmath79 where @xmath80 is the stellar surface and @xmath81 is the source surface .
figure [ twist ] shows the coronal magnetic field with boundary conditions ( [ bcs ] ) and free parameters @xmath82 and @xmath83 .
the effect of @xmath84 can be seen clearly . moving from the stellar surface upwards , the alignment angle of the magnetic arcades changes .
this shearing of the magnetic field is evidence of non - zero current . ) .
the free parameters are @xmath82 and @xmath83 .
low magnetic arcades are aligned at different angles to those that rise to greater radial distances .
this shearing of the field is due to a non - zero @xmath19 .
the coloured stellar surface shows the non - dimensional @xmath85 component of the dipolar field in equation ( [ bcs ] ) . ] the second term in equation ( [ current ] ) is perpendicular to gravity ( the radial direction ) .
it is this term that couples the magnetic field to the plasma pressure and density .
the second term depends on both of the free parameters @xmath19 and @xmath39 . as mentioned above , increasing the magnitude of @xmath19 increases the twist ( positive and negative values determine the sense of the twist ) .
the effect of changing @xmath39 requires more investigation .
@xcite perform several numerical experiments and show that increasing the value of @xmath39 causes magnetic arcades to expand .
physically , this expansion occurs if there is an increase in magnetic pressure .
consider equation ( [ pressure ] ) re - written as @xmath86 where @xmath7 is the plasma pressure , @xmath87 is the magnetic pressure perturbation and @xmath75 is the background hydrostatic pressure of the stellar atmosphere . by considering the expression for the pressure perturbation ,
the sign of @xmath87 depends on the sign of @xmath26 . from equation ( [ xi ] )
, the sign depends , in turn , on the choice of @xmath39 .
@xmath87 also depends on @xmath39 through equations ( [ u1 ] ) and ( [ u2 ] ) . to give an illustration of how the the pressure perturbation depends on @xmath39
, we calculate the signed maximum perturbation , @xmath88 .
this quantity can be found analytically as @xmath89 using the fact that the maximum magnitude of the pressure perturbation occurs at the stellar surface where @xmath90 .
figure [ fig_pm ] shows a plot of equation ( [ maxmag ] ) . as a function of @xmath39 for fixed @xmath19=0.3 .
values on the axes are non - dimensional as specified in equation ( [ nondim_p ] ) . ] for @xmath91 , the values are positive , indicating an expansion of the magnetic field ( e.g. * ? ? ? * ) .
when @xmath92 , @xmath87 is also negative , resulting in compression of the magnetic field .
what this analysis shows is that the choice of @xmath26 is crucial for determining how the pressure perturbation will behave .
the pressure perturbation , in turn , will influence the inflation of the magnetic field .
magnetic helicity is a topological invariant of ideal mhd ( e.g. * ? ? ?
* ) and can be interpreted as the average pairwise linkage of magnetic field lines @xcite .
the magnetic helicity is defined by @xmath93 where @xmath94 and @xmath95 is the volume of concern .
this quantity is only gauge - invariant , however , if the magnetic field is closed , that is if @xmath96 on @xmath97 . for the magnetic fields described in this paper , this condition is not met as field leaves the stellar surface and either returns back down to this level ( closed field ) or connects to the source surface above ( open , stellar wind bearing , field ) .
@xcite show that a gauge - invariant measure of helicity is also possible for open magnetic fields .
this form of helicity is known as relative helicity and can be written as @xmath98 where @xmath99 here @xmath95 is the volume between the stellar surface and the source surface and @xmath100 and @xmath101 represent volumes below and above @xmath95 respectively .
the magnetic fields in these two volumes match the field of @xmath95 at its boundaries and close to allow for helicity to be a topological invariant .
@xcite show that the magnetic fields in @xmath100 and @xmath101 can be ignored and that helicity only requires knowledge of the field in @xmath95 .
the required magnetic fields are @xmath102 , the coronal magnetic field found from equation ( [ globalb ] ) , and @xmath103 , a reference magnetic field with the same boundary conditions as @xmath102 .
the reference field is normally taken to be a potential magnetic field for simplicity .
relative helicity can then be thought of as a quantitative measure of how ` different topologically ' @xmath102 is from @xmath103 .
if @xmath104 and @xmath105 are chosen to be potential , the relative magnetic helicity reduces to @xmath106 in a spherical geometry . in order to determine an expression for @xmath107 , we proceed by decomposing the magnetic field into poloidal @xmath108 and toroidal @xmath109 components @xmath110 the magnetic field in equation ( [ poltor ] )
can be written as @xmath111.\ ] ] the vector potential is found to be @xmath112,\ ] ] where @xmath113 is an arbitrary gauge function .
the relative helicity can be expressed in the appealing form @xmath114 which reflects the fact that helicity arises due to the linkage of the poloidal and toroidal magnetic field .
equation ( [ rel_hel ] ) ( in a slightly different form ) is derived in @xcite and so we do not repeat the derivation here .
@xcite then goes on to find an expression for the relative helicity of linear force - free fields .
we shall now proceed in a similar manner and find an expression for this class of mhs equilibria . inserting equation ( [ poltor ] ) into equation ( [ current ] )
, we can derive the equations @xmath115 which can be reduced to @xmath116 using equation ( [ poltor ] ) , we can show that @xmath117 applying @xmath118 to equation ( [ p ] ) and using the distributive property of @xmath118 , we obtain @xmath119 which is just equation ( [ vec_schrod ] ) .
hence , one can expand @xmath120 as @xmath121 following @xcite , after integration by parts , equation ( [ rel_hel ] ) with suitable substitutions becomes @xmath122 where @xmath123 is the stellar radius and @xmath124 is the radius of the source surface . to investigate the dependence of the relative helicity on the free parameters , we determine equation ( [ rhe ] ) for the dipolar field with boundary conditions specified in equation ( [ bcs ] ) .
figure [ hr_fig ] displays how @xmath107 varies as a function of @xmath39 for different values of @xmath19 . as a function of @xmath39 for different values of @xmath19 .
key : solid , @xmath82 ; dashed , @xmath125 ; dotted , @xmath126 ; dot - dashed , @xmath127 .
values on the axes are non - dimensional as specified in equation ( [ nondim_p ] ) . ]
the curves in figure [ hr_fig ] can be interpreted as how much the field is ` different topologically ' from a potential field with the same boundary conditions . for @xmath126 , @xmath128 , as shown by the dotted line in figure [ hr_fig ] .
if there is no twist injected into the field , through @xmath19 , then the field is topologically equivalent to a potential field and so @xmath129 . for @xmath130 ,
the relative helicity increases monotonically with @xmath39 . for @xmath131
, @xmath132 also increases monotonically but now the sign of the relative helicity is negative . in the range of the parameter space displayed in figure [ hr_fig ] , the effect of increasing both @xmath39 and @xmath19 is to increase the magnitude of the relative helicity .
this result occurs because increasing the free parameters @xmath39 and @xmath19 increases the current density which , in turn , moves the field further from a potential state . for much higher values of @xmath19
, resonances can occur where the magnetic helicity and magnetic energy go to infinity @xcite .
these large values , however , are not applicable to our model which is concerned with large - scale stellar fields that are not twisted to such extents ( e.g. * ? ? ?
to illustrate the model , we apply it to the young single m - dwarf gj 182 ( @xmath133 , @xmath134 , @xmath135 ; @xcite ) .
this star shows evidence for surface differential rotation , with the equatorial regions rotating faster than the poles .
the magnetic field is dominantly toroidal with a mostly non - axisymmetric poloidal component @xcite , consistent with a partially convective interior @xcite . as input to the model
, we take gj 182 s magnetogram , the radial component of the magnetic field at the stellar surface . the magnetogram is shown in figure [ br_fig ] where the field is normalized with respect to its largest magnitude , 355 g. this star has been chosen to demonstrate the model because the star s magnetic field differs from the large - scale symmetric dipolar structure of the solar magnetic field .
also , gj 182 s coronal magnetic field has been modelled by @xcite using a potential field extrapolation .
this previous study allows us to make a comparison and highlight the extra features that appear in our model .
has been normalized with respect to its largest magnitude ( 355 g ) . ] to complete the model , an outer boundary condition , the source surface , is required . as mentioned previously ,
the radial position and form of the source surface are modelling choices . for simplicity , we take the dipolar radial field @xmath136 where @xmath137 . with these radial boundary conditions , the magnetic field in the region @xmath138 can be calculated using equation ( [ globalb ] ) .
we take @xmath139 , which is enough to capture the features of the magnetogram . in the last section
, we discussed some of the effects of the free parameters . for applying the model to gj 182 ,
we choose @xmath140 and @xmath141 .
this choice lies in the part of the parameter space where the magnetic pressure is increased , resulting in a pressure deficit due to the magnetic field .
a visualization of the coronal field is shown in figure [ fieldline_fig ] where several field lines are plotted and the sphere shows the map of @xmath85 given in figure [ br_fig ] . as mentioned above ,
@xcite plot a potential field extrapolation of the corona of gj 182 .
the orientation of the star in figure [ fieldline_fig ] is chosen to be similar to that shown in @xcite for ease of comparison . and @xmath141
. the contour map shows @xmath85 .
sigmoidal field lines are representative of @xmath142 .
field lines are multi - coloured to make them easier to visualize . ]
although the source surface will help to determine the behaviour of open field lines , its influence on the geometry of lower closed field lines is smaller as such is determined primarily by the form of the current density .
looking at the field lines between the large positive and negative regions of @xmath85 , the arcades are sheared .
that is , the low field lines at the polarity inversion line are at a different angle to those above . as mentioned before , this effect is due to a non - zero current . the twist of the field due to the linear force - free term can also be seen in the sigmoidal shape of the field lines .
this shape is most clearly visualized in figure [ fieldline_fig ] by the higher field lines of the sheared arcade .
reducing the magnitude of @xmath19 would result in field lines becoming straighter , i.e. losing their sigmoidal shape . in the potential extrapolation of @xcite
, the arcade field lines do not possess a sigmoidal shape as @xmath126 in their case .
the expansion of the field lines can be controlled by @xmath39 , as described previously . to see the effects of the coronal magnetic field on the background atmosphere
, we have constructed maps of the density and pressure perturbations at the stellar surface .
figure [ density_fig ] shows the density perturbation , the second term on the rhs of equation ( [ density ] ) .
$ ] , at the stellar surface .
values on the colour scale are non - dimensional as specified in equation ( [ nondim_p ] ) . ] in the magnetogram for gj 182 ( figure [ br_fig ] ) , the radial field in the northern hemisphere contains concentrations that alternate in sign . beginning with a positive concentration , this changes to negative with increasing longitude . moving further around the star in longitude , the positive to negative pattern repeats .
this pattern is also revealed in the density perturbation map . from equation ( [ density ] )
the density perturbation is proportional to @xmath143 .
hence , the stronger the magnetic polarity , the greater the density perturbation .
the pressure perturbation map , shown in figure [ pressure_fig ] , is very similar to the density perturbation map .
again , this quantity , the second term on the rhs of equation ( [ pressure ] ) , is proportional to @xmath143 .
the only significant difference between the two maps is the magnitude of the perturbations .
, at the stellar surface .
the values are negative as we are considering the perturbation to the hydrostatic pressure , shown in equation ( [ pressure ] ) .
values on the colour scale are non - dimensional as specified in equation ( [ nondim_p ] ) . ]
in this paper we have presented , analyzed and demonstrated a mhs model for stellar coronae .
the model extends the popular potential field model to allow for the magnetic field to contain current and couple to the background density and pressure .
the main assumption of the model is the choice for the current density , which is split into a linear force - free part and another part that links the field to the background atmosphere .
the model , given by equation ( [ current ] ) , has its particular form so that ampre s law ( [ ampere ] ) is linear and analytical progress is possible .
the mhs equations then boil down to solving bessel s equation subject to a model choice for @xmath26 . in this paper
, we have chosen a form that allows for analytical progress .
other choices that allow for analytical solutions are also possible ( e.g. * ? ? ?
* ) . of course
, more complicated expressions for @xmath26 can be specified but will require a numerical treatment ( e.g. * ? ? ?
a closed - form expression for the coronal magnetic field is found and the associated perturbations to the background hydrostatic pressure and density are also determined .
we present an analysis of the effects of the two free model parameters . in terms of the geometry of the field ,
the parameter @xmath19 increases the shear / twist in the magnetic field . increasing the magnitude of @xmath19 creates field lines with a sigmoidal shape .
the parameter @xmath39 is related to the inflation of magnetic field lines through the magnetic pressure .
the size of magnetic arcades can change depending on the choice of @xmath39 .
the topology of the coronal field is studied by considering the relative magnetic helicity .
this quantity gives a measure of how different , topologically , the coronal field is compared to a potential field with the same boundary conditions ( the choice of reference field in this paper ) .
an expression for the relative helicity , for the class of mhs equilibria presented here , is determined .
increasing the magnitude of @xmath19 also increases the relative helicity .
this increase occurs because the current , and hence twist , increases in the magnetic field . increasing the magnitude of @xmath39 increases the relative helicity .
finally , a simple demonstration of the model is given for the m0.5 dwarf star gj 182 .
field lines are plotted showing the non - potential features of the model .
the density and pressure perturbations at the stellar surface are also displayed and are proportional to @xmath143 so that shapes on the countour plots can be matched to corresponding ones on the magnetogram .
the model presented in this paper is suitable for determining the large - scale features of a star s corona .
the model also has the advantage of being simple , fast to implement and , unlike full mhd models , is not computationally expensive to implement .
the effort involved is only slightly more than a potential field model ( which is a special case of the mhs model ) . by a suitable choice of @xmath26 and the free parameters @xmath19 and @xmath39
, one can investigate stellar coronae with different geometries , topologies , pressure and density distributions and current density profiles .
in future work , we will apply our model to stellar observations and attempt to constrain the free parameters using the information provided by the observations , such as the toroidal component of the magnetic field .
sgg acknowledges support from the science & technology facilities council ( stfc ) via an ernest rutherford fellowship [ st / j003255/1 ] .
tn would like to acknowledge support from the stfc consolidated grant st / k000950/1 .
the authors would like to thank j. morin for helpful advice on the paper .
99 arnold , v.i . & khesin , b.a .
1992 , annu .
fluid mech . , 24 , 145 berger , m.a . & field , g.b .
1984 , j. fluid mech . , 147 , 133 berger , m.a .
1985 , apjs , 59 , 433 biskamp , d. 1997 , nonlinear magnetohydrodynamics , cambridge university press bogdan , t.j . &
low , b.c .
1986 , apj , 306 , 271 cohen , o. , drake , j.j . ,
kashyap , v.l . ,
hussain , g.a.j .
& gomnosi , t.i .
2010 , apj , 721 , 80 donati , j .- f . ,
morin , j. , petit , p. , et al .
2008 , mnras , 390 , 545 gary , a. 201 , sol .
, 203 , 71 gibson , s.e . , bagenal , f. & low , b.c .
1996 , j. geophys .
res . , 101 , 4813 gregory , s.g . & donati , j .- f .
2011 , an , 332 , 1027 gregory , s.g . ,
donati , j .- f . , morin , j. , et al .
2012 , apj , 755 , i d .
97 hood , a.w . ,
archontis , v. & mactaggart , d. 2011 , sol .
phys . , 278 , 3 hussain , g.a.j . , jardine , m. , donati , j .- f .
2007 , mnras , 377 , 1488 jackson , j.d .
1975 , classical electrodynamics , john wiley & sons , inc .
jardine , m. , collier cameron , a. & donati , j .- f .
2002 , mnras , 333 , 339 jardine , m. , gregory , s.g .
& donati , j .- f .
2008 , mnras , 386 , 688 jardine , m. , vidotto , a.a . ,
van ballegooijen , a. , donati , j .- f . ,
morin , j. , fares , r. & gombosi , t.i .
2013 , mnras , 431 , 528 lang , p. , jardine , m. , donati , j .- f . ,
morin , j. & vidotto , a. 2012 , mnras , 424 , 1077 low , b.c .
1982 , apj , 263 , 952 low , b.c . 1991 ,
apj , 370 , 427 mactaggart , d. , elsheikh , a. , mclaughlin , j.a .
& simitev , r.d .
2013 , a&a , 556 , a40 neukirch , t. 1995 , a&a , 301 , 628 neukirch , t. & rasttter , l. 1999 , a&a , 348 , 1000 petrie , g.j.d . & neukirch , t. 2000 , a&a , 356 , 735 priest , e.r .
1982 , solar magnetohydrodynamics , kluwer academic publishers rae , i.m .
2002 , quantum mechanics , iop publishing ltd rgnier , s. 2013 , solar phys .
, 288 , 481 riley , p. , linker , j.a . ,
miki , z. , lionello , r. , ledvina , s.a . & luhmann , j.g .
2006 , apj , 653 , 1510 romanova , m.m .
, long , m. , lamb , f.k . ,
kulkarni , a.k .
& donati , j .- f . 2010 , mnras , 411 , 915 ruan , p. , wiegelmann , t , inhester , b. , neukirch , t. , solanki , s.k .
& feng , l. 2008 , a&a , 481 , 827 vidotto , a.a . ,
jardine , m. , opher , m. , donati , j .- f .
& gombosi , t.i . 2011 , mnras , 412 , 351 wiegelmann , t. & sakurai , t. 2012 , living rev . solar phys . ,
9 , 5 zhao , x.p , hoeksema , j.t . &
scherrer , p.h .
2000 , apj , 528 , 932
the spherical components of the coronal magnetic field are @xmath144 | we introduce to the stellar physics community a method of modelling stellar coronae that can be considered to be an extension of the potential field . in this approach , the magnetic field
is coupled to the background atmosphere .
the model is magnetohydrostatic ( mhs ) and is a balance between the lorentz force , the pressure gradient and gravity .
analytical solutions are possible and we consider a particular class of equilibria in this paper .
the model contains two free parameters and the effects of these on both the geometry and topology of the coronal magnetic field are investigated .
a demonstration of the approach is given using a magnetogram derived from zeeman - doppler imaging of the 0.75@xmath0 m - dwarf star gj 182 .
[ firstpage ] mhd stars : coronae stars : magnetic field methods : analytical |
dusty or complex plasma is a low - temperature plasma , which includes dust particles with sizes ranging from @xmath0 to @xmath1 . due to the higher electron mobility
, particles acquire a considerable electric charge .
thus , a strongly coupled coulomb system is formed.@xcite in such plasma , various collective phenomena at the level of individual particles take place .
complex plasmas are usually studied in gas discharges at low pressures , e.g. , in the radio frequency ( rf ) discharges . under microgravity conditions , a large homogeneous bulk of the complex plasma can be observed .
the particles form a nearly homogeneous cloud around the center of the chamber , typically with a central void caused by the ions streaming outwards .
the microgravity conditions are realized either in parabolic flights@xcite or onboard the international space station ( iss).@xcite some experiments are carried out on an inhomogeneous system consisting of the particles with different diameters .
the simplest example of such a system is a large particle surrounded by a dense cloud of smaller particles .
usually , this particle called the projectile moves through the cloud with a supersonic or subsonic velocity .
such projectiles are generated using controlled mechanisms of acceleration,@xcite or they can appear sporadically.@xcite in the latter case , agglomerates or larger particles left over from previous experiments ( not removed during the cleaning procedure and accumulated at the periphery of the particle cloud ) are cracked upon illumination by the laser sheet or upon shaking the chamber .
detached individual particles acquire a negative charge and then they are accelerated due to the coulomb repulsion.@xcite in the work by havnes _
et al._,@xcite propagation of a long - wave nondispersive disturbance , which is usually called the sound , and formation of a cone corresponding to the mach cone in a continuous medium were predicted for the strongly coupled system of the dust particles . in the first experiments ,
the mach cones in the 2d plasma crystal were excited by a sphere moving faster than the lattice sound speed beneath the 2d lattice plane ( samsonov _ et al._@xcite ) and by applying a force from the radiation pressure of a moving laser beam ( melzer _ et al._@xcite ) .
later , nosenko _
et al._@xcite used the latter method of the disturbance excitation and observed several mach cones , which were attributed by the authors to propagation of the compressional and shear wakes and to their interference .
the shape of the mach cones formed by nondispersive linear sound waves was calculated analytically , and the curved wings of the mach cones were experimentally observed by zhdanov _ et al_.@xcite in the recent works by caliebe , arp , and piel,@xcite jiang _
et al._@xcite , and schwabe _
et al._,@xcite the excitation of the 3d mach cones by the supersonic projectiles moving in a strongly coupled cloud of charged particles was observed . in these studies ,
argon was used as a buffer gas .
the determined speed of sound is not much different in the performed experiments . in this work ,
we report the first measurement of the speed of sound in the dust cloud by observation of the mach cones in the case that neon is used as a buffer gas . the measured speed of sound proved to be more than twice as low as for argon , which means that this quantity depends rather sensitively on the sort of a gas
. we can account for observed dependences on the basis of a scaling law for the dust cloud obtained in ref . .
the paper is organized as follows . in sec .
[ s2 ] , we describe the experimental setup and the method of the mach cone visualization
. the details of experimental data processing and the results of speed of sound measurement are presented in sec .
a qualitative interpretation for our experiment is discussed in sec .
[ s4 ] , and the results of this study are summarized in sec .
the experiment was performed during the 13th mission of pk-3 plus on the iss .
the setup is described in detail in ref . .
the heart of this laboratory consists of a capacitively coupled plasma chamber with circular electrodes of 6 cm diameter and 3 cm apart .
a radio frequency ( rf ) voltage applied to these electrodes generates a bulk of plasma .
a dust cloud was formed by the microparticles injected into the main plasma with dispensers .
neon was used as a buffer gas at the pressures of 15 and 20 pa , and the main microparticle cloud was composed of the monodisperse silica particles with the diameter of @xmath2 .
the diameter of observed projectiles estimated as the diameter of larger particles , which are also present in the chamber , is most likely the same as in the experiment@xcite , i.e. , it is equal to @xmath3 .
the trajectories of the dust particles and the projectile were monitored using the optical particle detection system , which consisted of a laser illumination system and a recording system , containing three progressive scan ccd - cameras .
the illumination system is based on two laser diodes with @xmath4 and a continuous wave optical power of 40 mw , the light of which is focused to a thin sheet .
this laser light sheet has a full width at half maximum of about @xmath5 at the focal axis . the cameras with different magnifications and fields of view recorded the light scattered by the microparticles at @xmath6 .
to analyze the microparticle motion , we used three cameras with different fields of view and resolutions , which showed the entire microparticle cloud between the electrodes .
the plasma glow was filtered out .
the cameras follow the pal standard with a resolution of @xmath7 pixels .
each camera provides two composite time interlaced video channels with 25 hz frame rate .
both video channels from one camera were selected for recording , so they were combined to a 50 hz progressive scan video .
( a ) snapshot of the projectile moving through the cloud of dust particles with a supersonic speed ; ( b ) enlarged fragment of the snapshot ; and ( c ) the result of the mach cone visualization .
the neon gas pressure is 20 pa.,width=356 ] we observed two events of the projectile motion through the dust cloud .
figure [ f1 ] shows the first one .
the images were recorded by the quadrant view camera with the resolution of @xmath8 and @xmath9 per pixel in horizontal and vertical directions , respectively , at 50 frames / s .
the projectile moved with a supersonic velocity from the upper left to the lower right side of the dust cloud [ fig .
[ f1](a ) ] .
the track of the moving projectile is surrounded by a dust - free region ( cavity ) , which emerges as a result of a strong coulomb repulsion between the negatively charged dust particles and the projectile [ fig .
[ f1](b ) ] .
the cavity is elongated , the position of a projectile being eccentric . a comparison with [ fig .
[ f1](c ) ] shows that in the center of a perturbation propagating through the dust cloud , the number density of dust particles is continuous both in the vicinity of the cavity and far apart from it .
the perturbation proper has a typical form of the mach cone . on this basis
, we can conclude that observed perturbation is a _
contact discontinuity_.@xcite visualization of the mach cone included the comparison of corresponding pixels for each pair of two successive video frames converted to 8-bit grayscale mode negative images .
if a gray value for the latter image was within 10% of that for the former image , a corresponding pixel of the resulting image was left blank [ i.e. , it is white in fig .
[ f1](c ) ] .
otherwise , the pixel assumed the value of the former image .
although the dust particles form a strongly coupled system , they move around their equilibrium positions in the dust crystal . typically , the positions of a dust particle in the successive frames differ by at least one pixel . due to a low velocity , a point rather than
a track in the image represents a particle .
consequently , many more of the unperturbed dust particles appear in the resulting image .
however , if a particle finds itself in the contact discontinuity region , it is represented by a track due to a considerable velocity inside the perturbation .
the tracks of neighboring particles overlap .
thus , a small region of the pixels with almost equal gray values is formed . since the time interval between successive frames is 0.02 s , this region is shifted by 0.02 cm at the propagation velocity of @xmath10 .
if the perturbation thickness is larger than 0.02 cm ( in our experiment , it is about 0.04 cm ) , its images in the successive frames partially overlap .
this area of overlap is represented by the white pixels in the resulting image , which makes it possible to visualize the mach cone in fig .
[ f1 ] ( obviously , the dust - free regions are also represented by the white pixels ) . note that the efficiency of other methods used in refs
. would be insufficient for the visualization if neon was used as a carrier gas .
it is well - known that the mach angle @xmath11 is related to the speed of sound @xmath12 by the mach cone relation @xmath13 , where @xmath14 is the velocity of the perturbation source ( in our case , this is the projectile velocity ) , i.e. , @xmath15 the projectile velocity was determined by manual measurement of the positions of the projectile track centers in different frames .
the velocity proved to increase from @xmath16 to @xmath17 as the projectile crossed the dust cloud .
the estimates show that such acceleration along with a slight curvature of the projectile trajectory , which could lead to a bend of the rulings of a cone , would have a negligibly small effect on the result of determination of @xmath12 as compared to the measurement errors .
the mach angle can be determined using the vectors @xmath18 and @xmath19 that coincide with the lower and upper rulings of the mach cone , respectively : @xmath20 alternatively , one can measure the mach angle @xmath21 ( @xmath22 ) between the vector @xmath18 ( @xmath23 ) and the projectile displacement vector @xmath24 , which coincides with the projectile track .
then @xmath25^{1/2 } . } \\ \end{array } \label{e3}\ ] ] speed of sound as a function of the length of the projectile path in the dust cloud @xmath26 ( the first event )
. the methods of measurement are discussed in the text.,width=347 ] we determined the coordinates of the vectors @xmath27 , @xmath23 , and @xmath24 manually in each frame , which allowed one to measure @xmath28 .
figure [ f2 ] illustrates the results .
_ method 1 _ denotes the calculation by the formula ( [ e2 ] ) ; _ method 2 _ implies averaging of @xmath29 given by eqs .
( [ e2 ] ) and ( [ e3 ] ) .
it is seen that both methods yield close results and no apparent dependence on the coordinate can be revealed within experimental errors .
the averaging over the entire path of the projectile leads to the estimate @xmath30 .
the second event of the projectile motion through the dust cloud was detected at the neon pressure of 15 pa .
the other parameters were the same as for the first event . here
, the mach cone can be resolved only in three processed images , which increases the error . the average projectile velocity for this event @xmath31 is more than twice as low as for the above - discussed event , and the speed of sound still amounts to @xmath32 , which is very close to the previous estimate .
thus , @xmath12 is more than twice as low as for argon@xcite revealing the effect of the gas sort . damping rate as a function of the length of the projectile path in the dust cloud @xmath26 ( the first event).,width=355 ] a clear resolution of the mach cone rulings makes it possible to estimate the damping rate of the propagating perturbation as @xmath33 , where @xmath34 is the base radius of the mach cone . like @xmath28
, @xmath35 reveals no apparent dependence on the coordinate along the projectile path ( fig .
its average amounts to @xmath36 for the first event and to @xmath37 , for the second one . for the first event , the damping length @xmath38 , which is of the same order of magnitude as the visible perturbation wavelength @xmath39 .
both lengths have the scale of ca . three to four interparticle distances .
under the conditions of our experiment , the volume charge of electrons is negligibly small as compared to that of the ions and particles,@xcite and the complex plasma can be treated as a system of negatively charged particles on the uniform positive background of the ions .
such situation is characteristic of strongly coupled coulomb systems under high energy density .
similarly to the theory of ideal collisionless plasma , the perturbation treated in this work is commonly associated with the ion acoustic wave with the speed of sound@xcite @xmath40 where @xmath41 is the dust particle charge in units of the elementary charge @xmath42 , @xmath43 is the particle radius , @xmath44 is the electron temperature ( the boltzmann constant is set to unity ) , @xmath45 , @xmath46 is the electrostatic potential of a particle , @xmath47 is the particle temperature that is related to the average kinetic energy of a particle , @xmath48 is the particle mass , and @xmath49 is the density of the particle material . for our experiment ,
@xmath50 . if we use the orbital motion limited approximation@xcite for determination of @xmath51 and set @xmath52 , where @xmath53 is the temperature of a buffer gas ( room temperature ) , we arrive at the estimate @xmath54 .
obviously , this is yet a _
lower bound_. thus , we can ascertain at least one order of magnitude disagreement between experiment and theory , which can not be removed by existing alternative approaches to the calculation of the particle charge .
therefore , the formula ( [ e4 ] ) is _ fully inapplicable _ for strongly coupled coulomb systems . for such systems , we will derive an alternative estimate for the speed of sound @xmath28 .
obviously , @xmath55 is proportional to the ratio of the pressure , which is in the order of magnitude @xmath56 , to the mass density of a dust cloud @xmath57 .
hence , @xmath58 where @xmath59 is the ion number density and we used the quasineutrality condition @xmath60 .
it was shown in ref . that the overlap of potentials of the dust particles , which scatter streaming ions , leads to the scaling law for the dust cloud that relates the particle number density to the particle radius : @xmath61 , where @xmath62 is some constant ( the dust invariant " ) .
we substitute this in ( [ e5 ] ) to derive @xmath63 unfortunately , neither @xmath64 nor @xmath59 are available for the experiment with neon . for this reason ,
we test the relation ( [ e6 ] ) on the experiments with argon ( table [ t1 ] ) .
it is seen that for @xmath65 , which is close to typical values for argon,@xcite this relation yields a reasonable agreement with the experiment and reproduces the observed dependences of the speed of sound on the particle radius and electron temperature .
equation ( [ e6 ] ) reproduces the determined speed of sound for neon ( @xmath66 ) at @xmath67 , which also seems to be a reasonable value because the ion number density in neon is typically one order of magnitude lower than that in argon@xcite ( however , @xmath62 may be different for neon ) .
note that ( [ e4 ] ) disagrees with the experiments with argon as well , albeit the disagreement is not as great as for neon .
for example , it leads to @xmath68 for the experiment@xcite .
.speed of sound @xmath12 in experiments with argon as a buffer gas at different argon pressures @xmath69 and dust particle diameters vs. theoretical estimation eq .
( [ e6 ] ) at @xmath70 @xcite and @xmath71 . [ cols="^,^,^,^",options="header " , ] consider the damping of a propagating perturbation .
the damping due to the friction between the dust particles and a buffer gas ( neutral or the epstein drag@xcite ) , which always takes place in the complex plasma , is characterized by the friction coefficient@xcite @xmath72 , where @xmath73 is the accommodation coefficient , @xmath74 is the mass of a buffer gas molecule , @xmath75 and @xmath76 are the number density and thermal velocity of the buffer gas molecules , respectively . for our experiments , @xmath77 at the neon pressure of @xmath78 and @xmath79 , at @xmath80 .
the comparison with the above - discussed measurements of the damping coefficient @xmath35 shows that damping of the particle motion by the neutral drag must dominate but in this case , the wave extinction needs a more accurate treatment .
it is of interest to compare our results with those obtained from the mach cone observations in 2d systems.@xcite the first difference is a single 3d mach cone observed in this work ( single cones were also observed in an argon discharge@xcite ) vs. a double cone for a 2d system@xcite ( in refs . and , three to four cones were observed ) . a double cone structure
is sometimes associated with the propagation of the compressional and shear waves.@xcite while there is a principal possibility to excite the compressional and shear waves in a 2d system , the shear waves are unlikely to be observed in the 3d complex plasma .
indeed , in the vicinity of a projectile , a local melting of the dust crystal occurs,@xcite and the medium for the wave propagation becomes liquid .
this rules out the shear waves and accounts for a single cone structure in our case .
note that there is a controversy about the number of the mach cones in the 2d case and their nature as yet .
the second difference between the systems is the dissipation . in the experiments with 2d systems , the buffer gas pressure was an order of magnitude lower , and the particle radius , by several times larger than for our system ( the particle diameter varied from about @xmath81 to @xmath82 in ref . ) .
thus , in our case , the damping rate exceeds this quantity for the conditions of ref . by almost two orders of magnitude
apparently , a high damping is responsible for the absence of the interference patterns in our 3d case while these patterns are observed for a 2d system.@xcite as for the effect of dissipation , the epstein drag qualitatively accounts for the observed dependence @xmath83 in our case : the damping length decreases with the increase in @xmath35 ( see the discussion above ) .
instead , in the 2d case , the opposite trend was registered.@xcite a significant similarity between the two systems lies in the fact that solely the longest wavelength perturbation can be excited because @xmath84 , and both quantities have the scale of several interparticle distances .
this allows one to conclude that for both systems , the observed mach cones correspond to the nondispersive waves and define the speed of sound .
it is noteworthy that the speed of sound determined in ref
. amounts to ca .
@xmath85 and is independent both on the sort of a gas ( argon , xenon , krypton ) and on the particle diameter .
almost the same speed of sound is listed in table [ t1 ] ; likewise , @xmath12 is weakly dependent on the particle diameter .
it is then surprising that for neon that was not investigated in the previous studies we obtained the speed of sound , which is twice as low as for other gases .
to summarize , we used the excitation of the mach cones by large particles moving with the supersonic velocity for measurement of the speed of dust sound in a complex plasma with neon as a buffer gas . for this purpose , a high - definition method of the mach cone visualization
was developed .
the determined speed of sound proved to be more than one order of magnitude lower than that predicted by the theory of the ion acoustic waves .
we propose an interpretation of these results based on the similarity between a strongly coupled coulomb system and a solid .
using a scaling law that relates the dust particle number density to its radius , we obtained the theoretical estimate for the speed of sound , which describes the main regularities of sound propagation in complex plasmas .
the authors gratefully acknowledge the support from the russian science foundation ( project no .
14 - 12 - 01235 ) for theoretical interpretation of the experimental results and the support from dlr / bmwi ( grants nos .
50wm0203 and 50wm1203 ) for realizing our joint space experiments on the iss .
33ifxundefined [ 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 and , eds . , @noop _ _ , series in plasma physics ( , ) link:\doibase 10.1103/physrevlett.72.4009 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.73.652 [ * * , ( ) ]
link:\doibase 10.1143/jjap.33.l804 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1016/j.physrep.2005.08.007 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.81.25 [ * * , ( ) ] link:\doibase 10.1088/0034 - 4885/73/6/066501 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/8/1/007 [ * * , ( ) ] link:\doibase 10.1063/1.3606468 [ * * , ( ) ] link:\doibase 10.1103/physreve.77.026407 [ * * , ( ) ] link:\doibase 10.1103/physreve.83.016402 [ * * , ( ) ] link:\doibase 10.1103/physreve.83.066404 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/10/3/033037 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.83.1598 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.205001 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/10/3/033036 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/85/45002 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/96/55001 [ * * , ( ) ] link:\doibase 10.1063/1.3568839 [ * * , ( ) ] link:\doibase 10.1103/physreve.61.5557 [ * * , ( ) ] link:\doibase 10.1063/1.4881473 [ * * , ( ) ] link:\doibase 10.1116/1.580119 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.83.3649 [ * * , ( ) ] link:\doibase 10.1103/physreve.62.4162 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.88.135001 [ * * , ( ) ] link:\doibase 10.1103/physreve.68.056409 [ * * , ( ) ] link:\doibase 10.1103/physreve.66.026411 [ * * , ( ) ] link:\doibase 10.1103/physreve.69.026407 [ * * , ( ) ] @noop _ _ ( , , ) @noop _ _ ( , , ) link:\doibase 10.1103/physrev.23.710 [ * * , ( ) ] link:\doibase 10.1103/physreve.86.016401 [ * * , ( ) ] | we report the first observation of the mach cones excited by a larger microparticle ( projectile ) moving through a cloud of smaller microparticles ( dust ) in a complex plasma with neon as a buffer gas under microgravity conditions .
a collective motion of the dust particles occurs as propagation of the contact discontinuity .
the corresponding speed of sound was measured by a special method of the mach cone visualization .
the measurement results are incompatible with the theory of ion acoustic waves .
the estimate for the pressure in a strongly coupled coulomb system and a scaling law for the complex plasma make it possible to derive an evaluation for the speed of sound , which is in a reasonable agreement with the experiments in complex plasmas . |
the minimal supersymmetric standard model ( mssm ) is a promising extension of the standard model of particle physics ( sm ) @xcite .
if weak scale supersymmetry ( susy ) exists , it should be discovered at the lhc @xcite , which is scheduled to start in 2008 . however , detailed measurements of the masses , decay widths , couplings , and spins of the discovered particles are only possible at the international linear collider ( ilc ) @xcite . in the first stage of the ilc ,
the center - of - mass energy will be @xmath10 and the luminosity , @xmath11 , will be @xmath12 within the first four years .
in preparing for the ilc , there is an on - going debate over the extent of beam polarization to be included in the initial design @xcite .
it is clear that there will be at least 80% polarization of the electron beam , possibly even 90% @xcite .
a polarized positron beam is technically and financially more involved .
however , it is possible to achieve 30% polarization through the undulator based production of the positrons , already from the start @xcite .
depending on the undulator length , even a higher degree of positron polarization can be achieved . in light of this discussion , it is the purpose of this letter to reconsider the effect of various degrees of electron and positron polarization on a particular supersymmetric production process , namely the radiative production of the lightest neutralino mass eigenstate @xmath13 @xmath14 we shall focus on specific regions of the supersymmetric parameter space .
the signal is a single , highly energetic photon and missing energy , carried by the neutralinos .
the process ( [ eexxg ] ) was previously studied within the mssm and with general neutralino mixing in refs .
the additional effect of polarized beams was considered in refs .
@xcite . in ref .
@xcite , it was shown that polarized beams significantly enhance the signal and simultaneously suppress the main sm photon background from radiative neutrino production , @xmath15 moreover , it was pointed out that for certain regions of the mssm parameter space , the process ( [ eexxg ] ) is kinematically the _ only _ accessible susy production mechanism in the first stage of the ilc at @xmath16 @xcite . here the heavier electroweak gauginos and the sleptons are too heavy to be pair produced , _ i.e. _ their masses are above 250 gev
. other than the standard center - of - mass energy , @xmath17 , at the ilc , also lower energies are of particular interest , namely for higgs and top physics .
higgs strahlung , @xmath18 can be well studied at the threshold energy @xmath19 , which is @xmath20 , for a higgs boson mass of @xmath21 . the cp - quantum number and the spin of the higgs boson can be determined from an energy scan of the production cross section near the threshold @xcite . from a scan at the threshold energy of top pair production , @xmath22 , the top mass @xmath23 can be determined with an error @xmath24 @xcite .
thus the present error on the top mass , @xmath25 @xcite , and the foreseen error from lhc measurements , @xmath26 @xcite , can be reduced by one order of magnitude .
also the top width , @xmath27 , and the strong coupling constant , @xmath28 , can be precisely determined by a multi parameter fit of the cross section , top momentum distribution , and forward - backward charge asymmetry near threshold @xcite . in this letter
, we take these physics questions as a motivation to study the role of polarized beams in radiative neutralino production at the energies @xmath29 , @xmath30 , and @xmath31 at the ilc . for each beam energy , we shall focus on a specific supersymmetric parameter set within the context of minimal supergravity grand unification ( msugra ) @xcite .
we thus consider three msugra scenarios , which we label a , b and c , respectively , and which are listed below in table [ tab : scenarios ] together with the resulting spectra in table [ tab : masses ] . we restrict ourselves to msugra scenarios , in order to reduce the number of free parameters and since we find it sufficient to illustrate our point .
the specific scenarios are chosen such that radiative neutralino production is the _ only _ supersymmetric production mechanism which is kinematically accessible at the given center - of - mass energy .
it is thus of particular interest to learn as much about supersymmetry as is possible through this mechanism .
as we shall see , beam polarization is very helpful in this respect . in sect .
[ definitions ] , we define the significance , the signal to background ratio and define a first set of experimental cuts . in sect .
[ numerics ] , we study numerically the dependence of the signal cross section and the sm background , the significance , and the signal to background ratio on the beam polarization .
in particular , we compare the results for different sets of beam polarizations , @xmath32 , @xmath33 , @xmath34 , @xmath35 , @xmath36 and @xmath37 .
we summarize and conclude in sect .
[ summary ] .
we wish to study the beam polarization dependence of the cross sections for the signal process ( [ eexxg ] ) and the background process ( [ eenng ] ) . in the following ,
we define the theoretical significance as @xcite @xmath38 and the signal to background ratio ( or reliability ) as @xcite @xmath39 here @xmath40 and @xmath41 shall refer to the signal and background cross sections also after the cuts we impose on the final state phase space .
a significance of @xmath42 implies that the signal can be measured at the statistical @xmath43-level .
if the experimental error for the background is , for example , @xmath44 , the signal to background ratio must be greater than @xmath44 @xcite . for a signal to be detectable at the ilc requires at least @xmath45 we remark that these criteria can only be seen as rough estimates to judge whether an excess of signal photons can be measured over the background , since we do not include any detector simulation , with effective particle reconstruction efficiencies .
a detailed monte carlo study is beyond the scope of the present work .
for the tree - level calculation of the cross sections , we use the formul for the amplitudes squared as given in ref .
@xcite , where details of the neutralino mixing can also be found . to regularize the infrared and collinear divergences of the tree - level cross sections for signal and background , we apply cuts on the photon scattering angle @xmath46 @xcite @xmath47 and on the photon energy @xmath48 @xcite @xmath49 with the beam energy : @xmath50 .
the upper cut on the photon energy @xmath51 is the kinematical limit of radiative neutralino production , and also reduces much of the on - shell @xmath52 boson contribution to the background from radiative neutrino production @xcite .
we assume that the mass of the lightest neutralino @xmath53 is known from lhc measurements @xcite .
although @xmath53 can be measured at the percent - level at the lhc , such a precision will not be required for the upper cut @xmath54 , since the endpoint of the photon energy distribution is not densely populated with signal events .
finally , note that the ratios @xmath55 and @xmath56 only weakly depend on the absolute values of the cuts @xmath57 , and @xmath58 , since signal and background have similar distributions in energy @xmath48 and angle @xmath46 , see ref . @xcite .
we define three msugra scenarios a , b , and c for the three different energies @xmath59 , @xmath30 , and @xmath31 , respectively , see table [ tab : scenarios ] .
we choose the three scenarios in such a way , that only the lightest neutralinos can be radiatively produced for each of the @xmath60 values , respectively .
the other susy particles , _
i.e. _ the heavier neutralinos and charginos , as well as the sleptons and squarks are too heavy to be pair produced at the ilc .
it is thus of paramount interest to have an optimal understanding of the signature ( [ eexxg ] ) , in order to learn as much as possible about susy at a given ilc beam energy .
note that in the three scenarios ( a , b , c ) the squark and gluino masses are below @xmath61 gev , respectively and should be observable at the lhc @xcite .
scenario a is related to the snowmass point sps1a
@xcite by scaling the common scalar mass @xmath62 , the unified gaugino mass @xmath63 , and the common trilinear coupling @xmath64 by @xmath65 .
thus the slope @xmath66 remains unchanged . for scenarios b and c , we also choose @xmath67 , however we change the slopes to @xmath68 in scenario b , and @xmath69 in scenario c. for all scenarios , we fix the ratio @xmath70 of the vacuum expectation values of the two neutral higgs fields . in table
[ tab : scenarios ] , we explicitly give the relevant low energy msugra parameters for all scenarios .
these are the @xmath71 and @xmath72 gaugino mass parameters @xmath73 and @xmath74 , respectively , and the higgsino mass parameter @xmath75 .
the masses of the light neutralinos , charginos , and sleptons are given in table [ tab : masses ] .
all parameters and masses are calculated at one - loop order with the computer code spheno @xcite .
note that the lightest neutralino , @xmath76 , is mostly bino in all three scenarios ; @xmath77 in scenario a , @xmath78 in scenario b , and @xmath79 in scenario c. thus in our scenarios , radiative neutralino production proceeds mainly via right selectron exchange in the @xmath80 and @xmath81 channel . left selectron exchange and @xmath52 boson exchange are severely suppressed @xcite .
the background process @xmath82 mainly proceeds via @xmath83 boson exchange .
thus positive electron beam polarization @xmath84 and negative positron beam polarization @xmath85 should enhance the signal rate and reduce the background at the same time @xcite .
this effect is clearly observed in figs . [
fig : scenarioa ] , [ fig : scenariob ] , and [ fig : scenarioc ] for all scenarios . the signal cross section and
the background vary by more than one order of magnitude over the full polarization range .
.definition of the msugra scenarios a , b , and c. all values are given in gev .
we have fixed @xmath86 .
for completeness we have included the corresponding value of @xmath1 for each scenario . [ cols="^,^,^,^,^,^,^,^ " , ] for scenario a
, we show the beam polarization dependence of the signal cross section @xmath87 in fig .
[ fig : sigma220 ] , and the dependence of the background cross section @xmath88 in fig .
[ fig : back220 ] . in both cases
we have implemented the cuts of eqs .
( [ cuts1 ] ) and ( [ cuts2 ] ) .
the contour lines in the @xmath89-@xmath90 plane of the significance @xmath55 , eq . ( [ significance ] ) , and the signal to background ratio @xmath56 , eq .
( [ sigbck ] ) , are shown in figs .
[ fig : signi220 ] and [ fig : reli220 ] respectively .
the results for scenario b are shown in fig .
[ fig : scenariob ] , and those for scenario c are shown in fig .
[ fig : scenarioc ] . in order to quantify the behavior
, we give the values for the signal and background cross sections , the significance @xmath55 and the signal to background ratio @xmath56 for a specific set of beam polarizations @xmath91 , @xmath33 , @xmath34 , @xmath35 , @xmath92 , and @xmath93 in tables [ tab : xa ] , [ tab : xb ] , [ tab : xc ] for the scenarios a , b , and c , respectively . in the tables
, we also give the values for the effective beam polarization @xmath94 which is a helpful quantity for discussing combined polarizations when the electron and the positron beam polarizations have the opposite sign .
for example , a double beam polarization of @xmath95 and @xmath96 acts effectively as ( approximately ) @xmath97 of single beam polarization @xcite .
we find that an additional positron polarization @xmath98 enhances the significance @xmath55 by factors @xmath99 in scenarios @xmath100 , respectively , compared to beams with only @xmath101 polarization latexmath:[$(p_{e^+ } scenarios @xmath100 , respectively , for @xmath103 compared to @xmath104 .
the signal to background ratio @xmath56 is enhanced by @xmath105 for @xmath106 compared to @xmath107 and by @xmath108 for @xmath109 compared to @xmath110 .
if the positron beams would be polarized by @xmath111 , the enhancement factors for @xmath55 are @xmath112 , and for @xmath56 they are @xmath113 . for @xmath114
, it is only with positron polarization that we obtain values of @xmath56 clearly above 1% .
if we have @xmath115 , then @xmath56 exceeds @xmath44 without positron beam polarization .
since the neutralinos are mainly bino , the signal cross section also depends sensitively on the mass @xmath116 of the right selectron . in scenarios @xmath100 the masses
are @xmath117 gev , respectively , see table [ tab : masses ] . for larger masses , the signal to background ratio drops below @xmath118 . with @xmath119 ,
this happens for @xmath120 gev , and the significance would be @xmath121 .
these selectron masses correspond to the msugra parameter @xmath122 gev
. in anomaly mediated supersymmetry breaking ( amsb ) scenarios , the gaugino mass parameters are roughly related by @xmath123 at the weak scale @xcite .
as a consequence , the lightest neutralino is wino like , and left slepton exchange is enhanced in the reaction @xmath124 . therefore , it is no longer possible to increase the signal and to reduce the background simultaneously by a suitable choice of beam polarizations . in amsb scenarios , also @xmath125 gev
is bounded from below @xcite and rather large . for smaller @xmath126 ,
either the lightest stau is the lsp or sleptons would already have been observed .
the relatively heavy selectrons lead then to a suppression of the cross section for radiative neutralino production .
for example in the scenario sps9 @xcite with @xmath127 gev and @xmath128 gev , @xmath129 gev , @xmath130 gev , calculated with spheno @xcite , we find that the significance is @xmath131 and the signal to background ratio @xmath132 , for various values of the beam polarizations at @xmath133 gev . finally , in amsb scenarios the lightest neutralino @xmath13 is nearly mass degenerate with the lightest chargino @xmath134 .
thus , if the radiative production of neutralinos is kinematically accessible , also the pair production of charginos @xmath135 can be well observed . in this context it is interesting to note that the radiative production of charginos @xmath136 might be important , if the charginos decay almost invisibly @xcite .
this could happen , if the neutralino - chargino mass difference is such , that the charginos decay into rather soft pions @xmath137 .
the cross section of radiative neutralino production is also small in models , where the lightest neutralino is higgsino - like , _ i.e. _ @xmath138 .
for example , for @xmath139 gev and @xmath140 gev we find @xmath131 and @xmath141 fb , since the gaugino couplings of the neutralino are suppressed .
the couplings of the neutralino to the @xmath52-boson are proportional to the squared difference of the higgsino couplings , @xmath142 see ref .
@xcite , and thus too small to lead to an observable signal .
we have studied radiative neutralino production @xmath143 at the ilc with longitudinally polarized beams . for the center - of - mass energies
@xmath29 , @xmath30 , and @xmath31 , we have considered three specific msugra inspired scenarios . in our scenarios , only radiative neutralino production is kinematically accessible , since the other supersymmetric particles are too heavy to be pair produced . we have investigated the beam polarization dependence of the cross section from radiative neutralino production and the background form radiative neutrino production @xmath144 .
we have shown that polarized beams enhance the signal and suppress the background simultaneously and significantly . in our scenarios ,
the signal cross section for @xmath145 is larger than @xmath146 , the significance @xmath147 , and the signal to background ratio is about @xmath148 .
the background cross section can be reduced to @xmath149 . increasing the positron beam polarization to @xmath150 , both the signal cross section and the significance increase by about @xmath151 , in our scenarios . for @xmath152
the radiative neutralino production signature is observable at the ilc but both the significance and the signal to background ratio are considerable improved for @xmath153 , making more detailed investigations possible .
the electron and positron beam polarization at the ilc are thus essential tools to observe radiative neutralino production . for unpolarized beams
this process can not be measured .
we conclude that radiative neutralino production can and should be studied at @xmath0 , as well as at the lower energies @xmath29 and @xmath154 , which are relevant for higgs and top physics .
we have shown that for these energies there are scenarios , where other susy particles like heavier neutralinos , charginos and sleptons are too heavy to be pair produced . in any case , a pair of radiatively produced neutralinos is the lightest accessible state of susy particles to be produced at the linear collider .
we would like to thank gudrid moortgat - pick and klaus desch for very helpful discussions .
this work was supported by the sfb transregio 33 : the dark universe .
one of us ( ok ) would like to thank the organizers of the ecfa - ilc workshop valencia 06 , where the initial idea for this work was developed .
( -1.2,0.0)@xmath155 ( -8.0,6.5)@xmath90 ( -1.2,1.0)@xmath156 ( -1.8,1.5)@xmath157 ( -2.6,2.2)@xmath158 ( -3.4,3.1)@xmath159 ( -4.5,4.2)@xmath160 ( -5.5,5.2)@xmath161 ( -6.1,5.9)@xmath162 ( -1.2,0)@xmath155 ( -8,6.5)@xmath90 ( -2.0,1.7)@xmath163 ( -2.9,2.4)@xmath164 ( -3.8,3.2)@xmath165 ( -4.9,4.4)@xmath166 ( -5.8,5.3)@xmath167 ( -6.4,6.1)@xmath168 a. djouadi , j. lykken , k. mnig , y. okada , m. j. oreglia and s. yamashita , international linear collider reference design report volume 2 : physics at arxiv:0709.1893 [ hep - ph ] . j. a. aguilar - saavedra _ et al . _
[ ecfa / desy lc physics working group ] , arxiv : hep - ph/0106315 . t. abe _ et al . _
[ american linear collider working group ] , arxiv : hep - ex/0106055 . k. abe _ et al .
_ [ acfa linear collider working group ] , arxiv : hep - ph/0109166 . g. a. moortgat - pick _
et al . _ , arxiv : hep - ph/0507011 , see also http://www.ippp.dur.ac.uk/ @xmath169gudrid/ source/ s. ambrosanio , b. mele , g. montagna , o. nicrosini and f. piccinini , nucl .
b * 478 * ( 1996 ) 46 [ arxiv : hep - ph/9601292 ] .
h. k. dreiner , o. kittel and u. langenfeld , phys .
d * 74 * ( 2006 ) 115010 [ arxiv : hep - ph/0610020 ] .
s. y. choi , j. s. shim , h. s. song , j. song and c. yu , phys .
d * 60 * ( 1999 ) 013007 [ arxiv : hep - ph/9901368 ] .
h. baer and a. belyaev , `` associated neutralino neutralino photon production at nlc , '' in _ proc . of the aps / dpf / dpb summer study on the future of particle physics ( snowmass 2001 ) _ ed .
n. graf , econf * c010630 * ( 2001 ) p336 [ arxiv : hep - ph/0111017 ] .
a. h. chamseddine , r. arnowitt and p. nath , phys .
* 49 * ( 1982 ) 970 ; p. nath , r. arnowitt and a. h. chamseddine , nucl .
b * 227 * ( 1983 ) 121 . for
a review see also @xcite .
m. t. dova , p. garcia - abia and w. lohmann , arxiv : hep - ph/0302113 .
w. m. yao _ et al .
_ [ particle data group ] , j. phys .
g * 33 * ( 2006 ) 1 .
i. borjanovic _ et al .
_ , eur .
j. c * 39s2 * ( 2005 ) 63 [ arxiv : hep - ex/0403021 ] .
m. martinez and r. miquel , eur .
j. c * 27 * ( 2003 ) 49 [ arxiv : hep - ph/0207315 ] . j. a. aguilar - saavedra _ et al .
_ , eur .
j. c * 46 * ( 2006 ) 43 [ arxiv : hep - ph/0511344 ] . b. c. allanach _ et al .
_ , in _ proc .
of the aps / dpf / dpb summer study on the future of particle physics ( snowmass 2001 ) _ ed .
n. graf , _ in the proceedings of aps / dpf / dpb summer study on the future of particle physics ( snowmass 2001 ) , snowmass , colorado , 30 jun - 21 jul 2001 , pp p125 _ [ arxiv : hep - ph/0202233 ] .
m. dittmar and h. k. dreiner , phys .
d * 55 * ( 1997 ) 167 [ arxiv : hep - ph/9608317 ] ; m. dittmar and h. k. dreiner , arxiv : hep - ph/9703401 .
n. ghodbane and h. u. martyn , in _ proc . of the aps / dpf / dpb summer study on the future of particle physics ( snowmass 2001 ) _ ed .
n. graf , arxiv : hep - ph/0201233 .
w. porod , comput .
commun . * 153 * ( 2003 ) 275 [ arxiv : hep - ph/0301101 ] .
t. omori , `` a polarized positron beam for linear colliders , '' | we analyze the impact of electron and positron beam polarization on radiative neutralino production at the international linear collider ( ilc ) .
we focus on three different msugra scenarios in turn at the higgs strahlung threshold , the top pair production threshold , and at @xmath0 . in these scenarios at the corresponding @xmath1 , radiative neutralino production is the only supersymmetric production mechanism which is kinematically allowed . the heavier neutralinos , and charginos as well as the sleptons , squarks and gluinos are too heavy to be pair produced .
we calculate the signal cross section and also the standard model background from radiative neutrino production . for our scenarios ,
we obtain significances larger than @xmath2 and signal to background ratios between @xmath3 and @xmath4 , if we have electron beam polarization @xmath5 and positron beam polarization @xmath6 .
if we have electron beam polarization of @xmath7 , then the signal is observable with @xmath8 but both the significance and the signal to background ratio are significantly improved for @xmath9 . |
this paper considers the effect of primordial gravitational waves on the cosmic background radiation ( cbr ) .
we consider spatially flat , perturbed friedman - robertson - walker ( frw ) universes that begin with an early inflationary phase .
as the universe rapidly expands , perturbations of the spatial geometry that are local in origin ( eg . thermal fluctuations ) are quickly redshifted both in amplitude and wavelength . after sufficient inflation
these perturbations are no longer visible to an observer ; the only perturbations that remain visible within a hubble sphere are quantum - mechanical zero - point fluctuations .
because these perturbations extend to arbitrarily high frequencies they can not be redshifted away . since the only significant perturbations remaining after inflation are zero - point fluctuations , we assume that the initial state of the universe was the vacuum state appropriate to de sitter space , containing only the quantum fluctuations and no additional excitations .
as the universe continues to expand after inflation , these quantum fluctuations are redshifted to longer wavelengths and amplified ; one may think of this in terms of particle ( graviton ) production ( as we do ) , non - adiabatic amplification , or super - radiant scattering . in the present epoch
the occupation numbers of the graviton modes are large .
this is not surprising .
it has been shown that one may treat the collective effects of primordial gravitational waves as being due to the presence of a stochastic background of classical gravitational waves , and since gravitons are bosons , such an interpretation is only possible if the occupation numbers are large .
sachs and wolfe @xcite showed how gravitational wave perturbations result in cbr temperature anisotropy .
as photons from the cbr propagate , the paths they follow are perturbed by the metric perturbation @xmath6 , which in this discussion is due entirely to primordial gravitational waves .
the energies of the photons are perturbed , which results in temperature fluctuations from point to point on the celestial sphere .
these temperature fluctuations are usually characterized by the two - point angular correlation function @xmath3 defined on the celestial sphere . here
@xmath7 is the angle between two points on the sphere .
most often the two - point angular correlation function is expanded in terms of legendre polynomials , and the expansion coefficients or multipole moments are calculated . for a derivation of the angular correlation function for spatially flat cosmologies
see the recent paper by allen and koranda @xcite .
henceforth we will assume that the reader is familiar with this paper , which contains a detailed review of previous work on this problem , a comprehensive discussion of the physical motivation , and a detailed and self - contained `` first - principles '' calculation .
early work on this subject @xcite assumed that the universe was completely dust dominated at last - scattering when the cbr decoupled , as did recent work by white @xcite which presented a concise derivation of the formula for the multipole moments due to tensor perturbations . also recently
, grishchuk has adapted the terminology and techniques of quantum optics to the analysis @xcite .
grishchuk stresses the importance of the phase correlations between the modes of the metric perturbations . a similar analysis by allen and koranda @xcite using standard `` quantum field theory in curved space '' techniques found equivalent results .
we also showed that the now standard formula given by abbott and wise @xcite and starobinsky @xcite for the @xmath8th multipole moment is a long wavelength approximation to an exact formula . both the work by grishchuk , and by allen and koranda assumed that the universe was completely dust dominated at last - scattering .
the first authors to consider the cbr anisotropy ( within the framework of the sachs - wolfe effect ) for a universe that was _ not _ completely dust dominated at last - scattering were turner , white , and lidsey @xcite .
they used a `` transfer function '' to express the solutions of the wave equation for the gravitational wave amplitude in terms of the standard long wavelength formula given by abbott and wise .
they found that the standard formula , which assumes that the universe is _ completely _ dust dominated at last - scattering , consistently underestimates the contribution of gravitational waves to the cbr anisotropy .
more recent work by ng and speliotopoulos @xcite used numerical methods to integrate the wave equation and found similar results .
neither of these studies used analytic expressions for the gravitational wave amplitude ( or equivalently , the graviton mode function ) .
other work has been done which does not directly use the sachs - wolfe formula to calculate the anisotropy due to gravitational waves .
crittenden et .
@xcite numerically evolved the photon distribution function using first - order perturbation theory of the general relativistic boltzmann equation for radiative transfer , and included a thomson scattering source term .
dodelson , knox and kolb @xcite have done a similar numerical analysis .
both found that the standard formula of abbott and wise is only accurate for small @xmath8 multipole moments , and consistently underestimates the contribution of the gravitational waves to the cbr anisotropy for higher @xmath8 moments . in this paper
we give the first correct analytic expression for the graviton mode function in a cosmology that transforms smoothly from being radiation to dust dominated .
( we correct a minor error in earlier work by sahni @xcite and nariai @xcite which claimed to find an analytic expression for the mode function . )
we use this analytic expression and the sachs - wolfe formula to find an analytic expression for the multipole moments @xmath2 of the angular correlation function @xmath3 . the analytic expression for the multipole moments
is written in terms of two integrals ; we use numerical methods to evaluate these integrals and report numerical values for the multipole moments .
we compare our results with those mentioned above , and find good agreement .
the paper is organized as follows . in section [ angularcorrelationfunction ]
we reproduce general expressions for the angular correlation function derived in @xcite , and explain how one uses these formulae to calculate the multipole moments for any spatially flat , inflationary cosmological model . in section [ cosmologicalmodel ]
we introduce our cosmological model , which `` begins '' with an infinite de sitter phase followed by a mixed phase that contains both radiation and dust .
early in the mixed phase the universe is radiation dominated ; later it transforms smoothly from being radiation dominated to dust dominated . in section [ gravitonmodefunction ]
we solve the massless klein - gordon equation ( or wave equation ) for the graviton mode function . during the mixed phase the solutions to the wave equation
are expressed in terms of spheroidal wave functions .
the multipole moments are calculated in section [ multipolemoments ] using the graviton mode function determined in section [ gravitonmodefunction ] .
an analytic expression for the moments is given in terms of two integrals , which are then evaluated numerically . in appendix [ appendixa ]
we discuss the spheroidal wave functions .
the differential equation of spheroidal wave functions is introduced and its solutions examined .
we introduce a useful notation for the spheroidal wave functions , and give a practical method for evaluating them . finally in appendix
[ appendixb ] we describe the numerical techniques used to evaluate the spheroidal wave functions and the two multipole moment integrals . throughout this paper
we use units where the speed of light @xmath9 .
we retain newton s gravitational constant @xmath10 and planck s constant @xmath11 explicitly .
we consider only the anisotropy of the cosmic background radiation ( cbr ) arising via the sachs - wolfe effect @xcite from tensor perturbations ( gravitational waves ) .
the anisotropy is characterized by the two - point angular correlation function @xmath3 , where @xmath7 is the angle between two points located on the celestial sphere .
the correlation function may be expanded in terms of legendre polynomials as @xmath12 the expansion coefficients @xmath2 are referred to as the multipole moments .
the multipole moments are given in terms of an integral over graviton wavenumber @xmath13 , @xmath14 where the function @xmath15 is proportional to the sachs - wolfe integral along null geodesics , and is given by @xmath16 here @xmath17 is the time of last - scattering , and @xmath18 is the conformal time today .
the function @xmath19 is a spherical bessel function of the first kind @xcite .
the function @xmath20 is proportional to the first derivative of the graviton mode function @xmath21 , and is defined by @xmath22_{\eta=\ne+\lambda}. \label{fdefinition}\ ] ] the graviton mode function obeys the massless klein - gordon or wave equation @xmath23 where @xmath24 is the scale factor and @xmath25 the mode function must satisfy the wronskian normalization condition @xmath26 the only physical input required is the choice of an initial quantum state for the gravitational field , which amounts to a choice of boundary conditions for the wave equation ( [ kge ] ) .
these formula for the angular correlation function are very general ; a detailed and complete derivation is given in @xcite .
to calculate the correlation function ( or equivalently the multipole moments ) for any particular cosmological model , one need only to solve the wave equation ( [ kge ] ) for the graviton mode function , and substitute into the formulae ( [ correlationdefinition]-[fdefinition ] ) .
the spacetime considered here is a spatially flat , perturbed frw universe .
the metric is @xmath27 , \label{metric}\ ] ] where @xmath28 is the flat metric of @xmath29 , @xmath30 is the conformal time , and @xmath24 is the cosmological length scale or scale factor .
the scale factor satisfies the einstein equations @xmath31 where @xmath32 is the energy density and @xmath33 the pressure of the cosmological fluid .
the metric perturbation @xmath6 is assumed to be small ; in the limit as @xmath6 vanishes the spacetime is an unperturbed frw universe .
we have chosen a gauge so that the tensor perturbation @xmath6 has only spatial components . in the absence of the metric perturbation @xmath6 ,
the background spacetime is a spatially flat frw universe .
since the spacetime is spatially flat , the density parameter @xmath34 , which is the ratio of the present - day energy density @xmath35 to the critical energy density required to produce a spatially flat universe , is equal to unity : @xmath36 to specify the model , we need to give the scale factor @xmath24 . we do this in such a way that the model is completely defined by the minimal set of free parameters given in table [ table1 ] .
_ all _ our results , including the final expression ( [ moments ] ) , can be expressed in terms of this minimal set of parameters . for clarity ,
we often define auxiliary quantities and express results in terms of them ; the auxiliary quantities can be expressed in terms of the parameters in table [ table1 ] . in typical inflationary models ,
the free parameters in table [ table1 ] have values of order @xmath37 between 50 and 100 km @xmath38 @xmath39 , @xmath40 , @xmath41 , and @xmath42 . for a review of inflationary cosmology see reference @xcite .
our cosmological model passes through two phases .
the first phase is a de sitter or inflationary phase . during the de sitter
phase the universe expands exponentially ( expressed in terms of _ comoving _ time @xmath43 , the scale factor behaves like @xmath44 , where @xmath45 is the hubble constant ) . in terms of conformal time
, the scale factor is @xmath46 where @xmath47 is the conformal time at the end of the de sitter phase . in terms of the parameters listed in table [ table1 ] , @xmath48^{1/2}\approx h_0^{-1}{\sqrt{\zeq}\over \zend},\end{aligned}\ ] ] where we have set the scale factor today @xmath49 . during the de sitter
phase the energy density @xmath50 is constant and given by @xmath51 the pressure during the de sitter phase is negative and constant , and is given by @xmath52 the de sitter phase is followed by a mixed phase .
immediately following the de sitter phase is a mixed phase containing both dust and radiation .
the scale factor is @xmath53 \>\>\text{for}\>\>\eta\geq\nend . \label{mixedscalefactor}\ ] ] the constant @xmath54 is defined in terms of the free parameters by @xmath55 the stress - energy tensor is that of a perfect fluid with energy density @xmath56 where the energy density for the dust @xmath57 , and for the radiation @xmath58 , are @xmath59 here @xmath60 is the energy density at conformal time @xmath61 , when the dust and radiation energy densities are equal , and is given by @xmath62 the pressure of the cosmological fluid @xmath33 is @xmath63 where the pressure at dust / radiation equality @xmath64 is @xmath65 by inspection of ( [ radiationenergydensity]),([mixedpressure ] ) , and ( [ pressureeq ] ) one sees that @xmath66 which is the relation between the pressure and energy density one would expect since the dust has zero pressure and the pressure is due entirely to the radiation present .
the scale factor at @xmath61 is @xmath67 one can express @xmath61 in terms of the conformal time at the end of the de sitter phase @xmath47 and the constant @xmath54 as @xmath68 for @xmath69 the energy density ( [ mixedenergydensity ] ) varies as @xmath70 and the cosmological model is radiation dominated , and for @xmath71 the energy density varies as @xmath72 and the model is dust dominated .
the model makes a smooth transition at @xmath73 from being radiation to dust dominated .
to calculate the multipole moments for the cosmological model described above , one must first solve the wave equation ( [ kge ] ) for the graviton mode function .
we first solve for the `` natural '' positive- and negative - frequency mode functions during both the de sitter and the mixed phases .
we then make an appropriate choice of initial mode function during the de sitter phase .
this choice of initial mode function completely determines the graviton mode function for all later times .
we express the mode function at later times using bogolubov coefficient notation .
the solution to the wave equation ( [ kge ] ) for the de sitter phase can be expressed in terms of spherical hankel functions .
the scale factor during the de sitter phase is given in ( [ desitterscalefactor ] ) . by making the change of dependent variable @xmath74 and
the change of independent variable @xmath75 the wave equation can be expressed in the form @xmath76 this is bessel s differential equation , and the solutions are spherical bessel or hankel functions @xcite . using the normalization condition ( [ phinormalization ] )
one obtains for the positive - frequency mode function during the de sitter phase @xmath77 where @xmath78 is a spherical hankel function of the second kind @xcite , and @xmath79 is the planck energy density .
the negative - frequency mode function during the de sitter phase @xmath80 is just the complex conjugate of the positive - frequency mode function ( [ phidesitter ] ) .
the positive- and negative - frequency mode functions form a complete solution to the wave equation for the de sitter phase .
the wave equation during the mixed radiation and dust phase @xmath81 can be cast in the form of the spheroidal wave function differential equation . by making the change of dependent variable @xmath82
the wave equation ( [ kge ] ) becomes @xmath83 one may define a new independent variable @xmath84 so that @xmath85 using ( [ xdefinition ] ) and ( [ dxdn ] ) one may write the wave equation ( [ wavetwo ] ) in the form @xmath86{d
\chi \over dx}\nonumber\\ & & \nonumber\\ & + & \bigg[4 k^2 x^2 { a^2(\nq)\over { \dot a}^2(\eta ) } -x^2 { a^2(\nq)\over a^2(\eta)}-2x^2 { \ddot a(\eta)a^2(\nq)\over { \dot a}^2(\eta ) a(\eta)}\bigg]\chi=0 . \label{wavethree}\end{aligned}\ ] ] this expression can be simplified using the einstein equations ( [ einstein ] ) , along with ( [ mixedenergydensity ] ) and ( [ xdefinition ] ) one can show that @xmath87 and using this expression one may write the wave equation ( [ wavethree ] ) as @xmath88\chi=0.\ ] ] since from ( [ xdefinition ] ) @xmath89 and from ( [ mixedscalefactor ] ) and ( [ xidefinition ] ) @xmath90 the wave equation becomes @xmath91 here @xmath92 is defined by @xmath93 where @xmath94 is the _ present _ day wavelength of the mode with wavenumber @xmath13 . for the multipole moments of interest , @xmath95
$ ] , the contributions typically come from wavelengths in the range @xmath96 $ ] .
this form of the wave equation is the spheroidal wave function differential equation ( [ diffeq ] ) , which is discussed in detail in appendix [ appendixa ] .
the solutions to the wave equation ( [ kge ] ) during the mixed phase can be expressed in terms of spheroidal wave functions . by inspection of ( [ xdefinition ] )
one sees that @xmath97 , and so the possible solutions to the wave equation ( [ diffeq2 ] ) may be expressed as sums of any pair of @xmath98 where @xmath99 is a spheroidal wave function . using ( [ chidefinition ] ) ,
( [ chixk ] ) , the normalization condition ( [ phinormalization ] ) , and the wronskian relation ( [ wronskian ] ) , one obtains a positive - frequency mode function during the mixed phase @xmath100 note that @xmath101 is a function of the conformal time @xmath30 and @xmath92 depends on the wavenumber @xmath13 .
the negative - frequency mode function @xmath102^*$ ] is just the complex conjugate of the positive - frequency mode function .
the positive and negative frequency mode functions form a complete solution to the wave equation during the mixed radiation and matter phase .
note that the choice of graviton mode function during the de sitter phase completely determines the mode function at all later times .
thus , the choice ( [ mixedmodefunction ] ) of `` positive - frequency '' during the mixed phase is unimportant .
the choice of mode function during the de sitter phase completely determines the mode function at all later times . this is because a solution to the wave equation depends only on the values of @xmath103 and @xmath104 on a spacelike cauchy surface ( ie .
a surface of constant @xmath30 ) .
we choose the mode function during the de sitter phase to be the pure positive - frequency de sitter solution ( [ phidesitter ] ) .
this is the unique solution corresponding to a de sitter - invariant vacuum state with the same ( hadamard ) short distance behavior as one would find in minkowski space @xcite .
having made this choice for the mode function during the de sitter phase , the mode function during all times is @xmath105 where @xmath106 and @xmath107 are bogolubov coefficients .
the bogolubov coefficients are determined by requiring that the mode function and its first derivative be continuous at @xmath108 .
one obtains @xmath109 { i\over 2k\nend}\bigg(1+{i\over k\nend}\bigg ) -{^3 \s^1_2}(x_1,\kappa)\bigg\},\nonumber\\ & & \nonumber\\ \beta(k\nend)&=&k\nend\sqrt { { 1+\xi\over\xi } } \bigg\{\bigg[{\xi\over \sqrt{1+\xi } } { ^4 \s^1_2}'(x_1,\kappa)- { ^4 \s^1_2}(x_1,\kappa)\bigg ] { i\over 2k\nend}\bigg(1+{i\over k\nend}\bigg ) -{^4 \s^1_2}(x_1,\kappa)\bigg\}.\nonumber\\ & & \label{bogolubov}\end{aligned}\ ] ] the prime in ( [ bogolubov ] ) is defined by @xmath110_{z = x_1},\ ] ] where @xmath111 using the wronskian relation for the spheroidal wave functions ( [ wronskian ] ) one can easily verify that @xmath112 we stress again that _ the choice of graviton mode function during the de sitter phase completely determines the mode function at all later times_. thus , the choice of `` positive - frequency '' during the mixed phase is unimportant .
had we chosen a different solution to the wave equation during the mixed phase to call `` positive - frequency '' , then the bogolubov coefficients ( [ bogolubov ] ) would be different in such a way so that the mode function ( [ finalmodefunction ] ) _ would be the same_.
having determined the graviton mode function for our cosmological model , we may calculate the angular correlation function @xmath3 , or equivalently , the multipole moments @xmath2 .
we simply substitute the graviton mode function ( [ finalmodefunction ] ) into the formulae ( [ correlationdefinition]-[fdefinition ] ) .
the graviton mode function ( [ finalmodefunction ] ) is exact ; no approximations ( ie .
long wavelength ) have been made .
one may directly substitute the mode function into the formulae ( [ correlationdefinition]-[fdefinition ] ) to obtain an exact expression for the multipole moments ; because the arguments of the spheroidal wave functions in ( [ finalmodefunction ] ) are functions themselves , however , the result is complicated and not very illuminating . in typical inflationary models ,
the amount of expansion is very large . if one takes the limit @xmath113 , then @xmath114 .
this allows one to write a fairly compact expression for the multipole moments ; after substituting the mode function ( [ finalmodefunction ] ) into the formulae ( [ correlationdefinition]-[fdefinition ] ) , one may collect together terms in the expression for the multipole moments @xmath115 which are the same order in @xmath54 .
then in the limit as @xmath113 and @xmath114 , one may consider only the leading term .
the leading term , which is @xmath116 , is given below in ( [ moments ] ) .
the neglected terms are @xmath117 or greater . for typical inflationary models
one has @xmath118 and @xmath119 , so that @xmath120 .
so the expression ( [ moments ] ) is quite accurate for most cosmological models , since the neglected terms are very small .
the expressions for the multipole moments @xmath2 are fairly simple . after substituting the mode function ( [ finalmodefunction ] ) into the formulae ( [ correlationdefinition]-[fdefinition ] ) , and making the same changes of variable as in section [ mixed ] , ie .
@xmath121 one obtains for the multipole moments @xmath122 @xmath123 the integral above is simply an integral over the ( rescaled , dimensionless ) wavenumber @xmath92 .
the functions @xmath124 are the same functions defined in ( [ newtdefinition ] ) .
the functions @xmath125 are the ( reparametrized ) sachs - wolfe integrals over null geodesics , given by @xmath126 where the limits @xmath127 and @xmath128 on the integral are given in terms of redshift by @xmath129 these formulae , and especially the spheroidal wave functions , may be numerically evaluated using the techniques discussed in appendix [ appendixb ] .
henceforth we refer to ( [ moments ] ) and ( [ gldefinition ] ) as the `` mixed '' formulae for the multipole moments .
furthermore , we refer to multipole moments calculated using ( [ moments ] ) and ( [ gldefinition ] ) as the `` mixed '' multipoles .
we have numerically evaluated the mixed " formulae ( [ moments ] ) and ( [ gldefinition ] ) .
the second column of table [ table2 ] shows the mixed " multipoles .
the third column of table [ table2 ] shows the results obtained by evaluating eq .
( 6.2 ) of allen and koranda @xcite , which we refer to as the exact - dust " formula for the multipole moments .
the exact - dust " formula assumes ( 1 ) the universe begins with an initial de sitter phase , followed by ( 2 ) a pure radiation phase containing only radiation and no dust , followed by ( 3 ) a pure dust phase containing only dust and no radiation , during which ( 4 ) last - scattering takes place
. we refer to this type of universe as a dust " universe , to distinguish it from the model described by ( [ mixedscalefactor ] ) , ( [ mixedenergydensity ] ) , and ( [ mixedpressure ] ) , which we refer to as mixed " universe because it contains both dust and radiation at last - scattering .
the fourth column shows the results obtained by evaluating eq .
( 6.1 ) of allen and koranda , which we refer to as the approximate - dust " formula for the multipole moments .
the approximate - dust "
formula is a long - wavelength approximation to the exact - dust " formula , and is _ equivalent _ to the standard formulae for the multipole moments given by abbott and wise @xcite and starobinsky @xcite .
the fifth column shows recent results of ng and speliotopoulos @xcite , who numerically integrated the wave equation to obtain the amplitude of the gravitational wave ( or graviton mode function ) .
they also considered a universe model which contains both radiation and dust at last - scattering .
all of the results in table [ table2 ] were obtained with @xmath130 and @xmath131 .
figures [ figure1],[figure2 ] , and [ figure3 ] compare the
mixed " multipole moments to the multipole moments calculated using other techniques .
the quantity @xmath132 in figures [ figure1]-[figure3 ] is defined by @xmath133 ( note that in @xcite the equation defining @xmath132 contains an extraneous factor of @xmath134 . )
the points labeled `` mixed '' in figure [ figure1 ] show results obtained by numerically evaluating the mixed " formulae ( [ moments ] ) and ( [ gldefinition ] ) .
the points in figure [ figure1 ] labeled `` transfer function '' show results from turner , white , and lidsey @xcite , who numerically integrate the wave equation , and express the solutions in terms of the standard long - wavelength approximate mode functions @xcite using a `` transfer function '' .
the points in figure [ figure2 ] labeled `` mixed '' again show results obtained by numerically evaluating the mixed " formulae .
the points labeled `` exact - dust '' show results obtained using the exact - dust " formula , and the points labeled `` approximate - dust '' show results obtained using the approximate - dust " formula .
figure [ figure3 ] compares the mixed " results to results from dodelson , knox , and kolb @xcite ( labeled `` boltzmann '' ) , who do not use the sachs - wolfe formula to calculate the cbr anisotropy .
instead they use numerical methods to evolve the photon distribution function using first - order perturbation theory of the general relativistic boltzmann equation for radiative transfer .
all the multipole moments shown in figures [ figure1 ] , [ figure2 ] , and [ figure3 ] are for cosmological parameters @xmath135 and @xmath131 .
( the values of @xmath136 for the turner , white , and lidsey work and for the the dodelson , knox , and kolb results differs from @xmath137 by about one percent ) .
the ng and speliotopoulos multipoles agree quite well with the mixed " multipoles obtained here , using spheriodal wavefunctions , for a cosmology containing dust and radiation components .
this is expected because the two methods used to calculate the multipole moments should be essentially equivalent .
the graviton mode functions ng and speliotopoulos obtain by numerically integrating the wave equation ( with the correct boundary conditions ) must be equivalent to our analytic expressions ( [ finalmodefunction ] ) .
the cosmological model ng and speliotopoulos consider , however , is slightly different than our own .
they model the smooth transition during the mixed phase by a simpler scale factor than our own ( [ mixedscalefactor ] ) .
this may account for the small discrepancy between their results and our own .
the dodelson , knox , and kolb ( figure [ figure3 ] ) multipoles also appear consistent with the mixed " multipoles .
this is also expected since the boltzmann formalism @xcite and the sachs - wolfe formalism should yield equivalent results .
we do not understand at this time the small discrepancy between the two multipole spectrums . for @xmath138 , the exact - dust " multipoles and the approximate - dust " multipoles agree fairly well with the mixed " multipoles ( figure [ figure2 ] ) .
this is not surprising .
the small @xmath8 multipoles @xmath2 are most affected by longer wavelength perturbations .
these longer wavelength perturbations were redshifted outside the hubble sphere early in the inflationary phase , and only recently re - entered the hubble sphere ( the longest wavelength perturbations remain outside the hubble sphere even today @xcite ) . because they remained outside the hubble sphere until recently , the longer wavelength perturbations are insensitive to the details of the evolution of the universe before it became dust dominated .
thus , the mixed " universe and dust " universe models are essentially equivalent for longer wavelength perturbations .
so it s not surprising that the same cbr anisotropies are produced by the long wavelength perturbations in either model .
thus the exact - dust " , approximate - dust " , and
mixed " multipoles should , and do , agree for small @xmath8 .
furthermore , because the approximate - dust " formula for the multipole moments is equivalent to the standard formulae given by abbott and wise @xcite and starobinsky @xcite , one can conclude that the standard formulae for the multipole moments are accurate for small @xmath8 , whether or not the universe was completely dust dominated at last - scattering . for @xmath139 ,
the exact - dust " multipoles and the approximate - dust " multipoles differ significantly from the mixed " multipoles .
again this is not surprising .
the larger @xmath8 multipoles are more affected by shorter wavelength perturbations , which re - entered the hubble sphere before the universe became dust dominated , and are therefore sensitive to the details of the cosmological expansion before dust domination .
mixed " universe becomes dust dominated _ much more slowly _ than a dust " universe , which is dust dominated immediately after the radiation phase ends at dust / radiation equality @xmath61 .
therefore the shorter wavelength modes in a
mixed " universe and a dust " universe evolve very differently after they re - enter the hubble sphere .
this difference is evident in figure [ figure2 ] for the large @xmath8 multipole moments .
the turner , white , and lidsey ( twl ) multipole spectrum differs significantly from the mixed " multipole spectrum ( figure [ figure1 ] ) .
in particular their multipole moments are significantly greater than the mixed " multipoles for @xmath140 .
although twl consider a mixed " universe , they use the standard multipole moment formulae for a dust " universe .
however , they modify the standard formulae by including a time - independent transfer function @xmath141 , which depends only on the wavenumber @xmath13 .
twl give an explicit functional form for the transfer function , which they obtain by numerically integrating the wave equation @xcite : @xmath142^{1/2},\>\text{where}\>y = k / k_{\rm eq}. \label{numericaltransferfunction}\ ] ] since the transfer function @xmath143 appears in the expression for the multipole moments as @xmath144 ( see eqs .
( 22 ) and ( 23 ) of @xcite ) , one can see that the effect of the transfer function ( [ numericaltransferfunction ] ) is to _ increase _ the multipole moments .
furthermore , the contribution from shorter wavelength ( larger @xmath13 ) modes is enhanced more than the contribution from larger wavelength ( smaller @xmath13 ) modes , since @xmath145 as @xmath146 .
since the twl multipoles are significantly greater than the mixed " multipoles for @xmath147 , _ the transfer function overestimates the contribution to the multipole moments from longer wavelength modes_. this is easy to see . as discussed above ,
the standard formulae are accurate for small @xmath8 multipoles , and give essentially the same results as the mixed " formulae for @xmath138 .
the twl formulae for the multipole moments are equivalent to the standard formulae , except for the transfer function .
thus if the transfer function was set to unity , the twl multipoles would be the same as the standard multipoles , and hence would be equivalent to the
mixed " multipoles for @xmath138 .
since the twl multipoles are larger than the mixed " multipoles for @xmath147 , the transfer function must enhance the the contribution to the @xmath140 multipole moments too much . because the small @xmath8 moments are affected most by longer wavelength perturbations
, the transfer function must overestimate the contribution from longer wavelengths , or smaller wavenumber @xmath13 .
the twl multipole spectrum is also significantly different than the mixed " multipole spectrum for large @xmath8 .
the large @xmath8 multipoles are most affected by shorter wavelength modes .
the transfer function ( [ numericaltransferfunction ] ) significantly enhances the contribution of shorter wavelength modes ( larger @xmath13 ) to the multipole moments . because the twl transfer function is time - independent , however
, it can not alter the time evolution of the shorter wavelength modes . by comparing the dust " universe graviton mode function ( see eq .
( 4.28 ) of @xcite ) @xmath148 ( which is equivalent to the standard formulae for the gravity - wave amplitude @xcite , and hence equivalent to the twl formulae for the gravity - wave amplitude ) to the mixed " universe mode function ( [ finalmodefunction ] ) , one can see that the time evolution of the shorter wavelength modes is very different in a dust " universe when compared to the time evolution in a
mixed " universe .
thus one expects that for shorter wavelength modes , the integral along the null geodesics ( [ ildefinition ] ) in the sachs - wolfe formula will be very different for a dust " universe , when compared to a mixed " universe . since the transfer function is time - independent , it has no effect on the integral ( [ ildefinition ] ) .
so one expects that the multipole spectrum for a dust " universe will be very different than the multipole spectrum for a mixed " universe , even with the transfer function included .
figure [ figure1 ] shows that this is the case .
in particular , the `` bump '' in the twl multipole spectrum appears near @xmath149 , rather than @xmath150 , as for the mixed " multipole spectrum .
this paper examines the tensor perturbations of the gravitational field in a spatially flat , frw cosmology containing a mixture of radiation and dust , and shows that they may be expressed in terms of spheroidal wave functions .
although spheroidal wave functions have appeared in this context before , previous authors incorrectly determined the characteristic exponent which labels these functions . after explaining the correct method for determining the characteristic exponent , we show that spheroidal wave functions may be efficiently and accurately evaluated using standard numerical techniques .
we considered inflationary cosmological models , and used the spheroidal wave functions to find the spectrum of cbr temperature fluctuations resulting from primordial tensor ( gravitational radiation ) perturbations .
these temperature fluctuations are predicted by all inflationary models .
their existence follows from first principles : it is a consequence of the uncertainty principle and the einstein equation .
the temperature fluctuations have been previously studied by a number of authors ( including ourselves ) using a variety of approximations , and both analytic and numerical techniques . in hindsight ,
only two approximations remain in this work .
one is that the amplitude of the gravitational perturbation @xmath6 is very small .
this approximation is indeed well justified .
the second approximation is that the energy density and pressure of the universe correspond to a mixture of dust and radiation as given in ( [ mixedenergydensity]-[mixedpressure ] ) .
going back in time , this approximation is good until approximately the time of nucleosynthesis , @xmath151 seconds , when the number of effectively massless particles in the universe changed
. this should not affect the multipole moments @xmath2 which we consider , which have @xmath152 . for larger values of @xmath8 , other physical effects such as the finite thickness of the surface of last scattering may also become relevant .
it is likely that the spheroidal wave functions , which describe gravitational wave perturbations in realistic cosmological models , will find other useful applications .
we expect that the results and methods of this paper may prove applicable to a wider variety of calculations than the cbr temperatures perturbations considered here . in spatially flat frw universes , graviton mode functions in a linearized theory of gravity obey a minimally - coupled , massless , scalar wave equation @xcite like ( [ kge ] ) .
if the scale factor of the spatially flat frw universe transforms smoothly from being radiation to dust dominated , the solutions to the equation are spheroidal wave functions .
[ [ the - differential - equation - of - spheroidal - wave - functions - and - its - solutions ] ] the differential equation of spheroidal wave functions and its solutions ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ there is no generally accepted form for the differential equation of spheroidal wave functions .
we write the differential equation as @xmath153 the parameters @xmath154 , and @xmath155 and the variable @xmath156 can in general be complex . here
we take @xmath156,@xmath157 , and @xmath155 to be real .
we also consider only @xmath158 .
the differential equation ( [ diffeq ] ) has two regular singular points at @xmath159 and an irregular singular point at @xmath160 .
we only consider @xmath161 .
then the solutions of ( [ diffeq ] ) are the _ spheroidal wave functions _
@xcite @xmath162 the parameter @xmath163 is the _ order _ of the spheroidal wave function , and @xmath164 is the _
characteristic exponent_. in later sections we consider in detail the characteristic exponent @xmath164 and its relation to the order @xmath163 and the parameters @xmath165 and @xmath155 .
for now we simply note that @xmath164 is restricted so that @xmath166 for a very thorough and complete discussion of the differential equation ( [ diffeq ] ) and the solutions ( [ solutions ] ) see reference @xcite .
the spheroidal wave functions can be expressed in terms of more familiar special functions . if @xmath167 then the differential equation ( [ diffeq ] ) reduces to legendre s differential equation , suggesting that the spheroidal wave functions can be expressed in terms of legendre functions @xcite . for @xmath161 , however , it is more useful to express the spheroidal wave functions as infinite sums of bessel functions . for @xmath168
one can write the spheroidal functions as @xmath169 where @xmath170 is the infinite sum @xmath171 the expansion coefficients @xmath172 and the normalization factor @xmath173 , which are the same for any @xmath174 , are discussed below .
the functions @xmath175 are proportional to bessel or hankel functions : @xmath176 the spheroidal wave function of the first kind @xmath177 and the spheroidal wave function of the second kind @xmath178 form a complete solution to the differential equation ( [ diffeq ] ) . because the hankel functions @xmath179 can be written as linear combinations of bessel functions ( see equation ( 3.86 ) in reference @xcite ) , the spheroidal functions of the third @xmath180 and fourth @xmath181 kind can be written as linear combinations of spheroidal functions of the first and second kind : @xmath182 the spheroidal functions of the third and fourth kind also form a complete solution to the spheroidal differential equation .
the expansions ( [ sdefinition ] ) of the spheroidal wave functions in terms of bessel and hankel functions are only useful if the infinite sums ( [ tdefinition ] ) converge .
the convergence of the infinite sums depends on the expansion coefficients @xmath183 .
substituting ( [ sdefinition ] ) and ( [ tdefinition ] ) into the differential equation ( [ diffeq ] ) yields a three - term recurrence relation that the expansion coefficients must satisfy .
the recurrence relation can be written as @xmath184 where @xmath185 the solution to this recurrence relation , as well as the convergence of the infinite sums ( [ tdefinition ] ) , depends critically on the parameters @xmath186 , and @xmath155 , and is discussed below .
once a solution to the recurrence relation ( for which the infinite sums ( [ tdefinition ] ) converge ) is obtained , the normalization factor @xmath173 is given by @xmath187^{-1}. \label{snormalization}\ ] ] this normalization is chosen so that in the limit as @xmath156 becomes very large @xmath188=1 . \label{latetime}\ ] ] this relation and many more details of the solutions ( [ sdefinition ] ) are developed in reference @xcite .
although the order @xmath163 of the spheroidal wave function , along with the parameters @xmath155 and @xmath165 , appears directly in the differential equation ( [ diffeq ] ) , the characteristic exponent @xmath164 does not . in most investigations and applications of spheroidal wave functions
@xcite , however , the parameter @xmath165 is left unfixed ; one _ assumes _ a ( typically integer ) value for @xmath164 and considers @xmath165 to be a function of @xmath189 , and @xmath155 and writes @xmath190 @xmath191 is often referred to as an eigenvalue , especially when considering spheroidal wave functions as solutions to the three - dimensional wave equation @xcite . for a given choice of the parameters @xmath189 , and @xmath155 ,
the eigenvalue @xmath191 is that value of @xmath165 for which the recurrence relation ( [ recurrencerelation ] ) has a _ minimal _ solution .
a minimal solution , roughly speaking , is a set of coefficients @xmath172 that satisfy the recurrence relation _ and _ fall off for large @xmath192 @xcite .
dominant _ solution is a set of coefficients that satisfy the recurrence relation but do not fall off .
if the solution to the recurrence relation is a dominant solution , the coefficients @xmath172 do not fall off for large @xmath192 , and the infinite sums in ( [ tdefinition ] ) may not converge . in general a three - term recurrence relation
will have two independent solutions , much like a second - order , ordinary differential equation .
however , neither of the two solutions , nor any linear combination of the two solutions , need be a minimal solution @xcite . for a given set of parameters
@xmath189 , and @xmath155 , a minimal solution to the recurrence relation ( [ recurrencerelation ] ) exists only for a single , discrete value of @xmath165 .
that value for which the minimal solution exists is the eigenvalue @xmath191 . in our problem , we are given @xmath193 , and @xmath163 .
one can find @xmath164 ( modulo an integer ) by requiring that the recurrence relation ( [ recurrencerelation ] ) have a minimal solution .
the functional relationship between the parameters @xmath194 and the eigenvalue @xmath195 is complicated .
it has been shown @xcite that the functional relationship can be expressed as @xmath196 where a closed , analytic form for the function @xmath197 is usually unattainable .
the relation ( [ constraint ] ) only determines the characteristic exponent @xmath164 ( as a function of @xmath157 , and @xmath155 ) up to an integer ; a second constraint @xcite fixes @xmath164 : @xmath198 in section [ mixed ] the special case of the differential equation ( [ diffeq ] ) with @xmath199 is considered . the constraint ( [ constraint2 ] ) , along with the condition ( [ nohalfinteger ] ) that @xmath164 not be a half - integer , fixes @xmath164 so that @xmath200 for investigations of some of the analytic properties of @xmath201 see reference @xcite .
a more practical method for finding the functional relation between the parameters @xmath194 , and @xmath165 is discussed in section [ characteristicexponent ] .
the special case of the differential equation ( [ diffeq ] ) with @xmath202 and @xmath199 is of cosmological interest , and has been previously studied by sahni @xcite and nariai @xcite , who incorrectly take the solution to be @xmath203 . in spatially flat frw universes containing a mixture of dust and radiation , one may cast the wave equation for graviton mode functions in the form of the spheroidal wave function differential equation , as shown in section [ mixed ] . the differential equation
the graviton mode function obeys ( [ diffeq2 ] ) is the special case of the differential equation ( [ diffeq ] ) with @xmath202 , @xmath199 , and @xmath155 arbitrary .
( @xmath155 is arbitrary since in section [ mixed ] it is proportional to the wave number of a graviton mode function , and we desire expressions for the mode functions that are valid for an interesting range of wavenumbers . ) even if one takes @xmath165 in ( [ diffeq ] ) to be fixed and does not consider the eigenvalue problem for @xmath165 , the arguments and conclusion in the previous paragraph are still valid , especially the constraint ( [ constraint ] ) ; if @xmath204 and @xmath155 do not satisfy ( [ constraint ] ) a minimal solution to the recurrence relation does not exist and the infinite sums in ( [ tdefinition ] ) do not converge . if @xmath199 , @xmath202 , and @xmath155 is arbitrary , then the characteristic exponent @xmath164 is determined by the constraint ( [ constraint ] ) with @xmath199 and @xmath202 .
tables of eigenvalues @xmath191 for different values of @xmath189 , and @xmath155 have been published@xcite , however , and from these one can determine that the solution to ( [ constraint ] ) for @xmath202 , @xmath199 , and @xmath155 arbitrary is _ not _
@xmath205 , ie .
@xmath206 so for arbitrary @xmath155 the solution to the differential equation ( [ diffeq ] ) with @xmath202 and @xmath199 is _ not _ the spheroidal wave function ( [ solutions ] ) with @xmath202 and @xmath205 , ie .
@xmath207 this subtle point is missed in @xcite , where the solution to the spheroidal wave function differential equation with @xmath202 , @xmath208 and @xmath155 arbitrary is incorrectly given as @xmath203 .
the notation introduced above for the spheroidal wave functions ( [ solutions ] ) is most useful when the order @xmath163 and the characteristic exponent @xmath164 are fixed , and the eigenvalue @xmath195 is considered as a function of @xmath209 and @xmath155 . since we are most interested in the case when @xmath163 and @xmath165 are fixed , it is more descriptive to denote the characteristic exponent as @xmath210 .
this convention , however , would have one denote the spheroidal wave functions as @xmath211 , which is cluttered and inconvenient . in section [ mixed ] , where @xmath155 is itself a function ( of wavenumber )
, the notation would be even more unpleasant .
for this reason we adopt a new notation for the spheroidal wave functions . the solution to the differential equation ( [ diffeq ] ) with @xmath163 and @xmath165 fixed
is denoted @xmath212 the function on the left - hand side above is , of course , the same function as on the right - hand side , expressed using a different notation .
the indices @xmath163 and @xmath165 are those that appear in the differential equation , and again @xmath213 . in terms of bessel functions ,
the spheroidal wave function is now written as @xmath214 where @xmath215 with @xmath210 , and the @xmath216 are the same functions given in ( [ psidefinition ] ) .
the expansion coefficients @xmath217 are the same as those in ( [ tdefinition ] ) and obey the same recurrence relation ( [ recurrencerelation ] ) with the obvious change in notation .
the normalization factor @xmath218 is also the same as in ( [ tdefinition ] ) with the obvious change in notation .
we use this notation in sections [ mixed]-[analyticalresults ] , and the rest of this appendix .
two relations for the spheroidal wave functions are especially useful in section [ mixed ] .
the first involves the spheroidal wave functions of the third and fourth kind , and is obvious from the relations ( [ s3s4definition ] ) : @xmath219^*. \label{complexconjugate}\ ] ] here a @xmath220 denotes complex conjugation .
the second relation is the wronskian for the spheroidal wave functions of the third and fourth kind : @xmath221 \equiv { ^4 \s^\mu_\lambda}(z,\theta){d\over dz}{^3 \s^\mu_\lambda}(z,\theta)- { ^3 \s^\mu_\lambda}(z,\theta){d\over dz}{^4 \s^\mu_\lambda}(z,\theta ) = { i\over \theta^{1/2}(z^2 - 1)}. \label{wronskian}\ ] ] this relation , along with related results , is derived in reference @xcite ( see especially section 3.65 , equation ( 53 ) ) .
the constraints ( [ constraint ] ) and ( [ constraint2 ] ) determine @xmath222 for a given choice of @xmath154 , and @xmath155 . in practice , however , it is difficult to use these constraints since a closed , analytic form for the the function @xmath197 is usually not known .
recall , though , that the constraint ( [ constraint ] ) is equivalent to finding a minimal solution to the recurrence relation ( [ recurrencerelation ] ) .
a minimal solution exists only for the single , discrete value of @xmath222 that satisfies the constraints ( [ constraint ] ) and ( [ constraint2 ] ) .
one can use the theory of continued fractions to find the minimal solution to the recurrence relation and determine @xmath222 . divide ( [ recurrencerelation ] ) by @xmath223 and consider the infinite continued fraction @xmath224 defined for @xmath225 as @xmath226 where we have again suppressed for clarity the indices @xmath163 and @xmath165 as well as the argument @xmath155 . for a given choice of parameters @xmath227 , and @xmath164
the infinite continued fraction can be used to find the ratios @xmath228 of all the expansion coefficients , for @xmath225 , _ if the continued fraction converges_. there is no loss of generality if one assumes @xmath229 @xcite , ( note that this is compensated for in the normalization @xmath230 ) so the continued fraction ( if it converges ) can be used to find the expansion coefficients @xmath231 for @xmath225 .
a second infinite continued fraction for @xmath232 can also be derived from the recurrence relation .
we define the continued fraction @xmath233 assuming @xmath229 , the continued fraction ( if it converges ) can be used to find the expansion coefficients @xmath231 for @xmath232 .
whether or not the infinite continued fractions converge depends on the existence of minimal solutions to the recurrence relation .
theorem @xcite tells us that the infinite continued fraction @xmath234 converges for @xmath225 if and only if the recurrence relation has a _ minimal _ solution for @xmath225 .
further , if the infinite continued fraction does converge , it converges to a minimal solution .
likewise , @xmath235 converges to a minimal solution for @xmath232 if and only if the recurrence relation has a minimal solution for @xmath232 . a subtle , but important , point is that @xmath234 and @xmath235 may converge _ to different minimal solutions_. given a set of parameters @xmath227 and @xmath164 , the continued fraction ( [ fraction1 ] ) ( with @xmath229 ) may converge and determine a set of coefficients @xmath236 .
these coefficients will satisfy the recurrence relation for @xmath225 and will fall off for large @xmath237 . likewise , for the same set of parameters , the continued fraction ( [ fraction2 ] ) may converge and determine a set of coefficients @xmath238 , which will satisfy the recurrence relation for @xmath232 and will fall of for large @xmath192 . the recurrence relation for @xmath239 , however , _ may not be satisfied_. this is because the @xmath239 recurrence relation @xmath240 is not explicitly solved when calculating either the @xmath234 or the @xmath235 .
one can see this by examining the continued fractions ( [ fraction1 ] ) and ( [ fraction2 ] ) ; the three coefficients @xmath241 , and @xmath242 do not appear together anywhere in ( [ fraction1 ] ) and ( [ fraction2 ] ) for any @xmath237 .
so both the set of expansion coefficients @xmath236 found using @xmath234 and the set @xmath238 found using @xmath235 are minimal solutions for a certain range of the index @xmath237 , but not for all @xmath237 .
in order for the recurrence relation ( [ recurrencerelation ] ) to be satisfied it must be true for _ all _ @xmath237 . to ensure this
, one must match the two solutions found using the continued fractions .
this is accomplished by requiring that @xmath243 which is just the requirement that the recurrence relation be satisfied for @xmath239 .
so finding the minimal solution to the recurrence relation , and hence the characteristic exponent @xmath222 , is equivalent to finding the root of the the function @xmath244 defined in ( [ zdefinition ] ) . given a test value for @xmath222 ,
one calculates the continued fractions @xmath245 and @xmath246 , and the rational functions @xmath247 , and @xmath248 , to find @xmath244 .
if @xmath249 is not zero , one modifies the test value for @xmath222 using whichever root - finding algorithm one prefers . in practice , this method for determining @xmath222 is efficient .
figure [ figure4 ] shows the function @xmath250 ( of interest for the cosmological case ) for @xmath251 and @xmath252 .
this plot is typical for @xmath155 in the range @xmath253 ; the root @xmath222 is always located between two singularities , and @xmath254 is positive for @xmath255 and negative for @xmath256 where @xmath257 .
this assists implementing a root - finding algorithm to find the zeroes of @xmath250 for arbitrary @xmath155 .
figure [ figure5 ] shows the characteristic exponent @xmath258 for @xmath259 .
once the characteristic exponent @xmath222 is determined , one can use the continued fractions ( [ fraction1 ] ) and ( [ fraction2 ] ) , along with @xmath260 to calculate the remaining expansion coefficients .
since these coefficients are the minimal solution , the coefficients fall off for large @xmath192 .
one can then use ( [ snormalization ] ) to find @xmath218 . using ( [ sdefinition ] ) and
( [ tdefinition ] ) , along with an algorithm for computing bessel and hankel functions , one can evaluate the spheroidal wave functions .
further details of the numerical techniques used may be found in appendix [ appendixb ] .
figure [ figure6 ] shows the spheroidal wave functions @xmath261 and @xmath262 as functions of @xmath156 for two values of @xmath155 .
this appendix briefly describes the numerical techniques used to obtain the results in section [ numericalresults ] .
one may separate the problem of numerically evaluating the spheroidal wave functions @xmath263 and @xmath262 into two parts .
the first is to determine the characteristic exponent @xmath258 for a given @xmath155 .
as noted in section [ characteristicexponent ] the characteristic exponent @xmath258 is that value of @xmath164 for which the function @xmath250 vanishes , where @xmath264 to find the root of @xmath250 ( and hence the characteristic exponent @xmath258 ) one must evaluate the five functions on the right - hand side of ( [ zdefinition2 ] ) . from ( [ abcdefinition ] )
one can see that the three functions @xmath265 , and @xmath266 ( here @xmath199 , _ not _ @xmath164 ) are rational functions of @xmath164 for any @xmath155 , and are easily evaluated .
the continued fractions @xmath267 and @xmath268 , defined by ( [ fraction1 ] ) and ( [ fraction2 ] ) , are evaluated using the modified lentz s method ( see section 5.2 in reference @xcite ) .
both continued fractions usually converge within 10 iterations when evaluated using this method .
the root of @xmath250 is found using a simple bisection method .
although bisection may not be as efficient as other methods , it has the advantage that it is guaranteed to work once the root has been bracketed .
this is helpful since we are interested in finding the root @xmath258 of @xmath250 for many different @xmath155 ; for some @xmath155 the root @xmath258 lies _ very _ close to a singularity of @xmath250 , and in these cases other root - finding methods may not converge ( see section 9.2 in reference @xcite ) . once the characteristic exponent @xmath258 has been calculated , the remaining problem is to calculate the sums ( [ newtdefinition ] ) of expansion coefficients @xmath269 times bessel functions @xmath270 and @xmath271 .
the expansion coefficients are calculated using the continued fractions @xmath272 and @xmath273 . with @xmath274 ,
one has @xmath275 , @xmath276 , and in general @xmath277 likewise , the negative index expansion coefficients are given by @xmath278 the continued fractions @xmath272 and @xmath273 are calculated using the modified lentz s method as noted above .
the normalization @xmath279 , defined in ( [ snormalization ] ) , is calculated at the same time as the expansion coefficients .
the bessel functions are calculated most efficiently using recurrence relations .
the functions @xmath270 and @xmath271 with @xmath280 and @xmath281 are handled separately .
for @xmath280 , @xmath270 and @xmath282 are calculated for some large @xmath237 using the routine bessjy ( ) found in reference @xcite .
using the recurrence relations for bessel functions ( see 3.87 and 3.88 in @xcite ) , the @xmath270 are calculated using downward recursion to @xmath239 ( this is the direction in which the recursion is stable for bessel functions of the first kind @xcite ) . a similar procedure , using upward recursion , is used for the bessel functions of the second kind @xmath271 .
this gives the necessary bessel functions for @xmath280 .
the reflection formulae for bessel functions are used to calculate the bessel functions for @xmath283 . since the index @xmath284 the same procedure outlined above is used for @xmath285 and @xmath286 , and then the reflection formulae ( see 6.7.19 of @xcite ) @xmath287 are used to find @xmath270 and @xmath271 for @xmath283 .
so all the bessel functions needed are calculated with only a few time - consuming calls to the routine bessjy ( ) .
once the expansion coefficients @xmath269 and the bessel functions @xmath270 and @xmath271 are tabulated , the sums in ( [ newtdefinition ] ) are calculated .
care must be taken when terminating the sums .
although the expansion coefficients @xmath269 fall off _ very _ quickly as @xmath192 becomes large , the products @xmath288 and @xmath289 may not fall off as fast .
we terminate the sums when the _ products _ of the expansion coefficients and bessel functions no longer contribute ( at double - precision accuracy ) to the sums .
the primary numerical technique used to evaluate the multipole moments ( [ moments ] ) is numerical integration .
both the integral over @xmath92 in ( [ moments ] ) and the integral over @xmath101 in ( [ gldefinition ] ) were done using a fifth order embedded runge - kutta - fehlberg algorithm with adaptive stepsize control @xcite .
although formally the upper limit of the integral over @xmath92 extends to infinity , we only integrated until the remaining contribution became negligible .
the bessel function with index @xmath290 was evaluated with the routine bessjy ( ) .
bruce allen and scott koranda , phys .
d * 50 * , 3713 ( 1994 ) . note that the caption for table i of this reference should indicate that the multipoles listed in the table have been divided by @xmath134 .
also note that the definition of @xmath132 given in the caption of figure 2 contains a spurious factor of @xmath291 ; the correct definition is given by ( [ mldefinition ] ) of this paper . c. flammer , _ spheroidal wave functions _
( stanford university press , stanford , ca , 1957 ) ; j.a .
stratton , et .
al . , _ spheroidal wave functions , including tables of separation constants and coefficients _ ( published jointly by the technology press of m.i.t . and wiley and sons , new york , 1956 ) . | we examine stochastic temperature fluctuations of the cosmic background radiation ( cbr ) arising via the sachs - wolfe effect from gravitational wave perturbations produced in the early universe .
we consider spatially flat , perturbed frw models that begin with an inflationary phase , followed by a mixed phase containing both radiation and dust .
the scale factor during the mixed phase takes the form @xmath0 , where @xmath1 are constants . during the mixed phase the universe smoothly transforms from being radiation to dust dominated .
we find analytic expressions for the graviton mode function during the mixed phase in terms of spheroidal wave functions .
this mode function is used to find an analytic expression for the multipole moments @xmath2 of the two - point angular correlation function @xmath3 for the cbr anisotropy . the analytic expression for the multipole moments
is written in terms of two integrals , which are evaluated numerically .
the results are compared to multipoles calculated for models that are _ completely _ dust dominated at last - scattering .
we find that the multipoles @xmath4 of the cbr temperature perturbations for @xmath5 are significantly larger for a universe that contains both radiation and dust at last - scattering .
we compare our results with recent , similar numerical work and find good agreement .
the spheroidal wave functions may have applications to other problems of cosmological interest . |
many concepts in geometric group theory , including hyperbolic groups , @xmath0 groups , and biautomatic groups , were developed to capture the geometric and computational properties of examples such as negatively curved closed riemannian manifolds and closed topological @xmath1-manifolds . while the geometric and computational aspects of hyperbolic groups are closely interconnected , the relationship between the geometrically defined class of @xmath0 groups and the computationally defined class of biautomatic groups is much less clear .
examples of biautomatic groups that are not @xmath0 groups are known and there is an example of a @xmath2-dimensional piecewise euclidean @xmath0 group that is conjectured to be neither automatic nor biautomatic @xcite .
nevertheless , several specific classes of @xmath0 groups are known to be biautomatic , beginning with the following results by gersten and short @xcite .
[ thm : gs ] every compact , nonpositively curved triangle complex has a biautomatic fundamental group .
similarly , every compact , nonpositively curved square complex has a biautomatic fundamental group .
we conjecture that this result extends to nonpositively curved complexes built out a mixture of triangles and squares .
[ conj : biaut ] the fundamental group of a compact nonpositively curved triangle - square complex is biautomatic . as progress towards answering conjecture [ conj : biaut ] in the affirmative ,
we establish that for every compact nonpositively curved triangle - square complex @xmath3 there exists a canonically defined language of geodesics that reduces to the regular languages used by gersten and short when @xmath3 is a triangle complex or a square complex . on the other hand , by investigating the canonical language of geodesics within a single flat plane , we highlight one reason why the `` mixed '' case studied here is significantly more difficult than the `` pure '' cases analyzed by gersten and short .
the article is structured as follows .
the early sections review the necessary results about piecewise euclidean complexes , nonpositively curved spaces , biautomatic groups , and disc diagrams .
sections [ sec : geos ] and [ sec : intervals ] contain general results about combinatorial geodesics in @xmath0 triangle - square complexes and section [ sec : gs - geos ] constructs the collection of canonical geodesic paths alluded to above
. sections [ sec : flats ] , [ sec : flat - geos ] and [ sec : periodic - flats ] investigate the behavior of these geodesics in a single triangle - square flat .
finally , section [ sec : biaut ] outlines the remaining steps needed to establish conjecture [ conj : biaut ] .
we begin by reviewing the theory of nonpositively curved metric spaces built out of euclidean polytopes .
although most of the article focuses on @xmath2-dimensional complexes where simplified definitions are available , the general definitions are given since higher dimensions occasionally occur .
see @xcite for additional details . a _ euclidean polytope _
@xmath4 is the convex hull of a finite set of points in a euclidean space , or , equivalently , it is a bounded intersection of a finite number of closed half - spaces .
a _ proper face _ of @xmath4 is a nonempty subset of @xmath4 that lies in the boundary of a closed half - space containing @xmath4 .
it turns out that every proper face of a polytope is itself a polytope .
there are also two trivial faces : the empty face @xmath5 and @xmath4 itself .
the _ interior _ of a face @xmath6 is the collection of its points that do not belong to a proper subface , and every polytope is a disjoint union of the interiors of its faces .
the _ dimension _ of a face @xmath6 is the dimension of the smallest affine subspace that containing @xmath6 .
a @xmath7-dimensional face is a _ vertex _ and a @xmath8-dimensional face is an _
edge_. a @xmath2-dimensional polytope is a _
polygon _ , and an _ @xmath9-gon
_ if it has @xmath9 vertices .
a _ piecewise euclidean complex _ ( or _ complex _ ) is the regular cell complex that results when a disjoint union of euclidean polytopes are glued together via isometric identifications of their faces .
a basic result due
to bridson is that so long as there are only finitely many isometry types of polytopes used in the construction , the result is a geodesic metric space , i.e. the distance between two points in the quotient is well - defined and achieved by a path of that length connecting them .
this was a key result from bridson s thesis @xcite and is the main theorem of chapter i.7 in @xcite .
the examples under investigation are special types of @xmath2-complexes .
a _ triangle complex _ is a complex where every polytope used in its construction is isometric to an equilateral triangle with unit length sides .
note that these complexes are not necessarily simplicial since two edges of the same triangle can be identified .
they agree instead with hatcher s notion of a @xmath10-complex @xcite .
similarly , a _
square complex _ is a complex where every polytope used is a unit square and a _ triangle - square complex _ is one where every polytope used is either a unit equilateral triangle or a unit square .
euclidean polytopes and complexes have spherical analogs that are needed for our characterization of nonpositive curvature . a _
spherical polytope _ is an intersection of a finite number of closed hemispheres in @xmath11 , or , equivalently , the convex hull of a finite set of points in @xmath11 . in both cases
there is an additional requirement that the intersection or convex hull be contained in some open hemisphere of @xmath11 .
this avoids antipodal points and the non - uniqueness of geodesics connecting them . with closed hemispheres replacing closed half - spaces and lower dimensional unit subspheres replacing affine subspaces ,
the other definitions are unchanged . a _ piecewise spherical complex _ ( or _ ps
. complex _ ) is the regular cell complex that results when a disjoint union of spherical polytopes are glued together via isometric identifications of their faces .
as above , so long as there are only finitely many isometry types of polytopes used in the construction , the result is geodesic metric space .
let @xmath6 be a face of a euclidean polytope @xmath4 .
the _ link of @xmath6 in @xmath4 _ is the spherical polytope of unit vectors @xmath12 perpendicular to the affine hull of @xmath6 and pointing towards @xmath4 ( in the sense that for any point @xmath13 in the interior of @xmath6 there is a sufficiently small positive value of @xmath14 such that @xmath15 lies in @xmath4 ) .
for example , if @xmath4 is a polygon , and @xmath6 is one of its vertices , the link of @xmath6 in @xmath4 is a circular arc whose length is the radian measure of the interior angle at @xmath6 , and if @xmath6 is an edge of @xmath4 then the link of @xmath6 in @xmath4 is a single point representing one of the two unit vectors perpendicular to the affine hull of @xmath6 .
finally , when @xmath3 is a complex and @xmath6 is one of its cells , the _ link of @xmath6 in @xmath3 _ is the ps .
complex obtained by gluing together the link of @xmath6 in each of the various polytopal cells of @xmath3 containing it in the obvious way .
when @xmath6 is a vertex , its link is a ps .
complex isometric ( after rescaling ) with the subspace of points distance @xmath14 from @xmath6 .
a vertex link in a triangle - square complex is a metric graph in which every edge either has length @xmath16 or @xmath17 depending on whether it came from a triangle or a square .
we turn now to curvature conditions .
@xmath0 space _ is a geodesic metric space in which every geodesic triangle is `` thinner '' than its comparison triangle in the euclidean plane and a space is _ nonpositively curved _ if every point has a neighborhood which is @xmath0 .
a key consequence of the @xmath0 condition is that every @xmath0 space is contractible .
for a complex these conditions are characterized via geodesic loops in its links .
a _ geodesic _ in a metric space is an isometric embedding of a metric interval and a _ geodesic loop _ is an isometric embedding of a metric circle .
a _ local geodesic _ and _ local geodesic loop _ are weaker notions that only require the image curves be locally length minimizing .
for example , a path more than halfway along the equator of a @xmath2-sphere is a local geodesic but not a geodesic and a loop that travels around the equator twice is a local geodesic loop but not a geodesic loop .
a loop of length less than @xmath18 is called _
[ prop : char - npc ] a complex with only finitely many isometry types of cells fails to be nonpositively curved iff it contains a cell whose link contains a short local geodesic loop , which is true iff it contains a cell whose link contains a short geodesic loop . the first equivalence is the traditional form of gromov s link condition .
the second follows because brian bowditch proved in @xcite that when a link of a complex contains a short local geodesic loop , then some link ( possibly a different one ) contains a short geodesic loop .
see @xcite for a detail exposition of this implication .
we record one corollary for future use .
[ cor : char - npc ] let @xmath19 be a cellular map between @xmath2-complexes that is an immersion away from the @xmath7-skeleton of @xmath20 and an isometry on the interior of each cell .
if @xmath21 is nonpositively curved then so is @xmath20 . first note that in a @xmath2-complex , the vertex links are the only links that can contain short local geodesic loops .
next , by hypothesis , the vertex links of @xmath20 immerse into the vertex links of @xmath21 and , as a consequence , if @xmath20 contains a vertex @xmath22 whose link contains a short local geodesic loop , then the link of @xmath23 in @xmath21 contains a short local geodesic loop .
proposition [ prop : char - npc ] completes the proof .
[ prop : char - cat ] a complex with only finitely many isometry types of cells is @xmath0 iff it is connected , simply connected , and nonpositively curved .
this follows from the cartan - hadamard theorem : the universal cover of a complete , connected , nonpositively curved metric space is @xmath0 .
biautomaticity involves distances in cayley graphs and regular languages accepted by finite state automata .
the standard reference is @xcite . a _ path _ in a graph @xmath24
can be thought of either as an alternating sequence of vertices and incident edges , or as a combinatorial map @xmath25 from a subdivided interval @xmath26 .
every connected graph can be viewed as a metric space by making each edge isometric to the unit interval and defining the distance between two points to be the length of the shortest path connecting them . in this metric
the _ length _ of a path @xmath27 , denoted @xmath28 , is the number of edges it crosses .
define @xmath29 to be the start of @xmath27 when @xmath30 is negative , the point @xmath30 units from the start of @xmath27 when @xmath31 and the end of @xmath27 when @xmath32 .
let @xmath27 and @xmath33 be two paths in a graph @xmath24 .
there are two natural notions of distance between @xmath27 and @xmath33 .
the _ subspace distance _ between @xmath27 and @xmath33 is the smallest @xmath34 such that @xmath27 ( or more specifically its image ) is contained in a @xmath34-neighborhood of @xmath33 and @xmath33 is contained in a @xmath34-neighborhood of @xmath27 .
the _ path distance _ between @xmath27 and @xmath33 is the maximum distance between @xmath29 and @xmath35 as @xmath30 varies over all real numbers . when the path distance is bounded by @xmath34 , @xmath27 and @xmath33
are said to _
synchronously @xmath34-fellow travel_. for later use we note that when geodesics start close together , bounding their subspace distance also bounds their path distance .
[ prop : ft - geo ] if @xmath27 and @xmath33 are geodesic paths in @xmath24 that start distance @xmath36 apart , and the subspace distance between them is @xmath34 , then the path distance between them is at most @xmath37 . in particular , they synchronously @xmath38-fellow travel .
let @xmath29 be a point on @xmath27 and let @xmath39 be the closest point on @xmath33 . by hypothesis
the distance between @xmath29 and @xmath39 is at most @xmath34 .
moreover , because @xmath27 and @xmath33 are geodesics , @xmath40 and @xmath41 .
thus @xmath42 . as a result
, @xmath29 is within @xmath43 of @xmath35 .
let @xmath3 be a one - vertex cell complex with fundamental group @xmath44 and let @xmath45 index its oriented edges .
the oriented loops indexed by @xmath45 represent a generating set for @xmath44 that is closed under involution and the @xmath8-skeleton of the universal cover of @xmath3 is called the _
right cayley graph of @xmath44 with respect to @xmath45 _ and denoted @xmath46 . alternatively ,
if @xmath44 is a group and @xmath45 is generating set closed under involution , @xmath46 is a graph with vertices indexed by the elements of @xmath44 and an oriented edge connecting @xmath47 and @xmath48 labeled by @xmath49 whenever @xmath50 . actually , the oriented edges come in pairs and we add only one unoriented edge for each such pair with a label associated with each orientation .
it is the right cayley graph because the generators multiply on the right .
the second definition shows that the structure of @xmath46 is independent of @xmath3 .
there is a natural label - preserving left @xmath44-action on @xmath46 defined , depending on the definition , either by deck transformations or by sending @xmath51 to @xmath52 .
an _ alphabet _ is a set @xmath45 whose elements are _ letters_.
a _ word _ is a finite sequence of letters , and the collection of all words is denoted @xmath53
language @xmath54 over an alphabet @xmath45 _ is any subset of @xmath53 .
when @xmath45 is identified with the oriented edges of a one - vertex cell complex @xmath3 with fundamental group @xmath44 , there are natural bijections between languages over @xmath45 , collections of paths in @xmath3 , and collections of paths in the universal cover of @xmath3 that are invariant under the @xmath44-action . a language @xmath54 over @xmath45 _ maps onto @xmath44 _ if the corresponding @xmath44-invariant set of paths contains at least one path connecting every ordered pair of vertices , or equivalently contains a word representing every element of @xmath44 .
two key properties of languages are used to define biautomatic groups .
a _ finite state automaton _ is a finite directed graph with decorations .
the edges are labeled by a set @xmath45 , there is a specified _ start vertex _ , and a subset of the vertices designated as _ accept states_. concatenating the labels on the edges of any path starting at the start vertex and ending at an accept state creates an _
accepted word_. the collection of all accepted words is the _ language accepted _ by the automaton . finally , a language is _ regular _ if it is accepted by some finite state automaton .
let @xmath44 be a group generated by a finite set @xmath45 closed under inversion and let @xmath54 be a language over @xmath45 that maps onto @xmath44 .
the language @xmath54 _ fellow travels _ if there is a constant @xmath34 such that all pairs of paths in @xmath46 that start and end at most @xmath8 unit apart and corresponding to words in @xmath54 synchronously @xmath34-fellow travel .
the following characterization of biautomaticity is from ( * ? ? ?
* lemma 2.5.5 ) .
[ thm : biaut ] if @xmath44 is a group generated by a finite set @xmath45 closed under inversion and @xmath54 is a language over @xmath45 that maps onto @xmath44 , then @xmath54 is part of a biautomatic structure for @xmath44 iff @xmath54 is regular and fellow travels .
recall that a _
action of a group on a space is one that is proper , cocompact and by isometries and note that the action of @xmath44 on its cayley graph is geometric , but also free and transitive on vertices . by focusing on the paths in the cayley graph corresponding to the language @xmath54 and slightly modifying a few definitions
, jacek wiatkowski was able in @xcite to extend the characterization of biautomaticity given in theorem [ thm : biaut ] to groups merely acting geometrically on a connected graph .
let @xmath44 be a group acting geometrically on a connected graph @xmath21 , and let @xmath55 be a @xmath44-invariant collection of paths in @xmath21 .
the set @xmath55 is said to _ map onto @xmath44 _ if every path start and ends in the @xmath44-orbit of a particular vertex @xmath22 and every ordered pair of vertices in the @xmath44-orbit of @xmath22 are connected by at least one path in @xmath55 .
the @xmath44-invariant paths in @xmath55 correspond to a set of paths in a finite graph ( with loops when edges are inverted ) and we say @xmath55 is _ regular _ if the set of paths over the finite alphabet of oriented edges in the quotient is regular in the usual sense .
finally , the fellow traveler property is modified to consider pairs of paths that start and end within @xmath36 units of each other where @xmath36 has been chosen large enough so that any two vertices in the @xmath44-orbit of @xmath22 are connected by a sequence of other vertices in the orbit with adjacent vertices in the sequence at most @xmath36 units apart . when the @xmath44-orbit is vertex transitive , @xmath56 is sufficient .
[ thm : group ] let @xmath44 be a group acting geometrically on a connected graph @xmath21 , and let @xmath55 be a @xmath44-invariant collection of paths in @xmath21 that maps onto @xmath44 . if @xmath55 is regular and fellow travels in the modified sense described above , then @xmath44 is biautomatic .
the curvature of a @xmath0 @xmath2-complex impacts the distances between paths in its @xmath8-skeleton via the reduced disc diagrams that fill its closed loops . a _ disc diagram _
@xmath57 is a contractible @xmath2-complex with an implicit embedding into the euclidean plane and a _ nonsingular disc diagram _ is a @xmath2-complex homeomorphic the closed unit disc @xmath58 .
disc diagrams that are not nonsingular are called _ singular _ and the main difference between the two is that singular disc diagrams contain _ cut points _ , i.e. points whose removal disconnects the diagram . the boundary of @xmath57 is a combinatorial loop that traces the outside of @xmath57 in a clockwise fashion . when @xmath57 is nonsingular , this is clear and when @xmath57 is singular amibiguities are resolved by the planar embedding .
the boundary cycle is the essentially unique loop in @xmath57 that can be homotoped into @xmath59 while moving each point an arbitrarily small distance .
let @xmath57 be a disc diagram and let @xmath3 be a regular cell complex .
we say that a cellular map @xmath60 turns @xmath57 into a _ disc diagram over @xmath3 _ and its boundary cycle is the image of the boundary cycle of @xmath57 under the map @xmath61 . when a loop @xmath27 in @xmath3 is the boundary cycle of a disc diagram @xmath57 over @xmath3 we say that @xmath27 _ bounds _ @xmath57 and that @xmath57 _ fills _ @xmath27 .
let @xmath60 be a disc diagram over a regular cell complex @xmath3 . when there is a pair of @xmath2-cells in @xmath57 sharing an edge @xmath62 such that they are folded over along @xmath62 and identified under the map @xmath61 , we say @xmath57 contains a _
cancellable pair _ and a disc diagram over @xmath3 with no cancellable pairs is said to be _
reduced_. note that a disc diagram over @xmath3 is a reduced disc diagram over @xmath3 iff the map @xmath60 is an immersion away from the @xmath7-skeleton of @xmath57 , which is true iff for every vertex @xmath22 in @xmath57 , the induced map from the link of @xmath22 to the link of @xmath23 is an immersion .
the importance of reduced disc diagrams over @xmath2-complexes derives from the following classical result . for a proof see @xcite .
[ thm : vk ] every combinatorial loop in a simply - connected regular cell complex bounds a reduced disc diagram . in @xmath2-dimensions ,
curvature conditions on the cell complex @xmath3 propagate to reduced disc diagrams over @xmath3 .
[ cor : vk ] every combinatorial loop in a @xmath0 @xmath2-complex bounds a @xmath0 disc diagram . let @xmath27 be a combinatorial loop in a @xmath0 @xmath2-complex @xmath3 .
since @xmath0 spaces are contractible , @xmath27 bounds a reduced disc diagram @xmath57 by theorem [ thm : vk ] . by pulling back the metric of @xmath3 to @xmath57
, we can turn it into a disc diagram .
next , by corollary [ cor : char - npc ] , @xmath57 is nonpositively curved . and
finally , by proposition [ prop : char - cat ] @xmath57 is @xmath0 .
the fine structure of a @xmath0 disc diagram is highlighted when we focus on its local curvature contributions .
an _ angled @xmath2-complex _ is a @xmath2-complex @xmath3 where the corner of every polygon is assigned a number .
when @xmath3 is , this number is naturally the measure of its angle , or equivalently the length of the spherical arc that is the link of this vertex in this polygon .
the curvature of a @xmath9-gon in an angled @xmath2-complex @xmath3 is the sum of its angles minus the expected euclidean angle sum of @xmath63 .
when @xmath3 is , its polygon curvatures are all @xmath7 .
the curvature of a vertex @xmath22 in an angled @xmath2-complex is @xmath18 plus @xmath64 times the euler characteristic of its link minus the sum of angles sharing @xmath22 as a vertex . when the angled @xmath2-complex is a disc diagram @xmath57 this reduces to @xmath18 minus the angle sum for interior vertices and @xmath64 minus the angle sum for boundary vertices that are not cut points .
note that when @xmath57 is and nonpositively curved , the curvature of an interior vertex is nonpositive .
[ thm : cbg ] if @xmath3 is an angled @xmath2-complex then the sum of the vertex curvatures and the polygon curvatures is @xmath18 times the euler characteristic of @xmath3 .
a proof of this relatively elementary result can be found in @xcite .
the basic idea is that every assigned angle contributes to exactly one vertex curvature and exactly one polygon curvature but with opposite signs .
the curvature sum is thus independent of the angles assigned and minor bookkeeping equates it with @xmath18 times the euler characteristic .
the main result we need is a easy corollary .
[ cor : disc ] if @xmath57 is a @xmath0 disc diagram , then the sum of its boundary vertex curvatures is at least @xmath18 .
this follows from theorem [ thm : cbg ] once we note that being a disc diagram implies its euler characteristic is @xmath8 , implies its face curvatures are @xmath7 and @xmath0 implies its interior vertex curvatures are nonnegative .
in this section we use disc diagrams to turn an arbitrary path into a geodesic inside a @xmath0 triangle - square complex .
this procedure uses moves that modify paths by pushing across simple diagrams . after introducing these diagrams and moves ,
we describe the procedure .
[ def : doubly - based ] a _ doubly - based disc diagram _ is a nonsingular disc diagram @xmath57 with two distinquished boundary vertices @xmath12 and @xmath22 .
an example is shown in figure [ fig : sample - disc ] .
the vertex @xmath12 is its _ start vertex _ and @xmath22 is its _
end vertex_. when @xmath12 and @xmath22 are distinct there are two paths in the boundary cycle of @xmath57 that start at @xmath12 and end at @xmath22 .
the one that proceeds clockwise along the boundary is the _ old path _ and the one that proceeds counter - clockwise along the boundary is _
new path_. when illustrating doubly - based diagrams , we place @xmath12 on the left and @xmath22 on the right so that the old path travels along the top and the new path travels along the bottom .
the old path / new path terminology is extended to the case @xmath65 by defining the old path as the clockwise boundary cycle starting and ending at @xmath65 and the new path as the trivial path at @xmath65 .
disc diagram.[fig : sample - disc ] ] [ def : moves ] let @xmath66 be a path in a @xmath2-complex @xmath3 and let @xmath57 be a doubly - based disc diagram . if there is a map @xmath67 that sends the old path to a directed subpath of @xmath66 , then _ applying the move @xmath57 _ means replacing the image of the old path in @xmath66 with the image of the new path to create an altered path @xmath68 .
this can be thought of as homotoping @xmath66 across the image of @xmath57 .
the move is _ length - preserving _ when the old and new paths have the same length and _ length - reducing _ when the new path is strictly shorter .
the moves we are interested in are quite specific .
[ def : basic - moves ] we define five basic types of moves over a triangle - square complex that involve a single triangle , a single square , a pair of triangles , a finite row of squares with a triangle on either end , or a single edge .
we call these _ triangle moves _ , _ square moves _ , _ triangle - triangle moves _ , _ triangle - square - triangle moves _ , and _ trivial moves _ , respectively .
diagrams illustrating the first four are shown in figure [ fig : moves ] .
note that for each type of move the old path is at least as long as the new path .
the triangle move is length - reducing .
the start and end vertices for the square move are non - adjacent making it length - preserving .
the start and end vertices of a triangle - triangle move or a triangle - square - triangle move are the unique vertices of valence @xmath2 and they are also length - preserving . note that a triangle - square - triangle move can have any number of squares and that a triangle - triangle move can be viewed as _ degenerate triangle - square - triangle _ move with zero squares . finally , the trivial move ( defined below ) is not a move in the sense of definition [ def : moves ] , but it is included for convenience .
let @xmath57 be a diagram consisting of a single edge and identify both @xmath12 and @xmath22 with one of its vertices .
if we view the path of length @xmath2 starting and ending at @xmath65 as the old path and the trivial path at @xmath65 as the new path , then we can view the removal of a `` backtrack '' in a path @xmath66 as applying a trivial move across @xmath57 . because the only moves considered from this point forward are the basic moves described above , we drop the adjective .
thus a move now refers to one of these five basic types .
notice that for every move ( in this revised sense ) , the subspace distance between the old and new paths is @xmath8 since every vertex lies in one of the paths and , if not in both , it is connected to a vertex in the other path by an edge .
cccc [ cols="^ " , ] by theorem [ thm : flat - embedding ] geodesics in the image of a flat correspond exactly with geodesics in the domain , making the following corollary immediate .
[ cor : interval - images ] let @xmath6 be a colored crumpled triangle - square flat and let @xmath69 be its embedding into @xmath70 .
for all vertices @xmath12 and @xmath22 in @xmath6 with images @xmath71 and @xmath72 in @xmath3 , the image of the interval @xmath73 $ ] is the intersection of the interval @xmath74 $ ] with the embedded flat @xmath75 .
this corollary is interesting , at least in part , because intervals in @xmath76 are extremely simple in form . for all vertices @xmath12 and @xmath22 in the @xmath77-complex @xmath78 , the interval @xmath79 $ ] is a ( possibly degenerate ) @xmath77-dimensional parallelepiped .
because @xmath3 is a direct product , a vertex @xmath80 is between @xmath12 and @xmath22 iff the image of @xmath80 is between the images of @xmath12 and @xmath22 under both projections @xmath81 and @xmath82 .
in particular , the interval in @xmath3 is a direct product of two intervals in @xmath83 and these are ( possibly degenerate ) @xmath2-dimensional parallelograms . with these tools in hand ,
we now shift our attention to flats that are potentially periodic .
the self - similarity of the potentially perioidic flat means that its image under the embedding described above looks , roughly speaking , like an affine plane in @xmath84 , a description that can be made precise using the notion of a quasi - isometry .
let @xmath85 be a function between metric spaces . if there are constants @xmath86 and @xmath87 such that for all @xmath88 with images @xmath89 both @xmath90 and @xmath91 , then @xmath61 is called a _ quasi - isometric embedding _ of @xmath21 into @xmath20 . if @xmath61 is also _ quasi - onto _
( i.e. there is a constant @xmath92 such that the @xmath92-neighborhood of @xmath93 is all of @xmath20 ) then @xmath61 is called a _ quasi - isometry _ and @xmath21 and @xmath20 are _ quasi - isometric_. although it is not immediately obvious from this definition , the relation of being quasi - isometric in an equivalence relation on metric spaces .
[ prop : quasi - flat ] let @xmath6 be a colored triangle - square flat and let @xmath94 be the corresponding embedding . if @xmath6 is potentially periodic then its image @xmath75 is quasi - isometric to a flat affine plane in @xmath95 .
let @xmath12 be a vertex of @xmath6 embedded in @xmath96 . because @xmath6 is potentially periodic there are a pair of linearly independent translations from @xmath6 to itself that generate a cocompact action of @xmath97 on @xmath6
let @xmath22 and @xmath80 be the images of @xmath12 under these two maps . since the translations acting on @xmath6 do not change colors or orientations , they also induce translations of @xmath76 that preserve the image of @xmath6 .
in particular , if we let @xmath45 be the affine span in @xmath98 of the points @xmath99 , @xmath100 and @xmath101 , then each point of @xmath75 lies within a bounded distance of a point in @xmath45 ( since each point @xmath6 lies within a bounded distance of a point in the orbit of @xmath12 under the @xmath97 action ) .
thus , by choosing @xmath92 large enough , @xmath75 lies in a @xmath92-neighborhood of @xmath45 which implies that @xmath6 is quasi - isometric to @xmath45 .
finally , note that @xmath99 , @xmath100 and @xmath101 can not be collinear since euclidean planes and lines are not quasi - isometric .
thus @xmath45 is a @xmath2-dimensional affine plane . and
finally , in order to understand the behavior of gersten - short geodesic in crumpled triangle - square flats , we need to know that the direction of the next choke point can be determined locally .
[ lem : next - move ] let @xmath6 be a colored crumpled triangle - square flat , let @xmath69 be the corresponding embedding of @xmath6 into @xmath70 and let @xmath12 and @xmath22 be vertices in @xmath6 with images @xmath71 and @xmath72 in @xmath3 .
the first move of a gersten - short geodesic from @xmath12 to @xmath22 is completely determined by the vertex neighborhood of @xmath12 .
more specifically , the edges touching @xmath12 whose images lie in the parallelopiped @xmath74 $ ] span the initial cell of the first move . by theorem [ thm : first - move ] the first move is determined by its first cell which in turn is determined by the edges leaving @xmath12 that extend to geodesics connecting @xmath12 to @xmath22 and such edges are characterized by whether or not their images lie in @xmath74 $ ] .
more specifically , if an edge leaving @xmath12 is the first edge of a geodesic from @xmath12 to @xmath22 then by corollary [ cor : interval - images ] its image lies in @xmath74 $ ] .
conversely , any edge leaving @xmath12 whose image lies in @xmath74 $ ] is contained in @xmath74 \cap p(f)$ ] , by corollary [ cor : interval - images ] it is contained in @xmath102)$ ] , and thus it extends to a geodesic from @xmath12 to @xmath22 . as a consequence of local determination of the next move
, we can show that gersten - short geodesics in a flat do not cross in the following sense .
[ cor : noncrossing ] let @xmath6 be a colored crumpled triangle - square flat and let @xmath69 be the corresponding embedding of @xmath6 into @xmath70 .
if @xmath12 and @xmath22 are vertices in @xmath6 with images @xmath71 and @xmath72 in @xmath3 , then for every pair of vertices @xmath13 and @xmath103 within @xmath8 unit of @xmath12 and @xmath22 in @xmath6 , the next choke point in the gersten - short geodesic from @xmath13 to @xmath103 is locally determined by the neighborhood of @xmath13 and the straight line segment from @xmath13 to this next choke point in @xmath6 does not cross the straight line segment from @xmath12 to its next choke point . since the parallelepipeds @xmath74 $ ] and @xmath104 $ ] are nearly identical , except in situations where one of them is degenerate , we can assume by lemma [ lem : next - move ] that @xmath105 .
next , note that the edges crossed by the straight line segment from @xmath12 to its first choke point are edges with both endpoints the same distance from @xmath22 and also note that no other edges in the move have this property . because @xmath105 the edges crossed by the straight line segment from @xmath13 to its next choke point have the same property .
in particular , neither straight line segment can cross into the interior of the move containing the other .
the argument in the exceptional case is essentially the same .
combining lemma [ lem : next - move ] and corollary [ cor : noncrossing ] , we now show that image of a gersten - short geodesic in @xmath76 is approximately linear until the situation degenerates .
[ lem : quasi - linear ] let @xmath6 be a colored potentially periodic flat , let @xmath69 be the corresponding embedding of @xmath6 into @xmath70 , and let @xmath12 and @xmath22 be vertices in @xmath6 with images @xmath71 and @xmath72 in @xmath3 .
there exists a direction vector and a constant depending only on the structure of @xmath6 and the directions of the sides of the parallelepiped @xmath74 $ ] so that image under @xmath69 of any gersten - short geodesic from @xmath12 to @xmath22 stays with this constant of the line through @xmath71 with this direction until that point @xmath106 at which the interval @xmath107 $ ] degenerates into a lower dimensional parallelepiped .
the basic idea goes as follows .
pick @xmath12 and @xmath22 so that @xmath74 $ ] is far from degenerate ( i.e. all four side lengths are large ) . by lemma [ lem : next - move ]
the first choke point only depends on the neighborhood of @xmath12 and the directions of the sides of @xmath74 $ ] . repeat .
if the initial side lengths were large enough , then we pass through two vertices that lies in the same @xmath97 orbit before the remaining interval is able to degenerate .
the vector connecting these two vertices is the vector mentioned in the theorem . by lemma [ lem : next - move ] again , the same sequence of moves is now repeated until that point at which the interval remaining degenerates .
this shows that this portion of the gersten - short geodesic approximates a line in @xmath3 to within a specified constant .
finally , if we pick any other vertex @xmath13 close to @xmath12 and repeat the argument above , we find that the gersten - short geodesic from @xmath13 also has an image in @xmath3 that approximates a straight line .
because gersten - short geodesics in @xmath6 do not cross ( corollary [ cor : noncrossing ] ) the direction of this line must agree with the previous direction ( since it can be bounded on both sides by lines originating in @xmath12 and other nearby vertices in the same @xmath97 orbit as @xmath12 .
combining these results produces our final main result .
[ thm : ft - periodic ] if @xmath6 is a potentially periodic triangle - square flat , then there is a value @xmath34 such that pairs of gersten - short geodesics in @xmath6 that start and end within @xmath8 unit synchronously @xmath34-fellow travel . since this result
has already been established from pure and striped flats ( propositions [ prop : ft - pure ] and [ prop : ft - striped ] ) , we only need to consider flats that are thoroughly crumpled ( proposition [ prop : aperiodic - flats ] ) .
so assume that @xmath6 is a potentially periodic thoroughly crumpled triangle - square flat that has been colored with corresponding embedding @xmath69 into @xmath78 .
let @xmath27 and @xmath33 are gersten - short geodesics in @xmath6 that start and end within @xmath8 unit of each other .
we show that they synchronously @xmath34-fellow travel by showing that they both approximate a pair line segments in the same two directions . by lemma [ lem : quasi - linear ] , so long as the corresponding intervals in @xmath3 are not degenerate parallelopipeds , the images of @xmath27 and @xmath33 approximate straight lines in the same direction .
moreover , once the intervals do degenerate , the remaining portion must approximate the intersection of the affine plane @xmath45 approximating @xmath75 ( proposition [ prop : quasi - flat ] ) and the affine space containing the degenerate parallelopiped . because the projection maps from thoroughly crumpled planes are onto ( corollary [ cor : project - onto ] )
, the intersection of these two affine spaces is at most one - dimensional . combining the constant governing
how closely gersten - short geodesics travel along a straight line in the non - degenerate portion with the constant governing how closely @xmath75 approximates @xmath45 , yields a global constant @xmath34 that is , by construction , independent of @xmath27 and @xmath33 and only dependent on the periodic structure of @xmath6 itself .
in this final section we outline a sequence of steps which , if carried out , would fully establish that a compact nonpositively curved triangle - square complex @xmath3 has a biautomatic fundamental group ( conjecture [ conj : biaut ] ) . according to theorem [ thm : biaut ] , in order to establish the biautomaticity of @xmath108 , one needs to find a collection of paths and then to prove that they fellow travel and form a regular language .
we firmly believe that the gersten - short geodesics defined have satisfy these conditions .
next , since triangle - square complexes with no immersed flats are already known to be biautomatic ( theorem [ thm : flats - biaut ] ) , it is natural to focus attention on the flats that do exist .
a good first step would be to establish conjecture [ conj : aperiodic ] in order to rule out the existence of intrinsically aperiodic flat planes focusing attention on those potentially periodic flats that immerse into @xmath3 . for each individual potentially periodic flat plane
we have shown that there is a value @xmath34 , depending on the flat , so the gersten - short geodesics contained in the flat that start and end within @xmath8 unit of each other synchronously @xmath34-fellow travel ( theorem [ thm : ft - periodic ] ) .
the natural next step would be to prove the following : for any compact nonpositively curved triangle - square complex @xmath3 there are only finitely many distinct thoroughly crumpled planes that immerse into @xmath3 . as a consequence ,
there is a global value @xmath34 depending only on @xmath3
so that gersten - short paths in any flat inside @xmath109 that start and end within @xmath8 unit of each other synchronously @xmath34-fellow travel . once this point has been reached , one should use the fellow traveling constants within planes and the @xmath0 nature of @xmath109 overall to prove that all gersten - short paths in @xmath109 synchronously @xmath34-fellow travel for some uniform global @xmath34 .
since the constants in pure square , pure triangle and striped flats are small , and periodic crumpled flats have bounded intersection with other flats , this seems reasonable .
the necessary arguments might be similar to those used by chris hruska in @xcite . finally , once it is known that these paths fellow travel in @xmath109 , the falsification by fellow traveler property @xcite should be invoked to establish the regularity of the language they describe . | in the early 1990s steve gersten and hamish short proved that compact nonpositively curved triangle complexes have biautomatic fundamental groups and that compact nonpositively curved square complexes have biautomatic fundamental groups . in this article
we report on the extent to which results such as these extend to nonpositively curved complexes built out a mixture of triangles and squares . since both results by gersten and short have been generalized to higher dimensions , this can be viewed as a first step towards unifying januszkiewicz and wiatkowski s theory of simplicial nonpositive curvature with the theory of nonpositively curved cube complexes . |
statistical model selection in the high - dimensional regime arises in a number of applications . in many data analysis problems in geophysics , radiology , genetics , climate studies , and image processing ,
the number of samples available is comparable to or even smaller than the number of variables .
however , it is well - known that empirical statistics such as sample covariance matrices are not well - behaved when both the number of samples and the number of variables are large and comparable to each other ( see @xcite ) .
model selection in such a setting is therefore both challenging and of great interest . in order for model selection to be well - posed
given limited information , a key assumption that is often made is that the underlying model to be estimated only has _ a few degrees of freedom_. common assumptions are that the data are generated according to a graphical model , or a stationary time - series model , or a simple factor model with a few latent variables .
sometimes geometric assumptions are also made in which the data are viewed as samples drawn according to a distribution supported on a low - dimensional manifold .
a model selection problem that has received considerable attention recently is the estimation of covariance matrices in the high - dimensional setting . as the sample covariance matrix is poorly behaved in such a regime @xcite , some form of _ regularization _ of the sample covariance is adopted based on assumptions about the true underlying covariance matrix .
for example approaches based on banding the sample covariance matrix @xcite have been proposed for problems in which the variables have a natural ordering ( e.g. , times series ) , while `` permutation - invariant '' methods that use thresholding are useful when there is no natural variable ordering @xcite .
these approaches provide consistency guarantees under various sparsity assumptions on the true covariance matrix .
other techniques that have been studied include methods based on shrinkage @xcite and factor analysis @xcite .
a number of papers have studied covariance estimation in the context of _ gaussian graphical model selection_. in a gaussian graphical model the _ inverse _ of the covariance matrix , also called the concentration matrix , is assumed to be sparse , and the sparsity pattern reveals the conditional independence relations satisfied by the variables .
the model selection method usually studied in such a setting is @xmath0-regularized maximum - likelihood , with the @xmath0 penalty applied to the entries of the inverse covariance matrix to induce sparsity .
the consistency properties of such an estimator have been studied @xcite , and under suitable conditions @xcite this estimator is also `` sparsistent '' , i.e. , the estimated concentration matrix has the same sparsity pattern as the true model from which the samples are generated .
an alternative approach to @xmath0-regularized maximum - likelihood is to estimate the sparsity pattern of the concentration matrix by performing regression separately on each variable @xcite ; while such a method consistently estimates the sparsity pattern , it does not directly provide estimates of the covariance or concentration matrix . in many applications throughout science and engineering
, a challenge is that one may not have access to observations of all the relevant phenomena , i.e. , some of the relevant variables may be hidden or unobserved .
such a scenario arises in data analysis tasks in psychology , computational biology , and economics . in general latent variables
pose a significant difficulty for model selection because one may not know the number of relevant latent variables , nor the relationship between these variables and the observed variables .
typical algorithmic methods that try to get around this difficulty usually fix the number of latent variables as well as the structural relationship between latent and observed variables ( e.g. , the graphical model structure between latent and observed variables ) , and use the em algorithm to fit parameters @xcite .
this approach suffers from the problem that one optimizes non - convex functions , and thus one may get stuck in sub - optimal local minima .
an alternative method that has been suggested is based on a greedy , local , combinatorial heuristic that assigns latent variables to groups of observed variables , based on some form of clustering of the observed variables @xcite ; however , this approach has no consistency guarantees . in this paper
we study the problem of latent - variable graphical model selection in the setting where all the variables , both observed and hidden , are jointly gaussian .
more concretely let the covariance matrix of a finite collection of jointly gaussian random variables @xmath1 be denoted by @xmath2 , where @xmath3 are the observed variables and @xmath4 are the unobserved , hidden variables .
the marginal statistics corresponding to the observed variables @xmath3 are given by the marginal covariance matrix @xmath5 , which is simply a submatrix of the full covariance matrix @xmath2 .
however suppose that we parameterize our model by the concentration matrix @xmath6 , which as discussed above reveals the connection to graphical models .
in such a parametrization , the _ marginal concentration matrix _ @xmath7 corresponding to the observed variables @xmath3 is given by the schur complement @xcite with respect to the block @xmath8 : @xmath9 thus if we only observe the variables @xmath3 , we only have access to @xmath5 ( or @xmath10 ) .
the two terms that compose @xmath10 above have interesting properties .
the matrix @xmath11 specifies the concentration matrix of the _ conditional statistics _ of the observed variables given the latent variables .
if these conditional statistics are given by a sparse graphical model then @xmath11 is _
sparse_. on the other hand the matrix @xmath12 serves as a _ summary _ of the effect of marginalization over the hidden variables @xmath13 .
this matrix has small rank if the number of latent , unobserved variables @xmath13 is small relative to the number of observed variables @xmath14 ( the rank is equal to @xmath15 ) . therefore the marginal concentration matrix @xmath10 of the observed variables
@xmath3 is generally _ not sparse _ due to the additional low - rank term @xmath12 . hence standard graphical model selection techniques applied directly to the observed variables @xmath3 are not useful . a modeling paradigm that infers the effect of the latent variables @xmath4 would be more suitable in order to provide a simple explanation of the underlying statistical structure .
hence we _ decompose _
@xmath10 into the sparse and low - rank components , which reveals the conditional graphical model structure in the observed variables as well as the _ number _ of and effect due to the unobserved latent variables .
such a method can be viewed as a blend of principal component analysis and graphical modeling . in standard graphical modeling
one would directly approximate a concentration matrix by a sparse matrix in order to learn a sparse graphical model . on the other hand in principal component analysis
the goal is to explain the statistical structure underlying a set of observations using a small number of latent variables ( i.e. , approximate a covariance matrix as a low - rank matrix ) . in our framework based on decomposing a concentration matrix , we learn a graphical model among the observed variables _ conditioned _ on a few ( additional ) latent variables . notice that in our setting these latent variables are _ not _ principal components , as the conditional statistics ( conditioned on these latent variables ) are given by a graphical model .
therefore we refer to these latent variables informally as _
hidden components_. our first contribution in section [ sec : iden ] is to address the fundamental question of _ identifiability _ of such latent - variable graphical models given the marginal statistics of only the observed variables .
the critical point is that we need to tease apart the correlations induced due to marginalization over the latent variables from the conditional graphical model structure among the observed variables . as the identifiability problem is one of _ uniquely _ decomposing the sum of a sparse matrix and a low - rank matrix into the individual components , we study the algebraic varieties of sparse matrices and low - rank matrices .
an important theme in this paper is the connection between the tangent spaces to these algebraic varieties and the question of identifiability .
specifically let @xmath16 denote the tangent space at @xmath11 to the algebraic variety of sparse matrices , and let @xmath17 denote the tangent space at @xmath12 to the algebraic variety of low - rank matrices .
then the _ statistical _ question of identifiability of @xmath11 and @xmath12 given @xmath10 is determined by the _
geometric _ notion of _ transversality _ of the tangent spaces @xmath16 and @xmath17 .
the study of the transversality of these tangent spaces leads us to natural conditions for identifiability .
in particular we show that latent - variable models in which @xmath18 the sparse matrix @xmath11 has a small number of nonzeros per row / column , and @xmath19 the low - rank matrix @xmath12 has row / column spaces that are not closely aligned with the coordinate axes , are identifiable .
these two conditions have natural statistical interpretations .
the first condition ensures that there are no densely - connected subgraphs in the conditional graphical model structure among the observed variables @xmath3 given the hidden components , i.e. , that these conditional statistics are indeed specified by a sparse graphical model .
such statistical relationships may otherwise be mistakenly attributed to the effect of marginalization over some latent variable .
the second condition ensures that the effect of marginalization over the latent variables is `` spread out '' over many observed variables ; thus , the effect of marginalization over a latent variable is not confused with the conditional graphical model structure among the observed variables .
in fact the first condition is often assumed in some papers on standard graphical model selection without latent variables ( see for example @xcite ) .
we note here that question of parameter identifiability was recently studied for models with discrete - valued latent variables ( i.e. , mixture models , hidden markov models ) @xcite .
however , this work is not applicable to our setting in which both the latent and observed variables are assumed to be jointly gaussian . as our next contribution
we propose a _
regularized maximum - likelihood decomposition _
framework to approximate a given sample covariance matrix by a model in which the concentration matrix decomposes into a sparse matrix and a low - rank matrix .
a number of papers over the last several years have suggested that heuristics based on using the @xmath0 norm are very effective for recovering sparse models @xcite .
indeed such heuristics have been effectively used , as described above , for model selection when the goal is to estimate sparse concentration matrices . in her thesis
@xcite fazel suggested a convex heuristic based on the nuclear norm for rank - minimization problems in order to recover low - rank matrices .
this method generalized the previously studied trace heuristic for recovering low - rank positive semidefinite matrices .
recently several conditions have been given under which these heuristics provably recover low - rank matrices in various settings @xcite .
motivated by the success of these heuristics , we propose the following penalized likelihood method given a sample covariance matrix @xmath20 formed from @xmath21 samples of the observed variables : @xmath22 here @xmath23 represents the gaussian log - likelihood function and is given by @xmath24 for @xmath25 , where @xmath26 is the trace of a matrix and @xmath27 is the determinant .
the matrix @xmath28 provides an estimate of @xmath11 , which represents the conditional concentration matrix of the observed variables ; the matrix @xmath29 provides an estimate of @xmath12 , which represents the effect of marginalization over the latent variables .
notice that the regularization function is a combination of the @xmath0 norm applied to @xmath30 and the nuclear norm applied to @xmath31 ( the nuclear norm reduces to the trace over the cone of symmetric , positive - semidefinite matrices ) , with @xmath32 providing a tradeoff between the two terms .
this variational formulation is a _ convex optimization _ problem .
in particular it is a regularized max - det problem and can be solved in polynomial time using standard off - the - shelf solvers @xcite .
our main result in section [ sec : main ] is a proof of the consistency of the estimator in the high - dimensional regime in which both the number of observed variables and the number of hidden components are allowed to grow with the number of samples ( of the observed variables ) .
we show that for a suitable choice of the regularization parameter @xmath33 , there exists a range of values of @xmath32 for which the estimates @xmath34 have the same sparsity ( and sign ) pattern and rank as @xmath35 with high probability ( see theorem [ theo : main ] ) .
the key technical requirement is an identifiability condition for the two components of the marginal concentration matrix @xmath10 with respect to the fisher information ( see section [ subsec : fi ] ) .
we make connections between our condition and the irrepresentability conditions required for support / graphical - model recovery using @xmath0 regularization @xcite .
our results provide numerous scaling regimes under which consistency holds in latent - variable graphical model selection .
for example we show that under suitable identifiability conditions consistent model selection is possible even when the number of samples and the number of latent variables are on the same order as the number of observed variables ( see section [ subsec : scal ] ) .
[ [ related - previous - work ] ] related previous work + + + + + + + + + + + + + + + + + + + + + the problem of decomposing the sum of a sparse matrix and a low - rank matrix , with no additional noise , into the individual components was initially studied in @xcite by a superset of the authors of the present paper .
specifically this work proposed a convex program using a combination of the @xmath0 norm and the nuclear norm to recover the sparse and low - rank components , and derived conditions under which the convex program exactly recovers these components . in subsequent work cands et al .
@xcite also studied this noise - free sparse - plus - low - rank decomposition problem , and provided guarantees for exact recovery using the convex program proposed in @xcite .
the problem setup considered in the present paper is quite different and is more challenging because we are only given access to an inexact sample covariance matrix , and we are interested in recovering components that preserve both the sparsity pattern and the rank of the components in the true underlying model .
in addition to proving such a consistency result for the estimator , we also provide a statistical interpretation of our identifiability conditions and describe natural classes of latent - variable gaussian graphical models that satisfy these conditions . as such
our paper is closer in spirit to the many recent papers on covariance selection , but with the important difference that some of the variables are not observed .
[ [ outline ] ] outline + + + + + + + section [ sec : bg ] gives some background on graphical models as well as the algebraic varieties of sparse and low - rank matrices .
it also provides a formal statement of the problem .
section [ sec : iden ] discusses conditions under which latent - variable models are identifiable , and section [ sec : main ] states the main results of this paper .
we provide experimental demonstration of the effectiveness of our estimator on synthetic and real data in section [ sec : sims ] .
section [ sec : conc ] concludes the paper with a brief discussion .
the appendices include additional details and proofs of all of our technical results .
we briefly discuss concepts from graphical modeling and give a formal statement of the latent - variable model selection problem .
we also describe various properties of the algebraic varieties of sparse matrices and of low - rank matrices .
the following matrix norms are employed throughout this paper : * @xmath36 : denotes the spectral norm , which is the largest singular value of @xmath37 . *
@xmath38 : denotes the largest entry in magnitude of @xmath37 . *
@xmath39 : denotes the frobenius norm , which is the square - root of the sum of the squares of the entries of @xmath37 .
* @xmath40 : denotes the nuclear norm , which is the sum of the singular values of @xmath37 .
this reduces to the trace for positive - semidefinite matrices . *
@xmath41 : denotes the sum of the absolute values of the entries of @xmath37 . a number of _ matrix operator _
norms are also used .
for example , let @xmath42 be a linear operator acting on matrices
. then the induced operator norm @xmath43 is defined as : @xmath44 therefore , @xmath45 denotes the spectral norm of the matrix operator @xmath46 . the only vector norm used is the euclidean norm , which is denoted by @xmath47 .
a graphical model @xcite is a statistical model defined with respect to a graph @xmath48 in which the nodes index a collection of random variables @xmath49 , and the edges represent the conditional independence relations ( markov structure ) among the variables .
the absence of an edge between nodes @xmath50 implies that the variables @xmath51 are independent conditioned on all the other variables .
a _ gaussian graphical model _ ( also commonly referred to as a gauss - markov random field ) is one in which all the variables are jointly gaussian @xcite . in such models the sparsity pattern of the inverse of the covariance matrix , or the _ concentration _ matrix , directly corresponds to the graphical model structure .
specifically , consider a gaussian graphical model in which the covariance matrix is given by @xmath52 and the concentration matrix is given by @xmath53 .
then an edge @xmath54 is present in the underlying graphical model if and only if @xmath55 .
our focus in this paper is on gaussian models in which some of the variables may not be observed .
suppose @xmath14 represents the set of nodes corresponding to observed variables @xmath3 , and @xmath13 the set of nodes corresponding to unobserved , hidden variables @xmath4 with @xmath56 and @xmath57 .
the joint covariance is denoted by @xmath58 , and joint concentration matrix by @xmath59 .
the submatrix @xmath5 represents the marginal covariance of the observed variables @xmath3 , and the corresponding marginal concentration matrix is given by the schur complement with respect to the block @xmath8 : @xmath60 the submatrix @xmath11 specifies the concentration matrix of the conditional statistics of the observed variables conditioned on the hidden components .
if these conditional statistics are given by a sparse graphical model then @xmath11 is sparse . on the other hand
the marginal concentration matrix @xmath10 of the marginal distribution of @xmath3 is _ not _ sparse in general due to the extra correlations induced from marginalization over the latent variables @xmath4 , i.e. , due to the presence of the additional term @xmath12 .
hence , standard graphical model selection techniques in which the goal is to approximate a sample covariance by a sparse graphical model are not well - suited for problems in which some of the variables are hidden . however , the matrix @xmath12 is a low - rank matrix if the number of hidden variables is much smaller than the number of observed variables ( i.e. , @xmath61 ) .
therefore , a more appropriate model selection method is to approximate the sample covariance by a model in which the concentration matrix decomposes into the sum of a sparse matrix and a low - rank matrix .
the objective here is to learn a sparse graphical model among the observed variables _ conditioned _ on some latent variables , as such a model explicitly accounts for the extra correlations induced due to unobserved , hidden components . in order to analyze latent - variable model selection methods ,
we need to define an appropriate notion of model selection consistency for latent - variable graphical models .
notice that given the two components @xmath11 and @xmath12 of the concentration matrix of the marginal distribution , there are _ infinitely _ many configurations of the latent variables ( i.e. , matrices @xmath62 ) that give rise to the _ same _ low - rank matrix @xmath12 . specifically for any non - singular matrix @xmath63 , one can apply the transformations @xmath64 and still preserve the low - rank matrix @xmath12 . in _ all
_ of these models the marginal statistics of the observed variables @xmath3 remain the same upon marginalization over the latent variables @xmath4 . the key _
invariant _ is the low - rank matrix @xmath12 , which _ summarizes _ the effect of marginalization over the latent variables .
these observations give rise to the following notion of consistency : a pair of ( symmetric ) matrices @xmath65 with @xmath66 is an _ algebraically consistent _ estimate of a latent - variable gaussian graphical model given by the concentration matrix @xmath67 if the following conditions hold : 1 .
the sign - pattern of @xmath30 is the same as that of @xmath11 : @xmath68 here we assume that @xmath69 .
the rank of @xmath31 is the same as the rank of @xmath12 : @xmath70 3 .
the concentration matrix @xmath71 can be realized as the marginal concentration matrix of an appropriate latent - variable model : @xmath72 the first condition ensures that @xmath30 provides the correct structural estimate of the conditional graphical model ( given by @xmath11 ) of the observed variables conditioned on the hidden components .
this property is the same as the `` sparsistency '' property studied in standard graphical model selection @xcite .
the second condition ensures that the number of hidden components is correctly estimated .
finally , the third condition ensures that the pair of matrices @xmath65 leads to a realizable latent - variable model . in particular
this condition implies that there exists a valid latent - variable model on @xmath73 variables in which @xmath74 the conditional graphical model structure among the observed variables is given by @xmath30 , @xmath75 the number of latent variables @xmath15 is equal to the rank of @xmath31 , and @xmath76 the extra correlations induced due to marginalization over the latent variables is equal to @xmath31 . any method for matrix factorization
( see for example , @xcite ) can be used to factorize the low - rank matrix @xmath31 , depending on the properties that one desires in the factors ( e.g. , sparsity ) .
we also study parametric consistency in the usual sense , i.e. , we show that one can produce estimates @xmath65 that converge in various norms to the matrices @xmath77 .
notice that proving @xmath65 is close to @xmath77 in some norm does not in general imply that the support / sign - pattern and rank of @xmath65 are the same as those of @xmath77 .
therefore parametric consistency is different from algebraic consistency , which requires that @xmath65 have the same support / sign - pattern and rank as @xmath77 .
[ [ goal ] ] goal + + + + let @xmath78 denote the concentration matrix of a gaussian model .
suppose that we have @xmath21 samples @xmath79 of the observed variables @xmath3 .
we would like to produce estimates @xmath34 that , with high - probability , are both algebraically consistent and parametrically consistent ( in some norm ) .
given @xmath21 samples @xmath80 of a finite collection of jointly gaussian zero - mean random variables with concentration matrix @xmath81 , we define the sample covariance as follows : @xmath82 it is then easily seen that the log - likelihood function is given by : @xmath83 where @xmath84 is a function of @xmath85 .
notice that this function is strictly concave for @xmath25 .
now consider the latent - variable modeling problem in which we wish to model a collection of random variables @xmath3 ( with sample covariance @xmath86 ) by adding some extra variables @xmath4 .
with respect to the parametrization @xmath65 ( with @xmath30 representing the conditional statistics of @xmath3 given @xmath4 , and @xmath31 summarizing the effect of marginalization over the additional variables @xmath4 ) , the likelihood function is given by : @xmath87 the function @xmath88 is _ jointly concave _ with respect to the parameters @xmath65 whenever @xmath89 , and it is this function that we use in our variational formulation to learn a latent - variable model . in the analysis of a convex program involving the likelihood function
, the fisher information plays an important role as it is the negative of the hessian of the likelihood function and thus controls the curvature .
as the first term in the likelihood function is linear , we need only study higher - order derivatives of the log - determinant function in order to compute the hessian . letting @xmath90 denote the fisher information matrix
, we have that @xcite @xmath91 for @xmath92 .
if @xmath81 is a @xmath93 concentration matrix , then the fisher information matrix @xmath94 has dimensions @xmath95 .
next consider the latent - variable situation with the variables indexed by @xmath14 being observed and the variables indexed by @xmath13 being hidden .
the concentration matrix @xmath96 of the marginal distribution of the observed variables @xmath14 is given by the schur complement , and the corresponding fisher information matrix is given by @xmath97 notice that this is precisely the @xmath98 submatrix of the full fisher information matrix @xmath99 with respect to all the parameters @xmath100 ( corresponding to the situation in which _ all _ the variables @xmath101 are observed ) .
the matrix @xmath102 has dimensions @xmath103 , while @xmath104 is an @xmath98 matrix . to summarize
, we have for all @xmath105 that : @xmath106_{(i , j),(k , l ) } = { \mathcal{i}}(k^\ast_{(o~h)})_{(i , j),(k , l)}.\ ] ] in section [ subsec : fi ] we impose various conditions on the fisher information matrix @xmath104 under which our regularized maximum - likelihood formulation provides consistent estimates with high probability . an algebraic variety is the solution set of a system of polynomial equations .
the set of sparse matrices and the set of low - rank matrices can be naturally viewed as algebraic varieties . here
we describe these varieties , and discuss some of their properties .
of particular interest in this paper are geometric properties of these varieties such as the tangent space and local curvature at a ( smooth ) point .
let @xmath107 denote the set of matrices with at most @xmath108 nonzeros : @xmath109 the set @xmath107 is an algebraic variety , and can in fact be viewed as a union of @xmath110 subspaces in @xmath111 .
this variety has dimension @xmath108 , and it is smooth everywhere except at those matrices that have support size strictly smaller than @xmath108 . for any matrix @xmath112 ,
consider the variety @xmath113 ; @xmath37 is a smooth point of this variety , and the tangent space at @xmath37 is given by @xmath114 in words the tangent space @xmath115 at a smooth point @xmath37 is given by the set of all matrices that have support contained within the support of @xmath37 .
we view @xmath115 as a subspace in @xmath111 .
next let @xmath116 denote the algebraic variety of matrices with rank at most @xmath117 : @xmath118 it is easily seen that @xmath116 is an algebraic variety because it can be defined through the vanishing of all @xmath119 minors .
this variety has dimension equal to @xmath120 , and it is smooth everywhere except at those matrices that have rank strictly smaller than @xmath117 . consider a rank-@xmath117 matrix @xmath37 with svd given by @xmath121 , where @xmath122 and @xmath123 .
the matrix @xmath37 is a smooth point of the variety @xmath124 , and the tangent space at @xmath37 with respect to this variety is given by @xmath125 in words the tangent space @xmath126 at a smooth point @xmath37 is the span of all matrices that have either the same row - space as @xmath37 or the same column - space as @xmath37 . as with @xmath115
we view @xmath126 as a subspace in @xmath111 . in section [ sec : iden ]
we explore the connection between geometric properties of these tangent spaces and the identifiability problem in latent - variable graphical models .
the sparse matrix variety @xmath107 has the property that it has _
zero _ curvature at any smooth point .
consequently the tangent space at a smooth point @xmath37 is the _ same _ as the tangent space at any point in a neighborhood of @xmath37 .
this property is implicitly used in the analysis of @xmath0 regularized methods for recovering sparse models .
the situation is more complicated for the low - rank matrix variety , because the curvature at any smooth point is nonzero .
therefore we need to study how the tangent space changes from one point to a neighboring point by analyzing how this variety curves locally .
indeed the amount of curvature at a point is directly related to the `` angle '' between the tangent space at that point and the tangent space at a neighboring point . for any subspace @xmath127 of matrices ,
let @xmath128 denote the projection onto @xmath127 .
given two subspaces @xmath129 of the same dimension , we measure the `` twisting '' between these subspaces by considering the following quantity .
@xmath130 ( n)\|_2 .
\label{eq : rho}\ ] ] in appendix [ app : matper ] we briefly review relevant results from matrix perturbation theory ; the key tool used to derive these results is the resolvent of a matrix @xcite . based on these tools we prove the following two results in appendix [ app : rankcurv ] , which bound the twisting between the tangent spaces at nearby points .
the first result provides a bound on the quantity @xmath131 between the tangent spaces at a point and at its neighbor .
[ theo : tspace ] let @xmath112 be a rank-@xmath117 matrix with smallest nonzero singular value equal to @xmath132 , and let @xmath133 be a perturbation to @xmath37 such that @xmath134 .
further , let @xmath135 be a rank-@xmath117 matrix .
then we have that @xmath136 the next result bounds the error between a point and its neighbor in the normal direction .
[ theo : nspace ] let @xmath112 be a rank-@xmath117 matrix with smallest nonzero singular value equal to @xmath132 , and let @xmath133 be a perturbation to @xmath37 such that @xmath137 .
further , let @xmath135 be a rank-@xmath117 matrix
. then we have that @xmath138 these results suggest that the closer the smallest singular value is to zero , the more curved the variety is locally .
therefore we control the twisting between tangent spaces at nearby points by bounding the smallest nonzero singular value away from zero .
in the absence of additional conditions , the latent - variable model selection problem is ill - posed . in this section
we discuss a set of conditions on latent - variable models that ensure that these models are identifiable given marginal statistics for a subset of the variables .
suppose that the low - rank matrix that summarizes the effect of the hidden components is itself sparse .
this leads to identifiability issues in the sparse - plus - low - rank decomposition problem .
statistically the additional correlations induced due to marginalization over the latent variables could be mistaken for the conditional graphical model structure of the observed variables . in order to avoid such identifiability problems the effect of the latent variables must be `` diffuse '' across the observed variables . to address this point
the following quantity was introduced in @xcite for any matrix @xmath37 , defined with respect to the tangent space @xmath126 : @xmath139 thus @xmath140 being small implies that elements of the tangent space @xmath126 can not have their support concentrated in a few locations ; as a result @xmath37 can not be too sparse .
this idea is formalized in @xcite by relating @xmath140 to a notion of `` incoherence '' of the row / column spaces , where the row / column spaces are said to be incoherent with respect to the standard basis if these spaces are not aligned closely with any of the coordinate axes .
letting @xmath121 be the singular value decomposition of @xmath37 , the incoherence of the row / column spaces of @xmath37 ( initially proposed and studied by cands and recht @xcite ) is defined as : @xmath141 here @xmath142 denote projections , and projections onto matrix subspaces ( defined by a general linear operator ) by the calligraphic @xmath143 . ] onto the row / column spaces of @xmath37 , and @xmath144 is the @xmath145th standard basis vector . hence @xmath146 measures the projection of the most `` closely aligned '' coordinate axis with the row / column spaces . for any rank-@xmath117 matrix m we have that @xmath147 where the lower bound is achieved ( for example ) if the row / column spaces span any @xmath117 columns of a @xmath93 orthonormal hadamard matrix , while the upper bound is achieved if the row or column space contains a standard basis vector .
typically a matrix m with incoherent row / column spaces would have @xmath148 .
the following result ( proved in @xcite ) shows that the more incoherent the row / column spaces of @xmath37 , the smaller is @xmath149 .
[ theo : xiinc ] for any @xmath112 , we have that @xmath150 where @xmath140 and @xmath146 are defined in and .
based on these concepts we roughly require that the low - rank matrix that summarizes the effect of the latent variables be _ incoherent _ , thereby ensuring that the extra correlations due to marginalization over the hidden components can not be confused with the conditional graphical model structure of the observed variables .
notice that the quantity @xmath151 is not just a measure of the number of latent variables , but also of the overall effect of the correlations induced by marginalization over these variables .
* curvature and change in @xmath152 * : as noted previously an important technical point is that the algebraic variety of low - rank matrices is locally curved at any smooth point
. consequently the quantity @xmath152 changes as we move along the low - rank matrix variety smoothly .
the quantity @xmath153 introduced in also allows us to bound the variation in @xmath152 as follows .
[ theo : rhotspace ] let @xmath129 be two matrix subspaces of the same dimension with the property that @xmath154 , where @xmath131 is defined in .
then we have that @xmath155.\ ] ] this lemma is proved in appendix [ app : rankcurv ] .
an identifiability problem also arises if the conditional graphical model among the observed variables contains a densely connected subgraph .
these statistical relationships might be mistaken as correlations induced by marginalization over latent variables .
therefore we need to ensure that the conditional graphical model among the observed variables is sparse .
we impose the condition that this conditional graphical model must have small `` degree '' , i.e. , no observed variable is directly connected to too many other observed variables conditioned on the hidden components .
notice that bounding the degree is a more refined condition than simply bounding the total number of nonzeros as the _ sparsity pattern _ also plays a role . in @xcite
the authors introduced the following quantity in order to provide an appropriate measure of the sparsity pattern of a matrix : @xmath156 the quantity @xmath157 being small for a matrix implies that the spectrum of any element of the tangent space @xmath115 is not too `` concentrated '' , i.e. , the singular values of the elements of the tangent space are not too large . in @xcite
it is shown that a sparse matrix @xmath37 with `` bounded degree '' ( a small number of nonzeros per row / column ) has small @xmath158 .
[ theo : mudeg ] let @xmath112 be any matrix with at most @xmath159 nonzero entries per row / column , and with at least @xmath160 nonzero entries per row / column . with @xmath157
as defined in , we have that @xmath161 suppose that we have the sum of two vectors , each from two known subspaces .
it is possible to uniquely recover the individual vectors from the sum if and only if the subspaces have a transverse intersection , i.e. , they only intersect at the origin . this simple observation leads to an appealing algebraic notion of identifiability .
consider the situation in which we have the sum of a sparse matrix and a low - rank matrix .
in addition to this sum , suppose that we are also given the tangent spaces at these matrices with respect to the algebraic varieties of sparse and low - rank matrices respectively .
then a necessary and sufficient condition for _ local _ identifiability is that these tangent spaces have a transverse intersection .
it turns out that these transversality conditions on the tangent spaces are also sufficient for the regularized maximum - likelihood convex program to provide consistent estimates of the number of hidden components and the conditional graphical model structure of the observed variables conditioned on the latent variables ( without any side information about the tangent spaces ) . in order to quantify the level of transversality between the tangent spaces @xmath162 and
@xmath127 we study the _ minimum gain _ with respect to some norm of the addition operator restricted to the cartesian product @xmath163 .
more concretely let @xmath164 represent the addition operator , i.e. , the operator that adds two matrices .
then given any matrix norm @xmath165 on @xmath166 , the minimum gain of @xmath167 restricted to @xmath168 is defined as follows : @xmath169 where @xmath170 denotes the projection onto the space @xmath168 , and @xmath171 denotes the adjoint of the addition operator ( with respect to the standard euclidean inner - product ) .
the tangent spaces @xmath162 and @xmath127 have a _ transverse _ intersection if and only if @xmath172 . the `` level '' of transversality is measured by the magnitude of @xmath173 .
note that if the norm @xmath165 used is the frobenius norm , then @xmath174 is the square of the _ minimum singular value _ of the addition operator @xmath167 restricted to @xmath175 .
a natural norm with which to measure transversality is the dual norm of the regularization function in , as the subdifferential of the regularization function is specified in terms of its dual . the reasons for this will become clearer as we proceed through this paper .
recall that the regularization function used in the variational formulation is given by : @xmath176 where the nuclear norm @xmath177 reduces to the trace function over the cone of positive - semidefinite matrices .
this function is a norm for all @xmath178 .
the dual norm of @xmath179 is given by @xmath180 the following simple lemma records a useful property of the @xmath181 norm that is used several times throughout this paper .
[ theo : gg ] let @xmath162 and @xmath127 be tangent spaces at any points with respect to the algebraic varieties of sparse and low - rank matrices . then for any matrix @xmath37
, we have that @xmath182 and that @xmath183 .
further we also have that @xmath184 and that @xmath185 .
thus for any matrices @xmath186 and for @xmath163 , one can check that @xmath187 and that @xmath188 .
next we define the quantity @xmath189 as follows in order to study the transversality of the spaces @xmath162 and @xmath127 with respect to the @xmath181 norm : @xmath190 here @xmath191 and @xmath152 are defined in and .
we then have the following result ( proved in appendix [ app : rg ] ) : [ theo : rg1 ] let @xmath192 be matrices such that @xmath193 and let @xmath194 .
then we have that @xmath195 $ ] , where @xmath163 and @xmath189 is defined in . in particular
we have that @xmath196 .
the quantity @xmath189 being small implies that the addition operator is essentially isometric when restricted to @xmath163 .
stated differently the magnitude of @xmath189 is a measure of the level of transversality of the spaces @xmath162 and @xmath127 .
if @xmath197 then @xmath198 ensures that @xmath199 , which in turn implies that the tangent spaces @xmath162 and @xmath127 have a transverse intersection . * observation * : thus we have that the smaller the quantities @xmath200 and @xmath201 , the more transverse the intersection of the spaces @xmath162 and @xmath127 . the main focus of section [ sec : main ] is to analyze the regularized maximum - likelihood convex program by studying its optimality conditions .
the log - likelihood function is well - approximated in a neighborhood by a quadratic form given by the fisher information ( which measures the curvature , as discussed in section [ subsec : ll ] ) .
let @xmath202 denote the fisher information evaluated at the true marginal concentration matrix @xmath203 , where @xmath204 represents the concentration matrix of the full model ( see equation ) .
the appropriate measure of transversality between the tangent spaces @xmath205 and @xmath206 is then in a space in which the inner - product is given by @xmath207 .
specifically , we need to analyze the minimum gain of the operator @xmath208 restricted to the space @xmath163 .
therefore we impose several conditions on the fisher information @xmath207 .
we define quantities that control the gains of @xmath207 restricted to @xmath162 and @xmath127 separately ; these ensure that elements of @xmath162 and elements of @xmath127 are individually identifiable under the map @xmath207 .
in addition we define quantities that , in conjunction with bounds on @xmath200 and @xmath201 , allow us to control the gain of @xmath207 restricted to the direct - sum @xmath209 .
* @xmath207 restricted to @xmath162 * : the minimum gain of the operator @xmath210 restricted to @xmath162 is given by @xmath211 the maximum effect of elements in @xmath162 in the orthogonal direction @xmath212 is given by @xmath213 the operator @xmath207 is injective on @xmath162 if @xmath214 . the ratio @xmath215 implies the irrepresentability condition imposed in @xcite , which gives a sufficient condition for consistent recovery of graphical model structure using @xmath0-regularized maximum - likelihood .
notice that this condition is a generalization of the usual lasso irrepresentability conditions @xcite , which are typically imposed on the covariance matrix .
finally we also consider the following quantity , which controls the behavior of @xmath207 restricted to @xmath162 in the spectral norm : @xmath216 * @xmath207 restricted to @xmath127 * : analogous to the case of @xmath162 one could control the gains of the operators @xmath217 and @xmath218 .
however as discussed previously one complication is that the tangent spaces at nearby smooth points on the rank variety are in general different , and the amount of twisting between these spaces is governed by the local curvature
. therefore we control the gains of the operators @xmath219 and @xmath220 for all tangent spaces @xmath221 that are `` close to '' the nominal @xmath127 ( at the true underlying low - rank matrix ) , measured by @xmath222 being small .
the minimum gain of the operator @xmath220 restricted to @xmath221 ( close to @xmath127 ) is given by @xmath223 similarly the maximum effect of elements in @xmath221 in the orthogonal direction @xmath224 ( for @xmath221 close to @xmath127 ) is given by @xmath225 implicit in the definition of @xmath226 and @xmath227 is the fact that the outer minimum and maximum are only taken over spaces @xmath221 that are tangent spaces to the rank - variety .
the operator @xmath207 is injective on all tangent spaces @xmath221 such that @xmath228 if @xmath229 . an irrepresentability condition ( analogous to those developed for the sparse case ) for tangent spaces near @xmath127 to the rank variety would be that @xmath230 .
finally we also control the behavior of @xmath207 restricted to @xmath221 close to @xmath127 in the @xmath231 norm : @xmath232 the two sets of quantities @xmath233 and @xmath234 essentially control how @xmath207 behaves when restricted to the spaces @xmath162 and @xmath127 _ separately _ ( in the natural norms ) .
the quantities @xmath235 and @xmath236 are useful in order to control the gains of the operator @xmath207 restricted to the _ direct sum _ @xmath209 .
notice that although the magnitudes of elements in @xmath162 are measured most naturally in the @xmath231 norm , the quantity @xmath235 is specified with respect to the spectral norm .
similarly elements of the tangent spaces @xmath221 to the rank variety are most naturally measured in the spectral norm , but @xmath236 provides control in the @xmath231 norm .
these quantities , combined with @xmath200 and @xmath201 ( defined in and ) , provide the `` coupling '' necessary to control the behavior of @xmath207 restricted to elements in the direct sum @xmath209 . in order to keep track of fewer quantities ,
we summarize the six quantities as follows : @xmath237 * main assumption * there exists a @xmath238 $ ] such that : @xmath239 this assumption is to be viewed as a generalization of the irrepresentability conditions imposed on the covariance matrix @xcite or the fisher information matrix @xcite in order to provide consistency guarantees for sparse model selection using the @xmath0 norm . with this assumption we have the following proposition , proved in appendix [ app : rg ] , about the gains of the operator @xmath207 restricted to @xmath209 .
this proposition plays a fundamental role in the analysis of the performance of the regularized maximum - likelihood procedure .
[ theo : irr ] let @xmath162 and @xmath127 be the tangent spaces defined in this section , and let @xmath207 be the fisher information evaluated at the true marginal concentration matrix .
further let @xmath240 be as defined above .
suppose that @xmath241 and that @xmath32 is in the following range : @xmath242.\ ] ] then we have the following two conclusions for @xmath243 with @xmath244 : 1 .
the minimum gain of @xmath207 restricted to @xmath245 is bounded below : @xmath246 specifically this implies that for all @xmath247 @xmath248 2 .
the effect of elements in @xmath243 on the orthogonal complement @xmath249 is bounded above : @xmath250 specifically this implies that for all @xmath247 @xmath251 the last quantity we consider is the spectral norm of the marginal covariance matrix @xmath252 : @xmath253 a bound on @xmath254 is useful in the probabilistic component of our analysis , in order to derive convergence rates of the sample covariance matrix to the true covariance matrix .
we also observe that @xmath255
let @xmath204 denote the full concentration matrix of a collection of zero - mean jointly - gaussian observed and latent variables , let @xmath256 denote the number of observed variables , and let @xmath257 denote the number of latent variables .
we are given @xmath21 samples @xmath258 of the observed variables @xmath3 .
we consider the high - dimensional setting in which @xmath259 are all allowed to grow simultaneously .
the quantities @xmath260 defined in the previous section are accounted for in our analysis , although we suppress the dependence on these quantities in the statement of our main result .
we explicitly keep track of the quantities @xmath261 and @xmath262 as these control the complexity of the latent - variable model given by @xmath204 .
in particular @xmath191 controls the sparsity of the conditional graphical model among the observed variables , while @xmath152 controls the incoherence or `` diffusivity '' of the extra correlations induced due to marginalization over the hidden variables .
based on the tradeoff between these two quantities , we obtain a number of classes of latent - variable graphical models ( and corresponding scalings of @xmath259 ) that can be consistently recovered using the regularized maximum - likelihood convex program ( see section [ subsec : scal ] for details ) . specifically we show that consistent model selection is possible even when the number of samples and the number of latent variables are on the same order as the number of observed variables .
we present our main result next demonstrating the consistency of the estimator , and then discuss classes of latent - variable graphical models and various scaling regimes in which our estimator is consistent .
[ theo : main ] let @xmath78 denote the concentration matrix of a gaussian model .
we have @xmath21 samples @xmath79 of the @xmath272 observed variables denoted by @xmath14 .
let @xmath273 and @xmath274 denote the tangent spaces at @xmath275 and at @xmath276 with respect to the sparse and low - rank matrix varieties respectively .
* assumptions * : suppose that the following conditions hold : 1 .
the quantities @xmath200 and @xmath201 satisfy the assumption of proposition [ theo : irr ] for identifiability , and @xmath32 is chosen in the range specified by proposition [ theo : irr ] .
the number of samples @xmath21 available is such that @xmath277 3 .
the regularization parameter @xmath33 is chosen as @xmath278 4 .
the minimum nonzero singular value @xmath132 of @xmath276 is bounded as @xmath279 5 .
the minimum magnitude nonzero entry @xmath280 of @xmath275 is bounded as @xmath281 * conclusions * : then with probability greater than @xmath282 we have : 1 .
algebraic consistency : the estimate @xmath34 given by the convex program is algebraically consistent , i.e. , the support and sign pattern of @xmath28 is the same as that of @xmath275 , and the rank of @xmath29 is the same as that of @xmath276 .
2 . parametric consistency : the estimate @xmath34 given by the convex program is parametrically consistent : @xmath283 the proof of this theorem is given in appendix [ app : main ] .
the theorem essentially states that if the minimum nonzero singular value of the low - rank piece @xmath267 and minimum nonzero entry of the sparse piece @xmath275 are bounded away from zero , then the convex program provides estimates that are both algebraically consistent and parametrically consistent ( in the @xmath231 and spectral norms ) . in section
[ subsec : cov ] we also show that these results easily lead to parametric consistency rates for the corresponding estimate @xmath284 of the marginal covariance @xmath285 of the observed variables .
notice that the condition on the minimum singular value of @xmath267 is more stringent than on the minimum nonzero entry of @xmath275 .
one role played by these conditions is to ensure that the estimates @xmath34 do not have smaller support size / rank than @xmath286 .
however the minimum singular value bound plays the additional role of bounding the curvature of the low - rank matrix variety around the point @xmath267 , which is the reason for this condition being more stringent .
notice also that the number of hidden variables @xmath287 does not explicitly appear in the bounds in theorem [ theo : main ] , which only depend on @xmath288 .
however the dependence on @xmath287 is implicit in the dependence on @xmath262 , and we discuss this point in greater detail in the following section .
finally we note that algebraic and parametric consistency hold under the assumptions of theorem [ theo : main ] for a _ range _ of values of @xmath32 : @xmath242.\ ] ] in particular the assumptions on the sample complexity , the minimum nonzero singular value of @xmath267 , and the minimum magnitude nonzero entry of @xmath275 are governed by the lower end of this range for @xmath32 .
these assumptions can be weakened if we only require consistency for a smaller range of values of @xmath32 .
the following corollary conveys this point with a specific example : [ theo : maincorl ] consider the same setup and notation as in theorem [ theo : main ] .
suppose that the quantities @xmath200 and @xmath201 satisfy the assumption of proposition [ theo : irr ] for identifiability .
suppose that we make the following assumptions : 1 .
let @xmath32 be chosen to be equal to @xmath289 ( the upper end of the range specified in proposition [ theo : irr ] ) , i.e. , @xmath290 .
2 . @xmath291 .
3 . @xmath292 .
4 . @xmath293 .
5 . @xmath294 . then with probability greater than @xmath282 we have estimates @xmath34 that are algebraically consistent , and parametrically consistent with the error bounded as @xmath295 the proof of this corollary
, one can further remove the factor of @xmath201 in the lower bound for @xmath132 .
specifically the lower bound @xmath296 suffices for consistent estimation if @xmath297 bound the minimum / maximum gains of @xmath207 for _ all _ matrices ( rather than just those near @xmath127 ) , and @xmath227 bounds the @xmath207-inner - product for _ all _ pairs of orthogonal matrices ( rather than just those near @xmath127 and @xmath298 ) . ] is analogous to that of theorem [ theo : main ] .
we emphasize that in practice it is often beneficial to have consistent estimates for a range of values of @xmath32 ( as in theorem [ theo : main ] ) .
specifically the stability of the sparsity pattern and rank of the estimates @xmath34 for a range of tradeoff parameters is useful in order to choose a suitable value of @xmath32 , as prior information about the quantities @xmath261 and @xmath262 is not typically available ( see section [ sec : sims ] ) .
next we consider classes of latent - variable models that satisfy the conditions of theorem [ theo : main ] .
recall that @xmath21 denotes the number of samples , @xmath272 denotes the number of observed variables , and @xmath287 denotes the number of latent variables . recall the assumption that the quantities @xmath260 defined in section [ subsec : fi ] remain bounded , and do not scale with the other parameters such as @xmath259 or @xmath262 or @xmath261 .
in particular we focus on the tradeoff between @xmath262 and @xmath261 ( the quantities that control the complexity of a latent - variable graphical model ) , and the resulting scaling regimes for consistent estimation .
let @xmath299 denote the degree of the conditional graphical model among the observed variables , and let @xmath300 denote the incoherence of the correlations induced due to marginalization over the latent variables ( we suppress the dependence on @xmath21 ) .
these quantities are defined in section [ sec : iden ] , and we have from propositions [ theo : xiinc ] and [ theo : mudeg ] that @xmath301 since @xmath260 are assumed to be bounded , we also have from proposition [ theo : irr ] that the product of @xmath191 and @xmath152 must be bounded by a constant .
thus , we study latent - variable models in which @xmath302 as we describe next , there are non - trivial classes of latent - variable graphical models in which this condition holds .
* bounded degree and incoherence * : the first class of latent - variable models that we consider are those in which the conditional graphical model among the observed variables ( given by @xmath275 ) has constant degree @xmath303 .
recall from equation that the incoherence @xmath145 of the effect of the latent variables ( given by @xmath267 ) can be as small as @xmath304 . consequently latent - variable models in which @xmath305 can be estimated consistently from @xmath306 samples as long as the low - rank matrix @xmath276 is almost maximally incoherent , i.e. , @xmath307 so the effect of marginalization over the latent variables is diffuse across almost all the observed variables .
thus consistent latent - variable model selection is possible even when the number of samples and the number of latent variables are on the same order as the number of observed variables .
* polylogarithmic degree * the next class of models that we study are those in which the degree @xmath303 of the conditional graphical model of the observed variables grows poly - logarithmically with @xmath272 .
consequently , the incoherence @xmath145 of the matrix @xmath276 must decay as the inverse of poly-@xmath308 . using the fact that maximally incoherent low - rank matrices @xmath276 can have incoherence as small as @xmath304 , latent - variable models in which @xmath309 can be consistently estimated as long as @xmath310poly-@xmath308 .
the main result theorem [ theo : main ] gives conditions under which we can consistently estimate the sparse and low - rank parts that compose the marginal concentration matrix @xmath311 .
here we prove a corollary that gives rates for covariance matrix estimation , i.e. , the quality of the estimate @xmath284 with respect to the `` true '' marginal covariance matrix @xmath312 . under the same conditions as in theorem
[ theo : main ] , we have with probability greater than @xmath313 that @xmath314 ) \lesssim \frac{1}{\xi(t ) } \sqrt{\frac{p}{n}}.\ ] ] specifically this implies that @xmath315 . * proof * : the proof of this lemma follows directly from duality .
based on the analysis in appendix [ app : main ] ( in particular using the optimality conditions of the modified convex program ) , we have that @xmath316 ) \leq \lambda_n.\ ] ] we also have from the bound on the number of samples @xmath21 that with probability greater than @xmath317 ( see appendix [ app : final ] ) @xmath318 ) \lesssim \lambda_n\ ] ] based on the choice of @xmath33 in theorem [ theo : main ] , we then have the desired bound .
@xmath319 standard results from convex analysis @xcite state that @xmath34 is a minimum of the convex program if the zero matrix belongs to the subdifferential of the objective function evaluated at @xmath34 ( in addition to @xmath34 satisfying the constraints ) .
the subdifferential of the @xmath0 norm at a matrix @xmath37 is given by @xmath320 for a symmetric positive semidefinite matrix @xmath37 with svd @xmath321 , the subdifferential of the trace function restricted to the cone of positive semidefinite matrices ( i.e. , the nuclear norm over this set ) is given by : @xmath322 ~~ \leftrightarrow ~~ { \mathcal{p}}_{t(m)}(n ) = u u^t , ~ { \mathcal{p}}_{t(m)^\bot}(n ) \preceq i,\ ] ] where @xmath323 denotes the characteristic function of the set of positive semidefinite matrices ( i.e. , the convex function that evaluates to @xmath324 over this set and @xmath325 outside ) .
the key point is that elements of the subdifferential decompose with respect to the tangent spaces @xmath115 and @xmath126 .
this decomposition property plays a critical role in our analysis .
in particular it states that the optimality conditions consist of two parts , one part corresponding to the tangent spaces @xmath162 and @xmath127 and another corresponding to the normal spaces @xmath212 and @xmath298 . consider the optimization problem with the additional ( non - convex ) constraints that the variable @xmath30 belongs to the algebraic variety of sparse matrices and that the variables @xmath31 belongs to the algebraic variety of low - rank matrices . while this new optimization problem is non - convex , it has a very interesting property . at a globally optimal solution ( and indeed at any locally optimal solution ) @xmath326 such that @xmath327 and @xmath328 are smooth points of the algebraic varieties of sparse and low - rank matrices , the first - order optimality conditions state that the lagrange multipliers corresponding to the additional variety constraints must lie in the _ normal spaces _ @xmath329 and @xmath330 .
this fundamental observation , combined with the decomposition property of the subdifferentials of the @xmath0 and nuclear norms , suggests the following high - level proof strategy : 1 .
let @xmath326 be the globally optimal solution of the optimization problem with the additional constraints that @xmath65 belong to the algebraic varieties of sparse / low - rank matrices ; specifically constrain @xmath30 to lie in @xmath331 and constrain @xmath31 to lie in @xmath332 .
show first that @xmath326 are smooth points of these varieties .
the first part of the subgradient optimality conditions of the original convex program corresponding to components _ on _ the tangent spaces @xmath333 and @xmath334 is satisfied .
this conclusion can be reached because the additional lagrange multipliers due to the variety constraints lie in the normal spaces @xmath329 and @xmath330 .
3 . finally show that the second part of the subgradient optimality conditions of ( without any variety constraints ) corresponding to components in the normal spaces @xmath329 and @xmath330 is also satisfied by @xmath326 .
combining these steps together we show that @xmath326 satisfy the optimality conditions of the _ original convex program _ .
consequently @xmath326 is also the optimum of the convex program . as this estimate is also the solution to the problem with the variety constraints , the algebraic consistency of @xmath326
can be directly concluded .
we emphasize here that the variety - constrained optimization problem is used solely as an analysis tool in order to prove consistency of the estimates provided by the convex program .
these steps describe our broad strategy , and we refer the reader to appendix [ app : main ] for details .
the key technical complication is that the tangent spaces at @xmath328 and @xmath267 are in general different .
we bound the twisting between these tangent spaces by using the fact that the minimum nonzero singular value of @xmath267 is bounded away from zero ( as assumed in theorem [ theo : main ] and using proposition [ theo : tspace ] ) .
in this section we give experimental demonstration of the consistency of our estimator on synthetic examples , and its effectiveness in modeling real - world stock return data .
our choices of @xmath33 and @xmath32 are guided by theorem [ theo : main ] .
specifically , we choose @xmath33 to be proportional to @xmath335 . for @xmath32
we observe that the support / sign - pattern and the rank of the solution @xmath34 are the same for a _ range _ of values of @xmath32
. therefore one could solve the convex program for several values of @xmath32 , and choose a solution in a suitable range in which the sign - pattern and rank of the solution are stable .
in practical problems with real - world data these parameters may be chosen via cross - validation . for small problem instances
we solve the convex program using a combination of yalmip @xcite and sdpt3 @xcite , which are standard off - the - shelf packages for solving convex programs . for larger problem instances we use the special purpose solver logdetppa @xcite developed for log - determinant semidefinite programs . in the first set of experiments we consider a setting in which we have access to samples of the observed variables of a latent - variable graphical model .
we consider several latent - variable gaussian graphical models .
the first model consists of @xmath336 observed variables and @xmath337 hidden variables .
the conditional graphical model structure of the observed variables is a cycle with the edge partial correlation coefficients equal to @xmath338 ; thus , this conditional model is specified by a sparse graphical model with degree @xmath339 .
the second model is the same as the first one , but with @xmath340 latent variables .
the third model consists of @xmath341 latent variable , and the conditional graphical model structure of the observed variables is given by a @xmath342 nearest - neighbor grid ( i.e. , @xmath336 and degree @xmath343 ) with the partial correlation coefficients of the edges equal to @xmath344 . in all three of these models each latent variable is connected to a random subset of @xmath345 of the observed variables ( and the partial correlation coefficients corresponding to these edges are also random ) .
therefore the effect of the latent variables is `` spread out '' over most of the observed variables , i.e. , the low - rank matrix summarizing the effect of the latent variables is incoherent . for each model
we generate @xmath21 samples of the observed variables , and use the resulting sample covariance matrix @xmath86 as input to our convex program .
figure [ fig : fig1 ] shows the probability of recovery of the support / sign - pattern of the conditional graphical model structure in the observed variables and the number of latent variables ( i.e. , probability of obtaining algebraically consistent estimates ) as a function of @xmath21 .
this probability is evaluated over @xmath346 experiments for each value of @xmath21 . in all of these cases standard
graphical model selection applied directly to the observed variables is not useful as the marginal concentration matrix of the observed variables is not well - approximated by a sparse matrix .
these experiments agree with our theoretical results that the convex program is an algebraically consistent estimator of a latent - variable model given ( sufficiently many ) samples of only the observed variables . in the next experiment we model the statistical structure of monthly stock returns of 84 companies in the s&p 100 index from 1990 to 2007 ; we disregard 16 companies that were listed after 1990 .
the number of samples @xmath21 is equal to @xmath347 .
we compute the sample covariance based on these returns and use this as input to .
the model learned using for suitable values of @xmath348 consists of @xmath349 latent variables , and the conditional graphical model structure of the stock returns conditioned on these hidden components consists of @xmath350 edges .
therefore the number of parameters in the model is @xmath351 .
the resulting kl divergence between the distribution specified by this model and a gaussian distribution specified by the sample covariance is @xmath352 .
figure [ fig : fig2 ] ( left ) shows the _ conditional _ graphical model structure .
the strongest edges in this conditional graphical model , as measured by partial correlation , are between baker hughes - schlumberger , a.t.&t .
- verizon , merrill lynch - morgan stanley , halliburton - baker hughes , intel - texas instruments , apple - dell , and microsoft - dell .
it is of interest to note that in the standard industrial classification system for grouping these companies , several of these pairs are in different classes . as mentioned in section [ subsec : ps ] our method estimates a low - rank matrix that summarizes the effect of the latent variables ; in order to factorize this low - rank matrix , for example into sparse factors , one could use methods such as those described in @xcite .
we compare these results to those obtained using a sparse graphical model learned using @xmath0-regularized maximum - likelihood ( see for example @xcite ) , without introducing any latent variables .
figure [ fig : fig2 ] ( right ) shows this graphical model structure .
the number of edges in this model is @xmath353 ( the total number of parameters is equal to @xmath354 ) , and the resulting kl divergence between this distribution and a gaussian distribution specified by the sample covariance is @xmath355 . indeed to obtain a comparable kl divergence to that of the latent - variable model described above , one would require a graphical model with over @xmath356 edges .
these results suggest that a latent - variable graphical model is better suited than a standard sparse graphical model for modeling the statistical structure among stock returns .
this is likely due to the presence of global , long - range correlations in stock return data that are better modeled via latent variables .
we have studied the problem of modeling the statistical structure of a collection of random variables as a sparse graphical model conditioned on a few additional hidden components . as a first contribution we described conditions under which such latent - variable graphical models are identifiable given samples of only the observed variables .
we also proposed a convex program based on regularized maximum - likelihood for latent - variable graphical model selection ; the regularization function is a combination of the @xmath0 norm and the nuclear norm . given samples of the observed variables of a latent - variable gaussian model we proved that this convex program provides consistent estimates of the number of hidden components as well as the conditional graphical model structure among the observed variables conditioned on the hidden components .
our analysis holds in the high - dimensional regime in which the number of observed / latent variables are allowed to grow with the number of samples of the observed variables .
in particular we discuss certain scaling regimes in which consistent model selection is possible even when the number of samples and the number of latent variables are on the same order as the number of observed variables .
these theoretical predictions are verified via a set of experiments on synthetic data .
we also demonstrate the effectiveness of our approach in modeling real - world stock return data .
several research questions arise that are worthy of further investigation .
while the convex program can be solved in polynomial time using off - the - shelf solvers , it is preferable to develop more efficient special - purpose solvers that can scale to massive datasets by taking advantage of the structure of the formulation .
finally it would be of interest to develop a similar convex optimization formulation with consistency guarantees for latent - variable models with non - gaussian variables , e.g. , for categorical data .
we would like to thank james saunderson and myung jin choi for helpful discussions , and kim - chuan toh for kindly providing us specialized code to solve larger instances of our convex program .
given a low - rank matrix we consider what happens to the invariant subspaces when the matrix is perturbed by a small amount .
we assume without loss of generality that the matrix under consideration is square and symmetric , and our methods can be extended to the general non - symmetric non - square case .
we refer the interested reader to @xcite for more details , as the results presented here are only a brief summary of what is relevant for this paper .
in particular the arguments presented here are along the lines of those presented in @xcite .
the appendices in @xcite also provide a more refined analysis of second - order perturbation errors .
the resolvent of a matrix @xmath37 is given by @xmath357 @xcite , and it is well - defined for all @xmath358 that do not coincide with an eigenvalue of @xmath37 .
if @xmath37 has no eigenvalue with magnitude equal to @xmath359 , then we have by the cauchy residue formula that the projector onto the invariant subspace of a matrix @xmath37 corresponding to all singular values smaller than @xmath359 is given by @xmath360 where @xmath361 denotes the positively - oriented circle of radius @xmath359 centered at the origin .
similarly , we have that the weighted projection onto the invariant subspace corresponding to the smallest singular values is given by @xmath362 suppose that @xmath37 is a low - rank matrix with smallest nonzero singular value @xmath132 , and let @xmath133 be a perturbation of @xmath37 such that @xmath363 .
we have the following identity for any @xmath364 , which will be used repeatedly : @xmath365^{-1 } - [ m - \zeta i]^{-1 } = - [ m - \zeta i]^{-1 } \delta [ ( m+\delta ) - \zeta i]^{-1}. \label{eq : matexp}\ ] ]
we then have that @xmath366^{-1 } - [ m - \zeta i]^{-1 } d \zeta \nonumber \\ & = & \frac{1}{2 \pi i } \oint_{{\mathcal{c}}_\kappa } [ m - \zeta i]^{-1 } \delta [ ( m+\delta ) - \zeta i]^{-1 } d \zeta .
\label{eq : projexp}\end{aligned}\ ] ] similarly , we have the following for @xmath367 : @xmath368^{-1 } - [ m - \zeta i]^{-1}\right\ } d \zeta \nonumber \\ & = & \frac{1}{2 \pi i } \oint_{{\mathcal{c}}_\kappa } \zeta ~ \left\{[m - \zeta i]^{-1 } \delta [ ( m+\delta ) - \zeta i]^{-1}\right\ } d \zeta \nonumber \\ & = & \frac{1}{2 \pi i } \oint_{{\mathcal{c}}_\kappa } \zeta ~ [ m - \zeta i]^{-1 } \delta [ m - \zeta i]^{-1 } d \zeta \nonumber \\ & & - \frac{1}{2 \pi i } \oint_{{\mathcal{c}}_\kappa } \zeta ~ [ m - \zeta i]^{-1 } \delta [ m - \zeta i]^{-1 } \delta [ ( m+\delta ) - \zeta i]^{-1 } d \zeta .
\label{eq : wprojexp}\end{aligned}\ ] ] given these expressions , we have the following two results .
[ theo : projbound ] let @xmath112 be a rank-@xmath117 matrix with smallest nonzero singular value equal to @xmath132 , and let @xmath133 be a perturbation to @xmath37 such that @xmath369 with @xmath370 .
then we have that @xmath371 * proof * : this result follows directly from the expression , and the sub - multiplicative property of the spectral norm : @xmath372 here , we used the fact that @xmath373^{-1}\|_2 \leq \frac{1}{\sigma - \kappa}$ ] and @xmath374^{-1}\|_2 \leq \frac{1}{\sigma - \frac{3\kappa}{2}}$ ] for @xmath364 .
@xmath319 next , we develop a similar bound for @xmath367 .
let @xmath375 denote the invariant subspace of @xmath37 corresponding to the nonzero singular values , and let @xmath376 denote the projector onto this subspace .
[ theo : wprojbound ] let @xmath112 be a rank-@xmath117 matrix with smallest nonzero singular value equal to @xmath132 , and let @xmath133 be a perturbation to @xmath37 such that @xmath369 with @xmath370 .
then we have that @xmath377 * proof * : one can check that @xmath378^{-1 } \delta [ m - \zeta i]^{-1 } d \zeta = ( i - p_{{u}(m ) } ) \delta ( i - p_{{u}(m)}).\ ] ] next we use the expression , and the sub - multiplicative property of the spectral norm : @xmath379 as with the previous proof , we used the fact that @xmath373^{-1}\|_2 \leq \frac{1}{\sigma - \kappa}$ ] and @xmath374^{-1}\|_2 \leq \frac{1}{\sigma - \frac{3\kappa}{2}}$ ] for @xmath364 .
@xmath319 we will use these expressions to derive bounds on the `` twisting '' between the tangent spaces at @xmath37 and at @xmath135 with respect to the rank variety .
for a symmetric rank-@xmath117 matrix @xmath37 , the projection onto the tangent space @xmath126 ( restricted to the variety of symmetric matrices with rank less than or equal to @xmath117 ) can be written in terms of the projection @xmath380 onto the row space @xmath381 . for any matrix @xmath382 @xmath383 one
can then check that the projection onto the normal space @xmath384 @xmath385 ( n)= ( i - p_{u(m ) } ) ~ n ~ ( i - p_{u(m)}).\ ] ] * proof of proposition [ theo : nspace ] * : since both @xmath37 and @xmath135 are rank-@xmath117 matrices , we have that @xmath391 for @xmath389 .
consequently , @xmath392 where we obtain the last inequality from proposition [ theo : wprojbound ] with @xmath389 .
@xmath319 * proof of lemma [ theo : rhotspace ] * : since @xmath154 one can check that the largest principal angle between @xmath393 and @xmath394 is strictly less than @xmath395 . consequently
, the mapping @xmath396 restricted to @xmath393 is bijective ( as it is injective , and the spaces @xmath129 have the same dimension ) . consider the maximum and minimum gain of the operator @xmath397 restricted to @xmath393 ; for any @xmath398 : @xmath399(m)\|_2 \\ & \in & [ 1-\rho(t_1,t_2 ) , 1+\rho(t_1,t_2)].\end{aligned}\ ] ] therefore , we can rewrite @xmath400 as follows : @xmath401(n)\|_\infty ] \\
& \leq & \frac{1}{1 ~ - ~ \rho(t_1,t_2 ) } ~ \left[\xi(t_1 ) ~ + ~ \max_{n \in t_1 , \|n\|_2 \leq 1 } ~ \|[{\mathcal{p}}_{t_1 } - { \mathcal{p}}_{t_2}](n)\|_\infty \right ] \\ & \leq & \frac{1}{1 ~ - ~ \rho(t_1,t_2 ) } ~ \left[\xi(t_1 ) ~ + ~ \max_{\|n\|_2 \leq 1 } ~ \|[{\mathcal{p}}_{t_1 } - { \mathcal{p}}_{t_2}](n)\|_2 \right ] \\ & \leq & \frac{1}{1 ~ - ~ \rho(t_1,t_2 ) } ~ \left[\xi(t_1 ) ~ + ~ \rho(t_1,t_2 ) \right].\end{aligned}\ ] ] this concludes the proof of the lemma . @xmath319
* proof of lemma [ theo : rg1 ] * : we have that @xmath402 ; therefore , @xmath403 .
we need to bound @xmath404 and @xmath405 .
first , we have @xmath406 \\ & \subseteq & [ \|s\|_{\infty } - \|l\|_{\infty } , \|s\|_{\infty } + \|l\|_{\infty } ] \\ & \subseteq & [ \gamma - \xi(t ) , \gamma + \xi(t)].\end{aligned}\ ] ] similarly , one can check that @xmath407 \\
& \subseteq & [ 1 - 2\|s\|_2 , 1 + 2 \|s\|_2 ] \\ & \subseteq & [ 1 - 2 \gamma \mu(\omega ) , 1 + 2 \gamma \mu(\omega)].\end{aligned}\ ] ] thus , we can conclude that @xmath408.\ ] ] where @xmath189 is defined in .
@xmath319 * proof of proposition [ theo : irr ] * : before proving the two parts of this proposition we make a simple observation about @xmath409 using the condition that @xmath410 by applying lemma [ theo : rhotspace ] : @xmath411 here we used the property that @xmath412 in obtaining the final inequality .
consequently , noting that @xmath413 $ ] implies that @xmath414 * part @xmath415 * : the proof of this step proceeds in a similar manner to that of lemma [ theo : rg1 ] .
first we have for @xmath416 with @xmath417 : @xmath418 next under the same conditions on @xmath419 , @xmath420 combining these last two bounds with , we conclude that @xmath421 where the final inequality follows from the assumption that @xmath422 $ ] . *
part @xmath339 * : note that for @xmath416 with @xmath423 @xmath424 similarly @xmath425 combining these last two bounds with the bounds from the first part , we have that @xmath426 this concludes the proof of the proposition .
here we prove theorem [ theo : main ] . throughout this section
we denote @xmath427 .
further @xmath273 and @xmath206 denote the tangent spaces at the `` true '' sparse matrix @xmath428 and low - rank matrix @xmath429 .
we assume that @xmath242 \label{eq : grange}\ ] ] we also let @xmath430 denote the difference between the true marginal covariance and the sample covariance .
finally we let @xmath431 throughout this section . for @xmath32 in the above range we note that @xmath432 standard facts that we use throughout this section are that @xmath412 and that @xmath433 for any matrix @xmath37 .
we study the following convex program : @xmath434 - \log\det(s - l ) ~ + ~ \lambda_n [ \gamma \|s\|_{1 } + \|l\|_\ast ] \\
\mbox{s.t . } & ~~~ s - l \succ 0 . \end{aligned } \label{eq : sdpsy}\ ] ] comparing with the convex program ,
the main difference is that we do not constraint the variable @xmath31 to be positive semidefinite in ( recall that the nuclear norm of a positive semidefinite matrix is equal to its trace ) .
however we show that the unique optimum @xmath435 of under the hypotheses of theorem [ theo : main ] is such that @xmath436 ( with high probability ) .
therefore we conclude that @xmath435 is also the unique optimum of .
the subdifferential with respect to the nuclear norm at a matrix @xmath37 with ( reduced ) svd given by @xmath121 is as follows : @xmath437 1 .
we show that if we solve the convex program subject to the additional constraints that @xmath438 and @xmath439 for some @xmath221 `` close to '' @xmath127 ( measured by @xmath440 ) , then the error between the optimal solution @xmath435 and the underlying matrices @xmath441 is small .
this result is discussed in appendix [ app : errbound ] .
we analyze the optimization problem with the additional constraint that the variables @xmath30 and @xmath31 belong to the algebraic varieties of sparse and low - rank matrices respectively , and that the corresponding tangent spaces are close to the tangent spaces at @xmath442 .
we show that under suitable conditions on the minimum nonzero singular value of the true low - rank matrix @xmath443 and on the minimum magnitude nonzero entry of the true sparse matrix @xmath444 , the optimum of this modified program is achieved at a _ smooth _ point of the underlying varieties .
in particular the bound on the minimum nonzero singular value of @xmath443 helps bound the curvature of the low - rank matrix variety locally around @xmath443 ( we use the results described in appendix [ app : rankcurv ] ) .
these results are described in appendix [ app : var ] .
3 . the next step is to show that the variety constraint can be linearized and changed to a tangent - space constraint ( see appendix [ app : ts ] ) , thus giving us a _
convex program_. under suitable conditions this tangent - space constrained program also has an optimum that has the same support / rank as the true @xmath441 .
based on the previous step these tangent spaces in the constraints are close to the tangent spaces at the true @xmath441 .
therefore we use the first step to conclude that the resulting error in the estimate is small .
4 . finally we show that under the identifiability conditions of section [ sec : iden ] these tangent - space constraints are inactive at the optimum ( see appendix [ app : final ] ) .
therefore we conclude with the statement that the optimum of the convex program without any variety constraints is achieved at a pair of matrices that have the same support / rank as the true @xmath441 ( with high probability ) .
further the low - rank component of the solution is positive semidefinite , thus allowing us to conclude that the original convex program also provides estimates that are consistent .
consider the taylor series of the inverse of a matrix : @xmath445 where @xmath446.\ ] ] this infinite sum converges for @xmath133 sufficiently small .
the following proposition provides a bound on the second - order term specialized to our setting : * proof * : we have that @xmath452 where the second - to - last inequality follows from the range for @xmath32 and that @xmath422 $ ] , and the final inequality follows from the bound on @xmath453 .
therefore , @xmath454 here we apply the last two inequalities from above . since the @xmath455-norm is bounded above by the spectral norm @xmath456 , we have the desired result .
@xmath319 next we analyze the following convex program subject to certain additional tangent - space constraints : @xmath457 - \log\det(s - l ) ~ + ~ \lambda_n [ \gamma \|s\|_{1 } + \|l\|_\ast ]
\\ \mbox{s.t . } & ~~~ s - l \succ 0 , ~~ s \in \omega , ~~ l \in t ' , \end{aligned } \label{eq : sdpts}\ ] ] for some subspace @xmath221 .
we show that if @xmath221 is any tangent space to the low - rank matrix variety such that @xmath410 , then we can bound the error @xmath458 .
let @xmath459 denote the normal component of the true low - rank matrix at @xmath221 , and recall that @xmath430 denotes the difference between the true marginal covariance and the sample covariance .
the proof of the following result uses brouwer s fixed - point theorem @xcite , and is inspired by the proof of a similar result in @xcite for standard sparse graphical model recovery without latent variables .
[ theo : bfpt ] let the error @xmath449 in the solution of the convex program ( with @xmath221 such that @xmath228 ) be as defined above .
further let @xmath448 , and define @xmath460 , \|{\mathcal{c}}_{t'}\|_2 \right\}.\ ] ] if we have that @xmath461 for @xmath32 in the range given by , then @xmath462 * proof * : based on proposition [ theo : irr ] we note that the convex program is strictly convex ( because the negative log - likelihood term has a strictly positive - definite hessian due to the constraints involving transverse tangent spaces ) , and therefore the optimum is unique .
applying the optimality conditions of the convex program at the optimum @xmath463 , we have that there exist lagrange multipliers @xmath464 such that @xmath465 restricting these conditions to the space @xmath243 , one can check that @xmath466 = z_\omega , ~~ { \mathcal{p}}_{t ' } [ \sigma^n_o - ( \hat{s}_\omega-\hat{l}_{t'})^{-1 } ] = z_{t'},\ ] ] where @xmath467 and @xmath468 ( we use here the fact that projecting onto a tangent space @xmath221 increases the spectral norm by at most a factor of two ) . denoting @xmath469 $ ] , we conclude that @xmath470 = z , \label{eq : bfpteq1}\ ] ] with @xmath471 . since the optimum @xmath463 is unique , one can check using lagrangian duality theory @xcite that @xmath463 is the unique solution of the equation . rewriting @xmath472 in terms of the errors
@xmath449 , we have using the taylor series of the matrix inverse that @xmath473^{-1 } \nonumber \\ & = & e_n - r_{\sigma^\ast_o}({\mathcal{a}}(\delta_s,\delta_l ) ) + { \mathcal{i}}^\ast { \mathcal{a}}(\delta_s,\delta_l ) \nonumber \\ & = & e_n - r_{\sigma^\ast_o}({\mathcal{a}}(\delta_s,\delta_l ) ) + { \mathcal{i}}^\ast { \mathcal{a}}{\mathcal{p}}_{\mathcal{y}}(\delta_s,\delta_l ) + { \mathcal{i}}^\ast { \mathcal{c}}_{t'}. \label{eq : bfpteq2}\end{aligned}\ ] ] since @xmath221 is a tangent space such that @xmath228 , we have from proposition [ theo : irr ] that the operator @xmath474 from @xmath168 to @xmath168 is bijective and is well - defined . now consider the following matrix - valued function from @xmath475 to @xmath168 : @xmath476 - z \right\}.\ ] ]
a point @xmath475 is a fixed - point of @xmath477 if and only if @xmath478 = z$ ] .
applying equations and above , we then see that the only fixed - point of @xmath477 by construction is the `` true '' error @xmath479 restricted to @xmath168 .
the reason for this is that , as discussed above , @xmath463 is the unique optimum of and therefore is the
_ unique solution _ of .
next we show that this unique fixed - point of @xmath477 lies in the ball @xmath480 . in order to prove this step , we resort to brouwer s fixed point theorem @xcite . in particular
we show that the function @xmath477 maps the ball @xmath481 onto itself . since @xmath477 is a continuous function and @xmath481 is a compact set
, we can conclude the proof of this proposition . simplifying the function @xmath477
, we have that @xmath482 + z \right\}.\ ] ] consequently , we have from proposition [ theo : irr ] that @xmath483 - z \right ) \\ & \leq & \frac{4}{\alpha } ~ \left\{{g_\gamma}({\mathcal{a}}^\dag [ e_n - r_{\sigma^\ast_o}({\mathcal{a}}(\delta_s,\delta_l+{\mathcal{c}}_{t ' } ) ) + { \mathcal{i}}^\ast { \mathcal{c}}_{t ' } ] ) + \lambda_n \right\ } \\ & \leq & \frac{r}{2 } ~~ + ~~ \frac{4}{\alpha } ~ { g_\gamma}({\mathcal{a}}^\dag r_{\sigma^\ast_o}({\mathcal{a}}(\delta_s,\delta_l+{\mathcal{c}}_{t'}))),\end{aligned}\ ] ] where in the second inequality we use the fact that @xmath484 and that @xmath471 , and in the final inequality we use the assumption on @xmath117 .
we now bound the term @xmath485 using proposition [ theo : rem ] as @xmath447 : @xmath486 where we have used the fact that @xmath487 .
hence @xmath488 by brouwer s fixed - point theorem .
finally we observe that @xmath489 @xmath319 in order to prove that the solution @xmath435 of has the same sparsity pattern / rank as @xmath441 , we will study an optimization problem that explicitly enforces these constraints .
specifically , we consider the following _ non - convex _ constraint set : @xmath490 recall that @xmath491 and @xmath492 .
the first constraint ensures that the tangent space at @xmath30 is the same as the tangent space at @xmath444 ; therefore the support of @xmath30 is contained in the support of @xmath444 .
the second and third constraints ensure that @xmath31 lives in the appropriate low - rank variety , but has a tangent space `` close '' to the tangent space @xmath127 . the final constraint roughly bounds the sum of the errors @xmath493 ; note that this does not necessarily bound the individual errors .
notice that the only non - convex constraint is that @xmath494 .
we then have the following nonlinear program : @xmath495 - \log\det(s - l ) ~ + ~ \lambda_n [ \gamma \|s\|_{1 } + \|l\|_\ast ] \\
\mbox{s.t . } & ~~~ s - l \succ 0 , ~~ ( s , l ) \in \mathcal{m}. \end{aligned } \label{eq : nlp}\ ] ] under suitable conditions this nonlinear program is shown to have a unique solution .
each of the constraints in @xmath496 is useful for proving the consistency of the solution of the convex program .
we show that under suitable conditions the constraints in @xmath496 are actually inactive at the optimal @xmath497 , thus allowing us to conclude that the solution of is also equal to @xmath497 ; hence the solution of shares the consistency properties of @xmath497 .
a number of interesting properties can be derived simply by studying the constraint set @xmath496 . * proof * : we have by the triangle inequality that @xmath502 as @xmath503 . therefore , we have that @xmath504 , where @xmath163 .
consequently , we can apply proposition [ theo : irr ] to conclude that @xmath505 finally , we use the triangle inequality again to conclude that @xmath506 @xmath319 this simple result immediately leads to a number of useful corollaries .
for example we have that under a suitable bound on the minimum nonzero singular value of @xmath429 , the constraint in @xmath496 along the normal direction @xmath298 is locally inactive .
next we list several useful consequences of proposition [ theo : nlp ] .
[ theo : nlpcor ] consider _ any _ @xmath498 , and let @xmath499 .
suppose @xmath32 is in the range specified by , and let @xmath507 and @xmath508 ( where @xmath509 is as defined in proposition [ theo : nlp ] ) .
let the minimum nonzero singular value @xmath132 of @xmath510 be such that @xmath511 for @xmath512 , and suppose that the smallest magnitude nonzero entry of @xmath444 is greater than @xmath513 for @xmath514 .
setting @xmath515 and @xmath459 , we then have that : 1 .
@xmath31 has rank equal to @xmath516 , i.e. , @xmath31 is a smooth point of the variety of matrices with rank less than or equal to @xmath516 .
in particular @xmath31 has the same inertia as @xmath443 .
2 . @xmath517 .
3 . @xmath518 .
4 . @xmath519 .
5 . @xmath520 .
@xmath521 . * proof * : we note the following facts before proving each step . first @xmath522 .
second @xmath412 .
third we have from proposition [ theo : nlp ] that @xmath523 .
finally @xmath524 for @xmath525 $ ] .
we prove each step separately . for the fourth step let @xmath530 denote the minimum singular value of @xmath31 .
consequently , @xmath531 \geq 8 \|\delta_l\|_2.\ ] ] using the same reasoning as in the proof of the second step , we have that @xmath532 hence @xmath533 notice that this corollary applies to _ any _
@xmath498 , and is hence applicable to _ any solution _
@xmath497 of the @xmath496-constrained program .
for now we choose an arbitrary solution @xmath497 and proceed . in the next steps
we show that @xmath497 is _ the unique _ solution to the convex program , thus showing that @xmath497 is also the unique solution to .
given the solution @xmath538 , we show that the solution to the convex program with the tangent space constraint @xmath539 is the same as @xmath538 under suitable conditions : @xmath540 - \log\det(s - l ) ~ + ~ \lambda_n [ \gamma \|s\|_{1 } + \|l\|_\ast ] \\ \mbox{s.t . } & ~~~ s - l \succ 0 , ~~ s \in \omega , ~~ l \in t_\mathcal{m}. \end{aligned } \label{eq : sdptsm}\ ] ] assuming the bound of corollary [ theo : nlpcor ] on the minimum singular value of @xmath443 the uniqueness of the solution @xmath541 is assured .
this is because we have from proposition [ theo : irr ] and from corollary [ theo : nlpcor ] that @xmath207 is injective on @xmath542 .
therefore the hessian of the convex objective function of is strictly positive - definite at @xmath541 .
[ theo : avtots ] let @xmath32 be in the range specified by .
suppose that the minimum nonzero singular value @xmath132 of @xmath492 is such that @xmath511 ( @xmath544 is defined in corollary [ theo : nlpcor ] ) .
suppose also that the minimum magnitude nonzero entry of @xmath444 is greater than or equal to @xmath513 ( @xmath545 is defined in corollary [ theo : nlpcor ] ) .
let @xmath546 .
further suppose that @xmath547 then we have that @xmath548 1 .
first we can change the non - convex constraint @xmath494 to the linear constraint @xmath549 .
this is because the lower bound assumed for @xmath132 implies that @xmath550 is a smooth point of the algebraic variety of matrices with rank less than or equal to @xmath516 ( from corollary [ theo : nlpcor ] ) .
due to the convexity of all the other constraints and the objective , the optimum of this `` linearized '' convex program will still be @xmath551 .
2 . next we can again apply corollary [ theo : nlpcor ] ( based on the bound on @xmath132 ) to conclude that the constraint @xmath552 is _ locally inactive _ at the point @xmath551 .
consequently , we have that @xmath551 can be written as the solution of a _ convex program _ :
@xmath553 - \log\det(s - l ) ~ + ~ \lambda_n [ \gamma \|s\|_{1 } + \|l\|_\ast ] \\ \mbox{s.t .
} & ~~~ s - l \succ 0 , ~~ s \in \omega , ~~ l \in t_\mathcal{m } , \\ & ~~~ { g_\gamma}({\mathcal{a}}^\dag { \mathcal{i}}^\ast { \mathcal{a}}(s - { s^\ast } , { l^\ast}- l ) ) \leq 11 \lambda_n .
\end{aligned } \label{eq : sdptsmg}\ ] ] we now need to argue that the constraint @xmath554 is also inactive in the convex program .
we proceed by showing that the solution @xmath555 of the convex program has the property that @xmath556 , which concludes the proof of this proposition .
we have from corollary [ theo : nlpcor ] that @xmath557 .
since @xmath558 by assumption , one can verify that @xmath559 & \leq & \frac{8\lambda_n}{\alpha}\left[1 + \frac{\nu}{3(2-\nu ) } \right ] \\ & = & \frac{16 ( 3-\nu ) \lambda_n}{3 \alpha ( 2-\nu ) } \\ & \leq & \min\left\{\frac{1}{4 c_1 } , \frac{\alpha \xi(t)}{64 d \psi c_1 ^ 2}\right\}.\end{aligned}\ ] ] the last line follows from the assumption on @xmath33 .
we also note that @xmath560 from corollary [ theo : nlpcor ] , which implies that @xmath561 . letting @xmath562
, we can conclude from proposition [ theo : bfpt ] that @xmath563 .
next we apply proposition [ theo : rem ] ( as @xmath564 ) to conclude that @xmath565 from the optimality conditions of one can also check that for @xmath566 , @xmath567 \\ & \leq & 4 \left[\frac{2 ( 3-\nu ) \lambda_n}{3 ( 2-\nu)}\right].\end{aligned}\ ] ] here we used in the last inequality , and also that @xmath557 ( as noted above from corollary [ theo : nlpcor ] ) and that @xmath568 .
therefore , @xmath569 because @xmath422 $ ] . based on proposition [ theo : irr ] ( the second part )
, we also have that @xmath570 summarizing steps and , @xmath571 this concludes the proof of the proposition .
[ theo : avtotscor ] under the assumptions of proposition [ theo : avtots ] we have that @xmath572 and that @xmath573 .
moreover , @xmath574 actually has the same inertia as @xmath443 .
we also have that @xmath575 .
the following lemma provides a simple set of sufficient conditions under which the optimal solution @xmath555 of satisfies the optimality conditions of the convex program ( without the tangent space constraints ) .
[ theo : mainlemma ] let @xmath555 be the solution to the tangent - space constrained convex program .
suppose that the assumptions of proposition [ theo : avtots ] hold .
if in addition we have that @xmath576 then @xmath555 is also the unique optimum of the convex program . * proof * : recall from corollary [ theo : avtotscor ] that the tangent space at @xmath574 is equal to @xmath577 .
applying the optimality conditions of the convex program at the optimum @xmath555 , we have that there exist lagrange multipliers @xmath578 such that @xmath579 restricting these conditions to the space @xmath580 , one can check that @xmath581 = -\lambda_n \gamma \mathrm{sign}({s^\ast } ) , ~~ { \mathcal{p}}_{t_\mathcal{m } } [ \sigma^n_o - ( \hat{s}_\omega-\hat{l}_{t_\mathcal{m}})^{-1 } ] = \lambda_n u v^t,\ ] ] where @xmath582 is a reduced svd of @xmath574 . denoting @xmath583
$ ] , we conclude that @xmath584 = z , \label{eq : tscond}\ ] ] with @xmath585 .
it is clear that the optimality condition of the convex program ( without the tangent - space constraints ) on @xmath168 is satisfied .
all we need to show is that @xmath586 ) <
\label{eq : offts}\ ] ] rewriting @xmath587 in terms of the error @xmath588 , we have that @xmath589 restating the condition on @xmath168 , we have that @xmath590 .
\label{eq : tscond2}\ ] ] ( recall that @xmath591 . ) a sufficient condition to show and complete the proof of this lemma is that @xmath592).\ ] ] we prove this inequality next . recall from corollary [ theo : nlpcor ] that @xmath557 .
therefore , from equation we can conclude that @xmath593 ) ) \\
& \leq & \lambda_n + 2\left[\frac{3 \lambda_n \nu}{6 ( 2-\nu ) } \right ] \\ & = & \frac{2\lambda_n}{2-\nu}.\end{aligned}\ ] ] here we used the bounds assumed on @xmath594 and on @xmath595 . applying the second part of proposition [ theo : irr ] , we have that @xmath596 ) \\ & \leq & \lambda_n - { g_\gamma}({\mathcal{p}}_{{\mathcal{y}}^\bot } { \mathcal{a}}^\dag [ - e_n + r_{\sigma^\ast_o } ( { \mathcal{a}}(\delta_s,\delta_l ) ) - { \mathcal{i}}^\ast { \mathcal{c}}_{t_\mathcal{m}}]).\end{aligned}\ ] ] here the second - to - last inequality follows from the bounds on @xmath594 , @xmath595 , and @xmath597 , and the last inequality follows from lemma [ theo : gg ] .
this concludes the proof of the lemma .
@xmath319 all the analysis described so far in this section has been completely deterministic in nature .
here we present the probabilistic component of our proof . specifically
, we study the rate at which the sample covariance matrix converges to the true covariance matrix . the following result from @xcite plays a key role in our analysis : [ theo : davsz ] given natural numbers @xmath598 with @xmath599 ,
let @xmath600 be a @xmath601 matrix with i.i.d .
gaussian entries that have zero - mean and variance @xmath602 . then the largest and smallest singular values @xmath603 and @xmath604 of @xmath600 are such that @xmath605 , \pr\left[{s}_p(\gamma ) \leq 1 - \sqrt{\tfrac{p}{n } } - t \right ] \right\ } \leq \exp\left\{-\tfrac{n t^2}{2}\right\},\ ] ] for any @xmath606 . using this result
the next lemma provides a probabilistic bound between the sample covariance @xmath20 formed using @xmath21 samples and the true covariance @xmath285 in spectral norm .
this result is well - known , and we mainly discuss it here for completeness and also to show explicitly the dependence on @xmath607 defined in .
[ theo : problemma ] let @xmath607 . given any @xmath608 with @xmath609 ,
let the number of samples @xmath21 be such that @xmath610 .
then we have that @xmath611 \leq 2 \exp\left\{-\tfrac{n \delta^2}{128 \psi^2}\right\}.\ ] ] * proof * : since the spectral norm is unitarily invariant , we can assume that @xmath312 is diagonal without loss of generality .
let @xmath612 , and let @xmath613 denote the largest / smallest singular values of @xmath614 .
note that @xmath614 can be viewed as the sample covariance matrix formed from @xmath21 independent samples drawn from a model with identity covariance , i.e. , @xmath615 where @xmath600 denotes a @xmath601 matrix with i.i.d .
gaussian entries that have zero - mean and variance @xmath602 .
we then have that @xmath616 & \leq & \pr\left[\|\bar{\sigma}^n ~ - i\|_2 \geq \tfrac{\delta}{\psi}\right ] \\ & \leq & \pr\left[{s}_1(\bar{\sigma}^n ) \geq 1 + \tfrac{\delta}{\psi } \right ] + \pr\left[{s}_p(\bar{\sigma}^n ) \leq 1 - \tfrac{\delta}{\psi } \right ] \\ & = & \pr\left[{s}_1(\gamma)^2 \geq 1 + \tfrac{\delta}{\psi } \right ] + \pr\left[{s}_p(\gamma)^2 \leq 1 - \tfrac{\delta}{\psi } \right ] \\ & \leq & \pr\left[{s}_1(\gamma ) \geq 1 + \tfrac{\delta}{4\psi } \right ] + \pr\left[{s}_p(\gamma ) \leq 1 - \tfrac{\delta}{4\psi } \right ] \\ &
\leq & \pr\left[{s}_1(\gamma ) \geq 1 + \sqrt{\tfrac{p}{n } } + \tfrac{\delta}{8 \psi } \right ] + \pr\left[{s}_p(\gamma ) \leq 1 - \sqrt{\tfrac{p}{n } } - \tfrac{\delta}{8 \psi } \right ] \\ & \leq & 2 \exp\left\{-\tfrac{n \delta^2}{128 \psi^2}\right\}.\end{aligned}\ ] ] here we used the fact that @xmath617 in the fourth inequality , and we applied theorem [ theo : davsz ] to obtain the final inequality by setting @xmath618 .
@xmath319 [ theo : samples ] let @xmath86 be the sample covariance formed from @xmath21 samples of the observed variables . set @xmath620 . if @xmath621 , then we have with probability greater than @xmath313 that @xmath622 \geq 1 - 2 \exp\{-p\}.\ ] ] in this section we tie together the results obtained thus far to conclude the proof of theorem [ theo : main ] . we only need to show that the sufficient conditions of lemma [ theo : mainlemma ] are satisfied .
it follows directly from corollary [ theo : avtotscor ] that the low - rank part @xmath574 is positive semidefinite , which implies that @xmath555 is also the solution to the original regularized maximum - likelihood convex program with the positive - semidefinite constraint . as usual set @xmath624 , and set @xmath625 . 1 .
let @xmath626 , and let the number of samples @xmath21 be such that @xmath627 note that @xmath628 .
2 . set @xmath620 , and then set @xmath33 as follows : @xmath629 note that @xmath630 .
3 . let the minimum nonzero singular value @xmath132 of @xmath443 be such that @xmath631 where @xmath544 is defined in corollary [ theo : nlpcor ] .
note that @xmath632 .
4 . let the minimum magnitude nonzero entry @xmath280 of @xmath444 be such that @xmath633 where @xmath545 is defined in corollary [ theo : nlpcor ] .
note that @xmath634 .
* proof of theorem [ theo : main ] * : we condition on the event that @xmath635 , which holds with probability greater than @xmath313 from corollary [ theo : samples ] as @xmath621 by assumption .
we note that based on the bound on @xmath21 , we also have that @xmath636.\ ] ] in particular , these bounds imply that @xmath637 and that @xmath638 both these weaker bounds are used later . based on the assumptions above
, the requirements of lemma [ theo : mainlemma ] on the minimum nonzero singular value of @xmath443 and the minimum magnitude nonzero entry of @xmath444 are satisfied .
we only need to verify the bounds on @xmath33 and @xmath594 from proposition [ theo : avtots ] , and the bound on @xmath639 from lemma [ theo : mainlemma ] .
finally we provide a bound on the remainder by applying propositions [ theo : bfpt ] and [ theo : rem ] , which would satisfy the last remaining condition of lemma [ theo : mainlemma ] . in order to apply proposition [ theo : bfpt ] , we note that @xmath644 & \leq & \frac{8}{\alpha } \left[\frac{\nu}{3 ( 2-\nu ) } + 1 \right ] \lambda_n \nonumber \\ & = & \frac{16(3-\nu)\lambda_n}{3 \alpha ( 2-\nu ) } \nonumber \\ & = & \frac{32(3-\nu)d}{\alpha \xi(t ) \nu } \delta_n \label{eq : lambound}\\ & \leq & \min\left\{\frac{1}{4c_1},\frac{\alpha \xi(t)}{64 d \psi c_1 ^ 2 } \right\}. \nonumber\end{aligned}\ ] ] in the first inequality we used the fact that @xmath568 ( from above ) and that @xmath597 is similarly bounded ( from corollary [ theo : nlpcor ] due to the bound on @xmath132 ) . in the second equality we used the relation @xmath642 . in the final inequality we used the bound on @xmath640 from .
this satisfies one of the requirements of proposition [ theo : bfpt ] .
the other condition on @xmath645 is also similarly satisfied due to the bound on @xmath132 from corollary [ theo : nlpcor ] .
specifically , we have that @xmath646 from corollary [ theo : nlpcor ] , and use the same sequence of inequalities as above to satisfy the second requirement of proposition [ theo : bfpt ] .
thus we conclude from proposition [ theo : bfpt ] and from that @xmath647 this bound implies that @xmath648 , which proves the parametric consistency part of the theorem . since the bound also satisfies the condition of proposition [ theo : rem ] ( from the inequality following above we see that @xmath649 ) , we have that @xmath650 \frac{d \delta_n}{\xi(t ) } \\ & \leq & \frac{d \delta_n}{\xi(t ) } \\ & = & \frac{\lambda_n \nu}{6 ( 2-\nu)}.\end{aligned}\ ] ] in the final inequality we used the bound on @xmath640 , and in the final equality we used the relation @xmath642 . this concludes the algebraic consistency part of the theorem . | suppose we have samples of a _ subset _ of a collection of random variables .
no additional information is provided about the number of latent variables , nor of the relationship between the latent and observed variables .
is it possible to discover the number of hidden components , and to learn a statistical model over the entire collection of variables ?
we address this question in the setting in which the latent and observed variables are jointly gaussian , with the conditional statistics of the observed variables conditioned on the latent variables being specified by a graphical model . as a first step we give natural conditions under which such latent - variable gaussian graphical models are identifiable given marginal statistics of only the observed variables .
essentially these conditions require that the conditional graphical model among the observed variables is sparse , while the effect of the latent variables is `` spread out '' over most of the observed variables .
next we propose a tractable convex program based on regularized maximum - likelihood for model selection in this latent - variable setting ; the regularizer uses both the @xmath0 norm and the nuclear norm .
our modeling framework can be viewed as a combination of dimensionality reduction ( to identify latent variables ) and graphical modeling ( to capture remaining statistical structure not attributable to the latent variables ) , and it consistently estimates both the number of hidden components and the conditional graphical model structure among the observed variables .
these results are applicable in the high - dimensional setting in which the number of latent / observed variables grows with the number of samples of the observed variables .
the geometric properties of the algebraic varieties of sparse matrices and of low - rank matrices play an important role in our analysis .
* keywords * : gaussian graphical models ; covariance selection ; latent variables ; regularization ; sparsity ; low - rank ; algebraic statistics ; high - dimensional asymptotics |
let @xmath2 be an @xmath12-dimensional smooth projective variety over @xmath13 with polarization @xmath14 .
let @xmath15 denote the moduli space of rank @xmath16 @xmath17-stable torsion free sheaves on @xmath2 with chern classes @xmath18 .
see @xcite for the notions of torsion free sheaf , stability , the construction of the moduli space , and its history . in this paper , we do _ not _ consider strictly semistable sheaves so this moduli space may be non - compact . in this case , the moduli space is compact . ] .
fixing @xmath19 , one can consider the generating function @xmath20 where @xmath21 denotes topological euler characteristic .
this is a laurent series in @xmath22 when @xmath12 is even and @xmath23 when @xmath12 is odd ( * ? ? ?
these generating functions are fascinating objects with links to combinatorics , number theory , and representation theory .
we mention a few among many references . for @xmath24 ,
generating functions of poincar polynomials were calculated by g. harder and m. s. narasimhan @xcite , u.v .
desale and s. ramanan @xcite , and m. f. atiyah and r. bott @xcite . for @xmath25 and @xmath26 ,
generating functions for poincar polynomials were computed by l. gttsche @xcite . for @xmath25 and @xmath27 ,
these generating functions are related to ( quasi-)modular forms and have been computed in numerous examples @xcite .
this case is related to the s - duality conjecture of physics @xcite . for any @xmath12 and @xmath26
, these generating functions lead to counting @xmath12-dimensional boxes as shown by j. cheah @xcite . for @xmath28 and @xmath2 satisfying @xmath29
, r. p. thomas constructed a perfect obstruction theory on @xmath30 @xcite .
when the moduli space is compact , this leads to a virtual cycle , which can be used to define deformation invariants .
when @xmath2 is in addition calabi - yau , the perfect obstruction theory is symmetric so the virtual cycle is 0-dimensional .
its degree is known as a donaldson - thomas invariant .
k. behrend @xcite showed the dt invariant is also equal to the euler characteristic of @xmath30 weighted by a constructible function @xmath31 @xmath32^{{\mathrm{vir } } } ) = e({\mathcal m}_{x}^{h}(r , c_1 , c_2 , c_3 ) , \nu ) = \sum_{k \in { \mathbb z } } k \ e(\nu^{-1}(\{k\})).\ ] ] therefore , the generating function is related to dt theory .
when @xmath2 is a toric calabi - yau 3-fold and @xmath26 , d. maulik , n. nekrasov , a. okounkov , and r. pandharipande @xcite showed that the generating function of dt invariants ( defined by @xmath4-localization ) is given by counting the @xmath4-fixed points of @xmath33 with signs .
the signs can be seen as a consequence of weighing by the behrend function @xcite .
little is known about the generating function for @xmath34 and @xmath27 .
we study the case @xmath28 , @xmath35 , and @xmath2 is _ toric_. in this case , @xmath2 contains a dense open torus @xmath4 acting on @xmath2 and this action lifts to the moduli space .
elements of the fixed locus of the moduli space are essentially @xmath4-equivariant sheaves on @xmath2 .
equivariant sheaves on toric varieties were studied in depth by a. klyachko @xcite .
he used his methods to calculate for @xmath36 and @xmath35 obtaining the holomorphic part of a quasi - modular form of weight 3/2 @xcite .
other localization calculations on ( stacky / ordinary ) toric surfaces appear in @xcite .
let @xmath37 denote the dual .
any torsion free sheaf @xmath38 can be naturally embedded in its double dual @xmath39 , known as its reflexive hull .
this leads to a short exact sequence @xmath40 where @xmath41 is a sheaf of dimension @xmath42 .
reflexive hulls are _ reflexive sheaves_. a reflexive sheaf is an easy type of torsion free sheaf determined by its restriction to the complement of any closed subset of codimension @xmath43 .
reflexive sheaves on a 3-fold are locally free outside a 0-dimensional subscheme . for general theory on reflexive sheaves
, we refer to @xcite .
klyachko showed that @xmath4-equivariant rank 2 reflexive sheaves are very easy : they are described by attaching a flag of @xmath44 to each ray of the fan of @xmath2 ( his description works for any rank on any @xmath12-fold ) .
denote by @xmath45 the moduli space of rank 2 @xmath17-stable reflexive sheaves on @xmath2 with chern classes @xmath7 , @xmath8 , @xmath9 . taking double
dual gives a map to the scheme theoretic disjoint union @xmath46 reflexive hulls of members of a flat family do not need to form a flat family ( already on surfaces ) , so this map is _ not _ a morphism . nevertheless , by a result of j. kollr @xcite ,
this map is _ _ constructible _ _ such that on each @xmath47 the map @xmath48 is a morphism . ] .
this is very useful for calculating euler characteristics by cut - paste methods . on a surface ,
the cokernel @xmath41 is 0-dimensional and the double dual map leads to a product formula expressing as a `` point contribution '' times the generating function of reflexive ( i.e. locally free ) sheaves @xcite .
two interesting things happen when @xmath28 and @xmath35 : the cokernel @xmath41 need not be 0-dimensional and @xmath49 need not be locally free . since we are interested in euler characteristics , we only need to understand at the level of @xmath4-fixed points , where we can use klyachko s description of @xmath4-equivariant torsion free and reflexive sheaves .
as already mentioned , @xmath4-equivariant reflexive sheaves are very well - behaved . generating functions of @xmath50 were studied in @xcite .
equivariant torsion free sheaves are much harder to enumerate .
at the level of closed points , the fibre of over a reflexive sheaf @xmath49 is the quot scheme @xmath51 of cokernels @xmath52 where @xmath41 has dimension @xmath42 and chern classes @xmath53 , @xmath54 .
suppose @xmath49 is @xmath4-equivariant , then @xmath4 acts on @xmath55 .
the connected components of @xmath4-fixed locus of @xmath55 are indexed by a new type of combinatorial objects , which we call double box configurations ( definition [ doubleboxconf ] ) . moreover , each connected component is isomorphic to a product of @xmath6 s ( proposition [ prodp1s ] ) .
when @xmath41 is 0-dimensional , i.e. @xmath56 , we solve the combinatorics completely in a ( combinatorial ) companion paper @xcite . one basic building block for our computations is @xmath57 : the quot scheme of 0-dimensional quotients @xmath52 of length @xmath12 , where @xmath49 is a @xmath4-equivariant rank 2 reflexive sheaf on @xmath58 .
if @xmath49 is locally free , it decomposes . ] as a sum of two line bundles @xmath59 , which gives an additional @xmath60-action by scaling the factors . localization with respect to this action gives a well - known result @xmath61 where @xmath62 denotes the macmahon function counting 3d partitions .
when @xmath49 is singular ( i.e. not locally free , hence indecomposable@xmath63 ) , this generating function is much harder to obtain . then there are integers @xmath64 , which roughly speaking encode the weights of the three homogeneous generators of @xmath65 . in terms of these weights
, we have ( corollary [ quotc3 ] ) : a[a ] @xmath66 the triple product factor is the generating function of 3d partitions in a box of lengths @xmath67 by @xmath68 by @xmath69 and is therefore a polynomial in @xmath22 ( * ? ? ?
* ( 7.109 ) ) .
the length of the singularity of @xmath49 is @xmath70 ( proposition [ rank2singular ] ) .
therefore , in the singular case we get the `` locally free answer '' @xmath71 times a polynomial counting partitions `` in the singularity of @xmath49 '' .
the way we obtain this formula seemingly has nothing to do with counting inside the singularity of @xmath49 .
instead it is obtained using non - trivial combinatorics involving the double dimer model @xcite .
what does this result tell us about generating function for @xmath28 and @xmath35 ?
recall that we fix @xmath72 and sum over all @xmath9 .
if @xmath8 is chosen such that @xmath73 is minimal with the property that there exists a rank 2 @xmath17-stable torsion free sheaf on @xmath2 with chern classes @xmath74 , then @xmath41 is automatically 0-dimensional ( proposition [ minc2a ] ) . in this case
, we get a structure theorem for ( proposition [ minc2b ] ) . if in addition the polyhedron @xmath75 of the toric 3-fold @xmath2 is small in the sense of definition [ small ] ( e.g. when @xmath76 ) , then all @xmath4-equivariant reflexive hulls in the moduli space are isolated @xmath4-fixed points and the formula simplifies ( theorem [ eulermin ] ) .
a special case of this setting is @xmath77 and @xmath78 . in principle
, our approach also works for more general @xmath8 , but then the combinatorics becomes increasingly complex .
we illustrate this for @xmath77 and @xmath79 .
b[b ] for @xmath77 and @xmath80 , we have @xmath81 let @xmath11 be a smooth toric calabi - yau 3-fold . since @xmath11 is non - compact , we choose a toric compactification @xmath82 with @xmath29 and a polarization @xmath14 on @xmath2 .
define the @xmath4-invariant open subset @xmath83 of torsion free sheaves @xmath38 on @xmath2 such that @xmath84 is supported on @xmath11 .
openness is proved in proposition [ openess ] .
unlike the rank 1 case , this moduli space depends on the choice of @xmath2 and @xmath14 .
this reflects dependence of stability on choice of polarization for higher rank .
when @xmath85 is a compact connected component of @xmath86 , it has an induced virtual cycle by @xcite and one can consider the expression @xmath87^{{\mathrm{vir } } } } \frac{1}{e(n^{{\mathrm{vir}}})},\ ] ] where @xmath21 denotes @xmath4-equivariant euler class and @xmath88 denotes the virtual normal bundle of @xmath85 .
these integrals depend on equivariant parameters @xmath89 .
we are interested in connected components @xmath85 , which also satisfy @xmath90 .
unfortunately , the dt complex restricted to such a component @xmath85 is _ not _ automatically symmetric .
however , if we suppose in addition that all closed points of @xmath85 have the same ( i.e. isomorphic ) reflexive hull @xmath49 , then we can write as @xmath91^{{\mathrm{vir } } } } \frac{1}{e(n^{{\mathrm{vir } } } ) \ e(r{\operatorname{hom}}({\mathcal r},{\mathcal r})_0 \otimes { \mathcal o}_{\mathcal c})},\ ] ] where the second factor of this product is a _ number _ for @xmath92 ( proposition - definition [ propdef ] ) .
we denote this number by @xmath93 .
the reason for this is that after removing @xmath94 from the dt complex on @xmath85 in the @xmath95-group , it becomes symmetric ( proposition [ sd ] ) .
we think of @xmath93 as the contribution of @xmath85 to dt type invariants is by using `` framing at infinity '' as in the work of d. oprea @xcite . in this case , the generating function becomes a product of rank 1 generating functions by using an extra torus action on the framing . in our theory , such a splitting does not occur due to the presence of singularities in the reflexive hull . ] of @xmath11 .
we have a precise conjecture for all @xmath93 in all curve classes on any toric calabi - yau 3-fold ( main conjecture [ mainconj ] ) . for @xmath96 and
fixed @xmath4-equivariant rank 2 reflexive hull @xmath49 on @xmath58 , consider all connected components @xmath85 with constant reflexive hulls @xmath49 and 0-dimensional cokernels .
then the generating function of all @xmath93 is the degree 0 rank 2 equivariant vertex .
we conjecture an explicit formula for it ( conjecture [ edgelessconj ] ) : c[c ] @xmath97 this can be viewed as the degree 0 part of rank 2 dt theory on smooth toric 3-folds .
this formula is in accordance with the expectation that higher rank dt invariants can be expressed in terms of rank 1 dt invariants and `` point contributions '' .
note that unlike the rank 1 case , there are no signs in this formula ( essentially this is because rank 2 is even ) . comparing to and theorem a , conjecture c suggest @xmath98 , but this is _ not _ the case .
in fact , @xmath93 can be zero due to obstructions ( example [ abnormal ] ) .
the following is strong evidence for conjecture c. let @xmath99 denote the `` calabi - yau torus '' defined by @xmath100 .
let @xmath90 be a compact connected component of the @xmath101-fixed locus with constant reflexive hulls @xmath49 and cokernels of dimension 0 .
using @xmath101-localization and assuming two analogs of conjectures . ] used in the @xmath101-localization of stable pair theory @xcite , we prove ( remark [ finalrem ] ) @xmath102 here the sum is over all connected components of the @xmath103-fixed locus of @xmath85 .
combinatorially , this means that whenever @xmath104 for some @xmath105 , other buddy components @xmath106 in the same @xmath60-fixed locus will compensate for this . main conjecture [ mainconj ] is a generalization of conjecture [ c ] to arbitrary smooth toric calabi - yau 3-folds and curve classes .
we provide the following evidence for it : * explicit examples ( section [ nolegs ] , [ o(-1,-1 ) ] ) . * a proof in the case of `` expected obstructions '' ( theorem [ tthm ] ) . * a proof using @xmath101-localization based on analogs of two conjectures of @xcite ( theorem [ t0thm ] ) . in section 2
, we introduce the rank 2 @xmath4-equivariant and @xmath101-equivariant dt invariants we study .
we show the calabi - yau specialization @xmath92 is well - defined ( proposition - definition [ propdef ] ) and @xmath101-fixed dt theory is symmetric in @xmath95-theory ( theorem [ sym ] ) . in section 3 , we recall klyachko s description of @xmath4-equivariant torsion free and reflexive sheaves on a toric variety @xmath2 .
we use it to describe ( scheme - theoretically ) the @xmath4-fixed locus and deduce it is smooth ( theorem [ fixedlocithm ] ) .
this result is a direct consequence of @xcite .
we give a combinatorial description of the fixed locus of the quot scheme of a @xmath4-equivariant rank 2 reflexive hull in terms of new combinatorial objects called double box configurations ( section [ 3dpartitions ] , definition [ doubleboxconf ] ) . in section 4 ,
we discuss applications to euler characteristics .
we provide a general structure formula in the case `` @xmath8 is minimal '' ( proposition [ minc2b ] , theorem [ eulermin ] ) .
we discuss applications to @xmath107 and @xmath77 ( corollary [ quotc3 ] , section [ p3 ] ) . in section 5
, we set up a vertex / edge formalism la @xcite ( theorem [ vertexedgeformalism ] ) . in section 6 , we study the @xmath4-equivariant dt invariants of section 2 after the calabi - yau specialization @xmath92 .
we conjecture a precise formula for all invariants ( main conjecture [ mainconj ] ) .
we introduce a rank 2 equivariant vertex measure in the case of no legs and a single leg ( sections [ nolegs ] , [ o(-1,-1 ) ] ) .
evidence for the main conjecture is provided by many examples ( sections [ nolegs ] , [ o(-1,-1 ) ] , theorem [ tthm ] ) and @xmath101-localization ( theorem [ t0thm ] ) . + * notation . * for any sets @xmath108 , we denote projections to the factors by @xmath109 and @xmath110 . if @xmath108 are schemes with coherent sheaves @xmath111 , we write @xmath112
. + * acknowledgements .
* we would like to thank k. behrend , j. bryan , j. manschot , d. maulik , and r. p. thomas for useful discussions .
was partially supported by nsf grant dms-1406788 .
m.k . was supported by ep / g06170x/1 , `` applied derived categories '' ( while at imperial college ) , a pims postdoctoral fellowship ( crg geometry and physics , while at ubc ) , and nwo - gqt and marie skodowska - curie project 656898 ( invloccy ) ( while at utrecht ) .
let @xmath2 be a smooth toric 3-fold with dense open torus @xmath4 .
we recall some basic facts and notation from toric geometry following @xcite .
the fan of @xmath2 gives rise to a polyhedron @xmath75 with faces transversal to the rays .
denote the collection of vertices of @xmath75 by @xmath113 and the collection of edges of @xmath75 by @xmath114 .
the vertices @xmath115 correspond bijectively to the cones of dimension 3 of the fan of @xmath2 .
we denote the edge between two adjacent vertices @xmath116 by @xmath117 .
the edges @xmath117 correspond bijectively to the cones of dimension 2 of the fan of @xmath2 .
we denote by @xmath118 the @xmath4-invariant affine open subset corresponding to @xmath119 .
furthermore , we write @xmath120 for the fixed point corresponding to @xmath119 and @xmath121 for the @xmath4-invariant line joining @xmath120 and @xmath122 .
the toric divisors are labelled by the faces of @xmath75 .
we denote the collection of faces with the collection of rays of the fan of @xmath2 by sending a face to the ray it is transverse to .
] of @xmath75 by @xmath123 and label its elements by @xmath124 .
the toric divisor corresponding to @xmath124 is denoted by @xmath125 . for any @xmath115
, there are coordinates @xmath126 on @xmath118 and @xmath127 on @xmath128 such that @xmath129 in these coordinates , the fixed point @xmath120 has cotangent representation @xmath130 where @xmath131 denotes @xmath13-span .
we therefore have a @xmath4-equivariant isomorphism @xmath132 the normal bundle @xmath133 is of the form @xmath134 the transition function between the charts @xmath135 , @xmath136 is given by @xmath137 where @xmath138 is defined by @xmath139 .
let @xmath11 be a smooth toric calabi - yau 3-fold .
then @xmath11 is non - compact so has no good notion of moduli space of stable sheaves on it .
we therefore want to compactify @xmath11 .
[ toriccompact ] let @xmath11 be a smooth toric calabi - yau 3-fold .
let @xmath2 be a smooth projective toric 3-fold .
we say @xmath2 is a _
toric compactification _ of @xmath11 if @xmath82 is a union of some of the invariant affine open subsets @xmath140 of @xmath2 . in this context , we denote by @xmath141 the collection of vertices @xmath119 for which @xmath142 .
moreover , we write @xmath143 for the edges spanned by vertices of @xmath141 .
the subscript @xmath144 indicates we consider edges corresponding to _ compact _ lines in @xmath11 only .
let @xmath145 and use coordinates @xmath126 on @xmath135 as above .
define the subtorus preserves the calabi - yau volume form on @xmath11 . ]
@xmath99 by @xmath146 .
equation ( [ klocal ] ) implies there exists a @xmath101-equivariant isomorphism @xmath147 let @xmath148 .
then the degree of @xmath149 is @xmath150 by the calabi - yau property , i.e. @xmath151 [ extoriccy ] important examples of toric compactifications are @xmath152 where @xmath153 denotes the total space of the indicated vector bundle , @xmath154 , and @xmath155 is any smooth projective toric surface .
these compactifications satisfy @xmath29 .
we now introduce some notation for future @xmath95-group calculations .
let @xmath2 be a smooth projective variety , @xmath156 a @xmath157-flat family of coherent sheaves on @xmath2 over a base scheme @xmath157 , and @xmath158 a line bundle on @xmath159 .
we use the common notation @xmath160 moreover , we denote the corresponding @xmath95-group class by @xmath161 \in k_0(b).\end{aligned}\ ] ] the corresponding trace free parts are denoted by @xmath162 , @xmath163 . in the @xmath95-group ,
we often omit square brackets @xmath164 $ ] and write @xmath38 instead of @xmath165 $ ] for brievity .
note that in the case @xmath166 and @xmath167 , we have @xmath168 on @xmath169 , we have an involution @xmath170 , induced by taking the dual .
if @xmath2 is toric and @xmath156 is @xmath4-equivariant , then @xmath171 is an element of the @xmath4-equivariant @xmath95-group @xmath172 .
here we endow @xmath157 with the trivial @xmath4-action .
taking equivariant dual @xmath170 induces an involution on @xmath172 . for
any coherent sheaf @xmath38 on @xmath157 , we have @xmath173 let @xmath2 be a smooth projective @xmath174-fold over @xmath13 with polarization @xmath14 .
we denote by @xmath175 the moduli space of @xmath17-stable torsion free is called torsion free if it has no non - trivial subsheaves of dimension @xmath176 . ]
sheaves on @xmath2 of rank @xmath16 and chern classes @xmath177 .
we will _ not _ consider @xmath17-strictly semistable sheaves in this paper and work directly on the open subset of @xmath17-stable sheaves .
when @xmath2 is toric , the action of the dense open torus @xmath4 on @xmath2 lifts to @xmath178 and we consider the fixed locus @xmath179 . for @xmath180 ,
the fixed locus is in general non - isolated and rather complicated .
an explicit modular description of @xmath179 using git is given in @xcite ( this description holds for @xmath2 of any dimension ) . in section [ fixedloci ] , we explicitly describe the scheme @xmath179 in the case @xmath35 and show it is smooth ( theorem [ fixedlocithm ] ) .
suppose @xmath2 is a toric compactification of a toric calabi - yau 3-fold @xmath11 ( definition [ toriccompact ] ) .
we want to define a moduli space of stable sheaves on @xmath11 using the compactification @xmath2 .
given a torsion free sheaf @xmath38 on @xmath2 , the natural map to its double - dual @xmath181 is injective . here
@xmath182 and @xmath183 is called the reflexive hull of @xmath38 .
reflexive hulls are reflexive sheaves is called reflexive if the natural map @xmath184 is an isomorphism .
reflexive sheaves are torsion free . for more information on reflexive sheaves
see @xcite . ] .
we obtain a short exact sequence @xmath185 the cokernel @xmath41 has proper support of dimension @xmath42 .
moreover , @xmath183 is locally free outside a 0-dimensional subscheme of @xmath2 .
we define @xmath186 \in { \mathcal m}_x \ | \
{ \mathcal f}^ { * * } / { \mathcal f}\ \mathrm{has \ support \ in \ } y \}.\ ] ] we denote the closed points of a finite type @xmath13-scheme @xmath157 by @xmath187 .
[ openness ] @xmath188 is open and @xmath4-invariant .
we first prove @xmath189 is @xmath4-invariant .
let @xmath165 \in { \mathcal m}_{y\subset x}$ ] and @xmath190 be closed points .
then @xmath191 induces an automorphism @xmath192 . by definition
, @xmath38 fits into the short exact sequence @xmath193 where @xmath194 .
this sequence remains exact after applying @xmath195 .
since @xmath195 commutes with taking ( double ) dual , we have @xmath196 since @xmath197 and @xmath11 is @xmath4-invariant , the result follows .
next , we prove @xmath198 is open in @xmath178 by proving that its complement is closed .
suppose @xmath199 is a smooth quasi - projective curve and let @xmath200 be a point on @xmath199 .
suppose we have a morphism @xmath201 such @xmath202 maps to @xmath203 .
it suffices to show @xmath200 also maps to @xmath203 .
let @xmath38 be the family induced by this map .
we know @xmath204 intersects @xmath205 for all closed points @xmath206 and want to show @xmath207 intersects @xmath205 .
after possibly removing a finite number of points from @xmath202 , the reflexive hulls @xmath208 , @xmath206 form a flat family @xcite .
we can take the flat limit of this family inside the ( proper ) moduli space of gieseker semistable torsion free sheaves containing @xmath208 with @xmath206 to fill in the missing fiber over @xmath200 .
this gives a flat family @xmath49 over @xmath199 such that @xmath209 for all closed points @xmath206 , but @xmath210 may not be equal to @xmath211 ( because reflexive hulls of members of a flat family may jump ) . in particular , @xmath210 may not be reflexive .
a priori , @xmath210 is only gieseker semistable and hence @xmath17-semistable . over @xmath202
we also have the flat family of quotients given by the short exact sequence @xmath212 we take the flat limit of these quotients inside the corresponding relative quot scheme @xmath213 where @xmath214 is the hilbert polynomial of @xmath210 with respect to a polarization on @xmath2 . the flat limit exists , because the map is proper .
we obtain a short exact sequence @xmath215 extending .
we claim that @xmath216 .
since @xmath217 and @xmath210 only differ in codimension @xmath43 , @xmath217 is @xmath17-semistable .
there exists a projective moduli scheme of @xmath17-semistable sheaves @xmath218 constructed by d. greb and m. toma @xcite .
seperatedness of @xmath218 , implies equality of the equivalence classes of @xmath219 , @xmath217 in @xmath218 .
this translates into @xmath220^ { * * } \cong [ gr^{\mu}({\mathcal f}'_0)]^{**},\ ] ] where @xmath221 denotes the graded object associated to a choice of @xmath17-jordan - hlder filtration and @xmath164^{**}$ ] denotes double dual - semistable sheaf @xmath38 and choice of @xmath17-jordan - hlder filtration of @xmath38 , the graded object @xmath222 depends on the choice of filtration .
however its double dual @xmath223^{**}$ ] is independent of this choice ( up to isomorphism ) ( * ? ? ?
1.6.10 ) . ]
since @xmath219 is @xmath17-stable this equality implies @xmath217 can not be strictly @xmath17-semistable .
since @xmath217 is @xmath17-stable , separateness of @xmath224 implies @xmath225 .
although @xmath210 and @xmath211 may not be equal , we claim that @xmath226 .
we proved @xmath219 injects into @xmath210 with cokernel @xmath227 , so they are isomorphic on the complement of @xmath228 . since @xmath228 has codimension @xmath43 , their reflexive hulls are completely determined on @xmath229 and @xmath226 ( * ? ? ? * prop .
1.6 ) . from the chain of the inclusions @xmath230
, we deduce @xmath231 so if @xmath232 intersects @xmath205 , we are done . for this
, it suffices to show that the locus of quotients in @xmath233 whose support intersect @xmath234 forms a closed subset .
let @xmath235 be any @xmath155-flat family of torsion free sheaves on @xmath236 for arbitrary base @xmath155 .
suppose the hilbert polynomial of the members of the family is @xmath214 with respect to some polarization on @xmath2 .
consider the corresponding relative quot scheme @xmath237 let @xmath41 be the universal quotient over @xmath238 .
then both @xmath239 and @xmath240 are closed subsets @xmath241 and so is their intersection @xmath242 .
since @xmath2 is proper , the image of @xmath242 under projection to @xmath243 gives a closed subset .
[ openess ] when @xmath29 , openness of @xmath189 implies it inherits the donaldson - thomas perfect obstruction theory from @xmath178 .
since @xmath244 need not be compact , it does not have a virtual cycle .
therefore , instead we consider connected components @xmath85 of @xmath179 which are contained in @xmath244 .
in other words , _ compact _ connected components of @xmath245 .
we will see there are plenty such components .
we then induce the perfect obstruction theory on @xmath85 via @xmath179 and obtain a virtual cycle @xmath246^{{\mathrm{vir}}}$ ] ( see section [ tequivdt ] ) .
a similar story will apply using @xmath101 instead of @xmath4 ( see section [ t0equivdt ] ) . in the case
@xmath26 , the moduli space @xmath247 is independent of the polarization @xmath14 .
moreover , it is isomorphic to a hilbert scheme @xmath248 of ideal sheaves @xmath249 of closed subschemes @xmath199 of dimension @xmath42 with @xmath250=\beta$ ] and @xmath251 . here
@xmath252 is poincar dual to @xmath253 and @xmath254 is determined by @xmath7 , @xmath8 , @xmath9 . in this case
, @xmath244 is the open subset of closed subschemes @xmath199 with support in @xmath11 . from this description
, it is clear that @xmath244 is _ independent _ of the choice of toric compactification @xmath2 .
there is no such independence in the case @xmath27 . in general ,
not only does @xmath244 depend on the choice of toric compactification @xmath2 , it also depends on the choice of polarization the moduli space @xmath178 can be empty . for examples ,
see @xcite . ]
@xmath14 on @xmath2 .
we think of @xmath244 as a moduli space of rank @xmath16 @xmath17-stable torsion free sheaves on @xmath11 . on @xmath244
, we have a type of `` serre duality in the @xmath95-group '' as we will see in proposition [ sd ] below .
we start with a technical lemma on reflexive hulls .
[ hulls ] let @xmath2 be a smooth projective variety with polarization @xmath14 and let @xmath255 be the moduli space of @xmath17-stable torsion free sheaves on @xmath2 with indicated chern classes .
let @xmath156 be a @xmath157-flat family of torsion free sheaves in @xmath224 over a variety @xmath157 , such that all members @xmath256 , @xmath257 have the same reflexive hull @xmath49
. then there exists a line bundle @xmath258 on @xmath157 and a short exact sequence @xmath259 where @xmath260 is @xmath157-flat . for each @xmath257
, this induces a short exact sequence @xmath261 where the first map is the natural inclusion @xmath262 . for @xmath156
_ any _ @xmath157-flat family , kollr @xcite constructs a stratification @xmath263 of @xmath157 such that over each stratum @xmath264 , @xmath265 has a @xmath264-flat hull @xmath266 here @xmath267 is a @xmath264-flat quasi - coherent sheaf and the morphism restricts to the natural inclusion @xmath268 for all @xmath257 . in our case , all members of @xmath156 have isomorphic reflexive hulls , so kollr s stratification is trivial .
we claim @xmath269 is injective with @xmath157-flat cokernel @xmath260 .
consider the exact sequence defined by taking kernel and cokernel @xmath270 fibres over closed points @xmath271 give short exact sequences @xmath272 where the first map is the natural inclusion . since @xmath256 is @xmath157-flat and @xmath273 by assumption
, we conclude that the hilbert polynomials of @xmath274 are constant .
hence @xmath260 is @xmath157-flat .
this implies @xmath275 is @xmath157-flat .
therefore @xmath276 for all closed points @xmath271 , so @xmath277 and @xmath278 .
finally , we show @xmath279 . since @xmath256 is @xmath17-stable , its reflexive hull @xmath49 is @xmath17-stable , because they only differ in codimension @xmath43 .
therefore , @xmath49 is an element of a moduli space of @xmath17-stable reflexive sheaves @xmath280 on @xmath2 .
the @xmath157-flat family @xmath281 gives a morphism @xmath282 suppose @xmath280 has a universal family @xmath283 .
then @xmath284 for some line bundle @xmath258 on @xmath157 .
since @xmath285 factors through the closed point @xmath286\}$ ] the result follows from @xmath287 . in general , @xmath280 may not have a universal family . in this case ,
an tale cover of @xmath280 does have a universal family @xcite .
working on the tale cover , the same result can be obtained .
[ sd ] let @xmath11 be a smooth toric calabi - yau 3-fold and let @xmath2 be a toric compactification of @xmath11 .
let @xmath14 be a polarization on @xmath2 and let @xmath288 , @xmath289 .
let @xmath156 be a @xmath157-flat family of torsion free sheaves in @xmath178 over a variety @xmath157
. assume all members @xmath256 , @xmath257 lie in @xmath244 and have the same reflexive hull @xmath49 .
then @xmath290 if @xmath156 is a @xmath101-equivariant family , then the same equality holds in @xmath291 .
if @xmath156 is a @xmath4-equivariant family , then @xmath292 consider the short exact sequence of lemma [ hulls ] @xmath293 from this short exact sequence , we get exact triangles @xmath294 the natural map @xmath295 induces @xmath296 we define @xmath297 where the map is the natural composition @xmath298 this leads to the following equalities in the @xmath95-group @xmath299 - \langle { \mathbb f},{\mathbb q}\rangle \\ & = \langle { \mathcal r } , { \mathcal r}\rangle_0 \otimes { \mathcal o}_b - \langle { \mathbb q } , { \mathcal r}\boxtimes l \rangle - \langle { \mathbb f},{\mathbb q}\rangle.\end{aligned}\ ] ] the first equality follows from and the definition of @xmath300 .
the second equality follows from and the definition of @xmath300 .
repeating the same reasoning on tensored by @xmath301 gives @xmath302 where we used that @xmath260 is supported in @xmath303 .
we deduce @xmath304 ( equivariant ) serre duality gives the results .
when @xmath2 is a smooth projective 3-fold with @xmath29 ( e.g. calabi - yau or fano ) , the moduli space @xmath178 carries a natural perfect obstruction theory constructed in @xcite @xmath305 where @xmath306 is a 2-term complex of locally free sheaves on @xmath178 and @xmath307 is the truncated cotangent complex of @xmath178 . over a closed point @xmath165 \in { \mathcal m}_x$ ]
, we have @xmath308 } \cong r{\operatorname{hom}}({\mathcal f},{\mathcal f})_0[1].\ ] ] if @xmath2 is in addition toric , the above perfect obstruction theory is @xmath4-equivariant . in the case
@xmath2 is calabi - yau , the virtual dimension @xmath309 of this perfect obstruction theory is 0 .
if @xmath178 is compact , then the perfect obstruction theory @xmath306 gives rise to a virtual cycle @xmath310^{{\mathrm{vir}}}$ ] . in the case
@xmath2 is calabi - yau , the donaldson - thomas invariants are then defined as @xmath311^{{\mathrm{vir } } } ) \in { \mathbb z}.\ ] ] for other 3-folds , we can take any @xmath312 of degree @xmath309 and consider @xmath313^{{\mathrm{vir } } } } \alpha \in { \mathbb z}.\ ] ] when @xmath2 is in addition toric , @xmath119 can be chosen in @xmath314 and the invariant lies in @xmath315.\ ] ] in this case , one can use the virtual torus localization formula of t. graber and r. pandharipande to compute the invariant @xcite .
this works as follows .
the @xmath4-fixed part of @xmath316 induces a perfect obstruction theory on @xmath179 .
moreover , the @xmath4-moving part @xmath317 is a @xmath4-equivariant 2-term perfect complex @xmath318 known as the virtual normal bundle .
consider the insertion @xmath319 where @xmath320 denotes top @xmath4-equivariant chern class .
the virtual localization formula states @xmath313^{{\mathrm{vir } } } } \alpha = \int_{[{\mathcal m}_{x}^{t}]^{{\mathrm{vir } } } } \frac{1}{e(n^{\mathrm{vir } } ) } \ \alpha|_{{\mathcal m}_{x}^{t}}.\ ] ] the rank 1 dt theory and its connections to gromov - witten theory have been studied in the seminal papers @xcite .
let @xmath2 be a smooth projective toric 3-fold with @xmath29 .
consider the moduli space @xmath178 .
so far in this section , the rank @xmath16 was arbitrary .
from now on , @xmath35 . in section [ fixedloci ] , we study the fixed locus @xmath179 for @xmath35 and show it is smooth ( theorem [ fixedlocithm ] ) .
if @xmath178 is projective , then @xmath321^{{\mathrm{vir } } } & = \sum_{{\mathcal c}\subset { \mathcal m}_{x}^{t } } \iota _ * \big ( \frac{1}{e(n^{\mathrm{vir } } ) } \cap [ { \mathcal c}]^{{\mathrm{vir } } } \big ) \\ & = \sum_{{\mathcal c}\subset { \mathcal m}_{x}^{t } } \iota _ * \big ( e(t_{{\mathcal c } } - e^{{{\scriptscriptstyle\bullet}}\vee } ) \cap [ { \mathcal c } ] \big ) , \end{split}\end{aligned}\ ] ] where the sum is over all connected components @xmath322 , @xmath323 denotes inclusion , and @xmath324 denotes the tangent bundle of @xmath85 .
the first equality is the virtual localization formula and the second equality follows from smoothness of @xmath85 as in ( * ? ? ?
4.2 ) . in section [ vertex / edge ] , we develop a vertex / edge formalism for this class . recall that in what follows , we do not assume @xmath178 is projective .
suppose @xmath11 is a smooth toric calabi - yau 3-fold with toric compactification @xmath2 with @xmath29 .
unlike the rank 1 case , for each closed point @xmath165 \in { \mathcal m}_{y \subset x}$ ] , the class @xmath308 } = - \langle { \mathcal f},{\mathcal f}\rangle_0\ ] ] does _ not _ have rank zero even though @xmath325 .
however , the class @xmath308 } + \langle { \mathcal r},{\mathcal r}\rangle_0,\ ] ] has rank zero by proposition [ sd ] ( taking @xmath326 ) .
we therefore want to `` split off '' the part @xmath327 . in order to be able to do this in families
, we make the following assumption : [ as1 ] let @xmath322 be a connected component such that @xmath90 .
assume all closed points of @xmath85 have the same ( i.e. isomorphic ) reflexive hull @xmath49 .
[ unioncomp ] by remark [ a=3 ] , assumption [ as1 ] implies @xmath49 is an isolated reduced point in the @xmath4-fixed locus ( i.e. @xmath328 ) and @xmath85 is a product of @xmath6 s .
conversely , suppose @xmath49 is a @xmath4-equivariant rank 2 @xmath17-stable reflexive sheaf on @xmath2 with @xmath328 .
any connected component @xmath329 with all closed points of @xmath85 having reflexive hull @xmath49 is isomorphic to a product of @xmath6 s , hence compact .
this fact follows from the toric description of section [ fixedlocigeneral ] .
so from _ isolated _ @xmath4-fixed rank 2 reflexive hulls on @xmath2 , we can produce lots of components satisfying assumption [ as1 ] as claimed in remark [ openess ] .
[ ex1as1 ] let @xmath2 a smooth projective toric 3-fold with @xmath76 .
let @xmath14 be a polarization on @xmath2 and @xmath74 chern classes such that @xmath330 and @xmath73 is minimal such that there exist rank 2 @xmath17-stable torsion free sheaves on @xmath2 with chern classes @xmath74 .
then for any connected component @xmath85 of @xmath179 , all closed points of @xmath85 have the same reflexive hull ( theorem [ eulermin ] ) .
an example of this setting is @xmath77 , @xmath331 , and @xmath78 ( example [ c2=1 ] ) . in the previous example , the cokernels @xmath332 of closed points @xmath165 \in { \mathcal c}$ ] are always 0-dimensional by proposition [ minc2a ] .
an example with 1-dimensional cokernels is the following .
[ ex2as1 ] let @xmath77 , @xmath331 , and @xmath79 .
then for all connected components @xmath85 of @xmath179 , the closed points of @xmath85 have the same reflexive hull ( example [ c2=2 ] ) .
moreover , some of these connected components have cokernels @xmath332 of dimension 1 .
[ ex3as1 ] let @xmath77 , @xmath331 , and @xmath333 .
then for some connected components @xmath85 of @xmath179 , all closed points of @xmath85 have the same reflexive hull , whereas for some @xmath85 this is not the case .
this is discussed in detail in example [ c2is3 ] .
let @xmath322 be a connected component satisfying assumption [ as1 ] . recall that @xmath334 and @xmath335 since @xmath29 .
since by assumption @xmath328 , the following expression is well - defined @xmath336 the reflexive hull @xmath49 is not intrinsic to the problem and depends on the choice of compactification , so we factor it out .
the following expression is well - defined @xmath337^{{\mathrm{vir } } } & = e(t_{{\mathcal c } } - e^{{{\scriptscriptstyle\bullet}}\vee } ) \cap [ { \mathcal c } ] \\ & = e(\langle { \mathcal r},{\mathcal r}\rangle_0 \otimes { \mathcal o}_{\mathcal c } ) \cdot e(t_{{\mathcal c } } - e^{{{\scriptscriptstyle\bullet}}\vee } - \langle { \mathcal r},{\mathcal r}\rangle_0 \otimes { \mathcal o}_{\mathcal c } ) \cap [ { \mathcal c } ] .
\end{aligned}\ ] ] we therefore write somewhat sloppily @xmath338^{{\mathrm{vir } } } = e(t_{{\mathcal c } } - e^{{{\scriptscriptstyle\bullet}}\vee } - \langle { \mathcal r},{\mathcal r}\rangle_0 \otimes { \mathcal o}_{\mathcal c } ) \cap [ { \mathcal c}].\ ] ] here @xmath339 has rank zero by proposition [ sd ] . [ propdef ]
let @xmath11 be a smooth toric calabi - yau 3-fold with toric compactification @xmath2 with @xmath29 and polarization @xmath14 .
let @xmath340 , @xmath341 and let @xmath322 be a connected component satisfying assumption [ as1 ] with reflexive hull @xmath49 .
define @xmath342^{{\mathrm{vir } } } } \frac{1}{e(n^{{\mathrm{vir } } } ) \
e(\langle { \mathcal r},{\mathcal r}\rangle_0 \otimes { \mathcal o}_{\mathcal c})}. \end{split}\end{aligned}\ ] ] the specialization @xmath343 is well - defined .
as mentioned above , assumption [ as1 ] implies @xmath344 for some @xmath345 ( remark [ a=3 ] ) . in remark
[ a=3 ] , we also observe that the toric description of section [ fixedloci ] gives a universal family @xmath156 over @xmath85 which fits in a short exact sequence @xmath346 again from the toric description , the cokernel @xmath260 is a @xmath85-flat coherent sheaf and its class in @xmath347 can be expressed in terms of the classes @xmath348 of the factors of @xmath349 .
using this , we show in section [ vertex / edge ] ( theorem [ vertexedgeformalism ] ) that @xmath350 is a finite sum of elements @xmath351 , where each coefficient @xmath144 is the class of a line bundle ( possibly trivial ) on a factor @xmath6 of @xmath349 .
consider @xmath339 .
we are only interested in terms which are @xmath101-fixed ( all other terms have a good specialization under @xmath352 ) .
each @xmath101-fixed term has a weight of the form @xmath353 .
we first consider terms with @xmath354 .
the tangent bundle @xmath324 has weight @xmath355 and its dual @xmath356 has weight @xmath357 ( proposition [ sd ] ) .
denote all other terms with @xmath355 by @xmath358 , where @xmath359 for some factor @xmath6 of @xmath349 and @xmath360 runs through some ( possibly empty ) index set .
note that these terms must come from obstructions and hence have a minus sign .
each such term comes with a dual term @xmath361 by proposition [ sd ] . in conclusion ,
the contribution of all terms with weight @xmath354 to the integral is @xmath362 next consider a @xmath101-fixed term of @xmath339 with @xmath363 .
denote these terms by @xmath364 where @xmath365 for some factor @xmath6 of @xmath349 and @xmath366 runs through some ( possibly empty ) index set . by proposition [ sd ] ,
such a term comes with a dual term @xmath367 together their contribution to is @xmath368 \\ & = \prod_l \big[-\frac{(w_{l}+1)}{w_{l } } + \frac{1}{(w_{l})^2 s } c_1(b_{l } ) \big ] .
\end{split}\end{aligned}\ ] ] multiplying out and gives the contribution of all @xmath101-fixed terms .
we see no powers of @xmath369 can occur since @xmath370 and each factor @xmath371 comes with a class of degree 1 .
moreover after multiplying out and , each non - zero term containing a factor @xmath372 with @xmath373 and @xmath6 equal to the @xmath374th factor of @xmath85 also contains @xmath375 , which cancels the pole .
similarly for @xmath376 . here we use that the factors of @xmath85 are 1-dimensional .
we conclude that all poles cancel . as in the previous sections ,
we consider a smooth toric calabi - yau 3-fold @xmath11 with toric compactification @xmath2 with @xmath29 and polarization @xmath14 .
let @xmath377 be as before .
consider the subtorus @xmath99 defined by @xmath100 . in @xcite , pandharipande - thomas conjecture that the @xmath101-fixed locus of the moduli space of stable pairs is smooth .
we conjecture the analog in our setting : [ t0conj1 ] let @xmath2 be a smooth projective toric 3-fold with polarization @xmath14 and let @xmath378 .
then @xmath379 is smooth .
we consider connected components @xmath380 such that @xmath90 .
we want to endow @xmath85 with a symmetric perfect obstruction theory coming from @xmath101-localization . in the stable pairs setting of @xcite this part is automatic .
however , in our case one must factor out the part coming from reflexive hulls similar to the previous section . in order to achieve this
, we make an assumption : [ as2 ] let @xmath380 be a compact connected component such that @xmath90 .
assume all closed points of @xmath85 have the same ( i.e. isomorphic ) reflexive hull @xmath49 .
moreover , assume @xmath381 .
[ remas2 ] compared to the @xmath4-equivariant case ( assumption [ as1 ] ) , there are two more assumptions : @xmath382 , and @xmath383 . in the @xmath4-equivariant case ,
the first vanishing was a _ consequence _ and the second vanishing was not needed .
we have evidence ( from examples ) that here both vanishings are consequences .
let @xmath380 satisfy assumption [ as2 ] .
then @xmath85 has an induced perfect obstruction theory from @xmath101-localization . assuming conjecture [ t0conj1 ] , we obtain @xmath384^{{\mathrm{vir } } } & = e(t_{{\mathcal c } } - e^{{{\scriptscriptstyle\bullet}}\vee } ) \cap [ { \mathcal c } ] , \\ & = e(\langle { \mathcal r } , { \mathcal r}\rangle_0 \otimes { \mathcal o}_{{\mathcal c } } ) \cdot e(t_{{\mathcal c } } - e^{{{\scriptscriptstyle\bullet}}\vee } - \langle { \mathcal r},{\mathcal r}\rangle_0 \otimes { \mathcal o}_{\mathcal c } ) , \end{aligned}\ ] ] where @xmath385 denotes the virtual normal bundle and @xmath386 denotes the @xmath101-moving part . [ t0dt ]
let @xmath11 be a smooth toric calabi - yau 3-fold with toric compactification @xmath2 with @xmath29 and polarization @xmath14 .
suppose conjecture [ t0conj1 ] holds .
let @xmath340 , @xmath341 , and let @xmath380 be a connected component satisfying assumption [ as2 ] with reflexive hull @xmath49 . then @xmath387^{{\mathrm{vir } } } } \frac{1}{e(n^{{\mathrm{vir}},0 } ) \
e(\langle { \mathcal r } , { \mathcal r}\rangle_0 \otimes { \mathcal o}_{{\mathcal c } } ) } \end{aligned}\ ] ] is well - defined ( by assumption [ as2 ] ) . here
@xmath21 denotes the @xmath101-equivariant euler class .
recall that @xmath85 is a connected component of @xmath388 and not @xmath245 as in proposition - definition [ propdef ] .
therefore , we allow ourselves to use the same notation . a priori , these invariants have equivariant parameters , but theorems [ sym ] , [ t0thm1 ] below imply they are numbers . the @xmath101-equivariant dt invariants of definition [ t0dt ] and the @xmath4-equivariant dt invariants of proposition - definition [ propdef ] are related in the obvious way .
[ t0dt = tdt ] let @xmath11 be a smooth toric calabi - yau 3-fold with toric compactification @xmath2 satisfying @xmath29 and polarization @xmath14 .
suppose conjecture [ t0conj1 ] holds .
let @xmath340 , @xmath341 , and let @xmath380 be a connected component satisfying assumption [ as2 ] with reflexive hull @xmath49 .
then @xmath389 acts on @xmath85 and the connected components @xmath105 of @xmath390 are connected components of @xmath391 .
moreover @xmath392 the formula follows by using the @xmath389-action on @xmath85 and applying virtual localization to @xmath93 . [ sym ]
let @xmath11 be a smooth toric calabi - yau 3-fold with toric compactification @xmath2 with @xmath29 and polarization @xmath14 .
suppose conjecture [ t0conj1 ] holds .
let @xmath340 , let @xmath341 , and let @xmath380 be a connected component satisfying assumption [ as2 ] with reflexive hull @xmath49 .
then dt theory induces a symmetric perfect obstruction theory @xmath393 on @xmath85 .
moreover @xmath394 let @xmath156 be the universal family and smoothness of @xmath85 , one can show that @xmath85 is scheme - theoretically isomorphic to a connected component of the @xmath101-fixed locus of the quot scheme @xmath395 of quotients @xmath396 .
the universal sheaf @xmath156 is the kernel of the universal quotient @xmath397 restricted to @xmath398 . ] on @xmath398 .
then @xmath156 has a @xmath101-equivariant structure . by assumption [ as2 ] and lemma [ hulls ]
, we have a short exact sequence @xmath399 the argument starts as in proposition [ sd ] . from the above short exact sequence ,
we obtain exact triangles @xmath400 the natural map @xmath295 induces an exact triangle @xmath401 we define @xmath402 where the map is the composition @xmath403 from and the definition of @xmath300 , we obtain the exact triangle @xmath404.\ ] ] by assumption [ as2 ] , @xmath405 , hence @xmath406.\ ] ] combining this isomirphism with , gives an exact triangle @xmath407 \longrightarrow r{\curly h\!om}_{p_{\mathcal c}}({\mathbb f},{\mathbb q})^{t_0}.\ ] ] we can apply a similar reasoning to tensored by @xmath408 .
this gives an exact triangle @xmath409 \longrightarrow r{\curly h\!om}_{p_{\mathcal c}}({\mathbb f},{\mathbb q}\otimes p_{x}^{*}k_x ) ^{t_0},\ ] ] where we used @xmath410
\cong 0.\ ] ] since @xmath260 is scheme - theoretically supported on @xmath411 and @xmath11 is calabi - yau , we have @xmath412 we obtain the exact triangle @xmath413 \longrightarrow r{\curly h\!om}_{p_{\mathcal c } } ( { \mathbb f},{\mathbb q})^{t_0}.\ ] ] combining exact triangles , , and a non - zero section @xmath414 gives an isomorphism @xmath415.\ ] ] this isomorphism provides the non - degenerate symmetric bilinear form required for a symmetric perfect obstruction theory \cong \theta$ ] , where @xmath416 is the map to the cotangent complex @xcite . in our applications
, we only use the isomorphism @xmath417 \cong e^{{{\scriptscriptstyle\bullet}}}$ ] ( namely for calculations in the @xmath101-equivariant @xmath95-group ) .
therefore , we do not prove the duality of @xmath418 . ] @xmath419.\ ] ] next , the second part of the theorem .
exactly as in the proof of proposition [ sd ] , we obtain the following equalities in @xmath420 @xmath421 subtracting these equations and using @xmath101-equivariant serre duality gives the desired result .
in this section , @xmath2 is a smooth projective toric 3-fold with polarization @xmath14 and @xmath422 denotes the moduli space of rank @xmath16 @xmath17-stable torsion free sheaves on @xmath2 with indicated chern classes .
we give klyacho s description of @xmath4-equivariant torsion free sheaves on @xmath2 @xcite .
klyachko s work was further developed by m. perling @xcite and we will mostly follow his notation .
most constructions of sections 3.13.3 work for @xmath2 of any dimension , but we stick to dimension 3 for notational convenience .
let @xmath38 be a @xmath4-equivariant quasi - coherent sheaf on @xmath2 and @xmath115 .
let @xmath423 be coordinates on @xmath118 as in section [ toric3folds ] .
we first study the restriction @xmath424 .
the global section functor @xmath425 gives an equivalence between the categories of quasi - coherent sheaves on @xmath135 and @xmath426$]-modules .
it is not hard to extend this to an equivalence between the categories of @xmath4-equivariant quasi - coherent sheaves on @xmath135 and @xmath426$]-modules with regular @xmath4-action @xcite .
a regular @xmath4-action corresponds to a decomposition into weight spaces according to the character group @xmath427 .
given a @xmath4-equivariant quasi - coherent sheaf @xmath428 on @xmath135 , we have a decomposition into weight spaces @xmath429 the sheaf @xmath428 is coherent if and only if @xmath430 has a finite number of homogeneous generators
. moreover @xmath428 is torsion free if and only if the multiplication maps @xmath431 are injective on all weight spaces @xmath432 @xcite .
therefore , the category of @xmath4-equivariant rank @xmath16 torsion free sheaves @xmath428 on @xmath135 is equivalent to the category of collections of complex vector spaces @xmath433 satisfying : * the vector spaces @xmath432 form a triple filtration @xmath434 * for all @xmath435 we have @xmath436 .
* whenever @xmath437 or @xmath438 or @xmath439 we have @xmath440 .
we refer to @xmath433 as a rank @xmath16 @xmath441-family denotes the cone of maximal dimension corresponding to vertex @xmath119 . in our notation
@xmath119-family would perhaps be better terminology . ] .
the maps in the category of @xmath4-equivariant rank @xmath16 torsion free sheaves on @xmath135 are @xmath4-equivariant morphisms and the maps in the category of rank @xmath16 @xmath441-families are endomorphisms of @xmath442 compatible with the triple filtrations .
a @xmath4-equivariant torsion free sheaf @xmath38 on @xmath2 gives rise to @xmath4-equivariant torsion free sheaves @xmath443 and corresponding @xmath441-families @xmath444 conversely , given such a collection of @xmath441-families , when do the corresponding sheaves @xmath428 glue to a sheaf @xmath38 on @xmath2 ? note that for fixed @xmath445 , the vector space @xmath446 does not change for @xmath447 .
we denote this `` limiting vector space '' by @xmath448 .
it is not hard to show that the gluing conditions are given by the following equations @xcite @xmath449 for all @xmath116 and @xmath450 . here
the coordinates are chosen such that @xmath138 is defined by equations @xmath139 in @xmath135 and @xmath451 in @xmath136 .
this describes an equivalence between the category of @xmath4-equivariant rank @xmath16 torsion free sheaves on @xmath2 and collections of rank @xmath16 @xmath441-families satisfying the gluing conditions ( [ glue ] ) @xcite .
the maps in both categories are the obvious ones .
let @xmath38 be a @xmath4-equivariant torsion free sheaf on a smooth projective toric 3-fold @xmath2 with corresponding collection of @xmath441-families @xmath452 .
we define the _ characteristic function _
@xmath453 of @xmath38 as follows @xmath454 we denote the collection of characteristic functions of @xmath4-equivariant torsion free sheaves on @xmath2 by @xmath455 .
characteristic functions of members of a flat family of @xmath4-equivariant coherent sheaves are locally constant over the base @xcite .
one can construct projective coarse moduli spaces of @xmath4-equivariant torsion free sheaves with fixed characteristic function using git .
moreover , one can choose a git stability condition for which git stability and gieseker stability coincide @xcite .
let @xmath38 be a @xmath4-equivariant torsion free sheaf on @xmath2 .
then there is an explicit formula for the chern character @xmath456 in terms of the characteristic function @xmath453 .
this formula is due to klyachko @xcite .
an alternative way for obtaining such a formula is by @xmath4-equivariant dvissage .
concretely , this works as follows .
the cohomology ring @xmath457 is explicitly determined by the fan ( or polyhedron ) of @xmath2 .
it is the free @xmath458-module generated by the @xmath4-invariant closed subsets @xmath459 modulo certain relations @xcite .
for example @xmath460 for all @xmath116 is such a relation .
there are three linear relations among the toric prime divisors @xmath125 and certain relations among the @xmath138 .
the @xmath4-equivariant picard group @xmath461 is generated by @xmath125 _ without _ relations . inside the @xmath4-equivariant @xmath95-group @xmath462
, there is an explicit way of writing @xmath38 as a finite sum of elements of the form @xmath463 for @xmath464 .
the algorithm for getting such a decomposition is explicitly described in @xcite for the case @xmath77 , @xmath35 , and @xmath38 reflexive .
see example [ p3reflch ] below for the equations .
it is straightforward to generalize the algorithm to any @xmath2 and @xmath38 a @xmath4-equivariant rank @xmath16 torsion free sheaf on @xmath2 . for any @xmath465
, we write @xmath466 for the chern character determined by a characteristic function @xmath467 .
characteristic function determines chern character and hence hilbert polynomial ( by hirzebruch - riemann - roch ) . for a fixed chern character @xmath468 ,
we denote by @xmath469 the collection of characteristic functions @xmath467 satisfying @xmath470 . a @xmath4-equivariant rank @xmath16 torsion free sheaf @xmath38 on @xmath2 is @xmath17-stable if and only if its reflexive hull @xmath183 is @xmath17-stable .
klyachko @xcite found out that @xmath4-equivariant reflexive sheaves on @xmath2 have a particularly easy description .
moreover , stability is easily described in terms of the characteristic function @xmath471 @xcite .
we denote by @xmath366 the number of faces of the polyhedron @xmath75 .
note that @xmath366 equals the rank of @xmath461 .
klyachko @xcite gives an equivalence between the category of @xmath4-equivariant rank @xmath16 reflexive sheaves on @xmath2 and the category of collections of flags @xmath472 satisfying * @xmath473 for all @xmath474 , * @xmath475 for all @xmath476 .
the morphisms of both categories are the obvious ones .
we now describe how a family of flags @xmath477 gives rise to a @xmath4-equivariant reflexive sheaf @xmath49 on @xmath2 .
let @xmath115 and let @xmath118 be the corresponding @xmath4-invariant affine open subset .
write the coordinates on @xmath135 by @xmath423 as in section [ toric3folds ] .
suppose the toric divisor @xmath478 corresponds to ray @xmath479 .
define vector spaces @xmath480 then @xmath481 is a collection of rank @xmath16 @xmath441-families satisfying the gluing conditions .
the corresponding sheaf @xmath49 on @xmath2 is reflexive @xcite . in the case
@xmath35 , a family of filtrations @xmath477 is entirely determined by the integers where the dimensions jump together with a choice of 1-dimensional subspaces for the jumps .
more precisely , for each @xmath124 , there are unique @xmath482 , @xmath483 and @xmath484 such that @xmath485 where @xmath486 does not occur when @xmath487 .
[ toricdata ] let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath2 corresponding to a collection of filtrations @xmath488 .
we refer to the data @xmath489 described above as the _ toric data _ of @xmath49 and abbreviate it by @xmath490
. given toric data @xmath490 , we define @xmath491 we abbreviate the collections of these number by @xmath492 , @xmath493 . clearly the the numbers @xmath494 determine the characteristic function of @xmath495 and vice versa .
we treat both notions on equal footing .
[ p3reflch ] the toric 3-fold @xmath77 is described by the lattice @xmath496 .
the fan @xmath497 consisting of the 3-dimensional cones @xmath498 here @xmath499 is the standard basis of @xmath500 .
we denote the rays generated by @xmath501 , @xmath502 , @xmath503 , @xmath504 by @xmath505 , @xmath506 , @xmath507 , @xmath508 .
the corresponding polyhedron is a tetrahedron and @xmath509 .
we define @xmath510 , @xmath511 , @xmath512 , and @xmath513 .
let @xmath514 denote the hyperplane class on @xmath2 and let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath2 described by toric data @xmath490 . in @xcite the chern classes of @xmath49
are computed using @xmath4-equivariant dvissage as outlined in section [ fixedloci1 ] @xmath515 this expresses @xmath516 in terms of the characteristic function @xmath517 . by further removing chern characters of @xmath518 , @xmath519 with @xmath520
, we obtain explicit expressions for @xmath456 in terms of @xmath521 for any given @xmath4-equivariant rank 2 torsion free sheaf @xmath38 on @xmath522 .
the following proposition is very useful .
[ rank2singular ] let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath2 with toric data @xmath490 .
let @xmath115 and denote by @xmath523 the faces sharing vertex @xmath119 .
then @xmath524 is singular , i.e. not locally free , precisely if @xmath525 , @xmath526 , @xmath527 are all positive and @xmath528 , @xmath529 , @xmath530 are mutually distinct .
if @xmath524 is singular , then the length of its singularity equals the product @xmath531 .
moreover , the degree of @xmath532 equals the sum of the lengths of the singularities of @xmath533 . for @xmath77
this is immediate from ( * ? ? ?
2.6 ) and the formula for @xmath532 of example [ p3reflch ] . for general @xmath2 , the proof is easily adapted as is discussed in detail in the proof of ( * ? ? ?
3.6 ) . back to general @xmath2 .
let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath2 described by toric data @xmath490 .
like in the case of chern classes , @xmath17-stability of @xmath49 can be described in terms of the toric data only .
more precisely @xmath49 is @xmath17-stable if and only if for any 1-dimensional subspace @xmath534 @xmath535 where we recall that @xmath14 denotes the polarization and @xmath536 the toric divisor corresponding to @xmath124 ( section [ toric3folds ] ) .
a similar description exists for any rank on any polarized smooth projective toric @xmath12-fold @xcite .
let @xmath38 be a @xmath4-equivariant torsion free sheaf on @xmath2
. then @xmath38 is @xmath17-stable if and only if its reflexive hull @xmath537 is @xmath17-stable .
so in order to understand @xmath17-stability of such sheaves , we only need to write the family of filtrations @xmath477 corresponding to @xmath49 in terms of the collection of @xmath441-families @xmath452 corresponding to @xmath38 .
this can be done as follows : for all @xmath115 denote by @xmath505 , @xmath506 , @xmath507 the faces sharing vertex @xmath119 , then @xmath538 for all @xmath154 .
let @xmath465 be the characteristic function of a @xmath4-equivariant rank 2 torsion free sheaf on @xmath2 .
there exists a natural moduli functor of families of @xmath17-stable - stability is easier . ]
@xmath4-equivariant torsion free sheaves on @xmath2 with characteristic function @xmath467 @xcite . using the previous section , a fairly straight - forward git construction yields a projective @xmath13-scheme corepresenting this functor .
in this section we describe the outcome . for details and proofs ( in a much more general setting ) see @xcite .
[ comp1s ] let @xmath539 and consider @xmath540 .
let @xmath541 , then @xmath155 is called a _ connected component of 1 s _ associated to @xmath540 when @xmath542 [ vefcomps ] let @xmath539 and fix @xmath115 .
consider a connected component of 1 s , @xmath543 , associated to @xmath540 . *
if @xmath543 is bounded in @xmath500 , then we refer to it as a _ vertex component _ associated to @xmath540 . *
the connected component @xmath543 is called a _ local face component _ associated to @xmath540 if it satisfies the following property .
suppose @xmath124 is a face containing the vertex @xmath119 and @xmath544 .
suppose , without loss of generality , the toric divisor @xmath125 corresponds to @xmath545 and write @xmath546 , @xmath547 .
then for any integers @xmath548 and @xmath549 , the point @xmath550 lies in @xmath543 .
* in all other cases @xmath543 is called a _ local edge components _ associated to @xmath540 .
[ vefcomps2 ] let @xmath539 . the toric data @xmath490 of any reflexive hull of a @xmath4-equivariant torsion free sheaf on @xmath2 with characteristic function @xmath467 has the same values @xmath551 , @xmath552 .
consider all connected components of 1 s of all @xmath553 at once .
it is clear that connected components of 1 s associated to @xmath540 , @xmath554 glue just like @xmath441-families .
vertex components will never glue , but local face or local edge components associated to one @xmath540 can glue to local face or local edge components associated to another @xmath554 . note that local face components may glue to local edge components .
after identification , we refer to a global connected component of 1 s containing a local face component as a _
face component _ associated to @xmath555 .
all other global connected components of 1 s obtained from gluing are called _ edge components _ associated to @xmath555 .
we denote the number of face , edge , vertex components associated to @xmath555 by @xmath556 .
let the situation be as in the previous definition .
the closed points of @xmath557 are in bijective correspondence with all possible collections of @xmath441-families of @xmath4-equivariant torsion free sheaves on @xmath2 with characteristic function @xmath467 ( up to isomorphism ) .
moreover , two points correspond to isomorphic objects if and only if they lie in the same @xmath558-orbit . here
@xmath558 acts by matrix multiplication on each factor @xmath6 .
the @xmath558-equivariant line bundles on this space are in 1 - 1 correspondence with elements of @xcite @xmath559 via the correspondence @xmath560 we want to pick an @xmath558-equivariant line bundle for which the properly git stable points exactly correspond to the collections of @xmath441-families for which the corresponding sheaf is @xmath17-stable . using the previous section , one can show that the following choice works @xcite @xmath561 where the sum is over all faces @xmath124 for which @xmath562 and the corresponding local face component is contained in a ( global ) face component indexed by @xmath374 .
the stability does not depend on the edge and vertex components but only on the face components .
in particular , we get an open subset @xmath563 such that the closed points of @xmath564 precisely correspond to @xmath17-stable sheaves . unfortunately the above @xmath558-equivariant line bundle is not ample .
for ampleness , all @xmath565 must be positive .
therefore we fix @xmath566 and take @xmath567 where the sum over @xmath568 is as above .
then the properly git stable locus is still @xmath569 and the git quotient @xmath570 is a quasi - projective variety whose closed points exactly correspond to @xmath4-equivariant isomorphism classes of @xmath4-equivariant rank 2 @xmath17-stable torsion free sheaves on @xmath2 with characteristic function @xmath467 .
the space @xmath571 is projective when @xmath572 .
moreover , @xmath571 is smooth , because @xmath573 is smooth and @xmath574 acts without stabilizers .
one can show that @xmath571 indeed corepresents a natural moduli functor as follows @xcite : [ equivmod ] let @xmath2 be a smooth projective toric 3-fold with polarization @xmath14 .
let @xmath575 be the characteristic function of a @xmath4-equivariant rank 2 torsion free sheaf on @xmath2 .
then @xmath576 is a smooth quasi - projective variety corepresenting the moduli functor of families of @xmath4-equivariant @xmath17-stable torsion free sheaves on @xmath2 with characteristic function @xmath555 .
[ a=3 ] in the above theorem , the double dual map is given by @xmath577 we deduce that if @xmath578 , then @xmath579 ( or else the connected component of reflexive hulls @xmath580 ) .
we also deduce : all closed points of @xmath571 have the same reflexive hull if and only if @xmath581 .
suppose this is the case .
then @xmath582 .
moreover , @xmath583 is an isolated reduced point @xmath584\}$ ] of the @xmath4-fixed locus of the moduli space of @xmath17-stable reflexive sheaves on @xmath2 .
therefore @xmath585 we have a tautological line bundle @xmath586 on each factor @xmath6 of @xmath587 . using the tautological inclusions
@xmath588 it is easy to construct a universal family @xmath156 on @xmath587 fitting in a short exact sequence @xmath589 where the cokernel @xmath260 is @xmath587-flat .
let us go back to the full moduli space @xmath340 and its fixed locus @xmath179 .
denote by @xmath590 the collection of characteristic functions of @xmath4-equivariant rank 2 torsion free sheaves on @xmath2 with chern classes @xmath177 . forgetting the @xmath4-equivariant structure defines a map @xmath591 it is not hard to show this is a morphism @xcite .
any element of @xmath179 admits a @xmath4-equivariant structure , which is unique up to tensoring by a character of @xmath592 @xcite .
the map is therefore surjective but not injective .
it can be made injective as follows .
pick any vertex @xmath593 and denote by @xmath594 the faces having @xmath119 as a vertex .
define @xmath595 as the subset of characteristic functions @xmath467 for which the corresponding @xmath596 satisfy @xmath597 then @xmath598 is a bijective morphism .
in fact , it is an isomorphism of schemes and any smooth projective toric @xmath12-fold . ]
* cor . 4.10 ) : [ fixedlocithm ] let @xmath2 be a smooth projective toric 3-fold with polarization @xmath14 and let @xmath340 . then the forgetful map described above is an isomorphism of schemes @xmath599 in particular , @xmath179 is a smooth quasi - projective variety .
as discussed in the introduction , double duals of members of a flat family of torsion free sheaves do not need to form a flat family .
however , a result of kollr @xcite implies the double dual map is constructible @xmath600 at the level of closed points , the fibre over @xmath601 \in { \mathcal n}_{x}^{h}(2,c_1,c_{2}',c_{3}')$ ] is the quot scheme @xmath602 of quotients of @xmath49 of dimension @xmath42 , such that @xmath603 in this section , we develop a combinatorial way of indexing the connected components of @xmath604 by a new type of combinatorial objects called double box configurations .
each connected component is isomorphic to a product of @xmath6 s as we will show first .
suppose @xmath49 is any @xmath4-equivariant rank 2 reflexive sheaf on @xmath2 described by toric data @xmath490 ( definition [ toricdata ] ) .
denote by @xmath605 the collection of characteristic functions of @xmath4-equivariant subsheaves of @xmath49 such that the cokernel has chern classes @xmath606 , @xmath607 .
the closed points of @xmath604 are @xmath4-invariant submodules of @xmath608 satisfying the gluing conditions ( [ glue ] ) .
this gives a decomposition into connected components @xmath609 [ prodp1s ] each @xmath610 of is isomorphic to a product of @xmath6 s .
this can be proved using the methods of @xcite .
in fact , this case is much easier than theorem [ fixedlocithm ] .
step 1 : introduce a moduli functor of @xmath4-equivariant families of quotients of @xmath49 with chern classes @xmath611 as in ( * ? ? ?
step 2 : describe these families via a family version of klyachko s filtrations like in ( * ? ? ?
step 3 : construct the moduli spaces @xmath610 as git quotients and show they corepresent the previous moduli functors ( * ? ? ?
step 4 : lift the action of @xmath4 on @xmath2 to @xmath612 ( done in the proof of ( * ? ? ?
step 4 : show the forgetful morphism is an isomorphism .
this goes similar to ( * ? ? ?
4.3 ) but much easier : this time elements of @xmath604 admit a _ unique _ @xmath4-equivariant structure .
fix @xmath115 and @xmath613 .
our goal is to give a combinatorial description of characteristic functions @xmath254 of rank 2 @xmath4-equivariant subsheaves of @xmath49 .
we denote the collection of such characteristic functions by @xmath614 .
let @xmath594 be the faces with vertex @xmath119 and suppose the labeling is chosen such that the toric divisor @xmath615 in chart @xmath135 corresponds to coordinate hyperplane @xmath616 .
then @xmath254 gives rise to integers ( definition [ toricdata ] ) @xmath617 moreover , @xmath618 is described by toric data ( definition [ toricdata ] ) @xmath619 this data is represented by figure 1 .
the sheaf @xmath618 is singular if and only if @xmath620 and @xmath621 are mutually distinct ( proposition [ rank2singular ] ) .
we discuss the case @xmath618 is singular .
the case @xmath618 is locally free is treated in remark [ deco ] . in
what follows , we let @xmath622 for all @xmath117 .
@xmath623 we start with some definitions . given a point @xmath624 in @xmath625 , we write @xmath626 a _ 2d partition _ based at @xmath627 is a subset @xmath628 satisfying the following condition .
if @xmath629 then @xmath630 , @xmath631 , and @xmath632 , @xmath633 for all @xmath634 , @xmath635 .
we call @xmath636 _ finite _ if @xmath637 .
[ doublesquare ] let @xmath638 , @xmath639 , or @xmath640 .
consider a triple of _ finite _ 2d partitions @xmath641 based respectively at @xmath642 , @xmath643 , @xmath644 and subject to the following condition .
let @xmath645 , @xmath646 , and @xmath647 .
define @xmath648 _ condition _ : each element of @xmath649 belongs to _ exactly two _ of the partitions
we define an equivalence relation @xmath651 on the collection of such triples @xmath641 as follows .
write @xmath652 whenever : 1 .
@xmath653 for all @xmath654 , 2 .
@xmath655 , 3 .
@xmath656 .
we refer to such equivalence classes as _ double square configurations _ in @xmath657 and denote the collection of such equivalence classes by @xmath658 . for any @xmath659
, we define the _ size _ @xmath660 as @xmath661 note that @xmath660 is independent of the choice of representative of the equivalence class .
let @xmath662 and @xmath663 , then the following triples of 2d partitions are equivalent @xmath664{1 + 1 + 1 + 1 } , \includegraphics[width=.21in]{1 + 1 } , \includegraphics[width=.12in]{s } ) , ( \includegraphics[width=.4in]{1 + 1 + 1 + 1 } , \includegraphics[width=.21in]{2 + 1 } , \varnothing ) , ( \includegraphics[width=.3in]{1 + 1 + 1 } , \includegraphics[width=.21in]{2 + 1 } , \includegraphics[width=.12in]{s}).\ ] ] given a point @xmath665 in @xmath666 , we write @xmath667 referring to figure 1 , we are interested in the following cones @xmath668 for all @xmath669 , let @xmath670 , @xmath671 denote projection .
a _ 3d partition _ based at @xmath672 is a subset @xmath673 satisfying the following condition .
if @xmath674 , then @xmath630 , @xmath631 , @xmath675 , and @xmath676 , @xmath677 , @xmath678 for all @xmath634 , @xmath635 , @xmath679 .
we call @xmath680 _ finite _ if @xmath681 .
[ doubleboxconf ] fix outgoing double square configurations @xmath682 for all @xmath683 .
consider a triple of ( not necessarily finite ) 3d partitions @xmath684 based at @xmath685 , @xmath686 , @xmath687 subject to the two conditions described below .
define @xmath688 take @xmath689 sufficiently large such that the following projections do not depend on @xmath690 @xmath691 \times [ u_2+v_2 , \infty ] \times [ u_3+v_3,\infty ] ) , \\
\lambda_{2}(\pi_1 ) & : = p_2(\pi_1 \setminus [ u_1,\infty ] \times [ u_2+v_2 , u_2+v_2+n ] \times [ u_3+v_3,\infty ] ) , \\
\lambda_{3}(\pi_1 ) & : = p_3(\pi_1 \setminus [ u_1,\infty ] \times [ u_2+v_2 , \infty ] \times [ u_3+v_3,u_3+v_3+n ] ) , \\
\lambda_{1}(\pi_2 ) & : = p_1(\pi_2 \setminus [ u_1,u_1+v_1+n ] \times [ u_2 , \infty ] \times [ u_3+v_3,\infty ] ) , \\
\lambda_{2}(\pi_2 ) & : = p_2(\pi_2 \setminus [ u_1+v_1,\infty ] \times [ u_2 , u_2+n ] \times [ u_3+v_3,\infty ] ) , \\
\lambda_{3}(\pi_2 ) & : = p_3(\pi_2 \setminus [ u_1+v_1,\infty ] \times [ u_2 , \infty ] \times [ u_3+v_3,u_3+v_3+n ] ) , \\
\lambda_{1}(\pi_3 ) & : = p_1(\pi_3 \setminus [ u_1,u_1+v_1+n ] \times [ u_2+v_2 , \infty ] \times [ u_3,\infty ] ) , \\
\lambda_{2}(\pi_3 ) & : = p_2(\pi_3 \setminus [ u_1+v_1,\infty ] \times [ u_2+v_2 , u_2+v_2+n ] \times [ u_3,\infty ] ) , \\
\lambda_{3}(\pi_3 ) & : = p_3(\pi_3 \setminus [ u_1+v_1,\infty ] \times [ u_2+v_2 , \infty ] \times [ u_3,u_3+n ] ) .
\end{aligned}\ ] ] _ condition 1 _ : each element of @xmath692 belongs to _ exactly two _
partitions @xmath693 .
+ _ condition 2 _ : @xmath694 we define an equivalence relation @xmath651 on the collection of such triples as follows .
write @xmath695 whenever : 1 .
@xmath696 for all @xmath683 , 2 .
@xmath697 , 3 .
@xmath698 .
we refer to such equivalence classes as _
double box configurations _ in @xmath618 and denote the collection of such equivalence classes by @xmath699 . for any @xmath700
, we define the _ size _ of @xmath701 as @xmath702 where @xmath703 on the rhs denotes the renormalized volume of a 3d partition @xcite .
note that @xmath704 is independent of the choice of representative of the equivalence class .
[ modfactor ] to any @xmath700 we associate a space @xmath705 as follows .
analogous to definition [ comp1s ] , the partition @xmath706 can be written as a union of connected components .
define the regions @xmath707 \times [ u_3+v_3,\infty ] , \\
s_2 & : = [ u_1+v_1,\infty ] \times \{u_2+v_2 - 1\ } \times [ u_3+v_3,\infty ] , \\
s_3 & : = [ u_1+v_1,\infty ] \times [ u_2+v_2,\infty ] \times \{u_3+v_3 - 1\}. \end{aligned}\ ] ] let @xmath670 , @xmath683 denote projections as before . if @xmath543 is a vertex or local edge connected component of @xmath701 ( definitions [ vefcomps ] , [ vefcomps2 ] ) such that @xmath708 for all @xmath683 , then we associate a copy of @xmath6 to it . note that this is independent of choice of representative of the equivalence class .
otherwise the connected component has no moduli assigned to it . taking the product of all connected components satisfying ( [ proj ] )
, we define the _ moduli factor _ associated to @xmath701 as @xmath709 we illustrate these definitions in the following examples .
let @xmath710 be given by figure 2 .
two connected components fail ( [ proj ] ) so have no moduli .
one connected component satisfies ( [ proj ] ) , so @xmath711 .
.,width=288 ] [ moduliinlegs ] let @xmath712{1 + 1 + 1 } , \includegraphics[width=.21in]{1 + 1 } , \includegraphics[width=.12in]{s } ) , ( \varnothing , \includegraphics[width=.12in]{s } , \varnothing ) , ( \includegraphics[width=.3in]{2 + 2 + 2},\includegraphics[width=.2in]{4 + 4},\includegraphics[width=.2in]{2 + 1}))$ ] be given by figure 3 .
one connected component ( of vertex type ) fails ( [ proj ] ) so has no moduli .
the other two connected components ( of edge type ) satisfy ( [ proj ] ) so @xmath713 .
.,width=192 ] we formulate our main combinatorial proposition : [ lem1 ] let @xmath618 be a @xmath4-equivariant rank 2 reflexive sheaf on toric chart @xmath118 .
let @xmath714 , @xmath715 , and @xmath716 .
denote by @xmath614 the collection of characteristic functions of @xmath4-equivariant rank 2 torsion free sheaves on @xmath135 with @xmath4-equivariant reflexive hull @xmath618 .
then there is a natural bijection @xmath717 let @xmath718 be the characteristic function of a @xmath4-equivariant rank 2 torsion free sheaf on @xmath135 with reflexive hull @xmath618 described by toric data @xmath490 . consider the regions @xmath719 of .
the region @xmath720 has zeros and ones .
these zeros form a 3d partition @xmath721 .
the region @xmath722 has zeros , ones , and twos .
the zeros give rise to a 3d partition @xmath723 .
the partition @xmath723 has the property that @xmath724 are 3d partitions . finally , consider the connected components of 1 s of @xmath722 .
such a connected component @xmath543 must have the property that @xmath725 is a 3d partition .
moreover , at least two of @xmath726 are 3d partitions . if only two of these form a 3d partition , say ( without loss of generality ) @xmath727 , then define @xmath728 otherwise choose any two , say @xmath729 , and define @xmath728 proceeding in this way for all connected components @xmath543 , we end up with 3d partitions @xmath730 then @xmath731 forms a double configuration of boxes .
we described a map @xmath732 it is easy to check this is a bijection .
[ littlebox ] let @xmath733{1 + 1 + 1 + 1 } , \includegraphics[width=.21in]{1 + 1 } , \varnothing ) , ( \varnothing , \includegraphics[width=.12in]{s } , \varnothing ) , ( \includegraphics[width=.3in]{2 + 2 + 2},\includegraphics[width=.2in]{4 + 4},\includegraphics[width=.2in]{2 + 2}))$ ] be given by figure 4 ( left )
. then @xmath711 .
however , adding one box in the corner figure 4 ( right ) kills the moduli and results in @xmath734 .
note that in this example @xmath735 .
( left ) and @xmath734 ( right).,width=384 ] we learn an important lesson from example [ littlebox ] : _ if we are given a double box configuration @xmath736 , then we can not decide from the @xmath737 alone whether there are edge moduli_. informally : the `` corners '' influence the `` legs '' .
this means we `` can not split '' vertex and leg contributions .
a similar phenomenon occurs when discussing the vertex / edge formalism of rank 2 dt theory in section [ vertex / edge ] ( see example [ withlegs ] ) .
we are led to consider `` global '' double box configurations in the following sense .
[ z ] consider a smooth projective toric 3-fold @xmath2 and a @xmath4-equivariant rank 2 reflexive sheaf @xmath49 on @xmath2 defined by toric data @xmath738 such that in each chart @xmath613 is singular ( the general case , with some @xmath618 locally free , is discussed in remark [ deco ] ) .
for each edge @xmath117 , fix a double square configuration @xmath739 for each @xmath115 we denote its neighbouring vertices by @xmath740 .
the labeling is chosen such that @xmath741 is defined by @xmath742 for all @xmath743 and @xmath744 .
fix a double box configuration @xmath745 then the collection @xmath746 describes the characteristic function @xmath555 of a @xmath4-equivariant rank 2 torsion free sheaf on @xmath2 with reflexive hull @xmath49 by proposition [ lem1 ] .
we denote the collection of all such double box configurations by @xmath747 . to each @xmath748
we associated a moduli factor @xmath749 ( definition [ modfactor ] ) .
then @xmath750 here @xmath751 when for some @xmath117 , there is a local edge component containing @xmath752 in chart @xmath135 and a local edge component containing @xmath753 in chart @xmath136 and both components match under gluing kills the moduli . ] .
we define the _ weight _ of @xmath754 as @xmath755 where @xmath21 denotes topological euler characteristic .
recall that for each @xmath117 , we have ( section [ toric3folds ] ) @xmath134 for each @xmath117 , there are two faces @xmath756 , @xmath757 which share @xmath758 as an edge and two disjoint faces @xmath759 , @xmath760 connected by the edge @xmath758 .
suppose face @xmath759 contains vertex @xmath119 and @xmath760 contains @xmath252 .
we define the _ generating function _ associated to @xmath2 , @xmath49 , and @xmath761 @xmath762 for any @xmath763 .
there is one issue with this definition : @xmath764 , @xmath765 depend on choice of representative of @xmath766 . in the formula for @xmath767
, it is assumed that for all @xmath115 , we choose representatives @xmath768 of @xmath769 with the following property : + for any @xmath117 , suppose without loss of generality that @xmath138 is given by @xmath770 in chart @xmath135 and @xmath771 in chart @xmath136 .
suppose the toric data of @xmath618 in chart @xmath135 is given by @xmath772 , @xmath773 and in chart @xmath136 by @xmath774 , @xmath775 .
then for all @xmath683 @xmath776 \times [ u_2+v_2,\infty ] \times [ u_3+v_3,\infty ] ) \\ & = p_1(\pi_{\beta,1 } \setminus [ u_1',u_1'+n ] \times [ u_2'+v_2',\infty ] \times [ u_3'+v_3',\infty ] ) \\ & p_1(\pi_{\alpha,2 } \setminus [ u_1+v_1,u_1+v_1+n ] \times [ u_2,\infty ] \times [ u_3+v_3,\infty ] ) \\ & = p_1(\pi_{\beta,2 } \setminus [ u_1'+v_1',u_1'+v_1'+n ] \times [ u_2',\infty ] \times [ u_3'+v_3',\infty ] ) \\ & p_1(\pi_{\alpha,3 } \setminus [ u_1+v_1,u_1+v_1+n ] \times [ u_2+v_2,\infty ] \times [ u_3,\infty ] ) \\ & = p_1(\pi_{\beta,3 } \setminus [ u_1'+v_1',u_1'+v_1'+n ] \times [ u_2'+v_2',\infty ] \times [ u_3',\infty]),\end{aligned}\ ] ] where @xmath777 denotes projection as before and @xmath689 . +
even though @xmath764 , @xmath765 depend on choice of representatives , the formula for @xmath767 does not by the following proposition : [ formulachi ] let @xmath2 be a smooth projective toric 3-fold .
let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath2 described by toric data @xmath490 and let @xmath52 be a @xmath4-equivariant quotient of dimension @xmath42 .
let @xmath778 correspond to the characteristic function of the kernel of @xmath52 , where @xmath745 suppose for all @xmath115 , we choose representatives of @xmath769 as described in definition [ z ]
. then @xmath779 moreover @xmath780 this follows from a ech calculation using the @xmath4-invariant open affine cover @xmath135 , @xmath115 like in ( * ? ? ?
* lem . 5 ) .
the upshot is to choose representatives of each @xmath769 as in the previous definition and then calculate the @xmath4-representation @xmath781 cancelling infinite terms and finally setting @xmath782 gives the formula for @xmath783 .
the formula for @xmath784 follows from hirzebruch - riemann - roch .
[ openz ] in definition [ z ] , we fixed a smooth _ projective _ toric 3-fold @xmath2 .
suppose @xmath11 is any smooth toric 3-fold with vertices @xmath141 , edges @xmath785 , and @xmath4-equivariant rank 2 reflexive sheaf @xmath49 on @xmath11 .
denote by @xmath786 the edges which correspond to _ compact _ lines @xmath787 .
suppose for each @xmath788 , we fix a double square configuration @xmath761 , then one can still consider @xmath789 where @xmath759 and @xmath760 are defined as in definition [ z ] . in applications
, @xmath11 has toric compactification @xmath2 and @xmath49 is the restriction of a @xmath4-equivariant rank 2 reflexive sheaf on @xmath2 .
then @xmath790 is part of a generating function on @xmath2 . sometimes , for computational reasons , it is useful to consider an analog of @xmath790 , where one fixes a double square configuration @xmath761 for all ( not necessarily compact ) edges @xmath791 @xmath792 [ deco ] it is not hard to redo this section in the case @xmath618 is locally free , i.e. @xmath793 or @xmath794 for some @xmath795 ( proposition [ rank2singular ] ) .
we think of these as degenerate cases .
then @xmath618 is @xmath4-equivariantly decomposable . in this
setting , a double box configuration is defined as a pair of partitions @xmath796 without any conditions , where we impose an equivalence relation on such pairs as before ( and similar for double square configurations ) . after adapting the formulae for @xmath704
, @xmath660 accordingly , the formula of proposition [ formulachi ] holds for any @xmath4-equivariant rank 2 reflexive sheaf on @xmath2 .
the details are straight - forward and left to the reader .
let @xmath2 be a smooth projective toric 3-fold with polarization @xmath14 .
let @xmath797 be the moduli space of rank 2 @xmath17-stable torsion free sheaves on @xmath2 with indicated chern classes . recall that we do _ not _ consider strictly @xmath17-semistable sheaves .
consider the generating function @xmath798 in the introduction , we noted that the double dual map is a constructible morphism @xmath600 here @xmath799 is empty unless @xmath800 where the first is the bogomolov inequality ( * ? ? ?
7.3.1 ) and the second follows from ampleness of @xmath14 .
moreover , for a given @xmath801 there exists a constant @xmath802 such that @xmath799 is empty unless ( * ? ? ?
3.6 ) @xmath803 the generating function @xmath804 is a laurent series in @xmath23 . using the methods of the previous section ,
it can be expressed in terms of euler characteristics of quot schemes as follows : * find a git representation of all connected components @xmath85 of all reflexive fixed loci @xmath805 .
these are configuration spaces of points on @xmath6 or open subsets thereof .
this was shown in @xcite .
e.g. for @xmath806 , each connected component @xmath85 is isomorphic to either a reduced isolated point @xmath807 or @xmath808 @xcite . *
compute the euler characteristics @xmath809 .
euler characteristics of configuration spaces of points on @xmath6 have been computed in general by klyachko @xcite .
* let @xmath85 be a connected component of @xmath805 .
the toric data @xmath490 of any @xmath4-equivariant reflexive sheaf @xmath601 \in { \mathcal c}$ ] has fixed @xmath810 and ( in general ) varying @xmath811 .
we stress the fact that @xmath551 , @xmath552 are constant over the connected component @xmath85 by writing @xmath812 , @xmath813 .
the components of the vectors @xmath812 , @xmath813 are indexed by rays @xmath124 ( definition [ toricdata ] ) @xmath814 the chern class @xmath815 can be computed in terms of the toric data @xmath816 as explained in section [ equivsh ] .
more precisely , @xmath817 is the total length of the singularity of any @xmath601 \in { \mathcal c}$ ] and is computed in proposition [ rank2singular ] .
the third chern class is constant for all @xmath601 \in { \mathcal c}$ ] , so we denote it by @xmath818 .
( in fact , for any @xmath115 and @xmath601 \in { \mathcal c}$ ] , the length of the singularity of @xmath524 is the same . ) + at the level of closed points , the fibre of @xmath48 over @xmath601 \in { \mathcal c}$ ] is @xmath604 @xmath819 from the combinatorial description of section [ 3dpartitions ] , it is clear that all fibres @xmath604 over @xmath601 \in { \mathcal c}$ ] are isomorphic and we may write @xmath820 which is well - defined up to isomorphism
. the connected components of the fibres @xmath821 are all products of @xmath6 s and were described in section [ 3dpartitions ] . *
putting everything together gives the following structure formula @xmath822 in the rest of this section we work out this generating function in the following cases : * @xmath8 is minimal in the sense of proposition [ minc2a ] .
* @xmath107 and @xmath332 is 0-dimensional . *
@xmath77 and @xmath823 .
we start with two propositions .
[ minc2a ] let @xmath2 be a smooth projective toric 3-fold with polarization @xmath14 and fix @xmath16 and @xmath7 .
let @xmath8 be such that @xmath73 is minimal with the property that there exists a rank @xmath16 @xmath17-stable torsion free sheaf on @xmath2 with chern classes @xmath74 .
then for any such sheaf the cokernel @xmath824 is 0-dimensional .
let @xmath38 be as stated .
by the bogomolov inequality ( * ? ? ?
7.3.1 ) @xmath825 suppose @xmath41 is 1-dimensional .
use the same notation for chern classes @xmath826 as in the beginning of this section
. then @xmath827 by ampleness , so @xmath828 therefore @xmath183 is a rank @xmath16 @xmath17-stable torsion free ( reflexive ) sheaf on @xmath2 with chern classes @xmath829 contradicting minimality .
[ minc2b ] let @xmath2 be a smooth projective toric 3-fold with polarization @xmath14 and fix @xmath7 .
let @xmath8 be such that @xmath73 is minimal with the property that there exists a rank 2 @xmath17-stable torsion free sheaf on @xmath2 with chern classes @xmath74 .
for each @xmath115 , denote by @xmath523 the faces sharing vertex @xmath119 . then @xmath830 \in { \mathcal c } : { \mathcal r}|_{u_\alpha } \
\mathrm{is \ singular } \end{array } } } } \prod_{i=1}^{v_{\rho_{1,\alpha}}({\mathcal c } ) } \prod_{j=1}^{v_{\rho_{2,\alpha}}({\mathcal c } ) } \prod_{k=1}^{v_{\rho_{3,\alpha}}({\mathcal c } ) } \frac{1-(q^{-2})^{i+j+k-1}}{1-(q^{-2})^{i+j+k-2}}.\end{aligned}\ ] ] combining , prop .
[ minc2a ] , prop .
[ formulachi ] gives @xmath831 where @xmath832 was introduced in remark [ openz ] .
the latter generating function is discussed in theorem [ ben ] below . if the polyhedron @xmath75 is `` small '' ( defined below ) , the connected components @xmath85 in the previous proposition turn out to be isolated @xmath4-fixed points .
[ small ] let @xmath833 be a convex polyhedron such that each vertex has exactly 3 edges .
we view @xmath833 up to arbitrary stretching of edges and angles between edges ( but not contracting them to a point ) .
we say that @xmath833 is _ small _ if the following two properties hold : 1 . for any way of colouring one face red and
three faces blue ( and not colouring any of the other faces ) , the red face must share an edge with a blue face .
2 . for any way of colouring three faces yellow ( and not colouring any of the other faces ) ,
at least two yellow faces share an edge .
any convex polyhedron such that each vertex has exactly 3 edges and @xmath834 faces is small .
there are 4 of these : tetrahedron , cube , prism over a triangle , prism over a triangle with one corner cut off .
therefore @xmath76 ensures @xmath75 is small .
polyhedra with @xmath835 faces are generally not small . however , a prism over a 5-gon has 7 faces but is small .
we conclude 16 out of 18 smooth toric fano 3-folds have small @xmath75 ( the two with @xmath836 do not have small @xmath75 @xcite ) .
[ eulermin ] let @xmath2 be a smooth projective toric 3-fold such that @xmath75 is small .
let @xmath14 be a polarization and @xmath7 a first chern class such that @xmath837 .
let @xmath8 be such that @xmath73 is minimal with the property that there exists a rank 2 @xmath17-stable torsion free sheaf on @xmath2 with chern classes @xmath74 .
for each @xmath115 , denote by @xmath523 the faces sharing vertex @xmath119
. then the connected components of @xmath838 are isolated reduced points for all @xmath9
. moreover @xmath839 \in { \mathcal n}_{x}^{h}(2,c_1,c_{2},c_{3})^t } q^{c_{3}({\mathcal r } ) } \\ & \times \prod_{{\scriptsize{\begin{array}{c } \alpha \in v(x ) \\
{ \mathcal r}|_{u_\alpha } \
\mathrm{is \ singular } \end{array } } } } \prod_{i=1}^{v_{\rho_{1,\alpha}}({\mathcal r } ) } \prod_{j=1}^{v_{\rho_{2,\alpha}}({\mathcal r } ) } \prod_{k=1}^{v_{\rho_{3,\alpha}}({\mathcal r } ) } \frac{1-(q^{-2})^{i+j+k-1}}{1-(q^{-2})^{i+j+k-2}}.\end{aligned}\ ] ] let @xmath49 be a @xmath4-equivariant rank 2 @xmath17-stable reflexive sheaf on @xmath2 with chern classes @xmath7 , @xmath8 .
let @xmath490 be the toric data corresponding to @xmath49 and @xmath555 the characteristic function of @xmath49 .
suppose the characteristic function @xmath467 has @xmath840 face components ( definitions [ vefcomps ] , [ vefcomps2 ] ) .
we claim @xmath581 .
the case @xmath841 can not happen , because then @xmath49 is decomposable hence @xmath17-unstable .
if @xmath842 , then the connected component @xmath85 of the @xmath4-fixed locus containing @xmath49 has positive dimension .
we show this can not happen .
then the theorem then follows from proposition [ minc2b ] .
suppose at @xmath843 , then the moduli space @xmath844 of @xmath4-equivariant rank 2 @xmath17-stable reflexive sheaves on @xmath2 with characteristic function @xmath555 has dimension @xmath845 .
the space @xmath846 is realized as the git quotient of a non - empty zariski open subset @xmath847 by @xmath558 ( with respect to the polarization described in section [ fixedlocigeneral ] ) .
we denote elements of @xmath848 by @xmath849 . inside @xmath847
lies the zariski open subset @xmath850 of elements @xmath849 with all @xmath851 mutually distinct .
the git quotient @xmath852 is a zariski open subset of @xmath844 of dimension @xmath845 .
therefore , there is a @xmath9 such that @xmath852 gives a zariski open subset @xmath853 consider the classical git problem of @xmath840 ( ordered ) points @xmath849 on @xmath6 .
we assign weights @xmath854 to these points as follows .
the characteristic function @xmath467 has @xmath840 face components .
for each face component , there is a maximal number of faces @xmath855 such that the corresponding local face components are connected or glued together in @xmath467 .
in particular , @xmath856 .
to this face component we assign weight @xmath857 set @xmath858 , then each @xmath859 is smaller than @xmath860 .
this follows from git stability on the zariski open subset @xmath850 ( section [ equivsh ] ) .
next we use the assumption @xmath75 is small . by lemma [ lemmasmall ] below ,
there exists a git semistable point configuration @xmath849 such that there are @xmath795 with the following properties : * @xmath861 is distinct from all other @xmath862 and all other @xmath862 are mutually distinct , * there exists a chart @xmath135 containing the local face components corresponding to @xmath374 and @xmath863 .
informally : on the polyhedron @xmath75 , there are two adjacent faces whose corresponding points on @xmath6 can become equal .
note that @xmath864 has to be a strict inequality , because @xmath865 .
therefore , the above point configuration @xmath849 corresponds to a @xmath17-stable torsion free sheaf @xmath38 with characteristic function @xmath555 ( and hence chern classes @xmath74 ) .
( see also ( * ? ? ?
the reflexive hull @xmath183 is @xmath17-stable with chern classes @xmath829 for some @xmath866 .
since @xmath867 are equal , the characteristic function of @xmath183 is obtained from the characteristic function @xmath467 of @xmath38 by adding a ( non - empty ) infinite leg of 1 s .
therefore @xmath868 , or equivalently @xmath869 , contradicting minimality of @xmath8 . [ lemmasmall ] let @xmath870 , @xmath854 , @xmath871 , and assume @xmath872 for all @xmath873 .
let @xmath833 be a small polyhedron .
let @xmath874 be a collection , such that * @xmath875 is a union of faces of @xmath833 , for all @xmath873 .
* @xmath875 minus the vertices is connected , for all @xmath873 . *
@xmath875 and @xmath876 have no faces in common , for each @xmath877 , @xmath878 , * @xmath879 .
then there exist @xmath880 , which share an edge and @xmath881 it suffices to prove the case where each @xmath875 is equal to a single face .
assume without loss of generality @xmath882 .
then @xmath883 for all @xmath884 .
we first prove the lemma for @xmath885 .
colour the face with label @xmath886 red and those with label @xmath887 blue .
the result follows from property ( 1 ) of definition [ small ] . for @xmath888
we need a slightly more refined analysis .
+ case 1 : @xmath889 .
then we colour the face with label @xmath886 red and those with label @xmath890 blue .
the result again follows from property ( 1 ) of definition [ small ] .
+ case 2 : @xmath891
. then @xmath892 . in this case
, we colour the faces with label @xmath893 yellow .
the result follows from property ( 2 ) of definition [ small ] .
let @xmath96 and @xmath49 a @xmath4-equivariant rank 2 reflexive sheaf on @xmath11 with toric data @xmath490 .
the following generating function ( introduced in remark [ openz ] ) leads to rich combinatorics @xmath894 here @xmath895 are three double square configurations describing fixed asymptotics along the @xmath896- , @xmath897- , and @xmath898-axes .
these generating functions form building blocks of the generating functions @xmath804 .
recall that @xmath49 is singular if and only if all @xmath899 and all @xmath900 are mutually distinct ( proposition [ rank2singular ] ) . in the case
@xmath49 is locally free , @xmath901 decomposes . each double square configurations @xmath737
is given by a pair of 2d partitions @xmath902 .
then @xmath903 where @xmath904 denotes the rank 1 dt vertex .
this product structure follows from the fact that @xmath49 splits @xmath4-equivariantly as a sum of two @xmath4-equivariant line bundles ( proposition [ rank2singular ] ) , which gives an extra @xmath60-scaling action on the factors .
using this extra @xmath60 , follows .
the interesting case is when @xmath49 is singular .
the generating functions @xmath901 can then be computed by relating double box configurations to double dimers . in the case of no legs ( or small legs ) this leads to closed formulae .
the following two theorems are proved in a forthcoming paper @xcite .
[ ben ] let @xmath49 be a singular @xmath4-equivariant rank 2 reflexive sheaf on @xmath58 described by toric data @xmath905 .
then @xmath906 where @xmath907 denotes the macmahon function .
[ benwithlegs ] let @xmath49 be a singular @xmath4-equivariant rank 2 reflexive sheaf on @xmath58 described by toric data @xmath905 .
suppose @xmath908 and @xmath909{s},\varnothing,\varnothing),(\varnothing,\varnothing,\varnothing)).\ ] ] then @xmath910 recall that @xmath62 is the generating function of finite 3d partitions located at the origin and weighted by @xmath911 . for @xmath620 @xmath912
is the generating function of 3d partitions located at the origin , contained in the box @xmath913 \times [ 0,v_2 - 1 ] \times [ 0,v_3 - 1],\ ] ] and weighted by @xmath911 ( * ? ? ?
* ( 7.109 ) ) .
this suggests double box configurations weighted by @xmath914 mysteriously correspond to triples of three partitions , two of which are free and one of which is confined to the box associated to the singularity .
we can not find such a correspondence directly .
let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath58 described by toric data @xmath905 .
consider the quot scheme @xmath57 of quotients @xmath915 where @xmath41 is @xmath200-dimensional and has length @xmath12 @xcite .
the following is a direct corollary of theorem [ ben ] and the discussions of the previous sections .
[ quotc3 ] let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath58 described by toric data @xmath905 . then @xmath916 in this section , we calculate @xmath804 for @xmath77 and @xmath823 .
we also classify the reflexive hulls for @xmath333 .
the generating function does not depend on @xmath14 and we drop @xmath2 and @xmath14 from the notation .
we take @xmath331 ( so rank and degree are coprime and there are no strictly @xmath17-semistable sheaves ) and also suppress it from the notation .
for toric data @xmath490 , we use the same notation as in example [ p3reflch ] .
[ c2=1 ] let @xmath78 . by the methods of section [ fixedloci ] , it is easy to see that @xmath78 is the minimal value for which there exist rank 2 @xmath17-stable torsion sheaves on @xmath522 with chern classes @xmath7 , @xmath8 ( see also @xcite )
there are exactly four such sheaves which are @xmath4-equivariant and reflexive .
each of them has a length 1 singularity at one of the torus fixed points and no further singularities .
the toric data @xmath490 of these reflexive hulls is described as follows : * for any choice @xmath917 with @xmath918 mutually distinct @xmath919 recall that @xmath920 is forced by equation ( [ slice ] ) .
note that for fixed @xmath921 and @xmath922 , different choices of @xmath923 give @xmath4-equivariantly isomorphic reflexive hulls .
theorem [ eulermin ] gives @xmath924 we also note that any connected component @xmath85 of any @xmath925 has constant reflexive hulls .
this is useful for applications to dt theory .
[ c2=2 ] let @xmath79 .
we describe the connected components @xmath85 of @xmath926 containing reflexive hulls of elements of @xmath927 like the previous example , these connected components @xmath85 are isolated reduced points , but legs can appear ( i.e. @xmath8 , @xmath801 may differ ) .
there are three types of reflexive hulls appearing .
their toric data @xmath490 is described as follows : * for any choice @xmath917 with @xmath918 mutually distinct @xmath928 these reflexive hulls have @xmath929 .
* for any choice @xmath917 with @xmath918 mutually distinct @xmath930 these reflexive hulls have @xmath929 . *
the reflexive hulls of example [ c2=1 ] . recall that these reflexive hulls have @xmath931 .
any @xmath4-equivariant rank 2 @xmath17-stable torsion free sheaves @xmath38 on @xmath522 with chern classes @xmath331 , @xmath79 has one of the above as its reflexive hull . in cases 1 and 2
the cokernel @xmath332 is 0-dimensional , whereas in case 3 the cokernel is 1-dimensional .
using formula , proposition [ formulachi ] , and theorem [ benwithlegs ] , one can compute the following formula @xmath932 we also note that any connected component @xmath85 of any @xmath933 has constant reflexive hulls .
[ c2is3 ] let @xmath333 .
we describe the connected components of @xmath926 containing reflexive hulls of elements of @xmath934 this time , @xmath85 need no longer be isolated .
the toric data @xmath490 of these connected components is classified as follows : * for any choice @xmath917 with @xmath918 mutually distinct @xmath935 these reflexive hulls have @xmath936 and form a non - isolated connected components isomorphic to @xmath937 . see also @xcite .
* for any choice @xmath917 with @xmath918 mutually distinct @xmath938 these reflexive hulls have @xmath936 and correspond to isolated reduced points . * for any choice @xmath917 with @xmath918 mutually distinct @xmath939 these reflexive hulls have @xmath936 and correspond to isolated reduced points . * the reflexive hulls of example [ c2=2 ] .
recall that these reflexive hulls have @xmath931 or @xmath929 and correspond to isolated reduced points .
any @xmath4-equivariant rank 2 @xmath17-stable torsion free sheaf @xmath38 on @xmath522 with chern classes @xmath331 , @xmath333 has one of the above as its reflexive hull .
this time not all connected components @xmath85 of have constant reflexive hulls . connected components with an element having a reflexive hull appearing in ( 1 ) do not , whereas those having a reflexive hull as in ( 2 ) and ( 3 ) do . for ( 4 )
some do and some do not .
let @xmath2 be a smooth projective toric 3-fold with @xmath29 .
let @xmath14 be a polarization on @xmath2 and let @xmath940 let @xmath587 be a connected component of the fixed locus @xmath179 , where @xmath941 ( theorem [ fixedlocithm ] ) .
recall that the space @xmath587 is smooth and equal to the following git quotient ( theorem [ equivmod ] ) @xmath942 we will write @xmath943 . in this section , we assume @xmath85 satisfies : + _ all closed points of @xmath85 have the same ( i.e. isomorphic ) reflexive hull @xmath49 . _ + recall that there are plenty of examples of such @xmath85 ( typically for `` low '' values of @xmath8 as in theorem [ eulermin ] and examples [ ex1as1 ] , [ ex2as1 ] , [ ex3as1 ] ) . in applications
, @xmath2 is often a toric compactification of a toric calabi - yau 3-fold @xmath11 and @xmath329 as in section [ tequivdt ] .
assuming @xmath85 has constant reflexive hulls has several nice consequences : * @xmath85 has a universal family @xmath156 ( remark [ a=3 ] ) .
* we have a short exact sequence @xmath944 where the cokernel @xmath260 is @xmath85-flat and @xmath945 ( remark [ a=3 ] ) . + * @xmath85 is a connected component of the fixed locus of the quot scheme @xmath604 ( section [ 3dpartitions ] ) . therefore , @xmath85 is isomorphic to a product of @xmath6 s . in this section
we set up a vertex / edge formalism closely following @xcite in the rank 1 case and @xcite in the stable pair case .
since our fixed loci are non - isolated ( but products of @xmath6 s ) our situation is most similar to @xcite . in order to follow their method
, we need the existence of a universal family on @xmath85 , which is ensured by assuming @xmath85 has constant reflexive hulls . unlike the case of @xcite
, we can have `` edge moduli '' , i.e. `` moduli in the legs '' ( example [ moduliinlegs ] ) .
we want to compute the @xmath95-group class of the dt complex ( [ dtcx ] ) restricted to @xmath85 @xmath946 a torsion free sheaf @xmath165 \in { \mathcal c}_{\mathrm{cl}}$ ] has a @xmath4-equivariant structure as explained in section [ fixedloci ] .
the complex @xmath947 restricted @xmath165 $ ] is the virtual tangent space @xmath948}$ ] at @xmath165 $ ] @xmath949 } & = { \operatorname{ext}}^1({\mathcal f},{\mathcal f } ) - { \operatorname{ext}}^2({\mathcal f},{\mathcal f } ) \\ & = - \langle { \mathcal f},{\mathcal f}\rangle_0.\end{aligned}\ ] ] here we used @xmath950 and @xmath951 , because @xmath38 is stable and @xmath29 . also note that for toric varieties @xmath952 so tracelessness is automatic .
we write this as @xmath953 } = \langle { \mathcal r},{\mathcal r}\rangle - \langle { \mathcal f},{\mathcal f}\rangle - \langle { \mathcal r},{\mathcal r}\rangle_0.\ ] ] the term @xmath954 is well - behaved and can be described by a nice vertex / edge formalism .
the term @xmath327 only contributes an overall constant in @xmath955 $ ] .
let @xmath140 be the @xmath4-invariant open affine cover of @xmath2 and denote the double intersections by @xmath956 .
denote the restriction of any sheaf @xmath38 to @xmath135 , @xmath957 by @xmath428 , @xmath958 respectively .
denote the global section functor by @xmath959 .
using the local - to - global spectral sequence for ext groups and computing sheaf cohomology by a ech calculation gives @xmath960 where contributions of triple and higher intersections vanish because @xmath38 and @xmath49 only differ on a @xmath4-invariant closed subscheme of dimension @xmath42 .
our goal is to calculate @xmath961 this section calculates @xmath962 suppose the action of @xmath4 on @xmath118 is in standard form @xmath963 for all @xmath683 . denote the poincar polynomial of @xmath428 by @xmath964.\ ] ] consider the trace map @xmath965 which associates to a @xmath4-equivariant vector bundle its character ( i.e. decomposition into weight spaces ) . following the reasoning in (
4.4 ) , we obtain @xmath966 where the operation @xmath170 was introduced in section [ toric3folds ] .
we need a family version of the above .
this is easily done replacing @xmath38 by @xmath156 , @xmath49 by @xmath967 , @xmath135 by @xmath968 , and @xmath957 by @xmath969 .
this gives @xmath970 the poincar polynomial of @xmath971 can be expressed in terms of the equivariant parameters @xmath972 as follows .
let @xmath490 be the toric data associated to @xmath618 ( definition [ toricdata ] ) .
+ * case 1 * : @xmath618 is not locally free , or equivalently all @xmath899 and all @xmath900 are mutually distinct ( proposition [ rank2singular ] ) . then @xmath973 * case 2 * : @xmath618 is locally free . in this case
@xmath618 has two homogeneous generators and @xmath974 is the sum of their characters .
+ the character of @xmath975 is more complicated and involves moduli .
recall that @xmath344 for some @xmath345 .
each factor @xmath6 of @xmath85 has a tautological subbundle @xmath976 attached to it .
the character @xmath977 of @xmath975 can be expressed in terms of the equivariant parameters @xmath972 and the classes of these @xmath586 s .
let @xmath701 correspond to @xmath540 by proposition [ lem1 ] .
then @xmath978 where * @xmath979 when @xmath980 , * @xmath981 when @xmath550 lies in exactly one @xmath693 , * @xmath981 when @xmath550 lies in exactly two @xmath982 s but _ not _ in a region with moduli ( see section [ 3dpartitions ] ) , * @xmath983t_{1}^{k_1 } t_{2}^{k_2 } t_{3}^{k_3}$ ] when @xmath550 lies in exactly two @xmath982 s and in a region @xmath984 with moduli ( see section [ 3dpartitions ] ) , * @xmath985 otherwise . from the short exact sequence ( [ ses ] )
we obtain a completely explicit character @xmath986 which lies in @xmath987 this section calculates @xmath988 assume the coordinates are chosen such that the toric line @xmath989 is given by @xmath990 .
we work over the ring @xmath991\ ] ] and define the formal @xmath992-function @xmath993 let @xmath701 correspond to @xmath540 by proposition [ lem1 ] .
denote the limiting double square configuration along the @xmath896-axis arising from @xmath701 by @xmath763 . denoting the character of @xmath994 by @xmath995
, we obtain @xmath996 where * @xmath997 if @xmath998 , * @xmath999 if @xmath1000 lies in exactly one @xmath1001 , * @xmath999 if @xmath1000 lies in exactly two @xmath1001 s but not in a region with moduli , * @xmath1002t_{2}^{k_2 } t_{3}^{k_3}$ ] if @xmath1000 lies in exactly two @xmath1001 s and in a region @xmath1003 with moduli , * @xmath1004 otherwise .
recall from example [ littlebox ] that from the data of @xmath766 alone , we can not determine whether a square lying in exactly two @xmath1001 s has moduli or not .
this depends on what happens `` in the corners '' .
this already indicates we can not split the vertex and edge contributions as easily as in the rank 1 case .
we come back to this in the examples of the next section .
a reasoning similar to the previous section gives @xmath1005 exactly as in @xcite and @xcite , the terms of the previous two sections can be redistributed so they become laurent _ polynomials _ in @xmath1006 .
denote the vertices neighbouring @xmath115 by @xmath740 .
define @xmath1007 where @xmath1008 . here
the labelling is such that @xmath1009 is given by @xmath742 for all @xmath1010 mutually distinct .
recall the definition of @xmath1011 , @xmath1012 from section [ toric3folds ] .
this redistribution does not lead to over - counting due to the choice of the second term in @xmath1013 ( see ) .
the reason these are laurent polynomials is the same as ( * ? ? ?
* lem . 4 ) .
we summarize this section in one theorem .
[ vertexedgeformalism ] let @xmath2 be a smooth projective toric 3-fold with @xmath29 and polarization @xmath14 .
let @xmath340 .
suppose @xmath322 is a connected component of the fixed locus such that all its closed points have reflexive hull @xmath49 .
then the @xmath4-character of donaldson - thomas theory @xmath1014 is given by @xmath1015 where @xmath1016 $ ] and @xmath1017 \\ & { \mathsf e}_{\alpha\beta } \in k_0({\mathcal c } ) \otimes_{{\mathbb z } } { \mathbb z}[t_{1}^{\pm 1},t_{2}^{\pm 1},t_{3}^{\pm 1 } ] , \end{aligned}\ ] ] for all @xmath1018 .
let @xmath11 be a smooth toric calabi - yau 3-fold and let @xmath2 be a toric compactification with @xmath29 and polarization @xmath14 .
we fix @xmath74 and consider the moduli spaces @xmath1019 in section [ moduli ] , we introduced the @xmath4-invariant open subset @xmath1020 of torsion free sheaves @xmath38 with @xmath332 supported in @xmath11 .
we fix a connected component @xmath329 satisfying assumption [ as1 ] : + _ assume @xmath85 is compact and all its closed points have the same ( isomorphic ) reflexive hull @xmath49 . _ + we introduced invariants @xmath93 in proposition - definition [ propdef ] .
we are interested in the `` calabi - yau specialization '' @xmath92 of these invariants . in this section ,
we conjecture a formula for these invariants in terms of the generating functions @xmath1021 of remark [ openz ] .
suppose @xmath165 \in { \mathcal c}$ ] is a closed point with characteristic function corresponding to double box configurations @xmath754 ( proposition [ lem1 ] ) , where @xmath1022 for each @xmath115 with neighbouring vertices @xmath740 . since the cokernel @xmath1023 lies entirely in @xmath11 , the asymptotic double square configurations @xmath761 are zero unless @xmath788 ( i.e. @xmath1024 ) and @xmath1025 unless @xmath145 .
elements of the connected component @xmath1026 ( definition [ z ] ) all have the same third chern class @xmath1027 determined by proposition [ formulachi ] . [ mainconj ]
let @xmath11 be a smooth toric calabi - yau 3-fold with toric compactification @xmath2 satisfying @xmath29 .
fix a polarization @xmath14 on @xmath2 and chern classes @xmath7 , @xmath8 . let @xmath49 be a @xmath4-equivariant rank 2 @xmath17-stable reflexive sheaf on @xmath2 with toric data @xmath490 .
suppose any connected component @xmath85 of @xmath1028 satisfies assumption [ as1 ] with reflexive hull @xmath49 .
for each @xmath788 , there are two faces @xmath1029 , @xmath757 which share @xmath758 as an edge and two disjoint faces @xmath759 , @xmath760 connected by the edge @xmath758 .
let @xmath1030 be such that for any @xmath1031 , the elements of @xmath1032 have second chern class @xmath8 .
then @xmath1033 the assumptions of this conjecture are satisfied in several interesting examples , typically for `` low '' values of @xmath8 . for instance , when @xmath8 is minimal and @xmath75 is small as discussed in section [ sectionmin ] , then the assumption is satisfied ( theorem [ eulermin ] ) . in this case
there are no legs ( proposition [ minc2a ] ) .
an example with legs is example [ ex2as1 ] .
we provide the following evidence for main conjecture [ mainconj ] : * the examples of sections [ nolegs ] , [ o(-1,-1 ) ] . * for connected components @xmath85 , `` with expected obstructions ''
we compute the invariants @xmath1034 using methods of @xcite ( section [ good ] ) . the answers are compatible with main conjecture [ mainconj ] .
* assuming two analogs of conjectures in the stable pairs case @xcite , we prove main conjecture [ mainconj ] using @xmath101-localization ( section [ t0loc ] ) .
note that the formula of conjecture [ mainconj ] only depends on the compactification through whether @xmath49 is @xmath17-stable or not and through an overall multiplicative factor of @xmath1035 .
in particular , the formula @xmath1036 does not depend on the compactification .
we expect this formula to be of the form @xmath1037 times a rational function in @xmath22 , which is invariant under @xmath1038 ( up to an overall multiplicative factor ) .
we show these two statements in the case @xmath1039 , where the rational function is in fact a laurent polynomial ( remark [ invariance ] ) .
in this section , we consider @xmath96 and @xmath49 an arbitrary @xmath4-equivariant rank 2 reflexive sheaf on @xmath58 described by toric data @xmath490 .
we will see below that @xmath49 is always the restriction of a @xmath4-equivariant rank 2 @xmath17-stable reflexive sheaf on a ( polarized ) toric compactification @xmath2 of @xmath58 with @xmath29 ( lemma [ ( p^1)^3 ] ) .
nevertheless , we want to think of the results of this section as being intrinsic to @xmath58 . since @xmath1040 , we choose all legs empty .
let @xmath1041 .
then @xmath705 is the moduli space of @xmath4-equivariant rank 2 torsion free sheaves on @xmath58 with characteristic function corresponding to @xmath701 ( proposition [ lem1 ] ) .
even though we did not start with a compact 3-fold , the vertex expression still makes sense for @xmath705 and is denoted by @xmath1042 . [ normal ] suppose @xmath1043 and @xmath1044 are mutually distinct .
consider @xmath701 given by @xmath1045 so @xmath1046 .
the reader can easily reconstruct @xmath701 from this formula .
then we have @xmath1047 and hence @xmath1048 where @xmath1049 is the weight defined in definition [ z ] .
[ normal2 ] suppose @xmath1043 and @xmath1044 are mutually distinct .
we consider an example with @xmath1050 and @xmath711 , namely @xmath1051 t_1t_2t_3.\ ] ] the reader can easily reconstruct @xmath701 from this formula .
then we have @xmath1052 - 1 - 2\,{\frac{[{\mathcal o}_{{\mathbb p}^1}(-1)]}{t_{{1}}t_{{2}}t_{{3}}}}+{\frac { 1}{t_{{1}}t_{{2 } } t_3}}+\cdots,\ ] ] where @xmath1053 indicates the @xmath101-moving terms .
since @xmath1054 - 1 $ ] , we obtain @xmath1055}{t_{{1}}t_{{2}}t_{{3}}}}-{\frac { 1}{t_{{1}}t_{{2 } } t_3}}+\cdots.\ ] ] a direct calculation shows that @xmath1053 produces a sign @xmath1056 .
we conclude @xmath1057 [ abnormal ] suppose @xmath1043 and @xmath1044 are mutually distinct .
we consider @xmath701 and @xmath1059 given by @xmath1060 so @xmath1061 .
note that neither @xmath705 nor @xmath1062 has moduli .
we obtain @xmath1063 where @xmath1053 indicates the @xmath101-moving terms .
further calculation shows @xmath1064 however , we _ do _ have @xmath1065 in @xcite , we associate to any double box configuration @xmath701 a double dimer model @xmath1066 .
it turns out that in this example @xmath1067 [ measurevertex ] let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath58 .
for each @xmath1068 define the _ rank 2 equivariant vertex measure _ associated to @xmath701 by @xmath1069 define the _ rank 2 equivariant vertex _ as @xmath1070\!].\ ] ] [ ( p^1)^3 ] let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath1072 .
we can view @xmath11 as a @xmath4-invariant affine open subset of @xmath1073 .
there exists a polarization @xmath14 on @xmath2 and a @xmath4-equivariant rank 2 @xmath17-stable reflexive sheaf @xmath1074 on @xmath2 such that @xmath1075 and @xmath1074 restricted to any other @xmath4-invariant affine open subset is locally free .
let @xmath490 be the toric data of @xmath49 and denote the pull - back of the class of the point on each of the three factors of @xmath1073 by @xmath1076 .
we prove the case @xmath49 is singular , i.e. @xmath1077 and @xmath1044 are mutually distinct ( proposition [ rank2singular ] ) .
the degenerate cases can be treated in a similar fashion .
let @xmath1078 be any polarization on @xmath2 , i.e. @xmath1079 .
define @xmath1074 by the following toric data . for each of the faces
@xmath1080 corresponding to the chart @xmath96 , we take @xmath1081 , @xmath1082 , and @xmath486 equal to the corresponding value of @xmath1083 , @xmath1084 , and @xmath900 coming from @xmath49 . for each other @xmath1080
, we take @xmath1085
. then clearly @xmath1086 and @xmath1074 restricted to the other standard charts is locally free ( proposition [ rank2singular ] ) .
it suffices to ensure @xmath1074 is @xmath17-stable .
this is equivalent to ( section [ equivsh ] ) @xmath1087 for all @xmath1088 mutually distinct .
these inequalities are equivalent to @xmath1089 for all @xmath1088 mutually distinct . taking @xmath1090 , @xmath1091 , @xmath1092 proves the lemma . 1 .
@xmath49 is locally free , i.e. some @xmath1096 or some @xmath900 are equal ( proposition [ rank2singular ] ) .
we saw in equation that @xmath1097 .
therefore conjecture [ edgelessconj ] implies @xmath1098 2 .
@xmath49 is singular , i.e. all @xmath899 and @xmath900 are mutually distinct . in this case , only depends on @xmath552 ( by ) and we define @xmath1099 the right hand side of conjecture [ edgelessconj ] is computed in theorem [ ben ] .
conjecture [ edgelessconj ] is an interesting non - trivial _ combinatorial identity _ stating @xmath1100 [ invariance ] conjecture [ edgelessconj ] implies main conjecture [ mainconj ] in the case @xmath1101 ( no legs ) . in this case , the generating function of main conjecture [ mainconj ] is @xmath1102 up to an overall multiplicative factor of @xmath22 to some power . since there are no legs , the formula of the previous remark shows the generating function is @xmath1037 times a laurent polynomial in @xmath1103 . as can be checked by a small calculation , this laurent polynomial is invariant under @xmath1038 ( up to an overall multiplicative factor ) .
[ buddies1 ] in example [ abnormal ] , we saw that in general @xmath1104 .
however , we do expect the following refinement of conjecture [ edgelessconj ] to be true .
let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath58 . in @xcite
, we associate a double dimer model @xmath1066 to any @xmath1041 . for any double dimer model @xmath1105
, we conjecture @xmath1106 in @xcite , we refer to elements @xmath1107 with the same double dimer model as _
buddies_. in section [ t0loc ] , we give a ( conjectural ) geometric characterization of buddies : @xmath1108 are buddies if and only if @xmath1109 , @xmath1110 are @xmath1111-fixed connected components inside the same @xmath101-fixed connected component ( remark [ buddies2 ] ) .
for @xmath1112 , one can fix a double square configuration @xmath766 along @xmath60 ( definition [ doublesquare ] ) .
unlike the rank 1 case , the edge term along this leg is not determined by @xmath766 alone .
therefore , @xmath1113 is not well - defined .
the reason is that whether the leg has moduli or not depends on what happens in the `` corners attached to the leg '' ( example [ littlebox ] ) . in this section , we consider @xmath1114 let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath11 described by toric data @xmath1115 .
let @xmath1116 label the two vertices of @xmath1117 .
the polyhedron @xmath1117 has four non - compact faces .
two of them share the unique compact edge and we denote them by @xmath1118 .
the other two we denote by @xmath1119 , where @xmath507 has vertex @xmath200 and @xmath508 has vertex @xmath1120 .
we will see @xmath49 is always the restriction of a @xmath4-equivariant rank 2 @xmath17-stable reflexive sheaf on a ( polarized ) toric compactification @xmath2 of @xmath11 ( lemma [ f1 ] ) . in order for this extension of @xmath49 to have no @xmath4-fixed moduli , we assume @xmath1121 fix a double square configuration @xmath766 along the compact leg and @xmath1122 , where @xmath1123 then @xmath754 has at most 3 face components and can have several leg and vertex components of 1 s ( definition [ vefcomps2 ] ) .
therefore , there are no moduli associated to the face components .
in particular , the moduli space @xmath705 ( definition [ z ] ) decomposes according to vertex and edge components @xmath1124 note that the expression for the edge in , makes sense without reference to a compactification of @xmath2 .
we write the corresponding expression as @xmath1125 .
it is tempting to define an equivariant edge measure @xmath1126 however , this may not be well - defined as illustrated by the following example .
[ withlegs ] suppose @xmath49 is given by toric data @xmath1115 with @xmath1127 and @xmath621 mutually distinct .
we consider the following three double box configurations @xmath1128 with leg along the zero section of @xmath11 @xmath1129t_1t_2)}{1-t_3 } , \frac{t_1t_3^{-1}+t_2t_3^{-1}+[{\mathcal o}_{{\mathbb p}^1}(1)]t_1t_2t_3^{-2}}{1-t_3^{-1 } } \right ) , \\ ( { \mathsf q}_{{\boldsymbol \pi}_{0 } '' } , { \mathsf q}_{{\boldsymbol \pi}_{1}''})&= \left(\frac{t_3(t_1+t_2+t_1t_2)}{1-t_3 } , \frac{t_1t_3^{-1}+t_2t_3^{-1}+t_1t_2t_3^{-1}}{1-t_3^{-1 } } \right).\end{aligned}\ ] ] the leg of @xmath754 has no moduli , two columns of weight spaces of dimension 1 , and one column of weight spaces of dimension 2 .
the leg of @xmath1130 is the same as @xmath754 , but with the column of 2-dimensional weight spaces replaced by a column of 1-dimensional weight spaces with moduli @xmath6 .
the leg of @xmath1131 has three columns of weight spaces of dimension 1 without moduli
. one can compute @xmath1132}{t_1}-\frac{1}{t_2 ^
2}+\frac{1}{t_2}-\frac{[{\mathcal o}_{{\mathbb p}^1}(1)]}{t_2}-\frac{2}{t_1 t_2}+\frac{[{\mathcal o}_{{\mathbb p}^1}(1)]}{t_1 t_2}+\frac{2}{t_3 } \right .
\\ & \left .
-\frac{[{\mathcal o}_{{\mathbb p}^1}(-1)]}{t_3}-\frac{1}{t_1 t_3}+\frac{[{\mathcal o}_{{\mathbb p}^1}(-1)]}{t_1 t_3}-\frac{1}{t_2 t_3}+\frac{[{\mathcal o}_{{\mathbb p}^1}(-1)]}{t_2 t_3}+\frac{t_1}{t_2 t_3}+\frac{t_2}{t_1 t_3 } , 0 \right ) , \\
( { \mathsf v}_{{\boldsymbol \pi}_{0 } '' } , { \mathsf v}_{{\boldsymbol \pi}_{1}''})= & \left(-1-\frac{1}{t_1}-\frac{1}{t_2}+\frac{1}{t_1 t_3}+\frac{1}{t_2 t_3}+\frac{1}{t_1 t_2 t_3},0 \right ) , \\ { \mathsf e}_{\hat{{\boldsymbol \pi}}}=&\frac{2 t_1}{t_2}+\frac{2 t_2}{t_1}-\frac{2}{t_1 ^ 2 t_3}-\frac{2}{t_2 ^ 2 t_3 } , \\ { \mathsf e}_{\hat{{\boldsymbol \pi}}'}=&2 [ { \mathcal o}_{{\mathbb p}^1}(1)]-1+\frac{[{\mathcal o}_{{\mathbb p}^1}(1 ) ] t_1}{t_2}+\frac{[{\mathcal o}_{{\mathbb p}^1}(1 ) ] t_2}{t_1}-\frac{[{\mathcal o}_{{\mathbb p}^1}(-1)]}{t_1 ^ 2 t_3}-\frac{[{\mathcal o}_{{\mathbb p}^1}(-1)]}{t_2 ^ 2 t_3}\\ & -\frac{2 [ { \mathcal o}_{{\mathbb p}^1}(-1)]}{t_1 t_2 t_3}+\frac{1}{t_1 t_2 t_3 } , \\
{ \mathsf e}_{\hat{{\boldsymbol \pi}}''}=&1+\frac{t_1}{t_2}+\frac{t_2}{t_1}-\frac{1}{t_1 ^ 2 t_3}-\frac{1}{t_2 ^ 2 t_3}-\frac{1}{t_1 t_2 t_3}.\end{aligned}\ ] ] interestingly , @xmath1133 has a positive @xmath4-fixed part , so @xmath1134 is not well - defined . therefore is in general not well - defined . as a consistency check : an explicit calculation using these expressions leads to the sign predicted by main conjecture [ mainconj ] .
let @xmath1138 and label the vertices and faces of @xmath1117 as before .
let @xmath49 be a @xmath4-equivariant rank 2 reflexive sheaf on @xmath11 described by toric data @xmath1115 with @xmath1139 .
let @xmath766 be a double square configuration along the compact edge and let @xmath1140 , where @xmath1141 then @xmath1142 is the _ rank 2 equivariant vertex measure _ associated to @xmath754 .
the specialization @xmath1143 is well - defined .
define the _ rank 2 equivariant vertex _ as @xmath1144\!],\ ] ] where @xmath1145 , @xmath1146 were defined in definition [ z ] .
we have to prove the calabi - yau specialization @xmath92 is well - defined .
the proof is similar to that of proposition [ welldef1 ] and follows from the following lemma and proposition - definition [ propdef ] .
[ f1 ] let @xmath49 be any @xmath4-equivariant rank 2 reflexive sheaf on @xmath1138 described by toric data @xmath1115 ( not necessarily with @xmath1139 ) .
then there exists a polarized toric compactification @xmath2 of @xmath11 and a rank 2 @xmath17-stable reflexive sheaf @xmath1074 on @xmath2 , such that @xmath1147 and the restriction of @xmath1074 to any @xmath4-invariant affine open subset not contained in @xmath11 is locally free .
let @xmath1148 be the first hirzebruch surface and let @xmath199 be the @xmath4-fixed line with @xmath1149 .
let @xmath1150 be a @xmath4-fixed fibre .
define @xmath1151 the following divisors form a @xmath1152-basis for @xmath1153 @xmath1154 then @xmath199 is a @xmath1155 curve in @xmath2 contained in two toric charts @xmath1156 .
their union is isomorphic to @xmath11 .
we define @xmath1074 on @xmath2 by the following toric data @xmath1157 .
the faces @xmath1158 of @xmath1117 ( introduced at the beginning of this section ) correspond to faces of @xmath75 and we take @xmath1159 from the given toric data @xmath1115 . for all other faces @xmath568 of @xmath75 , we set @xmath1160
. then @xmath1075 and @xmath1074 restricted to any other other toric chart is locally free ( proposition [ rank2singular ] ) .
we are reduced to finding a polarization @xmath14 on @xmath2 such that @xmath1074 is @xmath17-stable . the intersection numbers among the toric divisors are @xmath1161 where @xmath807 denotes the class of a point .
a divisor @xmath1162 is ample if and only if @xmath1163 .
take @xmath1164 arbitrary and @xmath1165 , @xmath1166 , @xmath1092 . using the above intersection numbers , it is not hard to see that there exists a @xmath1164 such that the following four stability inequalities hold @xmath1167 after scaling @xmath14 such that all its coefficients are integer , we are done . [ ( -1,-1)conj ]
let @xmath1138 and label the vertices and faces of @xmath1117 as before .
let @xmath49 be any @xmath4-equivariant rank 2 reflexive sheaf on @xmath11 described by toric data @xmath1115 with @xmath1139 .
let @xmath766 be a double square configuration along the compact edge
. then @xmath1168 consider the statement of main conjecture [ mainconj ] .
if @xmath1169 is such that at each vertex at most one leg of type @xmath1155 comes in , then the formula of conjecture [ mainconj ] is implied by conjecture [ ( -1,-1)conj ] .
also note that conjecture [ ( -1,-1)conj ] implies conjecture [ edgelessconj ] by taking @xmath1170 and embedding @xmath58 in @xmath11 as one of the two toric charts . in this section ,
we explain where the sign in main conjecture [ mainconj ] comes from .
the following theorem discusses the contribution of connected components @xmath85 with `` expected obstructions '' .
the argument is a variation on ( * ? ? ?
* sect . 4.10 , 4.11 ) .
[ tthm ] let @xmath11 be a smooth toric calabi - yau 3-fold with toric compactification @xmath2 with @xmath29 and polarization @xmath14 .
let @xmath49 be a @xmath4-equivariant rank 2 @xmath17-stable reflexive sheaf on @xmath2 with toric data @xmath490 . for each @xmath788 , there are two faces @xmath1029 , @xmath757 which share @xmath758 as an edge and two disjoint faces @xmath759 , @xmath760 connected by the edge @xmath758 .
let @xmath1031 for some @xmath1030 and consider the connected component @xmath1026 .
assume the following : in the case @xmath85 is in addition isolated , the second bullet states that @xmath1171 has no @xmath101-fixed terms . in @xcite ,
the @xmath4-fixed locus automatically consists of isolated reduced points and the absence of @xmath101-fixed terms always holds ( this is the crucial technical point mentioned in ( * ? ? ? * sect . 4.11 ) ) . in our case
, @xmath85 could be non - isolated . moreover , even when @xmath85 is isolated there can be non - zero @xmath101-fixed terms as for @xmath701 and @xmath1059 in example [ abnormal ] .
examples [ normal ] , [ normal2 ] , and [ withlegs ] satisfy the assumptions of the theorem . from the definition and assumptions @xmath1174 where @xmath1053 stands for the contribution of all @xmath101-moving terms .
this is equal to @xmath1175 because of degree reasons , the contribution of @xmath1053 is simply @xmath1176 .
more precisely , it is @xmath1056 to the power the number of ( serre dual ) @xmath101-moving _ pairs _ in @xmath1171 .
we conclude @xmath1177 in order to determine the sign , we use the following claim . for all @xmath788 and @xmath145 , there exist splittings @xmath1178 such that .
our dualities also hold before specialization . ]
@xmath1179 if the claim is proved , then we conclude @xmath93 equals @xmath1180 where @xmath1181 denotes the rank of the @xmath101-fixed term .
moreover , @xmath1182 and @xmath1183 should be interpreted as first restricting the vertex and edge expression to ( any ) closed point of @xmath85 , thus getting rid of moduli , and then putting @xmath1184 .
so we need to prove the above splittings and compute the terms in the power modulo 2 . for ,
we define ( similar to ( * ? ? ?
* sect . 4.10 ) ) @xmath1185 in this definition , the coordinates are chosen such that @xmath1186 is given by @xmath139 .
we then define @xmath1187 as in expression , but with @xmath1188 replaced by @xmath1189 . also define @xmath1190 .
the identity can then be checked by direct calculation .
this uses @xmath1191 ( valid because @xmath788 and @xmath11 is calabi - yau ) . from the definition of @xmath1187
, we obtain @xmath1192 where the second equality follows from @xmath1193 ( @xmath49 has rank 2 ) . using the fact that @xmath1193 and @xmath1194 , one obtains @xmath1195 where @xmath1196 are the two faces with common edge @xmath758 .
the splitting is more complicated .
let @xmath1189 be defined as above .
naively , one would like to define @xmath1197 and @xmath1198 . here
the vertices ( in @xmath1117 ) neighbouring @xmath145 are denoted by @xmath740 , where the labelling is such that @xmath1009 is given by @xmath742 for all @xmath1010 mutually distinct . in the case
@xmath1199 ( no legs ) , this works and one obtains .
the general case is more complicated .
one introduces @xmath1200 by the equation @xmath1201 and defines @xmath1202 exactly analogous to ( * ? ? ?
* sect . 4.10 ) .
setting @xmath1198 , equation can be deduced .
the correct definition of @xmath1202 leads to @xmath1203 this equality follows after noting that various terms in @xmath1202 are zero or even for @xmath1204 .
a similar ( but easier ) calculation as for @xmath1205 shows @xmath1206 where @xmath1207 are the three faces with vertex @xmath119 labelled such that @xmath1208 are the faces sharing edge @xmath1209 . finally , we claim @xmath1210 define @xmath1211 take any @xmath101-fixed term @xmath285 in @xmath1212 _ not _ occurring in @xmath1172 .
this term must cancel after adding @xmath1213 ( by assumption ) .
therefore , @xmath1213 contains the term @xmath1214 . by duality and
, @xmath1212 must contain @xmath1215 . therefore , nonzero terms in @xmath1212 not occurring in @xmath1172 come in pairs ( and likewise for @xmath1213 ) . distributing the terms of @xmath1172 over @xmath1216 by using , and calculating modulo 2
shows the claim . in this section ,
we prove main conjecture [ mainconj ] from a @xmath101-localization argument similar to ( * ? ? ?
pandharipande and thomas need two conjectures for their argument .
we need completely similar analogs of their conjectures in our setting .
the first is smoothness of the @xmath101-fixed locus ( conjecture [ t0conj1 ] ) .
the second will be described below .
[ t0thm1 ] assume conjecture [ t0conj1 ] is true .
let @xmath11 be a smooth quasi - projective toric calabi - yau 3-fold with toric compactification @xmath2 satisfying @xmath29 . fix a polarization @xmath14 on @xmath2 and let @xmath378 and @xmath1217 .
suppose @xmath380 is a connected component satisfying assumption [ as2 ] with reflexive hull @xmath49 .
let @xmath1218 be the virtual normal bundle on @xmath85 w.r.t .
the @xmath101-action .
then the class @xmath1219 is of the form @xmath1220 with @xmath1221 .
moreover @xmath1222 by symmetry of the @xmath101-equivariant obstruction theory ( theorem [ sym ] ) @xmath1223 therefore , the invariant is given by @xmath1224 note that @xmath1225 only has @xmath101-moving terms by assumption [ as2 ] .
theorem [ sym ] therefore gives @xmath1226 consequently , @xmath1227 is of the form @xmath1220 with @xmath1228 and we conclude @xmath1229 [ buddies2 ] combining theorem [ t0thm1 ] and proposition [ t0dt = tdt ] gives @xmath1230 it is tempting to conclude @xmath1231 for any @xmath60-fixed component @xmath105 of @xmath85 .
this is certainly not true in general , because it would imply @xmath1232 is never zero contradicting example [ abnormal ] .
this is very reminiscent of the discussion of buddies in remark [ buddies1 ] .
this leads us to expect two elements @xmath1233 are buddies if and only if @xmath1109 , @xmath1110 lie in the same connected component of the @xmath101-fixed locus .
this would provide a geometric explanation for the ( combinatorially mysterious ) notion of buddies .
this also shows the classical count @xmath1234 and virtual count @xmath1232 are different .
remarkably , they become equal again after summing over all connected components @xmath105 of the same connected component of the @xmath101-fixed locus .
[ t0conj2 ] let @xmath11 be a smooth quasi - projective toric calabi - yau 3-fold with toric compactification @xmath2 satisfying @xmath29 . fix a polarization @xmath14 , let @xmath378 , and let @xmath1217 .
suppose @xmath380 is a connected component satisfying assumption [ as2 ] and let @xmath105 be a connected component of @xmath390 .
since @xmath105 is a connected component of @xmath179 , we have defined vertices @xmath1235 , @xmath1236 and edges @xmath1013 , @xmath1237 on it . for any @xmath165 \in { \mathcal c}_i$ ] ,
these give restrictions @xmath1238}$ ] , @xmath1239}$ ] , @xmath1240}$ ] , @xmath1241}$ ] .
then @xmath1242 } ) + \sum_{\alpha\beta \in e_c(y ) } \mathbf{constant}({\mathsf e}_{\alpha\beta}^{+}|_{[{\mathcal f } ] } ) \mod 2,\end{aligned}\ ] ] where @xmath1181 denotes the rank of the @xmath101 fixed term .
[ t0thm ] let @xmath11 be a smooth quasi - projective toric calabi - yau 3-fold with toric compactification @xmath2 satisfying @xmath29 . fix a polarization @xmath14 on @xmath2 and chern classes @xmath7 , @xmath8 .
let @xmath49 be a @xmath4-equivariant rank 2 @xmath17-stable reflexive sheaf on @xmath2 with toric data @xmath490 .
we assume : * @xmath381 . * conjectures [ t0conj1 ] ( smoothness @xmath101-fixed loci ) , [ t0conj2 ] ( parity @xmath101-fixed tangent spaces ) hold . * all the connected components @xmath1243 are compact and have constant reflexive hulls @xmath49 ( i.e. assumption [ as2 ] ) . for each @xmath788 , there are two faces @xmath756 , @xmath757 which share @xmath758 as an edge and two disjoint faces @xmath759 , @xmath760 connected by the edge @xmath758 .
let @xmath1030 be chosen such that for any @xmath1031 , the elements of @xmath1032 have second chern class @xmath8 .
then @xmath1033 let @xmath1244 be a connected component of the @xmath101-fixed locus for some @xmath1245 .
let @xmath105 be a connected component of the @xmath60-fixed locus of @xmath390 .
then @xmath105 is equal to a unique @xmath1032 for some @xmath1246 and @xmath1031 .
moreover , each @xmath705 in the sum arises in this way . by theorem [ t0thm1 ] @xmath1247 suppose @xmath1248 as above .
the proof follows if we show @xmath1249 it suffices to evaluate @xmath1250 on a single closed point @xmath165 \in { \mathcal c}_{\hat{{\boldsymbol \pi}}}$ ] . by definition of @xmath1251 , @xmath1252 , @xmath1253 @xmath1254 } )
= & \sum_{\alpha \in v(y ) } ( { \mathsf v}_\alpha^{+}|_{[{\mathcal f}]}(1,1,1 ) - \mathbf{constant}({\mathsf v}_\alpha^{+}|_{[{\mathcal f } ] } ) ) \\ & + \sum_{\alpha\beta \in e_c(y ) } ( { \mathsf e}_{\alpha\beta}^{+}|_{[{\mathcal f}]}(1,1,1 ) - \mathbf{constant}({\mathsf e}_{\alpha\beta}^{+}|_{[{\mathcal f } ] } ) ) .
\end{aligned}\ ] ] in the proof of theorem [ tthm ] , we computed @xmath1255}(1,1,1 ) + \sum_{\alpha\beta \in e_c(y ) } { \mathsf e}_{\alpha\beta}^{+}|_{[{\mathcal f}]}(1,1,1 ) \\ &
= \sum_{\alpha\beta \in e_c(y ) } |{\boldsymbol \lambda}_{\alpha\beta}| ( m_{\alpha\beta } ( v_{\rho_{1,\alpha\beta } } + v_{\rho_{2,\alpha\beta } } + 1 ) + v_{\rho_{3,\alpha\beta } } + v_{\rho_{4,\alpha\beta } } ) \mod 2.\end{aligned}\ ] ] the theorem follows from @xmath1256 } ) + \sum_{\alpha\beta \in e_c(y ) } \mathbf{constant}({\mathsf e}_{\alpha\beta}^{+}|_{[{\mathcal f } ] } ) \mod 2,\end{aligned}\ ] ] which is precisely conjecture [ t0conj2 ] .
[ finalrem ] let @xmath11 be a smooth toric calabi - yau 3-fold with toric compactification @xmath2 satisfying @xmath29 .
let @xmath49 be a rank 2 @xmath17-stable reflexive sheaf on @xmath2 with respect to some polarization .
assume : then @xmath1258 by the calculation in the proof of the previous theorem , @xmath1259 ( no legs ) .
we deduce that in the absence of legs @xmath1260 this formula holds even though in general @xmath1261 ( example [ abnormal ] ) .
mnop2 m. f. atiyah and r. bott , _ the yang - mills equations over riemann surfaces _
london a308 ( 1982 ) , 523615 .
v. v. batyrev , _ toroidal fano 3-folds _ , math .
ussr - izv . 19 ( 1982 ) , 1325
. k. behrend , _ donaldson - thomas type invariants via microlocal geometry _ , annals of math
. 170 ( 2009 ) , 13071338 .
k. behrend and b. fantechi , _ symmetric obstruction theories and hilbert schemes of points on threefolds _ , alg . and
theory 2 ( 2008 ) , 313345 .
j. cheah , _ on the cohomology of hilbert schemes of points _ , j. alg .
geom . 5 ( 1996 ) , 479511
. j. choi , _ genus zero bps invariants for local @xmath1262 _ , imrn ( 2014 ) , 418450 .
i. dolgachev , _ lectures on invariant theory _ , cambridge university press ( 2003 ) .
u. v. desale and s. ramanan , _ poincar polynomials of the variety of stable bundles _ , math .
ann . 216 ( 1975 ) , 233244 .
w. fulton , _ introduction to toric varieties _ , princeton university press ( 1993 ) .
a. gholampour , y. jiang , and m. kool , _
sheaves on weighted projective planes and modular forms _ , arxiv:1209.3922 .
a. gholampour and m. kool , _ stable reflexive sheaves and localization _ , arxiv:1308.3688 .
a. gholampour , m. kool , and b. young , work in progress .
l. gttsche , _ the betti numbers of the hilbert scheme of points on a smooth projective surface _ , math .
ann . 286 ( 1990 ) , 193207
. l. gttsche , _ theta functions and hodge numbers of moduli spaces of sheaves on rational surfaces _ , comm .
206 ( 1999 ) , 105136 . t. graber and r. pandharipande , _ localization of virtual classes _ , invent . math . 135 ( 1999 ) , 487518 .
a. grothendieck , _ techniques de construction et thormes dexistence en gometrie algbrique , iv : les schmas de hilbert _ , sem .
bourbaki 1960/1961 221 .
d. greb and m. toma , _ compact moduli spaces for slope - semistable sheaves _
, arxiv:1303.2480 .
r. hartshorne , _ algebraic geometry _
, springer - verlag ( 1977 ) .
r. harthorne , _ stable reflexive sheaves _ , math .
ann . 254 ( 1980 ) , 121176 .
d. huybrechts and m. lehn , _ the geometry of moduli spaces of sheaves _ , cambridge university press ( 2010 ) . g. harder and m. s. narasimhan , _ on the cohomology groups of moduli spaces of vector bundles on curves _ , math .
. 212 ( 1975 ) , 215248 .
a. a. klyachko , _ equivariant bundles on toral varieties _ , math .
ussr izvestiya 35 ( 1990 ) , 337375 .
a. a. klyachko , _ vector bundles and torsion free sheaves on the projective plane _ , preprint max planck institut fr mathematik ( 1991 ) .
a. a. klyachko , _ spatial polygons and stable configurations of points in the projective line _ , in : algebraic geometry and its applications , j. kollr , _ hulls and husks _ ,
m. kool , _ fixed point loci of moduli spaces of sheaves on toric varieties _ , adv .
math . 227 ( 2011 ) , 17001755 .
m. kool , _ euler characteristics of moduli spaces of torsion free sheaves on toric surfaces _ , geom . ded . 176 ( 2015 ) 241 - 269 .
j. manschot , _ the betti numbers of the moduli space of stable sheaves of rank 3 on @xmath1263 _ , lett . math .
phys . 98 ( 2011 ) , 6578 .
j. manschot , _ bps invariants of @xmath1264 gauge theory on a surface _ , comm .
phys . 6 ( 2012 ) , 497516 .
j. manschot , _ bps invariants of semi - stable sheaves on rational surfaces _ , lett .
phys . 103 ( 2013 ) , 895918 .
d. maulik , n. nekrasov , a. okounkov and r. pandharipande , _ gromov - witten theory and donaldson - thomas theory , i _ ,
142 ( 2006 ) , 12631285 .
d. maulik , n. nekrasov , a. okounkov and r. pandharipande , _ gromov - witten theory and donaldson - thomas theory , ii _ , compos .
142 ( 2006 ) , 12861304 .
s. mozgovoy , _ invariants of moduli spaces of stable sheaves on ruled surfaces _ , arxiv:1302.4134 .
d. oprea , _ framed sheaves and symmetric obstruction theories _ , doc .
math . 18 ( 2013 ) , 323342 .
m. perling , _ graded rings and equivariant sheaves on toric varieties _ , math .
nachr . 263264 ( 2004 ) , 181197 .
m. perling , _
moduli for equivariant vector bundles of rank two on smooth toric surfaces _ , math .
nachr . 265 ( 2004 ) , 8799 .
r. pandharipande and r. p. thomas , _
the 3-fold vertex via stable pairs _ , geom . and topol . 13 ( 2009 ) , 18351876 .
r. p. stanley , _ enumerative combinatorics , volume 2 _ , cambridge studies in adv . math . 62 ( 2001 ) r. p. thomas , _ a holomorphic casson invariant for calabi yau 3-folds , and bundles on k3 fibrations _ , j. diff . geom .
54 ( 2000 ) 367438 . c. vafa and e. witten ,
_ a strong coupling test of s - duality _ , nucl .
phys . b 431 3 ( 1994 ) , 377 .
k. watanabe and m. watanabe , _ the classification of fano 3-folds with torus embeddings _ , tokyo j. math . 5 ( 1982 ) , 3748
. k. yoshioka , _ the betti numbers of the moduli space of stable sheaves of rank 2 on @xmath1265 _ , j. reine angew .
math . 453 ( 1994 ) , 193220 . | let @xmath0 be the moduli space of rank 2 stable torsion free sheaves with chern classes @xmath1 on a smooth 3-fold @xmath2 .
when @xmath2 is toric with torus @xmath3 , we describe the @xmath4-fixed locus of the moduli space @xmath0 .
connected components of @xmath5 with constant reflexive hulls are isomorphic to products of @xmath6 .
we mainly consider such connected components , which typically arise for any @xmath7 , `` low values '' of @xmath8 , and arbitrary @xmath9 . using a new type of combinatorics called double box configurations , we provide a framework for computing the generating function @xmath10 of topological euler characteristics of @xmath0 ( summing over all @xmath9 ) .
the combinatorics is solved using the double dimer model in a companion paper .
this leads to explicit formulae for @xmath10 involving the macmahon function .
when @xmath11 is a toric calabi - yau 3-fold , we consider donaldson - thomas type invariants of @xmath11 defined by virtual localization .
the contribution to the invariant of an individual connected component of the @xmath4-fixed locus is in general _ not _ equal to its signed euler characteristic due to @xmath4-fixed obstructions .
nevertheless , the generating function of _ all _ invariants is still given by @xmath10 up to a sign . |
the evidence that our universe is experiencing a phase of expansion with accelerated rate has been well demonstrated by cosmological data obtained from different independent observations of sneia , cosmic microwave background radiation ( cmbr ) anistropies , x - ray experiments and large scale structures ( lss ) @xcite .
three main ideas have been suggested to give a reasonable explanation to the present day observed accelerated expansion of our universe : the cosmological constant @xmath3 model , dark energy ( de ) models and theories of modified gravity models .
thorough discussions on these three ideas are available in the reviews of @xcite .
the cosmological constant @xmath3 , which has eos parameter @xmath4 , represents the earliest and the simplest theoretical candidate suggested in order to give a plausible explaination to the observational evidences of the universe s present day accelerated expansion .
it is well - known , anyway , that there are two main problems associated with @xmath3 : the fine - tuning and the cosmic coincidence problems .
the former mainly asks why the vacuum energy density is so small ( about an order of @xmath5 lower than what we can observe ) while the latter asks why the vacuum energy and dm give a nearly equal contribution at the present epoch even if they had an evolution which is independent and they had also evolved from mass scales which are different ( this fact represents a really strange coincidence if some internal connections between them are not taken into account ) .
till now , many attempts have been done in order to find a possible plausible explanation to the coincidence problem ( see @xcite ) .
the second idea suggested in order to possibly explain the observed accelerated expansion of the universe involve de models ( reviewed in @xcite ) .
in relativistic cosmology , the cosmic acceleration we are able to observe can be described by the mean of a perfect fluid which pressure and energy density , indicated with @xmath6 and @xmath7 , satisfy the relation given by @xmath8 .
this kind of fluid is dubbed as dark energy ( de ) .
the relation @xmath8 also tells us that the eos parameter of the fluid @xmath9 must be in agreement with the condition @xmath10 , while , from an observational point of view , it is a difficult work to constrain its exact value . the most direct evidence we have for the detection of de
is obtained from observations of supernovae of a type ia ( sneia ) whose intrinsic luminosities can be safely considered practically uniform @xcite . if we assume that the de idea is the right one in order to explain the present expansion of the universe with accelerated rate , we must have that the largest amount of the total cosmic energy density @xmath11 must be concentrated in the two dark sectors , _
i.e. _ dark energy ( de ) and dark matter ( dm ) which represent , according to recent cosmological observations , about the 70@xmath12 and about the 25@xmath12 of the total energy density @xmath11 of the present day universe @xcite .
moreover , the ordinary baryonic matter we are able to observe with our scientific instruments contributes for only the 5@xmath12 of @xmath11 .
furthermore , the radiation density gives a contribution to the total cosmic energy density which we can safely consider negligible .
many different models have been well studied in recent times to understand the exact nature of de .
some of these models include tachyon , quintessence , k - essence , quintom , chaplygin gas , agegraphic de ( ade ) , nade and phantom .
the various candidates of de have been reviewed in @xcite . a model of de , motivated by the holographic principle
, was proposed by li @xcite and it has been further studied in the references like @xcite . the energy density of hde @xmath13 as follows : @xmath14 with @xmath15 indicating a dimensionless constant parameter which which value @xmath16 is evinced by observational data : for a flat universe ( i.e. for @xmath17 ) it is obtained that @xmath18 and in the case of a non - flat universe ( i.e. for @xmath19 or @xmath20 ) it is obtained thar @xmath21 @xcite .
@xcite used the hde model in order to drive inflation in the early evolutionary phases of the universe .
@xcite studied the eos parameter @xmath0 of the hde model considering not a constant but a time - dependent newton s gravitational constant , _
i.e. _ @xmath22 ; furthermore , they obtained that @xmath0 can be significantly modified in the low redshift limit . recently ,
the cosmic acceleration has been also well studied by imposing the concept of modification of gravity @xcite .
this new model of gravity ( predicted by string / m theory ) gives a very natural gravitational alternative to the idea of the presence of exotic components .
the explanation of the phantom , non - phantom and quintom phases of the universe can be well described using modified gravity theories without the necessity of the introduction of a negative kinetic term in de models .
the relevance of modified gravity models for the late acceleration of the universe has been recently studied by many authors .
some of the most famous and known models of modified gravity are represented by braneworld models , @xmath23 modified gravity ( where @xmath24 indicates the torsion scalar ) , @xmath25 modified gravity ( where @xmath26 indicates the ricci scalar curvature ) , @xmath27 modified gravity ( where @xmath28 indicates the gauss - bonnet invariant which is defined as @xmath29 , with @xmath26 representing the ricci scalar curvature , @xmath30 representing the ricci curvature tensor and @xmath31 representing the riemann curvature tensor ) , @xmath32 modified gravity , @xmath33 modified gravity , dgp model , dbi models , horava - lifshitz gravity and brans - dicke gravity . modified theories of gravity have been reviewed in @xcite . recently , the idea that our universe is a brane which is embedded in a higher - dimensional space obtained a lot of attention from scientific community @xcite .
the friedmann equation on the brane has some corrections with respect to the usual four - dimensional equation @xcite .
_ @xcite found a term @xmath34 , which is problematic from an observational point of view .
the model is consistent if the tension on the brane and a cosmological constant in the bulk are considered .
this leads to a cosmological version of the randall - sundrum ( rs ) scenario of warped geometries @xcite .
@xcite considered an interaction between the bulk and the brane , which can be considered as another non - trivial aspect of braneworld theories .
the main aim of this paper is to disclose the effects produced by the energy exchange between the brane and the bulk on the evolutionary history of the universe by taking into account the flow of energy onto ( or away ) from the brane . in this paper
, we will focus our attention to three particular de models , i.e the hde model with ir cut - off given by the recently proposed granda - oliveros cut - off , the modified holographic de ( mhrde ) model and a de model which is proportional to the hubble parameter @xmath1 squared and to higher time derivatives of @xmath1 .
moreover , we will consider two different scale factors , _
i.e. _ the power law and the emergent ones , in order to study the cosmological properties of the de models in the bulk - brane interaction .
both the de models and the scale factors considered will be described in details in the following sections .
this study is motivated by the works of @xcite . in an interaction between the bulk and the brane ,
setare @xcite considered the holographic model of de in non - flat universe under the assumption that the cdm energy density on the brane is conserved while the hde energy density on the brane is not conserved because of to brane - bulk energy exchange .
sheykhi @xcite considered the agegraphic models of de in the framework of a braneworld scenario with brane - bulk energy exchange under the assumption that the adiabatic equation for the dm is satisfied but it is violated for the agegraphic de ( ade ) model because of the energy exchange between the brane and the bulk . in the paper of sheykhi @xcite , it was obtained that the eos parameter can evolve from the quintessence regime to the phantom regime .
myung @xmath35 kim @xcite introduced the brane - bulk interaction in order to discuss a limitation of the cosmological cardy - verlinde formula which is useful for the holographic description of brane cosmology .
they also showed that if there is presence of the brane - bulk interaction , it is not possible to derive the entropy representation of the first friedmann equation .
saridakis @xcite studied a generalized version of the hde model arguing that it must be taked into account in the maximally subspace of a cosmological model ; moreover he showed that , in the framework of brane cosmology , it leads to a bulk hde which transfers its holographic nature to the effective @xmath36 de .
furthermore , saridakis @xcite applied the bulk hde in general @xmath37 two - brane models and he also extracted the friedmann equation on the physical brane , showing that in the general moving - brane case the effective @xmath36 hde has a quintom - like behavior for a large parameter - space area of a simple solution subclass . in this paper
, we consider an interaction between the bulk and the brane , which represents a non - trivial aspect of the braneworld theories .
we also discuss the flow of energy onto or away from the brane - universe .
we then apply this idea to a braneworld cosmology under the assumption that the de energy density on the brane is conserved , but the de energy density on the brane is not conserved because of the brane - bulk energy exchange .
the plan of the paper is the following . in section 2
, we describe the main features of bulk - brane interaction . in section 3
, we describe the main features of the de models considered in this paper ; moreover , we derive the expression of the eos parameter @xmath0 and the evolutionary form of the parameter @xmath38 ( defined as @xmath39 ) for the de models we are considering . in section 4 , we consider two different models of scale factors , ( in particular , the power law and the emergent ones ) in order to study the behavior of the expression of @xmath40 derived in the previous section .
finally , in section 5 , we write the conclusion of this work .
in this section , we want to describe the main features of the bulk - brane interaction , introducing the main quantities useful for the following part of the work .
+ the bulk - brane action @xmath41 is given by the following expression @xcite : @xmath42 where @xmath43 represents the 5d curvature scalar , @xmath44 denotes the bulk cosmological constant , @xmath45 stands for the 5d coupling constant , @xmath46 indicates the brane tension , @xmath28 and @xmath47 denote the determinant of the 5d and of the 4d metric tensors , respectively while @xmath48 and @xmath49 are the matter lagrangian in the bulk and the matter lagrangian in the brane .
+ we here consider the cosmological solution with a metric given by @xcite : @xmath50 where @xmath51 represents the metric for the maximally symmetric three - dimensional space .
the non - zero components of einstein tensor are given by @xcite : @xmath52+\frac{kn^2}{b^2}\right\}~,\\ \label{gij } g_{ij}&=&\frac{a^{2}}{b^{2}}\gamma_{ij}\left[\frac{a'}{a}\left(\frac{a'}{a}+\frac{2n'}{n}\right)- \frac{b'}{b}\left(\frac{n'}{n}+\frac{2a'}{a}\right)+\frac{2a''}{a}+\frac{n''}{n}\right]+ \nonumber \\ & & \frac{a^{2}}{n^{2}}\gamma_{ij}\left[\frac{\dot{a}}{a}\left(-\frac{\dot{a}}{a}+\frac{2\dot{n}}{n}\right ) -\frac{2\ddot{a}}{a}+\frac{\dot{b}}{b}\left(-\frac{\dot{2a}}{a}+\frac{\dot{n}}{n}\right ) -\frac{\ddot{b}}{b}\right]-k\gamma_{ij}~ , \\
\label{g05 } g_{05}&=&3\left(\frac{n'}{n}\frac{\dot{a}}{a}+\frac{a'}{a}\frac{\dot{b}}{b}-\frac{\dot{a}'}{a}\right)~,\\ \label{g55 } g_{55}&=&3\left\{\frac{a'}{a}\left(\frac{a'}{a}+\frac{n'}{n}\right)-\frac{b^2}{n^2 } \left[\frac{\dot{a}}{a}\left(\frac{\dot{a}}{a}-\frac{\dot{n}}{n}\right)+\frac{\ddot{a}}{a}\right ] -\frac{kb^2}{a^2}\right\},\end{aligned}\ ] ] where @xmath53 denotes the curvature parameter of space which possible values are @xmath54 which correspond , respectively , to a flat , a closed and an open universe .
moreover , the primes and the dots indicate , respectively , a derivative with respect to the variable @xmath55 and a derivative with respect to the variable @xmath56 . the 4d braneworld universe is assumed to be at @xmath57 .
the einstein equations are given by : @xmath58 where we have that the stress - energy momentum tensor @xmath59 has both bulk and brane components and it can be also written as follows @xcite : @xmath60 where : @xmath61 where @xmath6 and @xmath7 represent , respectively , the total pressure and the total density on the brane .
+ by integrating eqs .
( [ g00 ] ) and ( [ gij ] ) with respect to the variable @xmath55 around the point @xmath62 and assuming the @xmath63 symmetry around the brane , we derive the following jump conditions : @xmath64 the 2 subscripts + and - indicate , respectively , the sides corresponding to @xmath65 and @xmath66 , which represent the two sides of the brane embedded in the bulk .
moreover , the subscript 0 indicates quantities which are evaluated at @xmath62 .
+ starting from the results of eqs .
( [ g05 ] ) and ( [ g55 ] ) , we can obtain the following expressions : @xmath67 -k\frac{b_0 ^ 2}{a_0 ^ 2}\right\ } & = & -\kappa _ 5 ^ 2\lambda_5 b_0 ^ 2 + \kappa_5 ^ 2t_{55},\end{aligned}\ ] ] where the terms @xmath68 and @xmath69 represent , respectively , the @xmath70 and @xmath71 components of @xmath72 when evaluated on the brane .
+ moreover , using eqs .
( [ jump1 ] ) and ( [ jump2 ] ) , we obtain : @xmath73&+&\frac{k}{a_{0}^{2 } } \nonumber \\ & = & \frac{\kappa_{5}^{2}}{3}\left(\lambda_{5}+\frac{\kappa_{5}^{2}\sigma^{2}}{6}\right ) \nonumber \\ & & -\frac{\kappa_{5}^{4}}{36}\left[\sigma(3p-\rho)+\rho(3p+\rho)\right]-\frac{\kappa_{5}^{2}}{3}t^{5}_{5}.\end{aligned}\ ] ] considering an appropriate gauge with the coordinate frame @xmath74 , eqs .
( [ field1 ] ) and ( [ field2 ] ) can be also expressed in the following equivalent forms : @xmath75 where @xmath76 and @xmath77 .
the effective 4d cosmological constant @xmath3 on the brane , the bulk cosmological constant @xmath44 , and the brane tension @xmath46 are well known to be constrained by the fine - tuning relation @xcite : @xmath78 if we assume that the bulk matter ( relative to bulk vacuum energy ) is much less than the ratio of the brane matter to the brane vacuum energy , we can neglect the @xmath79 term : this can lead to the derivation of a solution that is largely independent of the bulk dynamics .
if we take into account this approximation and we concentrate on the low - energy region with @xmath80 , eqs .
( [ field1 ] ) and ( [ field2 ] ) can be simplified , leading to the following system of equations : @xmath81 the auxiliary field @xmath82 ( which appear in eqs .
( [ field22 ] ) and ( [ field33 ] ) ) incorporates non - trivial contributions of de which differ from the standard matter fields confined to the brane .
hence , with the energy exchange @xmath24 between the bulk and brane , the usual energy conservation is violated .
we shall denote the energy density of de by @xmath13 . since we will consider two dark components in the universe , namely , dm and de
, we will have @xmath83 .
+ in the following section , three different de models are concerned , namely , the hde model with granda - oliveros cut - off , the mhrde model and the de model proportional to the hubble parameter @xmath1 squared and to higher time derivatives of @xmath1 in the framework of bulk brane interaction .
it is accomplished by using some of the concepts introduced in this section and two choices of the scale factor , namely the power law and the emergent ones .
we now want to give a description of the de models considered in this work and to find some relevant cosmological quantities . we will also introduce some relevant equations which will be useful for the understanding of the work .
+ the bulk - brane interaction has been studied for various aspects , where in particular the effective de of the braneworld universe is dynamical , as a result of the non - minimal coupling , which gives a mechanism for bulk - brane interaction through gravity @xcite .
we assume here that the adiabatic equation for the dm is satisfied , while it results to be violated for de due to the energy exchange between the brane and the bulk @xcite .
then , we obtain the following continuity equations : @xmath84 we define the fractional energy densities for dm , de and @xmath82 , respectively , as follows : @xmath85 the planck data provide the values @xmath86 and @xmath87 at @xmath88 cl @xcite . the critical energy density @xmath89 ( _ i.e. _ the energy density required for flatness ) is defined as follows : @xmath90 or , assuming units of @xmath91 , as : @xmath92 using the definition of @xmath89 given in eq .
( [ rhocr ] ) , we can write the fractional energy densities given in eqs .
( [ fracden1 ] ) , ( [ fracden2 ] ) and ( [ fracden3 ] ) , respectively , as follows : @xmath93 the interaction between bulk and brane is given by the relation @xmath94 , where the parameter @xmath95 represents the rate of interaction .
the wilkinson microwave anisotropy probe ( wmap ) satellite is well known to have measured the curvature parameter @xmath96 in eq .
( [ fracden4 ] ) , and , along with baryon acoustic oscillation ( bao ) and hubble parameter measurement , it constrained the fractional energy density of the curvature parameter @xmath53 as @xmath97 , in 95% cl @xcite .
( [ fracden4 ] ) for @xmath96 is hence equal to zero in this context . considering the parameter @xmath98
, the above equations lead to @xcite : @xmath99.\ ] ] in this paper , we decided to consider the particular case corresponding to @xmath100 .
furthermore , following @xcite , we have chosen the following expression for @xmath95 : @xmath101 where @xmath102 represents a coupling parameter between dm and de , also known as transfer strength @xcite . from the observational data of the gold sneia samples , cmbr data obtained from the wmap and planck satellites and the baryonic acoustic oscillations ( bao ) obtained thanks to the sloan digital sky survey ( sdss )
, the coupling parameter between dm and de is estimated to assume a small positive value , satisfying the requirement for solving the cosmic coincidence problem and the constraints given by the second law of thermodynamics @xcite .
cosmological observations of the cmbr anisotropies and of clusters of galaxies indicate that @xmath103 @xcite .
this evidence is in agreement with the fact that @xmath102 must be taken in the range of values [ 0,1 ] @xcite , with @xmath104 representing the non - interacting flrw model . +
using the definitions of the fractional energy densities given in eqs .
( [ fractional1 ] ) , ( [ fractional2 ] ) and ( [ fractional3 ] ) , we can rewrite the first friedmann equation defined in eq .
( [ field22 ] ) as follows : @xmath105 eq .
( [ allfrac ] ) has the main property of relating all the fractional energy densities considered in this work .
+ moreover , using eqs .
( [ fractional1 ] ) , ( [ fractional2 ] ) and ( [ fractional3 ] ) along with the definition of @xmath38 and the relation @xmath106 , we can easily obtain the following relation between the parameter @xmath38 and the fractional energy densities : @xmath107 we now want to introduce three different energy density models for de , _
i.e. _ the hde with granda - oliveros cut - off , the modified holographic ricci de ( mhrde ) model and the de model proportional to @xmath108 and to higher time derivatives of @xmath1 . before proceeding with calculations , we briefly describe these three models .
+ recently , granda @xmath35 oliveros introduced a new ir cut - off based on purely dimensional ground which includes a term proportional to @xmath109 and one term proportional to @xmath108 .
this new ir cut - off is known as granda - oliveros ( go ) scale , indicated with the symbol @xmath110 and it is given by @xcite : @xmath111 where @xmath112 and @xmath113 represent two constant parameters . in the limiting case corresponding to @xmath114 and @xmath115 ,
the go scale @xmath110 becomes proportional to the average radius of the ricci scalar curvature ( i.e. , @xmath116 ) in the case the curvature parameter @xmath53 assume the value of zero ( i.e. @xmath17 ) , corresponding to a flat universe .
recently , wang @xmath35 xu @xcite have constrained the new hde model in non - flat universe using observational data .
the best fit values of the two parameters @xmath117 with their confidence levels they found are given by @xmath118 and @xmath119 for non flat universe , while for flat universe they found that are @xmath120 and @xmath121 .
+ we decided to consider the go scale @xmath110 as infrared cut - off for some specific reasons .
if the ir cut - off chosen is given by the particle horizon , the hde model can not produce an expansion of the universe with accelerated rate . if we consider as cut - off of the system the future event horizon , the hde model has a causality problem .
the de models which consider the go scale @xmath110 depend only on local quantities , then it is possible to avoid the causality problem , moreover it is also possible to obtain the accelerated phase of the universe .
+ granda @xmath35 oliveros considered that , since the origin of the hde model is still not known exactly up to now , the consideration of the term with the time derivative of the hubble parameter in the expression of the energy density of de may be expected since this term appears in the curvature scalar and it has the right dimension . + the expression of the hde energy density with @xmath110 cut - off is given by : @xmath122 we must underline here that we are considering the planck mass @xmath123 equal to one .
+ contrary to the hde model based on the event horizon , the de models which consider the go scale depend only on local quantities , then it is possible to avoid in this way the causality problem .
+ the second model we consider in this paper is the modified holographic ricci de ( mhrde ) model , which is given by the following expression : @xmath124 where @xmath125 and @xmath126 are the model parameters .
hereupon , we shall denote by @xmath127 any quantity @xmath128 related to the mhrde model .
this de model was studied for the non - interacting case in reference @xcite , and chimento _ et al . _
have analyzed this this type of de in interaction with dm as chaplygin gas @xcite . in the limiting case corresponding to @xmath129 and @xmath130 , the de energy density model given in eq .
( [ mhrde ] ) leads to the de energy density with ricci scalar curvature for a spatially at flrw space - time as ir cut - off .
+ the use of the mhrde is motivated by the holographic principle since we can relate the de with an uv cut - off for the vacuum energy with an ir scale such as the one given by the ricci scalar curvature @xmath26 .
moreover , it is possible to proceed in a different way taking into account that the ricci scalar curvature @xmath26 is a new kind of de , for example , a geometric de instead of evoking the holographic principle .
irrespective of the origin of the de component , it modifies the friedmann equation leading to a second order differential equation for the scale factor .
+ in this work , we decided to consider also a de energy density model which was recently proposed by chen @xmath35 jing @xcite .
this new model is function of the hubble parameter squared @xmath108 and of the first and second derivatives with respect to the cosmic time @xmath56 of the hubble parameter @xmath1 and it is given by the following expression : @xmath131 where @xmath125 , @xmath126 and @xmath132 represent three arbitrary dimensionless parameters .
the inverse of the hubble parameter , _
i.e. _ @xmath133 , is introduced in the first term of eq .
( [ lgo5 - 1higher ] ) so that the dimensions of each of the three terms are the same
. + the behavior and the main cosmological features of the de energy density model defined in eq .
( [ lgo5 - 1higher ] ) strongly depend on the three parameters of the model , _
i.e. _ @xmath125 , @xmath126 and @xmath132 .
( [ lgo5 - 1higher ] ) can be considered as a generalization of two previously proposed energy density models of de .
in fact , in the limiting case corresponding to @xmath134 , we recover the energy density of de in the case the ir cut - off of the system is given by the granda - oliveros cut - off .
moreover , in the limiting case corresponding to @xmath135 , @xmath136 and @xmath137 , we obtain the expression of the energy density of de with ir cut - off proportional to the average radius of the ricci scalar ( i.e. , @xmath138 ) in the case of curvature parameter @xmath53 assumes the value of zero ( @xmath17 ) .
+ using the expressions of the energy densities of de given in eqs . ( [ lgo5 - 1 ] ) , ( [ mhrde ] ) and ( [ lgo5 - 1higher ] ) in eq .
( [ fractional1 ] ) , we obtain the following expressions for @xmath139 , @xmath140 and @xmath141 : @xmath142 the final expression of @xmath143 can be derived by first solving the continuity equation for @xmath143 given in eq .
( [ matter ] ) , yielding : @xmath144 where @xmath145 indicates the present day of the energy density of dm .
+ using the expression of @xmath146 given in eq .
( [ rhoemme ] ) , we can write the fractional energy density of dm as follows : @xmath147 we now want to find the final expressions of the eos parameter @xmath0 and of @xmath40 for all the de models considered in this work .
+ differentiating eq .
( [ field22 ] ) with respect to the cosmic time @xmath56 and using eqs .
( [ field11 ] ) - ( [ field33 ] ) , we obtain ( considering units of @xmath91 ) the following expression of the time derivative of the hubble parameter for flat universe : @xmath148.\ ] ] moreover , using eqs .
( [ field22 ] ) and ( [ hdot ] ) in eqs .
( [ lgo5 - 1 ] ) , ( [ mhrde ] ) and ( [ lgo5 - 1higher ] ) , we obtain the following expressions for the eos parameters of the de models we are dealing with : @xmath149 using the relation between @xmath38 and @xmath82 given by @xmath150 , we can find the following expression for @xmath151 : @xmath152 then , inserting eq .
( [ u ] ) in the expressions of the eos parameters obtained in eqs .
( [ eosgo ] ) , ( [ eosmh ] ) and ( [ eoshigher ] ) , along with the relation @xmath153 , we can rewrite eqs .
( [ eosgo ] ) , ( [ eosmh ] ) and ( [ eoshigher ] ) as follows : @xmath154 we must underline that in eq .
( [ eoshigher1 ] ) we used the main definition of @xmath141 given in eq .
( [ fractional1higher ] ) .
+ moreover , using the relation @xmath155 ( which can be obtained from eq .
( [ ufra ] ) ) in eqs .
( [ eosgo1 ] ) , ( [ eosmh1 ] ) and ( [ eoshigher1 ] ) , we can write : @xmath156 - \left ( 1-\frac{2}{3}\frac{\alpha}{\beta } + \frac{2}{3c^2\beta } \right ) \nonumber\\ & & + \left ( \frac{2\alpha -3\beta}{3\beta } \right)\left [ \frac{1}{\left(1+u_{go}\right)\omega_{d_{go } } } -1 \right]\nonumber \\ & = & \left[\frac{2u_{go}}{3}\left ( \frac{\alpha - 2\beta}{\beta } \right ) + \frac{2}{3}\frac{\alpha}{\beta}-1 \right]\left [ \frac{1}{\left(1+u_{go}\right)\omega_{d_{go } } } \right ] - \frac{2}{3c^2\beta } , \label{eosgo2}\\ \omega_{d_{\tiny\circ } } & = & \left ( \alpha -1 \right)\left[\frac{1}{\left(1+u_{\tiny\circ}\right)\omega_{d_{\tiny\circ}}-1 } \right]+\left(\alpha - \frac{4}{3 } \right)u\left ( 1 + \frac{\omega_m}{\omega_{d_{\tiny\circ } } } \right ) + \beta -1,\label{eosmh2}\\ \omega_{d , higher } & = & \frac{2u_{higher}}{3}\left ( \frac{\alpha - 2\beta}{\beta } \right)\left [ \frac{1}{\left(1+u_{higher}\right)\omega_{d , higher } } \right ] - \left ( 1-\frac{2}{3}\frac{\alpha}{\beta } + \frac{2}{3c^2\beta } \right ) \nonumber\\ & & + \left ( \frac{2\alpha -3\beta}{3\beta } \right)\left [ \frac{1}{\left(1+u_{higher}\right)\omega_{d , higher } } -1 \right ] + \frac{2\varrho}{3\beta \omega_{d , higher}}\frac{\ddot{h}}{h^3}\nonumber \\ & = & \left[\frac{2u_{higher}}{3}\left ( \frac{\alpha - 2\beta}{\beta } \right ) + \frac{2}{3}\frac{\alpha}{\beta}-1 \right]\left [ \frac{1}{\left(1+u_{higher}\right)\omega_{d , higher } } \right ] - \frac{2}{3c^2\beta } \nonumber \\ & & + \frac{2\varrho}{3\beta \omega_{d , higher}}\frac{\ddot{h}}{h^3}. \label{eoshigher2}\end{aligned}\ ] ] using eqs .
( [ matter ] ) and ( [ energy ] ) along with the expression of @xmath95 we have chosen , we obtain the following expression for the time evolution of @xmath38 for the three different de models we are dealing with : @xmath157 , \label{uasa1}\\ \dot{u}_{\tiny\circ}&=&\left(\frac{3hu_{\tiny\circ}\omega_{d_{\tiny\circ}}}{\omega_{d_{\tiny\circ}}+\omega_{m}}\right)\left[\omega_{d_{\tiny\circ}}- \frac{1}{3}\left(\frac{\omega_{m } + \omega_{d_{\tiny\circ}}}{\omega_{d_{\tiny\circ}}}\right ) -\frac{b^2\left(1+u_{\tiny\circ}\right)^2}{u_{\tiny\circ}}\right ] , \label{uasa1 - 1}\\ \dot{u}_{higher}&=&\left(\frac{3hu_{higher}\omega_{d , higher}}{\omega_{d , higher}+\omega_{m}}\right)\left[\omega_{d , higher}- \frac{1}{3}\left(\frac{\omega_{m } + \omega_{d , higher}}{\omega_{d , higher}}\right)-\frac{b^2\left(1+u_{higher}\right)^2}{u_{higher } } \right ] . \label{uasahigher}\end{aligned}\ ] ] inserting the expressions of the eos parameters obtained in eqs .
( [ eosgo2 ] ) , ( [ eosmh2 ] ) and ( [ eoshigher2 ] ) into eqs .
( [ uasa1 ] ) , ( [ uasa1 - 1 ] ) and ( [ uasahigher ] ) and using the relation @xmath158 , we obtain the following expressions for the three different de models considered : @xmath159\left [ \frac{1}{\left(1+u_{go}\right)\omega_{d_{go}}}\right]\right . \nonumber \\ & & \left .
-\frac{2}{3c^2\beta } -\frac{b^2\left(1+u_{go}\right)^2}{u_{go}}\right\ } , \label{uasanew1}\\ \dot{u } _ { \tiny\circ}&= & \frac{3hu_{\tiny\circ}\left ( 1+u_{\tiny\circ } \right)}{\omega_{d_{\tiny\circ } } } \left\ { \left(\alpha - \frac{4}{3}\right ) \left ( \frac{1}{\omega_{d_{\tiny\circ}}}\right ) + \beta - \alpha -\frac{b^2\left(1+u_{\tiny\circ}\right)^2}{u_{\tiny\circ } } \right\ } , \label{uasanew2}\\ \dot{u}_{higher } & = & \frac{3hu_{higher}\left ( 1+u_{higher } \right)}{\omega_{d , higher } } \left\ { \left [ \frac{2}{3}u_{higher}\left ( \frac{\alpha - 2\beta}{\beta } \right ) + \frac{2\alpha}{3\beta}-\frac{4}{3 } \right]\times \right.\nonumber \\ & & \left .
\left [ \frac{1}{\left(1+u_{higher}\right)\omega_{d , higher}}\right ] + \frac{2\varrho}{3\beta \omega_{d , higher}}\frac{\ddot{h}}{h^3}\right .
\nonumber \\ & & \left .
-\frac{2}{3c^2\beta } -\frac{b^2\left(1+u_{higher}\right)^2}{u_{higher}}\right\ } .
\label{uasanew3}\end{aligned}\ ] ] in the following section , we will study the behavior of the evolutionary forms of @xmath160 , @xmath161 and @xmath162 obtained , respectively , in eqs .
( [ uasanew1 ] ) , ( [ uasanew2 ] ) and ( [ uasanew3 ] ) for two different choices of the scale factor , _
i.e. _ the power law and the emergent ones . using the reconstructed expressions of @xmath38
, we will use them in order to study the behavior of the eos parameters for the three de models we are considering and obtained , respectively , in eqs .
( [ eosgo1 ] ) , ( [ eosmh1 ] ) and ( [ eoshigher1 ] ) . we must also emphasize that will find the final expression of the term @xmath163 according to the choice of the scale factor we will make . +
in this section , we want to study the behavior of the reconstructed expressions of @xmath38 , determined from @xmath160 , @xmath161 and @xmath162 obtained , respectively , in eqs .
( [ uasanew1 ] ) , ( [ uasanew2 ] ) and ( [ uasanew3 ] ) , for two different choices of the scale factor , _ i.e. _ the power law and the emergent ones .
+ in order to find the final expressions of @xmath40 for the different choices of scale factor , we need to calculate the expressions of @xmath139 , @xmath140 and @xmath141 ( defined , respectively , in eqs .
( [ fractional1go ] ) , ( [ fractional1mh ] ) and ( [ fractional1higher ] ) ) and @xmath1 for the relevant case of the scale factor ( remembering that @xmath164 ) .
we will then plot the reconstructed expressions of @xmath38 derived from @xmath40 we obtained for some range of values of the parameters involved .
thanks to the reconstructed expression of @xmath38 , we can plot the behavior of the eos parameter @xmath0 for the relevant model and the specific scale factor .
we start the study of the different scale factors taking into account the power law scenario .
+ following setare @xcite , we consider the power law case of the scale factor in the following form : @xmath165 where @xmath166 , @xmath167 and @xmath168 are three constants .
the term @xmath167 indicates the finite future singularity time and the scale factor defined in eq .
( [ plscale ] ) is often used in scientific literature in order to check the type ii ( sudden singularity ) or type iv ( which corresponds to @xmath109 ) for positive values of the power law index @xmath168 .
+ we have that the derivative of the scale factor given in eq .
( [ plscale ] ) with respect to the cosmic time @xmath56 is given by : @xmath169 using the results of eqs .
( [ plscale ] ) and ( [ plscale2 ] ) , we obtain that the expression of the hubble parameter and its first and second time derivatives are given , respectively , by : @xmath170 using the expression of @xmath1 obtained in eq .
( [ hpl ] ) and the expressions of @xmath139 , @xmath140 and @xmath171 , obtained inserting in eqs .
( [ fractional1go ] ) , ( [ fractional1mh ] ) and ( [ fractional1higher ] ) the expressions of @xmath172 , @xmath173 and @xmath174 defined in eqs .
( [ lgo5 - 1 ] ) , ( [ mhrde ] ) and ( [ lgo5 - 1higher ] ) calculated for @xmath1 , @xmath109 and @xmath175 given in eqs .
( [ hpl ] ) , ( [ hdotpl ] ) and ( [ hddotpl ] ) , we derive the following expressions for @xmath160 , @xmath161 and @xmath176 : @xmath177\times \nonumber \\ & & \left\ { \left [ \frac{2}{3}u\left ( \frac{\alpha - 2\beta}{\beta } \right ) + \frac{2\alpha}{3\beta}-\frac{4}{3 } \right ] \frac{n}{c^2 ( -\alpha + n \beta ) \left(1+u_{go}\right ) } \right .
\nonumber \\ & & \left .
-\frac{2}{3c^2\beta } -\frac{b^2\left(1+u_{go}\right)^2}{u_{go}}\right\ } , \label{ugopl}\\ \dot{u}_{\tiny\circ } & = & 3u_{\tiny\circ}\left ( 1+u_{\tiny\circ } \right ) \left \ { \frac{3n^2\left
( -\alpha + \beta \right)}{\left ( t_s -t \right)\left ( 3n\alpha -2 \right ) } \right\ } \times \nonumber \\ & & \left\ { \left[\left(\alpha - \frac{4}{3}\right ) \frac{3n ( \alpha -\beta ) } { -2 + 3 n \alpha } \right]+\beta - \alpha -\frac{b^2\left(1+u_{\tiny\circ}\right)^2}{u_{\tiny\circ } } \right\ } , \label{umhpl}\\ \dot{u}_{higher } & = & 3u_{higher}\left ( 1+u_{higher } \right ) \left [ \frac{n^2}{c^2 ( t_s - t ) ( \alpha -n \beta ) } \right]\times \nonumber \\ & & \left\ { \left [ \frac{2}{3}u\left ( \frac{\alpha - 2\beta}{\beta } \right ) + \frac{2\alpha}{3\beta}-\frac{4}{3 } \right ] \frac{n}{c^2 ( -\alpha + n \beta ) \left(1+u_{higher}\right ) } \right .
\nonumber \\ & & \left .
+ \frac{4\varrho}{3c^2\beta \left [ n\left(n\alpha - \beta\right ) + 2\varrho \right ] } -\frac{2}{3c^2\beta } -\frac{b^2\left(1+u_{higher}\right)^2}{u_{higher}}\right\ } .
\label{uhigherpl}\end{aligned}\ ] ] by using numerical integration , the evolution for @xmath178 , @xmath179 and @xmath176 are depicted in figures [ ugopowerlaw ] , [ umh ] and [ uhigherpowerlaw ] . for the case pertaining to the hde model with go cut - off , we considered three different cases , _
i.e. _ for @xmath180 ( plotted in red ) , @xmath181 ( plotted in green ) and @xmath182 ( plotted in blue ) , while the other parameters involved have been chosen as @xmath183 , @xmath184 , @xmath185 , @xmath186 and @xmath187 .
it is worthwhile to emphasize that @xmath178 has a monotonically increasing behavior for all the three cases considered .
+ for the case corresponding to the mhrde model , three different cases are regarded , namely @xmath188 ( plotted in red ) , @xmath189 ( plotted in green ) and @xmath190 ( plotted in blue ) , while the other parameters involved have been chosen as @xmath191 , @xmath192 , @xmath193 , @xmath194 and @xmath187 . as for the previous case , an increasing profile of @xmath179 can be observed , for all the three cases considered .
+ for the model proportional to higher time derivatives of the hubble parameter @xmath1 , we have considered three different cases , corresponding to @xmath180 ( plotted in red ) , @xmath181 ( plotted in green ) and @xmath182 ( plotted in blue ) , while the other parameters involved have been chosen as @xmath183 , @xmath195 , @xmath184 , @xmath185 , @xmath186 and @xmath187 .
we can observe in figure [ uhigherpowerlaw ] that @xmath176 monotonically increases for all the cases considered , as also found for the other two de model considered .
+ these increasing behaviors of @xmath178 , @xmath179 and @xmath176 shown in figures [ ugopowerlaw ] , [ umh ] and [ uhigherpowerlaw ] clearly indicate a non - trivial contribution of de , contribution which increases with the temporal evolution of the universe . for power - law scale factor against the time @xmath56 .
the increasing pattern indicates that the non - trivial contribution of de increases with the evolution of the universe.,width=302 ] for power - law scale factor against the time @xmath56 . the increasing pattern indicates the non - trivial contribution of de increases with the evolution of the universe.,width=302 ] for power - law scale factor against the time @xmath56 .
the increasing pattern indicates that the non - trivial contribution of de increases with the evolution of the universe.,width=302 ] against the time @xmath56 for power - law scale factor .
we observe a decreasing behavior for all cases considered.for @xmath180 ( plotted in red ) , @xmath196 starts being @xmath197 , then it decreases and it can eventually cross @xmath198 . for the other two cases , we obtain that @xmath196 is always lower that @xmath199 . ,
width=302 ] against the time @xmath56 for the emergent scale factor .
for all the cases considered , we have that @xmath200 has a slowly decreasing pattern and it is always greater than @xmath199.,width=302 ] against the time @xmath56 for power - law scale factor .
for all the cases considered , we have that @xmath201 has a decreasing pattern and it is always greater than @xmath199 . , width=302 ] using the reconstructed expressions of @xmath178 , @xmath179 and @xmath176 obtained , respectively , from eqs .
( [ ugopl ] ) , ( [ umhpl ] ) and ( [ uhigherpl ] ) and plotted in figures [ ugopowerlaw ] , [ umh ] and [ uhigherpowerlaw ] , we can also derive and plot the profile of the eos parameters obtained in eqs .
( [ eosgo1 ] ) , ( [ eosmh1 ] ) and ( [ eoshigher1 ] ) for the three de models concerned .
+ for the de model with go cut - off , we obtain that , for @xmath180 , the eos parameter @xmath196 starts being @xmath197 , while with the passing of the time , it decreases and it asymptotically reaches the value @xmath199 and can eventually cross it . for the other two cases , we obtain that @xmath196 has a decreasing behavior , being always lower that @xmath199 .
+ instead , for the mhrde model , we obtain that the eos parameter @xmath200 has a slowly decreasing behavior for all the three cases considered , staying always greater than @xmath199 .
+ for the model proportional to higher time derivatives of the hubble parameter @xmath1 , we observe a slowly decreasing behavior of the eos parameter @xmath201 , with @xmath202 for the range of time considered .
it is possible that , for sufficiently high time , the three models can cross the value @xmath203 .
+ we now consider some particular values of the parameters involved . + for the de model with go cut - off , we study the case corresponding to the ricci scale , which is recovered for @xmath204 and @xmath130 ( plotted in green ) and we will also consider the case corresponding to @xmath205 and @xmath206 ( plotted in red ) . instead , for the mhrde model , we will consider the case corresponding to the ricci scale , which is recovered for @xmath129 and @xmath130 .
the values of the other parameters have been taken as the previous cases considered .
+ we can clearly see in figure [ ugopowerlawlim ] that , for both limiting cases , @xmath178 has a decreasing behavior while @xmath196 has a slowly increasing behavior .
moreover , we have that for the case pertaining to the ricci scale , @xmath196 is always greater than -1 while for the case with @xmath205 and @xmath206 it is always lower than -1 .
+ for the limiting case of the mhrde , we observe that @xmath179 has an increasing behavior while @xmath200 slowly decreases , being always greater than -1 . for power - law scale factor against the time @xmath56 for the limiting cases of @xmath204 and @xmath130 ( plotted in green ) and @xmath205 and @xmath206 ( plotted in red ) .
, width=302 ] for power - law scale factor against the time @xmath56 for the limiting case of @xmath129 and @xmath130 . ,
width=302 ] against the time @xmath56 for power - law scale factor for the limiting cases of @xmath204 and @xmath130 ( plotted in green ) and @xmath205 and @xmath206 ( plotted in red).,width=302 ] against the time @xmath56 for the emergent scale factor for the limiting case of @xmath129 and @xmath130 . ,
width=302 ] we now consider the second scale factor chosen in this work , _
i.e. _ the emergent one .
+ this form of scale factor @xmath207 , as stated in @xcite is given by : @xmath208 where @xmath166 , @xmath209 , @xmath210 and @xmath211 represents four positive constant parameters .
we can make some considerations about the values which can be assumed by the parameters present in eq .
( [ aeme ] ) : * if both @xmath212 and @xmath211 are negative defined , then the emergent scenario produces the big bang singularity at the infinity paste time , _
i.e. _ for @xmath213 * @xmath166 must be a positive defined quantity if we want to have the scale factor of the emergent scenario as a positive defined quantity * @xmath212 or @xmath211 must be positive defined in order to obtain an expanding model of the universe * @xmath209 must be a positive defined quantity if we want to avoid singularities ( like the big rip ) at finite time @xmath56 consequences of this choice are discussed in @xcite .
+ the emergent scenario of the universe in the framework of de has been taken into account in many recent papers .
@xcite studied the generalized second law of thermodynamics ( gslt ) for the emergent scenario of the universe for some particular models of @xmath23 modified gravity theory .
@xcite studied a general context for an emergent universe scenario and they derived that the emergent universe scenarios do not represent isolated solutions but they can occur for different combinations of matter and radiation .
del campo _
@xcite considered the emergent model of scale factor in the framework of a self - interacting jordan - brans - dicke modified gravity theory : they derived that this model is able to lead to a stable past eternal static solution which eventually is able to enter a phase where the stability is broken , which leads to a period of inflation .
+ the first time derivative of the scale factor for the emergent scenario given in eq .
( [ aeme ] ) is : @xmath214^{m-1}. \label{dotaeme}\end{aligned}\ ] ] using the definition of the scale factor given in eq .
( [ aeme ] ) along with its time derivative given in eq .
( [ dotaeme ] ) , we can easily derive that the hubble parameter @xmath1 and its first and second derivatives with respect to the cosmic time @xmath56 are given , respectively , by the following expressions : @xmath215 ^ 2 } , \label{dotheme}\\ \ddot{h } & = & \frac{\ddot{h}}{dt^2 } = \frac{m\lambda \mu^3e^ { \mu t } \left ( \lambda - e^ { \mu t } \right ) } { \left[e^ { \mu t } + \lambda\right]^3}\label{ddoteme}.\end{aligned}\ ] ] using the expression of @xmath1 obtained in eq .
( [ heme ] ) and the expressions of @xmath139 , @xmath140 and @xmath171 , obtained inserting in eqs .
( [ fractional1go ] ) , ( [ fractional1mh ] ) and ( [ fractional1higher ] ) the expressions of @xmath172 , @xmath173 and @xmath174 defined in eqs .
( [ lgo5 - 1 ] ) , ( [ mhrde ] ) and ( [ lgo5 - 1higher ] ) calculated for @xmath1 , @xmath109 and @xmath175 given in eqs .
( [ heme ] ) , ( [ dotheme ] ) and ( [ ddoteme ] ) , we derive the following expressions for @xmath160 , @xmath161 and @xmath176 : @xmath216 \times \nonumber \\ & & \left\ { \left [ \frac{2}{3}u\left ( \frac{\alpha - 2\beta}{\beta } \right ) + \frac{2\alpha}{3\beta}-\frac{4}{3 } \right ] \frac{e^{t \mu } m}{c^2 \left(e^{t \mu } m \beta + \alpha \lambda \right)\left(1+u_{go}\right ) } \right .
\nonumber \\ & & \left .
-\frac{2}{3c^2\beta } -\frac{b^2\left(1+u_{go}\right)^2}{u_{go}}\right\ } , \label{ugoeme}\\ \dot{u}_{\tiny\circ } & = & 3u_{\tiny\circ}\left ( 1+u_{\tiny\circ } \right ) \left\ { \frac{3e^{2\mu t}m^2 \mu \left ( \alpha -\beta \right)}{\left ( e^{\mu t } + \lambda \right ) \left ( 3 e^{t \mu } m \alpha + 2 \lambda \right ) } \right\}\times \nonumber \\ & & \left\ { \left[\left(\alpha - \frac{4}{3}\right)\frac{3e^{t \mu } m ( \alpha -\beta ) } { 3 e^{t \mu } m \alpha + 2 \lambda } \right]+\beta - \alpha -\frac{b^2\left(1+u_{\tiny\circ}\right)^2}{u_{\tiny\circ } } \right\ } , \label{umheme}\\ \dot{u}_{higher } & = & 3u_{higher}\left ( 1+u_{higher } \right ) \left [ \frac{e^{2 t \mu } m^2 \mu } { c^2 \left(e^{t \mu } + \lambda \right ) \left(e^{t \mu } m \beta + \alpha \lambda \right ) } \right ] \times \nonumber \\ & & \left\ { \left [ \frac{2}{3}u\left ( \frac{\alpha - 2\beta}{\beta } \right ) + \frac{2\alpha}{3\beta}-\frac{4}{3 } \right ] \frac{e^{t \mu } m}{c^2 \left(e^{t \mu } m \beta + \alpha \lambda \right)\left(1+u_{higher}\right ) } \right .
\nonumber \\ & & \left .
-\frac{2\varrho \left(e^{t \mu } -\lambda \right)\lambda}{3c^2\beta\left [ m^2\alpha e^{2 t \mu } + e^{t \mu } \left(m\beta - \varrho \right)\lambda + \varrho \lambda^2 \right ] } -\frac{2}{3c^2\beta } -\frac{b^2\left(1+u_{higher}\right)^2}{u_{higher}}\right\}. \label{uhighereme}\end{aligned}\ ] ] as accomplished for the previous model studied , we use a numerical integration in order to obtain the evolutionary forms of @xmath178 , @xmath179 and @xmath176 and we plot them , respectively , in figures [ ugoemergent ] , [ umhrdeemergent ] [ uhigheremergent ] . + for the to the hde model with go cut - off , three different cases have been considered , namely @xmath217 ( plotted in red ) , @xmath218 ( plotted in green ) and @xmath219 ( plotted in blue ) , while the other parameters involved have been chosen as @xmath220 , @xmath221 , @xmath222 , @xmath185 and @xmath223 .
we can clearly observe that @xmath178 has a decreasing behavior for all the three cases considered .
+ for the case corresponding to the mhrde model , we consider three different cases , @xmath224 ( plotted in red ) , @xmath225 ( plotted in green ) and @xmath226 ( plotted in blue ) , while the other parameters involved have been chosen as @xmath220 , @xmath221 , @xmath222 , @xmath185 and @xmath223 .
similarly to @xmath178 , @xmath179 has a decreasing behavior for all the three cases considered .
+ for the model proportional to higher time derivatives of the hubble parameter @xmath1 , we considered three different models , one with @xmath227 ( plotted in red ) , one with @xmath228 ( plotted in green ) and one with @xmath229 ( plotted in blue ) .
the other parameters have been chosen as follows : @xmath230 , @xmath231 , @xmath232 , @xmath222 , @xmath233 , @xmath234 and @xmath186 .
we can observe in figure [ uhigheremergent ] that @xmath176 has a decreasing behavior for all the cases considered .
+ therefore , we conclude that we find a decreasing behavior for all the three de models considered for all the range of values we considered
. + for scale factor emergent scenario .
the decreasing pattern indicates that the non - trivial contribution of de decreases with the evolution of the universe.,width=302 ] for scale factor emergent scenario .
the decreasing pattern indicates that the non - trivial contribution of de decreases with the evolution of the universe.,width=302 ] for scale factor emergent scenario .
the decreasing pattern indicates that the non - trivial contribution of de decreases with the evolution of the universe.,width=302 ] for scale factor emergent scenario .
@xmath196 has a decreasing behavior for all the three cases considered.,width=302 ] for scale factor emergent scenario .
@xmath200 can go beyond the phantom phase of the universe in all cases .
, width=302 ] for scale factor emergent scenario .
@xmath201 has a decreasing behavior for all the three cases considered .
only the case with @xmath229 ( which is plotted in blue ) is able to cross @xmath198 , instead for the other two models we always have @xmath235.,width=302 ] using the expression of @xmath178 , @xmath179 and @xmath176 obtained , respectively , from eqs .
( [ ugoeme ] ) , ( [ umheme ] ) and ( [ uhighereme ] ) and plotted in figures [ ugoemergent ] , [ umhrdeemergent ] and [ uhigheremergent ] , we can also plot the eos parameters for the three de models considered in this paper derived in eqs .
( [ eosgo1 ] ) , ( [ eosmh1 ] ) and ( [ eoshigher1 ] ) .
the eos parameter of the de model with go cut - off @xmath196 has a decreasing behavior , staying always in the region corresponding to @xmath236 .
moreover , @xmath196 assumes a constant value of @xmath237 $ ] ( according to the values of the parameters considered ) for @xmath238 .
+ studying the plot of @xmath200 , we observe an increasing behavior of the eos parameter for all the three cases considered .
therefore , we have that @xmath200 can go beyond the phantom phase of the universe in all cases we considered .
+ for the case pertaining to the de model proportional to @xmath108 and to higher time derivatives of the hubble parameter @xmath1 , we observe that all the cases considered have a decreasing behavior .
moreover , we observe that only the case with @xmath229 and plotted in blue can cross the line @xmath198 , while the other two models always stay in the region @xmath235 . +
as done for the power law scale factor studied in the previous section , we now consider some particular values of the parameters involved .
for the de model with go cut - off , we study the case corresponding to the ricci scale , which is recovered for @xmath204 and @xmath130 ( plotted in green ) , and we also consider the case corresponding to @xmath205 and @xmath206 ( plotted in red ) . instead ,
for the mhrde model , we consider the case corresponding to the ricci scale , which is recovered for @xmath129 and @xmath130 .
the values of all the other parameters are taken as the previous cases considered . for scale factor
emergent scenario against the time @xmath56 for the limiting cases of @xmath204 and @xmath130 ( plotted in green ) and @xmath205 and @xmath206 ( plotted in red ) .
, width=302 ] for scale factor emergent scenario against the time @xmath56 for the limiting case of @xmath129 and @xmath130 . , width=302 ] for scale factor emergent scenario against the time @xmath56 for the limiting cases of @xmath204 and @xmath130 ( plotted in green ) and @xmath205 and @xmath206 ( plotted in red ) . , width=302 ] for scale factor emergent scenario against the time @xmath56 for the limiting case of @xmath129 and @xmath130 .
, width=302 ] we can clearly observe in figure [ ugoemergentlim ] that @xmath178 has a decreasing behavior for both limiting cases considered . moreover , for the case with @xmath205 and @xmath206
, @xmath178 starts to assume a constant value for @xmath239 .
instead , the eos parameter @xmath196 has an initial increasing behavior for both case considered , becoming constant for @xmath240 .
moreover , for the case corresponding to the ricci scale , we have @xmath196 staying always greater than -1 while for the case with @xmath205 and @xmath206 , @xmath196 is always lower than -1 .
+ for the mhrde model , we observe that @xmath241 has a decreasing behavior while the eos parameter @xmath200 start with an increasing behavior and it becomes constant from @xmath242 , being always greater than -1 .
in this work , we have investigated and studied the effects which are produced by the interaction between a brane universe and the bulk in which the universe is embedded .
we have assumed that the adiabatic equation for the dm is satisfied , while it is violated for the de due to the energy exchange between the brane and the bulk .
taking into account the effects of the interaction between a brane universe and the bulk , we have obtained the eos parameter for the interacting hde model with granda - oliveros cut - off , the modified holographic ricci de ( mhrde ) model and the de model proportional to @xmath108 and to higher time derivatives of the hubble parameter having their energy densities given by @xmath243 , @xmath244 and @xmath245 , respectively .
moreover , we must underline that we are considering a flat universe , then @xmath17 .
+ we have considered two choices of scale factor , namely , the power - law and the emergent ones . the rate of interaction has been taken as @xmath246 .
we observed that , for the model pertaining to the power law scale factor , the parameter @xmath38 has an increasing pattern for all the three de energy density models considered while , for the scale factor pertaining to the emergent case , the parameter @xmath38 has a decreasing pattern for all the three de energy density models considered .
these observation are valid for all the values of the parameters considered .
+ we have also studied the behavior of the eos parameter using the reconstructed parameter @xmath38 .
we first considered the model with power law scale factor . for the de model with go cut - off
, we obtained that , for the case corresponding to @xmath180 , @xmath196 starts being @xmath197 , while with the passing of the time , it asymptotically reaches the point @xmath199 and can eventually cross it . for the other two cases considered
, we obtained that @xmath196 has a decreasing behavior , being always lower that @xmath199 .
instead , for the mhrde model , we obtained that @xmath200 has a slowly decreasing behavior for all the three cases considered , staying always greater than @xmath199 .
moreover , for the model proportional to higher derivatives of the hubble parameter @xmath1 , we obtain that @xmath201 has a decreasing behavior for all the cases considered , staying always in the region @xmath236 .
+ considering the case corresponding to the emergent scale factor , we obtained that @xmath196 has a decreasing behavior , staying always in the region @xmath236 .
furthermore , @xmath196 assumes a constant value in the region @xmath237 $ ] for @xmath238 . studying the plot of @xmath200 , we observed an increasing behavior of @xmath200 for all the three cases considered .
moreover , we have that @xmath200 can go beyond the phantom phase of the universe in all cases . for the model proportional to higher derivatives of the hubble parameter @xmath1
, we obtain a decreasing behavior for @xmath201 for all the cases considered .
furthermore , we have that only the case with @xmath229 ( which is plotted in blue ) can cross the phantom divide line corresponding to @xmath198 , instead the other two models always stay in the region @xmath235 .
+ we also considered the limiting cases corresponding to the ricci scale for the interacting hde model with granda - oliveros cut - off and for the modified holographic ricci de ( mhrde ) and also the interacting hde model with granda - oliveros cut - off for some particular values of the parameters @xmath125 and @xmath126 ( i.e. @xmath205 and @xmath206 ) derived in a recent work of wang @xmath35 xu . + for the case corresponding to the power law scale factor , we obtained that , for both limiting cases considered , @xmath178 has a decreasing behavior while @xmath196 has a slowly increasing behavior .
moreover , we have that for the case corresponding to the ricci scale , @xmath196 stays always greater than the value -1 , while for the case with @xmath205 and @xmath206 it is always lower than -1 . for the limiting case of the mhrde
, we observe that @xmath179 has an increasing behavior while @xmath200 slowly decreases , being always greater than the value of -1 .
+ for the case corresponding to the emergent scale factor , we can obtained that @xmath178 has a decreasing behavior for both limiting cases considered . moreover , for the case with @xmath205 and @xmath206 , @xmath178 starts to assume a constant value for @xmath239 .
instead , we have that the eos parameter @xmath196 has an initial increasing behavior for both limiting cases considered , becoming constant for @xmath240 . moreover , for the case corresponding to the ricci scale ,
we have @xmath196 staying always greater than -1 while for the case with @xmath205 and @xmath206 , @xmath196 is always lower than -1 .
+ for the mhrde model , we observe that @xmath241 has a decreasing behavior while the eos parameter @xmath200 start with an increasing behavior and it becomes constant from @xmath242 , being always greater than -1 .
sc acknowledges financial support from dst , govt of india under project grant no . sr / ftp / ps-167/2011 . 99 p. de
bernardis _ et al .
_ , nature * 404 * , 955 ( 2000 ) .
s. perlmutter _
_ , astrophys
. j. * 517 * , 565 ( 1999 ) .
a. g. riess _
et al . _ ,
j. * 116 * , 1009 ( 1998 ) .
p. j e. peebles , b. ratra , rev .
phys . * 75 * , 559 ( 2003 ) .
v. sahni , class .
* 19 * , 3435 ( 2002 ) . e. j. copeland , m. sami , m. tsujikawa , int . j. mod
d * 15 * , 1753 ( 2006 ) .
t. clifton , p. g. ferreira , a. padilla , c. skordis , phys .
rep . * 513 * , 1 ( 2012 ) .
s. del campo , r. herrera , d. pavon , j. cosmol .
phys . * 0901 * , 020 ( 2009 ) .
g. leon , e. n. saridakis , phys .
b * 693 * , 1 ( 2010 ) .
k. bamba , s. capozziello , s. i. nojiri , s. d. odintsov , astrophysics and space science * 342 * , 155 ( 2012 ) .
h. v. peiris _
_ , astrophys
. ser . * 148 * , 213 ( 2003 ) .
m. li , physics letters b * 603 * , 1 ( 2004 ) .
y. s. myung , physics letters b * 649 * , 247 ( 2007 ) .
y. s. myung , m. g. seo , physics letters b * 671 * , 435 ( 2009 ) .
q. g. huang , m. li , j. cosmol .
astropart . phys .
* 8 * , 13 ( 2004 ) .
g. t hooft , international journal of modern physics d * 15 * , 1587 ( 2006 )
. l. susskind , journal of mathematical physics * 36 * 6377 ( 1995 ) .
d. bigatti , l. susskind , strings , branes and gravity tasi * 99 * , 883 ( 2001 ) .
w. fischler , l. susskind , arxiv : hep - th/9806039 ( 1998 ) .
m. li , x. d. li , y. z. ma , x. zhang , z. zhang , j. cosmol .
* 9 * , 21 ( 2013 ) .
l. xu , phys .
d * 87 * , 043525 ( 2013 ) .
i. duran , l. parisi , phys .
d * 85 * , 123538 ( 2012 ) .
m. li , x. -d .
li , s. wang , y. wang , x. zhang , j. cosmol .
. phys . * 0912 * , 014 ( 2009 ) .
m. li , x. -d .
li , s. wang , x. zhang , j. cosmol .
. phys . * 0906 * , 036 ( 2009 ) .
b. chen , m. li , y. wang , nuclear physics b * 774 * , 256 ( 2007 ) .
m. jamil , e. n. saridakis , m. r. setare , physics letters b * 679 * , 172 ( 2009 ) .
s. nojiri , s. d. odintsov , 2008 , arxiv:0807.0685 s. nojiri , s. d. odintsov , int . j. geom .
phys . * 4 * , 115 ( 2007 ) .
t. p. sotiriou , v. faraoni , rev .
phys . * 82 * , 451 ( 2010 ) .
s. i. nojiri , s. d. odintsov , phys .
rep . * 505 * , 59 ( 2011 ) . v. a. rubakov , m. e. shaposhnikovubakov , phys lett b * 125 * , 139 ( 1983 ) .
i. antoniadis , phys lett b * 246 * , 377 ( 1990 ) .
p. binetruy , c. deffayet , u. ellwanger , d. langlois , phys lett b * 477 * , 285 ( 2000 ) .
c. van de bruck , m. dorcab , c. j. a. p. martins , m. parryd , phys lett b * 495 * , 183 ( 2000 ) .
s. nojiri , s. d. odintsov , s. ogushi , int . j. mod .
a * 17 * , 4809 ( 2002 ) .
c. deffayet , phys .
b * 502 * , 199 ( 2001 ) .
m. jamil , k. karami , a. sheykhi , int . j. theor
* 50 * , 3069 ( 2011 ) .
e. n. saridakis , phys .
b * 660 * , 138 ( 2008 ) .
m. r. setare , phys lett b * 642 * , 421 ( 2006 ) .
a. sheykhi , int .
d * 19 * , 305 ( 2010 ) .
y. s. myung , j. y. kim , class .
quant . grav .
* 20 * , l169 ( 2003 ) .
k. y. kim , h. w. lee , y. s. myung , gen .
. grav . * 43 * , 1095 ( 2011 ) .
e. n. saridakis , j. cosmol .
. phys . * 04 * , 020 ( 2008 ) .
r. g. cai , b. hu , y. zhang , commun .
. phys . * 51 * , 954 ( 2009 ) .
r. maartens , k. koyama , living rev .
* 13 * , 5 ( 2010 ) .
r. da rocha , j. m. hoff da silva , phys .
d * 85 * , 046009 ( 2012 ) .
d. bazeia , j. m. hoff da silva , r. da rocha , phys .
d * 90 * , 047902 ( 2014 ) .
r. da rocha , a. piloyan , a. m. kuerten , c. h. coimbra - araujo , class .
. grav . * 30 * , 045014 ( 2013 ) .
m. r. setare , e. n. saridakis , j. cosmol .
. phys . * 03 * , 002 ( 2009 ) .
planck collaboration collaboration , p. ade et .
al . , [ arxiv:1303.5076 ] .
wmap collaboration collaboration , e. komatsu et .
al . , astrophys .
* 192 * , 18 ( 2011 ) .
l. amendola , d. tocchini - valentini , phys .
d * 64 * , 043509 ( 2001 ) .
a. sheykhi , m. jamil , phys .
b * 694 * , 284 ( 2011 ) .
w. zimdahl , d. pavon , gen .
. grav . * 35 * , 413 ( 2003 ) .
m. r. setare , m. jamil , j. cosmol .
. phys . * 02 * , 010 ( 2010 ) c. feng _ et al .
b * 665 * , 111 ( 2008 ) .
k. ichiki k _
et al . _ , j. cosmol .
* 06 * , 005 ( 2008 )
. z. h. zhu , phys .
d * 73 * , 043518 ( 2006 ) .
l. n. granda , a. oliveros , physics letters b * 671 * , 199 ( 2009 ) . l. n. granda , a. oliveros , physics letters b * 669 * , 275 ( 2008 ) .
y. wang , l. xu , phys .
d * 81 * , 083523 ( 2010 ) .
l. p. chimento , m. g. richarte , phys .
d * 84 * , 123507 ( 2011 ) .
t. k. mathew , j. suresh , d. divakaran , int .
d * 22 * , 1350056 ( 2013 ) .
l. p. chimento , m. forte , m. g. richarte , eur .
j. c * 73 * , 10052 ( 2013 ) .
s. chen , j. jing , physics letters b * 679 * , 144 ( 2009 ) .
m. r. setare , phys .
d * 18 * , 419 ( 2008 ) .
u. debnath , classical and quantum gravity * 25 * , 205019 ( 2008 ) .
u. debnath , s. chattopadhyay , m. jamil , journal of theoretical and applied physics * 7 * , 25 ( 2013 ) .
r. ghosh , s. chattopadhyay , u. debnath , int . j. theor
. phys . * 51 * , 589 ( 2012 ) .
b. c. paul , p. thakur , s. ghose , mon . not .
r. astron .
soc . * 407 * , 415 ( 2010 ) .
r. ghosh , a. pasqua , s. chattopadhyay , european physical journal plus * 128 * , 12 ( 2013 ) .
s. mukherjee , b. c. paul , n. k. dadhich , s. d. maharaj , a. beesham , classical and quantum gravity * 23 * , 6927 ( 2006 ) . s. del campo , r. herrera , p. labraa , j. cosmol .
. phys . * 11 * , 30 ( 2007 ) . | in this paper , we study the effects which are produced by the interaction between a brane universe and the bulk in which the universe is embedded . taking into account the effects produced by the interaction between a brane universe and the bulk , we derived the equation of state ( eos ) parameter @xmath0 for three different models of dark energy ( de ) , _ i.e. _ the holographic de ( hde ) model with infrared ( ir ) cut - off given by the granda - oliveros cut - off , the modified holographic ricci de ( mhrde ) model and a de model which is function of the hubble parameter @xmath1 squared and to higher derivatives of @xmath1 .
moreover , we have considered two different cases of scale factor ( namely , the power law and the emergent ones ) . a nontrivial contribution of the de
is observed to be different from the standard matter fields confined to the brane .
such contribution has a monotonically decreasing behavior upon the evolution of the universe for the emergent scenario of the scale factor , while monotonically increasing for the power - law form of the scale factor @xmath2 . |
in the last four years the ape group has been extensively studying lattice qcd in the quenched approximation .
several simulations have been done to study weak matrix elements such as @xmath4 , @xmath5 , @xmath6 and to study semileptonic decays @xcite-@xcite .
these simulations have allowed to study the dependence of the results on the spacing @xmath7 and to investigate finite size effects . here
we present results for light meson masses and decay constants and baryon spectroscopy .
the results have been obtained from eight sets of data at @xmath8 , @xmath9 and @xmath10 using either the wilson action or the `` improved '' sw - clover action , in order to reduce @xmath11 effects .
the parameters used in each simulation are listed in table [ tab : latparams ] .
the ape group has performed extensive comparisons on data extracted from smeared and non - smeared propagators and found no real improvement for lattices with a time extent of 64 at @xmath8 and @xmath12 @xcite .
note that the simulations have been performed at @xmath13 values of @xmath14 or larger .
we have used the value of meson masses to read off the lattice baryon masses from the planes ( @xmath15 , @xmath16 ) and ( @xmath17 , @xmath16 ) . using the same values of @xmath18 used for meson masses
we have obtained the results reported in table [ tab : baryonsgev ] . for baryons we find very good agreement with the old ape @xcite data , while we find slightly larger values when comparing with jlqcd @xcite and lanl @xcite . also for baryon masses
we can conclude that we do not see a dependence on @xmath7 and that we have an agreement with the experimental data of @xmath3 .
jlqcd collaboration ( s. aoki et al . ) , ( proc . suppl . ) 47 ( 1996 ) 354 .
ukqcd collaboration ( c.r .
allton et al . ) , phys .
d 49 ( 1994 ) 474 - 485 .
ape collaboration ( s. cabasino et al . ) , phys .
b 258(1991 ) 195 - 201 . | we present a study of @xmath0 ( quenched ) lattice configurations from the ape collaboration , for @xmath1 using both the wilson and the sw - clover fermion action .
we determine the light hadronic spectrum and meson decay constants .
we extract the inverse lattice spacing using data at the _ simulated values _ of the quark mass .
we find an agreement with the experimental data of @xmath2 for mesonic masses and @xmath3 for baryonic masses and pseudoscalar decay constants .
a larger deviation is present for the vector decay constants . |
recent observations of the @xmath0 puppis system suggest that the asymptotic giant branch ( agb ) , mira - like variable may be in the early stages of transitioning to a planetary nebula ( pn ) @xcite . at a distance of 64 @xmath1 ,
@xmath0 puppis is one of the nearest and brightest agb stars .
it seems to be orbited by a close binary companion , and thus represents a unique laboratory in which to test models of the late - stages of stellar evolution .
adaptive optics imaging has revealed the presence of a optically thick , circumstellar disk with wide - bipolar outflows @xcite .
its m5iii spectral type @xcite implies that @xmath0 puppis has an effective temperature @xmath2 .
photometric studies show that it is a variable star with an apparent magnitude varying from 2.60 to 6.00 with a stable period of 140.6 days @xcite , consistent with a typical period associated with thermal pulsation @xcite .
the derived radius of @xmath0 puppis is @xmath3 @xcite thus its luminosity is @xmath4 . placing @xmath0 puppis on a hertzsprung - russell diagram using a zams evolution model computed by @xcite implies that @xmath0 puppis is now an agb star with mass of @xmath5 and an age of @xmath6 .
@xmath0 puppis has a large lobe structure that extends more than @xmath7 to the northeast of the disk in l - band images @xcite , likely due to the interaction of an agb wind with the secondary star .
a recent result by @xcite supports this hypothesis , revealing evidence of a close - in secondary source at a projected separation of @xmath8 .
in addition , the existence of an optically - thick circumstellar dusty disk hints at the presence of a secondary . if the disk were stable and in approximate keplerian motion , it would have a high specific angular momentum . however , agb stars are slow rotators @xcite and the large difference of specific angular momentum implies that there should be some mechanism that can transfer angular momentum to the gas and dust
. a companion can transfer angular momentum to the gas and shape the outflows seen in the post - agb and pn phases @xcite .
we have carried out a hydrodynamic simulation with astrobear and have used the result to produce synthetic images from monte carlo radiative transfer with the code radmc-3d .
@xcite used a similar method .
we describe the hydrodynamic simulation in detail in sect .
[ hydromodel ] and present the results in sect .
[ hydroresults ] .
we discuss the model used in our radmc-3d simulation to generate synthetic observations in sect .
[ radmcmodel ] . in section [ rtresults ]
we compare our synthetic observations to the observational data .
we have performed three - dimensional hydrodynamic adaptive - mesh - refinement ( amr ) simulations of a @xmath9 agb star with a @xmath10 secondary at a separation of @xmath8 in the published version which was wrong ] .
the agb star is represented by a sphere of radius @xmath11 with a steady mass - loss rate of @xmath12 and wind velocity of @xmath13 . after the simulation reaches steady - state ,
we eject a pulse of dense wind at a rate of @xmath14 that lasts @xmath15 years .
the speed of the pulsed material is @xmath16 @xmath17 which is above the escape velocity of the agb star but below the escape velocity of binary system .
our chosen value for mass loss rate is comparable to normal agb winds but our initial wind velocity is higher .
we choose a high velocity at the initial launch radius because we do not include wind launching in our models .
this means the gravitational force of the agb star will only decelerate the wind as it expands outwards and we seek to keep @xmath18 in the @xmath19 range at radii of interest where the interaction with the binary occurs . a model for driving agb winds @xcite will be included in our future work .
the properties of such a pulse are consistent with those expected from the ingestion of a planet by the agb star .
such events are expected to occur on the giant branches if planetary companions are present and initially orbiting within @xmath20 of the main - sequence progenitor @xcite .
radiation from the agb star is assumed to be isotropic at all radii with the dust and gas fully coupled . under such conditions ,
the radiation force can be expressed as @xmath21 where @xmath22 is the mass weighted opacity of gas and dust .
the radiation force is proportional to the gravitational force @xmath23 in the optically thin limit @xcite .
the magnitude of the total inward force acting on the gas is therefore , @xmath24 we assume that @xmath25 in our model .
note that once the disk forms , radiation may not be isotropic at all radii outside of the agb star .
the formation of an optically thick region , such as a torus or disk , would block photons and the local interior regions of these structures would experience a reduced or absent direct radiation force .
optically thin regions , such as those in the polar direction ( and for the majority of our computational domain ) would experience the full radiation force . despite our use of eq .
[ eq:1 ] everywhere , a prominent disk clearly forms suggesting that any case with a position dependednt reduced radiation force ( less total outward force ) would also incur disk formation .
future work should incorporate a more accurate local variation in the radiation force .
assuming that the dust and gas are in local thermal equilibrium ( lte ) and that the luminosity from the agb star is constant , the gas temperature varies as @xmath26 .
at @xmath27 @xmath28 the surface temperature of the agb star is @xmath29 and is approximately @xmath30 at @xmath31 away . since our simulation zone is @xmath32 we employ an isothermal @xmath33 for our simulation .
we use isothermal condition for practical purpose and the actual temperature in the disk might be close to @xmath30 . we know that the temperature near the actual agb star is much higher than @xmath30 and that temperature gradient is an important condition for driving agb winds , but these effects are are only important in the launch region which we do not model .
as noted above , we leave an exploration of the effect detailed driving mechanism of agb winds and cooling for future research . to carry out our three - dimensional hydrodynamic simulations , we employ astrobear .
astrobear is a multi - physics , adaptive - mesh - refinement ( amr ) code which employs a riemann solver to solve the fluid equations @xcite . in our simulation
, we use @xmath34 computational cells as the base grid and 3 levels of amr such that the total effective resolution is @xmath35 .
the equations of motion of fluid in the simulation are @xmath36 @xmath37 @xmath38 where subscript @xmath39 represents the agb star and subscript @xmath40 represents the secondary .
the mean atomic weight of the gas material is @xmath41 @xmath42 .
we run the simulation for @xmath43 years .
approximately , @xmath44 years after the ejection of the dense wind , material falls back with a disk forming in the orbital plane .
meanwhile , low - density outflows form in the polar directions . in figure 1 and 2
we show results of the simulations in terms of cuts of density in , and perpendicular , to the equatorial plane ( fig 1 ) as well as 3-d iso - density contours ( fig 2 ) .
to measure whether material is gravitationally bound to the binary system , we construct the parameter @xmath45 where @xmath46 is the specific gravitational potential energy of the material . when @xmath47 the material is bound to the system .
delineate the bound and unbound regions.,title="fig:",width=321 ] delineate the bound and unbound regions.,title="fig:",width=321 ] , width=321 ] the density cuts in figure [ fig:1 ] include conturs of @xmath48 and they demonstrate the formation of both unbound bipolar outflows and a gravitationally bound disk structure . in the top panel ,
we see high - density gas orbiting the binary in a torus / disk configuration .
the radius of the disk is roughly @xmath49 and there is a spiral - arm in its outer regions which sweeps clockwise .
the @xmath50 and @xmath51 iso - contours demonstrate that most of the gas in mid - plane is gravitationally bound , although there is some escaping gas in the lower right corner of the top panel .
this is to be expected as gas from the mass - losing agb star continually interacts with the disk .
the disk should be sustible to rossby wave instabilities @xcite which may be the cause of the episodic ejection . in the bottom panel , the @xmath52 the contours
take the form of bipolar cones indicating that the gas in the polar direction is not bound .
note the presence of `` ripples '' in the two cones induced by the orbital motion of the binary system .
the velocity of the bipolar outflows is roughly @xmath53 .
note that the disk is seen in the bottom panel of fig .
[ fig:1 ] as the dense gravitationally - bound region inside the red contour .
we note that in the equatorial region , some of the outflowing gas will fall back onto the the mid - plane and incorporate into the disk while some will escape the system . figure [ fig : contour ] shows the 3d iso - density plot with the disk spin axis inclined by @xmath54 to the direction toward the observer .
the bipolar lobes are seen in the red and yellow iso - density contours in this image .
the green regions demarcate a iso - density contour of @xmath55 with includes the disk .
we employ the radmc-3d code @xcite to post - process our hydrodynamic results and create synthetic maps for comparison with observations . given a prescribed dust distribution , we use a monte carlo method @xcite to determine the local dust temperature , which is then used to produce images of photon density with asymmetric scattering .
the results of our hydrodynamic simulation serve as the input for our radiation transfer simulation .
we adopt an agb photospheric model from @xcite which is shown as the blue curve in figure [ fig : sed ] .
the agb model has an effective temperature of @xmath29 , @xmath56 and @xmath57 = 0.0 $ ] .
the raw sed is subject to reddening of @xmath58 with a standard milky way @xmath59 interstellar dust model and shown as the green curve in figure [ fig : sed ] @xcite .
observational studies of @xmath0 puppis show that there is an infrared excess in the sed . @xcite .
such infrared excesses in evolved stars are associated with dust shells or disks @xcite .
furthermore , the @xmath60 @xmath61 m feature suggests the dust is primarily silicate - based .
as such , we choose two kinds of amorphous silicates : olivine and @xmath62 pyroxene @xcite which are the most common silicates around agb stars .
we assume that both types of dust grains dust are spherical and with size- number distributions of @xmath63 @xcite , where @xmath64 is the dust radius and @xmath65 is the number density .
we consider dust radii in the range from @xmath66 .
we choose this range of radii because : 1 .
the flat feature in sed between @xmath39 @xmath67 to @xmath68 @xmath67 implies that the disk may have dust grains with radii around @xmath69 @xmath67 .
larger dust grains need more time to evolve from small dust grains and large grains may not obey the size distribution law described in @xcite .
we thus set the cutoff at @xmath70 @xmath67 .
dust can not form when the environment is too hot . if we assume the dust to be in approximate lte with the star s effective temperature , given the luminosity @xmath71 and the dust sublimation temperature @xmath72 , the dust is unlikely for form within @xmath73 .
however , there are likely high temperature shocks around the star created by the agb winds and the orbital motion of the secondary .
the high temperature shocks will moderately extend the dust free region .
as such , we set the dust - to - gas mass ratio @xmath74 within @xmath75 of the center - of - mass of the binary .
for the rest of the computational domain , we set @xmath76 , since small dust grains are almost fully coupled with the fluid @xcite .
we assume that @xmath77 of the dust mass is olivine and @xmath78 mass of the dust is pyroxene .
the total mass of the dust in the simulation is @xmath79 .
given our chosen dust size distribution , dust species and optical properties of amorphous silicates , the optical properties of the dust mixture can be computed from mie theory @xcite .
we use the matlab code @xcite to calculate the absorption and scattering opacities and the results are plotted out in figure [ fig : dust ] .
we also take asymmetric scattering by dust into consideration . in our monte carlo simulation
, we use the henyey - greenstein @xcite scattering phase function @xmath80 where @xmath81 is the wavelength dependent asymmetric coefficient .
the actual asymmetric coefficient for dust is close to @xmath39 in short wavelength regime .
when the asymmetric coefficient is close to @xmath39 , the scattering is strongly forward directed
. in reality , dust tend to agglomerate into larger grains when the temperature drops @xcite and larger grains are more weakly coupled to the fluid @xcite .
therefore , by assuming a fixed dust distribution , composition , and gas - to - dust ratio , our treatment may underestimate the dust contribution in dense regions such as the disk itself @xcite .
however , this simplifying approach for present purposes makes the problem tractable with current computational resources .
we leave the self - consistent , multi - fluid , three - dimensional , radiation - hydrodynamic modeling for future work .
figure [ fig : sed ] shows our results for the broad - band sed computed from our hydrodynamic model and compared with existing observational data ( details can be found in @xcite ) . in figure 5 we also present synthetic observations in terms of n and v band images of models and compare these with recent zimpol observations .
consideration of fig [ fig : sed ] and figure 5 shows we successfully reproduce the @xmath82 to @xmath83 flat features as well as the @xmath84 bump in the sed .
we note that the naco data , indicating the flat , lower intensity regions of the spectrum between @xmath82 to @xmath83 were taken in 2013 and represents higher quality data than the photometeric data associated with the pre-2004 points .
the astrobear simulation , with our choice of dust model , thus is able to capture the shape of the sed across 2 orders of magnitude in radius . in figure
[ fig : vnbandside ] , we have compared synthetic v - band and n - band radmc-3d images with the corresponding zimpol observations @xcite .
the zimpol observations have an angular resolution of @xmath85 , which corresponds to @xmath86 in @xmath0 puppis . to foster the best comparison
, we made the resolution in the synthetic image the same as that of the observation .
the images reveals an optically thick circumstellar torus in v and n band . and @xmath87 respectively.,title="fig:",width=154 ] and @xmath87 respectively.,title="fig:",width=154 ] and @xmath87 respectively.,title="fig:",width=283 ] and @xmath87 respectively.,title="fig:",width=283 ] from top .
the image on the right shows the synthetic image of @xmath88 also from top.,title="fig:",width=156 ] from top .
the image on the right shows the synthetic image of @xmath88 also from top.,title="fig:",width=156 ] the fit between our simulation and observations is good as the simulation synthetic images show both the bipolar lobes and dark band indicating the presence of the disk .
the broad opening angle of the lobes is also recovered in the simulation . finally in figure
[ fig : vtop ] , the @xmath89 synthetic image shows a brighter center while the @xmath88 synthetic image has a brighter ring . the @xmath88 photon is a signature of silicate dusts ,
thus the @xmath88 synthetic image shows where the dusts are .
the contour lines show that the inner boundary of the ring is about @xmath75 which is the no - dust radii we have assumed .
the relatively dim center implies that there are very few dusts in the polar direction .
it tells us that the optical depth in the polar direction is much lower compared to the equatorial plane - otherwise we would have seen a brighter center .
given that there are many more dusts outside the center , the brighter center in the @xmath89 image indicates that the dusts in the ring are colder - otherwise they will emit more short wavelength photons .
it also proves that the polar direction is optically thin .
in this paper , we present a fully 3-d hydrodynamic simulation that models a pulse of dense wind followed by constant stellar wind in a binary system .
our simulation self - consistently forms a circumbinary disk with wide - bipolar outflows for the @xmath0 puppis system parameters . using the output of hydrodynamic modeling to produce synthetic observations of broad - band photometric and imaging data
significantly improves upon previous non - dynamical , morphological studies of @xmath0 puppis . throughout our computational domain
, we tracked the fluid to determine whether it is gravitationally bound or escaping .
the bulk of the bound material is located in the circumstellar disk in the orbital plane . along the poles ,
a wide angle low velocity outflows emerges with measured velocities of @xmath90 .
synthetic observations constructed by post - processing our hydrodynamic simulation with the radiation transfer code radmc-3d shows strong morphological similarity to the v - band and n - band sphere observations .
the broad - band sed computed from our results also matches the observations well , reproducing the `` flat '' range from @xmath82 to @xmath83 and the infrared excess seen long - ward of @xmath91 m .
the top - down synthetic images show that the polar direction is optically thin and the dust in the disk has low temperature .
this study is an initial step in fully dynamical three - dimensional simulations of evolved binary star systems and their photometric and imaging observations .
future work should include additional physical processes such as cooling , self - consistent radiative transfer and multi - dust species .
we gratefully acknowledge the help from dr .
eric lagadec for the useful discussion of the recent observations of @xmath0 puppis that led to this paper .
we also thank dr .
pierre kervella and dr . miguel montarges very much for sharing us the useful data and images .
our thanks should also be given to dr .
baowei liu and dr .
jonathan carroll - nellenback for helping us in developing the code .
this work used the computational and visualization resources in the center for integrated research computing ( circ ) at the university of rochester .
part of the work also used the extreme science and engineering discovery environment ( xsede ) , which is supported by national science foundation grant oci-1053575 . financial support for this project
was provided by the department of energy grant gr523126 , the national science foundation grant gr506177 , and the space telescope science institute grant gr528562 .
99 bedding t. r. , et al .
, 2002 , mnras , 337 , 79 bertelli g. , girardi l. , marigo p. , nasi e. , 2008 , a&a , 484 , 815 bjorkman j. e. , wood k. , 2001 , apj , 554 , 615 bowen g. h. , 1988 , apj , 329 , 299 carroll - nellenback j. j. , shroyer b. , frank a. , ding c. , 2013 , jcoph , 236 , 461 castelli f. , kurucz r. l. 2004 , arxiv e - prints [ arxiv : astro - ph/0405087 ] chen z. , frank a. , blackman e. g. , nordhaus , j. , 2016 , mnras , 457 , 3219 dorschner j. , begemann b. , henning t. , jaeger c. , mutschke , h. , 1995 , a&a , 300 , 503 dullemond c. p. , 2012
, radmc-3d : a multi - purpose radiative transfer tool , astrophysics source code library dumm t. , schild , h. , 1998 , new a. , 3 , 137 feuchtinger m. u. , dorfi e. a. , hofner s. , 1993 , a&a , 273 , 513 fitzpatrick e. l. , 1999 , pasp , 111 , 63 gail h .-
p . , wetzel s. , pucci a. , tamanai a. 2013 , a&a , 555 , a119 hendrix t. , keppens r. , camps , p. , 2015 ,
a&a , 575 , a110 henyey l. g. , greenstein j. l. , 1941 , apj , 93 , 70 jaeger c. , mutschke h. , begemann b. , dorschner j. , henning , t. , 1994 , a&a , 292 , 641 kervella p. , et al . , 2014 , a&a , 564 , 88 kervella p. , et al . , 2015 , a&a , 578 , 77 kervella p. , montarges m. , lagadec e. , 2015 , eas publications series , 71 , 211 krumholz m. r. , klein r. i. , mckee c. f. , offner s. s. r. , cunningham a. j. , 2009 , science , 323 , 754 matzler c. , 2002 , matlab functions for mie scattering and absorption . univ . of bern ,
bern , switzerland mastrodemos n. , morris m. , 1998 , apj , 497 , 303 mathis j. s. , rumpl w. , nordsieck k. , 1977 , apj , 217 , 425 meibom s. , mathieu r.d .
, stassun k.g . , 2009 , apj , 695 , 679 meheut h. , yu c. , lai d. , 2012 , mnras , 422 , 2399 nordhaus , j. , blackman , e. g. 2006 , mnras , 370 , 2004 nordhaus j. , blackman e. g. , frank , a. , 2007 , mnras , 376 , 599 nordhaus j. , et al . 2008 , mnras , 388 , 716 nordhaus j. , spiegel d. s. , ibgui l. , goodman j. , burrows a. , 2010 , mnras , 408 , 631 nordhaus j. , spiegel d. s. , 2013 , mnras , 432 , 500 samus n. n. , et al . , 2009 , vizier online data catalog , 1 , 2025 soker n. , rappaport s. , 2000 , apj , 538 , 241 van de hulst h. c. , 1957 , light scattering by small particles .
john wiley & sons , inc .
, new york , ny van marle a. j. , meliani z. , keppens r. , decin l. , 2011 , apjl , 734 , l26 woitke p. , 2006 ,
a&a , 452 , 537 | recent observations of the @xmath0 puppis system suggest that this mira - like variable may be in the early stages of forming a bipolar planetary nebula ( pn ) .
as one of nearest and brightest agb stars , thought be a binary , @xmath0 puppis serves as a benchmark object for studying the late - stages of stellar evolution .
we perform global , three - dimensional , adaptive - mesh - refinement hydrodynamic simulations of the @xmath0 puppis system with astrobear .
we use the radiative transfer code radmc-3d to construct the broad - band spectral - energy - distribution ( sed ) and synthetic observational images from our simulations . given the reported binary parameters , we are able to reproduce the current observational data if a short pulse of dense material is released from the agb star with a velocity sufficient to escape the primary but not the binary
. such a situation could result from a thermal pulse , be induced by a periastron passage of the secondary , or could be launched if the primary ingests a planet .
[ firstpage ] method : numerical stars : agb and post agb ism : structure stars : winds , outflows |
during the last decade , the development of non - volatile electronic memories based on the resistive switching ( rs ) effect in transition metal oxides made a huge progress , becoming one of the promising candidates to substitute the standard technologies in a near future .
rs refers to the reversible change of the resistance of a nanometer - sized media by the application of electrical pulses @xcite .
though immediately after the discovery of the reversible rs effect in transition metal oxides its application for multi - level cell memories ( mlc ) was envisioned @xcite , reports so far chiefly focused on single - level cell ( slc ) devices @xcite . unlike slc , which can only store one - bit per cell , mlc memories may store multiple bits in a single cell @xcite . mlcs of up to 4 bits ( 16 levels ) are currently available based in both , standard flash @xcite and phase - change @xcite technologies . in an @xmath0-bit rs - based mlc memory , one has to encode the @xmath1 memory levels as distinct resistance states .
thus , the @xmath1 memory states can be identified with each one of the consecutive and adjacent resistance bins of width @xmath2 , with @xmath3 . to store the @xmath4 memory state
the device has to be set to the corresponding bin , ie @xmath5 .
the requirement of non - volatility implies that the value should remain within that bin , even after the input is disconnected or the memory state is read out . in fig .
[ fig : zoom ] we plot data of a random sequence of stored memory states in our implementation of a 64-level ( 6-bit ) mlc . for practical convenience resistance levels
are translated into voltage levels by a fixed bias current .
the central component of the mlc is a rs device of a resistance @xmath6 , whose magnitude can be changed by application of a current pulse @xmath7 .
the resistance change depends on @xmath7 through a highly non - trivial function @xmath8 ; @xmath9 the function @xmath8 is usually unknown , however , an important general requirement for non - volatile memory applications using rs is that @xmath10 for currents below a given threshold .
this allows to sense ( ie , read - out ) a stored memory state , ie the so called remnant resistance @xmath11 , injecting a bias current @xmath12 without modifying the stored information .
thus , @xmath13 , @xmath12 .
a two level memory can be simply implemented by pulsing a rs - device strongly with opposite polarities and sensing the corresponding high and low @xmath6 states with a weak bias current .
however , the previous observation of a current threshold has an important consequence for the implementation of mlc memories .
in fact , in order to implement an mlc one should be able to tune the value of @xmath11 to any desired value .
the better the control on this tuning , the better the ability to define the bins and a larger number of bits could be coded in the mlc . in principle , a perfect knowledge of the function @xmath8 should allow for this , but in practice the function is not known .
yet , any reasonable form of @xmath8 would allow to tune any arbitrary memory state by applying a sequence of pulses of decreasing intensity , following a simple zero - finding algorithm as in standard control circuit .
nevertheless , the requirement of a threshold current difficult the fine tuning , as small corrections beneath @xmath14 have no effect . in practice , the dead - zone introduced by relatively large values of @xmath14
prevents a straightforward implementation of rs - device as mlc memories with a large number of levels .
below we shall demonstrate how this conceptual problem can be overcome , and exhibit the implementation of a 64-level mlc .
, and a write state where current pulses are applied in order to cancel the difference between @xmath15 and @xmath16
. when wr is deactivated , @xmath7 is force to zero . ]
we adopt a manganite - based rs - device , made by depositing silver contacts on a sintered pellet of la@xmath17pr@xmath18ca@xmath19mno@xmath20 ( lpcmo ) @xcite .
a rs - device is defined between the pulsed ag / lpcmo and a second non - pulsed contact . a third electrode ( earth )
is required in a minimal 3-contact configuration setup .
a requirement for our mlc is that the rs - device operates in bipolar rs mode @xcite , ie , depending on the pulse polarity the remnant resistance either increases or decreases .
it is now well established that the mechanism behind bipolar rs is the redistribution of oxygen vacancies , within the nanometer - scale region of the sample in contact with the electrodes @xcite . in the case of lpcmo
, the oxygen vacancies significantly increase its resistivity because the electrical transport relies on the double - exchange mechanism mediated by the oxygen atoms @xcite .
pulsing an electrical current trough the contact will produce a large electric field at the interface due to the high resistivity of the ag / lpcmo schottky barrier .
if the pulse is strong enough , it will enable the migration oxygen ions across the barrier , modifying the concentration of vacancies and , hence , changing the interface resistance .
the ionic migration remains always near the interface and does not penetrate deep into the bulk , since the much larger conductivity there prevents the development of high electric fields .
thus , the rs effect remains confined to a nanometer - sized region near the interface , as schematically depicted in fig .
[ phys ] .
we now introduce a practical implementation of an rs - based mlc that produced the results shown in fig .
[ fig : zoom ] .
we used an ag / lpcmo interface and off - the - shelf electronic components .
the block diagram of the concept is presented in fig .
[ f4 ] and the schematics and technical details are included in the appendix .
we envisage an implementation of an rs - based mlc memory chip where the storage core is a set of many rs - units with a single common control circuit .
the common control would set each of the individual rs - units , one at a time , thus resembling the concept of the refresh logic circuit in the drams . in this work
we demonstrate the implementation of the control circuit with a single rs memory unit .
key to our mlc implementation is to adopt a discrete - time algorithm that overcomes the problems discussed above @xcite , and which we describe next .
the required memory state @xmath21 is coded in terms of a @xmath15 , while @xmath16 indicates the actual stored value ( fig .
the system iteratively applies pulses @xmath7 of a strength that is an estimate of the required value to set the target state @xmath15 , eventually converging to it .
this discrete - time feedback loop continuously cycles between 2 stages ; probe and correct . in probe stage
the switch connects the rs - device to the current source @xmath22 in order to sense the remnant resistance . in the correct stage
, the switch connects the rs - device to a pulse generator that applies a corrective pulse @xmath7 , of a strength that is obtained from the difference between the delayed @xmath16 and the target value @xmath15 as follows , @xmath23=k_{p}\,e\left[k\right]+k_{i}\,\sum_{i=0}^{k}e\left[i\right],\label{eq : pi}\ ] ] where the error signal @xmath24=v_{out}\left[k-1\right]-v_{in}\left[k\right]$ ] is the change required in the output voltage and @xmath25:=i_{pulse}\left(k\ , f_{{\rm clk}}^{-1}\right)$ ] , being @xmath26 the frequency of the system clock and @xmath27 an integer . @xmath28 and @xmath29 are generic proportional and integral constants respectively with @xmath30 unit , being @xmath31 and @xmath32 the electric current and voltage units .
the first term of the equation represents a proportional estimator . the second term prevents the system to get stuck in a condition where @xmath33 .
in fact , for low @xmath24 $ ] a pulse of strength @xmath34 $ ] would lay below the threshold , thus not producing any further change in the state of the system .
the magnitude of the second term linearly increases in time , thus making @xmath7 to eventually overcome the threshold and correct the output voltage in the desired direction . notice that even if this approach resembles a standard proportional - integral ( pi ) control loop with dead - zone , there are substantial differences .
first , the remnant resistance reading and the correcting pulse application occur at different times .
this requires the addition of the continuously commuting switch .
second , we also needed the introduction of a delay , implemented as a sample and hold circuit , as is required for the feedback path . these differences and
the strong non - linearity in @xmath8 makes the stability analysis of this approach ( which also depends on specific values of @xmath28 and @xmath29 ) a significant issue that we describe below .
the study of the system stability is based on the discrete - time model presented in fig .
[ fig : concept ] . in the z - domain ,
[ eq : pi ] becomes @xmath35 is compared with the output signal @xmath36 after being delayed , generating the error signal @xmath37 .
based on this signal , the estimator generates the corrective pulses @xmath38 . ] the resulting @xmath11 after these corrective pulses is probed by connecting the rs - device to the bias current source ( @xmath22 ) , obtaining the output signal @xmath39$](@xmath36 in z - domain ) . in the next cycle , it will be compared to the reference input @xmath40 $ ] generating the error signal @xmath41 $ ] .
this fact implies the 1-cycle delay @xmath42 .
central to the present proposal is the following equation , a discrete - time model for the rs - device : @xmath43=r_{0}+r_{1}\,\sum_{j=-\infty}^{k}\mathrm{nl}\left(i_{pulse}\left[j\right]\right),\label{eq : rrem}\ ] ] where @xmath44 and @xmath45 , having resistance units , are an offset and a proportionality factor respectively .
@xmath46 is a non - linear function that has to be defined on the basis of the general behavior of a rs - device operating in bipolar mode . in this model
the @xmath11 is calculated by integrating the writing pulses after been weighted by the @xmath46 function . in this way
, we model the possible change of the device resistance after the @xmath27-th pulse as @xmath47=\mathcal{r}_{rem}\left[k\right]-\mathcal{r}_{rem}\left[k-1\right]= r_{1}\mathrm{nl}\left(i_{pulse}\left[k\right]\right)$ ] .
a concrete implementation of nl is presented below @xmath48=\begin{cases } \left(i_{pulse}\left[k\right]+i_{th}\right)\frac1a & i_{pulse}\left[k\right]<-i_{th}\\ 0 & -i_{th}<i_{pulse}\left[k\right]<i_{th}\\ \left(i_{pulse}\left[k\right]-i_{th}\right)\,u_{1 } & i_{pulse}\left[k\right]>i_{th}\ , .
\end{cases}\label{eq : nl}\ ] ] the unit of @xmath49 is @xmath50.when a negative signal exceeds the threshold @xmath25<-i_{th}$ ] , @xmath11 decreases as a linear function of @xmath25 $ ] with slope 1@xmath51 . for a positive signal greater than @xmath52 , @xmath11 increases with slope @xmath53 . for the sake of stability analysis
[ eq : rrem ] is simplified as : @xmath54=v\sum_{j=-\infty}^{k}\mathrm{nl}\left(i_{pulse}\left[j\right]\right).\label{eq : c k - domain}\ ] ] in this way , the analysis is not considering any instability when sensing @xmath11 with @xmath55 .
indeed , our actual implementation did not present any critical issue at this level ( see the appendix ) . from now on ,
unit - step sequence is considered as input signal @xcite .
the steady - state error for the system with @xmath56 ( just proportional control ) and @xmath57 is @xmath58=\frac{i_{th}}{k_{p}}$ ] , then the steady - state response is @xmath59 . in fact , fig .
[ fig : pi](a ) shows that the system converges to the required set - point only for @xmath60 .
the system arrives to this condition because when @xmath61\right|\leq\frac{i_{th}}{k_{p}}$ ] the excitation of the rs - device is @xmath62\right|\leq i_{th}$ ] , i.e. lower than the minimum voltage required to produce a change in @xmath63 .
a proportional control that enters into this condition , remains there indefinitely .
the integral term in eq .
[ eq : pi k - domain ] , that turns the system into a pi control , avoids the problem ; when continuously integrated , @xmath24 $ ] makes @xmath25 $ ] to eventually overcome the threshold @xmath64 .
[ fig : pi](b ) and ( c ) .
we begin by considering the simplified situation where the stability is analyzed by removing the nonlinearities introduced by @xmath65 $ ] ( @xmath57 and @xmath60 ) .
the transfer function of the system is @xmath66 . for @xmath67 ( typical value )
the system is critically damped when @xmath68 , and remains stable for @xmath69 .
increasing @xmath29 moves the location of the poles following @xmath70 , arriving to @xmath71 when @xmath72 , where the system becomes unstable for any @xmath28 ( see fig .
5 ( m ) ) .
although the introduction of nonlinearities in eq .
[ eq : nl ] does not allow to study the system analytically , as in the previous section , its response might be numerically simulated .
in fact , simulations where performed in the @xmath27-space by solving eq .
5 , after substituting @xmath25 $ ] ( eq . 1 ) and @xmath65 $ ] ( eq .
4 ) . for the computation of @xmath24 $ ]
, we assigned @xmath73=c\left[k\right]$ ] and the unitary step function at the input ( ie .
@xmath74=1v$ ] ) . fig .
[ fig : pi ] ( d - f ) shows this simulated response of the system for three sets @xmath75 corresponding to representative position of the poles of the equivalent system without nonlinearities . fig . [
fig : pi ] ( g - i ) shows the effect of introducing the threshold ( @xmath76 ) for the same set @xmath75 .
the system response is qualitatively the same with the addition of periods where @xmath39 $ ] is `` frozen '' because @xmath77\right|\leq i_{th}$ ] , while the integral term is growing .
[ fig : pi ] ( j - l ) shows the effect of further introducing asymmetry into the nonlinear function @xmath78 . the effect that this nonlinearity introduces when @xmath79 is equivalent to reducing the gain of the system for @xmath25>0 $ ] ( see eq .
[ eq : nl ] ) .
in fact , fig . [
fig : pi](j ) evidences a clear difference between the system speed when @xmath39 $ ] is increasing as compared to the speed when @xmath39 $ ] is decreasing .
the case for @xmath80 is equivalent to a case where @xmath81 and @xmath82 .
simulations also show a change in the stability limit as reported in the following table : [ cols="^,^,^",options="header " , ] summarizing , the simulations show that the system stability is not compromised after the introduction of nonlinearities , even if in some conditions , the system might require a considerably higher number of cycles to stabilize .
two tests were done to evaluate the performance of the mlc memory .
experimental results for the memory retention test are presented in fig .
[ results ] .
we emphasize that the relatively slow operation speed of our memory mlc would be dramatically improved upon device miniaturization and integration . for simplicity , we sampled a random subset of 16 out of 64 different voltages uniformly distributed along the operative range ( fig .
[ results](a ) ) .
each write and read cycle had a duration of 13 seconds .
in the first 5 s the wr signal was active and a memory value was set . after a memory value was set , we observed a small relaxation of @xmath16 with time constant of @xmath831.6 s. the value of @xmath16 remained essentially stable afterwards .
thus , for the sake of performing a large number measurements , we only monitored the memory retentivity during the 5 time - constants ( ie , 8 seconds ) that immediately followed the set . within that period ,
the wr signal was inactive , and the memory state stored in @xmath16 was probed every 0.1 s ( ie , the memory was read out 80 times ) .
continuously sampled at 10 reading per second . during the readings indicated with the red bars the signal wr was active ;
the output value accurately follows the sequence of input values that is randomly chosen on the grid . during the readings indicated with blue bars ,
no pulses were applied to the rs - device , resulting in a drift .
( b ) histogram of the last read value of the memory in each cycle ( ie , last value of blue sections ) .
the horizontal bars denote the @xmath84 intervals corresponding to a 16 , 32 and 64 level devices .
the location of the bars are chosen so to minimize the error probability ( ie , area of histogram outside the interval ) . ] in fig .
[ results](b ) we present the histogram of the distribution of observed errors in the reading of a given stored memory value .
the observed values correspond to the remnant resistance of the memory cell , after reading it 80 times at a 10 hz rate , following the set .
the histogram is constructed from a sequence of 460 memory state recordings .
the finite dispersion of the histogram is the main limitation for implementing a high number of memory levels . to easily visualize the retentivity performance of the memory for increasing number of levels , we indicate with horizontal bars the resistance interval @xmath85 ( = @xmath86 ) that corresponds to having a 16 , 32 and 64 mlc memory ( ie , 4 , 5 and 6 bits , respectively ) .
the probability of error in storing a memory state corresponds to the area of the histogram that lays outside the resistance interval normalized to the total area . with a confidence level of 95% , the probability that the mlc fails to retain a stored state is [email protected] for a 64-level system ( 6 bit mlc ) .
this error probability rapidly improves as the number of levels is decreased , being @xmath88 0.012 for a 16-level mlc .
the previous data on the performance of the memory can also be expressed in terms of the bit error rates ( ber ) . in our mlc , we find [email protected] for the 16-level system , [email protected] for the 32-level system and [email protected] for the 64-level system .
lower error probabilities for bits than for levels , reflects the fact that , typically , the error corresponds to the memory level drifting to an immediate neighbor level ( that often shares a large number of bits ) .
in conclusion , we have introduced a feedback algorithm to precisely set the remnant resistance of a rs - device to an arbitrary desired value within the device working range .
this overcomes the conceptual problem of fine tuning a resistance value in non - volatile rs - devices for mlc applications , due the presence of a threshold current behavior .
the feedback configuration is intrinsically time - discrete , since it is based on read / write sequences
. the overhead in the writing time introduced in this feedback system may be a limitation for its utilization as primary memory in a computer system , but it can easily compete with actual speed of the current mass storage devices ( flash memories ) .
the applicability of the concept to implement an n - bit mlc memory was successfully demonstrated for @xmath0 = 4 , 5 and 6 , with an lpcmo - based circuit ; which clearly illustrates the critical trade - off between ber , number of memory levels and ( power and area ) overhead of the control circuitry .
[ fig : simplified - circuit ] shows a simplified circuit of our implementation .
the estimator circuit ( see fig . [ f4 ] and fig .
[ fig : concept ] ) is implemented by ic100d , ic101a and the high - current buffer a3 ( and associated components ) .
it compares the output of the sample and hold ( s&h ) against the input signal divided by 2 , and then computes the pi function . @xmath28 and @xmath29 are set by means of p101 and p100 respectively . during the correct state , corrective pulses
are applied to the rs - devices ( lpcmo ) , by closing s1 during @xmath92s .
c102 and r109 introduce a time constant of @xmath93s , reducing the rise time of the pulses in order to avoid overshoots . in the probe state , ic103b closes first and @xmath91ms after , ic103a .
also , s4 clamps the inverting input of a3 to @xmath94v+ .
the timing of the circuit is generated by a timing circuit based on the master clock clk ( see fig .
a current @xmath95 flows trough the lpcmo ( v+ = 7.5v , v- = -7.5v ) .
the voltage drop at the switching interface of the rs - device ( ranging 20 - 50 mv ) is amplified by the instrumentation amplifier a1 ( gain=21 ) , eventually setting the voltage at the output of the s&h .
both interfaces of the rs - device behave complementary ( ie .
when one interface reduces its resistance , the other increases @xcite ) , then the total drop across the device is @xmath9470 mv and quite insensitive to the state , and therefore , @xmath55 .
m. j. rozenberg , m. j. snchez , r. weht , c. acha , f. gomez - marlasca , and p. levy , `` mechanism for bipolar resistive switching in transition - metal oxides , '' _ phys .
b _ , vol . 81 , p. 115101
, mar 2010 .
r. waser , r. dittmann , g. staikov , and k. szot , `` redox - based resistive switching memories - nanoionic mechanisms , prospects , and challenges , '' _ advanced materials _ , vol .
21 , no .
25 - 26 , pp . 26322663 , 2009 .
h. lee , p. chen , t. wu , y. chen , c. wang , p. tzeng , c. lin , f. chen , c. lien , and m. tsai , `` low power and high speed bipolar switching with a thin reactive ti buffer layer in robust hfo2 based rram , '' in _ electron devices meeting , 2008 .
iedm 2008 .
ieee international _ , pp . 14 , ieee , 2008 .
f. alibart , l. gao , b. d. hoskins , and d. b. strukov , `` high precision tuning of state for memristive devices by adaptable variation - tolerant algorithm , '' _ nanotechnology _ , vol .
23 , no .
7 , p. 075201
, 2012 . c. schindler , s. thermadam , r. waser , and m. kozicki , `` bipolar and unipolar resistive switching in cu - doped sio@xmath96 , '' _ electron devices , ieee transactions on _ , vol .
54 , no .
10 , pp . 27622768 , 2007 .
n. papandreou , h. pozidis , a. pantazi , a. sebastian , m. breitwisch , c. lam , and e. eleftheriou , `` programming algorithms for multilevel phase - change memory , '' in _ proc .
ieee int circuits and systems ( iscas ) symp _ , pp .
329332 , 2011 . | we study the resistive switching ( rs ) mechanism as way to obtain multi - level memory cell ( mlc ) devices . in a mlc more than one bit of information can be stored in each cell . here
we identify one of the main conceptual difficulties that prevented the implementation of rs - based mlcs .
we present a method to overcome these difficulties and to implement a 6-bit mlc device with a manganite - based rs device .
this is done by precisely setting the remnant resistance of the rs - device to an arbitrary value .
our mlc system demonstrates that transition metal oxide non - volatile memories may compete with the currently available mlcs . multilevel cell , resistive switching , non - volatile memory , reram . |
here we remind the main ingredients and predictions of the model introduced in ref.@xcite , and present new results obtained with a more generalized microscopic approach to study spin dynamics and relaxation effects induced by magnetic moments . in the latter model
, we describe the local moment formation using the self - consistent hubbard model on hydrogenated graphene @xcite .
this model is shown to correctly reproduce _ ab initio _
results for spin splittings obtained for small supercells , whereas it provides , by a scaling analysis , a more extended spreading of local magnetic moments around the h impurity when compared with the analytic model used in @xcite .
next , the hubbard tight - binding parameters are implemented into a real space wavepacket propagation method which gives direct access to spin dynamics and spin relaxation effects @xcite .
when a hydrogen atom is chemisorbed , graphene locally undergoes a structural deformation , breaking the @xmath52 symmetry in favor of an @xmath53-like hybridization .
after _ ab initio _ structural optimization , the hydrogenated carbon atom is found to be slightly displaced out of the plane ( @xmath54 ) , forming a c - h bond of length @xmath55@xmath56 . as measured experimentally , the @xmath57 rehybridization induced by the hydrogen adatom comes with the formation of a local magnetic moment of the order of 1@xmath58 .
recently , spin - unpolarized first - principles calculations suggested that the local @xmath53 distortion and its resulting pseudospin inversion asymmetry induce a giant enhancement of the spin - orbit coupling ( soc ) @xcite . in this section ,
spin - polarized first - principles calculations are used to investigate the interplay between the local magnetic moment and the enhanced spin - orbit interaction induced by hydrogen adatoms . using unconstrained spinors to represent the one - particle wave - functions , the respective contributions to the energy bands splitting of both soc and electronic exchange
are computed .
the spin textures associated with the low energy electronic spectrum are also investigated . hydrogenated graphene in the dilute limit is represented by a large supercell ( 5x5x1 ) containing a single hydrogen defect , leading to a hydrogen concentration of @xmath59
. the out - of - plane dimension of the cell ensures a distance @xmath60@xmath56 between neighboring graphene planes in order to avoid interaction between periodically repeated images . the one particle hamiltonian is computed within the framework of non - collinear spin - polarized density functional theory as implemented in the vienna ab initio simulation package ( vasp ) @xcite .
the projector augmented wave method is used to expand the one - particle wave - functions up to an energy cutoff of @xmath61ev @xcite .
the eigenstates of the self - consistent hamiltonian are populated according to an electronic temperature of 300k .
electronic exchange and correlation are treated within the generalized gradient approximation by means of the pbe functional @xcite .
integrals over the brillouin zone are performed using a 6x6x1 monkhorst - pack grid of k - points .
the geometry is fully optimized until remaining atomic forces and stresses are lower than @xmath62ev/@xmath56 and @xmath62ev/@xmath63 respectively .
various magnetization axes , either parallel or perpendicular to the graphene plane , have been considered .
the different configurations are energetically equivalent , i.e. the computed total energies vary by less than 1 @xmath4ev . in what follows ,
we focus on the most symmetrical and experimentally relevant configuration where the magnetization axis is chosen as the direction perpendicular to the graphene plane . ) including spin - orbit coupling calculated within both ( a ) spin - polarized and ( b ) spin - unpolarized frameworks .
the color scale is associated with the in - plane component of the spin expectation values .
hence , the blue and red extrema of the color scale correspond to eigenstates whose spin is parallel and perpendicular to the graphene plane .
black dots are determined by symmetry .
( c , d ) in plane projection of the spin texture of the highest valence band in both the spin - polarized ( c ) and -unpolarized ( d ) frameworks .
the arrows indicate the in - plane projections of the spin directions ( rescaled for eye convenience ) .
the color scale indicates the actual values of the in - plane projections .
red and blue arrows respectively correspond to spin momenta that are parallel and perpendicular to the graphene plane . ]
the computed low - energy band structure is depicted in fig.[figure1 ] ( a ) .
the hydrogen impurity opens an energy gap of @xmath64ev and produces two nearly - flat bands around the fermi energy , accounting for the computed total magnetic moment of @xmath65 . in this configuration ,
the spin degeneracy is broken and the energy splitting of the bands originates from both electronic exchange and spin - orbit interaction . for the sake of comparison , the low energy spectrum computed within a spin - unpolarized framework is also reported in fig.[figure1 ] ( b ) .
we note that spin - unpolarized means that the density matrix is constrained to be diagonal with same values for both spin - directions . in this case
, the energy splitting of the bands only originates from the spin orbit interaction . at the scale of fig.[figure1](b ) , the splitting is not visible to the eye .
the computed energy splittings corresponding to the low - energy bands are reported in fig.[figure2 ] ( a , c ) .
the energy splitting computed within the spin unpolarized framework [ see fig.[figure2 ] ( c ) ] is in perfect agreement with previous estimates @xcite . in order to clarify the impact of the spin - orbit interaction , the soc - induced splitting ( @xmath66 ) here defined as the difference between the energy splittings computed with and without the spin - orbit interaction
is also reported in fig.[figure2 ] ( b ) .
@xmath66 is shown to be at least two orders of magnitude smaller than the energy splitting associated with the stabilization of the local magnetic moment by the electronic exchange ( fig.[figure2 ] ( a ) ) .
this highlights the prevailing role of electronic exchange in ruling the low - energy spectrum in the presence of local magnetic moments .
the comparison of fig.[figure2 ] ( b ) and ( c ) further emphasizes the interplay between spin - orbit and exchange couplings .
the reduction of soc - induced spin splitting due to exchange coupling is explained by considering a simple effective hamiltonian .
consider electrons with momentum @xmath67 along @xmath68 and energy @xmath69 , experiencing rashba soc @xmath70 which is in general momentum dependent .
if the electrons also feel an exchange coupling @xmath71 with a magnetic moment oriented along @xmath72 , as is our case , the hamiltonian is @xmath73 the energy splitting of the eigenvalues of this operator is @xmath74 in the absence of exchange , the splitting is linear in @xmath70 , since rashba soc splits two degenerate levels . in the presence of exchange , and for realistic @xmath75 , as in our case , the rashba contribution to the splitting is only second order in @xmath70 ( @xmath76 ) .
this rough estimate gives for soc of 100 @xmath4ev and @xmath71 of 0.1 ev , a rashba coupling contribution to the splitting of approximately 0.1 @xmath4ev , significantly below what is observed in fig.[figure2 ] ( b ) .
it is then likely that the difference in the calculated exchange coupling with and without soc is due to the rashba dependence of the exchange itself .
this could be linear with @xmath70 , with a numerical factor .
the understanding of this dependence in terms of simple phenomenological models is still elusive and is currently the object of further investigation .
the spin texture of the valence band is illustrated in fig.[figure1 ] ( c ) . when the electronic density matrix is constrained to be spin - unpolarized
, the frustration of the electronic exchange `` artificially '' leads to a highly non - collinear spin texture entirely ruled by soc ( see fig.[figure1 ] ( d ) ) . on the contrary , in the unconstrained framework
, the stabilization of the magnetic moment dominates the energetics of the low energy spectrum and the spin texture is found to be nearly collinear with the spin - orbit interaction only accounting for @xmath77 of the deviation from collinearity ( fig.[figure1 ] ( c ) ) . while the local rehybridization of graphene induces an enhancement of the spin - orbit interaction , the magnetic properties of graphene as well as its spin texture are quasi - exclusively ruled by the electronic exchange in the presence of hydrogen - induced local magnetic moments .
hence , the contribution of the spin - orbit interaction to the hamiltonian is neglected in the following discussion on the spin lifetimes in hydrogenated graphene .
the electronic structure of graphene is modeled by a nearest - neighbor tight - binding model with a single @xmath78-orbital per site .
when a h atom is adsorbed on top of a carbon ( c ) atom , the @xmath52-symmetry is locally broken , and the electron from the c @xmath78 orbital is removed from the @xmath79 bands to form a @xmath80 bond with the h atom ( fig.[figure3 ] ( a ) ) .
figure [ figure3](b ) shows a 5x5 supercell where the absence of a @xmath78-orbital in the center of the green region ( yellow site ) represents hydrogen adsorption . to remove the @xmath78 orbital , we use a sufficiently large on - site potential @xmath81 ev . to properly describe magnetism in hydrogenated graphene
, we introduce on - site coulomb repulsion between electrons with opposite spins by means of the hubbard model in its mean - field approximation @xmath82 where @xmath83 is the first - neighbors hopping term , @xmath84 ( @xmath85 ) is the creation ( annihilation ) operator in the lattice site @xmath86 ( @xmath87 ) with spin @xmath80 , @xmath88 is the on - site coulomb repulsion , and @xmath89 , @xmath90 are the converged expectation values of the occupation numbers for spin - down and spin - up electrons , respectively . the ratio @xmath91 has been chosen to accurately reproduce first - principles calculations .
hybridization breakdown upon hydrogen adsorption , equivalent to the creation of a single vacancy .
( b ) single - vacancy 5x5 supercell used in the mean - field hubbard calculations with periodic boundary conditions .
the @xmath78 orbital at the center of the vacancy ( yellow site inside the green region ) is removed by adding a large on - site potential ( @xmath92 ) to simulate hydrogen adsorption . ]
hydrogen defects in graphene may induce sublattice symmetry breaking leading to the appearance of zero - energy modes in the density of states ( dos ) , as predicted by inui _
et al . _
@xcite , which are mainly localized around the impurities @xcite .
these zero - energy states become spin - polarized upon switching on the coulomb repulsion , leading to semi - local @xmath93 magnetic moments with a staggered spin density primarily localized on one sublattice @xcite . for a finite concentration of defects , the long range ordering of magnetic moments
is dictated by the type of sublattice functionalization , being co - polarized or ferromagnetic ( fm ) for the same sublattice and counter - polarized or antiferromagnetic ( afm ) otherwise .
the total spin @xmath94 of the macroscopic ground state is given by the excess of magnetic moments on one chosen sublattice @xcite , although @xmath95 is the most likely value on simple statistical grounds ( equal h occupation of both sublattices ) .
here we consider the dilute limit so that long range magnetic states are neglected and we get local paramagnetic ( pm ) impurities . ) .
the value converges apparently for @xmath96 where long range magnetic interactions start to vanish precluding the formation of fm or afm ordering and leading to isolated paramagnetic defects . on the right side of the figure
, we show the magnetic moments @xmath97 corresponding to each lattice site for different supercell sizes . ] in order to reach a pure paramagnetic state , we have studied the evolution of the local magnetic moments @xmath98 with the varying supercell size ( fig.[figure4 ] ) .
in particular , fig.[figure4 ] ( left panel ) shows the values of the magnetic moment only at nearest neighbor sites of the defect @xmath99 .
as the supercell size is increased and the defects move away from each other , @xmath99 decreases quickly and converge for a supercell size 14x14 , which corresponds to @xmath100 % of adsorbed h and @xmath101 .
this decay is related to the overlap between wavefunctions corresponding to neighboring defects . for strongly overlapping states ( large concentration of h ) , long - range interactions
induce a fm state between magnetic moments located in neighboring cells which increases the values of @xmath97 .
when the overlap is very small ( diluted limit ) , long - range interactions have no influence on the local magnetism leading to almost constant @xmath97 values around the defect ( pm state ) . of h adsorbed ( 28x28 supercell ) .
the spin degeneracy is broken at the fermi energy ( @xmath102 ) due to the formation of local magnetic fields around impurities .
( b ) evolution of @xmath103 at three symmetry points k , @xmath104 and m , for increasing values of the distance between impurities ( @xmath105 ) .
the inset shows the evolution of @xmath103 along the m - k-@xmath104 path in the hexagonal brillouin zone of the valence ( bottom ) , impurity ( center ) and conduction ( top ) bands .
dashed lines correspond to dft results . ]
figure [ figure5](a ) shows the band structure of a single hydrogenated 28x28 graphene supercell , corresponding to @xmath106 of adsorbed h , calculated using the mean - field hubbard model ( eq . [ mfhm ] ) .
the splitting ( @xmath103 ) of the mid - gap state formed during h adsorption is plotted in fig . [ figure5](b ) for different supercell sizes at different @xmath67-points . at the k
point @xmath103 decays as @xmath107 , while at the @xmath104 and m - points @xmath108 .
these results confirm the existence of very small splittings at tiny concentrations of h ( of the order of 1ppm ) . in order to compare the values of the energy splitting of spin polarized bands obtained using the mean - field hubbard model with published dft results @xcite ,
we have plotted @xmath103 along m - k-@xmath104 path in the inset of fig .
[ figure5](b ) .
bottom , middle and top panels correspond to the valence , mid - gap and conduction bands respectively .
although dft results for a 5x5 supercell ( dashed lines ) show a slightly higher splitting of the spin polarized bands with respect to our results , this methodology nicely capture the main physics of magnetic resonances induced by h adatoms in graphene .
we now investigate the spin dynamics and relaxation phenomena in hydrogenated graphene by comparing two complementary theoretical approaches and contrasting our results with state - of - the - art experimental data .
first , we consider the magnetic scattering problem in the single impurity limit , which can be simplified by considering only the spin - flip processes stemming from the exchange coupling on the resonant scatterer ( h atom ) site .
this was done in details in ref .
@xcite on the basis of the hamiltonian @xmath109 which involves the usual @xmath110-orbital hamiltonian for the pristine graphene @xmath111 together with the term @xmath112 describing graphene - hydrogen chemisorption including the interaction of electron spin @xmath113 and impurity moment @xmath114 ( exchange coupling ) @xmath115\\ & -j\,\hat{\mathbf{s}}_h\cdot\hat{\mathbf{s } } \end{aligned } \label{rgl}\ ] ] here @xmath116 ( @xmath117 ) and @xmath118 ( @xmath119 ) are fermionic operators creating ( annihilating ) electron with spin @xmath80 on the hydrogen and graphene carbon site @xmath120 , respectively .
the orbital chemisorption parameters entering the hamiltonian @xmath112 are extracted from the _ ab - initio _ calculation in refs .
@xcite the hydrogen on - site energy @xmath121 ev and graphene - hydrogen hybridization @xmath122 ev . the exchange term @xmath123 describes an effective interaction of the local magnetic moment @xmath124 induced on the hydrogen site with an itinerant electron spin @xmath125 when this hops into the hydrogen @xmath126-level that hybridizes with the graphene host . both , @xmath124 and @xmath125 are vectors of spin @xmath127-operators ( in our definition without conventional @xmath128 factor ) and @xmath129 is a constant with the dimension of energy . for the exchange
we take @xmath130 since this value is consistent with a more detailed parametrization of the magnetic impurity model as discussed in ref .
in fact the precise value and the sign of @xmath129 are not important as long as @xmath131 is greater that the orbital resonance width ( for the case of hydrogen the resonance width is @xmath132 mev ) . in the independent electron - impurity picture
( we do not discuss kondo physics ) , the exchange term @xmath123 can be diagonalized introducing the singlet ( @xmath133 ) and triplet ( @xmath134 ) composite spin states @xmath135 ( @xmath136 ) . transforming the initial hamiltonian @xmath112 , eq .
( [ rgl ] ) , to the new spin basis we arrive at @xmath137|\ell , m_\ell\rangle\langle \ell , m_\ell|\\ & + t\left [ h^+c^{\phantom{+}}_{c_h}+c_{c_h}^+h\right]\otimes\sum\limits_{\ell , m_\ell}|\ell , m_\ell\rangle\langle \ell , m_\ell|\ , .
\end{aligned } \label{rg2}\ ] ] down - folding the hydrogen @xmath138-state by means of the lwdin transformation we get for each spin component @xmath139 an independent effective delta - function problem @xmath140 located on the hydrogenated carbon site @xmath141 with the energy dependent coupling @xmath142 from here it is straightforward to compute t - matrix elements for bloch states @xmath143 and @xmath144 of the unperturbed graphene ( to shorthand the notation @xmath145 comprises band index and crystal momentum ) .
assuming the bloch states are normalized to graphene unit cell the result is as follows @xmath146 where @xmath147 is given by eq .
( [ rg4 ] ) and @xmath148 is green s function per site and spin for the unperturbed graphene , i.e. , @xmath149 the above green s function is valid near the graphene charge neutrality point in the energy window from -1ev to 1ev , where the linearized bandwidth @xmath150 ev . for practical reasons we need relaxation rates @xmath151
that characterize spin - conserving ( @xmath152 ) and spin - flipping ( @xmath153 ) processes in graphene at the given fermi energy @xmath154 in presence of magnetic active impurities . for
that we take into account all scattering processes @xmath155 at the given energy @xmath156 with the requested incoming and outgoing electron ( hole ) spins @xmath80 and @xmath157 and allowed impurity spins @xmath158 and @xmath159 ( a charge carrier in graphene flips its spin only if the magnetic moment does the same to conserve the total angular momentum ) . to get @xmath151 we start with the transition rate @xmath160 and trace out the @xmath158-spin degrees of freedom corresponding to magnetic moment @xmath124 induced on the hydrogen site
, then we integrate over all outgoing momenta @xmath144 and finally we average the result over all incoming states @xmath143 at the given fermi energy @xmath154 . this can be done because the angular dependence of the t - matrix is trivial , it does not depend on the angle of @xmath145 neither of @xmath161 . assuming
the distribution of magnetic moments is dilute and they are in average spin - unpolarized ( e.g. by interaction with phonons ) the result is as follows @xmath162\ , .
\label{rg7}\ ] ] here @xmath163 is the concentration of hydrogen impurities per carbon atom , @xmath164 is the graphene density of states per atom and spin , @xmath165 for @xmath166 is given by eq .
( [ rg4 ] ) and function @xmath167 $ ] is defined by @xmath168= \frac{1}{2}\delta_{\sigma\sigma'}\bigl|x\bigr|^2 + \frac{1}{8}\bigl|x+(\sigma\cdot\sigma')\,y\bigr|^2\,.\ ] ] the function @xmath169 originates from the decomposition of the actual spin states @xmath170 and @xmath171 with respect to the singlet - triplet basis @xmath135 after the tracing out the spin states of the induced magnetic moment . the symbol @xmath172 entering its definition equals 1 ( -1 ) for the parallel ( antiparallel ) spin alignments and @xmath173 stands for the absolute value . knowing the partial rates @xmath151 we can define the _ spin - relaxation rate _ @xmath174 and _ momentum - relaxation rate _
@xmath175 at zero temperature as functions of fermi energy . from eq .
( [ rg7 ] ) we immediately see that both @xmath176 and @xmath177 should go to zero at the charge neutrality point since there are no states that can participate in scattering , @xmath178 . secondly , the denominators @xmath179 and @xmath180 that enter the function @xmath169 can minimize at certain energies @xmath181 and @xmath182 and this would manifests as two sharp peaks ( singlet and triplet one ) in the spin and momentum relaxation rates .
both mentioned features are clearly seen in fig .
[ figure7 ] . to account for finite temperature effects one
should thermally broaden @xmath176 implementing rate equations for the graphene charge - carriers that obey fermi - dirac statistics .
the concise formula for the spin - relaxation rate at finite temperature @xmath183 then becomes @xmath184 1/\tau_s(\epsilon_k)}{\sum\limits_{k } \bigl[- \partial_{\epsilon_k } \mathrm{f}(\epsilon_k , e , t)\bigr]}\,,\ ] ] where @xmath185 is the equilibrium fermi - dirac distribution at temperature @xmath183 and fermi energy @xmath154 .
experiments show that charged impurities and electron - hole puddles within the sample still further affect system s fermi energy . to account for all that effects one usually convolves @xmath186 by the gaussian kernel with standard deviation @xmath187 , @xmath188 depending on the sample quality the typical value of @xmath187 to be taken for best fits is about 70 - 110mev ( see discussion next section ) . in our second approach to spin relaxation ,
we study the spin dynamics using the following tight - binding hamiltonian , @xmath189 which includes the self - consistent hubbard terms describing isolated magnetic moments ( dilute limit ) , and a random distribution of h - adatoms in the ppm range . the long range nature of magnetic moment induced by hydrogen is involved in this hamiltonian by considering up to ninth - nearest - neighbor exchange coupling term @xmath190 .
spin - orbit coupling is found to yield negligible corrections to the results , so it is neglected in the following .
the spin dynamics are investigated by computing the time - dependence of the spin polarization defined by @xcite @xmath191}{{\rm tr}\left[\delta(e-\hat{h})\right]}\nonumber\ ] ] where @xmath192 . using random phase states to perform the trace efficiently ,
we get @xmath193 and the initial wavepacket can be prepared in an arbitrary spin polarization as @xmath194 in what follows , the wavepackets are prepared ( at @xmath195 ) with an in - plane spin polarization and their time - dependent propagation and @xmath196 are evaluated applying the evolution operator ( schrdinger equation ) to the wavepackets . in absence of spin - orbit interaction , the spin dynamics
is influenced by the existence of local magnetism , with h - related magnetic moments pointing out - of - plane .
fig.[figure6 ] shows the time - evolution of the spin polarization at three different energies , namely the dirac point , an energy close to the expected resonance and some high energy value ( see inset ) .
the curves first exhibit an sudden drop ( especially for the dirac point ) followed by an exponential decay , which dictates the values of the spin relaxation times @xmath197 using @xmath198 ( fitting the numerical results from @xmath199ps ) . the initial fast decay of @xmath200
is understood as follows .
we study the propagation of wavepackets which ( for computational efficiency ) are at time @xmath195 in a random phase state , that is extended through the whole system . for smooth disorder
, this method well captures the main transport length scales as described in ref.@xcite . here
however , the hydrogen adatoms ( or vacancies ) produce strong magnetic moments localized around the impurity .
this introduces some transient decay of the initial spin polarization which is not representative of the long time exponential behavior . in fig.[figure7 ] , the inverse spin relaxation times obtained for the two models are superimposed . within the single impurity approximation , two sharp _ magnetic resonances _ are clearly formed ( for singlet and triplet states ) , at which @xmath201 is maximum at zero temperature ( green dashed lines ) .
this approximation of @xmath46 however diverges close to the dirac point , which disagrees with most low - temperature experimental data .
room temperature spin - relaxation experimental data can however be reproduced using thermal and charge - puddles broadenings . a choice of 0.36 ppm of h - impurities together with a finite temperature broadening ( 300 k ) and charge density fluctuations of 110 mev are necessary ( blue dashed line ) to reproduce experimental data ( black filled circles ) obtained for high - quality graphene spin - valve devices . @xcite .
inset : total density of state for the case of 1ppm of hydrogen on graphene . ]
the inverse spin lifetimes obtained by the real space spin propagation method are shown in fig.[figure7 ] for a h - density of 1 ppm ( black solid line ) , 2 ppm ( red solid line ) and 15 ppm ( blue solid line ) .
the presence of the two resonances is confirmed by this calculation . at the resonances ,
the inverse spin lifetime is greater than in the single impurity approximation , which is likely due to the different natures of the adatom models . in the single - impurity adatom model
the the magnetic moment is strongly localized and the resonance is sharp , while in the hubbard model we have an effective defect model with more delocalized magnetism is real space .
this broadening reduces the _ residency time " _ of the electrons on the _ magnetic virtual bound state _ ( as pictured in the inset of fig.[figure7]-inset ) , and thus yields longer spin lifetimes .
indeed for 0.36 ppm of h - impurities , the two models differ to three orders of magnitude in spin lifetime . by further contrasting the real - space spin results to the experimental data ( black dots in fig.[figure7 ] )
, we extrapolate that a density of 10 - 20 ppm of h - adatoms can reproduce the range of experimental values of the graphene samples measured in ref.@xcite .
derived from the t - matrix of the single impurity model at zero temperature ( green dashed lines ) ; and at 300 k broadened by puddles with energy fluctuations of 110 mev ( blue dashed lines ) .
same quantity obtained from the real space spin propagation method for 1 ppm ( solid black line ) , 2 ppm ( solid red line ) and 15 ppm ( solid blue line ) of h adatoms .
inset : schematic of the magnetic resonance process at the origin of enhanced spin relaxation . ] to discuss the relaxation mechanism , we compute the momentum relaxation time @xmath202 first from the sum of spin conserving rate and spin - flip rate , which both can be calculated by transforming the singlet and triplet t - matrix amplitudes via composite spin states of electron and impurity @xcite .
the second approach uses the full hamiltonian @xmath112 and the time dependence of the diffusion coefficient @xmath203 @xcite . for the hydrogen coverage as low as @xmath204
, we still can observe the saturation of the diffusion coefficient @xmath205 , allowing us to extract @xmath202 as @xmath206 where @xmath207 is the pristine graphene velocity .
we use the fermi golden rule and the expected scaling of @xmath208 , to extrapolate the values for each ppm concentration ( see fig.[figure8 ] ) .
the result is given in fig.[figure8 ] ( green solid lines ) , and shows a weak energy dependence , with slight increase close to the dirac point , with values in the range of 60 - 70 ps for 1 ppm of hydrogen impurities . in fig.[figure8](inset ) , @xmath202 for the two models are reported showing some discrepancy close to the dirac point , as expected from the approximations made .
the values match well at high enough energy .
we observe that @xmath209 , while the scaling of @xmath46 with impurity density clearly manifests an elliott - yafet spin relaxation mechanism , as predicted for such type of impurities @xcite . in some experiments , @xmath46 is found to increase with hydrogen concentration , suggesting differently a dyakonov - perel relaxation mechanism @xcite , whereas others report on the opposite trend associated to the elliott - yafet mechanism @xcite .
the origin of such inconsistency remains obscure at that point but could stem ( in particular ) from the segregation of h - adatoms and the alteration of fundamental effects induced by magnetic resonances . in conclusion
, we investigated the impact of hydrogen adatoms on charge and spin transport in transport in graphene in the dilute limit ( down to the ppm limit ) , for which no long range magnetic ordering develops .
the importance of magnetic resonances as a new spin relaxation mechanism , as pioneered in ref.@xcite has been consolidated and further quantified using extended models of more delocalized magnetism , described within the hubbard hamiltonian in the mean field approximation .
using efficient real space order n wavepacket propagation methods , spin relaxation times in the nanosecond regime were obtained for 1 ppm hydrogen impurities , and for momentum scattering time few orders of magnitude shorter .
more work is needed to describe larger hydrogen densities for which magnetic interaction and long range magnetic ground states could develop and interfere with relaxation effects .
the research leading to these results has received funding from the european union seventh framework programme under grant agreement number 604391 graphene flagship .
this work was also funded by spanish ministry of economy and competitiveness under contracts mat2012 - 33911 , the severo ochoa program ( mineco sev-2013 - 0295 ) .
s. r. thanks the secretaria de universidades e investigacin del departamento de economa y conocimiento de la generalidad de catalua . j.c.c . acknowledges the national fund for scientific research [ f.r.s .- fnrs ] of belgium for financial support and the research concerted action on graphene nanoelectromechanics ( no .
11/16 - 037 ) .
computational resources have been provided by the supercomputing facilities of the `` consortium des equipements de calcul intensif '' en fdration wallonie - bruxelles ( ceci ) , the ucl - cism , and prace . j.f .
acknowledges support from dfg sfb 689 and grk 1570 .
d. c. elias , r. r. nair , t. m. g. mohiuddin , s. v. morozov , p. blake , m.p .
halsall , a.c .
ferrari , d.w .
boukhvalov , m.i .
katsnelson , a.k .
geim , k.s .
novoselov , science * 323 * , 610 ( 2009 ) ; a. bostwick , j.l .
mcchesney , k.v .
emtsev , t. seyller , k. horn , s.d .
kevan , and e. rotenberg , phys .
lett . * 103 * , 056404 ( 2009 ) ; j. katoch et al .
b * 82 * , 081417(r ) ( 2010 ) ; m. jaiswal , c. haley y.x .
lim , q. bao , ch.t .
toh , k. p. loh , b. ozyilmaz , acs nano * 5 * , 888 ( 2011 ) ; r. jayasingha , a. sherehiy , s .- yu wu , and g. u. sumanasekera , nano lett . * 13 * , 5098 ( 2013 ) .
s. h. cheng , k. zou , f. okino , h.r .
gutierrez , a. gupta , n. shen , p.c .
eklund , j.o .
sofo , j. zhu , phys .
b * 81 * , 205435 ( 2010 ) ; f. withers , m. dubois , and a. k. savchenko , phys . rev .
b * 82 * , 073403 ( 2010 ) .
s. roche , d. mayou , phys .
* 79 * , 2518 ( 1997 ) ; s. roche , physical review b * 59 * , 2284 ( 1999 ) ; s. roche , n. leconte , f. ortmann , a. lherbier , d. soriano , j .- c .
charlier , solid .
152 * , 1404 ( 2012 ) .
j. guillemette , s. s. sabri , binxin wu , k. bennaceur , p. e. gaskell , m. savard , p. l. levesque , f. mahvash , a. guermoune , m. siaj , r. martel , t. szkopek , and g. gervais , phys .
lett . * 110 * , 176801 ( 2013 ) .
d. soriano , n. leconte , p. ordejn , j .- ch .
charlier , j .- j .
palacios , and s. roche , phys .
lett . * 107 * , 016602 ( 2011 ) ; n. leconte , d. soriano , s. roche , p. ordejn , j .- ch . charlier and j. j.
palacios acs nano * 5 * , 3987 ( 2011 ) .
p. j. zomer , m. h. d. guimares , n. tombros , and b. j. van wees , phys .
b * 86 * , 161416(r ) ( 2012 ) .
m. drgeler , f. volmer , m. wolter , b. terrs , k. watanabe , t. taniguchi , g. gntherodt , c. stampfer and b. beschoten , nano lett * 14 * , 6050 ( 2014 ) . | in this review we discuss the multifaceted problem of spin transport in hydrogenated graphene from a theoretical perspective .
the current experimental findings suggest that hydrogenation can either increase or decrease spin lifetimes , which calls for clarification .
we first discuss the spin - orbit coupling induced by local @xmath0 re - hybridization and @xmath1 c - h defect formation together with the formation of a local magnetic moment .
first - principles calculations of hydrogenated graphene unravel the strong interplay of spin - orbit and exchange couplings .
the concept of magnetic scattering resonances , recently introduced @xcite is revisited by describing the local magnetism through the self - consistent hubbard model in the mean field approximation in the dilute limit , while spin relaxation lengths and transport times are computed using an efficient real space order n wavepacket propagation method .
typical spin lifetimes on the order of 1 nanosecond are obtained for 1 ppm of hydrogen impurities ( corresponding to transport time about 50 ps ) , and the scaling of spin lifetimes with impurity density is described by the elliott - yafet mechanism .
this reinforces the statement that magnetism is the origin of the substantial spin polarization loss in the ultraclean graphene limit . _
introduction_.-remarkable electronic and transport properties of graphene ( see for instance @xcite ) can be further tailored towards higher functionality through the use of chemical functionalization @xcite , irradiation ( defect formation ) @xcite or structural patterning ( such as creating a nanomesh superlattice ) @xcite . among the wealth of possible modifications , hydrogenation , ozonization or fluorination
have demonstrated the tunability of graphene from a weakly disordered semimetal to a wide band - gap or anderson insulator depending on the nature and density of impurity atoms , varying typically from @xmath2 to few percent @xcite . together with raman spectroscopy which provides an estimate of defect density , analysis of the low - temperature conductance behaviour and weak localization regime enable evaluation of the main transport length scales ( such as the mean free path and localization length ) @xcite . surprisingly , strongly hydrogenated graphene , with anomalously large ioffe - regel ratio of @xmath3 still exhibits quantum hall effect features in the high magnetic field regime @xcite .
hydrogen defects are particularly interesting since , together with a resonant scattering state created locally in space , the breaking of the sublattice symmetry entails the formation of a zero - energy mode and a local magnetic moment on the order of 1 bohr magneton for an isolated hydrogen adatom @xcite . meanwhile , the coupling between different induced magnetic moments is either ferromagnetic or antiferromagnetic , depending on whether the h - adatoms correspond to the same or to different sublattices of the graphene lattice , respectively .
theoretical calculations have reported specific magnetoresistance signals for specific long range magnetic ordering situations @xcite , which could be realized by substrate - induced chemical reactivity as proposed in ref .
@xcite .
graphene spintronics has attracted a lot of attention since the pioneering demonstration that spin could be efficiently injected and propagated over long distances at room temperature @xcite .
this has opened an opportunity for the development of lateral spintronics @xcite , that would benefit from the unique features of graphene , such as a high mobility and remarkable electronic and transport features like relativistic energy dispersion , klein tunneling phenomenon , ... @xcite . in table 1
, we compile the typical values obtained for charge and spin transport in a variety of graphene devices , from graphene supported on silicon oxide or boron nitride to suspended graphene .
one observes that the variation of mobility varies by up to two orders of magnitude , whereas spin lifetime seems difficult to relate to the quality of the material .
0.6|x|c|c|c|c|c| substrate & @xmath4 ( @xmath5 ) & @xmath6 ( @xmath7 ) & @xmath8 ( @xmath7)&@xmath9 & @xmath10 + @xmath11@xcite & @xmath12 & @xmath13 & @xmath13 & @xmath14 & @xmath15 + @xmath11@xcite & @xmath16 & @xmath17 & @xmath18 & @xmath19 & @xmath20 + suspended@xcite & @xmath21 & @xmath22 & @xmath23 & @xmath24 & @xmath25 + @xmath26@xcite & @xmath12 & @xmath13 & @xmath27 & @xmath28 & @xmath29 + @xmath26@xcite & @xmath30 & @xmath17 & @xmath17 & @xmath17 & @xmath31 + @xmath32@xcite & @xmath33 & @xmath34 & @xmath34 & @xmath35 & @xmath36 + @xmath37@xcite & @xmath38 & @xmath39 & @xmath34 & @xmath40 & @xmath41 + understanding spin transport in hydrogenated graphene is a challenging and important issue .
the underlying goal is to use a low enough level of hydrogenation to preserve a sizable transport signal while also inducing local or long range magnetic ordering ( ideally a ferromagnetic ground state ) , and evaluate how spin diffusion is affected by interaction between itinerant spins and local magnetic moments .
ultimately , one could envision spin devices such as spin valves that do not use ferromagnetic materials to inject and detect spins , but rather utilize the varying signal response of `` magnetized graphene '' to manipulate the spin degree of freedom and engineer logic functions , a long - standing quest of spintronics @xcite . from that perspective , the first experimental result showing some interaction between spin diffusion and magnetic moments produced by h adatoms was reported in 2012 by mccreary and coworkers @xcite .
low - temperature spin transport measurements on graphene spin valves ( t=15k ) were shown to exhibit a dip in the non - local spin signal as a function of the external magnetic field .
this effect was tunable with hydrogen density , and was related to spin relaxation spin relaxation by exchange coupling with paramagnetic moments .
spin lifetimes were estimated to be @xmath42ps in absence of hydrogen , and were slightly enhanced upon increasing the h - adatom density up to @xmath43 ) , and accompanied by a reduction of charge mobility by one order of magnitude to @xmath44 . a subsequent experiment also reported a surprising increase of spin lifetime with hydrogen density , with values enhanced by a factor of two after hydrogenation of the samples for @xmath45 @xcite .
the upscaling of @xmath46 with hydrogen density suggests a dyakonov - perel relaxation mechanism @xcite . in these experiments ,
the effect of magnetic moments is proposed to be strong enough to counteract the expected increase of spin - orbit coupling , which would be expected to produce shorter spin lifetimes with increasing h - adatom density . in 2013 , j. balakrishnan and coworkers @xcite reported experimental evidence for a room temperature spin hall effect ( she ) in weakly hydrogenated graphene , with non - local spin signals up to @xmath47 ( orders of magnitude larger than in metals ) .
the non - local she revealed by larmor spin - precession measurements was assigned to a `` colossal enhancement '' of the spin - orbit interaction induced by h - adatoms for density in the range @xmath48 , mobilities @xmath49 ) and spin lifetime on the order of @xmath50ps .
the spin - orbit interaction was estimated to be about 2.5 mev , one order of magnitude higher than _ ab initio _
calculations @xcite .
the observation of the large she - signal was exclusively assigned to an enhancement of spin - orbit coupling by h - adatoms , which limits spin lifetimes .
it is however not clear whether the previously observed contribution of magnetic moments in spin transport @xcite remains marginal or not for the formation and strength of the she signal , and which mechanism ( dyakonov - perel @xcite or elliott - yafet @xcite ) can explain the variety of conflicting experimental data . in this context , it is of prime importance to clarify the specific impact of h - adatoms on spin relaxation in graphene , and to quantify the relative contribution of spin - orbit coupling and magnetic effects .
a first fundamental advance in this direction has been made by kochan , gmitra and fabian @xcite who introduced a new relaxation mechanism in hydrogenated graphene driven by resonant scattering by magnetic impurities . by neglecting the spin - orbit coupling effect ,
spin lifetimes of @xmath51ps were estimated for 1 ppm hydrogen at room temperature and for large contribution of electron - hole puddles . |
dk is supported by an alexander von humboldt professorship of the alexander von humboldt foundation and the bmbf .
ok thanks the imprs for geometric analysis , gravitation & string theory ( potsdam golm ) for support .
in this section , we give a short introduction as well as a presentation and discussion on the type of results we obtain in this paper .
the usual way to compute a physical probability amplitude is to replace each term of the perturbative series in the coupling @xmath1 by a sum of feynman graphs . applying renormalized feynman rules @xmath19 to all such graphs
translates this sum of graphs to the physically observable renormalized feynman amplitude , say a green function @xmath20 , at least as a formal series .
@xmath19 evaluated on a graph is a polynomial in a suitably chosen external scale parameter @xmath21 , @xmath22 .
@xmath2 includes for example the center of mass energy @xmath23 given by the underlying process , with @xmath24 fixing a reference scale for renormalization .
further dependences are scattering angles , collected in @xmath25 by a set of variables @xmath26 .
these are dimensionless parameters incorporating dependences on scalar products @xmath27 or masses @xmath28 . throughout ,
we assume we leave those scattering angles unchanged for the renormalization point .
a discussion of this point can be found in @xcite .
therefore , any renormalized green function @xmath29 can be written as a triangular expansion @xmath30 called ` log - expansion ' . here , @xmath1 denotes the perturbative parameter , a coupling constant say , and @xmath31 are some functions in @xmath26 .
we have @xmath32 . for @xmath33 ,
the @xmath34 may depend on @xmath26 and are obtained from the first @xmath16 terms of a perturbative expansion in the coupling .
@xmath35 is a series in @xmath36 with @xmath37 .
these series are determined from their counterparts in the universal enveloping algebra of feynman graphs that we detail below , see also @xcite .
such a form of the green functions in a renormalizable field theory is a consequence of the hopf algebra structure of feynman graphs @xcite .
one starts from the fact that all superficially divergent one - particle irreducible ( 1pi ) feynman graphs of any physical quantum field theory generate a hopf algebra @xmath38 .
this allows to introduce dyson - schwinger equations ( dses ) as fix - point equations for feynman graphs upon using hochschild cohomology @xcite .
see @xcite for an effective application of these mathematical structures to the automatization of perturbative renormalization , graph generation and graph counting . given a quantum field theory , one always finds dses whose solution is simply related to the log - expansion * * ( * * eq .
( [ eq : log - expansion ] ) * ) * by applying renormalized feynman rules @xmath19 .
loc foissy classified exhaustively the structure of possible dses @xcite .
we apply our approach to two exemplary cases : first , to the fermion propagator @xmath39 occurring in yukawa theory .
@xmath40 represents an infinite sum of graphs that satisfies the dse @xmath41 secondly , to the photon self - energy @xmath42 occurring in quantum electrodynamics ( qed ) .
the infinite sum of graphs @xmath43 satisfies the dse @xmath44 note that acting with renormalized feynman rules @xmath19 on @xmath40 and @xmath43 yields the log expansions * * ( * * eq .
( [ eq : log - expansion])*)*. in yukawa theory , one should consider systems of dses because there are no ward identities . here , we restrict to a truncation eliminating vertex divergences for purposes of presentation .
our paper consists of two parts .
first , in section [ sec : preliminaries ]
, we give a brief overview on hopf algebras and dses . in particular , we relate the hopf algebra of feynman graphs to the hopf algebra of words by a morphism of hopf algebras .
once the solution of a dse is given in the hopf algebra of words , we describe the filtration method in section [ sec : filtration - method ] .
there , we rely on properties of renormalized feynman rules that we derive in section [ sec : renormalized - feynman - rules ] . finally , we present the filtration algorithm in section [ sec : filtration - algorithm ] . in the second part of our paper , we describe a general method to derive the next - to@xmath45-leading log expansion in section [ sec : relat - betw - next ] . in particular , we exemplify up to @xmath46 for the yukawa fermion propagator . in appendix
[ sec : qed - results ] , we collect the respective results for the qed photon self - energy . in the filtrations of the yukawa fermion propagator @xmath40 and the qed photon self - energy @xmath43
, each term comes with a multiplicity .
we list some resulting series in these multiplicities in appendix [ sec : summary ] . in the remainder of this section , we summarize and discuss our results .
cc7.5cmcc + & * periods @xmath18 * & * yukawa generating functions @xmath17 * & * qed generating functions @xmath17 * + @xmath47 & @xmath48&@xmath49 & @xmath50 + @xmath51 & @xmath52&@xmath53 & @xmath54 + @xmath55 & @xmath56&@xmath57 & @xmath58 + @xmath59 & @xmath60&@xmath61 & @xmath62 + @xmath63 & @xmath64&@xmath65 & @xmath66 + @xmath67 & @xmath68&@xmath69 & @xmath70 + @xmath71 & @xmath72&@xmath73 & @xmath74 + @xmath75 & @xmath76&@xmath77 & @xmath78 + @xmath79 & @xmath80&@xmath81 & @xmath82 + @xmath83 & @xmath84 } } = \alpha \left ( \phi_r\left(b_+^{\gamma_1}(\gamma_2)\right ) - \phi_r\left(b_+^{\gamma_2}(\gamma_1)\right ) \right ) $ ] & @xmath85}(u ) = \frac{x^3}{4 } - \frac{x^5}{4 } + \frac{x^3}{4}\log x - \frac{3x^5}{4}\log x$ ] & @xmath85}(u ) = \frac{y}{2 } - \frac{y^2}{2}$ ] + @xmath86 & @xmath87 } } = \alpha \bigg ( 2\phi_r \left ( b_+^{\gamma_1 } \left ( \gamma_1 \cup \gamma_1
\right ) \right ) - \phi_r \left ( b_+^{\gamma_1}\left(b_+^{\gamma_1}\left(\gamma_1\right)\right ) \right ) - \phi_r(\gamma_1)\phi_r\left(b_+^{\gamma_1}\left(\gamma_1\right)\right ) \bigg ) $ ] & @xmath88}(u ) = -\frac{x^3}{8 } + \frac{x^5}{8 } - \frac{x^3}{8}\log x + \frac{3x^5}{8}\log x $ ] & @xmath88}(u ) = 0 $ ] + let us first concentrate on yukawa theory , eq .
( [ eq : dse - yuk ] ) . to find the leading log expansion we can simplify to @xmath89 feynman rules @xmath19 with massless internal propagators and a momentum scheme for subtraction turn eq .
( [ eq : dse - yuk ] ) into a differential equation for the corresponding anomalous dimension , which can be solved implicitly @xcite .
if we are only interested in the leading log expansion the situation is even simpler .
we only have to solve @xmath90 indeed , only inserting the one - loop correction for the yukawa fermion propagator , @xmath91 into itself in all possible ways give graphs that contain leading log contributions . in our paper
, we introduce a convenient matrix notation . @xmath92$]-bracketed matrices with a dot in the upper left entry denote the functions @xmath8 in one variable , say @xmath93 .
these functions occur in the log - expansions * * ( * * eq . ( [ eq : log - expansion ] ) * ) * setting @xmath94 in section [ sec : relat - betw - next ] , we explain the notation and derive ordinary first order differential equations for these objects .
the respective differential equations depend on the corresponding dses * * ( * * in our case , eq . ( [ eq : dse - yuk ] ) * ) * and are solved for yukawa theory in section [ sec : relat - betw - next ] . for the leading log order
, we define @xmath95 with corresponding period @xmath14 .
the leading log expansion * * ( * * @xmath13 in eq .
( [ eq : log - expansion ] ) * ) * finally yields @xmath96 @xmath97 and @xmath98 are also given in the first line of table [ tab : results ] .
( [ eq : ll - solution ] ) is obvious as feynman rules in a momentum scheme map the hochschild one - co - cycle @xmath99 to the hochschild one - co - cycle @xmath100 on polynomials in the variable @xmath36 @xcite . to get the next - to - leading log expansion , we consider two contributions : the insertion of the one - loop propagator graph into itself gives a contribution @xmath101 .
the first term correspond to the leading log and the second term to the next - to - leading log expansion .
there is also a contribution @xmath102 from the next hochschild one - co - cycle provided by @xmath103 .
the latter contribution occurs in eq .
( [ eq : dse - yuk ] ) that have a single appearance of @xmath103 ( beside several @xmath104 ) .
we can therefore simplify to @xmath105 this gives the next - to - leading log expansion * * ( * * @xmath106 in eq .
( [ eq : log - expansion ] ) * ) * , @xmath107 the functions @xmath108 and @xmath109 as well as the corresponding periods are listed in lines @xmath51 and @xmath55 of table [ tab : results ] . for the next - to - next - to - leading log expansion , there are several contributions , which we do not summarize here . in section [ sec : next - tonext - leading ] , we derive @xmath110 } } \begin{bmatrix}\bullet\end{bmatrix}_{[a_1,a_2]}(u ) + \alpha^2 p_2^{\begin{bsmallmatrix}\bullet\end{bsmallmatrix}_{[a_1,\theta(a_1,a_1 ) ] } } \begin{bmatrix}\bullet\end{bmatrix}_{[a_1,\theta(a_1,a_1)]}(u)\end{aligned}\ ] ] with generating functions and periods given in lines @xmath111 of table [ tab : results ] .
( [ eq : gr - ll],[eq : gr - nll],[eq : gr - nnll ] ) and the respective periods ( see the first column of table [ tab : results ] ) are in general , valid for any dse . however , the generating functions ( 2nd and 3rd column of table [ tab : results ] ) depend on the dse . _ for the qed photon self - energy ,
the corresponding dse is given in eq .
( [ eq : dse - qed ] ) .
the respective generating functions are listed in the third column of table [ tab : results ] and are much simpler .
this is because there is no insertion point in the hochschild one - co - cycle @xmath99 in qed .
it follows that the mentioned differential equations for generating functions turn out to be ordinary equations .
note that in qed , we have @xmath112 instead of @xmath113 .
the usual perturbative approach to quantum field theory does not suffice in high - energy regimes , where @xmath114 .
there , it becomes significant to consider the log - expansion instead of perturbation theory .
our results simplify the calculation of the log - expansion in eq .
( [ eq : log - expansion ] ) drastically .
one usually needs to compute an infinite number of feynman integrals of any loop - number . using our results ,
the complete next - to@xmath45-leading log expansion only depends on at most @xmath16-loop graphs . for example , the leading log order * * ( * * eq .
( [ eq : gr - ll ] ) * ) * only depends on the feynman graph @xmath104 . the next - to - leading log green function in eq .
( [ eq : gr - nll ] ) only depends on the graphs @xmath104 , @xmath103 and @xmath115 .
the next - to - next - to - leading log expansion only depends on the feynman graphs @xmath116 see eq .
( [ eq : gr - nnll ] ) and the periods in the first column of table [ tab : results ] . in summary , we filter the images of feynman graphs * * ( * * as in eq .
( [ eq : feynman - graphs ] ) * ) * in a suitable universal enveloping algebra that we construct below .
this decomposition then yields ordinary first order differential equations for generating functions @xmath8 .
we solve for the functions @xmath8 , which are the coefficients for the periods @xmath18 obtained in these filtrations .
the methods presented here are valid in general .
one can apply the described techniques to any dse in any quantum field theory .
one could also compute the log - expansion for systems of dses .
we expect that the generating functions then become solutions in systems of ordinary first order differential equations . here , we are content in exhibiting our approach .
a structural analysis of its mathematical underpinnings is left to future work .
we start with the introduction of necessary preliminaries .
let @xmath38 be the vector space of 1pi feynman graphs and their disjoint unions in a given quantum field theory .
@xmath117 forms an associative and unital algebra , where @xmath118 denotes the empty graph and @xmath119 is the disjoint union of graphs , serving as a product .
this algebra is graded by the loop number ( the first betti number ) as an infinite sum of finite dimensional vector spaces @xmath120 with @xmath121 and augmentation ideal @xmath122 note that cartesian products @xmath123 deliver a decreasing filtration @xmath124 .
the associated graded spaces @xmath125 contain @xmath126 as the linear span of graphs in first degree .
we set @xmath127 .
@xmath38 acquires a co - algebraic structure by introducing a co - product @xmath128 that acts on 1pi graphs @xmath129 as @xmath130 @xmath131 is the set of all proper sub - graphs @xmath132 such that @xmath133 is the disjoint union of divergent 1pi sub - graphs in @xmath129 .
the action of @xmath134 on @xmath118 and products of graphs is given by @xmath135 where @xmath136 flips the second and third element of the fourfold tensor product .
the co - product is co - associative @xcite . together with a co - unit @xmath137 that assigns a non - zero value only for @xmath118 and @xmath138 , @xmath139 forms a co - associative and co - unital co - algebra .
the given construction implies that @xmath140 forms a bi - algebra as well . finally , @xmath141 forms a hopf algebra , abbreviated by @xmath38 .
@xmath142 is the antipode that fulfills @xmath143 @xmath38 is called the hopf algebra of feynman graphs .
there have been various introductions to hopf algebras and in particular , to the hopf algebra of feynman graphs @xcite .
graphs without sub - divergences form primitive elements . by eq .
( [ eq : delta - fg ] ) , their reduced co - product , @xmath144 vanishes . for each such primitive element @xmath145 , @xmath146
, we define a grafting operator @xmath147 that linearly inserts its argument graph(s ) into @xmath129 . for example , @xmath148 see figure [ fig : graphs ] for the qed feynman graphs @xmath149 and @xmath150 .
@xmath151 furthermore , we have @xmath152 .
if the insertion is not unique , the result is a sum over all possibilities , for example @xmath153 the graphs @xmath149 , @xmath154 , @xmath155 and @xmath156 are defined in figure [ fig : graphs ] .
the @xmath157 are hochschild one - co - cycles by definition @xcite @xmath158 they generate the co - radical filtration and the associated grading by sub - divergences . to define the co - radical filtration , set @xmath159 and for @xmath160 set @xmath161
this is an increasing filtration , @xmath162 , and we set @xmath163 , so the first degree elements are given as @xmath164 in this filtration , @xmath165 so they are the same element in @xmath166 , an algebraic fact that by @xmath19 is the starting point for the existence of the renormalization group @xcite .
hochschild closedness * * ( * * eq.([eq : hochschild-1-co - cycle ] ) * ) * will also be essential when we relate a general dse and its solution to the hopf algebra of words in section [ sec : isom - betw - mathc ] .
this relation is reminiscent of the flag - decomposition @xcite which appears when analyzing renormalized amplitudes as a limiting mixed hodge structure .
we now turn to the hopf algebra of words , for which the decreasing and increasing filtrations @xmath167 and @xmath168 above exist analogously .
let @xmath169 be the vector space of words and @xmath170 be the sub - space of letters .
we need to collect some properties of @xmath169 . in the remainder of our paper , we abbreviate letters by @xmath171 and words by @xmath172 .
concatenation of letters creates words , e.g. , @xmath173 and @xmath174 form new words .
let furthermore @xmath175 be a commutative and associative map that assigns a new letter to any two given letters .
it is always assumed that @xmath176 is completed if necessary so that it contains all images of @xmath177 . using this map
, we can recursively define the shuffle product @xmath178 ( also denoted by @xmath179 ) as @xmath180 where on the rhs of eq .
( [ eq : shuffle1 ] ) any word in brackets is concatenated from right to the respective letter . the so defined shuffle product is commutative and associative @xcite .
thus , @xmath181 forms an associative and unital algebra .
@xmath169 acquires a co - algebraic structure by introducing the co - product @xmath182 and the co - unit @xmath183 acting on words as @xmath184 for example , the co - product of the word @xmath185 is @xmath186 @xmath187 forms a bi - algebra as well . finally , @xmath188 forms a hopf algebra , called the hopf algebra of words
. the antipode @xmath189 fulfills @xmath190 for more details on the hopf algebra of words , the reader may consult the textbook of reutenauer @xcite .
we finally introduce the grafting operators .
the primitive elements of @xmath169 are all letters , since @xmath191 .
we then define @xmath192 which means that @xmath193 concatenates the letter @xmath194 from left to its argument word ( or sum of words since @xmath193 acts linearly ) .
the reader may check using eqs .
( [ eq : deltaw],[eq : b+ ] ) , that the @xmath193 are indeed , hochschild one - co - cycles , @xmath195 note that @xmath196 is the linear span of words that can not be written as a shuffle product , in analogy to the definitions for @xmath38 .
similarly , @xmath197 can be identified with the completed set of letters .
we can now relate the hopf algebra of feynman graphs to the hopf algebra of words .
indeed , there exists a unique hopf algebra morphism @xmath198 that fulfills @xmath199 it respects the hopf algebraic structures .
the existence of such a morphism is guaranteed by the fact that the grafting operators are hochschild one - co - cycles @xcite .
for example , eqs .
( [ eq : morphism - begin ] , [ eq : morphism - end ] ) give @xmath200 ( @xmath201 assigns the letter @xmath202 to the primitive feynman graph @xmath203 ) .
we give another example .
consider the last graph in figure [ fig : graphs ] . with @xmath204 ,
we find @xmath205 the hopf algebra morphism @xmath201 allows us to translate any dse to the hopf algebra of words .
in particular , applying @xmath201 to eqs .
( [ eq : dse - yuk],[eq : dse - qed ] ) and using eqs .
( [ eq : morphism - begin]-[eq : morphism - end ] ) yields @xmath206 here , we drop any further subscript on letters , hence @xmath207 and @xmath208 .
we finally solve eqs .
( [ eq : dse - yuk - words ] , [ eq : dse - qed - words ] ) via the anstze @xmath209 and obtain @xmath210 the first orders are @xmath211 , @xmath212 and @xmath213 , all others are recursively given .
_ this ansatz can also be used for any other dse @xcite .
_ renormalized feynman rules linearly act on words as @xmath214 and @xmath215 in the log - expansion * * ( * * eq .
( [ eq : log - expansion])*)*. the remaining question is now : which part of @xmath216 maps to which power of the external scale parameter @xmath2 in the log - expansion * * ( * * eq .
( [ eq : log - expansion ] ) * ) * ? we will answer this question in the next section , in full accordance with the blow - ups needed for feynman integrands from the viewpoint of algebraic geometry @xcite .
in the following , we filter the coefficients @xmath217 in the solution of any dse * * ( * * as occurring in eq .
( [ eq : dse - solution])*)*. each filtered term then maps to a certain power of @xmath2 in the log - expansion * * ( * * eq .
( [ eq : log - expansion])*)*. we derive filtration rules for words ( in general , @xmath217 is a sum of words ) by considering their dual elements in the universal enveloping algebra @xmath218 .
we introduce @xmath219 as the dual hopf algebra to @xmath169 in the next section .
section [ sec : renormalized - feynman - rules ] states and proves the two most important properties of renormalized feynman rules .
this also explains how the filtration of words works . in section [ sec : filtration - algorithm ] , we finally present a canonical filtration algorithm for arbitrary words and prove that it is free of redundancies .
let @xmath220 be a vector space over a field @xmath221 .
let furthermore @xmath222 : \mathcal{l } \otimes \mathcal{l } \to \mathcal{l}$ ] be a bi - linear map that fulfills @xmath223 = 0 $ ] as well as the jacobi - identity @xmath224 + [ y,[z , x ] ] + [ z,[x , y ] ] = 0\ ] ] @xmath225 .
@xmath226)$ ] ( or shortly @xmath220 ) is called a lie algebra and the bi - linear map @xmath222 $ ] ( called lie bracket ) is antisymmetric .
@xmath220 acquires a descending series of sub - algebras @xmath227 , where @xmath228 is generated by all @xmath229 $ ] with @xmath230 and @xmath231 . in the following ,
we denote all basis elements of some lie algebra @xmath220 that are not in @xmath232 by @xmath233 and ask for a complete basis in terms of these @xmath234 . for example , is @xmath235 $ ] or @xmath236 = -[x_1,x_2]$ ] a basis element of @xmath220 ? both are linearly dependent .
we therefore consider the hall basis @xcite , which exists for any lie algebra .
it requires a lexicographical ordering of all elements in @xmath220 , for example , let @xmath237 < \ldots$ ] ( it does not matter which ordering we take as long as we choose one ) .
we then define @xmath238 $ ] to be a ( hall ) basis element of @xmath220 iff both , 1 .
@xmath239 are ( hall ) basis elements with @xmath240 , 2 . if @xmath241 $ ] , then @xmath242 are fulfilled .
for example , @xmath235 $ ] , @xmath243 $ ] and @xmath244 $ ] are ( hall ) basis elements while @xmath236 $ ] , @xmath245 $ ] and @xmath246,x_3]$ ] etc .
are not .
the bracket in @xmath226)$ ] does not comprise an associative product .
however , one can construct enveloping algebras , i.e. an _ algebra _ @xmath247 such that there exists a homomorphism @xmath248 fulfilling @xmath249\textbf{)}\ ] ] @xmath250 .
there may be several enveloping algebras but we can always find a unique _ universal enveloping algebra _ @xmath251 up to isomorphism : for each enveloping algebra @xmath252 there exists a unique algebra homomorphism @xmath253 such that the following diagram @xmath254^{\rho_\mathcal{u } } \ar[dr]_{\rho_\mathcal{a } } & \mathcal{u}(\mathcal{l } ) \ar[d]^{\rho_{\mathcal{u } \to \mathcal{a } } } \\ & \mathcal{a}_\mathcal{l } } \end{xy } \end{split}\ ] ] commutes .
@xmath255 is unique * * ( * * assume that there are two universal enveloping algebras @xmath256 and @xmath257 , then the homomorphism @xmath258 turns out to be an isomorphism**)**. let us now construct the universal enveloping algebra for any lie algebra @xmath220 , which will prove its existence as well . consider the tensor algebra @xmath259 , where @xmath260 we define the sub - space @xmath261 as @xmath262 \right ) \otimes t| x , y \in \mathcal{l } ; \ , s , t \in t(\mathcal{l})\}\ ] ] ( @xmath263 is a 2-sided ideal ) and introduce equivalent classes of @xmath264 , @xmath265 = \left\ { s \in t(\mathcal{l } ) | s - t \in i \right\}.\ ] ] for example , @xmath266 \textbf { ] } = \ { [ x_1,x_2 ] , x_1 \otimes x_2 - x_2 \otimes x_1 , \ldots \}$ ] and @xmath267 = \ { x_1 \otimes x_2 \otimes x_3 , [ x_1,x_2 ] \otimes x_3 + x_2 \otimes x_1 \otimes x_3 , \ldots \}$ ] etc .
all such equivalent classes together form a vector space , denoted by @xmath268 .
the sum of two elements @xmath269,[t ] \in t(\mathcal{l})/i$ ] is well defined as @xmath270 $ ] .
we carefully abbreviate @xmath268 by @xmath255 and define an associative product @xmath271 acting on equivalent classes as @xmath272 \otimes [ t]\textbf { ) } : = [ s \otimes t].\ ] ] thus , @xmath273 \textbf{)}$ ] forms an algebra and together with the homomorphism @xmath274 defined as @xmath275 $ ] @xmath276 , it is an enveloping algebra * * ( * * eq .
( [ eq : enveloping - algebra ] ) holds**)**. finally , @xmath255 is even the universal enveloping algebra of @xmath220 , since @xmath264 fulfills the universality property in eq .
( [ eq : commuting - diagram ] ) as well @xcite .
it turns out that the universal enveloping algebra of a lie algebra acquires a hopf algebra structure . upon setting @xmath277)=[x]\otimes\one+\one\otimes [ x ]
\qquad \forall x \in \mathcal{l},\ ] ] it is determined from compatibility with the product @xmath278 . for the hopf algebra of words @xmath169 , there exists a lie algebra @xmath220 such that @xmath255 is dual to @xmath169 ( milnor - moore theorem @xcite ) . the indicated lie algebra
is easily constructed . for each letter @xmath279
, we name one respective element @xmath280 .
note that @xmath220 contains more elements than @xmath176 ( for example , there is no element @xmath281 $ ] in @xmath176 , but @xmath235
\in \mathcal{l}$ ] ) .
this will be crucial in the following .
duality between @xmath169 and @xmath255 allows us to uniquely define a linear and invertable map @xmath282 ( see section [ sec : hall - words ] ) , such that @xmath283 , \qquad \eta\textbf{(}\theta(a_i , a_j)\textbf { ) } = [ \theta(x_i , x_j ) ] , \qquad \eta ( a_i a_j ) = [ x_i \otimes x_j ] , \ldots\ ] ] @xmath284 . in general , concatenation of words is the dual operation of multiplication in @xmath255 .
we now give two crucial properties of renormalized feynman rules , which we need for our filtration method .
the proofs are collected below . 1 .
let @xmath285 and @xmath286\in \mathcal{u}(\mathcal{l})$ ] be its dual element @xmath287\textbf{)}$ ] .
if @xmath288 is also an element of @xmath289 , then renormalized feynman rules map @xmath9 to the @xmath2-linear part of the log - expansion in eq .
( [ eq : log - expansion ] ) , @xmath290 2 .
renormalized feynman rules are character - like , @xmath291 @xmath292 , where the dot on the rhs of eq .
( [ eq : state2 ] ) represents usual multiplication .
furthermore , @xmath293 is a period @xmath294 .
let us consider some examples before we turn to the proofs ( in the following , @xmath295 ) .
each letter @xmath296 has a dual element @xmath297\in \mathcal{u}(\mathcal{l})$ ] such that @xmath230 * * ( * * by construction of the dual elements , see eq .
( [ eq : dual - map])*)*. thus , @xmath298 and @xmath299 .
more interesting is the dual element of @xmath300 .
it is @xmath301 - [ x_j \otimes x_i ] = \textbf { [ } [ x_i , x_j ] \textbf{]}\ ] ] and @xmath302 \in \mathcal{l}$ ] .
thus , @xmath303 although @xmath304 also contains terms @xmath305 @xcite . finally , we _ filter _
the word @xmath306 to @xmath307 - \frac{1}{2 } \theta(a_i , a_j).\ ] ] here , we abbreviated @xmath308 by the concatenation commutator @xmath309 $ ] discussed in section ( [ sec : hall - words ] ) .
we treat concatenation ( multi-)commutators of letters as a letter itself , since the respective dual hopf algebra element is primitive .
note the information content of our filtration .
the first term of eq .
( [ eq : filtration - example ] ) maps to @xmath310 under renormalized feynman rules and eq .
( [ eq : state2 ] ) tells us how to determine this @xmath310-term simply by calculating @xmath311 and @xmath312 .
the other two terms of eq .
( [ eq : filtration - example ] ) map to @xmath2 .
the possibility to calculate the @xmath310-term in eq .
( [ eq : filtration - example ] ) out of the renormalized feynman amplitudes for @xmath313 and @xmath314 * * ( * * eq.([eq : state2 ] ) * ) * finally leads to the desired relations between next - to@xmath45-leading log orders and terms up to @xmath315 in the log - expansion * * ( * * eq .
( [ eq : log - expansion])*)*. we explore this in great detail in section [ sec : relat - betw - next ] . however , we first give the proofs for the necessary properties of feynman rules stated above in light in particular of the duality of @xmath169 and @xmath255 .
terms linear in @xmath2 can be interpreted in @xmath316 , and higher powers in @xmath2 reflect terms in the quotient algebra @xmath255 . as we stated before , renormalized
feynman rules map an element @xmath9 that is dual to a lie algebra element as above to the @xmath2-linear part of the log - expansion in eq .
( [ eq : log - expansion ] ) , @xmath317 see eq .
( [ eq : state1 ] ) .
this is a direct consequence of the renormalization group action on a single graph .
in fact , let @xmath129 be a hopf algebra element of fixed co - radical degree @xmath318 , @xmath319 .
then , it allows for an expansion @xmath320 by the renormalization group @xmath321 where @xmath322 is the function @xmath323 .
here , we identified the tensor - product of values with their product ( @xmath324 ) : @xmath325 note that @xmath326 is a symmetric function by construction .
this leads to a strict inequality on the co - radical degrees @xmath327 } < r_{\gamma_1}+r_{\gamma_2},\ ] ] which implies the result , eq .
( [ eq : state1 ] ) .
we also have that renormalized feynman rules are character - like , * * ( * * eq . ( [ eq : state2 ] ) * ) * , @xmath328 @xmath292 .
this is a direct consequence of chen s lemma @xcite in this context . to show that @xmath293 is a period @xmath294 is non - trivial only for letters in the image of @xmath177 .
therefore , it suffices to consider feynman graphs ( with fixed labels on their edges ) that are nested insertions of primitive graphs into each other : a hopf algebra element @xmath129 is a flag if there exists a sequence of primitive graphs @xmath329 , @xmath330 , with @xmath331 similarly , we say that a sum @xmath332 of @xmath318 flags @xmath333 , @xmath334 is a symmetrized flag if there exists a sequence of primitive graphs @xmath329 , @xmath330 , with @xmath335 the @xmath336-sum is over @xmath337 unsigned permutations . instead of having the full permutation group acting , one could also make do with permutations so as to make the rhs co - commutative , if so desired . for a given flag @xmath129 , and
hence given sequence of primitive graphs @xmath329 , @xmath330 , let @xmath338 be the cardinality of the set @xmath339 so @xmath340 .
a symmetrized flag is complete if @xmath341 in eq.([symmflag ] ) .
finally , renormalized feynman rules are a forest sum @xcite in graph polynomials @xmath342 ( see @xcite for notation ) : @xmath343 here , @xmath344 .
the coefficient @xmath345 of the @xmath2-linear term is @xmath346 if the renormalization point preserves scattering angles .
we then have the following result on the angle - independence of symmetrized flags : for any symmetrized flag @xmath332 , @xmath347 this justifies that @xmath348 is primitive in the hopf algebra of feynman graphs : to any set @xmath23 of letters , we can assign a unique complete symmetrized flag @xmath349 of feynman graphs corresponding to the letters in @xmath23 .
we set @xmath350 such that @xmath351 > from eq .
( [ eq : result - sym - flag ] ) follows that @xmath350 can indeed be regarded as a new letter in @xmath169 because it is independent of scattering angles by construction .
we now prove eq .
( [ eq : result - sym - flag ] ) : it follows immediately from writing eq.([linl ] ) as elementary symmetric polynomials in the variables @xmath352 , with @xmath353 ranging over all necessary forests and co - forests , which is possible precisely for symmetrized flags .
indeed , the denominator in eq.([linl ] ) @xmath354 is symmetric under exchange of @xmath355 , while in symmetrized flags , we also have co - commutativity which ensures symmetry under @xmath356 .
hence , in the sum for symmetrized flags , the factor @xmath357 in eq.([linl ] ) turns to unity .
this proves eq.([eq : result - sym - flag ] ) .
let us consider an example : have a look at the first two graphs in figure [ fig : graphs ] .
the graph @xmath149 has three vertices , the graph @xmath358 has five .
accordingly , there are three graphs @xmath359 , @xmath360 , obtained by replacing a vertex of @xmath149 by @xmath358 , and five graphs @xmath361 , @xmath362 , obtained by replacing a vertex of @xmath358 by @xmath149 .
we have the reduced co - products @xmath363 set @xmath364 we have @xmath365 , so @xmath366 is a symmetrized flag , and it is complete . then , @xmath367 using eq.([linl ] ) and the reduced co - products above , we indeed find that the second symanzik polynomial appearing in eq.([linl ] ) drops out in this co - commutative sum of symmetrized graph insertions @xmath368 which is an example of the above result ( see also @xcite ) .
it is not difficult to filter @xmath369 and @xmath370 in the solution of any dse * * ( * * as occurring in eq .
( [ eq : dse - solution])*)*. however , to filter higher order coefficients ( @xmath217 for @xmath371 ) requires a canonical algorithm that we give here .
consider for example @xmath372 for @xmath373 * * ( * * see eq .
( [ eq : dse - solution - yuk ] ) * ) * , @xmath374 we already filtered @xmath375 and @xmath376 without much effort but it is not obvious to see the filtration for @xmath377 .
the required filtration algorithm is the following loop over the length @xmath378 of occurring words : 1 .
bring all words with length @xmath378 into lexicographical order using the concatenation commutator ( respect the hall basis ) .
this introduces words with length @xmath379 , as we treat concatenation ( multi-)commutators as own letters .
repeat step 1 . for
the full shuffle products of the @xmath378 corresponding letters and insert them into the expression .
all words with length @xmath378 drop out , the introduced full shuffle products remain untouched in the remainder .
we start with the maximal length of occurring words down to @xmath380 .
hence , in the case of @xmath217 , we perform the above loop for @xmath381 .
let us illustrate this for @xmath377 .
the only word with length @xmath382 in eq .
( [ eq : w3-yuk ] ) is @xmath383 and it is already given in lexicographical order .
we calculate the corresponding full shuffle product @xmath384 and insert it into @xmath377 such that the word @xmath385 drops out .
hence , @xmath386 and we proceed with @xmath380 .
the term @xmath387 remains untouched during the rest of the filtration .
the words @xmath388 and @xmath389 are not in lexicographical order , we write @xmath390 - 2 a_1 \theta(a_1,a_1 ) - 3 [ a_1 , a_2 ] + 4 a_1a_2 - \frac{1}{2 } \theta(a_1 , a_1 , a_1 ) + \frac{1}{2 } a_1 \shuffle_\theta a_1 \shuffle_\theta a_1,\ ] ] where we only introduced hall basis elements ( @xmath391 $ ] instead of @xmath392 $ ] etc . ) .
the respective shuffle products are @xmath393 + \theta(a_1,a_2 ) , \\ \label{eq : sh - example2 } a_1 \shuffle_\theta \theta(a_1,a_1 ) = & 2 a_1 \theta(a_1,a_1 ) - [ a_1 , \theta(a_1,a_1 ) ] + \theta(a_1,a_1,a_1)\end{aligned}\ ] ] ( they are already brought into lexicographical order using the concatenation commutator ) . inserting eqs .
( [ eq : sh - example1 ] , [ eq : sh - example2 ] ) into eq .
( [ eq : w3-yuk2 ] ) finally results in @xmath394 - [ a_1 , a_2 ] + \frac{1}{2 } \theta(a_1 , a_1 , a_1 ) \nonumber\\ & - 2 \theta(a_1,a_2).\end{aligned}\ ] ] the explicit filtration algorithm is the basis for section [ sec : relat - betw - next ] .
there , we derive the relations for next - to@xmath45-leading log terms in the log expansion * * ( * * eq .
( [ eq : log - expansion])*)*. however , we first give some basics to hall words and concatenation ( multi- ) commutator letters .
this explains , why the filtration algorithm described above is redundancy - free . in the hopf algebra of feynman graphs @xmath38
, any @xmath395 evaluates to @xmath396 . here , @xmath397 and the @xmath398
are given through eq.([rge ] ) .
in particular , this amounts for @xmath399 to an evaluation of products of primitive elements . through @xmath201
, we inherit the same properties for words @xcite : any @xmath285 evaluates to @xmath400 , where @xmath401 and @xmath402 in particular , this amounts for @xmath403 to an evaluation of products of letters .
note that @xmath201 preserves the co - radical degree .
the above filtration algorithm answers the question how to obtain non - leading logs , @xmath404 , from the letters that constitute @xmath9 . for this
, we first have to consider the lower central series filtration @xmath405 of the lie algebra @xmath220 , @xmath406 and its associated grading .
secondly , we have to consider the filtrations and gradings of the universal enveloping algebra : @xmath407 by its augmentation and @xmath408 by its co - radical .
we will use that @xmath409 is isomorphic to the @xmath378-fold symmetric tensor - power of @xmath220 by the poincar birkoff
witt theorem : @xmath410 let @xmath411 as before .
we have the commutative diagrams @xmath412^{\shuffle_\theta } \ar[d]_{\eta\otimes\eta } & \mathcal{h}_w \ar[d]^{\eta } \\ \mathcal{u}_{\mathcal{l}}\otimes \mathcal{u}_{\mathcal{l } } \ar[r]^ { m_{\mathcal{u } } } & \mathcal{u}_{\mathcal{l } } } \end{xy } \qquad\qquad\qquad \begin{xy } \xymatrix { \mathcal{h}_w\otimes \mathcal{h}_w \ar[d]_{\eta\otimes\eta } & \mathcal{h}_w\ar[l]^{\delta } \ar[d]^{\eta } \\ \mathcal{u}_{\mathcal{l}}\otimes \mathcal{u}_{\mathcal{l } } & \mathcal{u}_{\mathcal{l}}\ar[l]^ { \delta_{\mathcal{u } } } } \end{xy } \end{split}\ ] ] and two more , by replacing @xmath413 and downward pointing arrows by upward pointing ones .
these determine the action of @xmath414 once we have defined it on letters @xmath415 * * ( * * @xmath416 $ ] , @xmath417\ldots$]**)**. for example , the degree two image @xmath418 of the word @xmath419 with respect to @xmath407 is @xmath420 $ ] , with @xmath421 $ ] .
we define feynman rules for elements @xmath269 \in \mathcal{u}_{\mathcal{l}}$ ] by @xmath422 \to \psi_r\textbf{(}\eta^{-1}([s])\textbf{)}.\ ] ] in particular for homogeneous elements @xmath269 \in \mathfrak{gr}_j ( \mathcal{u}_\mathcal{l})$ ] we have @xmath423)\sim l^j$ ] , by construction .
_ the structure of renormalized feynman rules then allows us to regain the above filtration algorithm on words @xmath424 as _
@xmath425 here , @xmath426 is the projection into the grade @xmath7 piece , @xmath427 .
@xmath428 is the map @xmath429\to \eta\circ\theta\textbf{(}\eta^{-1}([x_1]),\ldots,\eta^{-1}([x_j])\textbf{)}$ ] , where the @xmath430)\in \mathcal{h}_l$ ] are letters by construction ( @xmath431 ) . the lower central series filtration @xmath406 filters in particular @xmath432 .
thus , using the hall basis of @xmath218 and the invertibility of @xmath414 finally allows us to write the filtration algorithm in the word algebra @xmath169 with concatenation multi - commutators .
indeed , the rhs of eq.([strucfil ] ) is of degree one by construction as it is in the image of @xmath428 .
this suffices , as the degree-@xmath7 piece is a product of the corresponding @xmath7 degree - one pieces obtained in @xmath433 .
let us consider an example .
words on three letters @xmath434 have a hall basis , which _ for their degree one part _ can be written in @xmath435),\quad y_3:=\theta(a_2,[a_1,a_3]),\quad y_4:=\theta(a_3,[a_1,a_2 ] ) , \nonumber\\ y_5:=[a_2,[a_1,a_3]],\quad y_6:=[a_3,[a_1,a_2]]\}.\end{gathered}\ ] ] in degree one this is the inverse image @xmath436 of the elements @xmath437 , x_2[x_1,x_3 ] , x_3[x_1,x_2],[x_2,[x_1,x_3 ] ] , [ x_3,[x_1,x_2]]\}\ ] ] in @xmath219 written in hall basis notation ( ordered and omitting the symmetric tensor product ) . these form a standard hall basis on three ` letters ' @xmath438 dual to @xmath434 in @xmath219 .
there are six words on three distinct letters @xmath434 . for their degree one
part , these can be written in the basis above : @xmath439 inverting these equations expresses the degree one elements @xmath440 through the six words on the left .
these correspond to linear combinations of feynman graphs @xmath441 , @xmath442 which map under @xmath201 to the corresponding words .
for example , @xmath443)=\eta^{-1}p_1\eta(a_1a_2a_3-a_1a_3a_2+a_2a_3a_1-a_3a_2a_1),\ ] ] with @xmath444 furthermore @xmath445 so there is no term @xmath446 , whilst the term in @xmath310 is @xmath447)-\frac{1}{2}\left(\psi_r^1(a_2)\psi_r^1([a_1,a_3 ] ) + \psi_r^1(a_3)\psi_r^1([a_2,a_1])\right).\ ] ] this gives us a definition in terms of feynman diagrams for @xmath448 which defines a @xmath2-linear term .
let us now describe the standard hall basis for a set of words on @xmath449 distinct letters in general .
the case of repeated letters follows easily .
there are @xmath450 words we can form .
first , we count with the help of the mbius function @xmath451 the number of available concatenation multi - commutators @xmath452 .
@xmath453 with @xmath454 the well - known number of multi - commutators of degree @xmath449 on an alphabet of size @xmath7 .
let @xmath455 be the set of partitions @xmath456 of the integer @xmath449 with the following properties : @xmath457 we allow for at most one such @xmath458 to be marked , which we indicate as @xmath459 .
furthermore , if @xmath460 , @xmath461 must be marked .
if @xmath456 contains such a distinguished marked @xmath462 , we say @xmath456 is marked at @xmath463 , else it is unmarked .
the marking reflects the fact that @xmath464 $ ] is distinguished amongst all @xmath465 .
we say that a partition @xmath466 of a set @xmath467 of @xmath449 letters is compatible with the partition @xmath456 of the integer @xmath449 , if it is a disjoint union of sets @xmath468 of @xmath458 letters according to @xmath456 .
letters in @xmath469 are completely symmetrized , while all other sit in multi - commutators of degree @xmath470 .
assume @xmath471 is unmarked .
then we assign a set of letters @xmath472 on @xmath378 multi - commutators @xmath473 on letters @xmath474 to it .
if @xmath456 is marked at @xmath463 , we assign a set of letters @xmath475 then , summing over all partitions @xmath456 and all partitions of letters @xmath466 compatible with it , we get @xmath450 different words which form a base for the degree one elements of @xmath218 . if a partition @xmath456 contains an unmarked integer @xmath458 say @xmath476 times , the symmetry factor @xmath477 of @xmath456 is @xmath478 .
then , we indeed count @xmath479 where @xmath480 if @xmath456 is not marked at @xmath463 , and @xmath481 if it is marked at @xmath463 .
we complete this section by giving a final example . for four distinct letters
, we can have the partition @xmath482 with @xmath456 unmarked .
so it will provide six elements in @xmath483 .
for the partition @xmath484 we have @xmath485 possibilities to choose three letters for @xmath486 which itself has two elements , whilst the fourth letter belongs to @xmath487 . for the partition @xmath488
we have @xmath489 possibilities to choose two letters for a one - element @xmath490 while the other two letters constitute @xmath491 .
for @xmath492 we get a non - trivial symmetry factor and have @xmath493 possibilities to form the product @xmath494 .
finally , we have the partition @xmath495 .
this gives a single element - the symmetric sum over all permutations of four letters .
counting , we get @xmath496 all such words are independent by construction and using eq.([mcs ] ) repeatedly there are @xmath450 of them .
they hence form a base .
we now present the main result of our work : how to write the next - to@xmath45-leading log order as a function of terms up to @xmath497 in the log - expansion * * ( * * eq .
( [ eq : log - expansion ] ) * ) * ? we first introduce a convenient notation for multiplicities of full shuffle products in a filtered word .
secondly , we derive generating functions for these multiplicities using the example of the yukawa propagator .
we derive the generating functions for the qed photon self - energy in appendix [ sec : qed - results ] .
we represent the multiplicity of shuffle products in a filtered word by @xmath92$]-bracketed matrices @xmath119 , for example @xmath498 each matrix denotes a number , in our case @xmath499 , @xmath500 * * ( * * see eq .
( [ eq : w2-yuk])*)*. in the following we say that a matrix @xmath119 _ belongs to _ a shuffle product @xmath23 , when it gives the multiplicity of @xmath23 in a filtered word .
we also say that @xmath23 is the _ respective shuffle product _ to @xmath119 .
each matrix @xmath119 with corresponding shuffle product @xmath23 is built as follows : the first row contains the numbers of letters @xmath501 , @xmath502,@xmath503 in @xmath23 .
the other rows represent one letter @xmath504 in @xmath23 each , s.t .
@xmath503 contains @xmath505 letters @xmath314 for the @xmath463-th row .
for example , the filtered word @xmath506 contains the term , @xmath507 we index matrices @xmath119 if the respective shuffle @xmath23 contains ( multi-)commutator letters @xmath222 $ ] by the commutator letters themselves . for example , the filtered word @xmath508 contains the term @xmath509 } a_1 \shuffle_\theta \theta(a_1,a_1)\shuffle_\theta [ a_1,a_2 ] + \ldots\ ] ] we treat indexed matrices separately in section [ sec : commutators ] .
first though , we only treat index - free matrices and hence , full shuffle products _ without ( multi-)commutator letters_. now , consider an _ unfiltered word _ @xmath217 that is recursively defined via a dse . in our case of the yukawa propagator ,
this is eq .
( [ eq : dse - solution - yuk ] ) .
a matrix with @xmath510 brackets represents the number of words in @xmath217 that consists of a given set of letters .
the matrix entries encode the particular set of letters in full analogy to the case of @xmath92 $ ] brackets above .
for example , @xmath511 represents the number of words with four letters @xmath501 , one @xmath502 , one @xmath512 and one @xmath513 in the unfiltered word @xmath217 .
as another example , @xmath377 contains the terms @xmath514 and @xmath515 * * ( * * see eq .
( [ eq : w3-yuk ] ) * ) * , hence @xmath516 .
we always represent a @xmath92$]-bracketed matrix by a lower case letter and the same matrix with @xmath510 brackets by the corresponding capital letter ( @xmath517 ) .
the defined matrices can have any sizes .
filling zeros does not change the multiplicity of the corresponding shuffle product .
we call a row ` @xmath177-row ' when it is not the first row of a matrix .
we consider two matrices with two @xmath177-rows interchanged to be the same object , since they represent the same number ( @xmath179 is symmetric ) .
for example @xmath518 let two matrices @xmath519 , @xmath520 with corresponding shuffle products @xmath521 and @xmath522 be given .
we define the matrix @xmath523 such that it belongs to the shuffle product @xmath524 .
this defines a special summation of matrices , denoted by @xmath525 .
this is realized as follows : @xmath525 adds up each first row as in ordinary matrix summation and writes the @xmath177-rows one below the other .
we can even add matrices with different sizes by filling zeros .
for example , @xmath526 we define matrices @xmath527 as well as the useful vectors @xmath528 with appropriate sizes .
we use these vectors to notate some properties of matrices .
we calculate products of matrices and vectors by ordinary matrix multiplication and not by replacing the matrix by the corresponding number . for example , a matrix @xmath119 that occurs in the filtered word @xmath217 of a dse fulfills @xmath529 .
let @xmath530 be the number of nonzero @xmath177-rows .
then , we define @xmath531 which is the number of letters in the respective shuffle product .
note that once , a matrix representing a number is contracted with a vector , the result is regarded as a real vector , @xmath532 we introduce a function @xmath533 acting on two matrices , say @xmath519 and @xmath520 with respective shuffle products @xmath521 and @xmath522 .
@xmath534 counts the number of combinatorial possibilities to work out some shuffle products in @xmath522 such that the resulting expression contains one term @xmath521 .
for example , @xmath535 we comment on the second example : the corresponding shuffle product of @xmath536 is @xmath537 and there are 4 possibilities to work out one of the shuffle products to derive @xmath538 the last term on the rhs is the respective shuffle product of @xmath539 .
we relate @xmath510-bracketed matrices to @xmath92$]-bracketed matrices .
consider for example @xmath540 and the respective shuffle product @xmath541 .
computing @xmath23 gives @xmath542 words that consists of the letters @xmath501 , @xmath502 and @xmath543 .
ordering these words using the filtration algorithm , @xmath544\ ] ] does not change the overall number of words with letters @xmath501 , @xmath502 and @xmath543 .
it only introduces new words with commutator letters .
we have thus , @xmath545 in general , @xmath546 where the sum is over all possible matrices @xmath547 . if @xmath119 has only one row , @xmath548 if @xmath549 and @xmath550 otherwise . eq .
( [ eq : curly - square-2 ] ) then reads @xmath551 , as in eq .
( [ eq : curly - square-1 ] ) .
if @xmath119 consists of more than one row , the reader may verify that indeed , eq . ( [ eq : curly - square-2 ] ) is the correct generalization . note that eq .
( [ eq : curly - square-2 ] ) only holds for index - free matrices .
indeed , the number of words that consist of a certain ( commutator - free ) set of letters does not change when introducing concatenation commutators during the filtration algorithm .
we introduce generating functions for the multiplicities ( matrices ) @xmath119 .
let @xmath8 be a matrix with integer entries except for the upper left entry , which is just a dot .
a matrix @xmath119 is said to be equivalent to @xmath8 , @xmath552 , if replacing the upper left entry to a dot yields @xmath8 .
we define @xmath553 @xmath554 is a function in @xmath555 , represented by a matrix .
it generates all @xmath552 , @xmath556 examples of generating functions are @xmath557 we translate two properties of matrices @xmath119 to generating functions @xmath8 .
first , we sum generating functions @xmath8 in the same special way as matrices @xmath119 .
the dot remains untouched , for example @xmath558 secondly , let @xmath559 with @xmath560 .
we define @xmath561 .
summation and absolute value of generating functions will be useful in section [ sec : index - free - n0 ] .
we find another relation between a @xmath510-bracketed matrix @xmath562 and @xmath92$]-bracketed matrices by use of the recursive equation eq .
( [ eq : dse - solution - yuk ] ) .
let the lhs of eq .
( [ eq : dse - solution - yuk ] ) be an unfiltered word @xmath217 and let the words @xmath563 on the rhs be filtered words .
let @xmath562 be the number of words consisting of a certain set of letters on the lhs , a @xmath510-bracketed matrix hence
. on the rhs , only terms that constitute these words are taken into account .
note that we have @xmath564 .
the recursive relation for @xmath562 is @xmath565 where @xmath119 is still , the same matrix as @xmath562 but with @xmath92 $ ] brackets .
@xmath566 sums integers @xmath567 as in eq .
( [ eq : dse - solution - yuk ] ) and matrices @xmath568 such that @xmath569 we explain the different terms in eq .
( [ eq : curly - square-3 ] ) individually .
first , @xmath570 corresponds to the @xmath202 term in eq .
( [ eq : dse - solution - yuk ] ) .
the integers @xmath7 and @xmath378 range over the same numbers as in eq .
( [ eq : dse - solution - yuk ] ) .
the term @xmath571 gives 1(0 ) if the respective words to @xmath562 do ( not ) contain the letter @xmath314 .
only if they do contain @xmath314 , they may arise from the term @xmath572 in eq .
( [ eq : dse - solution - yuk ] ) .
since we introduce filtered words @xmath563 into the rhs in eq .
( [ eq : dse - solution - yuk ] ) , we obtain expressions @xmath573 , where @xmath23 is a full shuffle product .
@xmath23 is built out of @xmath378 shuffles , namely terms in @xmath574 , @xmath575 , @xmath503 , @xmath576 .
we therefore claim that @xmath23 is the respective shuffle product to @xmath577 .
each matrix @xmath568 belongs to a full shuffle product in @xmath563 .
condition @xmath566 in eq .
( [ eq : star ] ) consists of two parts .
the first three relations correspond to the third sum in eq .
( [ eq : dse - solution - yuk ] ) . the fourth equation together with the factor @xmath578 in eq .
( [ eq : curly - square-3 ] ) ensure that the letters in @xmath23 together with the letter @xmath314 ( matrix @xmath470 ) constitute the set of letters in @xmath562 .
@xmath23 consists of @xmath579 letters , which gives rise to the factor of @xmath580 in eq .
( [ eq : curly - square-3 ] ) .
we now derive an inhomogeneous linear differential equation for the corresponding generating function to @xmath119 , i.e. @xmath554 * * ( * * see eq .
( [ eq : def - generating - function])*)*. therefore , inserting eq .
( [ eq : curly - square-2 ] ) into eq .
( [ eq : curly - square-3 ] ) yields @xmath581 the final step is to multiply with @xmath582 and to sum over all matrices that are equivalent to @xmath119 .
_ this gives the master differential equation .
_ indeed , on the lhs we obtain @xmath583 , @xmath584 see eq .
( [ eq : star ] ) for the summation @xmath566 of integers @xmath567 and matrices @xmath568 . from eq .
( [ eq : def - generating - function ] ) , we read off the initial condition @xmath585 let us consider a first example : for @xmath586 , the first term in eq .
( [ eq : master ] ) vanishes .
the last term only gives a non - zero value for @xmath106 .
@xmath587 , @xmath588 .
the @xmath533 function gives 1 for @xmath589 and 0 otherwise .
together with the initial condition in eq .
( [ eq : initial - condition ] ) , we find @xmath590 we can now derive the homogeneous part of the differential master equation eq .
( [ eq : master ] ) . on the rhs
, the only terms including the function @xmath554 itself occur in the sum for @xmath106 , when @xmath379 of the matrices @xmath568 are equivalent to @xmath591 and the @xmath378-th matrix is equivalent to @xmath119 . using eq .
( [ eq : bullet1 ] ) , we obtain @xmath592 hence , the differential equation eq .
( [ eq : master ] ) reduces to an integration using the ansatz @xmath593 we read off the initial condition for @xmath594 from eq .
( [ eq : initial - condition ] ) .
in particular , the integration is @xmath595 where we obtain @xmath596 from eq .
( [ eq : master ] ) .
@xmath597 is an appropriate integration constant such that @xmath598 .
further general simplifications of the differential master equation are not obvious .
the problem is that the functions @xmath533 in eq .
( [ eq : master ] ) give individual numbers that do not generalize and so have to be worked out case - by - case .
they result in an overall differential operator acting on whatever follows .
we demonstrate this in several examples , which will give us next - to and next - to - next - to - leading log generating functions .
one exception is the case that @xmath8 only contains one row , @xmath599 .
these generate the matrices @xmath119 that belong to full shuffle products @xmath23 without @xmath600 letters . `
index - free ' means that @xmath23 also does not contain @xmath222 $ ] letters . here
, @xmath601 reduces to 1 if @xmath602 and to 0 otherwise .
thus , the first term in eq .
( [ eq : master ] ) is zero .
the other @xmath533 term constrains @xmath603 hence , @xmath604 .
we denote the generating functions of @xmath568 and @xmath470 by @xmath605 and @xmath606 to be consistent .
we obtain @xmath607 where @xmath608 sums the generating functions @xmath609 , @xmath503 , @xmath610 such that @xmath611 in the following , we give some examples that constitute the next - to@xmath45-leading log expansions .
the first example is @xmath613 .
the respective shuffle products to the matrices @xmath614 are @xmath615 for @xmath616 .
the sum in eq .
( [ eq : master - special ] ) only gives non - zero values if @xmath617 or @xmath618 .
for @xmath617 , @xmath619 and eq .
( [ eq : starstar ] ) is only fulfilled if one @xmath605 matches @xmath8 .
this part belongs to the homogeneous differential equation . for @xmath618 , @xmath620
and we find @xmath621 @xmath587 from eq .
( [ eq : starstar ] ) .
thus , the inhomogeneous part of eq .
( [ eq : master - special ] ) reads @xmath622 we used eq .
( [ eq : bullet1 ] ) in the last line .
we insert this result into eq .
( [ eq : generating - function - final ] ) and obtain @xmath623 matrices @xmath625 belong to the shuffle products @xmath626 for @xmath627 .
the sum in eq .
( [ eq : master - special ] ) gives non - zero values only if @xmath617 or @xmath628 .
as in the previous example , the @xmath617 part belongs to the homogeneous differential equation . for @xmath628 , @xmath629 and we find @xmath621 @xmath587 from eq .
( [ eq : starstar ] ) .
thus , the inhomogeneous part of eq .
( [ eq : master - special ] ) reads @xmath630 we insert this result into eq .
( [ eq : generating - function - final ] ) and obtain @xmath631 @xmath633 is the final next - to - next - to - leading log generating function for matrices @xmath634 with @xmath599 .
the matrices @xmath119 belong to the shuffle products @xmath635 for @xmath627 .
thus , the sum in eq .
( [ eq : master - special ] ) gives non - zero values only if @xmath617 or @xmath618 .
if @xmath617 , @xmath636 .
there are two possibilities to choose the matrices @xmath605 such that eq .
( [ eq : starstar ] ) is fulfilled .
however , only one of these belongs to the inhomogeneous part of the differential equation : for two integers @xmath637 , @xmath638 and for all other @xmath463 , @xmath639 . if @xmath618 , eq .
( [ eq : starstar ] ) is only satisfied if one of the matrices @xmath605 is @xmath640 and the others are equal to @xmath591 .
we finally find the inhomogeneous part of eq .
( [ eq : master - special ] ) , @xmath641 we used the previous results in eqs .
( [ eq : bullet1],[eq : bullet2 ] ) in the last line . inserting this into eq .
( [ eq : generating - function - final ] ) and performing the integration finally results in @xmath642 the respective full shuffle products of index - free matrices @xmath119 with @xmath643 contain at least one letter @xmath600 but no @xmath222 $ ] letters . here
, we have to proceed from the master differential equation eq .
( [ eq : master ] ) to obtain @xmath596 in eq .
( [ eq : generating - function - final ] ) .
we treat the different next - to@xmath45-leading log generating functions in separate subsections .
the first example is the next - to - leading log generating function @xmath645 .
matrices @xmath634 belong to the shuffle products @xmath646 for @xmath647 . in eq .
( [ eq : master ] ) , we replace the sum over @xmath634 by a sum over @xmath647 such that @xmath648 . in the first term on the rhs ,
only @xmath649 yields a non - vanishing function @xmath533 .
in particular , @xmath650 the other sum only survives if @xmath106 and hence @xmath651 .
the integers @xmath567 in the @xmath566-sum range such that @xmath652 * * ( * * see eq .
( [ eq : star])*)*. furthermore , in the first argument of the function @xmath533 , @xmath653 . for the inhomogeneous part of the differential equation
, this implies that @xmath654 @xmath587 such that @xmath655 .
then , @xmath656 all together , we obtain @xmath657 in the third line , we used the explicit expression for the generating function @xmath591 * * ( * * eq .
( [ eq : bullet1])*)*. inserting this result into eq .
( [ eq : generating - function - final ] ) , we finally obtain @xmath658 @xmath640 * * ( * * eq . ( [ eq : bullet2 ] ) * ) * and @xmath659 are the only necessary generating functions to derive relations for the next - to - leading log order .
it is surprising that they are related by a factor of @xmath660 , @xmath661 now , consider the next - to - next - to - leading log generating function @xmath663 .
matrices @xmath634 belong to the shuffle products @xmath664 for @xmath647 . again in eq .
( [ eq : master ] ) , we replace the sum over @xmath634 by a sum over @xmath647 such that @xmath665 .
the first term consists of two parts , @xmath666 and @xmath667 .
for other @xmath547 , @xmath601 vanishes .
in particular , @xmath668 the sum over @xmath7 only gives a non - zero value for @xmath106 .
hence @xmath669 and the integers @xmath567 in the @xmath566-sum require @xmath670 * * ( * * see eq .
( [ eq : star])*)*. the first argument of the function @xmath533 becomes @xmath671 . for the inhomogeneous part of the differential equation , this implies that either @xmath672 or @xmath673 . the first case is realized if @xmath654 @xmath587 .
the second case implies that one of the matrices @xmath568 is equal to @xmath674 and the rest of the matrices @xmath654 .
we compute @xmath675 we use all these observations to obtain the inhomogeneous part of the differential equation eq .
( [ eq : master ] ) , @xmath676 in the last line , we used the explicit formulas for the generating functions @xmath591 and @xmath659 * * ( * * eqs .
( [ eq : bullet1],[eq : bullet5])*)*. inserting eq .
( [ eq : bullet6-pre ] ) into the integration in eq .
( [ eq : generating - function - final ] ) finally results in @xmath677 now , consider @xmath679 .
the respective shuffle products to the matrices @xmath634 are @xmath680 for @xmath647 . in eq .
( [ eq : master ] ) , we replace the sum over @xmath634 by a sum over @xmath647 such that @xmath681 . here
, only @xmath682 contributes to the first sum and we calculate @xmath683 in the second sum , @xmath106 and @xmath651 .
the integers @xmath567 in the @xmath566-sum are constraint by @xmath684 * * ( * * see eq .
( [ eq : star])*)*. the first argument of the function @xmath533 becomes @xmath685 .
the second argument must be @xmath686 since we only consider the inhomogeneous part of the differential equation .
thus , one of the matrices @xmath568 is equal to @xmath687 and the rest of the matrices @xmath654 .
in particular , @xmath688 we deduce the inhomogeneous part of the differential equation eq .
( [ eq : master ] ) , @xmath689 in the last line , we used the explicit formulas for the generating functions @xmath591 and @xmath640 * * ( * * eqs .
( [ eq : bullet1],[eq : bullet2])*)*. we plug eq .
( [ eq : bullet7-pre ] ) into eq .
( [ eq : generating - function - final ] ) and obtain @xmath690 the function @xmath692 generates the rationals @xmath552 to the respective shuffle products @xmath693 for @xmath647 .
consider eq .
( [ eq : master ] ) .
we replace the sum over @xmath634 by a sum over @xmath647 such that @xmath694 . here , only the matrices @xmath695 and @xmath696 contribute to the first sum and we find @xmath697 in the second sum , @xmath106 and @xmath651 as before .
the integers @xmath567 in the @xmath566-sum range over @xmath698 * * ( * * see eq .
( [ eq : star])*)*. the first argument of the function @xmath533 is @xmath699 . here , there are three possibilities for @xmath700 in the second sum of eq .
( [ eq : master ] ) .
first , @xmath701 and @xmath654 @xmath587 .
secondly , @xmath702 , which implies that one of the matrices @xmath568 is equal to @xmath674 and the rest of the matrices @xmath654 . in the third case , @xmath703 .
note that @xmath704 did not occur in the previous examples because it was part of the homogeneous differential equation eq .
( [ eq : master ] ) . here , we realize @xmath703 within the inhomogeneous part of the differential equation : two of the matrices @xmath568 are equal to @xmath674 and the rest of the matrices @xmath654 . the corresponding functions @xmath578 give @xmath705 we simplify the inhomogeneous part of the differential equation eq .
( [ eq : master ] ) as follows : @xmath706 we use the previous results in eqs . ( [ eq : bullet1],[eq : bullet5 ] ) and insert the resulting expression into eq .
( [ eq : generating - function - final ] ) .
a little calculation yields @xmath707 the only next - to - next - to - leading log generating function left is @xmath709 .
it generates the rationals @xmath552 with the respective shuffle products @xmath710 for @xmath647 . in eq .
( [ eq : master ] ) , we replace the sum over @xmath634 by a sum over @xmath647 such that @xmath711 .
the function @xmath533 in the first sum vanishes except for @xmath712 : @xmath713 in the second sum , either @xmath617 or @xmath618 .
this is the main difference to the previous examples .
the sum does not vanish for @xmath714 because @xmath715 . for @xmath106 , @xmath651 as before .
then , the integers @xmath567 in the @xmath566-sum range over @xmath698 * * ( * * see eq .
( [ eq : star])*)*. furthermore , @xmath716 in the first argument of the function @xmath533 , which implies that either @xmath717 or @xmath718 . in the first case ,
there is one @xmath637 such that @xmath719 and for all other @xmath637 , @xmath654 . in the second case ,
one of the matrices @xmath568 must be equal to @xmath720 and another one must be equal to @xmath674 . the rest of the matrices @xmath654 .
the corresponding @xmath533-function factors are @xmath721 for @xmath714 , @xmath722 .
the integers @xmath567 in the @xmath566-sum then require @xmath723 . furthermore , @xmath724 in the first argument of the function @xmath533 .
thus , either @xmath672 or @xmath673 . in the former case , @xmath654 @xmath587 . in the latter case ,
one of the matrices @xmath568 must be equal to @xmath674 and the other matrices @xmath654 .
the corresponding functions @xmath578 give @xmath725 using these observations yields the inhomogeneous part of the differential equation eq .
( [ eq : master ] ) , @xmath726 we insert this into the integration in eq .
( [ eq : generating - function - final ] ) .
we also need eqs .
( [ eq : bullet1],[eq : bullet2],[eq : bullet5 ] ) and after some calculation , we obtain @xmath727 in this paper , we do not give a general method to obtain the generating functions for indexed matrices .
in this section however , we derive the two ` indexed generating functions ' @xmath728 } = \sum_{n\geq0 } \begin{bmatrix } n \end{bmatrix } _ { [ a_1,a_2 ] } z^{n+1 } , \qquad \begin{bmatrix}\bullet\end{bmatrix}_{[a_1,\theta(a_1,a_1 ) ] } = \sum_{n\geq0 } \begin{bmatrix } n \end{bmatrix } _ { [ a_1,\theta(a_1,a_1 ) ] } z^{n+1}.\ ] ] these generate the indexed matrices @xmath729}$ ] and @xmath730}$ ] that belong to the full shuffle products @xmath731,\qquad a_1^{\shuffle_\theta\,n } \shuffle_\theta[a_1,\theta(a_1,a_1)]\ ] ] respectively .
these shuffles make part of the filtered words @xmath732 * * ( * * eq . ( [ eq : dse - solution - yuk ] ) * ) * and map to the next - to - next - to - leading log order in the log - expansion .
in particular , @xmath733}$ ] and @xmath734}$ ] complete the set of generating functions that are necessary to obtain the next - to - next - to - leading log expansion .
we will work on a general method to derive indexed generating functions in future work .
only words with @xmath736 letters @xmath501 and one letter @xmath502 in the _ unfiltered word @xmath737 _ can contribute to the term @xmath738 } a_1^{\shuffle_\theta\,n-3 } \shuffle_\theta[a_1,a_2]\ ] ] in the _ filtered word @xmath737_. consider eq .
( [ eq : dse - solution - yuk ] ) and let the words @xmath563 on the rhs be filtered .
then , @xmath739 } a_1^{\shuffle_\theta\,t_k-3 } \shuffle_\theta [ a_1,a_2 ] \right ) \nonumber\\ & + b_+^{a_1 } \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } k \begin{bmatrix } t_1 \end{bmatrix } a_1^{\shuffle_\theta\,t_1 } \shuffle_\theta \ldots \shuffle_\theta \begin{bmatrix } t_{k-1 } \end{bmatrix } a_1^{\shuffle_\theta\,t_{k-1 } } \shuffle_\theta \begin{bmatrix } t_k-2 & 1 \end{bmatrix } a_1^{\shuffle_\theta\,t_k-2 } \shuffle_\theta a_2 \nonumber\\ & + b_+^{a_2 } \sum_{k\geq1 } \begin{pmatrix } 2 + k \\ k \end{pmatrix } \sum_{\sum_i t_i = n-2 } \begin{bmatrix } t_1 \end{bmatrix } a_1^{\shuffle_\theta\,t_1 } \shuffle_\theta \ldots \shuffle_\theta \begin{bmatrix } t_k \end{bmatrix } a_1^{\shuffle_\theta\,t_k } + \ldots,\end{aligned}\ ] ] where the dots represent all missing terms in eq .
( [ eq : dse - solution - yuk ] ) , for example @xmath740 } a_1^{\shuffle_\theta\,t_k-4 } \shuffle_\theta [ a_1,[a_1,a_2 ] ] \right).\ ] ] all these other terms do not contribute to @xmath741 } a_1^{\shuffle_\theta\,n-3 } \shuffle_\theta[a_1,a_2]$ ] in the filtered word .
indeed , the only missing terms in eq .
( [ eq : wn - contribution-1 ] ) that give words with @xmath736 letters @xmath501 and one @xmath502 include one of the multi - commutator letters @xmath742\ldots]$ ] in the shuffle product .
computing all these shuffles and filtrating the resulting words will not give words with @xmath743 letters @xmath501 and one @xmath744 $ ] but words with multi - commutator letters .
given a filtered word @xmath737 , one regains the original unfiltered word by _ first computing all shuffle products in @xmath737 and then , computing all ( multi-)commutators_. this can be seen by a look at the filtration algorithm in section [ sec : filtration - algorithm ] .
for the first term in eq .
( [ eq : wn - contribution-1 ] ) , this implies that we must first compute the shuffle products in the bracket .
secondly , we have to replace the commutator letter @xmath744 $ ] by @xmath745 and finally , we must compute all remaining shuffle products to obtain the respective terms in the unfiltered word @xmath737 * * ( * * eq .
( [ eq : wn - contribution-1])*)*. let us calculate the unfiltered @xmath737 .
we are only interested in words that contribute to @xmath733}$ ] .
hence , when computing the shuffle products in eq .
( [ eq : wn - contribution-1 ] ) , we shift all words with @xmath600 letters to the @xmath503 terms . in the following ,
it is convenient to define the words @xmath746 \underbrace{a_1\ldots a_1}_{q\times}\end{aligned}\ ] ] for @xmath747 .
we note that @xmath748 = & p !
\textbf{(}a(p+1,0 ) - a(0,p+1)\textbf { ) } + \ldots , \\ a_1^{\shuffle_\theta p}\shuffle_\theta a(q , 0 ) = & p!\sum_{r = 0}^{p } \begin{pmatrix } q + r \\ r \end{pmatrix } a(q + r , p - r ) + \ldots,\\ a_1^{\shuffle_\theta p}\shuffle_\theta a(0 , q ) = & p!\sum_{r = 0}^{p } \begin{pmatrix } q + r \\
r \end{pmatrix } a ( p - r , q + r ) + \ldots , \\
a_1^{\shuffle_\theta p } \shuffle_\theta a_2 = & p ! \sum_{q = 0}^{p } a(q , p - q ) + \ldots\end{aligned}\ ] ] the dots collect all words that consist of other letters than one @xmath502 and some @xmath501 . using these relations in eq .
( [ eq : wn - contribution-1 ] ) , we obtain @xmath749 } ( t_k - 3 ) ! ( n-1-t_k ) !
\sum_{p=0}^{n-1-t_k } \begin{pmatrix } t_k - 2 + p \\
p \end{pmatrix } \nonumber\\ & \times \textbf { ( } a(t_k - 1 + p , n - 1 - t_k - p ) - a ( n - t_k - p , t_k - 2 + p ) \textbf{)}\nonumber\\ & + \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } \begin{bmatrix } t_k-2 & 1 \end{bmatrix } ( n-3 ) !
\sum_{p=0}^{n-3 } a(p + 1 , n - 3 - p ) \nonumber\\ & + \sum_{k\geq1 } \begin{pmatrix } 2 + k \\ k \end{pmatrix } \sum_{\sum_i t_i = n-2 } \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_k \end{bmatrix } ( n-2 ) !
a(0,n-2 ) + \ldots\end{aligned}\ ] ] @xmath737 is still unfiltered .
now , the filtration algorithm brings the words @xmath750 into lexicographical order using the concatenation commutator . in our case
, @xmath502 is sorted to the right , @xmath751 in this section , a filtration algorithm that sorts @xmath502 to the left would be more convenient because the last term in eq .
( [ eq : wn - contribution-2 ] ) would already be given in lexicographical order . since the resulting filtered words must not depend on the lexicographical order of letters , we now assume throughout this section that we work with a filtration algorithm that sorts @xmath502 to the left , @xmath752 this change does not effect the final generating functions @xmath733}$ ] and the respective generated matrices .
inserting eq .
( [ eq : sorting ] ) into eq .
( [ eq : wn - contribution-2 ] ) yields @xmath753 } ( t_k - 3 ) ! ( n-1-t_k ) !
\sum_{p=0}^{n-1-t_k } \begin{pmatrix } t_k - 2 + p \\
p \end{pmatrix } \nonumber\\ & \times \left ( \sum_{q=0}^{t_k - 2 + p}b(t_k - 2 + p - q , n - 1 - t_k - p + q ) - \sum_{q=0}^{n - t_k - p - 1}b ( n - t_k - p - 1 - q , t_k - 2 + p + q ) \right)\nonumber\\ & + \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } \begin{bmatrix } t_k-2 & 1 \end{bmatrix } ( n-3 ) !
\sum_{q = 0}^{p}b(p - q , n - 3 - p + q ) \nonumber\\ & + \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } \begin{bmatrix } t_k-2 & 1 \end{bmatrix } ( n-2 ) !
a(0 , n - 2 ) \nonumber\\ & + \sum_{k\geq1 } \begin{pmatrix } 2 + k \\
k \end{pmatrix } \sum_{\sum_i t_i = n-2 } \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_k \end{bmatrix } ( n-2 ) !
a(0,n-2 ) + \ldots\end{aligned}\ ] ] the last two terms together are equal to @xmath754 , see section [ sec : bullet2 ] .
step 2 of the filtration loop in section [ sec : pres - filtr - algor ] computes @xmath755 and brings all the words @xmath756 on the rhs into lexicographical order using eq .
( [ eq : sorting ] ) .
hence , in eq .
( [ eq : wn - contribution-3 ] ) , @xmath757 is replaced by @xmath758 thus , after step 2 of the respective filtration loop , @xmath759 } ( t_k - 3 ) ! ( n-1-t_k ) !
\sum_{p=0}^{n-1-t_k } \begin{pmatrix } t_k - 2 + p \\
p \end{pmatrix } \nonumber\\ & \times \left ( \sum_{q=0}^{t_k - 2 + p}b(t_k - 2 + p - q , n - 1 - t_k - p + q ) - \sum_{q=0}^{n - t_k - p - 1}b ( n - t_k - p - 1 - q , t_k - 2 + p + q ) \right)\nonumber\\ & + \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } \begin{bmatrix } t_k-2 & 1 \end{bmatrix } ( n-3 ) ! \sum_{p=0}^{n-3 } \sum_{q = 0}^{p}b(p - q ,
n - 3 - p + q ) \nonumber\\ & + \begin{bmatrix } n-2 & 1 \end{bmatrix } a_1^{\shuffle_\theta n-2}\shuffle_\theta a_2 - \frac{1}{n-1 } \begin{bmatrix } n-2 & 1 \end{bmatrix } \sum_{p=1}^{n-2}\sum_{q = 0}^{p-1 } b(p-1-q , n-2-p+q ) + \ldots\end{aligned}\ ] ] note that we do not take the filtered term @xmath760 into account because it does not contribute to the term in eq .
( [ eq : cterm ] ) .
we now proceed as in the derivation of the master differential equation in section [ sec : deriv - mast - equat ] .
we denote by @xmath761}$ ] the number of words on the rhs of eq .
( [ eq : wn - contribution-4 ] ) with @xmath743 letters @xmath501 and one letter @xmath744 $ ] . as in section
[ sec : notation ] , @xmath761}$ ] is related to the indexed matrix @xmath741}$ ] by @xmath762 } = \frac{1}{(n-2 ) ! } \begin{bmatrix } n-3 \end{bmatrix } _ { [ a_1,a_2]}.\ ] ] we obtain @xmath761}$ ] by setting all words @xmath763 on the rhs of eq .
( [ eq : wn - contribution-4 ] ) to 1 , hence @xmath764 } = & \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } \!\!\!\!\!\!\ ! k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix}\ ! \begin{bmatrix } t_k - 3 \end{bmatrix } _ { [ a_1,a_2 ] } ( t_k - 3 ) ! ( n-1-t_k ) ! \!\!\ ! \sum_{p=0}^{n-1-t_k}\!\ ! \begin{pmatrix } t_k - 2 + p \\
p \end{pmatrix } \ ! ( 2t_k + 2p - n - 1)\nonumber\\ & + \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } \frac{(n-1)!}{2 } k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } \begin{bmatrix } t_k-2 & 1 \end{bmatrix } - \frac{n-2}{2 } \begin{bmatrix } n-2 & 1 \end{bmatrix}.\end{aligned}\ ] ] standard combinatorial calculation yields @xmath765 furthermore , we find @xmath766 from eq .
( [ eq : curly - square-2 ] ) .
we divide eq .
( [ eq : wn - contribution-5 ] ) by @xmath767 and insert eq .
( [ eq : curly - square-4 ] ) on the lhs .
this results in @xmath768 } = & \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } \frac{1}{t_k(t_k-1 ) } \begin{bmatrix } t_k - 3 \end{bmatrix } _ { [ a_1,a_2 ] } \nonumber\\ & + \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } \frac{1}{2 } k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } \begin{bmatrix } t_k-2 & 1 \end{bmatrix } - \frac{n-2}{2 } \begin{bmatrix } n-2 & 1 \end{bmatrix}.\end{aligned}\ ] ] finally , we multiply with @xmath769 and sum over all @xmath770 .
this will give us @xmath771 } \mathrm{d}z$ ] on the lhs with zero integration constant .
on the rhs , we obtain @xmath772 } \mathrm{d}z \;=\ ; & \sum_{n=0}^\infty \bigg ( \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } k \begin{bmatrix } t_1 \end{bmatrix } z^{t_1 } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } z^{t_{k-1 } } \frac{1}{t_k(t_k-1 ) } \begin{bmatrix } t_k - 3 \end{bmatrix } _ { [ a_1,a_2]}z^{t_k } \nonumber\\ & + \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } \frac{z}{2 } k \begin{bmatrix } t_1 \end{bmatrix } z^{t_1}\ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } z^{t_{k-1 } } \begin{bmatrix } t_k-2 & 1 \end{bmatrix } z^{t_k - 1 } - \frac{z^2}{2}\frac{\mathrm{d}}{\mathrm{d}z}\frac{1}{z } \begin{bmatrix } n-2 & 1 \end{bmatrix}z^{n-1 } \bigg ) \nonumber\\ \ ; = \ ; & \sum_{k\geq 1 } k \begin{bmatrix}\bullet\end{bmatrix}^{k-1 } \int \left ( \int \begin{bmatrix}\bullet\end{bmatrix}_{[a_1,a_2 ] } \mathrm{d}z \right ) \mathrm{d}z + \sum_{k\geq 1 } \frac{z}{2 } k \begin{bmatrix}\bullet\end{bmatrix}^{k-1 } \begin{bmatrix } \bullet & 1 \end{bmatrix } - \frac{z^2}{2}\frac{\mathrm{d}}{\mathrm{d}z}\frac{1}{z } \begin{bmatrix } \bullet & 1 \end{bmatrix } \nonumber\\ \ ; = \ ; & \frac{1}{1 - 2z}\int \left ( \int \begin{bmatrix}\bullet\end{bmatrix}_{[a_1,a_2 ] } \mathrm{d}z \right ) \mathrm{d}z + \frac{1}{4\sqrt{1 - 2z } } - \frac{1}{4\sqrt{1 - 2z}^3 } + \frac{1}{2\sqrt{1 - 2z}}\log \left ( \frac{1}{\sqrt{1 - 2z } } \right),\end{aligned}\ ] ] where we used the explicit form of the generating functions @xmath591 and @xmath640 in the last line * * ( * * eqs .
( [ eq : bullet1],[eq : bullet2])*)*. again , the integration constants are zero .
( [ eq : dgl - commutator ] ) is an ordinary first order differential equation for @xmath773 } \mathrm{d}z \right ) \mathrm{d}z$ ] with the same homogeneous part as in the case of index - free generating functions , see section [ sec : deriv - mast - equat ] .
we can thus , use eq .
( [ eq : generating - function - final ] ) to obtain @xmath774 } \mathrm{d}z \right ) \mathrm{d}z = -\frac{\sqrt{1 - 2z}}{4 } + \frac{1}{4\sqrt{1 - 2z } } - \frac{\sqrt{1 - 2z}}{4}\log \left ( \frac{1}{\sqrt{1 - 2z } } \right ) - \frac{1}{4\sqrt{1 - 2z}}\log \left ( \frac{1}{\sqrt{1 - 2z } } \right).\ ] ] the second derivative finally results in the generating function @xmath775 } = \frac{1}{4\sqrt{1 - 2z}^3 } - \frac{1}{4\sqrt{1 - 2z}^5 } + \frac{1}{4\sqrt{1 - 2z}^3}\log \left ( \frac{1}{\sqrt{1 - 2z } } \right ) - \frac{3}{4\sqrt{1 - 2z}^5}\log \left ( \frac{1}{\sqrt{1 - 2z } } \right).\ ] ] in this section , we derive the generating function @xmath734}$ ] for the matrices @xmath777}$ ] .
these belong to the shuffle products @xmath778 $ ] in the filtered word @xmath737 . in full analogy to the previous section
, we derive an equivalent to eq .
( [ eq : wn - contribution-5 ] ) . in the partly filtered @xmath737 ,
we denote the number of words with @xmath743 letters @xmath501 and one letter @xmath779 $ ] by @xmath780 } = ( n-2 ) !
\begin{bmatrix } n-3 \end{bmatrix } _ { [ a_1,\theta(a_1,a_1)]}.\ ] ] we then find @xmath781 } = & \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } \begin{bmatrix } t_k - 3 \end{bmatrix } _ { [ a_1,\theta(a_1,a_1 ) ] } ( t_k - 3 ) ! ( n-1-t_k ) !
\sum_{p=0}^{n-1-t_k }
\begin{pmatrix } t_k - 2 + p \\
p \end{pmatrix } \nonumber\\ & \times ( 2t_k + 2p
- n - 1 ) + \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } \frac{(n-1)!}{2 } k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } \begin{bmatrix } t_k-2 \\ 2 \end{bmatrix } \nonumber\\ & + \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } \mathcal{s } \left ( \begin{bmatrix } n-3 \\ 2 \end{bmatrix } , \begin{bmatrix } n-1 \end{bmatrix } \right)\frac{(n-1)!}{2 } \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_k \end{bmatrix } - \frac{n-2}{2 } \begin{bmatrix } n-2 \\ 2 \end{bmatrix}.\end{aligned}\ ] ] compared to eq .
( [ eq : wn - contribution-5 ] ) , we made the obvious replacements @xmath782 } \to \begin{bmatrix } t_k - 3 \end{bmatrix } _ { [ a_1,\theta(a_1,a_1 ) ] } , \qquad \begin{bmatrix } t_k-2 & 1 \end{bmatrix } \to \begin{bmatrix } t_k-2 \\ 2 \end{bmatrix } , \qquad \begin{bmatrix } n-2 & 1 \end{bmatrix } \to \begin{bmatrix } n-2 \\ 2 \end{bmatrix}.\ ] ] the only new term is the third one .
it arises because on the rhs of eq .
( [ eq : wn - contribution-1 ] ) , one must also consider the term @xmath783 and calculate @xmath784 we divide eq .
( [ eq : wn - contribution-6 ] ) by @xmath767 and use eqs .
( [ eq : curly - square-2],[eq : combinatorics],[eq : curly - square-5 ] ) to obtain @xmath785 } = & \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } \frac{1}{t_k(t_k-1 ) } \begin{bmatrix } t_k - 3 \end{bmatrix } _ { [ a_1,\theta(a_1,a_1 ) ] } \nonumber\\ + & \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } \frac{1}{2 } k \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } \begin{bmatrix } t_k-2 \\ 2 \end{bmatrix}\nonumber\\ + & \sum_{k\geq 1 }
\sum_{\sum_i t_i = n-1 } \mathcal{s } \left ( \begin{bmatrix } n-3 \\ 2 \end{bmatrix } , \begin{bmatrix } n-1 \end{bmatrix } \right)\frac{1}{2 } \begin{bmatrix } t_1 \end{bmatrix } \ldots \begin{bmatrix } t_k \end{bmatrix } \nonumber\\ - & \frac{n-2}{2 } \begin{bmatrix } n-2 \\ 2 \end{bmatrix } - \frac{n-2}{2 } \mathcal{s } \left ( \begin{bmatrix } n-2 \\ 2 \end{bmatrix } , \begin{bmatrix } n \end{bmatrix } \right ) \begin{bmatrix } n \end{bmatrix } .\end{aligned}\ ] ] as in the previous section , we multiply with @xmath769 and sum over all @xmath786 .
with @xmath787 we find @xmath788 } \mathrm{d}z = & \ , \sum_{n=0}^\infty \bigg ( \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } k \begin{bmatrix } t_1 \end{bmatrix } z^{t_1 } \ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } z^{t_{k-1 } } \frac{1}{t_k(t_k-1 ) } \begin{bmatrix } t_k - 3 \end{bmatrix } _ { [ a_1,\theta(a_1,a_1)]}z^{t_k } \nonumber\\ & + \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } \frac{z}{2 } k \begin{bmatrix } t_1 \end{bmatrix } z^{t_1}\ldots \begin{bmatrix } t_{k-1 } \end{bmatrix } z^{t_{k-1 } } \begin{bmatrix } t_k-2 \\ 2 \end{bmatrix } z^{t_k - 1 } \nonumber\\ & + \sum_{k\geq 1 } \sum_{\sum_i t_i = n-1 } \frac{z^2}{4}\frac{\mathrm{d^2}}{\mathrm{d}z^2 } \begin{bmatrix } t_1 \end{bmatrix } z^{t_1}\ldots \begin{bmatrix } t_k \end{bmatrix } z^{t_k } - \frac{z^2}{2}\frac{\mathrm{d}}{\mathrm{d}z}\frac{1}{z } \begin{bmatrix } n-2 \\ 2 \end{bmatrix}z^{n-1 } - \frac{z^2}{4 } \frac{\mathrm{d}^3}{\mathrm{d}z^3 } \begin{bmatrix } n \end{bmatrix } z^n \bigg ) \nonumber\\ = & \ , \sum_{k\geq 1 } k \begin{bmatrix}\bullet\end{bmatrix}^{k-1 } \int \left ( \int \begin{bmatrix}\bullet\end{bmatrix}_{[a_1,\theta(a_1,a_1 ) ] } \mathrm{d}z \right ) \mathrm{d}z + \sum_{k\geq 1 } \frac{z}{2 } k \begin{bmatrix}\bullet\end{bmatrix}^{k-1 } \begin{bmatrix } \bullet \\ 2 \end{bmatrix } + \sum_{k\geq1 } \frac{z^2}{4 } \frac{\mathrm{d}^2}{\mathrm{d}z^2 } \begin{bmatrix}\bullet\end{bmatrix}^k \nonumber\\ & - \frac{z^2}{2}\frac{\mathrm{d}}{\mathrm{d}z}\frac{1}{z } \begin{bmatrix } \bullet \\ 2 \end{bmatrix } - \frac{z^2}{4 } \frac{\mathrm{d}^3}{\mathrm{d}z^3}\begin{bmatrix}\bullet\end{bmatrix } \nonumber\\ = & \ , \frac{1}{1 - 2z}\int \left ( \int \begin{bmatrix}\bullet\end{bmatrix}_{[a_1,\theta(a_1,a_1 ) ] } \mathrm{d}z \right ) \mathrm{d}z - \frac{1}{8\sqrt{1 - 2z } } + \frac{1}{8\sqrt{1 - 2z}^3 } - \frac{1}{4\sqrt{1 - 2z}}\log \left ( \frac{1}{\sqrt{1 - 2z } } \right).\end{aligned}\ ] ] note that the third and the fifth term on the rhs of the second equation cancel .
this is an interesting incidence . because of eq .
( [ eq : bullet5bullet2 ] ) , the inhomogeneous parts of the differential equations eqs .
( [ eq : dgl - commutator],[eq : dgl - commutator-2 ] ) only differ by a factor of @xmath660 .
we therefore obtain @xmath789 } = -\frac{1}{2 } \begin{bmatrix}\bullet\end{bmatrix}_{[a_1,a_2 ] } = -\frac{1}{8\sqrt{1 - 2z}^3 } + \frac{1}{8\sqrt{1 - 2z}^5 } - \frac{1}{8\sqrt{1 - 2z}^3}\log \left ( \frac{1}{\sqrt{1 - 2z } } \right ) + \frac{3}{8\sqrt{1 - 2z}^5}\log \left ( \frac{1}{\sqrt{1 - 2z } } \right).\ ] ] we now demonstrate the power of the generating functions derived in the previous sections : one can write the next - to@xmath45-leading log order as a function of terms up to @xmath315 in the log - expansion * * ( * * eq .
( [ eq : log - expansion])*)*. we will show this for the yukawa fermion propagator * * ( * * eq .
( [ eq : dse - yuk ] ) * ) * up to @xmath790 using the explicit generating functions obtained in the previous section .
they are also collected in the second column of table [ tab : results ] .
we discuss our results for @xmath791 separately .
consider the filtered solution @xmath792 of the dse eq .
( [ eq : dse - yuk - words ] ) , see eq .
( [ eq : dse - solution ] ) .
the leading log order is @xmath793 contributing terms in the filtered words @xmath217 map to @xmath794 under renormalized feynman rules since the leading log order is @xmath795 .
these are only the full shuffle products @xmath796 * * ( * * see section [ sec : renormalized - feynman - rules ] that @xmath797**)**. thus , @xmath798 see the first row of table [ tab : yukawa ] for the explicit multiplicities . acting with renormalized feynman rules @xmath799 on both sides results in @xmath800 we write this equation in terms of feynman graphs
. therefore , set @xmath801 .
we find that @xmath802 on the rhs using the properties of the hopf algebra morphism @xmath803 in eqs .
( [ eq : morphism - begin]-[eq : morphism - end ] ) . on the lhs , @xmath804 , which is the full green function of the fermion propagator .
hence , eq . ( [ eq : psiw - ll ] ) yields @xmath805 using eq .
( [ eq : bullet1 ] ) for the generating function @xmath591 , we finally obtain @xmath806 without this result , the computation of @xmath807 would be quite more complicated , even impossible .
computing @xmath808 the ordinary way includes to calculate an infinite number of feynman integrals with any number of loops .
for example , the graphs @xmath809 contribute to @xmath808 . using our formula in eq .
( [ eq : gr - ll2 ] ) , _ we only need to compute the one - loop feynman integral @xmath810 to derive the full leading log order green function @xmath808 . _
the next - to - leading log order of eq .
( [ eq : dse - solution ] ) is @xmath811 contributing terms of the filtered words @xmath217 map to @xmath812 under renormalized feynman rules since the next - to - leading log order is @xmath813 .
these are the full shuffle products @xmath814 and @xmath815 . indeed , renormalized feynman rules are character - like . @xmath799 acting on a full shuffle product of @xmath816 letters is @xmath817 in the log - expansion , see section [ sec : renormalized - feynman - rules ] .
thus , @xmath818 acting with renormalized feynman rules @xmath799 on both sides results in @xmath819 again , we write @xmath801 and obtain the full next - to - leading log order renormalized green function of the yukawa fermion propagator on the lhs . on the rhs ,
the only subtle point is that @xmath820 has no single corresponding feynman graph .
however , we find the period @xmath821 thus , @xmath822 using the generating functions in eqs .
( [ eq : bullet2],[eq : bullet5 ] ) , we finally derive @xmath823 this is an enormous simplification : _ we only need to compute the one - loop feynman integral @xmath810 as well as the two - loop integrals @xmath824 and @xmath825 to calculate the full next - to - leading log order green function @xmath826 . _ the next - to - next - to - leading log order of eq .
( [ eq : dse - solution ] ) is @xmath827 contributing terms of the filtered words @xmath217 must map to @xmath828 under renormalized feynman rules since the next - next - to - leading log order is @xmath829 .
these are the full shuffle products @xmath830 , \qquad a_1^{\shuffle_\theta\ , ( n-3)}\shuffle_\theta [ a_1,\theta(a_1,a_1)].\end{gathered}\ ] ] > from eq .
( [ eq : w - nnll ] ) , @xmath831 } ( \alpha a_1)^{\shuffle_\theta\ , ( n-3)}\shuffle_\theta \left(\alpha^3 [ a_1,a_2]\right ) + \begin{bmatrix } n-3 \end{bmatrix } _ { [ a_1,\theta(a_1,a_1 ) ] } ( \alpha a_1)^{\shuffle_\theta\ , ( n-3)}\shuffle_\theta \left ( \alpha^3 [ a_1,\theta(a_1,a_1 ) ] \right)\!\ ! \bigg).\end{aligned}\ ] ] acting with renormalized feynman rules @xmath799 on both sides results in @xmath832 } \alpha^3 \psi_r ( [ a_1,a_2 ] ) + \begin{bmatrix}\bullet\end{bmatrix}_{[a_1,\theta(a_1,a_1 ) ] } \alpha^3 \psi_r ( [ a_1,\theta(a_1,a_1 ) ] ) \bigg)\bigg|_{z \to \alpha \psi_r(a_1)}.\end{aligned}\ ] ] we write @xmath801 and obtain the full next - to - next - to - leading log order green function on the lhs . on the rhs , for example @xmath833 .
however , the letters @xmath834 , @xmath513 , @xmath744 $ ] , @xmath779 $ ] have no obvious corresponding feynman graphs . we therefore write @xmath835 = & \ ,
b_+^{a_1}(a_2 ) - b_+^{a_2}(a_1 ) , \\ [ a_1 , \theta(a_1,a_1 ) ] = & \ , 2b_+^{a_1 } \left ( a_1 \shuffle_\theta a_1 \right ) - b_+^{a_1}\left(b_+^{a_1}\left(a_1\right)\right ) + \frac{1}{2 } a_1 \shuffle_\theta \theta(a_1,a_1 ) - \frac{1}{2 } a_1\shuffle_\theta a_1 \shuffle_\theta a_1.\end{aligned}\ ] ] now , we act with @xmath801 and use eq .
( [ eq : psitheta1 ] ) .
we thus , find the periods @xmath836 ) = & \ , \phi_r\left(b_+^{\gamma_1}(\gamma_2)\right ) - \phi_r\left(b_+^{\gamma_2}(\gamma_1)\right ) , \\
\psi_r([a_1 , \theta(a_1,a_1 ) ] ) = & \ , 2\phi_r \left ( b_+^{\gamma_1 } \left ( \gamma_1 \cup \gamma_1 \right ) \right ) - \phi_r \left ( b_+^{\gamma_1}\left(b_+^{\gamma_1}\left(\gamma_1\right)\right ) \right ) - \phi_r(\gamma_1)\phi_r\left(b_+^{\gamma_1}\left(\gamma_1\right)\right).\end{aligned}\ ] ] inserting these identities together with eq.([eq : psitheta1 ] ) into eq .
( [ eq : psiw - nnll ] ) , we finally obtain the next - to - next - to - leading log order green function , @xmath837 } \left ( \phi_r\left(b_+^{\gamma_1}(\gamma_2)\right ) - \phi_r\left(b_+^{\gamma_2}(\gamma_1)\right ) \right ) \nonumber\\ & + \ , \begin{bmatrix}\bullet\end{bmatrix}_{[a_1,\theta(a_1,a_1 ) ] } \bigg ( 2\phi_r \left ( b_+^{\gamma_1 } \left ( \gamma_1 \cup \gamma_1 \right ) \right ) - \phi_r \left ( b_+^{\gamma_1}\left(b_+^{\gamma_1}\left(\gamma_1\right)\right ) \right ) \nonumber\\ & - \phi_r(\gamma_1 ) \phi_r \left ( b_+^{\gamma_1}\left(\gamma_1\right ) \right ) \bigg ) \bigg)\bigg|_{z \to \alpha\phi_r(\gamma_1)}.\end{aligned}\ ] ] we refer to the second column of table [ tab : results ] for the explicit expressions of the generating functions .
_ this shows the explicit dependence of the full next - to - next - to - leading log order green function @xmath838 on the feynman graphs @xmath839 these are at most , three - loop graphs . _
we relate the next - to@xmath45-leading log order to the first @xmath16 terms of perturbation theory in the qed photon self - energy green function @xmath840 .
there are two differences to the yukawa propagator green function . *
* ( * * consider the dses eqs .
( [ eq : dse - solution - yuk],[eq : dse - solution - qed ] ) .
first , there are no insertion points in the one - loop primitive propagator graph .
the sum in eq .
( [ eq : dse - solution - qed ] ) starts with @xmath618 rather than @xmath617 in eq .
( [ eq : dse - solution - yuk ] ) .
this will simplify the following calculations drastically .
secondly , the term @xmath841 in eq .
( [ eq : dse - solution - yuk ] ) is replaced by @xmath842 in eq .
( [ eq : dse - solution - qed ] ) .
this will change the structure @xmath843 in the generating functions of yukawa theory to @xmath844 in qed . in the first part
, we treat index - free matrices .
the corresponding shuffle products do not contain @xmath222$]-letters .
we repeat the same steps to derive the master differential equation as in yukawa theory .
the two mentioned differences in the respective dses eqs .
( [ eq : dse - solution - yuk],[eq : dse - solution - qed ] ) change eq .
( [ eq : master ] ) to @xmath845 this is an ordinary equation for @xmath583 and no differential equation because the sum starts with @xmath618 .
we therefore call eq .
( [ eq : masterqed ] ) _ master equation_. we must integrate the master equation to obtain @xmath554 such that @xmath846 * * ( * * eq . ( [ eq : initial - condition])*)*. eqs . ( [ eq : def - generating - function],[eq : matrices - of - generating - functions ] ) remain valid , @xmath847 consider for example the case @xmath586 .
the matrices @xmath848 belong to the shuffle products @xmath849 . since there is no @xmath850
, we already know that @xmath851 for @xmath852 . in the master equation ,
all terms vanish except for @xmath570 . integrating the remaining equation @xmath853 yields @xmath854 which generates the matrices @xmath855 as expected * * ( * * see eq .
( [ eq : def - generating - function - copy])*)*. in the following , we derive the generating functions up to next - to - next - to - leading log order for the qed photon self - energy . as in section
[ sec : relat - betw - next ] , we consider the cases @xmath599 and @xmath643 separately . in section [ sec : commutators-1 ] , we treat indexed matrices @xmath119 .
index - free matrices with @xmath599 belong to shuffle products without @xmath600- and @xmath222$]-letters . in full analogy to the yukawa propagator , only one row in @xmath554 reduces the master equation to @xmath856 where @xmath857 let @xmath613 . the sum in eq .
( [ eq : masterqed - special ] ) is non - zero only for @xmath859
. then , @xmath620 and @xmath860 @xmath587 .
we obtain @xmath861 let @xmath863 . the sum in eq .
( [ eq : masterqed - special ] ) is non - zero only for @xmath864
. then , @xmath629 and @xmath860 @xmath587 .
we obtain @xmath865 the term @xmath866 is the integration constant such that @xmath867 .
let @xmath869 .
the sum in eq .
( [ eq : masterqed - special ] ) is non - zero only for @xmath859
. then , @xmath870 .
( [ eq : starstarqed ] ) implies that for one @xmath871 , @xmath872 and for all other @xmath463 , @xmath873 .
we obtain @xmath874 we now treat full shuffle products that contain @xmath600-letters but no @xmath222$]-letters . here , we have to proceed from the master equation eq .
( [ eq : masterqed ] ) .
the following generating functions simplify to zero , @xmath875 the reason is that the sum in the master equation starts with @xmath618 .
there is another simplification due to the missing @xmath877-term .
if the first row of a matrix @xmath119 is of the form @xmath878 with arbitrary @xmath879 , then the second ( and more complicated ) term on the rhs of the master equation eq .
( [ eq : masterqed ] ) yields zero .
for example for the next - to - next - to - leading log generating function @xmath880 , eq .
( [ eq : masterqed ] ) reduces to @xmath881 integration with suitable initial conditions and using eq .
( [ eq : bulletqed2 ] ) result in @xmath882 for @xmath884 , the master equation eq .
( [ eq : masterqed ] ) reduces to @xmath885 in the last line , we used eqs .
( [ eq : bulletqed1],[eq : bulletqed2 ] ) .
we integrate with suitable initial conditions and obtain @xmath886 as in yukawa theory , we only calculate the generating functions @xmath733}$ ] and @xmath734}$ ] that belong to the shuffle products @xmath887 $ ] and @xmath888 $ ] .
we find @xmath889 } = 0\ ] ] because the sum in eq .
( [ eq : dse - solution - qed ] ) starts with @xmath714 . for the generating function @xmath733}$ ] , eq .
( [ eq : dgl - commutator ] ) reduces to @xmath890 } \mathrm{d}z = - \frac{z^2}{2}\frac{\mathrm{d}}{\mathrm{d}z}\frac{1}{z } \begin{bmatrix } \bullet & 1 \end{bmatrix}\ ] ] because in eq .
( [ eq : wn - contribution-1 ] ) , the first two terms are missing . here
, the derivation of the generating function is even simpler than in all previous cases .
one only needs to differentiate instead of integrating or solving a differential equation .
we differentiate eq .
( [ eq : integration - commutator ] ) and use the explicit form of @xmath640 in eq .
( [ eq : bulletqed2 ] ) .
this results in @xmath891 } = \frac{1}{2(1-z ) } - \frac{1}{2(1-z)^2}.\ ] ] we repeat the steps in section [ sec : results ] to write the next - to@xmath45-leading log order as a function of terms up to @xmath892 in the log - expansion * * ( * * eq . (
[ eq : log - expansion])*)*. we show this up to @xmath790 and use eqs .
( [ eq : gr - ll1],[eq : gr - nll1],[eq : gr - nnll1 ] ) , which are universally valid ( in the qed case below several terms vanish ) : @xmath893 } \left ( \phi_r\left(b_+^{\gamma_1}(\gamma_2)\right ) - \phi_r\left(b_+^{\gamma_2}(\gamma_1)\right ) \right ) \nonumber\\ & + \ , \begin{bmatrix}\bullet\end{bmatrix}_{[a_1,\theta(a_1,a_1 ) ] } \bigg ( 2\phi_r \left ( b_+^{\gamma_1 } \left ( \gamma_1 \cup \gamma_1
\right ) \right ) - \phi_r \left ( b_+^{\gamma_1}\left(b_+^{\gamma_1}\left(\gamma_1\right)\right )
\right ) \nonumber\\ & - \phi_r(\gamma_1)\phi_r\left(b_+^{\gamma_1}\left(\gamma_1\right)\right ) \bigg ) \bigg)\bigg|_{z \to \alpha\phi_r(\gamma_1)}.\end{aligned}\ ] ] for the leading log and next - to - leading log expansions , we use the explicit generating functions in the third column of table [ tab : results ] . for the next - to - next - to - leading log order
, we only discard the zero functions .
thus , we finally obtain @xmath894 } \left ( \phi_r\left(b_+^{\gamma_1}(\gamma_2)\right ) - \phi_r\left(b_+^{\gamma_2}(\gamma_1)\right ) \right ) \bigg)\bigg|_{z \to \alpha\phi_r(\gamma_1)}.\end{aligned}\ ] ] see the third column of table [ tab : results ] for the remaining generating functions .
we list some multiplicities @xmath119 that are generated by the generating functions obtained so far , see table [ tab : results ] .
we treat the yukawa fermion propagator and the qed photon self - energy separately .
if sequences are known , we say so explicitly and refer to @xcite .
the generating functions that are necessary to simplify the log - expansion of the yukawa fermion propagator up to next - to - next - to - leading log order are collected in the second column of table [ tab : results ] .
some of the corresponding multiplicities are collected in table [ tab : yukawa ] .
@xmath591 generates the exponential sequence of double factorial odd numbers a001147 , @xmath895 furthermore , @xmath640 generates the exponential series for the scaled sums of odd reciprocals a004041 .
the formula is @xmath896 we finally find @xmath897 ( exponential sequence a001879 ) . as far as we are concerned , there are no known sequences for the other rationals in table [ tab : yukawa ] .
c|cccccccc & 0 & 1 & 2 & 3 & 4 & 5 + + @xmath898 & 0 & 1 & @xmath899 & @xmath899 & @xmath900 & @xmath901 + @xmath902 & 1 & 2 & @xmath903 & @xmath904 & @xmath905 & @xmath906 + @xmath907 & @xmath908 & @xmath866 & @xmath909 & @xmath910 & @xmath911 & @xmath912 + @xmath913 & 1 & 3 & @xmath914 & @xmath915 & @xmath916 & @xmath917 + @xmath918 & @xmath919 & @xmath920 & @xmath921 & @xmath922 & @xmath923 & @xmath924 + @xmath925 & @xmath899 & @xmath926 & @xmath927 & @xmath928 & @xmath929 & @xmath930 + @xmath931 & @xmath932 & @xmath933 & @xmath934 & @xmath935 & @xmath936 & @xmath937 + @xmath938 & @xmath939 & @xmath940 & @xmath941 & @xmath942 & @xmath943 & @xmath944 + @xmath945 & @xmath946 & @xmath947 & @xmath948 & @xmath949 & @xmath950 & @xmath951 + @xmath952}$ ] & @xmath866 & @xmath953 & @xmath954 & @xmath955 & @xmath956 & @xmath957 + @xmath958}$ ] & @xmath899 & @xmath959 & @xmath960 & @xmath961 & @xmath962 & @xmath963 + the generating functions that are necessary to simplify the log - expansion of the photon self - energy green function up to next - to - next - to - leading log order are given in the third column of table [ tab : results ] .
some of the corresponding rationals are listed in table [ tab : photon ] .
the reader immediately checks that these numbers look much simpler than in the yukawa case .
indeed , four rows only contain zero numbers ( the first line is almost zero , @xmath964 ) . we find the trivial sequences @xmath965 } = - \frac{n+1}{2}.\ ] ] one also finds that @xmath966 is the exponential series of the generalized stirling numbers a001705 , @xmath967 c|cccccccc & 0 & 1 & 2 & 3 & 4 & 5 + + @xmath898 & 1 & 0 & 0 & 0 & 0 & 0 + @xmath902 & 1 & @xmath899 & @xmath968 & @xmath969 & @xmath970 & @xmath971 + @xmath907 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath913 & 1 & 1 & 1 & 1 & 1 & 1 + @xmath918 & @xmath899 & @xmath972 & @xmath973 & @xmath974 & @xmath975 & @xmath976 + @xmath925 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath931 & @xmath908 & @xmath977 & @xmath978 & @xmath979 & @xmath946 & @xmath980 + @xmath938 & 0 & 0 & 0 & 0 & 0 & 0 + @xmath945 & @xmath971 & @xmath969 & @xmath981 & @xmath968 & @xmath982 & @xmath940 + @xmath952}$ ] & @xmath908 & @xmath866 & @xmath983 & @xmath932 & @xmath984 & @xmath985 + @xmath958}$ ] & 0 & 0 & 0 & 0 & 0 & 0 +
99 d.kreimer _ quantum fields , periods and algebraic geometry , _ arxiv:1405.4964 [ hep - th ] .
icmat proceedings _ periods and motives _ ,
madrid , july 2012 , publ.by ams _ contemporary mathematics _ , luis lvarez - cnsul et.al . , eds .
, in press .
f.brown , d.kreimer , _ angles , scales and parametric renormalization , _ lett . math .
* 103 * ( 2013 ) 933 [ arxiv:1112.1180 [ hep - th ] ] .
a.connes , d.kreimer , _ hopf algebras , renormalization and noncommutative geometry , _ commun . math .
* 199 * ( 1998 ) 203 [ hep - th/9808042 ] .
c.bergbauer , d.kreimer , _ hopf algebras in renormalization theory : locality and dyson - schwinger equations from hochschild cohomology , _ irma lect .
* 10 * ( 2006 ) 133 [ hep - th/0506190 ] .
m.borinsky , _ feynman graph generation and calculations in the hopf algebra of feynman graphs , _ comput .
commun .
* 185 * ( 2014 ) 3317 [ arxiv:1402.2613 [ hep - th ] ] .
l.foissy , _ lie algebra associated to systems of dyson - schwinger equations , _ adv .
math.226 ( 2011 ) , no . 6 , 4702 - 4730 .
d.broadhurst , d.kreimer , _ exact solutions of dyson - schwinger equations for iterated one loop integrals and propagator coupling duality , _ nucl .
b * 600 * ( 2001 ) 403 [ hep - th/0012146 ] .
s.bloch , d.kreimer , _ mixed hodge structures and renormalization in physics , _ commun .
* 2 * ( 2008 ) 637 [ arxiv:0804.4399 [ hep - th ] ] .
t.k.chen , _ iterated path integrals _ , bull.amer.math.soc.83 ( 1977 ) 831 .
d.kreimer , _ anatomy of a gauge theory , _ annals phys .
* 321 * ( 2006 ) 2757 [ hep - th/0509135 ] .
d.kreimer , m.sars , w.van suijlekom , _ quantization of gauge fields , graph polynomials and graph homology , _ annals phys .
* 336 * ( 2013 ) 180 [ arxiv:1208.6477 [ hep - th ] ] .
c.reutenauer , _ free lie algebras _ , lms ( 1993 ) .
a.connes , d.kreimer , _ renormalization in quantum field theory and the riemann - hilbert problem .
2 . the beta function , diffeomorphisms and the renormalization group , _ commun .
* 216 * ( 2001 ) 215 [ hep - th/0003188 ] .
p.hall , _ the algebra of partitions _ , proceedings of the 4th canadian mathematical congress , banff , pp .
w.milnor , j.c.moore , _ on the structure of hopf algebras _ , ann . of math .
( 2 ) , 81 : 2 ( 1965 ) pp . 211264 .
oeis foundation inc .
( 2011 ) , _ the on - line encyclopedia of integer sequences _ , http://oeis.org | dyson - schwinger equations determine the green functions @xmath0 in quantum field theory .
their solutions are triangular series in a coupling constant @xmath1 and an external scale parameter @xmath2 for a chosen amplitude @xmath3 , with the order in @xmath2 bounded by the order in the coupling .
perturbation theory calculates the first few orders in @xmath1 . on the other hand , dyson
schwinger equations determine next - to@xmath4-leading log expansions , @xmath5 .
@xmath6 sums for any finite @xmath7 a finite number of functions @xmath8 in @xmath9 . here , @xmath9 is the one - loop approximation to @xmath10 , for example , for the ( inverse ) propagator in massless yukawa theory , @xmath11 .
the leading logs come then from the trivial representation @xmath12 at @xmath13 with @xmath14 .
all non - leading logs are organized by corresponding suppressions in powers @xmath15 .
we describe an algebraic method to derive all next - to@xmath4-leading log terms from the knowledge of the first @xmath16 terms in perturbation theory and their filtrations .
this implies the calculation of the functions @xmath17 and periods @xmath18 . in the first part of our paper
, we investigate the structure of dyson - schwinger equations and develop a method to filter their solutions . applying renormalized feynman rules maps each filtered term to a certain power of @xmath1 and @xmath2 in the log - expansion .
based on this , the second part derives the next - to@xmath4-leading log expansions .
our method is general . here
, we exemplify it using the examples of the propagator in yukawa theory and the photon self - energy in quantum electrodynamics . in particular , we give explicit formulas for the leading log , next - to - leading log and next - to - next - to - leading log orders in terms of at most three - loop feynman integrals .
the reader may apply our method to any ( set of ) dyson - schwinger equation(s ) appearing in renormalizable quantum field theories . |
this work was funded by science foundation ireland through a principal investigator award grant no .
13/ia/1956 . ass was supported by an irish research council graduate fellowship .
we acknowledge additional support from intel corporation and quantumwise .
atomistic visualizations were rendered using vesta @xcite .
batra , a. ; darancet , p. ; chen , q. ; meisner , j. s. ; widawsky , j. r. ; neaton , j. b. ; nuckolls , c. ; venkataraman , l. _ nano letters _ * 2013 * , _ 13 _ , 62336237 , pmid : 24274757 capozzi , b. ; xia , j. ; adak , o. ; dell , e. j. ; liu , z .- f . ; taylor , j. c. ; neaton , j. b. ; campos , l. m. ; venkataraman , l. _ nat nano _ * 2015 * , _ 10 _ , 522527 aviram , a. ; ratner , m. a. _ chemical physics letters _ * 1974 * , _ 29 _ , 277 283 ansari , l. ; fagas , g. ; colinge , j .- p . ; greer , j. c. _ nano letters _ * 2012 * , _ 12 _ , 22222227 , pmid : 22500745 paul , w. _ journal of applied physics _ * 1961 * , _ 32 _ , 20822094 farrow , r. ; robertson , d. ; williams , g. ; cullis , a. ; jones , g. ; young , i. ; dennis , p. _ journal of crystal growth _ * 1981 * , _ 54 _ , 507 518 hchst , h. ; hernndez - caldern , i. _ surface science _ * 1983 * , _ 126 _ , 25 31 john , p. ; miller , t. ; chiang , t .- c . _ phys .
b _ * 1989 * , _ 39 _ , 32233229 cheong , b. h. ; chang , k. j. _ phys . rev .
b _ * 1991 * , _ 44 _ , 41034108 hedin , l. _ phys . rev . _ * 1965 * , _ 139 _ , a796a823 sanchez - soares , a. ; odonnell , c. ; greer , j. _ to be published _ * 2016 * , leu , p. w. ; shan , b. ; cho , k. _ phys . rev .
b _ * 2006 * , _ 73 _ , 195320 nolan , m. ; ocallaghan , s. ; fagas , g. ; greer , j. c. ; frauenheim , t. _ nano letters _ * 2007 * , _ 7 _ , 3438 soler , j. m. ; artacho , e. ; gale , j. d. ; garcia , a. ; junquera , j. ; ordejon , p. ; sanchez - portal , d. _ journal of physics : condensed matter _ * 2002 * , _ 14 _ , 2745 _ atomistix toolkit version 2015 - 1 quantumwise a / s ( www.quantumwise.com ) _ perdew , j. ; burke , k. ; ernzerhof , m. _ physical review letters _ * 1996 * , _ 77 _ , 38653868 morrison , i. ; bylander , d. ; kleinman , l. _ physical review b _ * 1993 * , _ 47 _ , 67286731 ozaki , t. _ physical review b _ * 2003 * , _ 67 _ , 155108 ozaki , t. ; kino , h. _ physical review b _ * 2004 * , _ 69 _ , 195113 bardeen , j. _ phys
. rev . _ * 1947 * , _ 71 _ , 717727 landman , u. ; barnett , r. n. ; scherbakov , a. g. ; avouris , p. _ phys . rev .
_ * 2000 * , _ 85 _ , 19581961 datta , s. _ superlattices and microstructures _ * 2000 * , _ 28 _ , 253 278 brandbyge , m. ; mozos , j .- l . ;
ordejn , p. ; taylor , j. ; stokbro , k. _ phys .
b _ * 2002 * , _ 65 _ , 165401 momma , k. ; izumi , f. _ journal of applied crystallography _ * 2011 * , _ 44 _ , 12721276 | for semimetal nanowires with diameters smaller than a few tens of nanometers , a semimetal - to - semiconductor transition is observed as the emergence of an energy band gap resulting from quantum confinement .
quantum confinement in a semimetal results in either lifting of the degeneracy of the conduction and valence bands in a _ zero _ gap semimetal , or shifting of bands with a _ negative _ energy overlap to form conduction and valence bands . for semimetal nanowires with diameters below 10 nanometer ,
the magnitude of the band gap can become significantly larger than the thermal energy at room temperature resulting in a new class of semiconductors relevant for nanoelectronics with critical dimensions on the order of a few atomic lengths .
the smaller a nanowire s diameter , the larger its surface - to - volume ratio thus leading to an increasing impact of surface chemistry on its electronic structure .
energy level shifts to states in the vicinity of the fermi level due to the electronegativity of surface terminating species are shown to be comparable in magnitude to quantum confinement effects at nanowire diameters of a few nanometer ; these two effects can be used to counteract one another leading to semimetallic behavior for nanowire cross sections at which the quantum confinement effect would otherwise dominate . abruptly changing the surface terminating species along the length of a nanowire leads to an abrupt change in the surface electronegativity .
this can result in the formation of a semimetal - semiconductor junction within a _
_ mono__material nanowire , without the need for impurity doping nor requiring the formation of a heterojunction .
using density functional theory in tandem with a green s function approach to determine electronic structure and charge transport , respectively , current rectification is calculated for such a junction .
current rectification ratios of @xmath0 are predicted from bias voltages as low as 300 mv . it is concluded that rectification can be achieved at essentially molecular length scales with conventional voltage biasing , while rivalling the performance of macroscopic semiconductor diodes .
* keywords : * semimetal nanowire , surface chemical modification , quantum confinement , schottky junction + molecular electronics aspires to produce diode , transistor , or logic gate functions in single molecules or small molecular assemblies .
diode behavior is the most fundamental of these as rectifying junctions can be arranged and combined with electrostatic gating to provide a transistor or switching function , and hence the realization of all boolean operations , as well as signal amplification , modulation , and other basic operations that enable electronic circuits .
molecular diodes have been demonstrated with increasing success and significant progress in recent years using new schemes such as tunable coupling @xcite and rectification ratios ( @xmath1 ) greater than 200 have been reported using electrochemical methods @xcite . in general , operating voltages tend to be relatively high ( @xmath2 v ) and rectification ratios low ( @xmath3 ) rendering single molecular diodes mostly unsuitable for nanoelectronic designs , and reproducible placement and electrode coupling remains a challenge .
+ in the following , a strategy is presented which realizes a schottky junction in a _ _ mono__material using ideas derived from molecular electronics in the sense that chemistry is used to locally modify electron affinity in order to control current flows @xcite . applying this approach , surface terminating groups of differing electronegativity
are introduced and shown to create semimetallic and semiconducting regions in @xmath4-tin nanowires ( snnws ) . the semimetal @xmath4-tin ( @xmath4-sn ) is used as an exemplar to explore the effects of quantum confinement and altering of surface electronegativity through use of a combination of density functional theory ( dft ) for electronic structure and a green s function method for studying charge transport across the junction .
@xmath4-sn nanowires are selected as they display advantageous properties for nanoelectronics in terms of the band gaps that can be achieved at small nanowire diameters @xcite , and relative to other semimetal it is straightforward to passivate the nanowires without the introduction of surface defect states within a quantum confinement - induced band gap .
although not stable in bulk form at room temperature @xcite , the @xmath4 phase ( diamond crystal structure ) in thin films can be stabilized @xcite at room temperature . for a hydrogen terminated snnw with diameter of approximately 5 nm , a semimetal - to - semiconductor transition is expected to yield a band gap of the order of 100 mev , with the band gap energy increasing rapidly to the order of an electron - volt with reducing nanowire cross section @xcite .
a hydrogen terminated @xmath4-sn nanowire is used as a convenient reference system against which the electron affinity for different surface terminating species and band gap shifts are compared .
it should be noted that a snnw diameter of @xmath51 nm diameter is used in this study , however due to deficiencies in the approximations to the exchange and correlation potentials in dft , the quantum confinement induced band gaps are significantly under - predicted , as verified through _ gw _ @xcite calculations for snnw band gaps @xcite .
hence the effects as demonstrated in the following are expected to extend to snnws with diameters estimated to be in the range of 3 nm to 5 nm .
it is known that surface chemistry results in a significant shift to the band gaps of semiconductor thin films and nanowires for critical dimensions below 10 nm .
for example , in silicon nanowires increasing the electron affinity of the surface terminating groups results in a red shift to the band gap energy @xcite . just as surface chemistry can reduce the effects of quantum confinement in semiconductor nws , it will be shown that increasing the electronegativity of surface terminating groups or surface bonding atoms to a semimetal nw competes against the quantum confinement induced band gap at small dimensions . ) , electron affinity ( @xmath6 ) , and bandgap ( @xmath7 ) for 1 nm diameter @xmath8 @xmath4-sn nws passivated with fluorine and hydrogen .
the larger electronegativity of fluorine is found to counteract the quantum - confinement induced bandgap observed for hydrogen terminated nws . ] for the case of @xmath4-tin nws oriented along the @xmath9 direction with fluorine termination , the band gap is reduced to the extent that the nanowire is nearly semimetallic with only a small energy gap of @xmath10 appearing at a diameter of @xmath51 nanometer . as the quantum confinement effect reduces with increasing nw diameter ,
the fluorine terminated snnws are predicted to remain semimetallic at room temperature for diameters @xmath11 1 nm . the effect of surface termination versus quantum confinement can be understood in terms of surface dipoles induced by the charge transfer at the surface of a nw to the terminating species . for large electron affinities ,
the terminating atoms or molecules draw charge out of the nw and the resulting surface negative charge can , as a first approximation , be thought of as a charge density distributed along the surface of a cylinder .
the image potentials in the semimetal nanowire of opposite polarity may likewise be treated as a surface charge density but on the surface of an inner cylinder concentric to the outer cylindrical charge distribution .
hence the surface dipoles may be considered as two nested , hollow cylindrical charge distributions .
the charges give rise to a potential difference between the two cylinders which can be expressed as @xmath12 where @xmath13 is the surface charge density , @xmath14 is the radius of the outer cylinder , @xmath15 the radius of the inner cylinder , and @xmath16 is the permittivity of free space . for two concentric cylinders ,
gauss law reveals there is no electric field within the region enclosed by both cylinders ; the net effect of the cylinders is to produce a constant potential shift within the region enclosed by the inner cylinder .
hence , by analogy , the effect of increasing the magnitude of the surface dipole is to increase the potential shift at the center of the nanowire
. a larger electronegativity of the terminating species increases the surface dipole leading to an increase in the magnitude of the electrostatic potential at the core region of the nanowire , or equivalently a lowering of the energy for an electron .
the quantum confinement effect works to increase the energy of a confined electron and thus the partial cancellation between the two effects with increasing surface electronegativity .
+ @xmath4-sn nw of diameter 1 nm . the contour plot shows the calculated local density of states ( ldos ) across the junction , in which the work function of the semimetallic region ( @xmath17 ) and electron affinity of the semiconducting region ( @xmath6 ) are indicated . ] the electronic structure of snnws is computed using dft @xcite .
the pbe @xcite formulation of the generalized gradient approximation ( gga ) for the exchange - correlation functional has been employed in conjunction with norm - conserving pseudopotentials @xcite and double-@xmath18-polarized basis sets consisting of localized pseudo - atomic orbitals ( paos ) @xcite . the electronic structure for the 1 nm diameter @xmath19110@xmath20 @xmath4-sn nws
are calculated with hydrogen and fluorine surface terminations with the results of the calculations presented in figure [ fig1 ] for band gap energies and electron affinities . for the hydrogen
terminated snnw , the energy band gap and electron affinity are found to be 0.68 ev and 4.24 ev , respectively . upon replacing the surface terminating species by fluorine atoms , the electron affinity changes dramatically to 7.21 ev and
the band gap nearly closes , being reduced to 30 mev .
the large influence of the fluorine atoms electron affinity is a reflection of the large surface - to - volume ratio with surface atoms accounting for @xmath21 of the chemical composition of the nanowire .
the larger electronegativity of the fluorine atoms increases charge transfer at the nw surface relative to hydrogen termination with the correspondingly increased surface dipole offsetting the quantum confinement potential .
the 30 mev band gap is of the order of @xmath22 at room temperature and hence a significant population in the conduction band by thermally excited electrons results in semimetallic behavior .
+ to generate the atomistic model for the electronic structure and transport calculations , the total electronic energy is minimized with respect to atomic positions in a region of 5 nm around the junction until computed forces remain below @xmath23 ev/ .
figure [ fig2](a ) shows the relaxed nanowire structure . in the left hand side of the figure
the snnw surface is terminated by fluorine atoms and in the right hand side the nanowire is terminated by hydrogen atoms .
the largest structural rearrangements in the nw relative to the bulk diamond structure are caused by the surface terminating species with comparably little distortion to the nw core . in the region with hydrogen terminations the nw is semiconducting whereas the region terminated by fluorine atoms the local density of states ( ldos ) indicates semimetallic behavior .
hence where the two regions meet , an abrupt semimetal - semiconductor junction is formed as shown in figure [ fig2](b ) where the ldos is plotted in the absence of a voltage bias across the junction region .
the vertical dashed line indicates the _ metallurgical junction _ defined as the region across where the surface termination changes from fluorine to hydrogen .
the resulting electronic structure clearly shows the characteristics of a schottky junction with the upward band bending indicating charge transfer from the semiconducting region to the semimetallic side of the junction . on the other hand ,
the junction differs to the common engineering of schottky diodes in that the semiconducting region induced by the hydrogen termination is intrinsic , hence the fermi level aligns to mid gap in the semiconductor region at large distances away from the junction .
the conduction and valence band bending upward in energy at the junction reflects the mismatch between the electron affinities in each region consistent with transfer of electrons from the semiconducting region to the semimetallic region . in this sense ,
the band bending of the junction resembles a metal/_n_-type semiconductor junction .
however , as the semiconducting region is an intrinsic semiconductor , the upward band bending of the conduction and valence band edges directly at the junction leads to the fermi level aligning closer to the valence band edge . in this sense ,
the fermi level alignment at the junction resembles a metal/_p_-type semiconductor junction .
these differences will lead to unconventional electron transport characteristics .
a second characteristic of the junction is the presence of evanescent states or metal - induced gap states ( migs)@xcite below the fermi level and extending up to approximately 20 nm from the semimetallic region into the semiconducting region .
similar effects have been observed for an aluminum - silicon nw junction@xcite although in their study a short silicon nanowire is connected between two metal electrodes . due to the fermi level alignment near the valence band edge
, the migs directly impact the current voltage ( iv ) characteristics across the junction , as will be seen from the charge transport calculations . the alignment of the bands at the interface results in a schottky barrier height ( usually denoted as @xmath24 ) of approximately 0.9 ev .
the electron populations in the conduction band of the intrinsic semiconducting region and for the states in the semimetallic region above the fermi level in the tail of the thermal distribution with energies near the top of the schottky barrier are both negligible .
these two conditions effectively rule out any significant electron flow between the semimetal and the semiconducting region s conduction band for typical voltage bias across the junction .
+ [ t ] to investigate the charge flow across the junction with the application of an external voltage , the non - equilibrium green s function ( negf ) formalism is applied @xcite .
figure [ fig3](a ) shows the ldos with a forward bias of 0.75 v , defined as a positive voltage applied to the semimetallic region relative to the semiconducting region , is applied across the junction . for this voltage polarity , the electronic states in the semiconducting region shift towards higher energies relative to the semimetallic region . the built - in electric field arising from charge redistribution required to equilibrate the junction at zero bias
is offset by the application of the voltage across the junction .
the conduction mechanism for forward voltage bias arises from electron tunneling from the semiconducting region s valence band into the semimetallic region . for sufficiently high voltages , the migs at the semimetal s fermi level mix with the semiconducting region s valence band states and
the device becomes strongly _ on _ , as depicted in figure [ fig4](a ) .
this behavior is reflected in the energy - resolved current where a sharp peak is observed aligned near the fermi level where these states extend into the valence band of the semiconducting region .
this mechanism is distinct to the usual schottky diode formed with an _ n_-type semiconductor under forward bias conditions for which electrons can more easily tunnel from a doped semiconductor s conduction band to the metal region as the applied voltage reduces the schottky barrier height .
reversing the voltage polarity across the junction results in the electronic structure depicted in figure [ fig3](b ) : the semiconducting region becomes increasingly depleted of electrons as the reverse bias voltage is increased , leading to an increase in band bending near the junction relative to the zero bias band profile . in this case
the quasi - fermi level in the semiconducting region is lower relative to quasi - fermi level in the semimetallic region .
as shown in at the right in figure [ fig3](b ) , electron transport occurs through band - to - band tunneling ( btbt ) whereby electrons from the semimetallic region _ tunnel _ through the semiconducting region s band gap and into the semiconducting region s conduction band .
the largest contributions to the energy resolved current are seen to correspond to migs extending into the band gap within the voltage bias window .
the migs reduce the distance required for btbt in the structure .
[ t ] figure [ fig4](a ) presents the calculated iv characteristic for the junction for forward and reverse bias voltages self - consistently calculated within the negf formalism at a temperature @xmath25 .
the asymmetry in the electron transport mechanisms with the polarity of the applied bias results in a two terminal rectifier or diode .
the reverse bias leads to very low , btbt limited transport . in the forward direction , the increasing electric field across
the junction increases coupling between the valence band edge in the semiconducting region and states near the quasi - fermi level in the semimetallic region leading to large current flows for applied voltages @xmath26 0.6 v. notable in figure [ fig4](a ) is a structure above approximately @xmath27 = 0.3 v extending to approximately @xmath27= 0.6 v. to explore this structure in the iv characteristic , transport across the junction at temperature @xmath28 was also computed and the plot in figure [ fig4](b ) gives the difference between the currents obtained at the two temperatures .
there are two predominant regions where the currents differ significantly , in the range @xmath29 corresponding to the low current _ step - like _ structure in the iv characteristic and for @xmath30 where the diode is strongly _
on_. in figure [ fig1](b ) , the equilibrium ( zero applied voltage ) electronic structure for the junction is shown . at approximately @xmath31 v
, a flat band condition is achieved in the semiconducting region when the valence band edge aligns to the semimetal quasi - fermi level . at t=300k ,
electrons in the semimetal region are readily excited above the fermi level providing partially unoccupied states below @xmath32 into which the electrons in the semiconducting region s valence band maximum ( vbm ) can tunnel .
the vbm provides an ideal conduction channel with transmission proportional to a one - dimensional density of states ( @xmath33 ) , as is reflected in the iv characteristic . at @xmath28 , there is no thermal excitation of electrons in the semimetallic region and all states below the fermi level are blocked which results in lower currents as this channel is only open at higher temperatures figure [ fig4](b)(i ) . at
higher applied voltages , as depicted in figure [ fig3](a ) , the states above the quasi - fermi level in the semimetallic region are within the voltage bias window and electrons can tunnel from the valence band in the semiconducting region into these states which are unoccupied at t=0k . for t=300k , the states above the quasi - fermi level are partially blocked due to their thermal occupations . hence tunneling into lowest states within the voltage bias window
is reduced at higher temperatures .
therefore the overall current is smaller at t=300k relative to t=0 k for @xmath34 figure [ fig4](b)(ii ) .
this property again reflects that the current transport mechanism for the surface modified semimetal junction is not the same as for conventional schottky diodes due to the differing band alignments .
these results demonstrate that schottky barriers can be formed in a
_ _ mono__material nanowire by exploiting the effects of quantum confinement and the impact of surface dipoles created by chemical terminating species with varying electronegativities .
these two effects are of similar orders of magnitude in small cross section nanowires and can be manipulated to allow for semiconducting and semimetallic regions within a single nanowire or thin film leading to the possibility to form a semimetal - semiconductor junction , akin to schottky barrier junctions formed in metal - semiconductor junctions in larger dimensions devices .
the semiconducting region that is formed in a snnw by quantum confinement results in an intrinsic material and hence the band alignments are different to conventional schottky barrier diodes which rely on substitutional doping . however , the rectifying properties of the surface modified semimetal diode are comparable to the properties of conventional macroscopic diodes , yet are achievable at molecular scales and without the requirement for substutional doping in nanostructures . |
among the results obtained with inelastic neutron scattering is the detailed information on the magnetic susceptibility in yba@xmath2cu@xmath3o@xmath4 and lanthanum cuprates which reveals a considerable difference in their magnetic properties . a sharp magnetic collective mode
the so - called resonance peak was observed at frequencies @xmath10 mev in yba@xmath2cu@xmath3o@xmath4 and some other cuprates , in underdoped case both below and above @xmath9 @xcite . in the momentum space
the mode is strongly peaked at the antiferromagnetic wave vector @xmath8 .
contrastingly , no resonance peak was observed in lanthanum cuprates .
instead for low temperatures a broad feature was detected at @xmath11 mev and for this and lower frequencies the susceptibility was found to be peaked at incommensurate momenta @xmath12 and @xmath13 @xcite . recently the low - frequency incommensurate magnetic response was also observed in yba@xmath2cu@xmath3o@xmath4 @xcite .
presently the fundamental difference of the susceptibility frequency dependencies in these two groups of cuprates is actively debated , as well as the nature of the resonance peak and the low - frequency magnetic incommensurability .
the aim of this work is to demonstrate that the above - mentioned unusual properties of cuprates can be interpreted in the framework of the @xmath0-@xmath1 model of a cu - o plane which is a common structure element of these crystals .
the model was shown to describe correctly the low - energy part of the spectrum of the realistic extended hubbard model @xcite . to take proper account of strong electron correlations inherent in moderately doped cuprate
perovskites the description in terms of hubbard operators and mori s projection operator technique @xcite are used .
the self - energy equations for hole and spin green s functions obtained in this approach are self - consistently solved for the ranges of hole concentrations @xmath14 and temperatures 2 k@xmath151200 k. lattices with 20@xmath1620 sites and larger are used .
the calculations reproduce correctly the frequency and momentum dependencies of the resonance peak in yba@xmath2cu@xmath3o@xmath4 , and its variation with doping and temperature in normal and superconducting states .
the peak is connected with the excitation branch of localized cu spins and the peak frequency is close to the frequency of these excitations at * q*. the absence of the resonance peak in lanthanum cuprates is related to an increased damping of spin excitations which is possibly connected with a large hole damping in these crystals . for low frequencies and temperatures the susceptibility is peaked at incommensurate wave vectors @xmath12 and @xmath17 . the incommensurability is connected with a dip in the magnon damping at * q*. in agreement with experiment the incommensurability parameter @xmath18 is approximately proportional to @xmath19 for @xmath20 and saturates for larger concentrations .
the incommensurability disappears with increasing temperature .
generally the incommensurate magnetic response is not accompanied by an inhomogeneity of the carrier density .
the hamiltonian of the two - dimensional @xmath0-@xmath1 model reads @xmath21 where @xmath22 is the hole annihilation operator , * n * and * m * label sites of the square lattice , @xmath23 is the spin projection , @xmath24 and @xmath25 are site states corresponding to the absence and presence of a hole on the site . these states may be considered as linear combinations of the products of the @xmath26 copper and @xmath27 oxygen orbitals of the extended hubbard model @xcite .
we take into account nearest neighbor interactions only , @xmath28 and @xmath29 where the four vectors * a * connect nearest neighbor sites .
the spin-@xmath30 operators can be written as @xmath31 and @xmath32 .
properties of the model are determined from the hole and spin retarded green s functions @xmath33\rangle , \label{green}\ ] ] where @xmath34 and @xmath35 are the fourier transforms of the respective site operators , operator time dependencies and averaging are defined with the hamiltonian @xmath36 with the chemical potential @xmath37 . as mentioned , to obtain self - energy equations for these functions we used mori s projection operator technique @xcite . in this approach
the fourier transform of green s function @xmath38 is represented by the continued fraction @xmath39 the elements of the fraction @xmath40 and @xmath41 are determined from the recursive procedure @xmath42=e_na_n+a_{n+1}+v_{n-1}a_{n-1},\nonumber\\ & & e_n=|[a_n , h]\cdot a_n^\dagger|\,|a_n\cdot a_n^\dagger|^{-1},\label{lanczos}\\ & & v_{n-1}=|a_n\cdot a_n^\dagger|\,|a_{n-1}\cdot a_{n-1}^\dagger|^{-1},\quad v_{-1}=0,\quad n=0,1,2,\ldots\nonumber\end{aligned}\ ] ] the operators @xmath43 constructed in the course of this procedure form an orthogonal set , @xmath44 . in eqs .
( [ cfraction ] ) and ( [ lanczos ] ) the definition of the inner product @xmath45 depends on the type of the considered green s function .
for example , for functions ( [ green ] ) these are @xmath46 and @xmath47\rangle$ ] , respectively . the method described by eqs .
( [ cfraction ] ) and ( [ lanczos ] ) can be straightforwardly generalized to the case of many - component operators which is necessary , for example , to consider green s functions for nambu spinors in the superconducting state @xcite .
the residual term of fraction ( [ cfraction ] ) is the fourier transform of the quantity @xmath48 where the time evolution of the operator @xmath49 is determined by the equation @xmath50,\quad a_{n , t=0}=a_n\ ] ] with the projection operators @xmath51 defined as @xmath52 .
the residual term @xmath53 is a many - particle green s function which can be estimated by the decoupling .
the decoupling is also used in calculating the elements of the continued fractions ( [ cfraction ] ) . following ref .
@xcite this approximation is improved by introducing the vertex correction @xmath54 which is determined from the constraint of zero site magnetization @xmath55 since the statistical averaging includes samples with different ordering orientations , this condition is fulfilled also in magnetically ordered states . equations for green s functions obtained in this way for the normal state read @xcite @xmath56 ^ 2 [ n_b(-\omega')+n_f(\omega-\omega')]\nonumber\\[-1ex ] & & \label{se}\\[-1ex ] & \times&a({\bf k - k'},\omega-\omega')b({\bf k'}\omega'),\nonumber\\ { \rm im}\,\pi({\bf k}\omega)&=&\frac{9\pi t^2j^2(1-x)}{2n(1-\gamma_{\bf k})(j|c_1|+tf_1)}\sum_{\bf k'}(\gamma_{\bf k+k'}-\gamma_{\bf k'})^2\nonumber\\ & \times&\int^\infty_{-\infty}d\omega ' a({\bf k'}\omega')a({\bf k+k ' } , \omega+\omega')\frac{n_f(\omega+\omega')-n_f(\omega')}{\omega},\nonumber\end{aligned}\ ] ] where @xmath57 , @xmath58 $ ] , @xmath59 and @xmath60 are the spin and hole correlations on neighboring sites which , as well as the hole concentration @xmath19 , can be calculated from green s functions , @xmath61 are the energies of spin excitations and holes with the parameter @xmath62 describing the spin gap at * q * , @xmath63 and @xmath64 are the hole and spin spectral functions , @xmath65 is the number of sites , @xmath66^{-1}$ ] , @xmath67^{-1}$ ] with the temperature @xmath68 .
real parts of the self - energies are calculated from their imaginary parts and the kramers - kronig relation .
the self - energy equations for the superconducting state can be found in ref .
@xcite .
equations ( [ constraint])([energies ] ) form a closed set which can be solved by iteration for fixed chemical potential @xmath37 and temperature @xmath68 .
we carried out such calculations for the parameters of the model @xmath69 ev , @xmath70 ev which correspond to hole - doped cuprates @xcite . to stabilize the iteration procedure an artificial damping @xmath71 , @xmath72 , was added to the frequency in the hole green s function . the results of such calculations can be directly compared with the imaginary part of the magnetic susceptibility @xmath73 derived from the neutron scattering experiments , since this quantity is connected with @xmath74 , @xmath75 where @xmath76 is the bohr magneton .
equation ( [ chi ] ) is the most general form for the magnetic susceptibility which follows from the mori formalism . with the spin excitation damping @xmath77 described by the fermion bubble ( [ se ] ) this equation acquires some similarity with the susceptibility derived in the random phase approximation @xcite which is frequently used for the discussion of the magnetic properties of cuprates . for this similarity the term @xmath78 in the denominator of eq .
( [ chi ] ) has to be negligibly small in comparison with @xmath79 . for the hole spectrum
self - consistently calculated here this can happen only in the overdoped region @xmath80 when the spin correlation @xmath81 which determines the magnitude of @xmath82 [ see eq .
( [ energies ] ) ] tends to zero @xcite .
as follows from eq .
( [ chi ] ) , the frequencies of spin excitations satisfy the equation @xmath83 as seen from eq .
( [ se ] ) , @xmath84 is finite for @xmath85 , whereas @xmath82 vanishes in this limit .
therefore the spin green s function has a purely imaginary , diffusive pole near the @xmath86 point , in compliance with the result of the hydrodynamic theory @xcite .
the dispersion of spin excitations near * q * is of special interest , because this region gives the main contribution into the neutron scattering .
this dispersion is shown in fig .
[ fig_i ] . the dispersion of spin excitations near * q * along the edge of the brillouin zone .
the dispersion was calculated in a 20@xmath1620 lattice for @xmath87 and @xmath88 k ( filled squares , the solid line is the fit with @xmath89^{1/2}$ ] ) .
open squares are the experimental dispersion @xcite of the maximum in the frequency dependence of the odd @xmath90 , @xmath91 , in yba@xmath2cu@xmath3o@xmath92 ( @xmath93 @xcite ) at @xmath94 k.,width=264 ] as seen from this figure , in contrast to the usual spin - wave theory the frequency of spin excitations at the antiferromagnetic wave vector * q * is finite .
this frequency @xmath95 is directly connected with the spin correlation length @xmath96 . indeed , using eq .
( [ se ] ) and taking into account that the region near * q * gives the main contribution to the summation over * k * , we find for large distances and low temperatures @xmath97 where @xmath98 and @xmath99 is the distance between in - plane cu sites . as follows from the calculations , for low concentrations and temperatures @xmath100 and consequently @xmath101 .
an analogous relation between the spin correlation length and the concentration has been derived from experimental data in la@xmath5sr@xmath6cuo@xmath7 @xcite . in accord with the hohenberg - mermin - wagner theorem @xcite
the considered two - dimensional system is in the paramagnetic state for @xmath102 .
this result can also be obtained using eq .
( [ se ] ) .
also it can be shown @xcite that in the infinite crystal at @xmath103 the long - range antiferromagnetic order is destroyed for the hole concentrations @xmath104 .
the imaginary part of the spin susceptibility at the antiferromagnetic wave vector , obtained in our calculations , is shown in fig .
[ fig_ii ] .
-wave superconducting ( a , b ) and normal ( c , d ) states .
curves show the results of our calculations for @xmath105 , @xmath88 k ( a , b ) and @xmath106 k ( c , d ) , @xmath87 ( a , c ) and @xmath107 ( b , d ) .
squares are the odd susceptibility measured @xcite in yba@xmath2cu@xmath3o@xmath92 ( @xmath108 k , @xmath93 , a and c ) and in yba@xmath2cu@xmath3o@xmath109 ( @xmath110 k , @xmath111 @xcite , b and d ) at @xmath94 k ( a , b ) and at @xmath112 k ( c , d ) . to depict the experimental data in the same scale they were multiplied by the factor @xmath113.,title="fig : " ] -wave superconducting ( a , b ) and normal ( c , d ) states .
curves show the results of our calculations for @xmath105 , @xmath88 k ( a , b ) and @xmath106 k ( c , d ) , @xmath87 ( a , c ) and @xmath107 ( b , d ) .
squares are the odd susceptibility measured @xcite in yba@xmath2cu@xmath3o@xmath92 ( @xmath108 k , @xmath93 , a and c ) and in yba@xmath2cu@xmath3o@xmath109 ( @xmath110 k , @xmath111 @xcite , b and d ) at @xmath94 k ( a , b ) and at @xmath112 k ( c , d ) . to depict the experimental data in the same scale they were multiplied by the factor @xmath113.,title="fig : " ] the experimental data on @xmath73 in yba@xmath2cu@xmath3o@xmath4 @xcite
are also depicted here .
yba@xmath2cu@xmath3o@xmath4 is a bilayer crystal and the symmetry allows one to divide the susceptibility into odd and even parts .
the odd part , which for the antiferromagnetic intrabilayer coupling can be compared with our single - layer calculations , is shown in the figure .
the maximum in fig .
[ fig_ii ] is the resonance peak .
as seen from this figure , the model reproduces satisfactorily the evolution of the peak position and shape with doping and temperature . in the @xmath0-@xmath1 model the peak is connected with the coherent excitations of localized cu spins near the antiferromagnetic wave vector * q * and the peak frequency for @xmath114 is close to the frequency @xmath115 in fig .
[ fig_i ] .
it is also seen in this figure that the experimental dispersion of the peak is close to the dispersion of the mentioned spin excitations .
thus , we came to conclusion that the resonance peak in yba@xmath2cu@xmath3o@xmath4 and apparently in other cuprate perovskites where it was observed is a manifestation of the magnon branch modified in the short - range antiferromagnetic order .
the above results are related to the underdoped case where the resonance peak is observed both in the normal and superconducting states . in the overdoped case in yba@xmath2cu@xmath3o@xmath4
the peak is observed in the superconducting state only @xcite .
calculations for the @xmath0-@xmath1 model show that in the normal state on approaching the overdoped region the maximum in @xmath73 rapidly loses its intensity , broadens and shifts to higher frequencies @xcite .
this result correlates with the broad feature observed experimentally in these conditions .
the mechanism of the peak reappearance in the superconducting state in the overdoped region was considered in ref .
the opening of the @xmath116-wave superconducting gap with the magnitude @xmath117 decreases considerably the spin excitation damping near * q * which restores the peak in @xmath73 . as seen from fig .
[ fig_ii ] , the resonance peak has a low - frequency shoulder which is more pronounced for low concentrations and temperatures .
the shoulder stems from the frequency dependence of the magnon damping @xmath118 which in its turn is a consequence of the hole fermi surface nesting existing in the @xmath0-@xmath1 model for moderate doping @xcite .
an analogous nesting is expected in the two - layer yba@xmath2cu@xmath3o@xmath4 between the parts of the fermi surface belonging to the bonding and antibonding bends @xcite .
it is worth noting that for all four curves in fig .
[ fig_ii ] the value of @xmath119 is smaller than @xmath120 .
the satisfactory agreement between the experimental and calculated results allows us to conclude that near * q * the spin excitations are not overdamped in underdoped yba@xmath2cu@xmath3o@xmath4 . in calculating the curves in fig .
[ fig_ii ] the artificial broadening @xmath121 in the hole green s function was set equal to @xmath122 .
this broadening simulates contributions to the hole damping from mechanisms differing from the spin excitation scattering .
the magnitude of these interactions can essentially vary in different crystals .
the example of the magnetic susceptibility calculated with the larger hole damping @xmath123 is shown in fig .
[ fig_iii ] .
k , @xmath123 and @xmath124 . the dashed line is for @xmath106 k and the same other parameters .
symbols are experimental results @xcite in la@xmath125sr@xmath126cuo@xmath7 at @xmath127 k ( filled circles ) and @xmath128 k ( open circles ) .
vertical bars show experimental errors .
the calculated susceptibility is given for @xmath129 , the experimental data are for a wave vector of the incommensurate peak .
( b ) the solid line is the same as in part ( a ) .
the dashed line is calculated for @xmath105 , other parameters are the same as for the solid line.,title="fig : " ] k , @xmath123 and @xmath124 .
the dashed line is for @xmath106 k and the same other parameters .
symbols are experimental results @xcite in la@xmath125sr@xmath126cuo@xmath7 at @xmath127 k ( filled circles ) and @xmath128 k ( open circles ) .
vertical bars show experimental errors . the calculated susceptibility is given for @xmath129 , the experimental data are for a wave vector of the incommensurate peak .
( b ) the solid line is the same as in part ( a ) .
the dashed line is calculated for @xmath105 , other parameters are the same as for the solid line.,title="fig : " ] as seen from this figure , the spin excitation damping is very sensitive to the hole damping .
the increase of this latter damping leads to a marked growth of the spin excitation damping which results in the overdamping of these excitations . instead of the pronounced peak at the frequency @xmath115 a broad low - frequency feature
is observed in the susceptibility ( see fig . [
fig_iii]b ) .
this shape resembles that observed in la@xmath5sr@xmath6cuo@xmath7 @xcite .
the comparison was carried out in fig .
[ fig_iii]a .
it should be noted that the use of a comparatively small 20@xmath1620 lattice did not allow us to describe the incommensurability of the magnetic response
@xmath73 is peaked at * q * and our calculated data belong to this momentum ( the use of a larger lattice is too time - consuming for the self - consistent calculations ) . in la@xmath125sr@xmath126cuo@xmath7
the low - frequency susceptibility is peaked at incommensurate momenta @xmath130 and the experimental data @xcite shown in fig .
[ fig_iii]a belong to one of these momenta .
these data were increased by approximately 2.5 times to show them in the scale with the calculated results .
the need for scaling is apparently connected with the splitting of the commensurate peak into the four incommensurate maxima .
apart from this difference in the momentum dependencies , the calculated frequency and temperature dependencies are in good agreement with the experimental results .
thus , it can be concluded that the dissimilarity of the frequency dependencies of the susceptibility in yba@xmath2cu@xmath3o@xmath4 and la@xmath5sr@xmath6cuo@xmath7 is connected with different values of the damping of spin excitations , which are well defined at the antiferromagnetic wave vector in the former crystal and overdamped in the latter .
this property of spin excitations in la@xmath5sr@xmath6cuo@xmath7 is not changed in the superconducting state due to the small superconducting gap @xmath131 in this crystal @xcite .
for low frequencies susceptibility ( [ chi ] ) is essentially simplified , @xmath132 as seen from fig .
[ fig_i ] and eq .
( [ energies ] ) , @xmath133 is the increasing function of the difference @xmath134 and therefore the denominator of the fraction ( [ chiloww ] ) favors the appearance of a commensurate peak centered at the antiferromagnetic momentum in the susceptibility . however , if the spin excitation damping has a pronounced dip at @xmath114 the peak splits into several peaks shifted from * q*. to make sure that the damping may really have such a behavior let us consider the case of low hole concentrations and temperatures . in this case
the hole fermi surface consists of four ellipses centered at @xmath135 @xcite .
two of them are shown in fig .
[ fig_iv ] . .
dashed lines are the fermi surface contours shifted by @xmath136 . regions of @xmath137 and @xmath138 contributing to the damping ( [ apprdamping ] ) are shaded.,width=151 ] the spin excitation damping described by the fermion bubble ( [ se ] ) is simplified to @xmath139\nonumber\\ & \times&\delta(\omega+\varepsilon_{\bf k'}-\varepsilon_{\bf k'+k})\label{apprdamping}\end{aligned}\ ] ] in the considered case . here
@xmath140 is the hole dispersion . in the process described by eq .
( [ apprdamping ] ) a fermion picks up an energy and momentum of a defunct spin excitation and is transferred from a region below the fermi level to a region above it .
however , for @xmath114 this process is impossible , because for initial states @xmath141 interior to an ellipse in fig .
[ fig_iv ] all final states with momenta @xmath138 will be inside another ellipse , i.e. below the fermi level .
thus , in this simplified picture the damping vanishes for @xmath114 and grows with increasing the difference @xmath134 approximately proportional to the shaded area in fig .
[ fig_iv ] @xcite . with increasing the hole concentration the fermi surface of the @xmath0-@xmath1 model
is transformed to a rhombus centered at * q * @xcite .
this result is in agreement with the fermi surface observed in la@xmath5sr@xmath6cuo@xmath7 @xcite [ however , to reproduce the experimental fermi surface terms describing the hole transfer to more distant coordination shells have to be taken into account in the kinetic term of hamiltonian ( [ hamiltonian ] ) ] . for such @xmath19
another mechanism of the dip formation in the damping comes into effect .
the interaction constant @xmath142 in eq .
( [ apprdamping ] ) vanishes in the so - called `` hot spots ''
the crossing points of the fermi surface with the boundary of the magnetic brillouin zone . for small frequencies and @xmath114
the nearest vicinity of these points contributes to the damping .
consequently , the damping has a dip at * q*. with the inclusion of the hole transfer to more distant coordination shells the interaction constant is changed , however the conclusion about the dip at the antiferromagnetic momentum in the damping remains unchanged .
the dip disappears when the hole damping @xmath121 exceeds the frequency @xmath143 . since the main contribution to the damping
is made by states with energies @xmath144 and @xmath145 , this means that the above consideration is valid when there exist well defined quasiparticle excitations near the fermi surface .
results of our calculations in a 120@xmath16120 lattice and experimental data are shown in fig .
[ fig_v ] .
( solid line ) , 0.06 ( dashed line ) , and 0.12 ( short dash ) . for all curves @xmath146 k and @xmath147 mev .
inset : the structure factor measured @xcite in la@xmath5sr@xmath6cuo@xmath7 for @xmath148 , @xmath149 k , @xmath150 ( filled squares and solid line ) , @xmath87 , @xmath151 k , @xmath147 mev ( triangles and dashed line ) , and @xmath107 , @xmath152 k , @xmath147 mev ( open squares and short dash ) .
fitting curves in the inset are from refs
. @xcite .
( b ) the incommensurability parameter @xmath18 vs. @xmath19 for @xmath146 k and @xmath147 mev .
inset : experimental data @xcite for la@xmath5sr@xmath6cuo@xmath7 . the wave vector and the incommensurability parameter
are given in the reciprocal lattice units @xmath153 . ] in agreement with experiment @xcite in our calculations peaks in the dynamic structure factor @xmath154^{-1}\chi''({\bf k}\omega)$ ] along the edge of the brillouin zone [ the direction @xmath155 are more intensive than those along the diagonal [ the direction @xmath156 . as seen from fig .
[ fig_v ] , the calculations reproduce correctly the main known features of the incommensurate response except that the experimental @xmath18 is somewhat larger than the calculated one . in experiment with increasing @xmath68
the incommensurability decreases and finally disappears .
the calculations reproduce this peculiarity also . in the calculations
the concentration dependence of @xmath18 is mainly determined by the spin gap parameter @xmath62 in @xmath82 in eq .
( [ energies ] ) . for @xmath157 it grows with @xmath19 and saturates at larger concentrations @xcite . with increasing @xmath158 the dip in the spin excitation damping is shallowed and finally disappears .
besides , on approaching @xmath120 the denominator in eq .
( [ chi ] ) will favor the appearance of the commensurate peak at * q*. thus , the low - frequency incommensurate maxima of @xmath73 converge to the commensurate peak at @xmath159 .
the dispersion of the maxima in @xmath73 above @xmath115 is determined by the denominator in eq .
( [ chi ] ) and is close to that shown in fig .
[ fig_i ] .
consequently , the dispersion of the maxima in @xmath73 resembles two parabolas converging near the point @xmath160 .
the upper parabola with branches pointed up reflects the dispersion of spin excitations , while the lower parabola with branches pointed down stems from the momentum dependence of the spin excitation damping .
such kind of the dispersion is indeed observed in cuprates @xcite . generally the incommensurate magnetic response is not accompanied by an inhomogeneity of the carrier density . in works based on the stripe mechanism @xcite the magnetic incommensurability
is connected with the appearance of a charge density wave .
notice however , that the magnetic incommensurability is observed in all lanthanum cuprates , while the charge density wave is detected in neutron scattering in those cuprates which are in the low - temperature tetragonal or less - orthorhombic phases ( la@xmath5ba@xmath6cuo@xmath7 , la@xmath161nd@xmath162sr@xmath6cuo@xmath7 ) @xcite . in these phases the charge density wave
is stabilized by the corrugated pattern of the in - plane lattice potential .
it can be supposed that the magnetic incommensurability may be a precursor rather than a consequence of the charge density wave .
in this paper the projection operator technique was used for investigating the magnetic properties of the @xmath0-@xmath1 model of cuprate high-@xmath9 superconductors .
it was demonstrated that the calculations reproduce correctly the frequency and momentum dependencies of the experimental magnetic susceptibility and its variation with doping and temperature in the normal and superconducting states in yba@xmath2cu@xmath3o@xmath4 and lanthanum cuprates .
this comparison with experiment allowed us to associate the resonance peak in yba@xmath2cu@xmath3o@xmath4 with the magnon branch modified in the short - range antiferromagnetic order .
the lack of the resonance peak in la@xmath5sr@xmath6cuo@xmath7 was connected with an increased damping of spin excitations at the antiferromagnetic wave vector @xmath8 .
one of the possible reasons for the overdamped excitations is an increased hole damping in this crystal .
it was shown that for low frequencies the susceptibility is peaked at incommensurate momenta @xmath12 and @xmath13 .
the incommensurability is the consequence of the dip in the momentum dependence of the spin excitation damping at * q*. the dip appears due to the fermi surface nesting for low hole concentrations @xmath19 and due to small hole - magnon interaction constants for moderate concentrations . in agreement with experiment for @xmath163
the incommensurability grows nearly proportional to @xmath19 and tends to saturation for @xmath164 .
this is connected with the concentration dependence of the spin excitation frequency at * q*. also in agreement with experiment the incommensurability decreases with increasing temperature .
the dispersion of the maxima in the susceptibility resembles two converging parabolas .
the upper parabola with branches pointed up reflects the dispersion of spin excitations , while the lower parabola with branches pointed down stems from the momentum dependence of the spin excitation damping . in the considered mechanism the magnetic incommensurability
is not accompanied by the inhomogeneity of the carrier density .
p. bourges , in _ the gap symmetry and fluctuations in high temperature superconductors _ , edited by j. bok , g. deutscher , d. pavuna , and s. a. wolf ( plenum press , 1998 ) , p. 349 ; h. he , y. sidis , p. bourges , g. d. gu , a. ivanov , n. koshizuka , b. liang , c. t. lin , l. p. regnault , e. schoenherr , and b. keimer , phys .
rev.lett . * 86 * , 1610 ( 2001 ) .
m. arai , t. nishijima , y. endoh , t. egami , s. tajima , k. tamimoto , y. shiohara , m. takahashi , a. garrett , and s. m. bennington , phys .
* 83 * , 608 ( 1999 ) ; d. reznik , p. bourges , l. pintschovius , y. endoh , y. sidis , t. matsui , and s. tajima , cond - mat/0307591 .
a. k. mcmahan , j. f. annett , and r. m. martin , phys.rev .
b * 42 * , 6268 ( 1990 ) ; v. a. gavrichkov , s. g. ovchinnikov , a. a. borisov , and e. g. goryachev , zh . eksp .
118 * , 422 ( 2000 ) [ jetp ( russia ) * 91 * , 369 ( 2000 ) ] .
x. j. zhou , t. yoshida , d .- h .
lee , w. l. yang , v. brouet , f. zhou , w. x. ti , j. w. xiong , z. x. zhao , t. sasagawa , t. kakeshita , h. eisaki , s. uchida , a. fujimori , z. hussain , and z .- x .
shen , cond - mat/0403181 .
h. yoshizawa , s. mitsuda , h. kitazawa , and k. katsumata , j. phys .
* 57 * , 3686 ( 1988 ) ; r. j. birgeneau , y. endoh , y. hidaka , k. kakurai , m. a. kastner , t. murakami , g. shirane , t. r. thurston , and k. yamada , phys .
b * 39 * , 2868 ( 1989 ) . | the magnetic susceptibility of underdoped yttrium and lanthanum cuprates is interpreted based on the self - consistent solution of the @xmath0-@xmath1 model of a cu - o plane .
the calculations reproduce correctly the frequency dependencies of the susceptibility in yba@xmath2cu@xmath3o@xmath4 and la@xmath5sr@xmath6cuo@xmath7 attributing their dissimilarity to the difference in the damping of spin excitations . in yba@xmath2cu@xmath3o@xmath4
these excitations are well defined at the antiferromagnetic wave vector @xmath8 even in the normal state which manifests itself in a pronounced maximum the resonance peak in the susceptibility . in la@xmath5sr@xmath6cuo@xmath7
the spin excitations are overdamped which leads to a broad low - frequency feature in the susceptibility .
the low - frequency incommensurability in the magnetic response is attributed to a dip in the magnon damping at * q*. the calculated concentration and temperature dependencies of the incommensurability parameter conform with experimental observations .
generally the incommensurate magnetic response is not accompanied with an inhomogeneity of the carrier density .
high-@xmath9 superconductors , magnetic properties , the @xmath0-@xmath1 model . |
theoretical analyses of the tidal interaction between close binaries can be classified according to whether an equilibrium tide is assumed or the dynamical tide is taken into account .
the theory of the equilibrium tide is based on the assumption that a star subject to the tidal disturbance of a companion instantly adjusts to hydrostatic equilibrium ( darwin 1879 ) . a calculation including the dynamical tide takes into account the fact that gravity or g modes can be excited in the convectively stable layers of the star and that resonances between the tidal disturbance and the normal modes of the star can occur ( cowling 1941 ) .
so far , dynamical tides have been studied only in massive close binaries , which have a convective core and a radiative envelope ( zahn 1975 , 1977 ; savonije & papaloizou 1983 , 1984 , 1997 ; papaloizou & savonije 1985 , 1997 ; savonije , papaloizou & alberts 1995 ) . in this paper , we examine the effect of dynamical tides excited by a companion on a solar type star , in which a radiative core is surrounded by a convective envelope .
this is of particular interest in connection with circularization of solar type binaries .
it has been proposed that circularization occurs through the action of turbulent viscosity , originating in the convective envelope , on the tide .
however , according to claret & cunha ( 1997 ) ( see also goodman & oh 1997 ) , who have used the equilibrium tide formalism of zahn ( 1989 ) , the circularization rate resulting from this mechanism is too small by about two orders of magnitude to account for the circularization timescales required on the main sequence .
the tidal response calculation undertaken here is also of interest in connection with the newly discovered planets , some of which are found to orbit around solar type stars with a period comparable to that of the high order g modes of the star .
one such example is 51 pegasi ( mayor & queloz 1995 ; marcy & butler 1995 ) . in these binaries ,
g mode oscillations are excited by the companion in the radiative region beneath the convective envelope .
they become evanescent in the convection zone where they are damped by their interaction with the convective eddies .
this dissipation leads to an exchange of angular momentum between the star and the orbit if the stellar rotation and the orbital motion are not synchronized .
here we assume that the orbital frequency is initially larger than the rotational frequency of the star .
then tidal interaction results in the decay of the orbit and the spin up of the star .
if the mass of the secondary companion is considerably smaller than that of the primary , the timescale for orbital decay is smaller than the stellar spin up timescale , and the companion eventually plunges into the primary .
but if the mass of the companion is large enough , synchronization may occur before the binary has merged , stopping further orbital decay .
estimates based on the theory of the equilibrium tide ( rasio _ et al . _ 1996
; marcy _ et al . _
1997 ) suggest that the orbital decay timescale and the stellar spin up timescale for a system like 51 pegasi are longer than the inferred age of the primary if the companion is a jovian like planet . in this paper
we examine the effect of resonances on these timescales , and determine the potentially observable magnitude of dynamical tides at the photosphere of a solar type star .
we also compare the dynamical tide calculations with the results of an asymptotic analysis we carry out in the limit of small frequencies which should correspond to the adiabatic equilibrium tide theory .
the paper is organized as follows : in [ sec : analysis ] , we study the tidal response of the star to the perturbation by a companion in a circular orbit with a period in the range 413 days . in [ sec : linearized ] , we first consider the linear adiabatic response and then , away from resonance with a g mode , extend the analysis using first order perturbation theory to calculate the torque due to dissipation in the convective envelope . this mechanism is then found to be more important than non adiabaticity arising from heat transport in the radiative interior ( i.e. radiative damping ) . however
, this is not the case in the vicinity of a g mode resonance .
there we also calculate the torque due to non adiabaticity in the radiative core using a wkb treatment of the non adiabatic terms .
we find that the torque at effective resonances is mainly determined by radiative damping .
an analysis valid for very low frequencies ( _ equilibrium tide _ ) is given in
[ sec : asy ] .
the orbital circularization timescale for systems initially in non
circular orbits can be derived from the response calculations for companions in circular orbits .
this is done in [ sec : times ] .
we then discuss how this might be used to calibrate the magnitude of the turbulent viscosity required to fit the observations in
[ sec : cal ] .
numerical calculations are presented in [ sec : results ] .
the results assuming the equilibrium tide are given in
[ sec : eqres ] . in
[ sec : dyres ] we present the results of the dynamical tide calculations .
we describe the tidal response of the star to a companion in circular orbit , give the induced velocity at its surface and the tidal torque .
we describe the resonances and show that , for the periods of interest of several days , they are not expected to affect the orbital evolution of the binary . in [ sec : comp ] we compare the calculations based on the dynamical and equilibrium tides .
we find that , for the standard solar model , at the orbital periods of interest , because of the long timescale associated with convection , the equilibrium tide calculations give a relatively poor approximation to the results of the dynamical tide calculations .
we find that the viscosity that is required to provide the observed circularization rates is about 50 times larger than that simply estimated from mixing length theory and discuss the means by which this viscosity might be enhanced in [ sec : calibration ] .
however , we note that _ the strength of the resonances for orbital periods larger than @xmath0 days and the perturbed velocity at the surface of the star are insensitive to the magnitude of the turbulent viscosity assumed_. only for periods @xmath1 days is the strength of the resonances decreased by a factor @xmath1 .
the observable width of the resonances as well is reduced when the viscosity is increased .
we also give the relation between the orbital evolution , circularization and spin up timescales and the orbital frequency in [ sec : fitting ] . finally in
[ sec : discussion ] we discuss and summarize our results , applying them to 51 pegasi in [ sec : peg ] .
the calculations presented in this section will be applied to close binary systems where the primary is a solar type star and the secondary a stellar or planetary companion .
the orbital periods of interest lie in the range @xmath2@xmath3 days .
the rotational angular velocity of the primary is assumed to be small compared to the orbital frequency , so that it can be neglected .
quadrupolar tidal forcing thus occurs through potential perturbations with periods in the range @xmath4@xmath5 days .
when calculating the tidal response well away from a condition of resonance with a g mode , we firstly calculate the tidal response assuming it to be adiabatic throughout the star . first
order perturbation theory is then used to calculate the dissipation occurring in the convective envelope .
the idea here ( as is borne out by the numerical results ) is that although short wavelength g modes are excited in the radiative core , when they are away from resonance they do not play an important role in comparison to the global component of the tidal response .
also the variations in the convective envelope occur on a comparatively long length scale , making the adiabatic approximation a reasonable one .
when there is a resonance with a high order g mode , the response becomes one with a very short length scale such that non
adiabaticity in the radiative core can not be neglected .
however , the modes are of high order such that a wkb treatment of the non adiabatic effects is possible and this is used close to resonance where the normal mode dominates the response . such non adiabatic effects turn out to be more important than the action of turbulent viscosity in the convective envelope , with the torque at significant resonances being determined mainly by non adiabatic effects . the linearized momentum , mass , and energy equations governing the adiabatic response of the non rotating star to the perturbing potential @xmath6 may be written ( see , for example , unno _ et al . _ 1989 )
@xmath7 where @xmath8 is the pressure , @xmath9 is the density , @xmath10 is the entropy , @xmath11 is the lagrangian displacement vector , @xmath12 denotes the lagrangian perturbation and the primed quantities are eulerian perturbations .
we use the cowling ( 1941 ) approximation , applicable to stars with high central condensation , which neglects the perturbation to the stellar gravitational potential .
we also have the thermodynamic relation @xmath13 where @xmath14 is the temperature and @xmath15 and @xmath16 are the adiabatic exponents of chandrasekhar .
this relation together with equation ( [ energy ] ) leads to @xmath17 where @xmath18 with @xmath19 being the brunt - visl frequency and @xmath20 the acceleration due to gravity .
following cowling ( 1941 ) , only the dominant quadrupole term is considered in the perturbing potential due to the companion . for a binary system with a circular orbit
, this is given in spherical polar coordinates @xmath21 by the real part of @xmath22 where the spherical harmonic @xmath23 with @xmath24 , @xmath25 being the associated legendre polynomial with indices @xmath26 and @xmath27 . here
@xmath28 is the orbital angular velocity , @xmath29 where @xmath30 is the orbital separation , and @xmath31 is the mass of the companion .
we adopt the same angular and time dependence for the perturbations , so that @xmath32 @xmath33 and @xmath34 are proportional to @xmath35 .
the corresponding expression for the lagrangian displacement is @xmath36 y_{n , m}(\theta,\varphi ) e^{- i m \omega t } .
\label{xi}\ ] ] the factor @xmath37 will be henceforth taken as read , so that hereafter the perturbations will be taken to depend only on @xmath38 .
physical perturbations are then found by taking real parts after inserting this factor .
the horizontal displacement @xmath39 is given by the non radial equation of motion ( [ momentum ] ) : @xmath40 this relation together with equation ( [ rhop ] ) allow @xmath41 and @xmath33 to be eliminated from the system ( [ momentum])([energy ] ) .
the radial equation of motion ( [ momentum ] ) and the continuity equation ( [ mass ] ) can then be written as a pair of ordinary differential equations for @xmath42 and @xmath39 : @xmath43 \xi_h + \frac{f r^2 \rho}{\gamma_1 p } , \label{ode1 } \\
\frac{d \xi_h}{dr } & = & \frac{1}{r } \left ( 1 - \frac{ap}{m^2 \omega^2 \rho } \frac{dlnp}{dr } \right ) \xi_r - \left ( a + \frac{1}{r } \right ) \xi_h + \frac{afr}{m^2 \omega^2 } , \label{ode2}\end{aligned}\ ] ] the solution of which requires two boundary conditions . at the surface of the star we take a free boundary : @xmath44 , i.e. @xmath45 the boundary condition at @xmath46 , where equations ( [ ode1 ] ) and ( [ ode2 ] ) have a regular singularity , is that the solutions be regular . since at @xmath47 @xmath48 , @xmath49 and @xmath50 , where @xmath51 and @xmath52 are respectively the central pressure and density , this leads to @xmath53 .
the interaction between convective motions and the tidal flow is expected to lead to the dissipation of tidally excited waves ( e.g. zahn 1977 ) .
we model this effect as arising from a turbulent viscosity .
to do this we suppose there is an additional dissipative force per unit mass acting in the convection zone given in spherical coordinates by : @xmath54 where @xmath55 is the flow velocity and @xmath56 is the turbulent viscosity . here
, we assume that variation in the radial direction is the most significant and note that the viscous force is defined in such a way as to lead to a positive definite energy dissipation rate . for the turbulent viscosity coefficient
, we take ( see , for example , xiong , cheng & deng 1997 ) @xmath57 where @xmath58 and @xmath59 are two constants , @xmath58 being on the order of unity , @xmath60 is the orbital period , @xmath61 is the mixing length , and @xmath62 is the convective timescale .
the viscosity is then @xmath63 for small forcing frequency @xmath64 .
the factor @xmath65 , where we shall use @xmath66 , allows for a reduction of efficiency of the damping of high frequency oscillations to which the convection can not adjust .
a similar prescription with @xmath67 has been proposed by zahn ( 1966 ) and used by zahn ( 1989 ) , whereas @xmath66 has been considered by goldreich & keeley ( 1977 ) and used by campbell & papaloizou ( 1983 ) and goldman & mazeh ( 1991 ) .
goodman & oh ( 1997 ) have also recently put forward some arguments in favor of @xmath66 . for the mixing length
, we shall take the standard relation @xmath68 and set @xmath69 . in principle
, equations ( [ momentum])([energy ] ) should be solved with @xmath70 added on the right hand side of equation ( [ momentum ] ) in the convective envelope . however
, this would increase the order of the differential system to be solved and make the numerical calculations much more complicated .
instead , we have found it adequate to solve first the adiabatic problem and then to treat the dissipative effect using a first
order perturbation theory .
this is valid everywhere , except very close to resonances , because dissipation is weak enough so that the imaginary parts of @xmath42 and @xmath39 are much smaller in magnitude than their real parts .
thus we solve equations ( [ momentum])([energy ] ) without dissipative terms , and use these ( real ) solutions to calculate the rate of energy dissipation @xmath71 due to convection which is given by : @xmath72 where the integration is over the volume @xmath73 of the convective envelope and the angular dependence of @xmath70 and @xmath55 has to be taken into account . using the relation @xmath74 and the expression ( [ xi ] ) for @xmath11 , we get @xmath75 dr , \label{pt1}\ ] ] where @xmath76 and @xmath77 are respectively the inner and outer radius of the convective envelope . noting that the ratio of the rate of exchange of energy and angular momentum with the orbit is given by the pattern speed of the tidal disturbance , @xmath78 the torque exerted by the companion on the star is given by @xmath79 this torque is positive because the star is non rotating .
when the frequency of the tidal wave is equal to that of some adiabatic normal mode frequency of the star , we can no longer use first order perturbation theory because it would give an infinite torque .
however , when the frequency @xmath28 is very close to a resonant frequency @xmath80 , the torque will be given by an expression of the form @xmath81 where @xmath82 is an amplitude and @xmath83 is the damping rate for the mode . first
order perturbation theory assumes dissipative effects are small in the response calculation .
therefore it is valid only for frequencies such that @xmath84 .
however , the damping rate , if small , can be found from first
order perturbation theory applied , as described above , very close to resonance where the mode dominates the response .
then it is given by ( see , for example , goldstein 1980 ) @xmath85 very close to resonance , where @xmath86 is the kinetic energy of the mode : @xmath87 ^2 dv = \frac{24 \pi}{5 } m^2 \omega^2 \int_0^{r_{\sun } } \rho r^2 \left ( \xi_r^2 + 6 \xi_h^2 \right ) dr,\ ] ] the integration being over the volume @xmath88 of the star ( in the first integral , the angular dependence of @xmath55 has to be taken into account ) . in ( [ gamma ] ) , the total energy of the mode is @xmath89 because at resonance there is equipartition between kinetic and potential energy .
we calculate @xmath83 using ( [ gamma ] ) as outlined above making sure that @xmath28 is close enough to @xmath80 by checking that @xmath83 remains approximately constant when @xmath28 is slightly changed .
to get @xmath90 we fit the torque obtained from first
order perturbation theory to the expression ( [ restor ] ) in the region approaching the resonance where still @xmath91 .
this procedure works satisfactorily when @xmath83 is small with consequent strong resonances .
this appears to be the situation when turbulent viscosity alone is assumed to act .
however , radiative damping can not be neglected for resonances at low forcing frequency and this is discussed below .
non adiabatic effects become important when the radiative diffusion time across the length scale associated with the tidal response shortens to become comparable to the wave propagation time across the star . in principle , these effects should be taken into account both in the radiative core and above the convection zone . however , since g modes , which are excited in the radiative core , are evanescent in the convective envelope , we do not expect non adiabaticity to play an important role above the convection zone .
to take into account non adiabatic effects , equation ( [ energy ] ) has to be modified in the radiative core to : @xmath92 where @xmath93 is the perturbed radiative flux .
the radiative flux is given by the radiative diffusion equation @xmath94 where @xmath14 is the temperature and @xmath95 is the radiative conductivity , with @xmath96 being the stefan
boltzmann radiation constant , @xmath97 the velocity of light and @xmath98 the opacity .
therefore @xmath99 - \mbox{\boldmath $ \nabla$}_h^2 \left ( kt ' \right ) , \label{grflux}\ ] ] where @xmath100 is the non radial component of the operator @xmath101 .
we now suppose that close to resonance , the response behaves exactly like a free g mode with very large radial wavenumber @xmath102 so that wkb theory can be used together with the local dispersion relation to evaluate @xmath103 .
the dominant term in the right hand side of ( [ grflux ] ) is then @xmath104 , and all the other terms can be neglected . for a high order free g mode ,
the local dispersion relation gives ( see , for example , unno _ et al .
_ 1989 ) : @xmath105 this expression of @xmath102 is derived under the adiabatic approximation . however , since we want to incorporate non
adiabatic effects to the lowest order , we do not need to take them into account in evaluating @xmath102 .
we have @xmath106 in the radiative core because @xmath107 there .
the perturbed temperature @xmath108 can be expressed as a function of @xmath41 and @xmath33 using the thermodynamic relation @xmath109 where @xmath110 and @xmath111 .
the equation of state in the radiative core of a solar type star is primarly that of a perfect gas ( we neglect here the radiation pressure which is very small compared to the gas pressure ) .
we then have @xmath112 and @xmath113 . using the above , the fact that @xmath114 , @xmath115 and the relation ( [ thermo ] ) , we can recast ( [ nad ] ) under the form @xmath116 where @xmath117 for high order free g modes ,
we have ( see , for example , unno _ et al .
_ 1989 ) @xmath118 which means that @xmath119 in the radiative core of a solar type star at the frequencies of interest .
the non adiabatic correction of the term associated with pressure in equation ( [ nad1 ] ) is then very small compared to the non adiabatic correction of the term involving @xmath42 .
therefore equation ( [ nad1 ] ) can be approximated by @xmath120 where we have defined @xmath121 .
we note that because we have identified the tidal response with a normal mode , this calculation is valid only close to resonances .
equation ( [ rhopnad ] ) is similar to ( [ rhop ] ) but with @xmath122 being replaced by @xmath123 . the system of differential equations we have to solve to get the non
adiabatic response of the star is then the same as ( [ ode1])([ode2 ] ) but with the following modification in the radiative core . in equations
( [ ode1 ] ) and ( [ ode2 ] ) , @xmath124 where it appears as a coefficient of @xmath42 , has to be replaced by @xmath123 .
the system of differential equations so obtained is complex .
in general , for the periods we consider , @xmath125 is small .
we then calculate both the real and imaginary parts of the response , so that the torque can be computed directly from : @xmath126 where the integral is over the volume @xmath88 of the star .
the angular dependence of @xmath6 and @xmath33 has to be taken into account in this expression .
equation ( [ torgen ] ) can be recast under the form ( savonije & papaloizou 1983 ) : @xmath127 as in [ sec : torque ] , we can calculate the damping rate @xmath128 in resonances due to non
adiabatic effects using ( [ gamma ] ) . here
, the rate of energy dissipation is calculated from the torque ( see [ det ] ) . in the limit of small @xmath129
, the following relations may be obtained for the _ adiabatic equilibrium tide _ : @xmath130 and , if @xmath131 : @xmath132 where the ` _ eq _ ' subscript denotes the equilibrium value . using these displacements ,
the torque may be calculated using equations ( [ pt1 ] ) and ( [ det ] ) .
however , we comment that equation ( [ xieq ] ) , which states that the lagrangian perturbation to the pressure is zero , can only be derived in the adiabatic low @xmath129 limit if the brunt - visl frequency is not zero ( in practice , one also requires that the forcing period be short compared to the thermal timescale of the star , but the latter is so long that it can be assumed to be infinite in this context ) .
equations ( [ xieq])([weq ] ) do not apply in a finite region where strictly @xmath133 in that case the fluid is locally barotropic , and the displacement can be written as the gradient of a potential : @xmath134.\ ] ] the continuity equation then gives for low frequencies @xmath135 equation ( [ crp ] ) gives a second order differential equation for @xmath136 this applies inside the region where @xmath133 it is possible , using the two available boundary conditions for ( [ crp ] ) , to match @xmath137 given by ( [ xieq ] ) at the boundaries of such a region , but not in general @xmath138 given by ( [ weq ] ) .
this means that there will tend to be a discontinuity in the tangential displacement at the boundaries for low frequencies .
when @xmath139 is not zero but very small , in particular small compared to @xmath140 which corresponds physically to the convective timescale being much longer than the forcing period , the tidal response more closely matches that given by ( [ crp ] ) than that given by equations ( [ xieq])([weq ] ) .
this feature causes a very slow convergence towards the low frequency limiting solution ( _ equilibrium tide _ ) found here , as well as near discontinuous behavior near the inner convective envelope boundary .
this is borne out by our numerical results ( see [ sec : velocity ] and [ sec : comp ] ) .
the torque , @xmath141 gives the rate at which angular momentum is transferred from the orbit to the star .
we can then calculate a tidal evolution ( decay ) timescale of the circular orbit : @xmath142 where @xmath143 is the reduced mass . in principle , the variation of the torque with @xmath28 has to be taken into account for the total decay time to be calculated .
however , since the torque increases as the companion spirals in , @xmath144 is mainly determined by the initial position of the companion , and a good estimate can be obtained using the above formula .
this exchange of angular momentum also results in the spin up of the star , the timescale of which is given by @xmath145 ( savonije & papaloizou 1983 ) , with @xmath146 being the stellar moment of inertia .
out of resonance , angular momentum deposition _ initially _ occurs mainly in the convective envelope where the turbulent viscosity acts ( goldreich & nicholson 1989 ) .
it is then of interest to calculate the spin up timescale for the convective envelope alone , which is @xmath147 , where @xmath148 is the moment of inertia of the convection zone .
we note however that on the long timescales associated with tidal evolution , angular momentum may be redistributed throughout the star . in practice ,
we find that the torque arising from a companion in a circular orbit varies with frequency approximately @xmath149 this result can be used to relate the orbital circularization timescale to the initial orbital decay timescale provided that the eccentricity is not too large . in practice ,
both these timescales can be significantly longer than the spin up timescale of the star due to its relatively small moment of inertia .
we should then consider that the star is synchronized with the orbit .
the ratio between the orbital decay timescale and the circularization timescale @xmath150 is found to be about 6 for the calculated frequency dependence of the circular orbit torque ( see , for example , the expressions given in savonije & papaloizou 1983 , 1984 ) .
this appears to be independent of whether the star is assumed to be synchronously rotating or non rotating in that the circularization rate calculated assuming no rotation , as we do here , gives a reasonable estimate in the synchronous case also .
in addition , to evaluate @xmath150 for an equal mass system , we have to take into account the reciprocal torque exerted by the primary on the companion .
to do this , we take @xmath150 to be proportional to a factor which is @xmath151 when @xmath152 and 1 when @xmath153 . to a reasonable approximation , we then get : @xmath154 one of the purposes of this paper is to calculate the tidally induced velocities on the star . in order to do this , the processes responsible for dissipating the disturbance should be included as accurately as possible .
an important dissipation process is that associated with turbulent friction in the convection zone .
as this process has been suggested as being responsible for circularizing the orbits of main sequence binaries of sufficiently short period ( see mathieu 1994 ) , we investigate whether reasonable assumptions about the behavior of turbulent viscosity can lead to the observed circularization rates .
zahn & bouchet ( 1989 ) have investigated the pre
main sequence evolution of late type binaries in which the stars are fully convective .
the main conclusion of their work was that circularization occurs at the very beginning of the hayashi phase , with hardly any decrease of the eccentricity on the main sequence .
the cutoff period they predict is between 7.3 and 8.5 days . according to them ,
observations show a cutoff period around 8 days independent on the age of the star and are then in agreement with their results .
however , we note that for this agreement to be reached , some observations had to be discarded . those by mayor & mermilliod ( 1984 ) , which indicated that the cutoff period of late type binaries was at most 5.7 days , and those by mathieu & mazeh ( 1988 ) , which showed that the cutoff period in the 4 @xmath155 old cluster m 67 was more than 10 days . in a review article ,
mathieu ( 1994 ) ( see also mathieu 1992 and mathieu _ et al . _
1992 ) has confirmed that `` cutoff periods increase with age , consistent with active main sequence tidal circularization '' .
main sequence cutoff period is very likely to be 4.3 @xmath156 ( an upper limit being 6.4 @xmath156 ) , and cutoff periods for solar mass binaries are 7.05 @xmath156 in the pleiades ( 0.1 @xmath155 ) , 8.5 @xmath156 in the hyades ( 0.8 @xmath155 ) , 12.4 @xmath156 in m 67 ( 4 @xmath155 ) and 18.7 @xmath156 in the halo field ( 16 @xmath155 ) .
we therefore conclude that active circularization does take place for main sequence binaries and that it is less efficient than proposed by zahn & bouchet ( 1989 ) on the pre main sequence .
recently , claret & cunha ( 1997 ) have applied the formalism developed by zahn ( 1989 ) to different stellar models .
they have computed the parameters which enter into the expression for the circularization timescale , which is based on treatment of the equilibrium tide , for a wide grid of stellar models as a function of mass and time .
their conclusion is that turbulent dissipation is too low by a factor 100200 during the main sequence to fit the observed cutoff periods .
the calculations presented in the following section are applied to the standard solar model described by christensen
et al . _ ( 1996 ) .
the torque associated with the equilibrium tide was calculated as indicated in
[ sec : asy ] .
as mentioned there , this calculation is only expected to apply at very low frequencies . for periods between 4.23 and 12.4 @xmath156 , and @xmath157
, the calculated torque can be interpolated by the following power law : @xmath158 for the calculation of the dynamical tide , we solve numerically the differential equations ( [ ode1 ] ) and ( [ ode2 ] ) using a shooting method to an intermediate fitting point ( press _ et al . _
1986 ) . to evaluate non
adiabatic effects in the radiative core close to resonances , we modify these equations in the way described in [ sec : torquenad ] .
we define the dimensionless quantity @xmath159 where @xmath77 is the outer radius of the convective envelope . with this notation , the equations are integrated from @xmath160 to @xmath161 the radiative core extends from @xmath162 to @xmath163 .
the results presented below have been obtained with the values @xmath157 and @xmath69 ( @xmath164 being the ratio of the mixing length @xmath61 to the pressure scale height ) .
we discuss the effect of changing these parameters . for illustration purposes , we plot the horizontal and radial displacements induced in the star at orbital periods of @xmath165 @xmath156 and 8.46 @xmath156 away from resonance in the adiabatic approximation .
the first period is that inferred for the system 51 pegasi .
the spatial distribution of the real parts of @xmath166 and @xmath167 are shown in figure [ fig1 ] .
these represent typical values of the radial and horizontal velocities , the maximum values being three and six times larger respectively .
since these quantities depend on the perturbing mass through the ratio @xmath168 , they have been represented in units of this factor .
as expected , the stellar response shows oscillations between turning points near the center and the inner radius of the convection zone , where @xmath169 otherwise it is evanescent .
the horizontal displacement varies rapidly in the photosphere because the temperature drops to zero rapidly there .
we see from figure [ fig1 ] that , when the perturbing mass @xmath152 , the maximum radial velocity at the surface of the star is about 6 and 1 @xmath170 for @xmath165 @xmath156 and 8.46 @xmath156 respectively .
these numbers drop to @xmath171 and @xmath172 respectively when the perturbing mass is one jupiter mass ( @xmath173 ) .
the radial displacement and the perturbed pressure at the surface of the star are well approximated by the equilibrium values ( [ xieq ] ) and ( [ peq ] ) respectively .
these quantities are found to be insensitive to the existence of resonances with the consequence that the radial velocity at the surface of the star never differs much from the values given above .
for the smallest periods considered , the ratio @xmath174 at the surface of the star can vary by up to one order of magnitude on passage through resonance .
this is due to the fact that @xmath39 is proportional to @xmath175 ( see [ xih ] ) . even though this ratio remains small
, it can vary by up to an order of magnitude as a resonance is passed through .
the numerical results indicate that both the amplitude and the wavelength of the response increase with the orbital frequency , in agreement with the theoretical expectation of a smaller radial order for higher frequencies ( christensen dalsgaard & berthomieu 1991 and references therein ) . finally , as expected ( see [ sec : asy ] ) , the plots shown in figure [ fig1 ] ( see also figure [ fig3 ] ) indicate that at the boundary of the radiative core and the convection zone there is a near discontinuity in the mean value of @xmath39 obtained after averaging out the interior oscillations . _
i ) resonances _ : figure [ fig2 ] shows @xmath176 versus @xmath28 in a log log representation for three different small frequency intervals in the vicinity of @xmath165 , 8.5 and 12.4 @xmath156 . on each plot , the dotted line gives the values obtained from the theory of the equilibrium tide as given by ( [ asyint ] ) .
since the torque depends on the perturbing mass through the factor @xmath177 , it has been represented in units of this .
these plots show several resonances , which occur when the frequency of the tidal wave is equal to the frequency of some normal mode of the star .
the left panels show the torque arising from convective dissipation , through turbulent viscosity , alone ( see [ sec : torque ] ) .
these plots have been displayed for comparison with the right panels , for which radiative damping has been taken into account in the resonances ( see [ sec : torquenad ] ) . as indicated by the table [ table1 ] ,
the normal mode damping rate due to radiative damping ( @xmath128 ) is much larger than that due to convective dissipation ( @xmath83 ) .
ccccccccccccccccccccccccccccccccc @xmath178 & @xmath28 & @xmath83 & @xmath128 + ( days ) & ( @xmath179 ) & ( @xmath179 ) & ( @xmath179 ) + 4.23 & @xmath180 & @xmath181 & @xmath182 + 8.5 & @xmath183 & @xmath184 & @xmath185 + 12.4 & @xmath186 & @xmath187 & @xmath188 + [ table1 ] thus the torque in the center of resonances , where they are significant , is predominantly determined by radiative damping . for the frequencies we consider ,
this contribution to the torque becomes much smaller than that due to turbulent viscosity in the tail of the resonances .
this means that non
adiabatic effects in the radiative core are negligible away from resonances .
therefore , in the right panels of figure [ fig2 ] , we have plotted the torque resulting from radiative damping acting alone in the center of resonances , and that resulting from convective dissipation acting alone away from resonances .
a comparison between the strength of the resonances shown in the right and left panels indicates the importance of non adiabatic effects in the radiative core .
as expected , the resonances are weakened and broadened , this effect being marginally important for @xmath189 @xmath156 .
we now discuss the properties and the effect of resonances on the tidal torque . from now on , when resonances are discussed
, we shall refer to the calculations which take into account radiative damping . in the neighborhood of @xmath165 , 8.5 and 12.4 @xmath190 the relative separation ,
@xmath191 between successive resonances is respectively @xmath192 @xmath172 and @xmath193 the relative width , @xmath194 of the resonances is respectively @xmath195 @xmath196 and @xmath197 here we have arbitrarily defined the width of a resonance as being the frequency interval over which @xmath176 is at least 3 times larger than the minimum torque obtained just out of this resonance . to calculate the probability of the companion being in a resonance , we have to take into account the fact that it drifts away from the resonances much more rapidly than elsewhere .
the relevant quantity for calculating the tidal evolution timescale is @xmath198 ( see expression [ torb1 ] ) . for a fixed oscillation spectrum , we can approximate this probability by @xmath199 times the ratio of the mean value of @xmath198 over a resonance to the mean value of @xmath198 between two resonances , where the mean value is defined by @xmath200 with the integrals being taken over the relevant frequency interval .
this gives a probability of being in a resonance which is close to 0.7% for @xmath165 and 8.5 @xmath156 and 2% for @xmath201 @xmath156 .
the fact that the probability of being in a resonance increases with @xmath178 is not significant , because resonances get weaker when the period increases ( see figure [ fig2 ] ) .
we note that this discussion applies only if the _ a priori _ probability of being in any frequency interval of a given width is independent of the frequency as might be expected to be a reasonable assumption if the normal mode spectrum is fixed .
however , different circumstances may apply if the combined effect of orbital and stellar evolution were to lock the companion in a resonance with changing location .
but we shall not consider the possibility of this process here .
as the companion spirals inwards , it goes through a succession of resonances .
however , for a fixed normal mode spectrum , the above calculation tells us that its migration is controlled essentially by the non - resonant interaction .
this can be seen by comparing @xmath202 evaluated over a large frequency range , both taking into account and neglecting the resonances .
such a comparison shows that neglecting resonances changes @xmath202 by at most a few per cent .
_ ii ) relation between the mean torque , @xmath28 and the circularization timescale _
: here we interpolate the numerical results to express the torque as a power of the frequency . to begin with ,
we consider the three frequency intervals described above .
we take the appropriate torque to be @xmath203 where the mean values are taken over the frequency intervals displayed in figure [ fig2 ] .
the results can be interpolated with the following relation : @xmath204 we have checked that the above formula gives a good estimate of the torque at other non resonant frequencies between 4.23 and 12.4 @xmath156 . since the index of the power law ( [ dynint ] ) is close to 4 , the circularization timescale @xmath150 is given by ( [ tcircb ] ) . at @xmath201
@xmath156 , @xmath150 is found from the above formula to be 56 times larger than the timescale of 4 @xmath155 that is indicated by the observations .
we note that both the dynamical and equilibrium tide calculations give a power law with an index close to 4 , which results in the circularization timescale being proportional to the binary period @xmath178 to the 13/3 . for comparison , zahn ( 1977 ) , zahn ( 1989 ) and
goldman & mazeh ( 1991 ) , using equilibrium tide calculations , found @xmath150 to be proportional to @xmath178 raised to the power 16/3 , 13/3 and 10/3 respectively .
the difference between these results can be related to a different choice of @xmath205 in the expression ( [ nu ] ) for @xmath206 these authors used an expression similar to ( [ nu ] ) with respectively @xmath207 , 1 and 2 .
the fact that we obtain an index close to 13/3 by setting @xmath66 or even @xmath67 ( see below ) , in contrast to the results above , is at least partially due to the effectively smaller value of @xmath59 we used ( see below ) .
we comment further that tassoul ( 1988 ) found @xmath208 for his postulated alternative hydrodynamical mechanism for tidal circularization .
the results presented above show that the torque corresponding to the dynamical tide is smaller than that given by the adiabatic equilibrium tide for all the frequencies we have computed .
however , the difference tends to decrease as the frequency gets smaller . from expressions ( [ asyint ] ) and ( [ dynint ] )
, we calculate that the ratio of these torques is indeed about 6.0 and 4.8 for @xmath165 and 12.4 @xmath156 respectively .
figure [ fig3 ] shows @xmath209 , @xmath210 and @xmath211 in units @xmath168 versus @xmath212 in the range @xmath213 for @xmath165 and 12.4 @xmath156 .
it is clear from these plots that @xmath42 departs from the asymptotic value in the convective envelope .
the difference is not large , being about 17% , but the derivatives of @xmath42 and @xmath39 from which the torque is calculated ( see [ pt1 ] ) depart more from their asymptotic values . in the limit where the magnitude of the brunt - visl frequency is everywhere large compared to the tidal forcing frequency
, calculations based on the dynamical tide should converge towards those based on the equilibrium tide .
this is because the convective timescale is small enough that the convective motions adjust essentially instantaneously to the tidal forcing .
we have checked this expectation by artificially increasing the magnitude of the brunt - visl frequency in the convection zone . except in the part of the convective envelope just below its outer radius , wherever @xmath214 , @xmath215 being an arbitrary constant , we make the replacement @xmath216 . in figure [ fig4 ]
we plot @xmath217 versus @xmath212 in the range @xmath218 for @xmath201 @xmath156 and for @xmath219 100 and 400 . for comparison we also plot the case corresponding to the original solar model .
as expected , @xmath217 converges towards 1 when the magnitude of the brunt - visl frequency is increased .
we note that @xmath220 , which corresponds to @xmath221 , gives @xmath222 smaller than 5% .
this confirms that the asymptotic limit is reached when the magnitude of the brunt - visl frequency is very large compared to the frequency of the tidal wave .
we note that the very slow convergence towards the equilibrium tide was predicted from the arguments presented in
[ sec : asy ] . also expected was the discontinuity in the mean value of @xmath223 obtained after averaging out the interior oscillations , at the boundary of the radiative core and the convection zone that is observed in figure [ fig3 ] ( see also figure [ fig1 ] ) .
we shall limit the comparison of our results with observations of main sequence binaries because our calculations do not apply to pre
main sequence stars , which have a much larger convective envelope than the sun .
as mentioned above , the observed circularization timescale we have to fit is then 4 @xmath155 for @xmath201 @xmath156 ( mathieu 1994 ) . as indicated above
, when using a simple estimate of the turbulent viscosity based on mixing length theory for non rotating stars , the circularization timescale we get from our calculations for this period is 56 times larger than 4 @xmath155 .
this indicates that either + ( i ) solar type binaries are not circularized through turbulent viscosity acting on tidal perturbations ( but see tassoul 1988 and kumar & goodman 1996 for other suggested tidal mechanisms ) , or + ( ii ) dissipation in the convective envelope of solar like stars is significantly more efficient than is currently estimated .
tassoul ( 1995 ) postulates that efficient tidal dissipation occurs in a very thin eckman layer close to the surface of a tidally deformed star and that this process greatly increases the efficiency of tidal interactions . but
a refutation of the notion that the free surface boundary condition appropriate to the tidally deformed star , rather than the more common rigid boundary condition , leads to such an effective boundary layer , has been given by rieutord & zahn ( 1997 ) .
further tassoul & tassoul ( 1997 ) state that their mechanism is inapplicable to extreme mass ratio cases such as 51 pegasi that we consider later in the paper .
we now consider briefly here the mechanism proposed by kumar & goodman ( 1996 ) , namely enhanced dissipation associated with high order oscillation modes excited through parametric instability .
the growth rate for the most rapidly growing modes is expected to be @xmath224 , where @xmath42 is the radial displacement in the primary oscillation , which we shall assume to be the equilibrium tide , evaluated at @xmath225 , and @xmath226 is its frequency .
if we assume that the non linear development of the parametric instability and subsequent dissipation of the excited modes leads to an effective viscosity , and frictional dissipation rate @xmath227 which is big enough to suppress the linear instability , then we expect from the classical theory of darwin ( 1879 ) that there will be a phase lag @xmath228 associated with the tide given by @xmath229 for a binary of unit mass ratio and period @xmath230 @xmath156 , synchronization occurs on a timescale very much shorter than that required for circularization , so that we assume that the stellar rotation is synchronized with the orbit and @xmath231 in the calculation of @xmath232 for small eccentricity , the circularization timescale @xmath150 is approximately given by @xmath233 , where the apsidal motion frequency is @xmath234 , with @xmath235 being the apsidal motion constant ( cowling 1938 ) .
we use the equilibrium tide value ( [ xieq ] ) to estimate @xmath236 at @xmath225 as @xmath237 for a mass ratio of unity .
we then get @xmath238 for @xmath228 given by ( [ darwin ] ) . for and
orbital period of 12.4 @xmath156 , this gives @xmath239 @xmath240 since @xmath241 , it is not very likely that this mechanism will be able to explain the observed circularization rates .
however , we stress that this has not been shown from a full non linear calculation of the development of parametric instability . in general , a large increase in the simply estimated turbulent viscosity coefficient
is needed in order to explain the observed circularization rate .
we now investigate what is needed to achieve this and give the numerical results of tests we have carried out . in all cases , we have checked that _ the velocity at the surface of the star is not sensitive to the magnitude of the turbulent viscosity assumed_. in addition , as indicated above , the resonances are essentially controlled by radiative damping in the radiative core as long as @xmath242 .
when the turbulent viscosity is enhanced , @xmath83 is increased
. however , in the tests we present below , for orbital periods larger than @xmath0 days , @xmath128 stays large enough compared to @xmath83 so that , although the non resonant torques increase , the central structure and strength of the resonances is determined by radiative damping . for orbital periods on the order of @xmath1 days
, @xmath83 can become comparable to @xmath128 .
then the damping factor in [ restor ] has to be replaced by @xmath243 , and the strength of the resonance is reduced by a factor @xmath1 .
we first consider the effect on the circularization timescale of _ varying the parameters _
@xmath58 , @xmath59 , @xmath205 and @xmath61 in the expression ( [ nu ] ) for @xmath56 .
note that the denominator we used in this expression is @xmath244 rather than @xmath245 using the latter with @xmath246 is equivalent to setting @xmath247 in the former . at present
, it seems that our knowledge of convection does not allow discrimination between these possibilities ( see , for example , the discussion in zahn 1989 ) .
however , we note that goodman & oh ( 1997 ) have recently put forward some arguments in favor of @xmath247 . using @xmath248 in ( [ nu ] ) results in a circularization time for @xmath201 @xmath156 that is 10 times larger than that obtained with @xmath246 . if we set @xmath249 , @xmath150 is decreased by only a factor 2 .
this is because when @xmath246 the factor @xmath250 is already smaller than , or even very small compared to , unity in a large part of the convective envelope .
thus it seems that adjusting the way in which turbulent viscosity responds to short period forcing can not produce the required enhancement in this case . in table
[ table2 ] we summarize the results obtained for different values of @xmath205 and @xmath59 , and in table [ table3 ] we indicate the corresponding index of the power law in expression ( [ dynint ] ) .
we note that @xmath246 corresponds to the denominator in ( [ nu ] ) being @xmath251 , whereas @xmath252 corresponds to @xmath253 .
ccccccccccccccccccccccccccccccccc @xmath178 & @xmath59 & @xmath205 & @xmath150 + ( days ) & & & ( @xmath155 ) + 4.23 & 1 & 2 & 2.46 + & & 1 & 1.37 + 12.4 & & 2 & 220 + & & 1 & 211 + & & & + 4.23 & @xmath254 & 2 & 39.4 + & @xmath255 & 1 & 5.98 + 12.4 & @xmath254 & 2 & 2010 + & @xmath255 & 1 & 685 + [ table2 ] ccccccccccccccccccccccccccccccccc @xmath59 & @xmath205 & index of the power law + 1 & 2 & 3.85 + & 1 & 4.3 + & & + @xmath254 & 2 & 3.3 + @xmath255 & 1 & 4.1 + [ table3 ] we note that setting @xmath256 with @xmath66 decreases the index of the power law in ( [ dynint ] ) down to @xmath257 this gives @xmath150 proportional to the orbital period to the @xmath258 , which is similar to the value found by goldman & mazeh ( 1991 ) .
zahn & bouchet ( 1989 ) have argued that the prescription @xmath66 suggested by goldreich & keeley ( 1977 ) and used later by campbell & papaloizou ( 1983 ) and goldman & mazeh ( 1991 ) ( see also goodman & oh 1997 ) would lead to cutoff periods in clear conflict with the observational data which they claim require @xmath259 the results presented for the model adopted here do not support this statement .
if @xmath246 , @xmath150 hardly changes when @xmath205 is changed from 2 to 1 .
if @xmath252 , the difference between @xmath67 and @xmath66 is not dramatic for @xmath260 @xmath156 . within the uncertainties associated with convection , our results do not allow a distinction to be made between @xmath67 and @xmath66 . finally , taking the mixing length to be the distance to the top boundary of the convective envelope rather than 3 times the pressure scale height does not affect significantly the circularization timescale .
we note that the torque is directly proportional to @xmath58 and @xmath261 .
however , @xmath58 is expected to be on the order of unity , and @xmath164 is usually taken to be between 1 and 4 . the magnitude of the turbulent viscosity given by ( [ nu ] ) would be increased if the magnitude of the brunt - visl frequency was larger in the convection zone .
this is because the convective timescale decreases .
we comment that a reduction in the convective timescale , while maintaining the same length scale , implies larger convective velocities which would have to occur without increasing the heat flux .
this is the essential feature of the modification . to illustrate the effect of increasing the brunt - visl frequency we consider the following dimensionless number : @xmath262 which is the superadiabatic temperature gradient when
radiation pressure and variations of the mean molecular weight are neglected .
the accuracy with which @xmath263 is known from helioseismic observations is not better than @xmath264 ( gough 1984 ) .
however , in most of the convective envelope , this parameter , estimated from mixing length theory applied to a non rotating stars , is much smaller than @xmath265 we have made a numerical investigation in which we increased @xmath266 in the convection zone by replacing @xmath263 by min@xmath267 @xmath268 being an arbitrary constant , wherever @xmath269 ( except just below the outer radius of the convective envelope ) .
we have considered @xmath270 and @xmath271 .
it is doubtful that consideration of present helioseismic data could preclude such an increase of @xmath266 ( thompson 1997 ) .
figure [ fig5 ] shows @xmath263 in the convective envelope versus @xmath212 for the original solar model and for @xmath270 and @xmath271 .
we also display the factor by which @xmath266 has been increased in each case .
the circularization timescales we find with @xmath246 and @xmath66 for @xmath201 @xmath156 when @xmath270 and @xmath271 are respectively 13 and 7.7 @xmath155 , which are now larger than the observed value by factors of 3 and 2 respectively .
small discrepancies of this magnitude could be dealt with by adjustments to the mixing length or @xmath272 when either @xmath270 or @xmath271 , the circular orbit torque is found to be proportional to @xmath273 .
if , keeping @xmath66 , we adopt @xmath256 , @xmath150 is increased only by a factor 1.3 compared to the case @xmath246 for @xmath271 .
this is because the factor @xmath250 is smaller than , or even very small compared to , unity in almost all the convection zone whatever @xmath59 between 1 and @xmath254 .
although an increase in @xmath274 of the magnitude we consider might be thought to be unrealistic , we note that such an increase in the deep layers of the convection zone has also been considered by dsilva ( 1995 ) as a means of explaining the dynamics of sunspots without invoking too strong a magnetic field . in his model , which applies strictly to a star rotating at the same rate as the sun , @xmath274 has to be larger than @xmath275 @xmath276 , which means that it has to be multiplied on average by a factor @xmath277 in the solar model we use , we need to multiply @xmath274 by a factor between 100 and 400 for @xmath278 , between 10 and 100 for @xmath279 and between 1 and 10 for @xmath280 in order to get such a minimum value .
if we do this , the circularization timescale we get for @xmath201 @xmath156 is only 6 times larger than the observed one . in this context , it is possible that if the proposed increase in @xmath274 is related to the stellar rotation , this may be even greater for the more rapidly synchronously rotating star that is expected in the equal mass binary case .
we note that numerical simulations of turbulent convection in the presence of rotation show an increase of @xmath274 with the effect of rotation ( brummell , hurlburt & toomre 1996 ) .
this is because , as pointed out by brummel _
et al . _ ( 1996 ) , rotation influences the thermodynamic mixing properties of the convection in such a way that it leads to a decrease in correlation between temperature fluctuations and vertical velocities .
the efficiency of the vertical convective transport is then weakened , with a subsequent enhanced superadiabatic mean stratification in the interior of the fluid ( see their figure 8.a ) . +
this suggests that the magnitude of the brunt - visl frequency in the convective envelope of rotating stars is actually larger than the values given by the solar model we have been using here .
however , it seems questionable that the extremely large increase required to account for the observed circularization rates can be achieved .
another means of increasing the total amount of dissipation is to assume that turbulent viscosity acts down to some depth below the inner boundary of the convective envelope .
this might be expected if convective overshooting takes place .
however , this might not be a very effective process because of the slow convective motions expected and the rapid increase in @xmath274 that occurs as the radiative zone is entered .
estimates based on the observed solar oscillation frequencies give an upper limit between 0.05 ( basu 1997 ) and @xmath281 ( christensen dalsgaard , monteiro & thompson 1995 ) times the pressure scale height on the extent of overshoot below the convection zone . here
we consider a simple illustrative situation in which convective blobs or some other turbulent motions are able to penetrate into the stratified radiative core over some fraction @xmath282 of the pressure scale height @xmath283 producing a turbulent viscosity .
we model this by setting @xmath56 to be constant from a distance @xmath284 above the inner radius of the convective envelope down to the same distance below this radius , equal to its value at the top of this zone . because such a viscosity is able to act on the short wavelength part of the tidal response associated with g modes it has a dramatic effect . for @xmath285 and @xmath157 , the circularization timescale for @xmath201
@xmath156 is decreased by a factor about 400 , being now 8 times smaller than the observed timescale .
if we set @xmath256 ( see discussion above ) , @xmath150 is increased by a factor 15 , being about 4 times larger than the observed timescale . for @xmath286 and @xmath157
, we get a circularization timescale for @xmath201 @xmath156 about 30 times larger than the observed one . for the calculated timescale to be in agreement with the observations
, we need @xmath282 between 0.4 and 0.5 with @xmath157 . in the model
we have adopted , the effect is of less importance for shorter periods because the number of oscillations of the response in the region of the radiative core where turbulent dissipation is introduced decreases with forcing frequency .
therefore , the torque does not vary with @xmath28 as a simple power law with an index close to 4 .
however , in reality , the effectiveness of the turbulent viscosity should be reduced for short wavelength disturbances giving a compensating effect to make it relatively less effective at low frequencies .
the above calculations show that overshooting is not likely to be efficient enough to decrease the circularization timescales by a factor of about 50 . to get such an effect , we indeed require an extent of overshoot below the convection zone at least 5 or 10 times larger than that deduced from the observations .
in addition , we have not taken into account the fact that overshooting leads to an increase of the g modes length scale through a decrease in the buoyancy or the magnitude of @xmath287 this in turn would decrease the amount of turbulent dissipation associated with overshooting . as we have already mentioned above , _
calculations based on both the dynamical and equilibrium tide theories give a torque proportional to the orbital frequency raised to a power @xmath1 _ ( see [ asyint ] and [ dynint ] ) .
if circularization of solar type binaries does occur through the action of turbulent viscosity on the tides , then its magnitude has to be calibrated so as to account for the observed timescale .
we have discussed in the previous section , somewhat speculatively , how the required enhancement of the magnitude of the viscosity above that obtained from simple estimates might be envisaged to occur .
since the enhancement might depend on forcing frequency , it is not clear that the resulting torque will still be proportional to the frequency to the @xmath1 .
however , the increase to @xmath288 described above gave torques that approximately preserved this power law so that in the absence of additional information , we shall suppose it holds .
then the calibration acts only to adjust the coefficient of the power law .
we note that the observations do not rule out any exponent between 3 and 5 ( see below ) .
since our calculations can only strictly be applied to solar type stars , we calibrate our results using @xmath289 @xmath155 for @xmath201 @xmath156 .
this gives ( in cgs ) :
@xmath290 or , equivalently : @xmath291 the corresponding formul for the orbital and spin up timescales ( in giga
years ) are : @xmath292 @xmath293 and @xmath294 ( for the solar model we use , @xmath295 @xmath296 and @xmath297 @xmath296 ) .
since the torque is proportional to @xmath298 we can use the relation ( [ tcircb ] ) for @xmath150 , so that the circularization time is given by @xmath299 even though our calculations can only be applied to solar type stars , it is of interest to compare the circularization timescales we get from ( [ tcirc ] ) with the observed ones . for @xmath300 , 7.05 , 8.5 and 18.7 @xmath156 , ( [ tcirc ] ) gives respectively @xmath301 , 0.3 , 0.8 and 24 @xmath155 , to be compared with the observed timescales 0.003 , 0.1 , 0.8 and 16 @xmath155 respectively .
the agreement for @xmath302 @xmath156 is within a factor 1.5 . for smaller periods
, circularization is expected to occur when the convective envelopes of the stars are larger , making turbulent dissipation more efficient .
we note that mathieu _
et al . _ ( 1992 ) have already pointed out that a power law @xmath303 provides a close fit to the slope of the observed cutoff periods .
however , the observations are equally well fitted with an index of 10/3 and an index of 16/3 can not be ruled out ( mathieu _ et al . _
1992 ) .
if the simple estimate of the turbulent viscosity based on mixing length theory for non rotating stars is used , the coefficients in the formul for the torque have to be divided by a factor @xmath304 whereas those in the formul for the timescales have to be multiplied by the same factor . in figure [ fig6 ]
, we have plotted @xmath144 in units @xmath305 and @xmath150 in units @xmath306 versus @xmath28 and @xmath178 in a log log representation .
it is of interest to apply these results to the system 51 pegasi , for which the orbital period ( assuming the observed oscillations are due to a companion ) is @xmath165 @xmath307 if the companion is a jupiter mass planet ( @xmath308 ) , then the tidal orbital evolution timescale given by ( [ torb ] ) is @xmath309 @xmath310 the star spin up timescale ( [ tsp ] ) is @xmath311 @xmath155 and the spin up timescale of the convective envelope is @xmath312 @xmath155 .
all of these timescales are long compared with the inferred age of 51 pegasi ( edvardsson _ et al . _
if the companion is a low - mass star of @xmath313 , as has been recently suggested , @xmath144 is 100 times smaller while @xmath314 and @xmath315 are @xmath316 times smaller .
we then expect the primary star to be synchronized with the orbit , in which case exchange of angular momentum is no longer taking place . synchronization is actually expected if the mass of the companion is larger than about 10 jupiter masses .
the orbital decay timescale is also smaller than the age of the system , but since @xmath317 , tidal interaction stops before the companion has plunged into the central star .
if the simple estimate of the turbulent viscosity based on mixing length theory for non rotating stars is used , all these timescales have to multiplied by @xmath318 . in that case
, synchronization is expected if the mass of the companion is larger than about 70 jupiter masses .
the planetary companion interpretation has been questioned recently by the reported 4.23 @xmath156 modulation in the line profile of 51 pegasi ( gray 1997 ) , and the possibility that this modulation may be due to g mode oscillations has been considered ( gray & hatzes 1997 ) .
we note that , according to our results , such a modulation could not be due to g mode oscillations tidally driven by a companion . for the oscillation to have a period of @xmath319 @xmath156
, the orbital period would have to be 8.46 @xmath156 .
the maximum perturbed radial velocity at the surface of the star induced by the companion would then be between @xmath172 and 1 @xmath170 for a perturbing mass between @xmath320 and 1 solar mass .
_ these numbers do not depend on the magnitude of the turbulent viscosity assumed , and are not expected to be affected by the possibility of resonance_. these velocities are at least about 50 times smaller than the observed ones . in this paper , we have studied the dynamical response of the star to the tidal perturbation of a companion . we have computed the torque due to dissipation in the convective envelope using first
order perturbation theory . in the vicinity of resonance
, we have also calculated the torque due to non adiabaticity in the radiative core using a wkb treatment .
we have found that the torque at effective resonances is mainly determined by radiative damping .
we have carried out an analysis based on the adiabatic equilibrium tide , and showed that agreement with the dynamical tide calculations can be rather poor . for the unmodified stellar model and the periods of interest of several days , the torque derived using the equilibrium tide is 4 to 6 times larger than that corresponding to the dynamical tide .
we have found that the presence of fixed resonances do not affect the long term orbital evolution of the binary , so that the different timescales ( orbital evolution , circularization and spin up ) are mainly determined by the non resonant interaction .
our calculations show that the viscosity that is required to provide the observed circularization rates of solar type binaries is @xmath318 times larger than that simply estimated from mixing length theory for non rotating stars .
we have explored some means by which this viscosity might be enhanced .
we have found that it could become large enough if the magnitude of the brunt - visl frequency in the deep convective envelope were increased sufficiently .
such an increase is expected to be produced by the effect of rotation on convection , but it is questionable whether it can be of sufficient magnitude .
we note that _ the strength of the resonances for orbital periods larger than @xmath0 days and the perturbed velocity at the surface of the star are insensitive to the magnitude of the turbulent viscosity assumed_. only for periods @xmath1 days is the strength of the resonances decreased by a factor @xmath1 .
the effective widths of the resonances affecting the tidal torques as well are reduced when the viscosity is increased .
we have applied our results to 51 pegasi , and showed that the oscillations which have been observed at the surface of this star can not be a tidally driven non
radial g mode .
also we have found that the stellar rotation and the orbital motion of this system are expected to be synchronized if the mass of the companion exceeds 0.1 solar mass .
we are grateful to m.j .
thompson for supplying us a solar model , and thank d.o .
gough , r.d .
mathieu and m.j .
thompson for helpful discussions .
this work is supported by pparc through grant gr / h/09454 and by nsf and nasa through grants ast 9315578 and nag 54277 .
acknowledge support by the center for star formation studies at nasa / ames research center and the university of california at berkeley and santa - cruz .
mathieu , r. d. , duquennoy , a. , latham , d. w. , mayor , m. , mazeh , t. , & mermilliod , j.c .
1992 , in binaries as tracers of stellar formation , ed .
a. duquennoy , & m. mayor ( cambridge : cambridge university press ) , 278 | we calculate the dynamical tides raised on a non rotating solar type star by a close stellar or planetary companion .
dissipation arising from a turbulent viscosity operating in the convection zone and radiative damping in the radiative core are considered .
we compute the torque exerted on the star by a companion in circular orbit , and determine the potentially observable magnitude of the tidally induced velocity at the stellar photosphere .
these calculations are compared with the results obtained by assuming that a very small frequency limit can be taken in order to calculate the tidal response ( _ equilibrium tide _ ) . for a standard solar model ,
the latter is found to give a relatively poor approximation at the periods of interest of several days , even when the system is far from resonance with a normal mode .
this behavior is due to the small value of the brunt - visl frequency in the interior regions of the convection zone .
it is shown that although the companion may go through a succession of resonances as it spirals in under the action of the tides , for a fixed spectrum of normal modes its migration is controlled essentially by the non resonant interaction .
we find that the turbulent viscosity that is required to provide the observed circularization rates of main sequence solar type binaries is about fifty times larger than that simply estimated from mixing length theory for non rotating stars .
we discuss the means by which this enhanced viscosity might be realized .
these calculations are applied to 51 pegasi .
we show that the perturbed velocity induced by the tides at the stellar surface is too small to be observed .
this result is insensitive to the magnitude of the turbulent viscosity assumed and is not affected by the possibility of resonance . for this system ,
the stellar rotation and the orbital motion are expected to be synchronized if the mass of the companion is as much as one tenth of a solar mass . |
observations indicate that roughly 70% of the energy density in the universe is in the form of an exotic , negative - pressure component , dubbed dark energy @xcite .
( see ref .
@xcite for a recent review . )
if @xmath3 and @xmath4 are the density and pressure , respectively , of the dark energy , then the dark energy can be characterized by the equation - of - state parameter @xmath5 , defined by @xmath6 it was first noted by caldwell @xcite that observational data do not rule out the possibility that @xmath7 . such
phantom " dark energy models have several peculiar properties .
the density of the dark energy _ increases _ with increasing scale factor , and both the scale factor and the phantom energy density can become infinite at a finite @xmath8 , a condition known as the big rip " @xcite .
it has even been suggested that the finite lifetime for the universe in these models may provide an explanation for the apparent coincidence between the current values of the matter density and the dark energy density @xcite . while @xmath9 as @xmath10 extends into the future is a necessary condition for a future singularity , it is not sufficient .
in particular , if @xmath0 approaches @xmath11 sufficiently rapidly , then it is possible to have a model in which @xmath3 increases with time , but in which there is no future singularity .
conditions which produce such an evolution ( specified in terms of @xmath4 as a function of @xmath3 ) were explored in refs .
@xcite . in this paper
, we examine such models in more detail . in particular
, we will extend the parameter space discussed in refs .
@xcite in both directions , showing that there are nonsingular models in which @xmath3 increases more rapidly than the nonsingular models discussed in those references , and , conversely , that there are singular models with @xmath3 increasing less rapidly than the singular models discussed in refs .
models without a future singularity in which @xmath3 increases with time will nonetheless eventually lead to a dissolution of bound structures at some point in the future , a process we have dubbed the little rip .
" we discuss the time scales over which this process occurs . finally , we consider observational constraints on these models . in the next section ,
we examine the conditions necessary for a future singularity in models with @xmath1 . in secs .
iii and iv , specific little - rip models and disintegration of bound systems are studied . finally , in sec .
v , there is discussion .
we limit our discussion to a spatially flat universe , for which the friedmann equation is [ fried1 ] ( ) ^2 = , where @xmath12 is the total density , @xmath10 is the scale factor , the dot will always denote a time derivative , and we take @xmath13 throughout .
we will examine the future evolution of our universe from the point at which the pressure and density are dominated by the dark energy , so we can assume @xmath14 and @xmath15 , and for simplicity we will drop the @xmath16 subscript .
then the dark energy density evolves as [ fried2 ] = -3()(+ p ) .
the simplest way to achieve @xmath1 is to take a scalar field lagrangian with a negative kinetic term , and the conditions necessary for a future singularity in such models have been explored in some detail @xcite . here
, however , we explore the more general question of the conditions under which a dark energy density that increases with time can avoid a future singularity , and the consequences of such models .
one can explore this question from a variety of starting points , by specifying , for example , the scale factor @xmath10 as a function of the time @xmath8 ( an approach taken , for example , in refs .
alternately one can specify the pressure @xmath17 as a function of the density @xmath12 , as in refs .
note that this is equivalent to specifying the equation - of - state parameter @xmath0 as a function of @xmath12 , since @xmath18 .
finally , one can specify the density @xmath12 as a function of the scale factor @xmath10 .
since we are interested specifically in nonsingular models for which @xmath12 increases with @xmath10 , we shall adopt this last approach , but we will briefly examine the other two starting points . of course , given any one of these three functions , the other two can be derived uniquely , but not always in a useful form .
for example , suppose that we specify @xmath19 . in order to avoid a big rip , it is sufficient that @xmath19 simply be a nonsingular function for all @xmath8 .
writing @xmath20 where @xmath21 is a nonsingular function , the density is given by equation ( [ fried1 ] ) as @xmath22 , and the condition that @xmath12 be an increasing function of @xmath10 is simply @xmath23 , which is satisfied as long as @xmath24 thus , all little - rip models are described by an equation of the form ( [ a(t ) ] ) , with nonsingular @xmath25 satisfying equation ( [ ddf ] ) .
now consider the approach of refs .
@xcite , who expressed the pressure as a function of the density in the form @xmath26 where @xmath27 ensures that the @xmath12 increases with scale factor . in order to determine the existence of a future singularity
, one can integrate equation ( [ fried2 ] ) to obtain @xcite @xmath28 and equation ( [ fried1 ] ) then gives @xcite @xmath29
the condition for a big - rip singularity is that the integral in equation ( [ trho ] ) converges .
taking a power law for @xmath30 , namely @xmath31 we see that a future singularity can be avoided for @xmath32 @xcite .
we examine this boundary in more detail below , noting that one can have @xmath30 increase more rapidly than @xmath33 without a future singularity .
now consider the third possibility : specifying the density @xmath12 as an increasing function of scale factor @xmath10 .
we will seek upper and lower bounds on the growth rate of @xmath34 that can be used to determine whether or not a big - rip singularity is produced . defining @xmath35 , we can rewrite equation ( [ fried1 ] ) as @xmath36 and the condition for avoiding a future big - rip singularity is @xmath37 the case @xmath38 from refs .
@xcite corresponds to @xmath39 where @xmath40 requires @xmath41 , and we take @xmath42 and @xmath43 at a fixed time @xmath44 . expressing this density as a function of time rather than scale factor gives a much simpler expression : @xmath45 the equation - of - state parameter @xmath0 corresponding to equation ( [ model1 ] ) can be derived from the relation @xmath46 : @xmath47 and the corresponding expansion law is @xmath48}. \label{aaa}\ ] ] however , we can find @xmath34 for which @xmath12 increases more rapidly with @xmath10 , but for which equation ( [ condition ] ) is still satisfied .
for example , writing @xmath49 as @xmath50 satisfies equation ( [ condition ] ) .
an example of such a @xmath12 , with a free parameter b , is @xmath51 where the choice @xmath52 leads to a real , nonnegative @xmath12 and an analytic form for the behavior of @xmath19 : @xmath53 } } } -1 ) } -b .
\label{bbb}\ ] ] this argument can be extended further . in general ,
if we denote @xmath54 , where the logarithm on the right - hand side is iterated @xmath55 times , then any function of the form @xmath56 satisfies equation ( [ condition ] ) as @xmath57 and avoids a big - rip singularity .
a density increasing as in equation ( [ nestedlog ] ) leads to an expansion law of the form @xmath58 where there are @xmath59 exponentials .
we have omitted the constants in equations ( [ nestedlog ] ) and ( [ nestedexp ] ) for the sake of clarity .
equation ( [ nestedexp ] ) , while growing extraordinarily rapidly , is manifestly nonsingular . while an expansion law of this sort might seem absurd ,
it is probably less so than a big - rip expansion law , and in any case our goal is to try to determine the boundary between little - rip and big - rip evolution for @xmath34 . in this spirit , consider the slowest growing power - law modification to equation ( [ nestedlog ] ) : @xmath60 where @xmath61 is a constant .
no matter how small @xmath62 is , and despite the fact that it modifies an extraordinarily slowly growing nested logarithm function , the growth law in equation ( [ nestedrip ] ) leads to a future big - rip singularity .
note that the bounds specified by equations ( [ nestedlog ] ) and ( [ nestedrip ] ) are not sharp ; we can always find forms for @xmath34 that interpolate between these two behaviors and produce either a little rip or a big rip .
however , as we take @xmath63 to be arbitrarily large , nearly any function of interest will increase more rapidly than equation ( [ nestedlog ] ) or more slowly than equation ( [ nestedrip ] ) , allowing us a practical , if not a rigorously sharp , bound .
this lack of a sharp bound is due to the fact that there is no bound on the fastest growing function @xmath19 which is nonsingular at finite @xmath8 .
if one is willing to place other restrictions on the form of @xmath34 , then more stringent bounds apply .
barrow @xcite demonstrated that if @xmath64 is a rational function of @xmath10 and @xmath8 , and @xmath19 is nonsingular at finite @xmath8 , then @xmath19 can grow no more rapidly than the double exponential of a polynomial in @xmath8 .
our equation ( [ nestedlog ] ) violates this condition because of the logarithmic functions .
here we shall examine in more detail the two specific little - rip models given by equations ( [ model1 ] ) and ( [ model2 ] ) , which we will call model 1 and model 2 , respectively . note that we do _ not _ make use of equations ( [ aaa ] ) and ( [ bbb ] ) here , as these are valid only when the matter density can be neglected in comparison to the dark energy density .
model 1 is characterized by a single free parameter @xmath65 , and the scale factor behaves asymptotically as a double exponential in @xmath8 , as in equation ( [ aaa ] ) : @xmath66 the parameter @xmath65 is chosen to make a best fit to the latest supernova data from the supernova cosmology project @xcite , and has the best - fit value @xmath67 , while a @xmath68 c.l .
fit can be found for the range @xmath69 .
model 2 is characterized by the free parameter @xmath70 and has a scale factor that behaves asymptotically as a triple exponential in @xmath8 , as in equation ( [ bbb ] ) : @xmath71 the parameter @xmath70 is chosen to make a best fit to @xcite as well , and it has the value @xmath72 . the confidence interval for @xmath70 at the @xmath68 c.l .
is @xmath73 . in fitting both models ,
@xmath74 , @xmath75 , and @xmath76 km s@xmath77 mpc@xmath77 , which are consistent with the best - fit ranges for these values given by wmap @xcite .
the resultant hubble and residual @xmath78cdm ( @xmath79 ) plots of distance modulus @xmath80 versus redshift @xmath81 for both models are displayed in fig
+ not surprisingly , the best - fit models closely resemble the @xmath78cdm model , which is known to be an excellent fit to the data @xcite . to see this more clearly , note that our models will resemble a cosmological constant at low redshift as long as @xmath82 for @xmath83 . for model 1
, this condition is satisfied when @xmath84 in equation ( [ model1 ] ) , while for model 2 , we require @xmath85 in equation ( [ model2 ] ) . to see that @xmath70 should be close to this value
, one should expand equation ( [ model2 ] ) around @xmath43 .
the zeroth - order term is @xmath86 , and the coefficient for the first - order term is @xmath87 when @xmath88 .
a comparison with our best - fit values indicates that these conditions are , indeed , satisfied .
furthermore , in the limit where these conditions are satisfied , these little - rip models closely resemble , at low redshift , big - rip models close to @xmath78cdm , i.e. , models with constant @xmath1 and @xmath89 . to see this , recall that constant-@xmath0 big - rip models have a density varying with scale factor as @xmath90 for @xmath91 and @xmath92 not too far from 1 , equation ( [ rhorip ] )
behaves as @xmath93.\ ] ] equation ( [ model1 ] ) reduces to equation ( [ rhorhip2 ] ) for @xmath84 , with @xmath94 .
a feature of a big rip is that all bound - state systems disintegrate before the final singularity @xcite . here
we show that little - rip models , despite not having a final singularity , also produce the disintegration of bound structures . as a first approximation ,
the disintegration time is when the dark energy density equals the mean density of the system .
a more accurate method was presented in @xcite .
we shall employ both methods to estimate the disintegration of the sun - earth system .
gyrs , it will envelope mercury and venus , and ( maybe ) earth @xcite . here , for the sake of making a point , we assume the earth will continue to orbit the sun until unbound by dark energy . ] for the little - rip models 1 and 2 , with the best - fit parameters derived in the previous section , we find the time @xmath95 from the present time @xmath44 until the earth ( @xmath96 ) - sun ( @xmath97 ) system is disintegrated to be : @xmath98 @xmath99 note that the disintegration time for model 2 is less than that of model 1 , which is expected since @xmath12 for model 2 grows faster than @xmath12 for model 1 .
it is straightforward to estimate the corresponding @xmath95 for big - rip models with constant @xmath0 to be @xcite @xmath100 and it is almost identical to @xmath101 , which is about one year later .
clearly , little - rip models can produce this disintegration either earlier or later than big - rip models , depending on the exact parameters of each model .
for example , by putting , @xmath102 in equation ( [ br ] ) , we find a value of @xmath103 tyrs for @xmath95 , which is larger than that of models 1 and 2 in eqs.([lra ] , [ lrb ] ) . in this case
, disintegration occurs earlier in the little - rip model than in the big - rip model . the five energy conditions ( weak , null , dominant , null dominant , strong ) ( see , e.g. , ref .
@xcite ) are all violated by all little - rip and big - rip models . a simple way to see
this is that if @xmath1 , which occurs for any rip , a boost is allowed with @xmath104 to an inertial frame with negative energy density .
having said that , if general relativity itself fails for length scales bigger than that of galaxies , we may not be constrained by the same energy conditions .
in the big rip , the scale factor and density diverge in a singularity at a finite future time . in the @xmath78cdm model
, there is no such divergence and no disintegration because the dark energy density remains constant .
the little rip interpolates between these two cases ; mathematically it can be represented as an infinite limit sequence which has the big rip and the @xmath78cdm model as its boundaries .
such models can be represented generically by a density varying with scale factor as in equation ( [ nestedlog ] ) .
physically , in the little rip , the scale factor and the density are never infinite at a finite time .
nevertheless , such models generically lead to structure disintegration at a finite time . for models consistent with current supernova observations ,
such disintegration can occur either earlier or later in a little - rip model than in a big - rip model , depending on the parameters chosen for the models . however , for a given present - day value of @xmath0 , the big - rip model with constant @xmath0 will necessarily lead to an earlier disintegration than the little - rip model with the same present - day value of @xmath0 .
this results from the fact that @xmath0 increases monotonically in the little rip models , resulting in a smaller value for @xmath12 at any given @xmath10 than in the corresponding constant-@xmath0 big - rip model , and therefore , a lower expansion rate .
thus , supernova bounds on the epoch of disintegration for constant-@xmath0 big - rip models also apply to little - rip models ; one can not simultaneously satisfy supernova constraints and hasten the onset of disintegration to an arbitrarily early time simply by iterating exponentials in the expansion law . furthermore , supernova data force
both big - rip and little - rip models into a region of parameter space in which both models resemble @xmath78cdm . in this limit ,
big - rip and little - rip models produce essentially the same expansion law up to the present , despite having very different future evolution .
thus , current data already make it essentially impossible to determine whether or not the universe will end in a future singularity .
finally , we remark that since the novel and speculative cyclic cosmology proposed in ref .
@xcite requires only disintegration and not a singularity , such cyclicity would seem to be possible within a little - rip model instead of the big rip considered in @xcite .
this is one potentially fruitful direction for future research @xcite .
we thank ryan m. rohm for useful discussions .
p.h.f . and k.j.l .
were supported in part by the department of energy ( de - fg02 - 05er41418 ) .
r.j.s . was supported in part by the department of energy ( de - fg05 - 85er40226 ) .
knop , et al .
* 598 * , 102 ( 2003 ) .
riess , et al .
, ap.j . * 607 * , 665 ( 2004 ) .
copeland , m. sami , and s. tsujikawa , int .
d * 15 * , 1753 ( 2006 ) .
caldwell , phys .
b * 545 * , 23 ( 2002 ) .
caldwell , m. kamionkowski , and n.n .
weinberg , , 071301 ( 2003 ) .
frampton and t. takahashi , phys .
b * 557 * , 135 ( 2003 ) .
s. nesseris and l. perivolaropoulos , , 123529 ( 2004 ) .
scherrer , , 063519 ( 2005 ) .
s. nojiri , s.d .
odintsov , and s. tsujikawa , , 063004 ( 2005 ) ; s. nojiri and s.d .
odintsov , , 023003 ( 2005 ) . h. stefancic , , 084024 ( 2005 ) .
hao and x .- z .
li , , 043529 ( 2004 ) .
m. sami and a. toporensky , mod .
a * 19 * , 1509 ( 2004 ) .
v. faraoni , class .
* 22 * , 3235 ( 2005 ) .
j. kujat , r.j .
scherrer , and a.a .
sen , , 083501 ( 2006 ) . j.d .
barrow , class .
. grav . * 21 * , l79 ( 2004 ) .
barrow , class .
* 21 * , 5619 ( 2004 ) . c. cattoen and m. visser , class .
quant . grav .
* 22 * , 4913 ( 2005 ) .
barrow and s.z.w .
lip , , 043518 ( 2009 ) .
barrow , class .
quant . grav .
* 13 * , 2965 ( 1996 ) .
r. amanullah , _ et al . _
( the supernova cosmology project ) ap.j , * 716 , * 712 - 738 ( 2010 ) .
e. komatsu _ et al .
* 192 * , 18 ( 2011 ) .
p. serra , a. cooray , d.e .
holz , a. melchiorri , s. pandolfi , and d. sarkar , phys . rev . *
d80 , * 121302 ( 2009 ) .
k.p . schrder and r.c .
smith , mon . not .
* 386 , * 155 ( 2008 ) .
carroll , m. hoffmann and m. trodden , phys . rev .
* d68 , * 023509 ( 2003 ) .
l. baum and p.h .
frampton , phys rev .
lett . * 98 , * 071301 ( 2007 ) .
frampton , k.j .
ludwick and r.j .
scherrer , in preparation . | we examine models in which the dark energy density increases with time ( so that the equation - of - state parameter @xmath0 satisfies @xmath1 ) , but @xmath2 asymptotically , such that there is no future singularity .
we refine previous calculations to determine the conditions necessary to produce this evolution .
such models can display arbitrarily rapid expansion in the near future , leading to the destruction of all bound structures ( a little rip " ) .
we determine observational constraints on these models and calculate the point at which the disintegration of bound structures occurs .
for the same present - day value of @xmath0 , a big rip with constant @xmath0 disintegrates bound structures earlier than a little rip . |
one of the most intriguing observations made over the years in financial data analysis concerns the tendency of financial forecasters to imitate each other .
an account of this phenomenon is found in @xcite , where earning forecasts from different markets over a 17 year period are analyzed , showing ( among other results ) that : ( a ) forecasts are typically optimistic , that is the difference between the forecast and the actual earning is positive on average ; and ( b ) the spread of forecasts among analysts is significantly smaller on average than the forecast error , i.e. the typical spread of forecasts around the real earning . in other words ,
analysts forecasts are more similar to each other than they are to the variable they are trying to forecast .
more recently , a similar data set has been studied to estimate the fraction of herding analysts @xcite , with the surprising conclusion that around 75% of the analysts in the data set displayed a marked tendency to herd as a forecasting strategy .
interestingly , about 10% of analysts were instead found to be anti - herding .
stock prices in turn tend to react more strongly to forecasts that differ from the consensus ( see e.g. @xcite for a recent study ) . while psychological factors like social pressure , reputation issues and ( for anti - herding ) a desire for visibility can be crucial in determining this scenario , it is quite difficult to explain these results within the efficient market hypothesis ( emh ) . in the emh world , one would expect different forecasters to use their respective partial information to produce a proxy for the target variable that is unbiased and such that the forecast dispersion and the forecast error are roughly the same , as would result from forecasters that are independent and fully heterogeneous with respect to information and forecasting ability .
different models in the economic literature have addressed the problem of the origin of herding behavior among financial forecasters in a bayesian game theoretic setting ( see e.g. @xcite ) , proving that if an analyst aims at maximizing the value of his reputation with investors ( or the chances that investors believe he s a good forecaster ) then it may actually be more profitable for him to replicate other agents forecasts rather than putting forward a guess based on his private information . a recent agent - based model proposed by curty and marsili @xcite focuses instead on the limits that herding imposes on the efficiency with which information is aggregated .
specifically , it was shown that when the fraction of herders in a population of agents increases , the probability that herding produces the correct forecast ( i.e. that individual information bits are correctly aggregated ) undergoes a transition to a state in which either all herders forecast rightly , or no herder does . in this note
we study a variation on the theme by curty and marsili , aiming at characterizing further the dynamical interplay between learning and heterogeneity of information in a population of agents aiming to predict a continuous exogenous random variable ( learning was briefly considered in a discrete forecasting setting also in @xcite ) . at each time step
, every agent is required to formulate a forecast either using his private information or by herding with a group of peers and selects the strategy to adopt based on his past performance .
we show that the structure of the agents choices changes significantly depending on the heterogeneity of information .
in particular , herding becomes increasingly preferred by agents as information becomes more and more unevenly spread across the population .
however , the herding coefficient ( measured by the ratio of the forecast error to the forecast dispersion ) peaks roughly where informational inhomogeneity is maximal , implying that learning in such conditions does not allow for an efficient aggregation of the available information .
the results we discuss are mostly obtained by computer simulations .
deeper analytical progress ( beyond the simple considerations made here ) could be possible either along the lines of @xcite or by reasonably simplifying the coupled herding and learning mechanisms .
following @xcite , we consider a population of agents ( labeled @xmath2 ) who have to forecast at each time step @xmath3 a continuous random variable @xmath0 drawn independently at each @xmath4 from a uniform distribution in @xmath5 $ ] .
the forecast @xmath6 of agent @xmath7 at time @xmath4 is taken to be correct if @xmath8 , in which case we shall write @xmath9 . in
what follows , the resolution parameter @xmath10 will be fixed to the value @xmath11 so as to focus the study on the remaining parameters . in order to fomulate his forecast , every agent must choose between using his private information ( strategy lebeled @xmath12 ) and herding ( label @xmath13 ) .
in the former case , agents simply propose a forecast @xmath14 which is correct with probability @xmath15 , initially unknown to @xmath7 .
note that @xmath16 is the probability with which a random uniform guess in @xmath5 $ ] is within @xmath10 from @xmath0 , hence we are assuming that the private information has better - than - random predictive power .
larger values of @xmath17 correspond to higher forecasting abilities .
we assume , along the lines of @xcite , that the @xmath17 s are sampled independently from @xmath18\label{distr}\ ] ] tuning @xmath19 one passes from a situation in which almost all agents are well informed ( small @xmath19 ) , to one in which almost all agents have no forecasting ability ( large @xmath19 ) ; as @xmath19 increases , the information heterogeneity ( or the a priori forecasting ability ) reaches a maximum when @xmath20 , corresponding to a uniform distribution .
we shall denote by @xmath21 the average value of @xmath17 , given by @xmath22 .
when herding , an agent uses instead a prediction @xmath23 obtained by pooling a group @xmath24 of @xmath25 peers chosen randomly and uniformly for each @xmath7 ( our results do not appear to depend significantly on @xmath25 as long as @xmath25 is not extensive ; we shall use @xmath26 here ) .
this represents , in analogy with e.g. @xcite , the `` contact network '' of the agent . our choice for a plain erds - renyi topology parallels that made in @xcite , but results
are expected to change if different topologies are employed .
the herding forecast is defined as the fixed point of the iterative process @xmath27 with @xmath28 . in words , at time step @xmath4 every agent only interacts with his peers whose initial ( private ) guess is sufficiently close , within a range measured by @xmath29 , to his @xcite .
the `` effective interaction range '' is a crucial parameter to model social interaction and has proved to play a non trivial role in other contexts @xcite .
note that the number @xmath30 of such peers is obviously bounded by @xmath25 but it fluctuates in time ; furthermore , in the above sum @xmath31 denotes the forecast of agent @xmath32 who may be herding ( in which case @xmath33 changes with @xmath34 ) or not ( in which case it is fixed to @xmath35 ) .
this defines a dynamical `` imitation network '' onto the contact network , close in spirit to that defined in @xcite in the context of minority games . in this case , however , the averaging operation through which herding is performed does not allow for a straightforward identification of an imitation hierarchy . as in many other instances of games with learning @xcite
, we take agents to be inductive : they monitor the performances of their two strategies over time via scoring functions indexed by @xmath36 that are updated in time according to @xmath37 and the agent s chosen strategy @xmath38 is selected by a logit rule with learning rate @xmath39 : @xmath40 the different parameters appearing above have all been introduced and discussed at length in the context of minority games and related models ( see e.g. @xcite for a broad review ) . in the present case , @xmath41 denotes a memory length parameter , roughly corresponding to the inverse of the time scale over which agents preserve a memory of the past performance of their strategies .
@xmath42 denote instead the cost ( or the incentive ) faced by each player to get his private information or to herd .
@xmath43 denotes the profit faced by agent @xmath7 at time step @xmath4 . for the sake of simplicity , we reward agents who guess correctly with one point , while we take one point from agents who guessed incorrectly .
finally , @xmath44 is a parameter that encodes for a tunable stochasticity in the agents choice rules , with deterministic behavior recovered for @xmath45 .
note that at every time step both the strategy scores of every agent are updated .
starting from initial conditions @xmath46 , we are interested in observing the steady state behavior of the following quantities : * the herding probability , measured by the time - averaged fraction of herders : @xmath47 * the probability of success of strategy @xmath13 ( averaged over time and agents ) : @xmath48 where @xmath49 if the event @xmath50 is true , and zero otherwise * the time- and agent - averaged forecast error : @xmath51 * the time - averaged forecast dispersion : @xmath52 * the _ herding ratio _
@xmath53 the way in which agents produce their forecasts , and as a consequence the herding ratio will be affected by all of the above parameters but , most importantly , will depend on the information heterogeneity , measured by @xmath19 .
note that @xmath54 .
we remind , finally , that , under the emh , @xmath55 , so one would expect @xmath56 .
we begin by analyzing the case in which every agent consults all his peer group ( @xmath57 or @xmath58 for all times ) , has infinite score memory ( @xmath59 ) and learning rate ( @xmath60 , corresponding to a deterministic choice rule ) and faces no costs ( or receives no incentives ) to use his strategies ( @xmath61 ) .
this will be used as a reference situation to evaluate the impact of the above parameters on the game .
[ uno ] shows the average herding success probability @xmath62 as well as the fraction of herders @xmath63 as a function of @xmath19 .
probability @xmath62 of successful herding versus @xmath19 .
markers denote results for individual samples ; the continuous line stands for the average .
left inset : @xmath21 ( dashed line ) and @xmath62 ( continuous line ) versus @xmath19 .
right inset : @xmath63 versus @xmath19 .
simulation parameters : @xmath64 , @xmath26 , @xmath65.,width=377 ] we see that when agents are well informed ( @xmath19 small ) herding outperforms the private information strategy and the majority of agents correctly learn to herd to increase their success probability well above @xmath21 . for large @xmath19 ,
on the other hand , when agents have limited predictive ability , they learn to use their private signal as it slightly outperforms the herding forecast .
this behavior , including the decay of @xmath62 that is observed for large @xmath19 ( @xmath66 as @xmath67 ) essentially parallels that observed in @xcite .
the behavior of @xmath63 , by contrast , displays a sharp drop as informational heterogeneity increases beyond the point where , as is clear from the inset , the private signal outperforms herding .
note that for large @xmath19 sample - to - sample fluctuations become larger and larger as @xmath68 , so that realizations with many herders ( up to a 70% fraction of the population ) coexist with realizations with few herders ( less than 10% ) .
it is worthwhile to inspect the dependence of sample - to - sample fluctuations on @xmath19 more closely , see fig .
[ fluc ] .
sample - to - sample variance of @xmath63 versus @xmath19 for different values of @xmath69 .
simulation parameters : @xmath26 , @xmath65 .
the dashed vertical line marks the position of @xmath70 in this case .
, width=377 ] besides the increase for large @xmath19 , a sharp peak ( roughly independent of the system size ) appears for intermediate @xmath19 , marking a qualitative change in the game s macroscopic properties .
again , this signals a strong sample - dependence of the fraction of herders and occurs when the payoffs of the two strategies become comparable .
such an effect is absent in the one - shot game and is induced by learning .
note that the fraction of herders for @xmath19 s close the the peak is consistent with the results of @xcite . in turn
, the herding ratio displays the behavior shown in fig . [ due ] .
inverse herding ratio @xmath71 versus @xmath19 .
inset : @xmath72 ( continuous line ) and @xmath73 ( dashed line ) versus @xmath19 .
simulation parameters : @xmath74 , @xmath65.,width=377 ] one observes a sharp minimum in @xmath71 around @xmath75 , where it attains a value consistent with that found empirically in @xcite , where @xmath76 . away from this crossover region ( where ,
we remind , information heterogeneity is maximal ) , @xmath77 is instead closer to the emh limit @xmath78 .
this suggests that inductive agents indeed manage to aggregate information quite efficiently when it is distributed uniformly across them , regardless of the quality of the information .
this process is no longer possible in presence of a more diversified information landscape .
a naive analytical estimate of the value of @xmath19 where the game undergoes a crossover can be obtained by arguing as in @xcite . denoting by @xmath79 the probability of a correct forecast
, one has @xmath80 where @xmath62 denotes the probability of a correct forecast by herding . neglecting both the learning and the herding dynamics
, @xmath62 is given by the probability that the majority of the peer group members have a correct forecast , i.e. @xmath81 a qualitative change is expected to occur when @xmath82 .
one easily sees from ( [ b ] ) that this condition is satisfied ( besides the trivial solutions @xmath83 or @xmath84 ) when @xmath85 , implying @xmath86 for @xmath65 the prediction @xmath70 agrees well with the numerical experiments .
( [ bstar ] ) also characterizes the role of the resolution parameter : as @xmath10 increases , the number of effective alternative `` states '' of the variable @xmath0 ( given by @xmath87 ) decreases and @xmath88 becomes larger , extending the range of effectiveness of herding as a forecasting strategy .
( the fact that @xmath10 can not exceed @xmath89 was implicit in the model s definition . ) note that the nave guess for the fraction of herders given by @xcite @xmath90 according to which agents with @xmath91 are assumed to herd asymptotically , does not produce the correct results in the present model , suggesting that the specific forms of @xmath63 and of its fluctuations , including the large @xmath19 increase , are likely due to the learning dynamics itself .
we now focus our analysis on the additional model parameters described in sec .
2 . for the sake of simplicity , we will study one of them at a time as variations of the basic case investigated above .
figures [ param ] and [ param2 ] show specifically how the fraction of herders and the herding ratio change when these parameters are varied for different values of @xmath19 .
fraction of herders @xmath63 versus @xmath95 for @xmath96 ( triangles ) , @xmath20 ( circles ) and @xmath97 ( squares ) .
simulation parameters : @xmath98 , @xmath65.,width=453 ] starting with @xmath99 $ ] , which as said above introduces a finite memory of past scores ( for @xmath59 there is no loss of memory while for @xmath100 the score is exponentially discounted in time with a characteristic time given by @xmath101 , see @xcite for a solvable model that highlights its role in the context of minority - like games ) , we see both @xmath63 and @xmath77 are essentially insensitive to @xmath92 except in the `` critical region '' where informational heterogeneity is maximal . in this case , a larger @xmath92 ( or a shorter memory ) leads to a decrease of both @xmath63 and @xmath77 , implying that a finite memory may lead closer to the rational aggregation of the available information when the latter is distributed in a highly in - homogeneous way .
similar conclusions can be reached by analyzing the learning rate @xmath44 that tunes the amount of stochasticity in agents choice .
its impact appears to be most remarkable close to @xmath20 , when a ( small ) degree of randomness in the strategy selection process helps to reduce the fraction of herders and hence to bring @xmath77 closer to the rational value of @xmath1 .
inverse herding ratio @xmath71 versus @xmath95 for @xmath96 ( triangles ) , @xmath20 ( circles ) and @xmath97 ( squares ) .
simulation parameters : @xmath98 , @xmath65.,width=453 ] the role of parameter @xmath102 $ ] in the context of opinion dynamics models has been discussed e.g. in @xcite .
it takes into account the chance that people may prefer to interact , and possibly accept to adjust their beliefs , only with opinions not too far from their initial ones . for @xmath57 all opinions in the peer group are treated equally .
for other values of the parameter , instead , the sum in is restricted only to agents whose forecasts are within @xmath29 from each agent s initial choice .
it is interesting to note that when herding represents the best strategy , i.e. around @xmath20 , the fraction of herders is almost insensitive to different values of @xmath29 ( which possibly suggests also a weak dependence of the results on @xmath25 ) . nevertheless reducing @xmath29
the herding ratio appears to get smaller , indicating that a more rational aggregation of information may occur .
by contrast , for larger @xmath19 agents may be unable to identify @xmath12 as the most likely successful strategy if @xmath29 is too small .
coming finally to the incentives @xmath103 ( which are known to have a far from trivial impact on minority games , see e.g. @xcite ) , we focus on the dependence of @xmath63 on the parameter @xmath104 , which is easily understood to be the relevant quantity in this case .
as to be expected , incentives to herding , or higher costs for using the private signal , shift agents towards the @xmath13 strategy , and lead to an increase of @xmath77 ( and vice versa for incentives to the @xmath12 strategy ) .
this crossover appears to be smooth only when @xmath19 is sufficiently large . for smaller
@xmath19 it sharpens and one observes a steep jump at @xmath105 in both @xmath63 and @xmath77 , reminiscent of similar effects induced by incentives or tobin taxes in minority games @xcite . in this case , agents appear to polarize on the herding strategy as soon as a small incentive is available , leading to disastrous consequences for @xmath77 and suggesting the existence of a first - order transition in @xmath106 ( though a more refined numerical analysis would be needed to clarify this point ) .
the qualitative outlook is however essentially unchanged with respect to the the case of larger @xmath19 s .
we have studied here a simple forecasting game with inductive agents who must formulate a forecast by either using their private information or by herding with a group of peers .
the quality of the information at an agent s disposal is measured by the a priori predictive ability of his private signal , and we investigate how the game s overall properties are affected by increasing informational heterogeneity in the population .
our main result is that inductive agents may be unable to produce rational forecasts when the heterogeneity is large . in this situation ,
the herding ratio becomes significantly larger than one , taking on values similar to those measured empirically in the financial literature @xcite .
we have also observed that the efficacy of herding depends strongly on the distribution of information .
the role of several parameters of interest in the context of games with inductive agents has finally bee analyzed .
generically speaking , a finite learning rate , a shorter memory or a smaller interaction range may all contribute to reduce the herding ratio when the informational heterogeneity is large .
the interest in forecasting games is based on the fact that they present a simple outlook and a rich phenomenology that allows to shed some light on the process of information aggregation and its limits .
it is however difficult to establish a direct contact between these toy models and financial markets , which still represent the main source of empirical data .
one possible step in this direction that is worth exploring would consist in coupling the event to be forecasted , @xmath0 here , with the agents choices , as done for example in @xcite .
@xmath0 would play in these models the role of a ` price ' . depending on the form of the payoff one
would then observe a dynamical feedback between learning and ` price ' leading to a rich outlook possibly similar to that studied in minority games . | we investigate a toy model of inductive interacting agents aiming to forecast a continuous , exogenous random variable @xmath0 .
private information on @xmath0 is spread heterogeneously across agents .
herding turns out to be the preferred forecasting mechanism when heterogeneity is maximal .
however in such conditions aggregating information efficiently is hard even in the presence of learning , as the herding ratio rises significantly above the efficient - market expectation of @xmath1 and remarkably close to the empirically observed values .
we also study how different parameters ( interaction range , learning rate , cost of information and score memory ) may affect this scenario and improve efficiency in the hard phase . |
the existence of dark energy and dark matter is indirectly shown by the observations of cosmic microwave background ( cmb ) , large scale structure ( lss ) of the universe , super novae of type ia ( snia ) , and related observational probes .
dark energy is usually regarded as a cosmological constant @xmath4 in the standard model of cosmology , @xmath4cdm model . however ,
in general it represents some energy which have a negative pressure with equation of state parameter being less than @xmath5 . on the other hand , dark matter , in particular cold dark matter ( cdm ) means the nonrelativistic unknown matter .
there are many candidates for dark energy ( for review , see@xcite ) and dark matter@xcite .
however , it is not yet clear which model is viable one . in this paper
, we will consider mimetic matter model @xcite , which was first proposed as a model of dark matter .
however , it was realized later that it can be treated as a model of dark energy by adding the potential term of the scalar field @xcite .
mimetic matter model is a conformally invariant theory by making the physical metric as a product of an auxiliary metric and the contraction of an auxiliary metric and the kinetic term of the scalar field . whereas , the equivalence between mimetic matter model and a scalar field model with a lagrange multiplier @xcite is shown in @xcite .
therefore , we treat mimetic matter model as a scalar field model with a lagrange multiplier in this paper .
the contents of the paper are the following . in sec .
ii , we consider the background evolution of the universe in general mimetic matter model .
the cosmological reconstruction of the model , which is the way to construct the action which may realize the given evolution history of the universe by the choice of potential , is done and @xmath4cdm - like era is explicitly executed for corresponding scalar potential .
the evolution of the matter density perturbation is discussed in sec .
first , mimetic matter model without potential is used to evaluate whether mimetic dark matter also behaves as dark matter at the cosmological perturbations level .
next , we investigate the matter density perturbation in general mimetic matter model with scalar potential . in sec .
iv , the possibility of the phantom non - phantom transition is explicitly demonstrated in mimetic matter model .
the behavior of the cosmological perturbations is also discussed , especially the behavior close to finite - time future singularity .
concluding remarks are given in sec . v. the units of @xmath6 are used and gravitational constant @xmath7 is denoted by @xmath8 with the planck mass of @xmath9gev in this paper .
we consider the action of mimetic matter model@xcite : @xmath10+s_\mathrm{matter } , \label{10}\end{aligned}\ ] ] where @xmath11 is an arbitrary function of the scalar field @xmath12 and @xmath13 is a lagrange multiplier .
the einstein equation obtained from eq .
( [ 10 ] ) is @xmath14 where @xmath15 and @xmath16 is the energy momentum tensor of the usual matter . on the other hand ,
the equation given by the variation with respect to the scalar field @xmath12 is the following one : @xmath17 where @xmath18 @xmath19 @xmath20 .
the constraint equation given by the variation of @xmath13 is @xmath21 when the spatially flat friedmann - lemaitre - robertson - walker ( flrw ) metric , @xmath22 , is taken , flrw equations are written by @xmath23 where @xmath24 .
@xmath25 is the energy - density of the matter and @xmath26 is the equation of state ( eos ) parameter expressed by @xmath27 . from eqs .
( [ fe ] ) and ( [ lambda ] ) we have @xmath28 the equation of continuity of the matter is @xmath29 if @xmath26 is constant , one can integrate eq .
( [ 15 ] ) as @xmath30 where @xmath31 is an integration constant . while , combining eqs .
( [ fl00 ] ) , ( [ flii ] ) , and ( [ bla ] ) it gives the equations @xmath32 and @xmath33 therefore , all the solutions for the variables will be found if eq .
( [ sol3 ] ) can be solved with respect to @xmath34 . from the other side
, one can say that this model is able to realize arbitrary evolution history of the universe by using the arbitrariness of the function @xmath35 .
indeed , one can choose the form of @xmath35 for the arbitrary function @xmath36 such that @xmath37 as @xmath38 some forms of the function @xmath35 are considered in ref.@xcite .
here , we consider some other examples of @xmath35 , which are adequate to the cosmological observations and yield the accelerating universe . in
particularly , the form of @xmath35 which can yield the phantom non - phantom transition is treated in sec .
[ phantom ] . in the following , we assume the value of eos parameter of the usual matter as @xmath39 .
wcdm model consists of cold dark matter and a fluid with eos parameter @xmath40 . for the case @xmath41
, wcdm model is equivalent to the @xmath4cdm model , so its background evolution can be realized by choosing the form of the potential as @xmath42 . in the case of @xmath43 , we can reproduce the background evolution of the universe in wcdm model if the effective energy density @xmath44 and the effective pressure @xmath45 satisfy the following conditions : @xmath46 and @xmath47 where @xmath48 and @xmath49 are the current energy - density of dark matter and that of a fluid , respectively . here , the current value of the scale factor is chosen as @xmath50 .
thus , @xmath51 the specific form of the potential is very complicated even if it can be obtained by solving eq .
( [ fl00 ] ) .
however , one can easily obtain some approximated forms of @xmath35 . when @xmath52 , an approximated solution of wcdm model is given by @xmath53 ^{\frac{1}{3 } }
t^{\frac{2}{3 } } \left \ { 1 + \left [ ( 1+w)^2 \frac{\rho
_ { w0}}{\rho _ { b0 } + \rho _ { dm0 } } \left ( \frac{3}{4}\kappa ^2 ( \rho _ { b0 } + \rho _ { dm0 } ) t^2 \right ) ^{-w } \right ] ^\frac{1}{3(1+w ) } \right \ } , \label{appa}\ ] ] where @xmath54 is the current energy - density of the baryonic matter , because the solutions of eq .
( [ fl00 ] ) in the matter dominant universe and in the fluid matter dominant universe are given by @xmath55 and @xmath56^{1/(3(1+w))}t^{2/(3(1+w))}$ ] , respectively .
therefore , an approximated form of potential @xmath35 is expressed as @xmath57^{-\frac{1}{3(1+w ) } } } \nonumber \\
\times \left \ { - \frac{3w + w^2}{(1+w)^2 } + \frac{w^2}{(1+w)^2 } \frac{1 } { 1 + \left [ ( 1+w)^2 \frac{\rho _ { w0}}{\rho _ { b0 } + \rho _ { dm0 } } \left ( \frac{3}{4}\kappa ^2 ( \rho _ { b0 } + \rho _ { dm0 } ) \phi ^2 \right ) ^{-w } \right ] ^{-\frac{1}{3(1+w ) } } } \right \ } .
\label{appv}\end{aligned}\ ] ] eq .
( [ appa ] ) is a very rough approximation as a kind of the interpolation . to improve the accuracy of the approximation
, one can use the current value of the hubble rate @xmath58 as : @xmath59^{\frac{1}{3 } } + \left [ \frac{3}{4}(1+w)^2\kappa ^2 \rho _ { w0 } t^2 \right ] ^{\frac{1}{3(1+w ) } } + c ( h_0 t)^{\frac{2+w}{3(1+w ) } } , \label{imdapp}\end{aligned}\ ] ] where @xmath60 is a constant tuned to satisfy @xmath61 when @xmath62 .
the value of @xmath60 is evaluated by solving a simultaneous equation @xmath63 , @xmath64 with respect to @xmath65 and @xmath60 . in the case of @xmath66 ,
@xmath67 years and @xmath68 are obtained .
the comparisons between the exact wcdm model and eq .
( [ imdapp ] ) when @xmath66 and @xmath69 are shown in fig .
it seems that eq .
( [ imdapp ] ) is not a bad approximation when @xmath26 is large , although only three points @xmath70 are used as the interpolation points . ) .
thick solid line , solid line , dashed line , and dotted line represent eq .
( [ imdapp ] ) with @xmath66 , eq .
( [ imdapp ] ) with @xmath69 , wcdm model with @xmath66 , and wcdm model with @xmath69 , respectively .
@xmath71 km / s / mpc , @xmath72 , and @xmath73 are assumed . ]
one can rewrite eqs .
( [ sol2 ] ) and ( [ sol3 ] ) by using @xmath0 and @xmath75 instead of @xmath76 and @xmath77 as @xmath78 where @xmath54 is current energy - density of the baryons .
@xmath0 is defined by @xcite @xmath79 in the case of the @xmath4cdm model , @xmath0 is in accord with the current matter fraction @xmath80 .
however , in the recent paper@xcite , a discrepancy between cmb observations ( planck+wp ) and bao observations is pointed out .
it is pointed out that the value of @xmath81 is larger for larger redshift .
the explicit values of @xmath82 for planck+wp and @xmath83 for bao observations are @xmath84@xcite and @xmath85@xcite , respectively . of course
, these values are incompatible in the @xmath4cdm model , because @xmath83 is constant with respect to redshift .
however , we can construct a model that satisfies these conditions . if we consider the following scale factor as a solution of eqs .
( [ fl00 ] ) and ( [ flii ] ) : @xmath86 , \nonumber \\ g(t)=\left ( 0.122 + 0.04 \frac{t_{cmb}}{t } - 0.02 \frac{t_{cmb}^2}{t^2 } \right ) /h^2 , \label{om4}\end{aligned}\ ] ] where @xmath87 is a time of @xmath88 , then @xmath89 for small redshift and @xmath90 at @xmath88 is realized . the potential which causes such an expansion history
is given by using eq .
( [ rec ] ) , however , it is not written down explicitly , being very complicated .
similarly , we can reconstruct any requested universe history . in the next section ,
the evolution of the matter density perturbations is studied .
in this section , we investigate the growth rate of the matter density perturbations ( for general review of cosmological perturbations theory , see @xcite ) .
when we use the metric of the newtonian gauge , @xmath91 , then @xmath92 , @xmath93 , @xmath94 , @xmath95 , and @xmath96 elements of the linearized einstein equation in the fourier space are represented as follows , respectively : @xmath97 where we defined the sound speed by @xmath98 , the matter density perturbation by @xmath99 and the scalar perturbation of the four velocity as @xmath100 .
the linearized equations of ( [ fe ] ) and ( [ lambda ] ) are given as follows : @xmath101 from the perturbation of the equation of continuity , @xmath102 , one gets @xmath103 in the following , we treat the equation of state parameter and the sound speed of the matter as @xmath104 by assuming the era after matter dominance .
first , we consider the case @xmath105 as a simple example .
then , the field @xmath13 behaves as dark matter in the flat flrw space - time@xcite .
let us investigate whether the field @xmath13 also behaves as dark matter when we consider cosmological perturbations .
first , eqs .
( [ pla ] ) and ( [ rij ] ) give the relation @xmath106 . by applying the subhorizon approximation , @xmath107 and @xmath108 , eqs .
( [ r00 ] ) and ( [ pfe ] ) , then , yield the following equations : @xmath109 differentiating eq .
( [ lfesr ] ) with respect to @xmath110 gives @xmath111 where @xmath112 . in the case of @xmath113
, we can obtain the following equation by using eqs .
( [ t0 ] ) , ( [ ti ] ) , @xmath114 taking into account @xmath115 , we obtain @xmath116 where @xmath117 and @xmath118 . by using eq .
( [ r00sr ] ) , we finally obtain @xmath119 now , one can recognize the equivalence between mimetic dark matter and fluid dark matter at the linear perturbation level because eq .
( [ totsr ] ) is same as that of the @xmath4cdm model .
the equivalence of mimetic dark matter and fluid dark matter had been shown in ref .
@xcite .
in this section , we consider more involved case @xmath120 .
the perturbation equations ( [ r0i ] ) , ( [ rii ] ) , and ( [ t0 ] ) can be simplified as follows by using eqs .
( [ rii ] ) and ( [ pla ] ) : @xmath121 eliminating the term proportional to @xmath122 from eqs .
( [ 410 ] ) and ( [ 430 ] ) gives @xmath123 then , we can derive the following evolution equation of the matter density perturbation : @xmath124 eq .
( [ 450 ] ) is quite different from that in quintessence model .
there are no oscillating solutions , moreover , the quasi - static solutions are modified .
in fact , if we assume @xmath125 then @xmath126 therefore , the solution which satisfies eq .
( [ 460 ] ) is not the solution of eq .
( [ 450 ] ) except for the cases @xmath127 or @xmath128 or @xmath129 .
it is necessary to be careful because there is an ambiguity to express eq .
( [ 470 ] ) due to adding the terms proportional to eq .
( [ 460 ] ) or the derivatives of eq .
( [ 460 ] ) are allowed . in @xmath130-essence model ,
the growth equation of the matter density perturbation is also represented by fourth order equation@xcite as in eq .
( [ 450 ] )
. however , in general , the sound speed read off from the field equation ( [ pfe ] ) is not zero in the @xmath130-essence model , so there are no modifications to the quasi - static solution .
thus , the modification to the quasi - static solution in eq .
( [ 450 ] ) is the peculiar feature in this theory .
however , in the matter dominant era or in @xmath13 dominant era , the term @xmath131 is less than the other terms in eq .
( [ 470 ] ) .
therefore , we need to consider the era where there is almost same amount of dark energy and dark matter such as the current universe to distinguish eq .
( [ 450 ] ) from eq .
( [ 470 ] ) . whereas , in the case of @xmath132 , eq . ( [ 450 ] )
recovers the differentiated equation of ( [ totsr ] ) , correctly . the procedure to obtain eq .
( [ 450 ] ) is 1 .
differentiating eq .
( [ 440 ] ) with respect to @xmath110 2 .
eliminating the terms proportional to @xmath133 from the equation derived at the first step by using eq .
( [ 420 ] ) 3 . differentiating the equation derived at the second step with respect to @xmath110 4 . eliminating the terms proportional to @xmath133 from the equation derived at the third step by using eq .
( [ 420 ] ) 5 . differentiating the equation derived at the fourth step with respect to @xmath110 6 . eliminating the terms proportional to @xmath133 from the equation derived at the fifth step by using eq .
( [ 420 ] ) 7 . eliminating the terms proportional to @xmath134 from the equations derived at the second , fourth , and sixth step by using eq .
( [ 440 ] ) 8 . eliminating the terms proportional to @xmath135 and @xmath136 by using the equations derived at the seventh step .
let us evaluate the difference between eqs .
( [ 450 ] ) and ( [ 460 ] ) in the case that mimetic matter behaves like wcdm model . for a correct evaluation
, we will not use eq .
( [ appv ] ) but use eqs .
( [ wlambda ] ) and ( [ wpotential ] ) .
if we utilize @xmath137 instead of time @xmath110 , we can describe all the functions in eq .
( [ 450 ] ) as the functions of @xmath138 by using eqs .
( [ fl00 ] ) , ( [ flii ] ) , ( [ sol1 ] ) , ( [ wlambda ] ) , and ( [ wpotential ] ) .
therefore , eq . ( [ 450 ] ) can be solved at least by numerical calculations . in mimetic wcdm .
the dot - dash line , the solid line , and the dashed line express @xmath139 in the cases @xmath140 , @xmath41 , and @xmath141 , respectively .
@xmath142 , @xmath143 , and @xmath144 are assumed to depict the above figure .
initial conditions are @xmath145 and @xmath146 . ]
[ f1 ] shows the changes of the growth rate function @xmath147 , with respect to @xmath148 , where @xmath149 .
[ f2 ] and [ f3 ] express the differences between mimetic wcdm and wcdm model in the growth rate function .
here , @xmath142 , @xmath143 , and @xmath144 , where @xmath150 is the current energy fraction of the matter x , are assumed to depict the figures . as the initial conditions , @xmath145 and @xmath146 are used , because , in the matter dominant era , i.e. @xmath151 , we could use eq .
( [ totsr ] ) and can obtain @xmath152 and @xmath153 . as noted in sec .
[ mdm ] and at the beginning of sec .
[ mmp ] , there is no difference between mimetic model and wcdm model when @xmath41 .
however , in the case @xmath43
@xmath154 @xmath40 , there is difference which come from the term @xmath155 in eq .
( [ 470 ] ) .
one can see in figs .
[ f2 ] and [ f3 ] that the differences are small but surely exist .
the term @xmath155 is not relevant at the matter dominant era , so we can also find that the differences become small as redshift z becomes large . when @xmath141 .
] when @xmath140 . ]
the behavior of the growth rate function @xmath139 when the background solution of the universe is given by eq .
( [ om4 ] ) is expressed in fig .
[ f4 ] . when the evolution of the background space - time is described by eq .
( [ om4 ] ) .
solid , dashed , and dot - dashed curves represent @xmath139 in this model , @xmath139 in the @xmath4cdm model with @xmath156 , and @xmath139 in the @xmath4cdm model with @xmath157 , respectively .
initial conditions are @xmath145 and @xmath146 . ]
@xmath158 , @xmath159 is assumed to depict fig .
although @xmath0 is a smoothly changing function from @xmath160 to @xmath161 when we consider @xmath162 , however , the growth rate function @xmath139 is larger than that of the @xmath4cdm model with @xmath163 .
this effect comes from the difference between eqs .
( [ 450 ] ) and ( [ 460 ] ) .
in fact , if we evaluate @xmath139 for a constant @xmath74 model with @xmath164 in eq .
( [ 450 ] ) , we have a larger value than that of the dynamical @xmath165 ( [ om4 ] ) .
in the quintessence model , the value of eos parameter @xmath166 is constrained as @xmath167 so that phantom behavior @xmath168 can not be realized . here
, we formally consider the possibility to construct a model , which can describe the transition from @xmath169 to @xmath168 or from @xmath168 to @xmath169 . if the hubble rate @xmath76 is given by the following function , @xmath170 @xmath171 @xmath172 or @xmath173 then for the case , @xmath174 or @xmath175 , @xmath176 , @xmath177 , or @xmath178 or @xmath179 , where @xmath180 and @xmath181 , effective eos parameter @xmath182 satisfies @xmath183 , respectively@xcite . here ,
@xmath184 , and @xmath185 are real valued constants and @xmath186 and @xmath187 are positive constants .
( [ uid ] ) could also describe the unified model of the inflation and dark energy because there are two regimes @xmath174 and @xmath175 satisfying @xmath183 , the era @xmath174 and the era @xmath175 correspond to the inflation and a dominance of dark energy , respectively .
the forms of the potential which cause the above expansion rates of the universe are obtained by using eq .
( [ rec ] ) as @xmath188,\end{aligned}\ ] ] @xmath189,\ ] ] @xmath190 , \label{ex3v}\ ] ] and @xmath191 , \ ] ] respectively . here
, eos parameter of the matter is assumed to be @xmath192 . as seen in eq .
( [ wlambda ] ) , the phantom crossings in these examples are realized by the transition from @xmath193 to @xmath194 when there is no usual matter .
hence , in mimetic matter model with potential the phantom - divide transitions maybe realized by proper choice of the scalar potential .
let us consider the evolution behavior of the matter density contrast in the model with eq .
( [ ex3v ] ) as an example . in this model
, the universe will experience the singular evolution in @xmath76 if @xmath195 is larger than the current time .
it is known that there may occur several finite - time future singularity scenarios of the universe as @xcite which are classified as @xcite : * type i ( big rip ) : for @xmath196 , @xmath197 , @xmath198 , and @xmath199 .
this also includes the case of @xmath200 , @xmath201 being finite at @xmath195 . *
type ii ( sudden ) : for @xmath196 , @xmath202 , @xmath203 , and @xmath199 . *
type iii : for @xmath196 , @xmath202 , @xmath198 , and @xmath199 .
* type iv : for @xmath196 , @xmath202 , @xmath204 , @xmath205 , and higher derivatives of @xmath206 diverge .
this also includes the case in which @xmath201 ( @xmath200 ) or both of @xmath201 and @xmath200 tend to some finite values , while higher derivatives of @xmath206 diverge .
singularity in the case under consideration corresponds to type i or type iii .
when @xmath207 and @xmath208 years , we have a solution of eq .
( [ ex3 ] ) as @xmath209 where @xmath210 is an integration constant .
it is chosen as @xmath211 to comply with the observed background evolution of the universe . in this case , @xmath212 is simply expressed as @xmath213 .
the evolution behavior of the matter density perturbation is changed around the singularity because the approximation @xmath214 no longer holds in eq .
( [ 450 ] ) . in the model ( [ ex3v ] ) ] in the model ( [ ex3v ] ) ] in fig .
[ f5 ] , the behavior of the growth rate function @xmath215 until the current time is expressed . on the other hand ,
the future evolution of @xmath216 is expressed in fig .
[ f6 ] shows singular behavior in @xmath216 around @xmath217 years , although the singularity in @xmath34 and @xmath76 occurs at @xmath218 years .
the reason why such a behavior of @xmath216 occurs is instability of the perturbations of phantom matter even at the classical level . to be concrete , the flip of the sign of the kinetic term would cause the instability .
a decay of perturbations much earlier of future singularity found in @xcite is just the same as this instability ( for discussion of problems of phantom - divide transitions in perturbations , see @xcite ) .
in this paper , we have investigated the evolution of the matter density perturbation in mimetic matter model . in sec .
ii , the reconstruction , the way to construct a model which can describe arbitrary evolution history of the universe , has been explicitly shown and some examples , including mimetic wcdm model and evolving om(z ) model , have been given . at the beginning of sec .
iii , the growth rate of the matter density perturbation has been investigated when potential term is not dynamical . as a result
, it has been shown that mimetic dark matter also behaves as cdm at the perturbation level . in the case of dynamical potential
, it has been shown that the differential equation of the matter density perturbation is changed .
in general , this equation is different from that in the @xmath4cdm model and the quasi - static equation of the matter density perturbation in quintessence model .
we have presented the behavior of the growth rate function in mimetic wcdm model and in a model of dynamical @xmath0 as examples .
the difference seems not to happen in mimetic wcdm , but it clearly appears in dynamical @xmath0 . in sec .
iv , we have proved the possibility to realize phantom non - phantom transition in mimetic matter model .
it has been shown that phantom non - phantom transition is possible because the sign of kinetic term of the scalar field is determined by the sign of the lagrange multiplier .
however , usually phantom brings an instability .
such an instability occurs not only as quantum effect but also as classical effect@xcite , and it breaks the perturbation theory as is explicitly demonstrated
. it would be interesting to study the accelerating universe evolution and cosmological perturbations in the generalizations of mimetic matter model which include mimetic @xmath219 gravity @xcite , higher - derivative mimetic model@xcite or its further generalizations@xcite . in this case
, we have not only scalar potential but also extra function like @xmath219 or higher - derivative term .
hence , the cosmological reconstruction can be executed quite easily .
this will be studied elsewhere .
jm would like to acknowledge a suggestion from a. a. starobinsky .
the work was supported by the russian government program of competitive growth of kazan federal university and , partially , by the russian foundation for basic research grants no .
14 - 02- 00598 .
the research by sdo has been supported in part by mineco ( spain ) , projects fis2010 - 15640 and fis2013 - 44881 .
e. a. lim , i. sawicki , and a. vikman , jcap * 1005 * , 012 ( 2010 ) [ arxiv:1003.5751 [ astro-ph.co ] ] ; s. capozziello , j. matsumoto , s. nojiri and s. d. odintsov , phys .
b * 693 * ( 2010 ) 198 [ arxiv:1004.3691 [ hep - th ] ] .
v. mukhanov , `` physical foundations of cosmology , '' cambridge , uk : univ .
( 2005 ) 421 p ; v. f. mukhanov , h. a. feldman and r. h. brandenberger , phys .
* 215 * ( 1992 ) 203 .
d. s. gorbunov and v. a. rubakov , `` introduction to the theory of the early universe : cosmological perturbations and inflationary theory , '' hackensack , usa : world scientific ( 2011 ) 489 p h. kodama and m. sasaki , prog .
* 78 * ( 1984 ) 1 .
s. nojiri and s. d. odintsov , gen .
. grav .
* 38 * , 1285 ( 2006 ) [ hep - th/0506212 ] .
a. a. starobinsky , grav . cosmol .
* 6 * , 157 ( 2000 ) [ astro - ph/9912054 ] .
s. nojiri , s. d. odintsov and s. tsujikawa , phys .
d * 71 * , 063004 ( 2005 ) [ arxiv : hep - th/0501025 ] .
r. r. caldwell , m. kamionkowski and n. n. weinberg , phys .
* 91 * ( 2003 ) 071301 [ astro - ph/0302506];j . d. barrow , class .
* 21 * ( 2004 ) l79 [ gr - qc/0403084];class .
quant . grav .
* 21 * ( 2004 ) 5619 [ gr - qc/0409062 ] ; v. gorini , a. y. kamenshchik , u. moschella and v. pasquier , phys .
d * 69 * ( 2004 ) 123512 [ hep - th/0311111 ] ; s. nojiri and s. d. odintsov , phys .
b * 562 * ( 2003 ) 147 [ hep - th/0303117];phys . rev .
d * 72 * ( 2005 ) 023003 [ hep - th/0505215 ] ; p. singh , m. sami and n. dadhich , phys .
d * 68 * ( 2003 ) 023522 [ hep - th/0305110 ] ; v. faraoni , phys .
d * 68 * ( 2003 ) 063508 [ gr - qc/0307086 ] ; m. p.
dabrowski , t. stachowiak and m. szydlowski , phys .
d * 68 * ( 2003 ) 103519 [ hep - th/0307128 ] ; m. bouhmadi - lopez , p. f. gonzalez - diaz and p. martin - moruno , phys .
b * 659 * ( 2008 ) 1 [ gr - qc/0612135 ] ; l. fernandez - jambrina and r. lazkoz , phys .
d * 74 * ( 2006 ) 064030 [ gr - qc/0607073 ] ; k. bamba , s. nojiri and s. d. odintsov , jcap * 10 * , 045 ( 2008 ) [ arxiv:0807.2575 [ hep - th ] ] ; h. ghodsi , m. a. hendry , m. p. dabrowski and t. denkiewicz , mon . not .
* 414 * ( 2011 ) 2 , 1517 [ arxiv:1101.3984 [ astro-ph.co ] ] .
a. vikman , phys .
d * 71 * , 023515 ( 2005 ) [ astro - ph/0407107 ] .
w. hu , phys .
d * 71 * , 047301 ( 2005 ) [ astro - ph/0410680 ] .
r. r. caldwell and m. doran , phys .
d * 72 * , 043527 ( 2005 ) [ astro - ph/0501104 ] .
s. dubovsky , t. gregoire , a. nicolis and r. rattazzi , jhep * 0603 * , 025 ( 2006 ) [ hep - th/0512260 ] .
r. p. woodard , lect .
notes phys .
* 720 * , 403 ( 2007 ) [ astro - ph/0601672 ] .
i. sawicki and a. vikman , phys .
d * 87 * , no .
6 , 067301 ( 2013 ) [ arxiv:1209.2961 [ astro-ph.co ] ] .
j. m. cline , s. jeon and g. d. moore , phys .
d * 70 * , 043543 ( 2004 ) [ hep - ph/0311312 ] .
j. garriga and a. vilenkin , jcap * 1301 * , 036 ( 2013 ) [ arxiv:1202.1239 [ hep - th ] ] .
e. elizalde , s. nojiri and s. d. odintsov , phys .
d * 70 * , 043539 ( 2004 ) [ hep - th/0405034 ] .
s. nojiri and s. d. odintsov , phys .
d * 72 * , 023003 ( 2005 ) [ hep - th/0505215 ] .
a. v. astashenok and s. d. odintsov , phys .
b * 718 * , 1194 ( 2013 ) [ arxiv:1211.1888 [ gr - qc ] ] .
s. nojiri and s. d. odintsov , arxiv:1408.3561 [ hep - th ] . l. mirzagholi and a. vikman , arxiv:1412.7136 [ gr - qc ] .
d. momeni , a. altaibayeva and r. myrzakulov , arxiv:1407.5662 [ gr - qc ] .
g. leon and e. n. saridakis , arxiv:1501.00488 [ gr - qc ] . | we investigate the cosmological evolution of mimetic matter model with arbitrary scalar potential .
the cosmological reconstruction , which is the way to construct a model for arbitrary evolutions of the scale factor , is explicitly done for different choices of potential . the cases that mimetic matter model shows the evolution as cold dark matter(cdm ) , wcdm model , dark matter and dark energy with dynamical @xmath0 , where @xmath1/[(1+z)^3 -1],$ ] or phantom dark energy with phantom - non - phantom crossing are presented in detail .
the cosmological perturbations for such evolution are studied in mimetic matter model .
for instance , the evolution behavior of the matter density contrast which is different from usual one , i.e. @xmath2 is investigated . the possibility of peculiar evolution of @xmath3 in the model under consideration is shown .
special attention is paid to the behavior of matter density contrast near to future singularity where decay of perturbations may occur much earlier the singularity . |
detecting and measuring the properties of objects in astronomical images in an automated fashion is a fundamental step underlying a growing proportion of astrophysical research .
there are many existing tasks , some quite sophisticated , for performing such analyses .
regardless of the wavelength at which an image has been made however , each of these tasks has one thing in common : a threshold needs to be defined above which pixels will be believed to belong to real sources .
defining an appropriate threshold is a complex issue and , owing to the unavoidable presence of noise , _ any _ chosen threshold will result in some true sources being overlooked and some false sources measured as real . varying the chosen threshold to one extreme or the other will minimise one of these types of error at the expense of maximising the other .
clearly choosing a threshold to jointly minimise both types of error is not trivial , but even more problematic is that it is not even clear that one can , _ a priori _ , make a well defined estimate of the magnitude of each type of error .
typically this is done by comparing measured source counts with existing estimates for the expected number of sources .
this is an unsatisfactory solution for surveys reaching to new sensitivity limits , or at previously uninvestigated wavelengths ( where there can be no estimate of the expected number of sources from existing studies ) , and also for small area imaging where the properties of large populations are affected by clustering or small number statistics . throughout this paper
we shall use the terms ` source - pixel ' to mean a pixel in an image which is above some threshold , and thus assumed to be part of a true source .
the term ` source ' shall be used to mean a contiguous collection of ` source - pixels ' which corresponds to an actual astronomical object , a star or galaxy for example , whose properties we are interested in measuring . in this work we concentrate on radio images , specifically images produced by radio interferometers .
we emphasise , though , that all of the conclusions presented here are valid for any pixelized map where the ` null hypothesis ' is known at each pixel . throughout this paper the null hypothesis is taken to be the ` sky background ' at each pixel ( after the image is normalized ) .
now consider the following possible criteria applied to some chosen threshold : ( 1 ) that there be no falsely discovered source - pixels ; ( 2 ) that the proportion of falsely discovered source - pixels be some small fraction of the total number of pixels ( background plus source ) ; ( 3 ) that the fraction of false positives ( i.e. the number of falsely discovered sources over the total number of discovered sources ) be small .
the first of these can be achieved by using a very high threshold called the bonferroni threshold @xcite .
this threshold is rarely used since , although guaranteeing no false detections , it detects so few real sources .
the second criterion is most often applied in astronomy , and can be achieved by simply choosing the appropriate significance threshold .
a threshold of @xmath0 , for example , ensures that @xmath1 of the total number of _ pixels _ are falsely discovered .
unfortunately this is not the same as a constraint on the fraction of false detections compared to total number of _
detections_. this quantity is a more meaningful measure to use in defining a threshold for the following reason . consider a @xmath0 threshold in an image composed of @xmath2 pixels ( @xmath3 ) and containing only gaussian noise .
this would yield , on average , 1000 pixels above the threshold .
if real sources are also present , these 1000 pixels appear as false source - pixels , and if it happens that only 2000 pixels are measured as source - pixels , then half the detections are spurious ! if many more true source - pixels are present , of course , this threshold may be quite adequate .
the third criterion defines a more ideal threshold .
such a threshold allows one to specify _ a priori _ the maximum number of false discoveries , on average , as a fraction of the total number of discoveries .
such a method should be independent of the source distribution ( i.e. , it will adapt the threshold depending on the number and brightness of the sources ) .
the false discovery rate ( fdr ) method @xcite does precisely this , selecting a threshold which controls the fraction of false detections .
we have implemented the fdr technique in a task for detecting and measuring sources in images made with radio telescopes .
radio images were chosen for the current analysis for several reasons , including previous experience at coding of radio source detection tasks , but also since the conservative nature of constructing many radio source catalogues allows the value of this method to be emphasised .
traditionally radio source catalogues are constructed in a fashion aimed at minimising spurious sources , accomplished by selecting a very conservative threshold , which is usually @xmath4 or even @xmath5 .
this is partly driven by the difficulty of completely removing residual image artifacts such as sidelobes of bright sources , even after applying very sophisticated image reconstruction methods , and the desire to avoid classifying these as sources .
many of the issues surrounding radio source detection are described by @xcite .
while such a conservative approach may minimise false detections , it has the drawback of not detecting large numbers of real , fainter , sources important in many studies of the sub - mjy and @xmath6jy populations , for example .
an fdr defined threshold may allow many more sources to be included in a catalogue while providing a quantitative constraint to be placed on the fraction of false detections .
table [ extasks ] briefly describes the algorithms used in some common radio source detection and measurement tasks . there is much similarity in the selection of tasks available within the two primary image analysis packages , aips ( astronomical image processing system ) and miriad ( multichannel image reconstruction , image analysis and display ) , as alluded to in the summaries given in this table , although the specifics of each such task certainly differ to a greater or lesser extent .
in addition to these packages , aips++ also has a image measurement task , _ imagefitter _ , similar in concept to the imfit tasks in aips and miriad .
a stand - alone task , sextractor @xcite , is used extensively in image analysis for detecting and measuring objects , primarily in images made at optical wavelengths .
given the flexibility of the sextractor code provided by the many user - definable parameters , it is possible to use this task also for detecting objects in radio images . in the analysis below
we compare the effectiveness of sfind 2.0 , which runs under the miriad package , with that of imsad ( also a miriad task , and which employs essentially the same algorithm as sad and vsad in aips ) and sextractor .
section [ sfind2 ] describes the operation of the sfind 2.0 task and the implementation of the fdr method .
section [ tests ] describes the monte - carlo construction of artificial images on which to compare sfind 2.0 , imsad and sextractor , the tests used in the comparison and their results .
subsequent implementation using a portion of a real radio image is also presented .
section [ disc ] discusses the relative merits of each task , the effectiveness of the fdr method , and raises some issues regarding the validity of the null hypothesis ( i.e. the background level ) and the correlation between neighbouring pixels .
section [ conc ] presents our conclusions along with the strong suggestion that implementing fdr as a threshold defining method in other existing source detection tasks is worthwhile .
lll imfit / jmfit & imfit & fits multiple gaussians to all pixels in a defined area .
+ sad / vsad / happy & imsad & defines ` islands ' encompassing pixels above a user + & & defined threshold , and fits multiple gaussians within these areas .
+ & sfind 2.0 & defines threshold using fdr to determine pixels belonging + & & to sources , and fits those by a gaussian .
here we describe the algorithm used by sfind 2.0 , which implements fdr for identifying a detection threshold .
we include the version number of the task simply to differentiate it from the earlier version of sfind , also implemented under miriad , as the source detection algorithm is significantly different .
subsequent revisions of sfind will continue to use the fdr thresholding method .
the elliptical gaussian fitting routine used to measure identified sources has not changed , however , and is the same as that used in imfit and imsad .
an example of the use of the earlier version of sfind can be found in the source detection discussion of @xcite .
the first step performed is to normalise the image .
a gaussian is fit to the pixel histogram in regions of a user - specified size to establish the mean and standard deviation , @xmath7 , for each region .
then for each region of the image , the mean is subtracted and the result is divided by @xmath7 .
ideally this leaves an image with uniform noise characteristics , defined by a gaussian with zero mean and unit standard deviation . in practice
the finite size of the regions used may result in some non - uniformity , although a judicious choice of size for these regions should minimise any such effect .
we note that radio interferometer images often contain image artifacts such as residual sidelobes arising from the image - processing , sampling effects , and so on . with adequate sampling
these effects should be statistically random with zero mean , and simply add to the overall image noise . next the fdr threshold is calculated for the whole image .
the null hypothesis is taken to be that each pixel is drawn from a gaussian distribution with zero mean and unit standard deviation .
this corresponds to the ` background pixels ' . in the absence of real sources ,
each pixel has a probability @xmath8 , ( which varies with its normalised intensity ) , of being drawn from such a distribution . in images known to contain real sources , a low @xmath8-value for a pixel ( calculated under the assumption that no sources are present )
is often used as an indicator that it is a ` source - pixel . ' the @xmath8-values for all @xmath9 pixels in the image are calculated and ordered .
the threshold is then defined by plotting the ordered @xmath8-values as a function of @xmath10 ( where @xmath9 is the total number of pixels and @xmath11 is the index , from 1 to n ) and finding the @xmath8-value , @xmath12 say , corresponding to the last point of intersection between these and a line of slope @xmath13 .
here @xmath14 is the maximum fraction of falsely detected source - pixels to allow , on average ( over multiple possible instances of the noise ) , and @xmath15 if the statistical tests are fully independent ( the pixels are uncorrelated ) .
if the tests are dependent ( the pixels are correlated ) then @xmath16 since most radio images ( and indeed astronomical images in general ) show some degree of correlation between pixels , but tend not to be _ fully _ correlated , i.e. the intensity of a given pixel is not influenced by that of _ every _ other pixel , we have chosen to take an intermediate estimate for @xmath17 reflecting the level of correlation present in the image .
this is related to the synthesised beam size , or point - spread function ( psf ) .
if @xmath18 is the ( integer ) number of pixels representing the psf we define @xmath19 .
this will be discussed further in section [ cnchoice ] .
a diagrammatic example of the threshold calculation is shown in figure [ fdreg ] .
it becomes obvious from this figure that increasing or decreasing the value chosen for @xmath14 corresponds to increasing or decreasing the resulting @xmath8-value threshold , @xmath12 , and the number of pixels thus retained as ` source - pixels ' .
the fdr formalism ensures that the average fraction of false ` source - pixels ' will never exceed @xmath14 .
as described in @xcite , this explanation of implementing fdr does not explain or justify the validity of fdr .
the reader is referred to @xcite for a heuristic justification , and to @xcite and @xcite for the detailed statistical proof .
finally , once the fdr threshold is defined , the pixels with @xmath20 , corresponding to ` source - pixels ' , can be analysed .
each of the source - pixels are investigated in turn as follows .
a hill - climbing procedure starts from the source - pixel and finds the nearest local maximum from among the contiguous source - pixels . from this local peak ,
a collection of contiguous , monotonically decreasing pixels are selected to represent the source . at this point , it is possible to either ( 1 ) use all of the pixels around the peak which satisfy these criteria , or ( 2 ) use only those which are themselves above the fdr threshold .
the latter is the default operation of sfind 2.0 , but the user can specify an option for choosing the former method as well . the former method , which allows pixels below the fdr threshold to be included in a source measurement , may be desirable for obtaining more reasonable source parameters for sources close to the threshold . in either case , the resulting collection of source - pixels is fit in a least - squares fashion by a two - dimensional elliptical gaussian . if the fitting procedure does not converge , the source is rejected .
this is typically the case when the potential source consists of too few pixels to well - constrain the fit .
it is likely that most such rejections will be due to noise - spikes , which typically contain a small number of pixels , although some may be due to real sources lying just below the threshold such that only a few true source - pixels appear above it .
if the fit is successful the source is characterised by the fitted gaussian parameters , and the pixels used in this process are flagged as already ` belonging ' to a source to prevent them from being investigated again in later iterations of this step .
on completion , a source catalogue is written by the task , and images showing ( 1 ) the pixels above the fdr threshold , and ( 2 ) the normalised image , may optionally be produced .
to compare the effectiveness of sfind 2.0 , imsad and sextractor , one hundred artificial images @xmath21 pixels in size were generated .
each of these contained a different instance of random gaussian noise and the same catalogue of 72 point sources with known properties ( position and intensity ) .
the intensity distribution of the sources spans a little more than 2 orders of magnitude , ranging from somewhat fainter than the noise level to well above it .
many fewer bright sources were assigned than faint sources , in order to produce a realistic distribution of source intensities .
the artificial images were convolved with a gaussian to mimic the effects of a radio telescope psf .
the test images have @xmath22 pixels , and the gaussian chosen to represent the psf has fwhma of @xmath23 with a position angle of @xmath24 .
the ( convolved ) artificial sources in the absence of noise can be seen in figure [ artimages ] , along with one of the test images . on each image
, sfind 2.0 was run with @xmath25 , @xmath26 and @xmath27 , and the resulting lists of detected sources compared with the input catalogue . by way of an example
, the sources detected in a single test of sfind 2.0 are shown in figure [ artannimages ] , which indicates by an ellipse the location , size and position angle of each detected source .
this example demonstrates the difficulty of detecting faint sources , and the ability of noise to mimic the characteristics of faint sources .
@xcite examined the simplest possible scenario , source - pixels placed on a regular grid in the presence of uncorrelated gaussian noise . here
we investigate a much more realistic situation . the ` source - pixels ' now lie in contiguous groups comprising ` real ' sources in the sense that the whole image ( background and sources ) has been convolved with a psf .
the number of pixels in each source above a certain threshold will vary depending on the intensity of the source .
we now confirm that the fdr method works consistently on these realistic images . to verify the reliability of the fdr defined threshold ,
the number of pixels detected above the fdr threshold in each test were recorded along with the number which were unassociated with any true source .
the distribution of this fraction of false fdr pixels should never exceed the value specified for @xmath14 , and this can be seen in the histogram in figure [ fdrhist ] .
this figure also shows how the distribution of falsely detected pixels changes with the form chosen for @xmath17 , emphasising that the assumption of complete independence of the pixels is not justified ( as expected ) , but neither is the image fully correlated , evidenced by the conservative level of false detections seen under this assumption .
our choice for the form of @xmath17 , which is not in fact a result of the rigorous statistical proof , appears to be a feasible and reliable intermediate for such ` partially correlated ' images .
the fdr formalism ensures that the average fraction of falsely detected _ pixels _ will be less than @xmath14 .
the connection between numbers of pixels and numbers of sources is complex , however , and the same criterion can not be said to be true for the fraction of falsely detected _
sources_. the number of source - pixels per source will vary according to both instrumental effects , such as the sampling and the psf , as well as intrinsic source sizes compared to the instrumental resolution and the source brightnesses compared to the noise level in an image .
even if all sources are point - like , and hence should appear in the image as a psf , the number of source - pixels above a given threshold for a given source depends on its brightness , and the number of source - pixels per source would not be expected to be constant . to investigate the effect of this complex relation ,
we explore empirically the results of applying fdr thresholding to our simulated images .
the fraction of falsely detected _ sources _ in each image , as well as the fraction of true sources not detected , are shown in figure [ comparison ] as distributions for each tested value of @xmath14 . by construction ,
a number of the artificial sources have intensities comparable to or lower than the noise level in the images , so not all sources will be able to be recovered in every image .
this is reflected in the fact that somewhat more than @xmath28 of sources are missed ( by all tasks tested ) even with very liberal thresholds .
the histograms in figure [ comparison](a ) show that for @xmath29 , where up to @xmath30 of pixels could be expected to be false , the fraction of false sources is not much more .
the result for @xmath31 , is also quite reasonable , although for @xmath25 the outliers are further still , relatively speaking , from the expected fraction .
while the strict constraint applicable to the fraction of false source - pixels no longer holds for false sources , it still seems to be quite a good estimator .
for the case where only the peak pixel is required to be above the fdr threshold , figure [ comparison](c ) , the fractions of falsely detected sources are not so strongly constrained . for @xmath29 , the fraction may be almost twice that expected . in both cases , although with greater reliability in the former , this allows the fdr method to provide an estimate of the fraction of false sources to expect . even though the constraint may not be rigorous , and
clearly the estimate will be much more reliable in the case where all source - pixels are required to be above the fdr threshold , the fdr method allows a realistic _ a priori _ estimate of the fraction of false detections to be made .
this feature is not possible with the simple assumption of , say , a @xmath4 threshold . to test imsad and sextractor in the same fashion as sfind 2.0
, a choice of threshold value as a multiple of the image noise level ( @xmath7 ) was required . simply selecting a canonical value of @xmath0 , @xmath4 or @xmath5 ,
for example , would complicate the comparison between these tasks and sfind 2.0 , as this would be testing not only different source measurement routines but also potentially different thresholds .
the values of @xmath14 selected for testing sfind 2.0 result in threshold levels which correspond approximately , ( since the noise level varies minimally from image to image ) , to @xmath32 , @xmath33 and @xmath34 .
a @xmath4 threshold in these simulations would correspond to a value of @xmath35 . in a brief aside , it should be emphasised that this particular correspondence between a choice of @xmath14 and a particular @xmath7-threshold is only valid for the noise and source characteristics of the images used in the present simulations . for images with different noise levels or different source intensity distributions
, any particular value of @xmath14 will correspond to some different multiple of the local noise level .
the primary advantage of specifying an fdr @xmath14 value over choosing a @xmath4 threshold , say , is that the fdr threshold is adaptive .
the fdr threshold will assume a different value depending on the source intensity distribution relative to the background .
this point is made very strongly in @xcite in the diagrams of their figure 4 .
as a specific example in the context of the current simulations , we investigated additional simulated images containing the same noise as in the current simulations but containing sources having very different intensity distributions .
we chose one intensity distribution such that every source was 10 times brighter than in the current simulations and one such that every source was 10 times fainter . in the brighter case , the fdr threshold for @xmath25 , ensuring that on average no more than @xmath36 of source - pixels would be falsely detected , corresponded not to @xmath32 , but to about @xmath37 .
the reason here is that as the source distribution becomes brighter , many more pixels will have low @xmath8-values . to retain the constant fraction of false pixels more background pixels
must be included , so the threshold becomes lower . in the fainter case ,
where the artificial sources are very close in intensity to the noise level , the same fdr threshold corresponds to about @xmath38 .
of course in this case very few sources are detected , for obvious reasons , but the same constraint on the fraction of falsely detected pixels applies . here
fewer pixels will have low @xmath8-values , thus fewer background pixels are allowed , maintaining a constant fraction of false pixels , and the threshold increases . in the brighter case the simple assumption of , say , a @xmath39 threshold would give a lower rate of false pixels , while in the fainter case it would give a higher fraction .
the importance of these examples is to emphasise that fdr provides a consistent constraint on the fraction of false detections in an adaptive way , governed by the source intensity distribution relative to the background , which can not be reproduced by the simple assumption of , for example , a @xmath39 threshold .
while the source distributions in most astronomical images typically lie between the two extremes presented for this illustration , the adaptability of the fdr thresholding method still presents itself as an important tool .
returning now to the comparison between sfind 2.0 , imsad and sextractor , the imsad and sextractor thresholds were set to correspond to those derived from the @xmath14 values used in sfind 2.0 .
the distributions of falsely detected and missed sources were similarly calculated .
these are shown in figures [ comparison](e ) to [ comparison](h ) .
one of the features of sextractor is the ability to set a minimum number of contiguous pixels required before a source is considered to be real , and obviously the number of detected sources varies strongly with this parameter . after some experimentation we set this parameter to 7 pixels , as this resulted in a distribution of false detections most similar to that seen with sfind 2.0 , for the case where only pixels above the fdr threshold are used in source measurements .
values larger than 7 reduced the number of false detections at the expense of missing more true sources , and vice - versa . from this comparison
, sfind 2.0 appears to miss somewhat fewer of the true sources than sextractor when sextractor is constrained to the same level of false detections .
this is also true if sextractor is constrained to a similar distribution of false detections as obtained by sfind 2.0 for the case where only the peak pixel is required to be above the fdr threshold ( corresponding to 4 contiguous pixels ) . in both cases , allowing sextractor to detect more true sources by lowering the minimum pixel criterion introduces larger numbers of false detections .
sfind 2.0 also performs favourably compared to imsad .
while sfind 2.0 seems to miss a few percent more real sources than imsad , imsad seems to detect many more false sources than sfind 2.0 in either of its source - measurement modes . of primary importance in source measurement
is the reliability of the source parameters measured .
figure [ measvstrue ] shows one example from the one hundred tests comparing the true intensities and positions of the artificial sources with those measured by sfind 2.0 .
similar results are obtained with imsad , which uses the same gaussian fitting routine .
as expected , the measured values of intensity and position become less reliable as the source intensity becomes closer to the noise , although they are still not unreasonable . a comprehensive analysis of gaussian fitting in astronomical applications has been presented by @xcite , and the results of the gaussian fitting performed by sfind 2.0 ( and imsad ) are consistent with the errors expected . additionally , the assumption that a source is point - like , or only slightly extended , and thus well fit by a two - dimensional elliptical gaussian , is clearly not always true .
complex sources in radio images , as in any astronomical image , present difficulties for simple source detection algorithms such as the ones investigated here .
it is not the aim of the current analysis to address these problems , except to mention that the parameters of such sources measured under the point - source assumption will suffer from much larger errors than indicated by the results of the gaussian fitting . as a final test
, sfind 2.0 was used to identify sources in a real radio image , a small portion of the _ phoenix deep survey _ @xcite .
this image contains sources with extended and complex morphologies as well as point sources .
figure [ realfdr ] shows the results , with sfind 2.0 reliably identifying point source and extended objects as well as the components of various blended sources and complex objects .
the main aims of this analysis have been to ( 1 ) investigate the implementation of fdr thresholding to an astronomical source detection task , and ( 2 ) compare the rates of missed and falsely detected sources between this task and others commonly used .
implementation of the fdr thresholding method is very straightforward , ( evidenced by the seven step idl example of @xcite , in their appendix b ) .
the fdr method performs as expected in providing a statistically reliable estimate of the fraction of falsely detected _
pixels_. performing source detection on a set of pixels introduces the transformation of pixel groups into sources .
this ultimately results in the strong constraint on the false fraction of fdr - selected pixels becoming a less rigorous , but still useful and reliable estimate of the fraction of false sources .
as already mentioned , this is still a more quantitative statement than can be made of the rate of false sources in the absence of the fdr method .
it is possible that rigorous quantitative constraints on the fraction of false source detections may be obtained empirically for individual images or surveys . by performing monte - carlo source detection simulations with artificial images having noise properties similar to the ones under investigation ,
the trend of falsely detected sources with @xmath14 may be able to be reliably characterised .
this has not been examined in the current analysis , but will be included in subsequent work with sfind 2.0 . while constraints on the fraction of falsely detected sources may be possible , neither fdr nor any other thresholding method provides constraints on the numbers of true sources remaining undetected .
a study is also ongoing into whether a more sophisticated fdr thresholding method for defining a source may be feasible .
this would involve examining the combined size and brightness properties of groups of contiguous pixels to define a new @xmath8-value .
this would represent the likelihood that such a collection of pixels comes from a ` background distribution ' or null - hypothesis , corresponding to the properties exhibited by noise in various types of astronomical images .
using this new @xmath8-value an fdr threshold , now in the size - brightness parameter space , could be applied for defining a source catalogue .
clearly much care will need to be taken to avoid discriminating against true sources which may lie in certain regions of the size - brightness plane , such as low surface - brightness galaxies .
the assumption in many existing source - detection algorithms that sources are point - like already suffers from such discrimination , though , so even if such bias is unavoidable , some progress may still be achievable .
the form assumed for @xmath17 in this analysis is not in fact a rigorous result of the formal fdr proof .
instead it is a ` compromise ' estimate that seems mathematically reasonable , and gives reliable results in practice . to be strictly conservative
, the form of @xmath17 given by equation [ cn ] should be adopted to ensure that the fraction of falsely detected ` source - pixels ' is strictly less than @xmath14 .
this rigorous treatment , however , is dependent on the number of pixels present in the image .
now consider analysis of a sub - region within an image . as the size of this sub - region
is changed , the number of pixels being considered similarly changes , and this will have the effect of changing the threshold level , and the resulting source catalogue .
this , perhaps non - intuitive , aspect of the fdr formalism is the adaptive mechanism which allows it to be rigorous in constraining the fraction of false detections .
there are additional complicating factors which must be taken into account when performing source detection .
the null hypothesis assumed for the fdr method ( and indeed for all the source detection algorithms ) is that the background pixels have intensities drawn from a gaussian distribution ( or other well - characterised statistical distribution such as a poissonian ) .
this is not strictly true for radio images , where residual image processing artifacts may affect the noise properties , albeit at a low level . in all cases
, such deviations will result in a larger fraction of false pixels than expected , some of which may be clumped in a fashion sufficient to mimic , and be measured as , sources , thus increasing the fraction of falsely detected sources .
this comment is simply to serve as reminder to use caution when analysing images with complex noise properties .
we have implemented the fdr method in an astronomical source detection task , sfind 2.0 , and compared it with two other tasks for detecting and measuring sources in radio telescope images .
sfind 2.0 compares favourably to both in the fractions of falsely detected sources and undetected true sources .
the fdr method reliably selects a threshold which constrains the fraction of false pixels with respect to the total number of ` source - pixels ' in realistic images .
the fraction of falsely detected sources is not so strongly constrained , although quantitative estimates of this fraction are still reasonable .
more investigation of the relationship between ` source - pixels ' and ` sources ' is warranted to determine if a more rigorous constraint can be established . with the ability to quantify the fraction of false detections provided by the fdr method
, we strongly recommend that it is worthwhile implementing as a threshold defining method in existing source detection tasks .
we would like to thank the referee for several suggestions which have improved this paper .
amh acknowledges support provided by nasa through hubble fellowship grant hst - hf-01140.01-a awarded by the space telescope science institute ( stsci ) .
amh and ajc acknowledge support provided by nasa through grant numbers go-07871.02 - 96a and nra-98 - 03-ltsa-039 from stsci , and aisr grant nag-5 - 9399 .
stsci is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
benjamini , y. , hochberg , y. 1995 , j. r. stat .
b , 57 , 289 benjamini , y. , yekutieli , d. 2001 , ann .
statist . , ( in press ) bertin , e. , arnouts , s. 1996 , , 117 , 393 condon , j. j. 1997 , , 109 , 166 hopkins , a. , afonso , j. , cram , l. , mobasher , b. 1999 , , 519 , l59 miller , c. j. , genovese , c. , nichol , r. , wasserman , l. , connolly , a. , reichart , d. , hopkins , a. , schneider , j. , moore , a. 2001 , aj ( accepted ) ( astro - ph/0107034 ) white , r. l. , becker , r. h. , helfand , d. j. , gregg , m. d. , 1997 , , 475 , 479 | the false discovery rate ( fdr ) method has recently been described by @xcite , along with several examples of astrophysical applications .
fdr is a new statistical procedure due to @xcite for controlling the fraction of false positives when performing multiple hypothesis testing .
the importance of this method to source detection algorithms is immediately clear . to explore the possibilities offered
we have developed a new task for performing source detection in radio - telescope images , sfind 2.0 , which implements fdr .
we compare sfind 2.0 with two other source detection and measurement tasks , imsad and sextractor , and comment on several issues arising from the nature of the correlation between nearby pixels and the necessary assumption of the null hypothesis .
the strong suggestion is made that implementing fdr as a threshold defining method in other existing source - detection tasks is easy and worthwhile .
we show that the constraint on the fraction of false detections as specified by fdr holds true even for highly correlated and realistic images . for the detection of true sources , which are complex combinations of source - pixels
, this constraint appears to be somewhat less strict .
it is still reliable enough , however , for a priori estimates of the fraction of false source detections to be robust and realistic . |
since the discovery of graphene in 2004 , @xcite its application to nanoelectronic devices has been regarded as one of the most significant tasks in the field of nanotechnology sciences . however , the absence of a band gap in graphene restricts its utilization as a channel material in field - effect transistors .
so far , many efforts have been devoted to open a band gap in graphene .
@xcite specifically , the chemical modification approach with adatoms @xcite has been intensively employed for the gap opening of graphene .
for instance , fluorinated graphene ( hereafter termed f - graphene ) has been attracted much attention as a promising material for nanoelectronic device applications because of its tunable band gap with respect to the amount of fluorination and the fluorination patterns .
@xcite in order to design the structure of f - graphene in atomic scale , it is necessary to understand the diffusion and adsorption of the f atom on graphene .
it was found that the f atom prefers to adsorb on the on - top site of graphene with a single f@xmath0c bond .
@xcite recently , sadeghi @xmath1 @xmath2 .
@xcite reported a first - principles density - functional theory ( dft ) study of the f diffusion both on monolayer ( ml ) graphene and between the layers of bilayer ( bl ) graphene , showing a drastic difference between the two cases : i.e. , the mobility of the f atom within bl graphene increases by about an order of magnitude compared to that on ml graphene .
although sadeghi @xmath1 @xmath2 .
@xcite predicted a facilitated diffusion process of the f atom within bl graphene by hopping from one layer to the other layer , the microscopic underlying mechanism of the drastic difference of f diffusion between ml and bl graphenes is still lacking .
as another chemical modification of graphene , oxidized graphene ( hereafter termed o - graphene ) has been widely adopted because of its low cost and manufacturing feasibility for mass production of graphene or graphene nano - platelets .
@xcite in particular , o - graphene which can be generated from graphite oxide contains various o - containing functional groups , thereby offering a broad applicability to electronic devices .
it is noted that the o atom adsorbs on the bridge site of graphene with the c@xmath0o@xmath0c bond , @xcite which is different from the on - top adsorption site of hydrogen , halogen , and f atoms .
@xcite this different binding of o with graphene may cause a distinctive diffusion behavior compared to the above - mentioned f diffusion . indeed , earlier dft calculations @xcite reported that the energy barrier for the o diffusion on ml graphene ranges between 0.73 and 0.81 ev , which is relatively larger than that ( ranging between 0.13 and 0.29 ev ) for the f diffusion on ml graphene .
@xcite however , a comparative study of the o diffusions on ml graphene and within bl graphene is yet to be explored . in this paper , we perform a first - principles theoretical study of the diffusion behaviors of the o and f adatoms both on ml graphene and between the layers of bl graphene by using a van der waals ( vdw ) energy - corrected dft calculation . @xcite our calculated potential energy surfaces ( pess ) for such ml and bl o(f)-graphene systems show that ( i ) the diffusion barrier @xmath3 for o ( f ) on ml graphene is 0.81 ( 0.30 ) ev , and ( ii ) @xmath3 for o ( f ) within bl graphene increases ( decreases ) to 0.85 ( 0.18 ) ev .
such contrasting variations of the o and f diffusions between ml and bl graphenes can be associated with their different bonding natures : i.e. , the o adatom has a strongly covalent c@xmath0o@xmath0c bonding on the bridge site of the c@xmath0c bond , while the f atom has a weakly ionic f@xmath0c bonding on top of a c atom .
consequently , the o atom diffuses on one graphene layer with a slight interference of the other layer , while the f atom easily diffuses by hopping between two graphene layers with accepting more electron charges from the two layers .
our findings shed more light on understanding the contrasting diffusion behaviors of the o and f atoms on ml graphene and within bl graphene .
our first - principles dft calculations were performed using an accurate all - electron , full - potential scheme with numeric atom - centered orbital basis functions , as implemented in the fritz - haber - institute @xmath4 @xmath5 molecular simulations ( fhi - aims ) package .
@xcite all the calculations were carried out with `` tight '' computational settings and accurate tier-2 basis sets . the generalized gradient approximation ( gga ) functional of perdew - burke - ernzerhof ( pbe ) @xcite
was employed for the exchange - correlation energy . to include the effects of vdw interactions
, we used the pbe+vdw scheme developed by tkatchenko and scheffler , @xcite which has been demonstrated to be accurate in layered crystals as well as various adsorbates on surfaces .
@xcite the o - graphene or f - graphene system was simulated by containing one o or f atom within a @xmath6@xmath7@xmath6 unit cell , thereby minimizing the spurious interactions between o or f adatoms in the periodic unit cells .
the * k*-space integration was done with the 12@xmath712 uniform meshes in the two - dimensional brillouin zone .
all atoms were allowed to fully relax until all the residual force components on each atom were less than 0.01 ev / . using the pbe ( pbe+vdw ) calculation ,
the lattice constant of ml graphene and the interlayer distance of bl graphene were obtained as 2.47 ( 2.46 ) and 4.06 ( 3.24 ) , respectively .
.calculated pbe binding energies ( @xmath8 ) of the o - graphene and f - graphene systems , together with the pbe+vdw results in parentheses^_a_^ [ cols="^,^,^,<,^ " , ] + ^_a_^ for the definitions of @xmath9 and @xmath10 , see figure 1 . due to the interlayer vdw interactions in bl graphene
, the geometries and binding energies of the bl o - graphene and bl f - graphene systems are expected to be different between the pbe and pbe+vdw results . to examine these differecnes
, we perform the pbe+vdw calculation for various adsorption sites .
figure 1a - d show the most stable structures of ml o - graphene , ml f - graphene , bl o - graphene , and bl f - graphene , respectively , obtained using the pbe+vdw calculation .
the calculated binding energies of these o - graphene and f - graphene systems are also listed in table 1 .
we find that the most stable b site for ml or bl o - graphene and t ( th ) site for ml ( bl ) f - garphene are invariant between the pbe and pbe+vdw calculations .
however , the inclusion of vdw interactions in ml o - graphene and ml f - graphene slightly increases @xmath8 by 0.06 and 0.07 ev , respectively . on the other hand , such vdw effects in bl o - graphene and bl f - graphene decrease @xmath8 as large as 0.49 and 0.84 ev , respectively ( see table 1 ) .
this huge vdw - induced reduction of binding energy in bl o - graphene and bl f - graphene is caused by the fact that the inclusion of vdw interactions produces the large lattice deformation of bl graphene .
as shown in table 2 , for ml o - graphene and ml f - graphene , the pbe and pbe+vdw calculations hardly change the bond lengths @xmath11 and @xmath12 .
however , the pbe+vdw geometry of bl o(f)-graphene shows a dramatic decrease in the interlayer distance @xmath10 by 0.74 ( 1.20 ) compared to the corresponding values obtained using pbe ( see table 2 ) .
consequently , the pbe+vdw calculations for bl o - graphene and bl f - graphene give rise to a significant structural deformation of the upper graphene layer around o and f adatoms [ see figure 1c , d ] , thereby leading to a decrease in their binding energies compared to those computed from pbe . in these respects , we can say that vdw interactions play an important role in determining the geometries and binding energies of bl o - graphene and bl f - graphene .
next , we investigate the diffusions of o and f atoms on ml graphene using the pbe+vdw scheme . to obtain the minimum - energy diffusion pathway , we calculate the pess for ml o - graphene and ml f - graphene by optimizing the structure at the uniformly separated adsorption sites within the graphene unit cell .
figure 2a , b ) show the calculated pess for ml o - graphene and ml f - graphene , respectively .
we find that , for ml o - graphene , the o adatom has a diffusion path from the b site to the neighboring b site through the t@xmath13 ( close to t ) site with @xmath3 = 0.81 ev , in good agreement with a previous dft result @xcite where the transition state is slightly deviated from the t site with @xmath3 = 0.73 ev . here , the present angles of the c@xmath0c@xmath0o bonds at the t@xmath13 site are 101.5@xmath14 , 102.5@xmath14 , and 106.3@xmath14 ( see figure s1 of the supporting information ) , close to those ( 99.9@xmath14 , 101.0@xmath14 , and 108.1@xmath14 ) of a previous dft calculation .
@xcite meanwhile , for ml f - graphene , the f adatom diffuses from the t site to the neighboring t site through the b site with @xmath3 = 0.30 ev , in good agreement with that ( 0.29 ev ) of previous dft calculations .
@xcite thus , @xmath3 for the o diffusion of ml o - graphene is @xmath152.5 times larger than that for the f diffusion of ml f - graphene , indicating that the f atom can diffuse much faster than the o atom . to understand the contrasting diffusion behaviors of the o and f atoms between ml o - graphene and ml f - graphene
, we examine the bonding natures of the o and f adatoms on ml graphene .
figure 3a , b display the charge density difference calculated from ml o - graphene and ml f - graphene , respectively , which is defined as @xmath16 here , @xmath17 is the charge density of o - graphene or f - graphene and @xmath18+@xmath19 is the superposition of the charge densities of the separated systems , i.e. , isolated o or f atom and clean graphene . in figure 3a , @xmath20 of ml o - graphene represents a covalent bonding character for the c@xmath0o@xmath0c bond with @xmath11 = 1.46 . along the diffusion path b@xmath21t@xmath13@xmath21b
, one o@xmath0c bond is broken while the other o@xmath0c bond is shortened to be @xmath11 = 1.41 at the transition site . meanwhile , as shown in figure 3b , @xmath20 of ml f - graphene represents an ionic bonding character for the f@xmath0c bond with @xmath12 = 1.54 . note that f has a relatively larger electronegativity of 3.98 compared to that ( 2.55 ) of c. @xcite
this ionic f@xmath0c bond character for an isolated f adatom on graphene is consistent with a recent combined dft and x - ray photoelectron spectroscopy study .
@xcite interestingly , the f atom at the transition state ( equivalently at the b site ) along the diffusion path t@xmath21b@xmath21 t is far away from two neighboring c atoms with @xmath12 = 2.29 .
this longer separation between the f and c atoms at the transition state is drastically different from a decrease of @xmath11 at the transition state of ml o - graphene , reflecting the different bonding natures between ml o - graphene and ml f - graphene .
as shown in the right panels of figure 3a , b , @xmath20 for the transition states of ml o - graphene and ml f - graphene clearly show the covalent and ionic bonding characters , respectively . we continue to investigate the diffusion of the o and f atoms between the layers of bl graphene using the pbe+vdw scheme .
figure 2c , d display the calculated pess for bl o - graphene and bl f - graphene , respectively .
we find that the o atom in bl o - graphene is covalently bound to one graphene layer during the diffusion process .
compared with the result for ml o - graphene , the o diffusion in bl o - graphene is somewhat influenced by the presence of the other layer . for bl o - graphene , there are the two different diffusion paths i and ii with @xmath3 = 0.85 and 0.86 ev , respectively , which is slightly larger than ( 0.81 ev ) in ml o - graphene . here , the path i
( ii ) has the transition state passing the tt ( th ) site [ see figure 2c ] .
note that the hopping of the o atom from one layer to the other layer has a high energy barrier of @xmath221.3 ev .
by contrast , the pes for bl f - graphene shows that the f atom easily diffuses by hopping from one layer to the other layer : see figure 2d .
we find that the diffusion path along th@xmath21b@xmath21tt@xmath21b@xmath21ht involves double barriers with @xmath3 = 0.18 ev [ see figure 2f ] .
this value in bl f - graphene is much reduced compared to that ( 0.30 ev ) in ml f - graphene , indicating that the diffusion of f is facilitated within bl graphene .
these contrasting diffusion behaviors of o and f between bl o - graphene and bl f - graphene can be also traced to their different bonding natures , as discussed below . of ( a )
ml o - graphene , ( b ) ml f - graphene , ( c ) bl o - graphene , and ( d ) bl f - graphene .
the left ( right ) panels represent the results at the equilibrium ( transition ) state . in the contour plots ,
the first solid ( dashed ) lines are drawn at 5.0@xmath710@xmath23 @xmath24/@xmath25 ( @xmath05.0@xmath710@xmath23 @xmath24/@xmath25 ) and the contour spacing is 3.3@xmath710@xmath23 @xmath24/@xmath25 ( @xmath03.3@xmath710@xmath23 @xmath24/@xmath25 ) .
the solid and dashed lines indicate accumulated and depleted electrons , respectively .
the numbers represent the bond lengths ( in ) . ]
as shown in figure 3a , c , the values of @xmath11 at the equilibrium and transition states of bl o - graphene are similar to those in ml o - graphene , implying that o adsorption on one layer is little affected by the other layer . such small changes of the geometries of ml o - graphene and bl o - graphene is consistent with their similarity of @xmath20 : see figure 3a , c . these results in turn give the slight change in @xmath3 between ml o - graphene and bl o - graphene . on the other hand , @xmath12 at the transition state of bl f - graphene becomes 2.16 or 2.17 , much shorter than that ( 2.29 ) in ml f - graphene : see the right panels of figure 3b , d .
this decrease of @xmath12 in bl f - graphene can be associated with more ionic character due to its enhanced charge transfer : i.e. , the f adatom in bl f - graphene accepts more electrons simultaneously from the two layers compared to ml f - graphene , as shown in figure 3b , d .
such more charge accumulation around the f atom is likely to increase the binding energy due to an ionic bonding to its neighboring c atoms , leading to a decrease of @xmath3 in bl f - graphene .
it is finally worth noting that the adsorption of alkali and transition metal elements on graphene occurs at the h site , forming strong chemical bonds with surrounding c atoms .
@xcite such strong bindings on graphene can not be expected to an easy diffusion .
indeed , some transition metal elements favoring adsorption on the h site showed relatively larger diffusion barriers compared to other elements favoring adsorption on the b or t site @xcite .
it is thus likely that the initial adsorption structure of adatoms on graphene plays a crucial role in determining their diffusion behaviors .
we have presented a first - principles theoretical study of the diffusion behaviors of the o and f adatoms both on ml graphene and between the layers of bl graphene .
our calculated potential energy surfaces for the o - graphene and f - graphene systems showed that the diffusion barrier for the o atom slightly increases from 0.81 ev on graphene to 0.85 ev within bilayer graphene , while that for the f atom significantly decreases from 0.30 ev on graphene to 0.18 ev within bilayer graphene .
it is revealed that such contrasting variations of the o and f diffusions between ml and bl graphenes are attributed to their different bonding natures : i.e. , the o adatom that has a strongly covalent c@xmath0o@xmath0c bonding on the bridge site of the c@xmath0c bond diffuses on one graphene layer with a slight interference of the other layer , while the f adatom that has a weakly ionic f@xmath0c bonding on top of a c atom easily diffuses by hopping between two graphene layers with accepting more electron charges from the two layers .
our findings shed more light on understanding the contrasting diffusion behaviors of the o and f atoms on ml graphene and within bl graphene , which will provide a useful information for the chemical modification of graphene sheets .
* supporting information * + the geometry of the transition state of ml o - graphene is given .
this material is available free of charge via the internet at http://pubs.acs.org .
* corresponding author * + * email : [email protected] + * email : [email protected] + * notes * + the authors declare no competing financial interest .
this work was supported by national research foundation of korea ( nrf ) grant funded by the korean government ( 2015r1a2a2a01003248 ) and kisti supercomputing center through the strategic support program for the supercomputing application research ( ksc-2016-c3 - 0001 ) .
novoselov , k. s. ; geim , a. k. ; morozov , s. v. ; jiang , d. ; zhang , y. ; dubonos , s. v. ; grigorieva , i. v. ; firsov , a. a. electric field effect in atomically thin carbon films .
_ science _ * 2004 * , _ 306 _ , 666 .
giovannetti , g. ; khomyakov , p. a. ; brocks , g. ; kelly , p. j. ; van den brink , j. substrate - induced band gap in graphene on hexagonal boron nitride : _ ab _ initio density functional calculations .
b : condens . matter mater .
phys . _ * 2007 * , _ 76 _ , 073103 .
zanella , i. ; guerini , s. ; fagan , s. b. ; filho , j. m. ; filho , a. g. s. chemical doping - induced gap opening and spin polarization in graphene .
b : condens . matter mater .
phys . _ * 2008 * , _ 77 _ , 073404 .
gmez - navarro , c. ; weitz , r. t. ; bittner , a. m. ; scolari , m. ; mews , a. ; burghard , m. ; kern , k. electronic transport properties of individual chemically reduced graphene oxide sheets .
_ nano lett . _ * 2007 * , _ 7 _ , 3499 .
cheng , s .- h .
; zou , k. ; okino , f. ; gutierrez , h. r. ; gupta , a. ; shen , n. ; eklund , p. c. ; sofo , j. o. ; zhu , j. reversible fluorination of graphene : evidence of a two - dimensional wide bandgap semiconductor .
b : condens . matter mater .
* 2010 * , _ 81 _ , 205435 .
georgakilas , v. ; otyepka , m. ; bourlinos , a. b. ; chandra , v. ; kim , n. ; kemp , k. c. ; hobza , p. ; zboril , r. ; kim , k. s. functionalization of graphene : covalent and non - covalent approaches , derivatives and applications .
rev . _ * 2012 * , _ 112 _ , 6156 .
robinson , j. t. ; burgess , j. s. ; junkermeier , c. e. ; badescu , s. c. ; reinecke , t. l. ; perkins , f. k. ; zalalutdniov , m. k. ; baldwin , j. w. ; culbertson , j. c. ; sheehan , p. e. ; snow , e. s. properties of fluorinated graphene films , _
nano lett .
_ * 2010 * , _ 10 _ , 3001 . blum , v. ; gehrke , r. ; hanke , f. ; havu , p. ; havu , v. ; ren , x. ; reuter , k. ; scheffler , m. @xmath27 @xmath5 molecular simulations with numeric atom - centered orbitals .
commun . _ * 2009 * , _ 180 _ , 2175 .
ruiz , v. g. ; liu , w. ; zojer , e. ; scheffler , m. ; tkatchenko , a. density - functional theory with screened van der waals interactions for the modeling of hybrid inorganic - organic systems .
lett . _ * 2012 * , _ 108 _ , 146103 .
ganesh , p. ; kim , j. ; park , c. ; yoon , m. ; reboredo , f. a. ; kent , p. r. c. binding and diffusion of lithium in graphite : quantum monte carlo benchmarks and validation of van der waals density functional methods .
_ j. chem .
theory comput .
_ * 2014 * , _ 10 _ , 5318 .
yao , f. ; gne , f. ; ta , h. q. ; lee , s. m. ; chae , s. j. ; sheem , k. y. ; cojocaru , c. s. ; xie , s. s. ; lee , y. h. diffusion mechanism of lithium ion through basal plane of layered graphene .
soc . _ * 2012 * , _ 134 _ , 8646 . | the chemical modification of graphene with adatoms is of importance for nanoelectronics applications .
based on first - principles density - functional theory calculations with including van der waals interactions , we present a comparative study of the diffusion characteristics of oxygen ( o ) and fluorine ( f ) atoms both on graphene and between the layers of bilayer graphene . we find that o and f atoms have lower binding energies between the layers of bilayer graphene compared to on graphene .
interestingly , the calculated diffusion barrier for the o atom slightly increases from 0.81 ev on graphene to 0.85 ev within bilayer graphene , while that for the f atom significantly decreases from 0.30 ev on graphene to 0.18 ev within bilayer graphene
. such contrasting behaviors of the o and f diffusions within bilayer graphene can be traced to their different bonding natures : i.e. , the o adatom that has a strongly covalent c@xmath0o@xmath0c bonding on the bridge site of the c@xmath0c bond diffuses on one graphene layer with a slight interference of the other layer , while the f adatom that has a weakly ionic f@xmath0c bonding on top of a c atom easily diffuses by hopping between two graphene layers with accepting more electron charges from the two layers .
the present findings have important implications for understanding the diffusion processes of f and o atoms on graphene and within bilayer graphene . |
collapse models were formulated to describe in a unified framework the behavior of microscopic systems , as accounted for by quantum mechanics , and the emergence of the objective macroscopic world , described by classical mechanics .
after the pioneering works by pearle @xcite , the first consistent collapse model was put forward by ghirardi , rimini and weber ( grw ) @xcite ; see @xcite for more details and a list of references about the historical development of collapse models .
the crucial feature of the grw model is that the wavefunction associated with the state of a physical system undergoes sudden and random localization processes .
the latter do not practically affect microscopic systems , while they become relevant already on very short time scales for macroscopic systems , by virtue of the amplification mechanism .
the localization processes prevent macroscopic systems from being in a superposition of states centered around macroscopically distinct positions .
in addition , since any measurement process consists in an interaction between a microscopic system and a macroscopic measurement apparatus , the localization processes give a dynamical explanation of the collapse of the wavefunction to one of the eigenstates of the measured observable , without the need of introducing an ad - hoc reduction postulate @xcite .
the non - linear and stochastic nature of the reduction postulate in standard quantum mechanics is replaced within collapse models by a modification of the schrdinger equation which includes proper non - linear and stochastic terms .
one of the main advantages of collapse models is that , besides the relevant conceptual differences with respect to the standard theory , they provide experimentally testable predictions which depart from those of quantum mechanics @xcite ; see @xcite for a detailed list of references .
in particular , collapse models call into question the universal nature of the superposition principle , as they set a fundamental limit above which no physical system can exhibit position superpositions , except for a negligibly small time .
the investigation of collapse models thus plays a significant role in the experimental tests on the boundaries between the classical and the quantum description of reality at the mesoscopic and macroscopic scale .
will it be possible to prepare quantum superpositions for more and more complex systems with the future advances of the experimental techniques , or is there an intrinsic limit which will prevent from this , as predicted by collapse models ?
the renewed interest in collapse models and the demand to compare their predictions with actual experimental data coming from very different setups has further motivated the formulation of more realistic models .
the initial idea that the non - linear and stochastic modification of the schrdinger equation represents an intrinsic property of nature , was then superseded by the view that the collapse of the wavefunction is induced by a physical field filling space , which acts as a universal noise .
indeed , the precise definition of such a field needs the formulation of a new fundamental theory going beyond standard quantum mechanics @xcite .
hence , collapse models should be understood as phenomenological models , which capture the main effects of the interaction with the above - mentioned noise .
basic physical principles set some general constraints on the features of the admissible models .
this is the case , e.g. , for the absence of faster - than - light signaling @xcite or the principle of energy conservation .
an important drawback of the grw model is the violation of the energy conservation : the stochastic action of the noise induces larger and larger fluctuations in the momentum space , so that the energy of the system diverges for asymptotic times , although with a rate which is very small @xcite .
this is a common feature of the first collapse models @xcite and it traces back to the absence of any dissipative mechanism within the interaction between the system and the noise @xcite , as also witnessed by the structure of the master equations which can be associated with these models . in this paper , we modify the grw model in order to avoid the divergence in time of the energy , thus making an important step toward the reestablishment of the energy conservation within the model .
we explicitly show that this can be achieved via the introduction of new localization operators , without changing the other defining features of the model .
this provides us with a more realistic collapse model , while we keep the original effectiveness and physical transparency of the grw model .
the convergence of the energy of the system to a finite value allows us to associate the noise with a finite temperature @xmath0 , such that the grw model is recovered in the high temperature limit @xmath1 .
we follow a strategy similar to that exploited for a simplified collapse model @xcite and based on the formal analogy with a lindblad master equation @xcite including dissipation . besides discussing in detail the physical consequences of the extension of the model
, we also express both the original and the generalized grw models in terms of a stochastic differential equation .
this analysis , which relies on the framework of the stochastic differential equations of jump type in hilbert spaces @xcite , fills a gap with respect to collapse models with localization continuous in time @xcite .
the paper is organized as follows . in sec .
[ sec : tgm ] , we briefly recall the main features of the grw model . in sec . [
sec : egm ] , we introduce the dissipative grw model , characterizing the evolution in the hilbert space induced by the new jump operators , in the same spirit as it was done for the original grw model @xcite . in sec .
[ sec : sfotm ] , we show how the extended grw model , and the original one as well , can be formulated in terms of a stochastic differential equation and we study some relevant features of its solutions , focusing on the occurrence of the localization . in sec .
[ sec : me ] , we derive the master equation associated with the model , which is shown to be equivalent to an equation exploited in the description of collisional decoherence @xcite . after studying the solution of the equation , we prove explicitly , in sec .
[ sec : therm ] , that it implies dissipation and an exponential relaxation of the energy to a finite value . in sec .
[ sec : ampl ] , we study the amplification mechanism in the presence of dissipation : we first deal with a macroscopic rigid body and show that its center of mass behaves for all the practical purposes as a classical object , then we discuss the difficulties which are unavoidably encountered when more general situations are considered . finally , the conclusions and final remarks are given in sec .
[ sec : concl ] .
let us start by briefly recalling the main features of the collapse model introduced by ghirardi , rimini and weber @xcite .
the grw model can be formulated in terms of discrete jumps of the wavefunction which represents the state of the system taken into account . for the sake of simplicity
, we consider a particle in one dimension and we neglect spin and other internal degrees of freedom , so that the wavefunction @xmath2 is an element of the hilbert space @xmath3 .
the dynamics of the particle is then characterized through the following assumptions : * at random times the particle experiences a sudden jump described by @xmath4 where @xmath5 is the state immediately before the jump , which occurs at time @xmath6 and position @xmath7 , and @xmath8 is the self - adjoint contractive linear operator defined as @xmath9 with @xmath10 the position operator of the particle and @xmath11 a new parameter of the model . * the overall number of jumps is distributed in time according to a poisson process with rate @xmath12 , which is the second new parameter of the model .
the standard value of the rate is @xmath13 , while a higher rate was proposed more recently @xcite , @xmath14 . * if there is a jump at time @xmath6 , the probability density that it takes place at the position @xmath15 is @xmath16 * in the time interval between two consecutive jumps , the state vector evolves according to the usual schrdinger equation .
as will be shown in the following , see sec.[sec : this ] , this dynamics can be equivalently formulated through a stochastic differential equation .
moreover , the jump operators @xmath17 satisfy the relation @xmath18 which corresponds to the normalization of the probability distribution @xmath19 defined in eq.([eq : py ] ) .
the jump operator @xmath17 describes a localization process around the position @xmath15 and with width @xmath20 .
consider a gaussian wavefunction @xmath21 , @xmath22 where @xmath23 is the mean position , @xmath24 the mean momentum , while @xmath25 determines the position variance @xmath26 and the momentum variance @xmath27 ; @xmath28 is the normalization constant .
the wavefunction @xmath29 after the jump , see eq.([eq : jj ] ) , is still gaussian : apart from an irrelevant global phase , one has @xmath30 , where @xmath31 with @xmath32 the mean value of the momentum does not change , while the mean value of the position is shifted toward the localization position @xmath15 .
the position variance decreases and hence after the jump the particle is more localized : the reciprocals of @xmath33 and @xmath34 sum up and give the reciprocal of the new variance @xmath35 . importantly , the latter does not depend on the position of the localization process .
for large gaussian wave packets , such that @xmath36 , one has @xmath37 and @xmath38 .
the effects of the localization process on the gaussian wavefunction are summarized in fig.[fig:1]*(a)*. in addition , the probability that the localization process occurs around @xmath15 is , according to eq .
( [ eq : py ] ) , @xmath39 * ( a ) * + , see eq.([eq : gaus ] ) , and * ( b ) * a superposition of gaussian wavefunctions @xmath40 , see eq.([eq : sovr ] ) .
the position probability distribution @xmath41 before the localization is given by the blue dashed line , the probability distribution after the localization @xmath42 is given by the black dot dashed line , while the red line represents the normalized gaussian distribution associated with the jump operator , see eq.([eq : locop ] ) .
the states after the localization are given by eq.([eq : sxgrw ] ) , * ( a ) * , and by eq.([eq : cmp ] ) with @xmath43 , * ( b)*.,title="fig : " ] , see eq.([eq : gaus ] ) , and * ( b ) * a superposition of gaussian wavefunctions @xmath40 , see eq.([eq : sovr ] ) .
the position probability distribution @xmath41 before the localization is given by the blue dashed line , the probability distribution after the localization @xmath42 is given by the black dot dashed line , while the red line represents the normalized gaussian distribution associated with the jump operator , see eq.([eq : locop ] ) .
the states after the localization are given by eq.([eq : sxgrw ] ) , * ( a ) * , and by eq.([eq : cmp ] ) with @xmath43 , * ( b)*.,title="fig : " ] the role of the localization operators can be further understood by means of this simple example @xcite , which clarifies how the action of the localization operators can prevent the system from being in a position superposition .
consider a particle which is in the state @xmath44 given by the superposition of two gaussian wavefunctions with null mean momentum and the same variance , one centered around the position @xmath45 , the other around @xmath46 : @xmath47 where @xmath48 and @xmath49 is a normalization constant .
the state after the localization around @xmath15 , @xmath50 is still the superposition of two gaussian functions ; the normalization constant @xmath51 depends on where the localization takes place . now , consider a localization process around @xmath45 and assume that the distance between the two gaussians is much greater than the localization amplitude , while their width is much smaller than it , i.e. @xmath52 . the previous formula directly gives @xmath53 so that the gaussian centered around the localization position @xmath45 is left almost unchanged , while the gaussian centered around @xmath46 is suppressed by a factor @xmath54 .
the localization process practically destroys the superposition between the two gaussian wavefunctions , leading to a single gaussian state localized around @xmath45 , see fig.[fig:1]*(b)*. in addition , by further exploiting @xmath55 , one finds that the probability of a localization in a neighborhood of @xmath56 is given by @xmath57 : the definitions in eqs .
( [ eq : locop ] ) and ( [ eq : py ] ) allow to recover the usual born s rule for the probability distributions @xcite ; see also appendix [ app : sovr ] .
the grw model is fully determined by the stochastic evolution of the wavefunction @xmath58 previously presented .
however , it is often convenient to deal with the dynamics of the statistical operator which describes the state of the system averaged over all the possible trajectories built up by the different combinations of schrdinger evolutions and localization processes , see also sec .
[ sec : me ] .
the equation of motion of the statistical operator @xmath59 , i.e. , the master equation for the grw model reads @xmath60 + \lambda\left(\int { \mathrm{d}}y \ , l_y(\widehat{x } ) \hat{\rho}(t ) l_y(\widehat{x } ) -\hat{\rho}(t ) \right ) \nonumber \\ & = & - \frac{i}{\hbar}\left[\widehat{h } \,,\ , \hat{\rho}(t)\right ] + \lambda\left(\left(\pi { r_c}^2 \right)^{-1/2 } \int { \mathrm{d}}y \,e^{- ( \widehat{x } - y)^2/(2 { r_c}^2 ) } \hat{\rho}(t)e^{- ( \widehat{x } - y)^2/(2 { r_c}^2 ) } -\hat{\rho}(t ) \right).\end{aligned}\ ] ] the first term describes the standard quantum evolution induced by the hamiltonian @xmath61 , while the second term accounts for the occurrence of the localization processes .
equation ( [ eq : megrw ] ) establishes a semigroup evolution @xcite with pure decoherence in position : the off - diagonal terms in the position representation are suppressed , and distant superpositions are suppressed faster than closer ones .
the master equation associated with the collapse model allows to investigate relevant features , such as the extension to an @xmath62-particle system and then the amplification mechanism , as well as the asymptotic behavior of the energy of the system .
now , consider an @xmath62-particle system in which the localization processes occur individually for each constituent , so that the master equation associated with the @xmath62-particle statistical operator @xmath63 is simply @xmath64 + \sum_j \lambda_j\left(\int { \mathrm{d}}y \ , l_y(\widehat{x}_j ) \hat{\varrho}(t ) l_y(\widehat{x}_j ) -\hat{\varrho}(t ) \right),\ ] ] where @xmath65 is a shorthand notation for @xmath66 , @xmath67 being the identity operator on the hilbert space associated with the @xmath68-th particle , and @xmath69 is the position operator of the @xmath70-th particle , see eq.([eq : locop ] ) . as we will see in sec .
[ sec : ampl ] , the hypothesis of individual localization processes has to be considered with a certain caution .
it is useful to introduce the center of mass coordinates through the invertible linear transformation @xmath71 with @xmath72 , where @xmath73 is the mass of the @xmath70-th particle and @xmath74 the total mass , .
accordingly @xmath75 is the center - of - mass coordinate , while @xmath76 are the relative coordinates .
the position of the @xmath70-th particle can be expressed as @xmath77 one can then easily prove the relation @xmath78 where @xmath79 denotes the partial trace with respect to the relative degrees of freedom . hence ,
if we assume that the total hamiltonian is the sum of a term associated with the center of mass and a term associated with the internal motion , i.e. @xmath80 we find that the state of the center of mass , @xmath81 satisfies the same master equation as that in eq.([eq : megrw ] ) , with the one particle hamiltonian @xmath61 replaced by @xmath82 and , most importantly , the localization rate @xmath12 replaced by @xmath83 .
this is a direct manifestation of the amplification mechanism , which , along with localization , is the crucial feature of the grw model .
it explains why the model describes both microscopic and macroscopic systems . the localization rate @xmath12 ( assume @xmath84 for the sake of simplicity ) of microscopic systems is negligible and therefore the predictions of the grw model about the wavefunction of microscopic systems reproduce for all practical purposes the predictions of standard quantum mechanics .
on the other hand , if we consider a macroscopic object , the rate @xmath85 , with @xmath62 of the order of the avogadro s number , induces a localization of the center of mass on very short time scales : the wavefunctions of macroscopic objects are almost always well - localized in space , so that their centers of mass behave , for all the practical purposes , according to classical mechanics .
a well - known drawback of the grw model is that it predicts an infinite increase of the energy of the system .
this energy divergence is due to larger and larger fluctuations of the momentum induced by the interplay between the schrdinger evolution and the localization mechanism , as will be explicitly discussed in sec.[sec : lmaei ] .
nevertheless , the infinite increase of the energy can be inferred directly from the master equation ( [ eq : megrw ] ) .
in fact , this master equation predicts a linear increase of the mean value of the energy of the system with a rate @xcite @xmath86 this rate of the energy increase is actually very small , @xmath87 for a nucleon , even if one considers the @xmath62-particle case @xcite .
however , it is clear that from a fundamental point of view one would like to avoid the energy divergence and point to a reestablishment of the energy conservation principle within the model .
this forces us to put forward a more realistic description of the interaction between the system and the noise , also in order to test whether and how possible mechanisms excluding infinite energy increase modify the testable predictions of the model @xcite .
the use of the master equation formalism suggests a way out from the problem of the energy divergence . despite the deep conceptual differences between collapse models and the notion of decoherence , the same master equation associated with the grw model can be also derived in a specific model of collisional decoherence @xcite .
this correspondence clarifies that the origin of the energy divergence in the grw model can be ascribed to the lack of a dissipation mechanism , which would account for the energy loss of the system due to the action of the noise . by taking into account the master equation which generalizes eq.([eq : megrw ] ) to include dissipation @xcite , we have thus been led to a possible structure of a new jump operator replacing that in eq.([eq : locop ] ) and excluding the energy divergence .
indeed , this was a preliminary benchmark , but the choice of the jump operator in a collapse model is subjected to further constraints , the most relevant being the induction of localization . in addition , the definition of a collapse model in terms of the different trajectories within the hilbert space can be done without reference to any subsequent master equation or decoherence model . for these reasons ,
we present our results by first postulating a new localization operator and , as a consequence , a new collapse model and only after that we derive the corresponding master equation , see sec.[sec : me ] .
in this paper , we propose the following extension of the grw model : the jump operators @xmath17 defined in eq.([eq : locop ] ) are replaced with @xmath88 where @xmath89 is the momentum operator of the system , @xmath90 the mass of the particle and @xmath91 a new parameter of the model , which is related with the temperature of the noise inducing the localization .
this will be explicitly shown in sec.[sec : therm ] , where the specific choice of @xmath92 , as well as its peculiar role in the definition of @xmath93 , will be discussed .
note that the jump operators of the model are no longer self - adjoint .
an equivalent way to express @xmath93 is given by @xmath94 by which one immediately sees how the original grw jump operator @xmath17 , see eq.([eq : locop ] ) , is obtained in the limit @xmath95 . on the other hand , the general structure provided by items 1 - 4 in the previous section is left untouched .
explicitly , our collapse model can be formulated as follows : * the sudden jumps are now described by @xmath96 * the overall number of jumps is still distributed in time according to a poisson process with rate @xmath12 . *
if there is a jump at time @xmath6 , the probability density that the jump takes place at the position @xmath15 is now given by @xmath97 * still , between two consecutive jumps the state vector evolves according to the schrdinger equation .
it is important to observe that the jump operators satisfy a normalization condition as in eq.([eq : normgrw ] ) ; explicitly , @xmath98 this property guarantees that the probability distribution @xmath19 associated with the localization position is properly normalized , see eq.([eq : pyd ] ) , and its role within the model will be further discussed at the end of sec .
[ sec : edsme ] .
as will be shown extensively in the following , the replacement of @xmath17 with @xmath93 preserves all the desired features of the resulting collapse model , while the dependence on the momentum operator @xmath89 prevents the infinite energy increase and thus leads to a more realistic description of the action of the noise .
the role of the jump operator @xmath93 is illustrated directly by evaluating its action on a gaussian wavefunction @xmath21 , see eq.([eq : gaus ] ) .
the gaussian structure of the wavefunction is preserved and , specifically , one has that the state after the jump , see eq.([eq : jjd ] ) , is @xmath99 , with @xmath100 where we introduced the adimensional quantity @xmath101 which will be crucial in the following analysis , as well as @xmath102 the jump shifts the mean value of the position toward @xmath15 , and , now , it also damps the mean value of the momentum .
the variance @xmath103 after the jump does not depend on where the jump takes place and it is given by the reciprocal of the sum of the reciprocals of @xmath104 and @xmath105 .
the contribution due to the @xmath89-dependent term causes a slight increase of the wave - function width , which partially counterbalances the decrease due to the usual grw contribution . for gaussian wavefunctions such that @xmath106 is smaller than the threshold value @xmath107
the jump induced by @xmath93 will increase overall the position variance .
contrary to the original grw model , the repeated action of the jump operators does not induce an unlimited contraction of the wavefunction : there is a lower threshold under which the jump processes cease to be localization processes .
in realistic situations , this threshold value is not reached by the evolution , see eq.([eq : asvar ] ) and the following discussion .
however , @xmath108 plays a crucial role in fixing the asymptotic finite value of the energy , see sec .
[ sec : therm ] .
let us note that , as in the original grw model , the wavefuction after the localization process is non - vanishing over the whole space .
this is the so - called problem of the tails in collapse models @xcite and , indeed , the introduction of dissipation leaves it unaltered . the probability density for a jump to take place at the position @xmath15 is , see eq.([eq : pyd ] ) , @xmath109 finally , by taking into account the action of @xmath93 on a superposition of two gaussian wavefunctions as in eq.([eq : sovr ] ) , with @xmath110
, one can show along the same lines as for the original grw model , see appendix [ app : sovr ] , that also in this case the localization process destroys the superposition and selects a single gaussian wavefunction , see fig.[fig:1 ] * ( b ) * , with a probability which corresponds to the usual born s rule .
in the previous section , we have extended the grw model by introducing new jump operators , while leaving the general structure of the collapse model untouched .
the dynamics of the wavefunction consists in a unitary evolution interrupted by sudden discontinuous transformations ( jumps ) at random and separated times .
the stochastic differential equations which usually define collapse models are governed by wiener processes @xcite , so that they do not supply the piecewise deterministic evolution now recalled .
however , we will show in this section how also the generalized grw model can be formulated by postulating a stochastic differential equation .
the latter determines the trajectories in the hilbert space of the system through a random field , i.e. a family of stochastic processes , one for each point of space @xmath7 . as for the other collapse models ,
this equation has to be understood as a phenomenological equation , whose fundamental motivation has to be looked for by some underlying theory beyond standard quantum mechanics @xcite . in the proper limit , we will also get a stochastic differential equation for the original grw model .
the reader is referred to @xcite for further details and a rigorous treatment of the stochastic differential equations with jumps in hilbert spaces .
first , let us introduce a family of stochastic processes @xmath111 such that the counting process @xmath112 counts the jumps taking place at a position within @xmath15 and @xmath113 .
the stochastic processes are defined on a common probability space @xmath114 and @xmath115 $ ] indicates the statistical mean with respect to the probability @xmath116 .
furthermore , we denote as @xmath117 a generic sequence of instants and positions in which the jumps occur up to time @xmath6 ; indeed , this corresponds to specifying the trajectories of the counting processes up to time @xmath6 .
we assume that the processes @xmath111 are independent and satisfy @xmath118 & = & \lambda \| l_y(\widehat{x } , \widehat{p } ) { \vert \psi(t ) \rangle}\|^2 { \mathrm{d}}t , \label{eq : edn3}\end{aligned}\ ] ] where @xmath119 is the increment of @xmath120 in a time @xmath121 .
we introduced the short - hand notation @xmath122 to indicate the state of the system depending on the trajectory up to time @xmath6 : the wavefunction is itself a stochastic process , which has values in the hilbert space associated with the system and is determined by the sequences of jumps , i.e. by the trajectories of the counting processes
. equations ( [ eq : edn1 ] ) and ( [ eq : edn2 ] ) tell us that the probability of one count in a time interval @xmath121 is of order @xmath121 , while the probability of more than one count is of higher order @xcite . equation ( [ eq : edn3 ] ) yields the expected value of the increment of the counting processes _ conditioned _ on the occurrence of the sequence of jumps @xmath123 up to time @xmath6 @xcite .
this conditional expected value depends both on the ( stochastic ) wavefunction at time @xmath6 and on the ( deterministic ) jump operator @xmath93 . finally , the wavefunction @xmath5 is fixed by the following non - linear stochastic differential equation : @xmath124 the evolution of the wavefunction in a time interval @xmath121 has a deterministic contribution due to the hamiltonian @xmath61 and a stochastic contribution due the jumps described by eq.([eq : jjd ] ) ; a jump around the position @xmath15 corresponds to a non - zero increment of the counting process @xmath112 .
the solution of eq.([eq : edsnl ] ) can be represented straightforwardly : given the sequence of jumps @xmath125 and the initial condition @xmath126 , the corresponding trajectory in the hilbert space is @xmath127 where @xmath128 is the normalization factor .
this equation formally characterizes all the possible evolutions of the system s state within our model .
the deterministic evolution of the wavefunction induced by the group of unitary operators @xmath129 is interrupted by the jumps described by the operators @xmath130 ; the dynamics introduced in sec .
[ sec : egm ] is then recovered , compare with items @xmath131 and @xmath132 . furthermore , all the other features of the collapse model can be retrieved by the properties of the stochastic processes in eqs.([eq : edn1])-([eq : edn3 ] ) .
let us say that the system is in the state @xmath5 at time @xmath6 and recall that the probability of more than one jump in a time interval @xmath121 is negligible .
hence , the probability density @xmath133 of a jump at the position @xmath15 and a time between @xmath6 and @xmath134 is simply given by the conditional expectation of the increment of the corresponding process @xmath135
, i.e. , @xcite @xmath136 = \lambda \| l_y(\widehat{x } , \widehat{p } ) { \vert \psi(t ) \rangle}\|^2 { \mathrm{d}}t.\ ] ] now , as the jump operators satisfy the normalization condition in eq.([eq : norml ] ) , the probability @xmath137 to have a jump within @xmath6 and @xmath138 at any position is simply @xmath139 i.e. the overall jump rate does not depend on the state of the system and is given by @xmath12 , according to item @xmath140 as a matter of fact , this corresponds to the rate of the poisson process @xmath141 which counts the total number of jumps up to time @xmath6 . finally , the probability density that , if there is a jump at a time between @xmath6 and @xmath134 , it takes place at the position @xmath15 is @xmath142 , so that eq.([eq : pyd ] ) and item @xmath143 are recovered .
the results of this paragraph apply to the original grw model in the limit @xmath95 , i.e. @xmath144 . for example , we can associate the grw model with the non - linear stochastic differential equation @xmath145 by following @xcite , one can also introduce a linear equation equivalent to eq.([eq : edsnl ] ) after a proper change of probability on the measurable space @xmath146 . in appendix
[ sec : gs ] , we study in detail the gaussian solutions of the stochastic differential equation ( [ eq : edsnl ] ) . here , we focus on the evolution of the position variance , thus confirming the effectiveness of the localization mechanism ruling the collapse model .
we further characterize the finite values of the position and the momentum variances in the asymptotic time limit . given a gaussian solution of eq.([eq : edsnl ] ) , i.e. @xmath147 as in eq.([eq : gaus ] ) , with @xmath148 and the normalization @xmath49 in eq.([eq : cc ] ) , the position variance @xmath149 is defined as , compare with eq.([eq : apvar ] ) , @xmath150 as shown in appendix [ sec : gs ] , @xmath149 depends on the instants of the jumps , but not on their position , and hence it is a function of the trajectories @xmath151 of the poisson process @xmath152 . to illustrate in a compact way the evolution of the position variance
, we deal with its statistical mean . for a poisson process with rate @xmath12 , the probability that there is one count between @xmath153 and @xmath154 , @xmath155 , one count between @xmath156 and @xmath157 and no other counts up to time @xmath6 is @xcite @xmath158 :
the probability density associated with @xmath159 only depends on the overall time @xmath6 and number of jumps @xmath160 .
thus , the expected value of the position variance reads @xmath161 = \sum^{\infty}_{m=0 } \lambda^m e^{- \lambda t } \int^t_0 { \mathrm{d}}t_m \ldots \int^{t_2}_0 { \mathrm{d}}t_1 \frac{\left| \mathcal{g}_m\left(\mathcal{g}_{m-1}\left(\ldots \mathcal{g}_1(\gamma)\right)\right ) + \frac{i \hbar ( t - t_m)}{m } \right|^2}{2 \text{re } [ \mathcal{g}_m\left(\mathcal{g}_{m-1}\left(\ldots \mathcal{g}_1(\gamma)\right)\right ) + \frac{i \hbar ( t - t_m)}{m}]},\ ] ] where @xmath162 is defined in eq.([eq : f ] ) , and the integrand expresses the position variance at time @xmath6 on the trajectory @xmath151 , see eqs .
( [ eq : fff ] ) and ( [ eq : varx ] ) . * ( a ) * + ) , as a function of time ( blue dots ) , compared with the deterministic unitary evolution ( red line ) ; the statistical mean is obtained over a sample of @xmath163 trajectories .
( * a * , * b * ) microscopic system : @xmath164 and jump rate given by @xmath165 , with @xmath166 ( * a * ) and @xmath167 ( * b * ) .
( * c * , * d * ) macroscopic system : @xmath168 and jump rate given by @xmath169 , with @xmath170 ( * c * ) and @xmath171 ( * d * ) .
the initial variance is @xmath172.,title="fig : " ] ) , as a function of time ( blue dots ) , compared with the deterministic unitary evolution ( red line ) ; the statistical mean is obtained over a sample of @xmath163 trajectories .
( * a * , * b * ) microscopic system : @xmath164 and jump rate given by @xmath165 , with @xmath166 ( * a * ) and @xmath167 ( * b * ) .
( * c * , * d * ) macroscopic system : @xmath168 and jump rate given by @xmath169 , with @xmath170 ( * c * ) and @xmath171 ( * d * ) .
the initial variance is @xmath172.,title="fig : " ] + + ) , as a function of time ( blue dots ) , compared with the deterministic unitary evolution ( red line ) ; the statistical mean is obtained over a sample of @xmath163 trajectories .
( * a * , * b * ) microscopic system : @xmath164 and jump rate given by @xmath165 , with @xmath166 ( * a * ) and @xmath167 ( * b * ) .
( * c * , * d * ) macroscopic system : @xmath168 and jump rate given by @xmath169 , with @xmath170 ( * c * ) and @xmath171 ( * d * ) .
the initial variance is @xmath172.,title="fig : " ] ) , as a function of time ( blue dots ) , compared with the deterministic unitary evolution ( red line ) ; the statistical mean is obtained over a sample of @xmath163 trajectories .
( * a * , * b * ) microscopic system : @xmath164 and jump rate given by @xmath165 , with @xmath166 ( * a * ) and @xmath167 ( * b * ) .
( * c * , * d * ) macroscopic system : @xmath168 and jump rate given by @xmath169 , with @xmath170 ( * c * ) and @xmath171 ( * d * ) . the initial variance is @xmath172.,title="fig : " ] in fig .
[ fig : loc ] , we can observe the evolution of @xmath173 $ ] for different values of the initial variance @xmath33 and for both the microscopic and the macroscopic regime .
the former refers to the evolution of a single particle with a mass of the order of the nucleon mass @xmath174 , while the latter describes the evolution of the center of mass of a system composed by an avogadro s number @xmath62 of particles . as will be shown in sec .
[ sec : ampl ] , we can apply our collapse model to an @xmath62-particle system by simply replacing the jump rate @xmath12 with @xmath85 and referring @xmath90 to the total mass of the system , at least as long as a rigid body is considered . in the microscopic regime ,
[ fig : loc].(*a * ) and ( * b * ) , the evolution of the expected value of the position variance strictly follows the deterministic unitary evolution up to very long time scales and then saturates to a finite value , see the next paragraph .
the jump rate is @xmath175 and therefore the probability to have a jump will be negligible up to , say , @xmath176 : the action of the noise does not induce any observable localization process on the microscopic systems . on the other hand , when macroscopic systems are taken into account , fig .
[ fig : loc].(*c * ) and ( * d * ) , the evolution described by eq.([eq : pv ] ) strongly departs from the unitary one from the very beginning of the dynamics .
the repeated occurrence of the jumps rapidly reduces the position spread of the wavefunction , so that the localization mechanism is clearly manifested .
the time scale of the wavefunction localization is @xmath177 and it is the same as that for the grw model .
the total rate of events in our extended grw model is in fact the same as in the original one , compare items 2 of secs .
[ sec : egm ] and [ sec : gsotm ] .
we conclude that the modification of the jump operators put forward with eq.([eq : lyxp2 ] ) does not introduce any significant change in the localization mechanism compared to the original grw model , as also shown by the results of the next paragraph .
the trajectories of the model are made up of a sequence of deterministic unitary evolutions and random jumps .
these two transformations have opposite effects on the wavefunction , as long as the position variance is concerned .
the free evolution induces a spread of the position variance , which is the faster the narrower the wavefunction . on the contrary , the jumps
shrink the wavefunction , at least as long as @xmath178 , see eq.([eq : soglia ] ) . at some point of the evolution
the two opposite effects balance each other and thus the position variance reaches a finite and non - zero equilibrium value @xcite . as shown in appendix [ app : axp ] , the asymptotic value of the position variance can be evaluated via the relation @xmath179 with @xmath180 this asymptotic value is in general much higher than the value @xmath181 , which would correspond to the threshold in eq.([eq : soglia ] ) . for a macroscopic system , with @xmath182 ,
one has @xmath183 , while @xmath184 , which is also in agreement with the estimate given in @xcite for @xmath185 . as a matter of fact , due to the specific choice of @xmath186 ,
the threshold value is very small , so that the free evolution and the jumps balance each other before the spread of the wavefunction can reach it .
analogously , see appendix [ app : axp ] , the asymptotic value of the momentum variance is given by @xmath187 for a macroscopic system with @xmath188 , one gets @xmath189 , still perfectly compatible with the value for @xmath185 @xcite , so that @xmath190 is approximately twice the minimum value allowed by the uncertainty relation .
it is worth noting that , as the position variance , also the momentum variance reaches a finite asymptotic value , both for @xmath191 and for @xmath192 , which naturally leads to the following remark .
the reason for the energy divergence in the original grw model is quite a subtle one .
it is often understood by saying that the jump operator in eq.([eq : locop ] ) induces an indefinite contraction of the width of the wavefunction , so that @xmath193 and therefore , in accordance with the uncertainty relation , @xmath194 , implying the divergence of the energy .
however , it is clear how this picture is not the end of the story and it is , to some extent , misleading .
a crucial role here is played by the schrdinger evolution between the jumps .
as now recalled , the balance between unitary evolution and jumps implies a finite asymptotic value of the momentum variance , also for @xmath185 .
this means that the energy divergence in the grw model is actually due to fluctuations of the mean value of the momentum , as @xmath195 . to be more explicit ,
consider an initial gaussian wavefunction , @xmath196 .
the unitary evolution up to the first jump at time @xmath153 does not modify the mean value of the momentum @xmath197 and shifts the mean value of the position as @xmath198 , see eqs.([eq : free ] ) , ( [ eq : meanx ] ) and ( [ eq : momav ] ) for @xmath185 .
moreover , the unitary evolution introduces an imaginary component in @xmath199 , according to @xmath200 . because of such an imaginary component
, the jump at time @xmath153 and position @xmath15 actually modifies the mean value of the momentum , which after the jump will be : @xmath201}{{\gamma}^{'\text{r}}_{t_1 } } ( \langle x \rangle_{t_1 } - y ) = { \beta}+ \frac{\hbar m \tau_1}{\hbar^2 \tau^2_1 + m^2 { \gamma } ( { r_c}^2+{\gamma } ) } ( y- \langle x \rangle_{t_1}),\label{eq : lpp}\ ] ] where @xmath202 has reduced to @xmath203 , see eq.([eq : fsig ] ) , since we are now considering the limit @xmath185 , i.e. , the original grw model .
the system varies its momentum proportionally to the distance between the position of the jump and the mean value of the position before the jump .
this shift of the momentum , in turn , contributes to the change in position after the jump : between the first and the second jump the mean value of the position evolves as @xmath204 .
the iteration of these two transformations , according to the different spatial distribution of the jumps , will generate some trajectories such that both the mean position and the mean momentum will diverge to @xmath205 , and some other trajectories where they will diverge to @xmath206 : in both cases the mean kinetic energy will asymptotically diverge . due to the symmetric probability distribution of the location of each jump , see for example eq.([eq : pygrw ] ) , the effect on the mean momentum now described will be on average null , and then @xmath207 = { \beta}$ ] . however , the statistical average of the squared mean value of the momentum will diverge , @xmath208 \rightarrow + \infty$ ] , and thus the average of the mean energy will diverge with it .
the introduction of a dissipative mechanism , through a small @xmath209 , only slightly modifies the action of the jump operators , see sec .
[ sec : egm ] and appendix [ sec : gs ] . nevertheless , this tiny modification is enough to damp the long - time momentum fluctuations , thus leading to an asymptotic finite value of the energy , as will be shown and discussed in sec.[sec : therm ] .
up to now , we have dealt with the stochastic evolution of the wavefunction , as fixed by eq.([eq : edsnl ] ) .
the latter provides a complete characterization of the collapse model , as it yields all the possible piecewise deterministic trajectories which can be obtained according to items 1 - 4 in sec .
[ sec : egm ] . nevertheless , it is often convenient to study the predictions of the model related with the statistical mean of relevant physical quantities ,
i.e. , compare with eq.([eq : avo ] ) , @xmath210 = \mathbbm{e}[{\langle \psi(t ) \vert}\widehat{o}{\vert \psi(t ) \rangle } ] = \text{tr}\left\{\mathbbm{e}[{\vert \psi(t ) \rangle}{\langle \psi(t ) \vert } ] \widehat{o}\right\ } = \text{tr}\left\{\hat{\rho}(t ) \widehat{o}\right\ } .\label{eq : avo2}\ ] ] here , we have introduced @xmath211,\ ] ] which is by construction a statistical operator on the hilbert space associated with the system .
incidentally , since the stochastic wavefunction @xmath5 is uniquely determined by the trajectories @xmath123 , the statistical mean in eq.([eq : erho ] ) corresponds to the mean over the different trajectories @xmath123 , each one weighted with its @xmath116-probability density . in the following , we will focus on the evolution of the statistical operator @xmath59 , which will allow us to describe the evolution of relevant physical quantities , as well as to further characterize the dissipation and the amplification mechanism in the model .
the equation of motion satisfied by @xmath59 , i.e. , the master equation associated with the extended grw model , is easily determined by using the product rule @xmath212 and eqs.([eq : edn1])-([eq : edn3 ] ) .
explicitly , eqs.([eq : edn1 ] ) and ( [ eq : edn2 ] ) imply that the stochastic differential equation ( [ eq : edsnl ] ) gives ( using the notation @xmath213 ) @xmath214 { \mathrm{d}}t + \int { \mathrm{d}}y \left(\frac{\widehat{l}_y ( { \vert \psi(t ) \rangle } { \langle \psi(t ) \vert})\widehat{l}^{\dag}_y } { \|\widehat{l}_y { \vert \psi(t ) \rangle}\|^2 } - { \vert \psi(t ) \rangle } { \langle \psi(t ) \vert}\right ) { \mathrm{d}}n_y(t),\ ] ] which can be written as @xcite @xmath215 { \mathrm{d}}t+ \lambda \left ( \int { \mathrm{d}}y \widehat{l}_y ( { \vert \psi(t ) \rangle } { \langle \psi(t ) \vert } ) \widehat{l}^{\dag}_y - { \vert \psi(t ) \rangle } { \langle \psi(t ) \vert } \| \widehat{l}_y { \vert \psi(t ) \rangle } \|^2 \right ) { \mathrm{d}}t \nonumber\\ & & + \int { \mathrm{d}}y \left(\frac{\widehat{l}_y ( { \vert \psi(t ) \rangle } { \langle \psi(t ) \vert } ) \widehat{l}^{\dag}_y } { \| \widehat{l}_y { \vert \psi(t ) \rangle } \|^2 } - { \vert \psi(t ) \rangle } { \langle \psi(t ) \vert}\right)\left ( { \mathrm{d}}n_y(t ) - \lambda \| \widehat{l}_y { \vert \psi(t ) \rangle } \|^2 { \mathrm{d}}t \right ) .
\label{eq : prelo}\end{aligned}\ ] ] since the defining properties of the conditional expected value imply @xmath216 = \mathbbm{e}[{\vert \psi(t ) \rangle } { \langle \psi(t ) \vert}]$ ] , we can get an equation for @xmath59 by taking the expectation of eq.([eq : prelo ] ) conditioned upon the trajectory @xmath123 .
the conditional expected value of @xmath217 , see eq.([eq : edn3 ] ) , implies that the second line in eq.([eq : prelo ] ) does not give any contribution . by exploiting eq.([eq : norml ] ) and further taking the stochastic average , we end up with the master equation @xmath218 + \lambda\left(\int { \mathrm{d}}y \ , l_y(\widehat{x } , \widehat{p } ) \hat{\rho}(t ) l^{\dag}_y(\widehat{x } , \widehat{p } ) -\hat{\rho}(t ) \right ) \nonumber\\ & = & - \frac{i}{\hbar}\left[\widehat{h } \,,\ , \hat{\rho}(t)\right ] + \lambda\left ( \frac{{r_c}(1+k)}{\sqrt{\pi } \hbar } \int { \mathrm{d}}{q}\ , e^{\frac{i}{\hbar } { q}\widehat{x } } e^{-\frac{{r_c}^2}{2\hbar^2 } \left((1+k ) { q}+ 2 k \widehat{p}\right)^2 } \hat{\rho}(t ) e^{-\frac{{r_c}^2}{2\hbar^2 } \left((1+k ) { q}+ 2 k \widehat{p}\right)^2}e^{-\frac{i}{\hbar } { q}\widehat{x } } -\hat{\rho}(t ) \right ) .
\nonumber \\\end{aligned}\ ] ] indeed , this is a lindblad master equation @xcite , which reduces to the master equation associated with the grw model in the limit @xmath219 @xcite .
[ [ role-of-the-normalization-condition-in-eq.eqnorml ] ] role of the normalization condition in eq.([eq : norml ] ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in the next two paragraphs , we will discuss more in detail the physical meaning of eq.([eq : megrwdiss ] ) ; before that , let us make the following remark .
suppose to define a collapse model via an equation as eq.([eq : edsnl ] ) , but with different jump operators @xmath220 , which do not satisfy eq.([eq : norml ] ) .
first , this would imply a total rate of localization dependent on the state of the system and proportional to @xmath221 , as well as a probability density for the localization position defined as @xmath222 , compare with eq.([eq : pyd ] ) .
even more importantly , as can be easily seen by repeating the calculations of this paragraph , eq.([eq : megrwdiss ] ) should be replaced with @xmath223 + \lambda\left(\int { \mathrm{d}}y \tilde{l}_y(\widehat{x } , \widehat{p})\hat{\rho}(t)\tilde{l}^{\dag}_y(\widehat{x } , \widehat{p } ) - \mathbbm{e}\left [ \|\tilde{l}_y(\widehat{x } , \widehat{p}){\vert \psi \rangle } \|^2 { \vert \psi(t ) \rangle}{\langle \psi(t ) \vert}\right ] \right).\end{aligned}\ ] ] in order to get a closed equation for the statistical operator @xmath59 , which would also be in the lindblad form , one has to change the stochastic differential equation ( [ eq : edsnl ] ) by replacing the hamiltonian @xmath61 with the effective non - hermitian hamiltonian @xmath224 as well as adding a term which guarantees the norm preservation of the wavefunction @xcite .
the evolution between the jumps is then no longer unitary , but it is given by a more general completely positive map .
we conclude that the choice of the jump operators , and , in particular , their property expressed by eq.([eq : norml ] ) , is crucial to define the collapse model in terms of the usual unitary evolution interrupted by sudden jumps , at least as long as we want the evolution of @xmath59 to be described by a closed linear master equation . in order to understand the meaning of the master equation ( [ eq : megrwdiss ] ) ,
let us first note that a master equation of the same form appears within the description of the collisional decoherence , and also recall what has been said at the end of sec .
[ sec : amaed ] .
explicitly , the dynamics of a test particle interacting through collisions with a free low density background gas in the weak coupling regime ( and restricting to the one dimensional case for the sake of comparison ) , can be characterized through the equation @xcite : @xmath225 + ( 2 \pi)^2 n_{\text{gas } } \int { \mathrm{d}}{q}|\tilde{t}({q})|^2 \left ( \ , e^{\frac{i}{\hbar } { q}\widehat{x } } \sqrt{s({q } , \widehat{p } ) } \hat{\rho}(t ) \sqrt{s({q } , \widehat{p } ) } e^{-\frac{i}{\hbar } { q}\widehat{x } } - \frac{1}{2}\left\{s({q } , \widehat{p } ) \ , \hat{\rho}(t)\right\ } \right ) .
\label{eq : medic}\ ] ] here , @xmath226 is the density of the gas , while @xmath227 is the fourier transform of the two - body interaction potential between the test particle and the gas particles .
finally , @xmath228 is a two point correlation function , which is usually called dynamic structure factor and the operator valued function @xmath229 is defined through the relation @xmath230 .
the dynamic structure factor describes the energy and momentum exchange between the test particle and the gas , and for a free gas of maxwell - boltzmann particles it can be written as @xmath231 where @xmath232 is the inverse temperature and @xmath160 the mass of the gas particles , while @xmath233 is the energy exchanged in a collision such that @xmath234 .
now , if we choose an interaction potential @xmath235 , implying @xmath236 eq.([eq : medic ] ) exactly corresponds to eq.([eq : megrwdiss ] ) upon performing the following identifications between the parameters of the two equations : @xmath237 the new parameter of the model @xmath92 precisely corresponds to the most probable velocity of the gas particles @xmath238 .
in addition , the localization width is fixed by the thermal wavelength @xmath239 of the gas particle .
the same correspondence is present between the original grw model and the collisional master equation without dissipation @xcite .
finally , note that the master equation ( [ eq : medic ] ) , and then eq.([eq : megrwdiss ] ) as well , fall into the class of translation - covariant lindblad master equation , whose full characterization was given by holevo @xcite .
now that the link with a collisional model has been fixed , the interpretation of eq.([eq : megrwdiss ] ) is quite straightforward .
the lindblad operators , which describe the action of the environment on the system , are made up of two terms . the boost operator
@xmath240 describes the exchange of a momentum @xmath241 between the test particle and the background gas , according to @xmath242 .
the operator @xmath243 provides the probability amplitude that the change of momentum is equal to @xmath244 if the test particle has momentum @xmath245 , as seen by @xmath246 . by taking @xmath219 , the dependence on the momentum operator @xmath89 in eq.([eq : lmw ] )
disappears : the probability density that the system undergoes a momentum variation @xmath241 is fixed , i.e. , it does not depend on the momentum of the system itself .
therefore , the mean value of the momentum is constant , while the momentum variance steadily increases in time .
any dissipative effect is excluded from the dynamics , thereby leading to a divergence of the energy @xcite . the master equation ( [ eq : medic ] )
reduces to the master equation introduced in @xcite , which accounts for the recoil - free decoherence dynamics of a massive particle in the limit @xmath247 and reproduces the original grw master equation @xcite .
the introduction of a dependence on @xmath89 within the lindblad operators is just what introduces the energy relaxation in the collisional dynamics of the test particle , thus keeping the mean value of the energy finite , see sec .
[ sec : therm ] . as well - known ,
each master equation in the lindblad form can be obtained as the statistical mean of infinite different stochastic differential equations .
this fact is often conveyed by saying that any lindblad master equation has infinite different unravellings @xcite , essentially one for each way of writing it in terms of different lindblad operators . in particular , the master equation of our model could be obtained also by starting from a stochastic differential equation with a family of counting processes @xmath248 , one for each variation of the system s momentum , and with jump operators given by @xmath249 , i.e. by the lindblad operators in eq.([eq : megrwdiss ] ) .
this kind of unravelling was introduced in @xcite to study numerically the solutions of the master equation .
however , one should keep in mind that the stochastic differential equation can be associated with a collapse model only if the localization mechanism is present .
the position spread of the wavefunction has to be reduced in the different trajectories of the model . in a nutshell ,
not every unravelling of the master equation can be a candidate to describe a collapse model .
one can easily see that jump operators such as @xmath249 would not induce a localization of the wavefunction , while their fourier transform defines the jump operators of our model , i.e. , @xmath250 [ sec : sitpe ] the solution of the master equation can be obtained by exploiting the characteristic function @xcite @xmath251 as the matrix elements of the statistical operator in the position representation can be obtained through @xmath252 in appendix [ app : dote ] , we show that eq.([eq : megrwdiss ] ) is equivalent to the following equation for the characteristic function : @xmath253 since @xmath254 , see eq.([eq : kg ] ) , in general the off - diagonal elements @xmath255 will not vary significantly by replacing @xmath256 with @xmath257 , while keeping @xmath258 constant . under this condition , one can neglect the dependence on @xmath186 within the second term at the right hand side of eq.([eq : char ] ) , see eq.([eq : inv ] ) , thus getting @xmath259 with @xmath260 this first order partial differential equation is solved by @xcite @xmath261 where the function @xmath262 satisfies the free equation @xmath263 . by using eq.([eq : inv ] ) , one ends up with the solution of the master equation in the position representation , @xmath264 where @xmath265 is the solution of the free schrdinger equation .
note that @xmath266 guarantees the trace preservation of @xmath267 , while @xmath268 guarantees its hermiticity . by means of the expression of the statistical operator at time @xmath6
, we can explicitly evaluate the dynamics of relevant physical quantities , such as the statistical mean value of the position , see eq.([eq : avo2 ] ) , as well as its variance .
the variance @xmath269 of the observable represented by @xmath270 , if the average state of the system is @xmath59 , reads @xmath271 for the position mean value , by eq.([eq : mesol ] ) we have @xmath272 where @xmath273 denotes the mean value under free schrdinger evolution . as for the case without dissipation @xcite , the average effect of the localization processes does not influence the evolution of the mean value of the position .
analogously , one finds @xmath274 thus , the position variance is , see eq.([eq : frak ] ) , @xmath275 where @xmath276 is the position variance under the schrdinger evolution .
the dissipation introduces a spread of the position variance which is linear in time , in addition to the term proportional to @xmath277 , already present in the original grw model @xcite .
it is worth noting that the position variance referred to the statistical operator @xmath59 is not the statistical average of the position variance for the stochastic wavefunction , see eq.([eq : pv ] ) . while for the mean values of the observables one has eq.([eq : avo2 ] ) , an analogous relation does not hold for the variances , so that in general @xmath278 \neq ( \delta_{\rho_t } o)^2 $ ] .
the reason for that is essentially that the stochastic average @xmath279 and the square of the trace do not commute @xcite , see eq.([eq : frak ] ) .
by virtue of the dynamics for the statistical operator , we can now show in an explicit way how the modification of the jump operators leads to energy relaxation .
to do so , we study the evolution in time of the mean energy @xmath280 , focussing on the case @xmath281 . instead of using the solution in the position representation given by eq.([eq : mesol ] ) , we can compute directly the dynamics of the mean value of any operator @xmath282 , which is a function of the momentum operator only , as follows .
since @xmath283 = 0 $ ] and @xmath284 , eq.([eq : megrwdiss ] ) implies @xmath285 it is worth mentioning how this result is a direct consequence of the translation covariance of the master equation @xcite . for the momentum operator , one has @xmath286 the mean value of the momentum is then damped exponentially in time with the rate given by the product between the grw rate @xmath12 and the dissipation parameter @xmath287 ; namely @xmath288 by removing dissipation , i.e. setting @xmath185 , one recovers a constant value of the mean momentum .
for the kinetic energy , one has @xmath289 so that @xmath290 the mean value of the energy relaxes with rate @xmath291 to the asymptotic value @xmath292 indeed , since @xmath293 $ ] reaches an asymptotic finite value , @xmath294 for @xmath295 will be almost surely finite on the trajectories of the collapse model , see also sec .
[ sec : lmaei ] .
the asymptotic value of the mean energy given by eq.([eq : has ] ) corresponds to a temperature of the noise @xcite @xmath296 where we exploited the definition of @xmath186 in eq.([eq : kg ] ) .
most importantly , we have obtained a value of the noise temperature which is independent from the mass of the system .
let us stress that this is a consequence of the choice of the jump operator @xmath93 in eq.([eq : lyxp2 ] ) , including the specific dependence on the new parameter @xmath92 .
in addition , the estimate in eq .
( [ eq : t ] ) justifies our initial choice for the numerical value of @xmath92 : it yields the same order of magnitude of the temperature as the continuous collapse model analyzed in @xcite .
crucially , this value also points out how a proper collapse noise does not need suspicious and ad hoc properties .
a classical noise with typical cosmological features ( low temperature ) can guarantee the collapse of the wavefunction , along with the thermalization to a finite energy @xcite . finally , as expected , in the limit @xmath95 we recover an infinite temperature of the noise , corresponding to the original grw model .
the dissipation rate is @xmath297 , see eqs .
( [ eq : ht ] ) and ( [ eq : rrate ] ) and therefore its ratio with the collapse rate is @xmath298 the condition @xmath299 guarantees that the collapse occurs on a time scale @xmath300 much shorter than the time scale @xmath301 of dissipation .
in addition , recall that the collapse rate @xmath12 is not modified by the introduction of the dissipation , see secs .
[ sec : egm ] and [ sec : this ] .
any experimental investigation on the effects of collapse models on short time scales will thus not be able to highlight a significant role of the extended dissipative grw model with respect to the original one .
on the other hand , the exponential relaxation of the energy to a finite value drastically changes the predictions of the model for the experiments which involve the secular behavior of the energy @xcite .
the most important example is provided by the heating of the intergalactic medium ( igm ) , which up to now yields the second strongest upper bound to the localization rate @xcite .
such a bound has been derived by considering the continuous spontaneous localization ( csl ) collapse model , which allows to deal with the fermi or bose statistics for identical particles .
hence , the influence of dissipation on the predictions on the secular behavior of the energy has been evaluated in @xcite , where we extended the csl model in order to include dissipation .
it is worth noting how the asymptotic value of the energy is fixed by the threshold value introduced in sec .
[ sec : egm ] , see eq.([eq : soglia ] ) .
recall that the jump operators cease to induce a localization of the wavefunction for gaussian wave packets with a position spread smaller than @xmath302 as already noticed , the hamiltonian part of the master equation does not give any contribution to the evolution equation of the mean value of the energy .
this means that the same equation would be obtained by starting from a master equation as eq.([eq : megrwdiss ] ) , but without the hamiltonian term @xcite .
such a master equation can be thought as given by the statistical mean of the stochastic differential equation , compare with eq.([eq : edsnl ] ) , @xmath303 the trajectories provided by this equation are fixed by the action of the jumps operators only .
now , for the sake of simplicity , consider an initial minimum uncertainty gaussian state as in eq.([eq : gaus ] ) .
its position variance @xmath172 is increasingly contracted by the action of the jumps operator , see eq.([eq : xps ] ) , until it reaches the threshold value in eq.([eq : thrd ] ) , which represents the asymptotic value of @xmath149 for the trajectories of eq.([eq : aux ] ) .
but then , since the initial minimum uncertainty gaussian state has remained a minimum uncertainty gaussian state due to the absence of the free schrdinger evolution , the asymptotic value of the momentum variance will be @xmath304 .
as the mean value of the momentum @xmath305 relaxes to zero , @xmath306 determines the asymptotic value of the mean energy as @xmath307 , which is the value in eq.([eq : has ] ) .
let us emphasize once more that this description does not correspond to what actually happens in the collapse model fixed by eq.([eq : edsnl ] ) , but characterizes an auxiliary model , fixed by eq.([eq : aux ] ) , which gives the same predictions as the extended grw model , as long as the statistical mean value of the energy is concerned .
as recalled in sec . [ sec : amaed ] , the amplification mechanism is a basic feature of any collapse model .
the localization of the wavefunction due to the jumps has to increase with the number of the constituents of the system , as well as with its overall mass . in this section ,
we show that such a mechanism can be proved also in the presence of dissipation , if a rigid body is taken into account .
we also argue that the description of more complex systems , in which the internal dynamics plays a significant role , calls for a more realistic and detailed characterization of the @xmath62-particle evolution than that regularly used in collapse models .
let us start by dealing with an @xmath62-particle system subject to localization processes occurring individually for each constituent , which means that the overall effect of the noise consists simply in the sum of the independent effects on the single particles , see sec .
[ sec : amaed ] . in the next paragraph
, we will see how the validity of this assumption breaks down if the internal motion of the system has to be taken into account .
hence , we consider now the master equation for the total state @xmath63 given by , compare with eq.([eq : megrwn ] ) [ we are using the analogous short - hand notation ] , @xmath308 + \sum_j \lambda_j \left(\int { \mathrm{d}}y \ , l_y(\widehat{x}_j , \widehat{p}_j ) \hat{\varrho}(t ) l^{\dag}_y(\widehat{x}_j , \widehat{p}_j ) -\hat{\varrho}(t ) \right),\ ] ] where @xmath69 ( @xmath309 ) is the position ( momentum ) operator of the @xmath70-th particle , see eq.([eq : lyxp2 ] ) .
again , we introduce the center of mass coordinates through eq.([eq : rj ] ) , and we assume a total hamiltonian as in eq.([eq : hhh ] ) .
hence , the state of the center of mass , @xmath310 , satisfies @xmath311 + \sum_j \lambda_j \left ( \frac{{r_c}(1+k_j)}{\sqrt{\pi } \hbar } \int { \mathrm{d}}{q}\ , e^{\frac{i}{\hbar } { q}\widehat{x}_{\text{cm}}}\right .
\nonumber\\ & & \left.tr_{\text{rel } } \left\{e^{-\frac{{r_c}^2}{2\hbar^2 } \left((1+k_j ) { q}+ 2 k_j \widehat{p}_j\right)^2 } \hat{\varrho}(t ) e^{-\frac{{r_c}^2}{2\hbar^2 } \left((1+k_j ) { q}+ 2 k_j \widehat{p}_j\right)^2 } \right\}e^{-\frac{i}{\hbar } { q}\widehat{x}_{\text{cm } } } -\hat{\rho}_{\text{cm}}(t ) \right),\nonumber\end{aligned}\ ] ] where we used the cyclicity of the partial trace over the relative degrees of freedom .
the parameters @xmath312 are proportional to @xmath313 according to the relation , see eq.([eq : kg ] ) , @xmath314 and @xmath315 . as a consequence
, the previous equation can be well approximated by @xmath311 + \sum_j \lambda_j \left ( \frac{{r_c}}{\sqrt{\pi } \hbar } \int { \mathrm{d}}{q}\ , e^{\frac{i}{\hbar } { q}\widehat{x}_{\text{cm}}}\right .
\nonumber\\ & & \left.tr_{\text{rel } } \left\{e^{-\frac{{r_c}^2}{2\hbar^2 } \left ( { q}+ 2 k_j \widehat{p}_j\right)^2 } \hat{\varrho}(t ) e^{-\frac{{r_c}^2}{2\hbar^2 } \left({q}+ 2 k_j \widehat{p}_j\right)^2 } \right\}e^{-\frac{i}{\hbar } { q}\widehat{x}_{\text{cm } } } -\hat{\rho}_{\text{cm}}(t ) \right).\nonumber\end{aligned}\ ] ] now , we introduce the crucial assumption that we are dealing with a rigid body , so that @xmath316 , where @xmath317 is the total momentum .
note that the same assumption plays a relevant role also in order to define a proper measure for the macroscopicity of quantum superpositions @xcite .
thus , we finally get @xmath318 \nonumber\\ & & + \lambda_{\text t } \left ( \frac{{r_c}}{\sqrt{\pi } \hbar } \int { \mathrm{d}}{q}\ , e^{\frac{i}{\hbar } { q}\widehat{x}_{\text{cm } } } e^{-\frac{{r_c}^2}{2\hbar^2 } \left ( { q}+ 2 k_{\text t } \widehat{p}_{\text t}\right)^2 } \hat{\rho}_{\text{cm } } ( t ) e^{-\frac{{r_c}^2}{2\hbar^2 } \left({q}+ 2 k_{\text t } \widehat{p}_{\text t}\right)^2 } e^{-\frac{i}{\hbar } { q}\widehat{x}_{\text{cm } } } -\hat{\rho}_{\text{cm}}(t ) \right),\label{eq : ampldiss}\end{aligned}\ ] ] where , indeed , @xmath319 this equation is equivalent to the one - particle equation , see eq.([eq : megrwdiss ] ) , with the replacements @xmath320 apart from the factor @xmath321 which should multiply @xmath244 in the argument of the square exponential , as well as the coefficient @xmath322 .
this difference is safely negligible : one has , e.g. , @xmath323 for a macroscopic body of mass @xmath324 .
of course , the remaining dependence on @xmath325 in eq.([eq : ampldiss ] ) can not be neglected , since this would modify the asymptotic behavior of the momentum and energy of the system , see sec .
[ sec : therm ] .
the center of mass of a rigid body behaves as a single particle of mass @xmath326 , but with a rate of localization increased according to the number of components of the body . as shown in sec .
[ sec : paml ] , see especially fig .
[ fig : loc ] , the replacements in eq.([eq : repl ] ) induce the expected amplification mechanism : the localization processes do not affect the microscopic systems on observable time scales , while they allow to treat the center of mass of a macroscopic rigid body as a classical , well - localized , object
. the analysis of the previous paragraph can be applied only if the influence of the internal dynamics on the evolution of the center of mass is negligible . actually , we think that the simple procedure usually exploited in order to show the amplification mechanism , which describes the evolution of the center of mass through the same master equation as for the one - particle system with the replacements in eq .
( [ eq : repl ] ) , has a proper physical meaning just if one takes into account a rigid body . in different situations
, one would generally need and also expect a more complex analysis , which encompasses the interrelationship between the mutual interaction of the @xmath62 particles and the action of the noise .
we will develop such an analysis in a future and dedicated work .
once again , the purpose of our investigation should be clear : in our view collapse models are to be understood as phenomenological models and the limits of the first simpler models can be highlighted and overcome by searching for more realistic characterizations .
first , let us examine a simple example , which starts from the study put forward in @xcite ; we will draw , however , different conclusions . consider the two - particle state given by @xmath327 , so that @xmath328 with @xmath329 the normalization constant .
this is a product state with respect to the partition of the total hilbert space in terms of the center of mass and relative degrees of freedom , while it describes an entangled state of the two particles .
we set @xmath330 , so that @xmath331 , while , as usual , @xmath332 .
now , suppose that a localization process centered at a position @xmath15 affects particle 1 and that we can describe such a process through the operator @xmath333 .
the localization process is assumed to influence individually each constituent of the @xmath62-particle system : its action on particle 1 is independent from particle 2 .
hence , by using eq.([eq : braxl ] ) one has @xmath334 and therefore the state after the collision , @xmath335 , can be written as @xmath336 which now also includes entanglement between the center of mass and the relative degrees of freedom .
the mean values of the center of mass and relative positions after the localization process are @xmath337 now , we can associate the initial relative state @xmath338 with the ground state of a one - dimensional harmonic oscillator . in the spirit of @xcite
, we can describe a two - atom or a two - nucleon ground state by setting properly the wavefunction width @xmath35 .
we take @xmath339 for the atoms and @xmath340 for the nucleons , so that the level spacing @xmath341 of the corresponding harmonic oscillator is of the order of , respectively , @xmath342 and @xmath343 , as it must be . in both cases
one has @xmath344 and thus , if we also assume @xmath345 , the mean values in eq.([eq : meanbgrw ] ) can be well approximated as @xmath346 analogously , one can show that the two position variances are left almost unchanged ; namely , after the localization one has @xmath347 and @xmath348 .
the probability for the localization process to take place at the position @xmath15 is @xmath349 this provides a gaussian distribution of the localization position , centered around the initial mean position @xmath350 of particle 1 .
the width @xmath351 of this gaussian distribution , which is independent from the initial state of the system , implies that there is a non - negligible probability to have a localization process within , say , @xmath352 away from the initial mean position .
but now , let us focus on the case of a two - nucleon state . according to eq.([eq : xcmrel ] )
, the mean value of the relative position can thus be increased by a localization process from @xmath353 up to approximately @xmath354 , which clearly indicates that @xmath355 can be no longer associated with a nucleon bound state . in the original grw model , i.e. , for @xmath185 , in this regime @xmath356
would not be modified by the localization process , see eq.([eq : xcmrel ] ) .
one could think that the inclusion of dissipation within the grw model necessarily leads to sudden internal transitions or even dissociations of nuclei , which would be of course an unacceptable feature of the model .
this was actually the conclusion drawn in @xcite , where a similar example was considered @xcite .
we think , however , that this is not the case and that the previous example is pointing out something different .
one should in fact realize that the description of the localization mechanism has been carried out independently from the presence of a mutual interaction between the components of matter .
when one says that the localization on a 2-particle system is described by the jump operator @xmath333 ( and by @xmath357 ) , this means that the localization process on each particle is the same whether or not they mutually interact .
however , it is clear that such a characterization is in general not realistic and thus possibly leads to unphysical predictions . in the previous example
, one concludes that the two nucleons are shifted apart much farther than the nuclei length scale by the localization mechanism , simply because the latter has been described without taking into account the effects of the interaction between the two nucleons , e.g. their binding energy . in order to estimate the average variation of the energy @xmath358 due to the localization process ,
let us consider for simplicity the action of the localization operator on a one - particle gaussian state @xmath359 , see eq.([eq : xps ] ) .
hence , under the assumption @xmath360 , the average exchanged energy @xmath358 is of the order of @xmath361 this value is 5 orders of magnitude smaller than the energy needed for an internal transition in the case of a nucleon bound state , so that it should be clear how the localization processes of the model can not induce matter dissociation .
once again , the analogy with decoherence turns out to be useful .
if one considers the scattering of a tracer particle with a particle of its environment , the effects of the collision will be radically different whether or not the tracer particle is in a bound state .
for instance , think about a gas of free tracer particles interacting with a background gas , in the low density regime : to a good extent , the collisions of each tracer particle with the environmental gas can be treated independently , so that the overall dynamics can be acquired by summing up the individual ones , see eq .
( [ eq : megrwndiss ] ) . on the other hand ,
if the tracer particles are in a bound state , their interaction hamiltonian will combine with the interaction hamiltonian between each of them and the environmental particles , thus modifying the scattering processes .
more in general , consider the derivation of a master equation for an @xmath62-particle open system starting from the @xmath131-particle master equation , but including a mutual interaction between the @xmath62 particles during the interaction with the environment .
usually , the proper master equation can not be obtained by simply adding the mutual interaction hamiltonian to the hamiltonian term and summing up the @xmath131-particle lindblad equations , but the lindblad operators should be modified , as well .
actually , in general one can not even expect to obtain a lindblad master equation for the single particles , not to mention the center of mass of the @xmath62-particle system .
a non - markovian description of these dynamics will likely come into play @xcite .
for the interested readers , in @xcite a microscopic derivation of the master equation of two interacting qubits in a common environment is given .
this work highlights in a clear way how the mutual interaction between the two qubits modifies significantly the resulting lindblad structure , influencing , e.g. , the populations dynamics , as well as the fluorescence spectrum .
summarizing , the usual description of the amplification mechanism relies on the idea that the noise acts on each constituent without being influenced by the presence of the others .
however , in general this can not be taken for granted a priori .
of course , this does not mean that a more realistic description of the @xmath62-particle system is incompatible with the amplification mechanism .
on the contrary , we think that such a description will help to clarify to what extent and how the action of the noise induces the localization of the center of mass on more complex systems .
this collective property of an @xmath62-particle system seems to be more motivated if the effects of the noise on the different components are correlated .
indeed , a limited , but consistent way to take into account the mutual interaction between the constituents of the system is just obtained by dealing with a rigid body .
the details of the internal dynamics are neglected , but each particle is considered as a component of a single @xmath62-particle bound system . the action of the noise on each constituent is turned directly into the corresponding action on the center of mass : the overall effect of the noise is correctly described by the sum of the action on each constituent , so that the results of the previous paragraph are in this case justified .
in this paper , we have extended the grw model @xcite in order to avoid the infinite growth of the energy of the system .
we have introduced new jump operators , while leaving the other defining features of the model unchanged .
the jump operators correspond to the fourier transform of the lindblad operators of a model for collisional decoherence @xcite and depend on the momentum operator of the system .
hence , they induce a dissipative evolution , which leads to a damping of the momentum , as well as to a finite asymptotic value of the energy . the inclusion of a dissipative mechanism within the collapse model has called for the definition of a new parameter , which is directly linked to the temperature @xmath0 of the noise .
the original grw model is retrieved in the high temperature limit @xmath1 .
we have proved that the collapse model can be formulated equivalently by means of a proper stochastic differential equation , defined in terms of a random field made up of one counting process for each point of space .
the stochastic differential equation fully determines the possible trajectories in the hilbert space of the system . by focusing on the case of gaussian initial states
, we have investigated the main features of these trajectories , showing in particular the occurrence of the position and momentum localization .
the latter clarifies that the energy divergence in the original grw model has to be traced back to fluctuations in momentum , rather than to an unlimited increase of the momentum variance .
as long as a rigid body is taken into account , the amplification mechanism still holds , so that the center of mass of a macroscopic system is localized on very short time scales and behaves for all the practical purposes according to classical mechanics . as usual , this has been shown by considering a master equation for the center of mass which is given by the sum of the one - particle master equations and taking the partial trace over the relative degrees of freedom .
nevertheless , we have also argued that a more general and realistic analysis of the @xmath62-particle system s evolution should be given if the interaction between the components of the system can affect the action of the noise .
our results provide an important step toward the establishment of a realistic jump collapse model , which generalizes the original grw proposal . indeed , important advances are still to be made . a promising possibility to recover
entirely the principle of the energy conservation is to consider the noise as a real physical field influenced by the presence of the system : the energy variations of the latter could be then explained in terms of an energy exchange with the noise .
let us emphasize how this approach could be put forward by associating the random field to the physical field , as proposed for the continuous - time collapse models @xcite .
furthermore , also in this paper we have assumed the hypothesis of a white noise : this greatly simplifies the characterization of the model and can provide a satisfactory description of the interaction between the system and the noise in certain regimes .
however , there is no reason to exclude a priori the general case of a colored noise , such that , e.g. , the time distribution of the localization events is no longer memoryless . finally , the inclusion of dissipation has allowed us to avoid the energy divergence also in the csl model @xcite , by properly modifying the corresponding stochastic differential equation in the fock space associated with the system , and to investigate the overheating problem in the disi - penrose model @xcite .
the authors acknowledge financial support from the eu project nanoquestfit , infn , the funding fra-2014 of the university of trieste and cost ( mp1006 ) .
they also thank g.c .
ghirardi and g. gasbarri for very useful discussions .
in this appendix we study the action of the jump operator @xmath93 on a single - particle state @xmath44 given by the superposition of two gaussian wavefunctions , with null mean momentum , equal variance and opposite mean positions , @xmath56 ( @xmath362 ) , see eq.([eq : sovr ] ) and fig .
[ fig:1 ] * ( b ) * , such that @xmath363 .
first , note that the matrix elements of @xmath93 in the position representation are given by @xmath364 where @xmath186 is defined in eq.([eq : kg ] ) . as long as @xmath191 , where @xmath185 corresponds to the original grw model
, the off - diagonal matrix elements of the jump operator can be different from @xmath365 . as a consequence
, one has that the state @xmath366 after the localization is , see eq.([eq : jjd ] ) , @xmath367 compare with eq.([eq : cmp ] ) .
let s now focus on a localization process taking place at @xmath368 , so that @xmath369 where we have used @xmath370 . still , the second term is strongly suppressed since @xmath371 : the localization mechanism leaves us with only the gaussian wavefunction centered around @xmath45 .
let us now evaluate the probability that the localization position @xmath15 is such that @xmath372 ( so that it is far from @xmath46 ) . proceeding exactly as before , one finds @xmath373 i.e. a gaussian centered around @xmath45 and with width @xmath374 .
then , if we consider the probability to have a jump in a neighborhood of @xmath45 , say , within @xmath375 , we end up with @xmath376 : on a proper coarse grained spatial scale ( fixed by @xmath377 ) the predictions of quantum mechanics are recovered for all the practical purposes .
in this appendix , we are going to study more in detail the trajectories described by eq.([eq : sol ] ) , focusing on the case of gaussian wavefunctions .
we will show that the stochastic differential equation preserves the gaussian structure of the wavefunction , whose evolution in time will be explicitly characterized .
let us take into account a gaussian initial state @xmath196 , see eq.([eq : gaus ] ) , and assume @xmath281 , so that the unitary evolution describes the usual spreading of the gaussian wave packet .
the transition from the initial state to the gaussian state immediately before the jump at time @xmath153 can be summarized as @xmath378 the jump at time @xmath153 and position @xmath15 modifies the wavefunction according to eq.([eq : xps ] ) : explicitly @xmath379 where @xmath380 is defined as in eq.([eq : kg2 ] ) .
hence , the wavefunction has still a gaussian form after the jump , but since the free evolution implies a complex coefficient @xmath381 , both @xmath382 and @xmath383 will be in general complex numbers .
the full solution can be built up by iterating these two steps , where , of course , @xmath384 ( @xmath153 ) has to be replaced with the instant @xmath385 ( @xmath386 ) of the @xmath387-th ( @xmath70-th ) jump and @xmath388 in eq.([eq : free ] ) have to be replaced with the parameters @xmath389 after the @xmath387-th jump . finally ,
after the last jump at time @xmath156 , eq.([eq : free ] ) has to be used once more , with the instants @xmath6 and @xmath156 , as well as the parameters @xmath390 . in conclusion , the solution @xmath5 is the gaussian wavefunction @xmath147 as in eq.([eq : gaus ] ) , with @xmath148 and @xmath391 the normalization constant , with @xmath392 and @xmath393 , respectively , real and imaginary part of @xmath394 ; moreover , @xmath395 implies @xmath396 , while @xmath397 .
the parameters @xmath398 are stochastic quantities depending on the specific trajectory @xmath123 ( i.e. they are short - hand notation for @xmath399 ) .
in particular , @xmath400 only depends on the number of jumps up to time @xmath6 : given a trajectory with @xmath160 jumps , one has @xmath401 on the other hand , @xmath199 depends also on the instants at which the jumps take place , but not on their positions : explicitly , @xmath402 with @xmath403 where @xmath404 and @xmath405 .
finally , @xmath406 depends also on the specific positions where the jumps take place : one has @xmath407 for the sake of completeness ,
let us mention that the parameters @xmath398 satisfy the following stochastic differential equations : @xmath408 the deterministic contributions describe the evolution of the parameters under the unitary part of the dynamics , while the stochastic terms describe the action of the jumps .
the stochastic contribution in the equations of @xmath400 and @xmath199 traces back to the poisson process @xmath152 defined in eq.([eq : np ] ) , which counts the overall occurrence of the jumps , without any discrimination about their position . from the evolution of the parameters which fix the gaussian wavefunction
, one can directly infer the evolution of relevant physical quantities . given a wavefunction @xmath147 as in eq.([eq :
gaus ] ) , with @xmath148 , as well as the normalization @xmath49 in eq.([eq : cc ] ) , let us denote as @xmath409 the mean value of the observable @xmath270 on such a state at time @xmath6 , i.e. , @xmath410 while we denote as @xmath411 the corresponding variance , @xmath412 the mean value of the position at time @xmath6 is @xmath413 while the variance is @xmath414 with @xmath415 .
moreover , the mean value of the momentum at time @xmath6 is @xmath416 while the variance is @xmath417 hence , one has @xmath418 so that @xmath147 will not be a minimum uncertainty state as long as @xmath419 .
indeed , the quantities in eqs.([eq : meanx])-([eq : varp ] ) do depend on the trajectory @xmath123 : the variances depend on the instants in which the jumps occur , while the mean values also depend on the positions of the jumps , see eqs.([eq : pzt])-([eq : xzt ] ) .
the explicit evaluation of these quantities then calls for a numerical analysis , which can be easily achieved starting from the results of this appendix .
in order to estimate the asymptotic value of , respectively , position and momentum variance , it is convenient to examine the trajectories in which all jumps are equally spaced in time @xcite , so that @xmath420 for any @xmath70 ( where , of course , @xmath12 has to be replaced with @xmath421 for a macroscopic system ) .
since the position variance is fixed by @xmath199 , see eq.([eq : varx ] ) , a stable condition will be reached when a cycle composed of a free evolution and the subsequent jump does not modify the value of @xmath106 .
explicitly , the equilibrium value @xmath422 satisfies , see eqs.([eq : fff ] ) and ( [ eq : f ] ) , @xmath423 with @xmath424 .
this equation is solved by @xmath425 where @xmath108 is the threshold value defined in eq.([eq : soglia ] ) . given a complex number @xmath426 ,
its square root can be expressed as @xmath427 with @xmath428 if @xmath429 , @xmath430 if @xmath431 and @xmath432 , @xmath433 if @xmath434 and @xmath435 is the signum function . by using eq.([eq : equ ] ) , we thus get eq.([eq : asvar ] ) for the asymptotic value of the position variance . moreover , replacing eq.([eq : equ ] ) in the expression of the momentum variance in eq.([eq : varp ] ) , we get the asymptotic value in eq.([eq : deltap ] ) .
in this appendix , we show that eq.([eq : megrwdiss ] ) for the statistical operator @xmath59 is equivalent to eq.([eq : char ] ) for the characteristic function @xmath436 defined in eq.([eq : chid ] ) . by eq.([eq :
megrwdiss ] ) , we get @xmath437 + \mathcal{b}\left[\chi(\nu,\mu , t)\right ] - \lambda \chi(\nu,\mu , t).\ ] ] let us start with the hamiltonian contribution : @xmath438 & = & -\frac{i}{\hbar}\left(\text{tr}\left\{\widehat{h}\hat{\rho}(t)e^{\frac{i}{\hbar}\left(\nu \hat{x}+\mu \hat{p}\right ) } \right\ } -\text{tr}\left\{\hat{\rho}(t)\widehat{h}e^{\frac{i}{\hbar}\left(\nu \hat{x}+\mu \hat{p}\right ) } \right\ } \right ) = \frac{\nu}{m}\partial_{\mu } \chi(\nu,\mu , t),\end{aligned}\ ] ] where we used @xmath439 = \nu \left ( \widehat{p } ^{\frac{i}{\hbar}\left(\nu \hat{x}+\mu \hat{p}\right ) } + e^{\frac{i}{\hbar}\left(\nu \hat{x}+\mu \hat{p}\right ) } \widehat{p } \right)\ ] ] and the cyclicity of the trace .
the dissipative term provides us with @xmath440 & = & \frac{{r_c}(1+k ) \lambda}{\sqrt{\pi } \hbar } \int { \mathrm{d}}{q}\ , \text{tr}\left\ { e^{\frac{i}{\hbar } { q}\widehat{x } } e^{-\frac{{r_c}^2}{2\hbar^2 } \left ( ( 1+k ) { q}+2 k \widehat{p}\right)^2 } \hat{\rho}(t ) e^{-\frac{{r_c}^2}{2\hbar^2 } \left ( ( 1+k ) { q}+ 2 k \widehat{p}\right)^2}e^{-\frac{i}{\hbar } { q}\widehat{x } } e^{\frac{i}{\hbar}\left(\nu \hat{x}+\mu \hat{p}\right ) } \right\ } \nonumber\\ & = & \frac{{r_c}(1+k ) \lambda}{\sqrt{\pi } \hbar } \int { \mathrm{d}}{q}{\mathrm{d}}x \ , { \langle x \vert}\left ( e^{\frac{i}{\hbar } { q}\widehat{x } } e^{-\frac{{r_c}^2}{2\hbar^2 } \left ( ( 1+k ) { q}+2 k \widehat{p}\right)^2}\hat{\rho}(t ) e^{-\frac{{r_c}^2}{2\hbar^2 } \left ( ( 1+k ) { q}+2 k \widehat{p}\right)^2 } e^{-\frac{i}{\hbar } { q}\widehat{x } } e^{\frac{i}{\hbar } \mu \hat{p } } e^{\frac{i}{\hbar } \nu \hat{x}}\right){\vert x \rangle } e^{-\frac{i}{2 } \frac{\mu \nu}{\hbar } } \nonumber\\ & = & \frac{{r_c}(1+k ) \lambda}{\sqrt{\pi } \hbar } \int { \mathrm{d}}{q}{\mathrm{d}}x { \mathrm{d}}x ' { \mathrm{d}}x '' \ , e^{\frac{i}{\hbar } { q}\mu } f(x , x')\rho(x ' , x '' , t ) f(x '' , x-\mu ) e^{\frac{i}{\hbar } \nu x } e^{-\frac{i}{2 } \frac{\mu \nu}{\hbar}},\label{eq : prev}\end{aligned}\ ] ] where we used @xmath441 and we have introduced @xmath442 inserting this relation and eq.([eq : inv ] ) in eq.([eq : prev ] ) , we obtain @xmath440 & = & \frac{\lambda(1+k)}{16 \pi^{5/2 } \hbar^2 k^2 { r_c } } \int { \mathrm{d}}{q}{\mathrm{d}}x { \mathrm{d}}x ' { \mathrm{d}}x '' { \mathrm{d}}\nu ' \ , e^{\frac{i}{\hbar } ( { q}-\nu/2 ) \mu } e^{- \frac{i}{2 \hbar } \nu ' ( x'+x '' ) } \chi(\nu ' , x'-x'',t ) \nonumber\\ & & \times e^{- \frac{1}{8 k^2 { r_c}^2}\left[(x - x')^2+(x''-x + \mu)^2\right ] } e^{-\frac{i(1+k ) { q}}{2\hbar k}(x''-x ' + \mu ) } e^{\frac{i}{\hbar } \nu x } \nonumber\\ & = & \frac { \lambda(1+k)}{8 \pi^{2 } \hbar^2 k } \int { \mathrm{d}}{q}{\mathrm{d}}x ' { \mathrm{d}}x '' { \mathrm{d}}\nu ' \ , e^{\frac{i}{\hbar } { q}\mu } e^{- \frac{i}{2 \hbar } \nu ' ( x'+x '' ) } \chi(\nu ' , x'-x'',t ) \nonumber\\ & & \times e^{-\frac{i(1+k ) { q}}{2 \hbar k}(x''-x ' + \mu ) } e^{-\frac{\nu^2 { r_c}^2 k^2}{\hbar^2 } } e^{-\frac{1}{16 k^2 { r_c}^2}(x'-x''-\mu)^2}e^{i\frac{\nu}{2\hbar}(x'+x'')}\nonumber\\ & = & \frac { \lambda(1+k)}{16 \pi^{2 } \hbar^2 k } \int { \mathrm{d}}{q}{\mathrm{d}}s { \mathrm{d}}\delta { \mathrm{d}}\nu ' \ , e^{\frac{i}{\hbar } { q}\mu } e^{- \frac{i}{2 \hbar } ( \nu'-\nu ) s } \chi(\nu ' , \delta , t ) e^{\frac{i(1+k ) { q}}{2\hbar k}(\delta - \mu ) } e^{-\frac{\nu^2 { r_c}^2 k^2}{\hbar^2 } } e^{-\frac{1}{16 k^2 { r_c}^2}(\delta-\mu)^2}\nonumber\\ & = & \frac { \lambda(1+k)}{4 \pi \hbar k } \int { \mathrm{d}}{q}{\mathrm{d}}\delta \ , e^{\frac{i}{\hbar } { q}\mu } \chi(\nu , \delta , t ) e^{\frac{i(1+k ) { q}}{2 \hbar k } ( \delta - \mu ) } e^{-\frac{\nu^2 { r_c}^2 k^2}{\hbar^2 } } e^{-\frac{1}{16 k^2 { r_c}^2}(\delta-\mu)^2}\nonumber\\ & = & \lambda \chi\left(\nu , \mu\left(\frac{1-k}{1+k}\right),t\right ) e^{-\frac{\nu^2 { r_c}^2 k^2}{\hbar^2 } } e^{-\frac{\mu^2}{4 { r_c}^2(1+k)^2}}.\end{aligned}\ ] ] this concludes the proof . in @xcite a one particle state , which corresponds to our @xmath338 ,
was directly taken into account .
the authors evaluated the superposition probability between the states , respectively , after and before the localization , averaged over the possible localization positions .
this probability goes to zero for the values of @xmath103 corresponding to atoms or nucleons .
note that two momentum - dependent jump operators , which are actually different from ours , were investigated . | in this paper we present an extension of the ghirardi - rimini - weber model for the spontaneous collapse of the wavefunction . through the inclusion of dissipation
, we avoid the divergence of the energy on the long time scale , which affects the original model .
in particular , we define new jump operators , which depend on the momentum of the system and lead to an exponential relaxation of the energy to a finite value .
the finite asymptotic energy is naturally associated to a collapse noise with a finite temperature , which is a basic realistic feature of our extended model .
remarkably , even in the presence of a low temperature noise , the collapse model is effective .
the action of the new jump operators still localizes the wavefunction and the relevance of the localization increases with the size of the system , according to the so - called amplification mechanism , which guarantees a unified description of the evolution of microscopic and macroscopic systems .
we study in detail the features of our model , at the level of both the trajectories in the hilbert space and the master equation for the average state of the system .
in addition , we show that the dissipative ghirardi - rimini - weber model , as well as the original one , can be fully characterized in a compact way by means of a proper stochastic differential equation . |
quantum interference effects in quantum wires ( qws ) are potentially useful in nanotechnology since coupling to the continuum states shows an even - odd parity effect in the conductance when the fermi energy is localized at the center of the energy band @xcite .
consequently , a fine control of the electron transport can be achieved by varying the external parameters of the qw . in this context , we have recently considered new quantum devices based on an array of quantum dots ( qds ) @xcite , a double qd @xcite and nanorings @xcite coupled to a qw .
the attached device acts as scatterer for electron transmission through the qw and allows to tune its transport properties .
it was found that the conductance at zero temperature through the qw shows a complex behavior as a function of the fermi energy : far from the center of the band the conductance depends smoothly on the fermi energy , while around the center it develops an oscillating band with resonances and antiresonances due to quantum interference in the ballistic channel .
moreover , the transmission phase of the electron carries information complementary to the transmission probability .
this phase has been measured in qds @xcite and recently it was reported @xcite the experimental observation of the fano - kondo antiresonance in a qw with a side - coupled qd .
these experiments proved that transport through the system has a coherent component . in this work
we report further progress along the lines indicated above .
in particular , we study theoretically transport properties of a set of side - coupled double qds attached to a perfect qw . we find an analytical expression for the transmission probability and transmission phase .
the transmission probability at the center of the energy spectrum shows an energy spectrum with gap or allowed band depending of the symmetry up - down , to be explained below . in a symmetry
up - down , an even - odd parity effect in the transmission phase at the center of the band is demonstrated .
moreover , we show that an allowed band is formed in the asymmetric case and that the width of this band can be controlled by suitable gate voltages .
this phenomenon is in analogy to the dicke effect in quantum optics , that takes place in the spontaneous emission of two closely - lying atoms radiating a photon into the same environment @xcite . in the electronic case , however , the decay rates ( level broadening ) are produced by the indirect coupling of the up - down qds , giving rise to a fast ( _ superradiant _ ) and a slow ( _ subradiant _ ) mode .
this close analogy opens the way to exploit new electronic effects that usually arise in atomic physics . in this regard
, it has been shown that coupled qds display the electronic counterpart of fano and dicke effects that can be controled via a magnetic flux @xcite .
recently , brandes reviewed the dicke effect in mesoscopic systems @xcite .
the system under consideration is shown in fig .
the qw is attached to a @xmath0 side - coupled double qds .
the system , assumed in equilibrium , is modeled by a noninteracting anderson tunneling hamiltonian @xcite that can be written as @xmath1 , where @xmath2 describes the dynamics of the qw , @xmath3 being the hopping between neighbor sites of the qw , and @xmath4 ( @xmath5 ) creates ( annihilates ) an electron at site @xmath6 . side - coupled double qds attached to a perfect qw.,width=226 ] on the other side , @xmath7 , given by @xmath8 is the hamiltonian for the @xmath0 side - coupled double qds , where @xmath9 ( @xmath10 ) is the annihilation ( creation ) operator of an electron in the qd ( @xmath11 , @xmath12 is the corresponding single energy level . here
the index @xmath13 refers to the up ( @xmath14 ) and down ( @xmath15 ) qd , attached at site @xmath16 of the qw .
the coupling between the two subsystems ( qw and qds ) is described by the hamiltonian @xmath17 where @xmath18 is the coupling between the qw and one of the qds .
the hamiltonian for the qw , @xmath19 , corresponds to a free - particle hamiltonian on a lattice with spacing unity and whose eigenfunctions are expressed as bloch solutions @xmath20 where @xmath21 is the momentum eigenstate and @xmath22 is a wannier state localized at site @xmath16 .
the dispersion relation associated with these bloch states reads @xmath23 consequently , the hamiltonian supports an energy band from @xmath24 to @xmath25 and the first brillouin zone expands the interval @xmath26 $ ] .
the stationary states of the entire hamiltonian @xmath27 can be written as @xmath28 where the coefficient @xmath29 ( @xmath30 ) is the probability amplitude to find the electron at site @xmath16 of the qw [ at the qd @xmath31 ) ] in the state @xmath32 , that is , @xmath33 and @xmath34 .
the amplitudes @xmath29 obey the following linear difference equation @xmath35 the amplitudes @xmath36 can be expressed in terms of @xmath29 as follows @xmath37 from eq .
( [ bj - aj ] ) above , eq .
( [ eq - aj ] ) becomes @xmath38 where the site energy @xmath39 $ ] depends on the electron energy @xmath40 .
thus , the problem reduces to a linear chain of a @xmath0 sites of effective energies @xmath41 . in order to study the solutions of eq .
( [ newdif2 ] ) , we assume that the electrons are described by a plane wave incident from the far left with unity amplitude and a reflection amplitude @xmath42 , and at the far right by a transmission amplitude @xmath43 .
that is , @xmath44 the solution in the region @xmath45 can be written as @xmath46\ , \nonumber \\
a_{j}^{(k ) } & = ce^{\kappa j}+de^{-\kappa j}\ , & \text{if } \left\vert ( \omega-\tilde{\varepsilon})/2v\right\vert & > 1\ , \quad \kappa=-\cosh^{-1 } \left[-(\omega-\tilde{\varepsilon})/2v\right ] \ . \label{solut}\end{aligned}\ ] ] inserting ( [ solut2 ] ) and ( [ solut ] ) into ( [ newdif2 ] ) , we get a inhomogeneous system of linear equations for @xmath47 and @xmath42 , leading to the following result : if @xmath48 , @xmath49 with @xmath50 given by @xmath51 on the contrary , when @xmath52 @xmath53 with @xmath50 given by @xmath54 the transmission probability is given by @xmath55 , and it is related to the linear conductance at the fermi energy @xmath56 by the one - channel landauer formula at zero temperature , @xmath57 @xcite .
we also can obtain the transmission phase as @xmath58 .
to uncover the main features of the electron transport through the qw and the effects of the attached qds , we now consider several physical situations . if @xmath59 , the transmission probability and transmission phase reduces to @xmath60 with @xmath61 and @xmath62 .
we note that the transmission probability oscillates as a function of both @xmath0 and @xmath63 . on the other hand , when @xmath64 we get @xmath65 where @xmath66 and @xmath67 .
therefore , in this energy region @xmath0 tends exponentially to zero when @xmath0 is large , namely @xmath68 . to avoid the profusion of free parameters , for the sake of clarity we set the energies of the up and down qds as , @xmath69 and @xmath70 hereafter .
we first consider the case @xmath71 . in this case
the transmission exhibits a forbidden band ( gap ) .
figure [ fig2 ] shows the transmission probability versus @xmath40 for different values of @xmath0 .
it is apparent that @xmath72 tends to zero within a range @xmath73 $ ] , with @xmath74 , and the system shows a gap of width @xmath75 . inside the energy region @xmath73 $ ] the conductance does not present any feature that depends on @xmath0 .
however , at the center of the band ( @xmath76 , @xmath77 , @xmath78 ) the transmission phase takes on the following values , depending on the parity of @xmath0 : when @xmath0 is odd , @xmath79 for @xmath80 , while for @xmath0 even , @xmath81 for @xmath82
. we can understand this result as follows .
each time that the electron passes near a double qds , it undergoes a phase change equal to @xmath83 due to the destructive interference between the discrete levels in the double qd and the continuum states of the qw .
consider now the situation with @xmath84 .
figure ( [ fig3 ] ) shows the transmission probability for different values of @xmath85 , for @xmath86 and @xmath87 .
we note that an allowed band develops at the center of the gap .
it is straightforward to show that for @xmath88 the width of this allowed band is @xmath89 .
moreover the transmission probability becomes always unity at the center of the allowed band , independently of the value of @xmath90 figure [ fig4 ] displays the transmission probability for several values of @xmath85 when @xmath86 ( solid line ) and @xmath87 ( dashed line ) .
we note that the shape of the allowed band is independent of the value of @xmath85 .
additionally , fig .
[ fig5 ] shows the transmission phase for a similar set of parameters in the region of the allowed band .
we note a series of discontinuities of the transmission phase at some values of the energy @xmath40 .
it is worth to note that the drops in the transmission phase do not imply vanishing transmission probability . at the energies of the drops of the transmission phase , the real part of the transmission amplitude vanishes and at the same time its imaginary part changes its sign .
additionally from eq.([phase1 ] ) , we obtain that the transmission phase is zero ( or @xmath91 , @xmath92 integer ) at the center of the allowed band , independent of the value of @xmath0 and @xmath85 . at the same time
, the transmission probability becomes unity .
this phenomenon resembles the dicke effect in optics , which takes place in the spontaneous emission of a pair of atoms radiating a photon with a wave length much larger than the separation between them @xcite .
the luminescence spectrum is characterized by a narrow and a broad peak , associated with long and short - lived states , respectively .
the former state , weakly coupled to the electromagnetic field , is called _ subradiant _ , and the latter , strongly coupled , _ superradiant _ state . in the present case this effect is due to the indirect coupling between up - down qds through the qw .
the states strongly coupled to the qw yield a forbidden band with width @xmath75 and the states weakly coupled to the qw give an allowed dicke band with width @xmath93 .
in this work we studied the transmission probability and transmission phase through a qw with an array of side - coupled double qds .
we found that the transmission probability displays a gap or a band at the center of the energy spectrum , due to destructive and constructive interference in the ballistic channel , respectively .
for an array with symmetry up - down ( @xmath71 ) only a gap develops at the center of the band . for @xmath94
an allowed band arises at the center of the energy spectrum independently of the value of @xmath85 .
this phenomenon is in analogy to the dicke effect in optics .
the set of side coupled double qds seems a suitable system to study the dicke effect in experiments on quantum transport .
this effect could be used to develop nanodevices where an extremely fine control of electron transport could be achieved just by varying the gate potential of the qds .
p. a. o. would like to thank financial support from millennium science nucleus condensed matter physics and fondecyt under grants nos .
1020269 and 7020269 .
work in madrid was supported by mcyt ( mat2003 - 01533 ) .
e. diez acknowledge support by junta de castilla y len ( sa007b05 ) and mec ( ramn y cajal and fis2005 - 01375 ) .
additionally , p. a. o. would like to thank the hospitality of departamento de fsica de materiales ( ucm ) and departamento de fsica fundamental ( usal ) during his visits . | a system of an array of side - coupled quantum - dots attached to a quantum wire is studied theoretically .
transport through the quantum wire is investigated by means of a noninteracting anderson tunneling hamiltonian .
analytical expressions of the transmission probability and phase are given .
the transmission probability shows an energy spectrum with forbidden and allowed bands that depends on the up - down asymmetry of the system . in up - down symmetry
only the gap survives , and in up - down asymmetry an allowed band is formed .
we show that the allowed band arises by the indirect coupling between the up and down quantum dots .
in addition , the band edges can be controlled by the degree of asymmetry of the quantum dots .
we discuss the analogy between this phenomenon with the dicke effect in optics . |
in the case where the electrostatic coupling is symmetric with respect to the two electrode polarities , the force on a segment of membrane floating above the electrodes can be described by @xmath39 with the area of the segment @xmath10 , distance between fixed electrodes and membrane @xmath5 , dc voltage @xmath11 between fixed electrodes , and an electrostatic force constant @xmath12 ; @xmath13 is a scaling function with units of inverse area . in writing the force law ( [ eq : forcesi ] ) we have assumed the membrane dynamics to take place under fixed voltage conditions .
this requires the voltage source to supply and absorb charge at the timescale on which the membrane motion modulates the capacitance .
however , even if we operate in the opposite limit , i.e. fixed charge on the capacitor , the appropriate correction to the force law ( [ eq : forcesi ] ) is negligible for our setup @xcite ; this is due to stray capacitance ( in the cables from the voltage source ) that is much larger than that of the electrode - membrane arrangement .
intuitively , the large stray capacitance acts as a charge reservoir for the modulated electrode - membrane capacitor , effectively keeping the electrode voltage constant .
the equilibrium of forces for an infinitesimal piece of membrane with the area @xmath40 and thickness @xmath18 is @xmath41 with the displacement function @xmath42 , the tensile pre - stress @xmath43 , and the mass density @xmath19 .
@xmath44 is a heaviside step function ( or a product of such functions ) taking into account electrode gaps and edges , as well as the hole in the membrane metallization of some of the membranes used .
the deflection of a membrane can be described by @xmath45 with the mode shape function @xmath46 considering the first order taylor approximation of the electrostatic force ( [ eq : forcesi ] ) , the equation of motion can be written as @xmath47 in a linear system , the static force term @xmath48 causes a static deflection of the membrane .
this static deflection does not influence the eigenfrequency and can thus be neglected . following galerkin s method , ( [ eq : equationofmotion ] ) can be solved for the fundamental normal mode by multiplying it with @xmath49 and integrating over the entire membrane area @xmath50 with @xmath51 @xmath52 which with ( [ eq : drum_mode_expansion ] ) can be written as @xmath53 where we have ignored the induced couplings to other membrane modes assumed to be weak .
the frequency can now be isolated from ( [ eq : galerkin ] ) @xmath54 with the eigenfrequency @xmath55 of a membrane with zero voltage applied ( @xmath56 ) @xmath57 and introducing the overlap factor @xmath58 between the membrane mode @xmath49 and the electrode mask @xmath59 the first order taylor approximation of @xmath60 becomes @xmath61 which results in a relative frequency shift of @xmath62 for a membrane with a continuous electrode ( @xmath63 ) the frequency shift due to the electrostatic spring softening effect becomes @xmath64 in our experiments , the electrostatic force factor @xmath12 was determined by measuring the relative frequency shift of a membrane as a function of @xmath11 .
the measured data was fitted with @xmath65 and @xmath12 can then be obtained from @xmath66^{-1 } h \rho \omega_0 ^ 2 \eta_{1,1}^{-1}\ ] ] or @xmath67^{-1 } h \rho \omega_0 ^ 2\ ] ] for @xmath68 .
dismissing fringe fields , the force between the plates of a parallel plate capacitor with distance @xmath5 is @xmath69 where @xmath70 is the vacuum permittivity .
the potential of the floating conductive membrane is @xmath71 , given that it is placed symmetrically in relation to the two electrode polarities .
thus , the force on a conductive membrane segment of area @xmath10 to coplanar quarter - segment electrodes is @xmath72 and therewith the force constant ( [ eq : forcesi ] ) becomes @xmath73 using the scaling function @xmath74 . in our experiments with sin - g membranes on quarter - segment electrodes ,
it is possible that we may have additional capacitance contributions from the excess graphene outside the membrane area ( as opposed to the case of sin - al membranes where the al layer is strictly confined to within the sin membrane area ) .
specifically , such extra contributions may occur in a highly asymmetric manner with regard to the two electrode polarities .
this prompts us to derive the theoretical value of @xmath12 for the case of highly asymmetric capacitance between the conductive membrane and the two electrode polarities . in this case
the membrane will be essentially equipotential with one of the electrode polarities whereby the electrostatic force vanishes on membrane segments above electrodes of this polarity .
hence , under these circumstances , the membrane coupling to the electrodes is extremely asymmetric wrt . the two polarities and @xmath12 acquires a dependence on position in the membrane plane : the force constant vanishes for the part of the membrane above electrodes with which it is equipotential , while for membrane segments above the other polarity , with which the potential difference is @xmath11 , eq .
( [ eq : force_parallel ] ) leads to @xmath75 which is four times the symmetrical value , eq .
( [ eq : c_sym_theory ] ) .
we will now discuss how the above effect , if present in our setup , would affect the force constants extracted from experiment . in the simple case where the membrane _ mode _ is symmetric wrt . the two electrode polarities ( while the electrostatic force is not ) , the experimental force constant , as calculated using eq .
( [ eq : c_exp_eta11 ] ) , samples the effective and ineffective electrodes equally , resulting in @xmath76 ( given that the perfect conductor model is valid for sin - g ) .
this is a factor of 2 enhancement relative to the theoretical value for the case of symmetrical capacitances , eq .
( [ eq : c_sym_theory ] ) . due to a somewhat asymmetric electrode mask ( featuring a centrally placed connection bar not shown in fig . 1 of the main text ) and , more importantly , the possibility of misalignment ( estimated to be up to 25% of the membrane side length ) , the simple result for the symmetrical membrane mode does not apply in general .
in fact , by means of eq .
( [ eq : eta_general_4seg ] ) presented below , we find that the addition of a large , asymmetrical capacitance wrt .
one polarity can enhance the fundamental mode frequency shift by a factor of @xmath77 for alignments within the cited range . by assuming a flat distribution within this range , we find that the effect would enhance the averaged force constant by a factor of 2.1 . a related mechanism
, that would result in exactly the same enhancement as just described , is involuntary charging of the membrane by short - circuiting with one of the electrode polarities with @xmath78 time much less than the membrane oscillation period @xmath79 . on the other hand ,
consider a similar short - circuiting but with @xmath78 time longer than the membrane period while still much less than the scan time for the dc voltage @xmath11 .
in this situation the dynamics take place with fixed charge on the membrane and without access to the charge reservoir of the large stray capacitance as the latter now acts in series with the much smaller membrane - electrode capacitance ( compare to the situation described in section [ sec : force_constant ] ) .
we find that the fundamental mode frequency shift can be modified by a factor of @xmath80 within the alignment range stated earlier . assuming a flat alignment distribution once again , we find that the averaged force constant would be enhanced by a factor of 1.4 in case of such membrane charging
. regarding sin - g on quarter - segment electrodes , either of the above effects could serve to explain , at least partially , the large error bars of the average force constant ( presented in fig .
3c of the main text ) .
moreover , the effects of strongly asymmetrical capacitance and charging with short @xmath78 time enhance the frequency shift in a way that is compatible with the average experimental force constant lying above the theoretical value for symmetrical coupling by a factor of 2.7 .
we discussed in section [ sec : theory_c_4seg ] the two limits of coupling a conductive membrane to quarter - segment electrodes : symmetric and strongly asymmetric . in calculating the experimental force constants presented in the main text for the quarter - segment geometry , we assume that the capacitance between the electrodes and the conductive layer on the membrane only has contributions from the membrane area ( however , see remarks regarding excess graphene in section [ sec : theory_c_4seg_asym ] ) . to be able to account for electrode mask asymmetry and misalignment ,
we model the coupling as two capacitors in series : the capacitance between the positive electrode segments with the membrane in series with that between the negative electrodes and the membrane ( note that @xmath5 is much smaller than the gap between fixed electrodes ) .
this leads to the following generalization of the overlap factor @xmath20 , eq .
( [ eq : eta_sym ] ) , for the quarter - segment geometry @xcite : @xmath81,\ ] ] where we have introduced the symbol @xmath82 with @xmath83 and @xmath84 being the area of the membrane above electrodes of polarity @xmath85 . the terms in eq .
( [ eq : eta_general_4seg ] ) can be interpreted as follows : the first ( second ) term represents the frequency shift contribution from the membrane area above positive ( negative ) electrodes with the local potential difference fixed at its equilibrium value , whereas the third term is a correction due to the modulations of the membrane potential around the equilibrium value .
we use eq .
( [ eq : eta_general_4seg ] ) in calculating experimental @xmath12 values in order to account for gaps , edges and a slight asymmetry in the electrode mask as well as the central hole in the al layer of the sin - al membranes .
the formula was also used to estimate the uncertainty in @xmath12 due to lateral misalignment of the membrane wrt .
the electrodes ( estimated to be up to 25% of the membrane side length ) .
the graphene on the sin - g membranes was characterized by raman spectroscopy .
the graphene flakes are covering the entire sin membranes plus a large part of the surrounding frame . in fig .
[ fig : raman ] the raman spectra of a sin - g membrane chip is shown .
the graphene spectra clearly show the typical spectrum of single layer graphene with a 2d - peak having double the amplitude of the g - peak . in all 22 point measurements we made the d - peak at 1300 - 1400 @xmath86 was missing which indicates that the graphene layers is of good quality with probably only few defect .
in the case of charging , a linear force adds up to the quadratic electrostatic force .
thus , a linear term has to be added to the relative frequency shift as a function of dc voltage ( [ eq : softening ] ) @xmath87 fig .
[ fig : chargingsi]a shows the charging induced hysteresis recorded during the charging experiment presented in fig .
4a in the main document . in the experiment , the dc voltage was swept 4 times between 0 and + 30 v. from sweep to sweep the spring softening becomes continuously stronger .
this is a clear sign for the accumulation of charge on the bare sin membrane . in the 5th sweep ,
the voltage is swept from 0 to -30 v and a spring hardening is observed . in fig .
[ fig : charging]b , the initial sweep is fitted with the charge free model ( [ eq : softening ] ) and the last continuous sweeps 4 + 5 are fitted with ( [ eq : softeningwithcharge ] ) which contributes for a linear force from free charge on the membrane .
the fit is of good quality and strongly supports the hypothesis of charge accumulation on the membrane .
the force constant of the linear force due to free charge is one order of magnitude larger than the force constant of the electrostatic force .
hence , free charge on the membrane significantly contributes to the electromechanical membrane behavior especially at low dc voltages . 1 mm@xmath8 , d = 4.0 @xmath4 m ) on interdigitated electrodes ; electrode fingers are 4 @xmath4 m wide with a gap of 2 @xmath4 m between the fingers .
the voltage was swept 4 times between 0 to + 30 v and backwards . in the end
the dc voltage was swept from 0 to -30 v. b ) fig .
[ fig : charging]b shows the fitting of the initial voltage sweep 1 with ( [ eq : softening ] ) and the last sweep 4 + 5 with ( [ eq : softeningwithcharge ] ) .
the fitting parameters are for sweep 1 : @xmath88 v@xmath8 , and for sweep 4 + 5 : @xmath89 v@xmath90 while @xmath16 was kept constant at @xmath88 v@xmath8.,scaledwidth=70.0% ] | due to their exceptional mechanical and optical properties , dielectric silicon nitride ( sin ) micromembrane resonators have become the centerpiece of many optomechanical experiments .
efficient capacitive coupling of the membrane to an electrical system would facilitate exciting hybrid optoelectromechanical devices .
however , capacitive coupling of such dielectric membranes is rather weak . here
we add a single layer of graphene on sin micromembranes and compare electromechanical coupling and mechanical properties to bare dielectric membranes and to membranes metallized with an aluminium layer .
the electrostatic coupling of graphene coated membranes is found to be equal to a perfectly conductive membrane .
our results show that a single layer of graphene substantially enhances the electromechanical capacitive coupling without significantly adding mass , decreasing the superior mechanical quality factor or affecting the optical properties of sin micromembrane resonators .
hybrid devices capable of coupling different systems are presently one of the hot topics in quantum technologies @xcite .
recent proposals @xcite outline optoelectromechanical systems , where a mechanical resonator is strongly coupled to an optical and an electrical resonator at the same time .
the centerpiece of these proposals is a mechanical micromembrane inside an optical resonator which is capacitively coupled to an lc circuit @xcite .
recently , such a hybrid optoelectromechanical system has been realized based on sin - al membranes @xcite . for these applications
it is essential to have a strong electromechanical coupling relative to the mechanical properties of the membrane .
a key feature of sin micromembranes is their ultrahigh quality factor ( q ) reaching @xmath0 , low mass and excellent optical transparency with losses less than @xmath1 in the near infrared @xcite . in order to not downgrade these essential properties
it was proposed to use dielectric polarization forces on bare sin membranes for the electromechanical coupling @xcite .
such forces on a dielectric are however , inherently weaker than the forces on conductors . in this work ,
we deposit a single layer of graphene onto sin membranes ( sin - g ) @xcite and compare their mechanical properties and electromechanical coupling to bare sin membranes and to membranes covered with an aluminium layer ( sin - al ) @xcite .
we show that this single layer of graphene allows superior electromechanical coupling without deteriorating the exceptional properties of sin membranes .
importantly , our setup uses graphene in a floating electrode configuration .
thereby graphene is not in contact with a metal and the interaction happens electrostatically . in this fashion
we avoid the large contact resistance associated with connecting graphene and a metal electrode @xcite .
we used commercial high and low - stress 50 nm thick si@xmath2n@xmath3 membranes ( norcada inc . ) for both the sin and sin - g resonators .
single layer graphene was grown on copper foil using standard cvd techniques @xcite .
the graphene on copper was cut to size , the copper wet - etched and the graphene transferred to the surface of the membrane in the aqueous environment .
a thin layer of pmma was used to support the graphene during the etching and transfer process .
once dry , the pmma layer was removed from the graphene using acetone vapor .
the final structure was robust enough for subsequent fabrication steps , for instance for making a round opening window in the graphene for future optical cavity applications or lithographically defining gates for control of the carrier density and type .
in the experiments described below we used membranes without the opening in the graphene layer . a characterization by raman spectroscopy of the single layer graphene on sin - g chips is presented in the supplementary information .
the al covered membranes were fabricated in - house by standard cleanroom processing .
the high - stress stoichiometric sin layer is 100 nm thick .
the aluminium layer is 50 nm thick and it is patterned by a lift - off step . the al layer could also be fabricated with a round hole in the center of the membrane for optical access . a small rim ( 5% of the membrane size ) along the anchor of the membrane was spared out in order to minimize damping @xcite .
two types of coplanar electrode chips were used : interdigitated electrodes and quarter - segment electrodes .
schematic drawings of two types of electrodes are depicted in fig .
[ fig : fig1 ] . the electrodes fabricated by standard cleanroom processing
are made of a 200 nm thick gold layer sitting on borosilicate glass substrate or sin covered silicon substrate .
the membranes are placed membrane downwards onto the electrode .
the electrode chips feature pillars with a height of 600 nm and 1 @xmath4 in order to define a small gap between membrane and electrodes .
optical measurements of the gap distance @xmath5 and the membrane vibrations have been made with a white light interferometer ( vibrometer msa-500 from polytec gmbh ) .
the gap distance ranged from 3.5 to 14 @xmath4 m which is larger than the height of the dedicated pillars .
the larger measured distance can be ascribed to electrode chip unevenness coming from dirt and chip stress gradients and has been reduced down to the pillar height size in subsequent experiments . the experimental setup is schematically depicted in fig .
[ fig : fig1 ] .
the membrane - electrode sandwich is placed in a vacuum chamber ( pressure below @xmath6 mbar ) and is electrically connected to an external voltage source . for quality factor measurements ,
the membranes were placed on a piezo for stimulation .
the quality factors were extracted from the membrane ring - down time and the -3db bandwidth of the resonance peak .
a key feature of micromechanical membranes is their q of up to several million .
such high quality factors are a requirement for strong optomechanical coupling and for high resolution measurements . in the first experiment we compared the quality factors of the 3 different types of membranes .
fig .
[ fig : fig2 ] shows the measured quality factors of a bare sin , sin - al , and sin - g membrane . on the sin - al membrane a rim of 5% of the membrane dimension
remained uncoated , whereas graphene is fully covering the entire sin - g membrane .
the quality factors are highly mode dependent and the measured values are in correspondence with values measured by yu et al .
@xcite .
there are clearly two sets of q - values . according to @xcite , the lower set ( below @xmath7 )
is limited by clamping losses .
the higher set is limited by intrinsic damping , such as bulk or surface losses .
it can be seen from the measurements that there is no significant difference between the different membrane types in both sets .
the graphene sheet seems to be mechanically invisible and it does not significantly contribute to the energy loss .
thus with graphene it is not required to spare out the membrane area close to the clamping edge as it is for metal layers .
the additional metal layer of the sin - al membrane downshifts the resonance frequencies compared to the other two membranes , as can be clearly seen in fig .
[ fig : fig2 ] for low mode numbers .
this divergence increases for higher mode numbers , which makes a direct comparison difficult . 0.5 mm@xmath8 and are all made of high - stress stoichiometric sin.,scaledwidth=45.0% ] it is crucial for an efficient hybrid optoelectromechanical device to have strong electromechanical coupling between the membrane and the electrical circuit .
the electrostatic interaction for sin membranes is due to dielectric polarization forces ( the electric field interacts with a dielectric ) @xcite and in some of the experiments presented here it is also due to the quasi - permanent electric charges in a dielectric .
the interdigitated electrodes have been design to generate strong electric field gradients that are required for the dielectric polarization force .
the interaction for conductive sin - al @xcite is due to the electrostatic force between conductors ( the electric field interacts with charges on a conductor ) . for this case
the quarter - segment electrodes have been designed .
using the two electrode geometries allows us to characterize the comparative performance of both types of membranes .
it also allows us to conclude to which type the sin - g membranes belong . neglecting the effect of free charges on the membrane
, the electrostatic force between a dielectric or conductive thin membrane and electrodes can generally be described by @xmath9 with @xmath10 - the membrane area , @xmath11 - the potential difference over the electrodes , @xmath12 - the electrostatic force constant that characterizes the coupling performance , and @xmath13 - a function describing the distance dependence of the force .
we extract the intrinsic coupling strength @xmath12 of each membrane type from measuring the mechanical frequency shift due to the so - called spring softening effect which results in a quadratic frequency drop with the dc - bias voltage @xmath14 .
the relative frequency shift is given by ( see supplementary information ) @xmath15 using @xmath16 found from the frequency shift data ( [ eq : springsoftening ] ) , the electrostatic force constant @xmath12 can be calculated from @xmath17^{-1 } h \rho\omega_0 ^ 2 \eta^{-1}\ ] ] with the membrane thickness @xmath18 and the membrane mass density @xmath19 ; while @xmath20 ( of order unity ) quantifies the spatial overlap between the membrane mode shape and the fixed electrodes , in particular compensating for the gap between quarter - segment electrodes and for the circular hole in the sin - al membranes center ( see supplementary information ) . here
the coupling performance of sin - g membranes is compared to that of the sin and sin - al membranes using the two types of capacitor+membrane chips described above . in the first set , sin and sin - g membranes
are coupled to interdigitated electrodes ; in the second set , sin - al and sin - g membranes are coupled to quarter - segment electrodes . for the interdigitated electrode geometry and the range of membrane - capacitor distances @xmath5 used in the experiments , the electrostatic force for both dielectric and perfectly conducting membranes ( an appropriate model for our sin - g membranes )
is well approximated by @xmath21 with @xmath22 - a scaling constant with units of area .
for the data presented below , we use @xmath23 @xmath4m@xmath8 for our specific setup , we determined numerically that @xmath24 for both dielectric and perfectly conducting membranes . in fig .
[ fig : c]a the electrostatic spring softening of a sin - g membrane on top of an interdigitated electrode is shown . from the fit parameter @xmath16 obtained from this data the force constant @xmath12 is calculated for the membranes using eq .
( [ eq : forceconstant ] ) and the exponential force law ( [ eq : forcelaw - inter ] ) . the resulting average force constants for the bare sin and sin - g membranes on interdigitated electrodes
are shown in the left part of fig .
[ fig : c]c .
sin - g is seen to outperform sin by a factor of @xmath25 .
also shown are the theoretical values determined semi - analytically for dielectric ( sin , @xmath26 ) and perfectly conducting membranes ; the average force constant of sin - g is seen to be compatible with the latter . .
a ) resonance frequency of fundamental mode of a @xmath27 mm@xmath8 sin - g membrane on interdigitated electrodes as a function of dc voltage .
the distance between membrane and electrodes is @xmath28 @xmath4 m .
b ) resonance frequency fundamental mode of a @xmath27 mm@xmath8 sin - g membrane on quarter - segment electrodes as a function of dc voltage .
the distance between the membrane and electrodes is @xmath29 @xmath4 m .
c ) comparison of force constants @xmath12 ( according to ( [ eq : forceconstant ] ) using ( [ eq : forcelaw - inter ] ) and ( [ eq : forcelaw - seg ] ) ) for different combinations of electrodes and membranes . on the left , @xmath12 extracted for bare sin ( 4 experiments ) and sin - g ( 3 experiments ) membranes on interdigitated electrodes are shown .
the electrode fingers are 4 @xmath4 m wide with a gap of 2 @xmath4 m between the fingers . to the right , @xmath12 of sin - al ( 3 experiments ) and sin - g ( 4 experiments ) on quarter - segment electrodes
are shown , with a gap between the segments of 60 @xmath4 m .
the error bars represent the standard deviation of the conducted experiments .
the solid and dashed grey lines represents the theoretical values for the scenario of a pure dielectric polarization force and an electrostatic force on the conductive membrane , respectively.,scaledwidth=45.0% ] next , we use the quarter - segment coplanar electrodes with a sin - al or a sin - g membrane as a floating electrode .
the results of this set of experiments are well described by another force law ( in the regime where @xmath5 is smaller than the inter - electrode gap ) : @xmath30 the measured spring softening of a sin - g membrane placed over quarter segment electrodes is shown in fig .
[ fig : c]b . from the fit parameter @xmath16 obtained from this data the force constant @xmath12 is calculated for the conducting membranes again with eq .
( [ eq : forceconstant ] ) but now using the quadratic force law ( [ eq : forcelaw - seg ] ) . on the right side of fig .
[ fig : c]c the extracted average force constants of sin - g and sin - al membranes are shown .
the experimental values again agree well with the theoretical value for a perfect conductor .
the large standard deviation of the force constant can be assigned to uncertainties in the distance @xmath5 measurement and lateral misalignment which can contribute up to 20% error .
the relatively large average @xmath12 value of the sin - g membranes might be an effect due to the excess graphene on the frame , as discussed in the supplementary information .
mm@xmath8 , @xmath31 m ) on interdigitated electrodes ; electrode fingers are 4 @xmath4 m wide with a gap of 2 @xmath4 m between the fingers .
b ) response of a sin - g membrane ( @xmath27 mm@xmath8 , @xmath32 m ) on interdigitated electrodes ; the electrode fingers are 4 @xmath4 m wide with a gap of 5 @xmath4 m between the fingers .
c ) response of a sin - al membrane ( @xmath33 mm@xmath8 , @xmath34 @xmath4 m ) on quarter - segment electrodes . the resonance frequency was determined from the thermomechanical resonance peak.,scaledwidth=50.0% ] an additional complication with sin membranes which is eliminated by using sin - al or sin - g membranes is the accumulation of free charges on the dielectric sin membrane .
the charging has been investigated by increasing the dc voltage and then inverting the polarity . if quasi - static charges are present , a spring hardening ( increase in oscillation frequency ) instead of softening will be observed .
the results of this experiment are shown in fig .
[ fig : charging ] with a bare sin , a sin - al and a sin - g membrane on interdigitated electrodes . from the observed spring hardening , shown in fig .
[ fig : charging]a , it can be concluded that free electric charges are indeed available on the bare sin membranes .
we note that the force on a charge is a function of the inverse of the distance ( @xmath35 ) and may therefore be partly responsible for the somewhat higher value of the coupling constant c for sin shown in fig .
[ fig : c]c .
the charging of the bare sin membrane is investigated in more detail in the supplementary information .
fig .
[ fig : charging]b and fig .
[ fig : charging]c present the results of the same charging experiment conducted with a sin - g on similar interdigitated electrodes and with sin - al membranes on quarter - segment electrodes , respectively .
assuming that the electrode design does not influence the charging effect , we conclude that the absence of spring hardening for sin - g shows that the single layer of graphene as well as the al layer eliminates charging .
mm@xmath8 sin - g membrane on interdigitated electrodes .
response of a ) resonance peak amplitude and b ) quasi - static displacement to dc bias voltage .
the distance between membrane and electrodes is @xmath36 @xmath4 m .
the electrode fingers are 4 @xmath4 m wide with a gap of 5 @xmath4 m between the fingers .
the resonance peak amplitude was measured for a white noise signal of @xmath37 mv .
the quasi - static displacement was measured with a rectangular signal at 10 khz ( which is far below the fundamental resonance frequency and can therefore be considered to be quasi - static ) .
the error bars represent the standard deviation of 3 consecutive measurements .
the grey lines represent a linear fit in a ) and a quadratic fit in b).,scaledwidth=45.0% ] we performed further characterization of the sin - g membranes by studying the dependence of the ac amplitude and the dc displacement on the coupling voltage ( fig .
[ fig : sin - g ] ) .
if the ac driving voltage is small compared to the dc bias voltage , the electrostatic force becomes @xmath38 thus the mechanical resonance peak amplitude is approximately a linear function of the dc voltage , for a fixed ac driving voltage , and the static deflection of the membrane is a quadratic function of the dc voltage . as seen from fig .
[ fig : sin - g ] the experimentally observed behavior supports the model .
in conclusion we have shown that sin - g membranes with a single layer of graphene are promising canditates for efficient optoelectromechanical hybrid coupling devices .
furthermore , by using the graphene sheet on sin in a floating electrode configuration , we overcome the unsolved problem of high contact resistance between metal electrodes and graphene .
they show electromechanical coupling which is as good as for an ideal conductor .
the enhanced electromechanical coupling of conductive membranes lowers the threshold for strong coupling , as predicted for dielectric membranes @xcite , accordingly .
the single graphene layer is mechanically invisible and does not negatively influence the membrane performance . unlike the real metal coating , graphene does not require patterning in order to maintain the highest mechanical quality factors of a bare sin membrane . also , unlike the metal coating , graphene does not add any noticeable mass to the membrane which is an advantage for high sensitivity applications , such as , e.g. , the optical detection of radio waves with metallized silicon nitride membrane resonators @xcite .
the lower mass achieved by sin - g membranes as compared to sin - al membranes used in the reference would improve the quantum efficiency of rf - to - optical conversion by up to 4 times ( by the ratio of the masses ) .
finally , the use of graphene overcomes the complications of charging effects in sin membranes , which has shown to be difficult to control .
we acknowledge funding from the darpa program quasar , the eu project qessence and the erc projects interface and qios ( under grant agreement n. 306576 ) .
the research leading to these results has received funding from the european community s seventh framework programme ( fp7/2007 - 2013 ) under grant agreement n211464 - 2 and it was further supported by the villum kann rasmussen centre of excellence `` namec '' under contract no .
65286 .
we would like to thank louise jrgensen for cleanroom support and martin benjamin b. s. larsen for support with the raman measurements .
10 j. d. teufel , d. li , m. s. allman , k. cicak , a. j. sirois , j. d. whittaker , and r. w. simmonds , nature * 471 * , 204 ( 2011 ) .
e. gavartin , p. verlot , and t. j. kippenberg , nat .
nanotech . * 7 * , 509 ( 2012 ) . j. m. taylor , a. s. srensen , c. m marcus , and e. s. polzik , phys .
rev .
lett . * 107 * , 273601 ( 2011 ) . c. a. regal and k. w. lehnert , j. phys .
. ser . * 264 * , 012025 ( 2011 ) .
t. bagci , a. simonsen , s. schmid , l. g. villanueva , e. zeuthen , j. appel , j. m. taylor , a. srensen , k. usami , a. schliesser , and e. s. polzik , arxiv:*1307.3467 * [ physics.optics ] , ( 2013 ) .
j. d. thompson , b. m. zwickl , a. m. jayich , f. marquardt , s. m. girvin , and j. g. e. harris , nature * 452 * , 72 ( 2008 ) .
d. j. wilson , c. a. regal , s. b. papp , and h. j. kimble , phys .
rev .
lett . * 103 * , 207204 ( 2009 ) .
t. j. kippenberg and k. j. vahala , opt .
express * 15 * , 17172 ( 2007 ) .
s. lee , v. p. adiga , r. a. barton , a. van der zande , g .- h .
lee , b. r. ilic , a. gondarenko , j. m. parpia , h. g. craighead , and j. hone , arxiv:*1305.4852 * [ cond-mat.mtrl-sci ] , ( 2013 ) .
p .-
l .
yu , t. p. purdy , and c. a. regal , phys . rev .
lett . * 108 * , 083603 ( 2012 ) .
s. russo , m. f. craciun , m. yamamoto , a. f. morpurgo , and s. tarucha , physica e * 42 * , 677 ( 2010 ) .
a. venugopal , l. colombo , and e. m. vogel , appl .
. lett . * 96 * , 013512 ( 2010 ) .
k. s. novoselov , v. i. falko , l. colombo , p. r. gellert , m. g. schwab , and k. kim , nature * 490 * , 192 ( 2012 ) .
x. li , w. cai , j. an , s. kim , j. nah , d. yang , r. piner , a. velamakanni , i. jung , e. tutuc , s. k. banerjee , l. colombo , and r. s. ruoff , science * 324 * , 1312 ( 2009 ) .
s. schmid , k. d. jensen , k. h. nielsen , and a. boisen , phys . rev .
b * 84 * , 165307 ( 2011 ) .
s. schmid , m. wendlandt , d. junker , and c. hierold , appl .
. lett . * 89 * , 163506 ( 2006 ) .
s. schmid , c. hierold , and a. boisen , j. appl .
phys . * 107 * , 054510 ( 2010 ) .
q. p. unterreithmeier , e. m. weig , and j. p. kotthaus , nature * 458 * , 1001 ( 2009 ) . |
low - energy compton scattering off the nucleon is traditionally used to access the _
nucleon polarisabilities _ , but the relation between the observables and the polarisabilities is not straightforward ; see refs .
@xcite for recent reviews .
a direct relation only exists at the level of the low - energy expansion ( lex ) , which is a polynomial expansion in the photon energy @xmath15 .
the lex validity is limited to very low energies , well below the pion production threshold ( @xmath16 ) .
most of the experimental data are , however , obtained at energies above 100 mev .
chiral perturbation theory ( @xmath17pt ) @xcite in the single - nucleon sector @xcite , as a low - energy effective - field theory of qcd , can provide the necessary model - independent link between the nucleon compton - scattering data and polarisabilities beyond the lex applicability . in recent years
@xmath17pt calculations have been performed using both the heavy - baryon ( hb@xmath17pt ) and manifestly - lorentz - covariant ( b@xmath17pt ) formulations ; see respectively @xcite and @xcite ; some details of the differences between the two approaches can be found in @xcite .
these calculations are not only useful for precision determination of the polarisabilities from experiment they also provide a testing ground for @xmath17pt in the single nucleon sector.pt is contained in refs .
@xcite , and in a chiral framework in @xcite . ] in this paper we follow up on the manifestly - covariant calculation of ref .
@xcite , where the next - to - next - to - leading order ( nnlo ) calculation of proton compton scattering was carried out in su(2 ) b@xmath17pt with pion , nucleon and delta(1232 ) degrees of freedom .
this nnlo calculation provides a prediction for all the nucleon polarisabilities and scattering observables ; however only the proton scalar dipole polarisabilities and the unpolarised differential cross sections were examined in ref .
@xcite . in the present paper
we consider a number of the other predictions .
we correct the numerical results for the dipole polarisabilities of the proton and present the corresponding results for the neutron .
we present the nnlo results , both analytically and numerically , for the quadrupole polarisabilities and the four dipole spin polarisabilities .
we also extend the calculation of @xcite to the delta(1232)-resonance region and present results for the compton multipoles and dynamical polarisabilities . in doing so
, we shall remark on a subtlety in the matching of the multipole expansion and static polarisabilities .
finally , we shall consider the b@xmath17pt predictions for the polarised proton observables , namely the asymmetries @xmath18 , @xmath19 , @xmath9 , and @xmath10 .
these results should be of particular interest to the ongoing compton - scattering experiments at mami ( mainz ) and higs ( duke ) facilities , which aim to determine the scalar and spin polarisabilities of the proton and neutron in polarised measurements @xcite .
the paper is structured as follows . in sect .
[ sec : amp ] we recall the main ingredients of the nnlo b@xmath17pt calculation of compton scattering . in sect .
[ sec : amp : tensor ] we introduce various parametrisations of the compton amplitude and establish relations between them .
we provide the predictions of b@xmath17pt for the nucleon static polarisabilities in sect .
[ sec : pols : static ] , and the compton multipoles and dynamical polarisabilities in sect .
[ sec : pols : dyn ] , and compare them with various empirical and theoretical results .
section [ sec : obs ] presents our predictions for some of the polarised observables , compared with experimental data where possible .
we conclude with sect . [
sec : concl ] .
following refs .
@xcite , where one can find details such as the relevant @xmath17pt lagrangians , we consider low - energy compton scattering in b@xmath20pt , i.e. , a manifestly - covariant formulation of @xmath17pt with pion , nucleon and delta(1232 ) isobar degrees of freedom ; see ref .
4 ) for review .
our present calculation of the compton amplitude follows that of ref .
@xcite below photoproduction threshold ( and hence for the static polarisabilities ) , but improves the treatment of the delta - excitation near the resonance , as described below .
our eft expansion uses the @xmath21-counting @xcite , where the mass difference between the nucleon and the delta , @xmath22 , is considered an intermediate scale , so that m_// _ , and hence the usual chiral expansion scale @xmath23 is counted as @xmath24 .
an important feature of the @xmath21-counting is that the characteristic momentum @xmath25 distinguishes two regimes : _ low energy : _
@xmath26 , and _ delta - resonance : _ @xmath27 .
since the delta propagators go as @xmath28 , rather than simply @xmath29 as for the nucleon , the counting of the graphs with deltas is different in the two regimes .
for example , replacing a nucleon propagator by a delta changes the power of the graph by @xmath30 , which is @xmath31 in the low - energy regime and @xmath32 in the resonance region .
in addition , the counting demands that the one - delta - reducible ( @xmath33r ) loop graphs , which contain the pole at @xmath34 , are resummed in the regime @xmath35 , which gives rise to the delta - resonance width and results in a natural description of the resonance peak . in general , a graph with @xmath36 loops , @xmath37 pion propagators , @xmath38 nucleon propagators , @xmath39 delta propagators and @xmath40 vertices of order @xmath41 counts in the low - energy regime as @xmath42 , with n = 4l - 2n_-n_n -n_+ k v_k . in the resonance regime , one identifies the number of the @xmath33r propagators , @xmath43 , then n = 4l - 2n_-n_n - n_- 2 n_1 + k v_k . in the low - energy regime , which is most relevant for the extraction of nucleon polarisabilities , the diagrams with @xmath39 delta propagators are suppressed by @xmath44 , or equivalently by @xmath45 , with respect to the same diagram but with nucleon propagators .
the leading order ( lo ) contribution in this regime hence comes from the nucleon born term ( first two diagrams in ) which are responsible , for instance , for the correct thomson limit .
inclusion of the nucleon anomalous magnetic moment in these graphs is warranted at @xmath46 , which here is the next - to - leading order ( nlo ) . also of @xmath46
is the @xmath47-pole graph in , and all of the pion - nucleon loops in fig .
[ fig : pinloops ] .
these graphs give the lo contribution to the polarisabilities , as the born graphs do not contribute to polarisabilities by definition .
the delta - pole and pion - delta loop graphs , fig .
[ fig : pidloops ] , enter at @xmath48 , here the next - to - next - to - leading order ( nnlo ) , and give the nlo contribution to polarisabilities .
we note that the delta - pole term , despite being formally suppressed compared to the pion - nucleon loops , has long been known to give a large contribution to the magnetic polarisabilities of the nucleon .
the nnlo amplitude thus receives contributions from the nucleon born and @xmath47 pole graphs of fig .
[ fig : born ] , the @xmath49 loop graphs of fig .
[ fig : pinloops ] and the delta graphs in fig .
[ fig : pidloops ] ; all of these were computed in @xcite . in order to extend this calculation to the delta - resonance region ,
we resum the @xmath50 loop contributions to the delta propagator in the delta - pole graphs , as explained in ref .
in addition we take into account the leading pion - loop corrections to the @xmath51n@xmath52 vertex , shown in fig . [
fig : vpinloops ] .
those additions are important for unitarity above the pion threshold . at nlo in the regime @xmath35 ,
only the imaginary part of these vertex loops is relevant ( specific remarks on the real part are given in sect . )
; this means that these corrections vanish below the pion production threshold and hence do not affect the static polarisabilities .
we will , however , investigate their importance for the dynamical polarisabilities and compton multipoles ; see sect .
[ sec : pols : dyn ] .
the technical details of the counting and the evaluation of these @xmath53-vertex corrections are found in refs .
@xcite .
n@xmath52 vertex with nlo loop corrections .
only the imaginary parts of the loops shown here lead to contributions at next - to - leading order around the delta peak . ]
the effects of running of the @xmath51n@xmath52 coupling constants @xmath54 and @xmath55 , arising due to the real parts of the @xmath51n@xmath52 vertex corrections fig .
[ fig : vpinloops ] , are beyond the order we are working at in the delta region @xcite .
nevertheless , we briefly consider the potential impact of these effects , together with the effects of the running of the delta mass and field renormalisation constants .
none of these effects are included in subsequent section ; they are also not the only higher - order effects which could impact the polarisabilities . of the latter
, the inclusion of the anomalous magnetic moments of the proton and neutron within @xmath50 loop graphs may well be significant @xcite .
the running of the real part of the @xmath56 couplings from the graphs of fig .
[ fig : vpinloops ] was calculated in @xcite .
the corresponding numerical results are shown in the left panel of fig .
[ fig : running ] . both coupling
are seen to be reduced , in absolute value , at low energy , and hence the delta - pole contribution to the polarisabilities will be reduced .
n@xmath52 coupling constants @xmath54 and @xmath55 ( left panel , solid / dashed line , respectively ) , of the delta mass ( centre panel ) , and of the field renormalisation constant ( right panel ) , as functions of @xmath57 in gev .
grid lines represent the values renormalised at the delta peak and the corresponding value of @xmath57 . ] on the other hand , this effect is partially compensated by the one - loop running of the delta propagator , which reads : s_(p)=-^3/2_(p ) , where @xmath58 is the spin-@xmath59 projection operator , @xmath60 is the running delta mass , and @xmath61 is the width of the delta .
the renormalised self - energy and field renormalisation are calculated as shown in appendix [ sec : remarks - details ] .
the resulting @xmath60 and @xmath62 are shown fig .
[ fig : running ] and are seen thus to strengthen the delta - pole effects at lower energies .
a complete @xmath63 b@xmath17pt calculation will need to be done to correctly quantify the effect of these contributions .
this calculation is , however , beyond the scope of the present work . with these ingredients , our calculation of the amplitudes is nnlo [ @xmath48 ] at low energies and nlo [ @xmath65 in the delta resonance region , as summarised in table [ tab : counting ] .
0.7|c|y|y|y|y|y|y| & @xmath66 pole & @xmath50 loops & @xmath67 loops & @xmath52 pole & @xmath52-pole corr .
+ @xmath68 & @xmath69 & @xmath70 & @xmath71 & @xmath71 & @xmath72 + @xmath73 & @xmath69 & @xmath70 & @xmath74 & @xmath75 & @xmath76 + the numerical values of the physical constants used in the present calculation are given in table [ tab : constants ] .
they are only marginally different from the ones used in @xcite , and the same as in @xcite .
.parameters used in our calculation ( the @xmath47 mass is only used for the computation of the @xmath77-channel pion - pole graph ) .
most of the values are from particle data group @xcite .
the @xmath78n@xmath52 coupling constant @xmath79 is fit to the the experimental delta width and the magnetic and electric @xmath51n@xmath52 coupling constants @xmath54 and @xmath55 are taken from the pion photoproduction study of ref .
more details can be found in ref .
[ cols= " < , < , < " , ]
here we collect the analytic expressions for the various contributions computed in this work .
the corresponding numerical results are shown in tables [ tab : polsscalar][tab : polsspin_n ] .
the spin polarisabilities @xmath80@xmath81 provided here are related to those defined in sect .
[ sec : pols : static : def ] by @xmath82 the complete @xmath83 chiral - loop contribution to the dipole , spin and quadrupole polarisabilities of the proton are ( with @xmath84 ) : _ e1^()= & + & , _ m1^()=- & + & , _ 1^()= & + & , _ 2^()= & + & , _ 3^()=- & + & , _ 4^()= & + & , _ e2^()= & -_m1^()+ + & , _ m2^()=&- _
e1^()+ + & .
the complete @xmath83 chiral - loop contribution to the dipole , spin and quadrupole polarisabilities of the neutron are ( with @xmath84 ) : _ e1^(,n)= & + & , _ m1^(,n)=- & + & , _ 1^(,n)= & + & , _ 2^(,n)= & + & , _ 3^(,n)=&-
+ & , _ 4^(,n)= & + & , _ e2^(,n)= & -_m1^(,n ) - + & , _ m2^(,n)=&- _ e1^(,n)+ + & .
our expressions for the @xmath67-loop contributions are sufficiently compact when we leave one of the integrations over the feynman parameters undone . the integrands involve the following function of the feynman parameter @xmath85 : d_(x)=x^2 + ( 1 + ) ^2 - x ( 2 + 2 + ^2 - ^2 ) , where @xmath84 , @xmath86 .
furthermore , the divergent parts of the polarisabilities , absorbed by higher - order contact terms , are renormalised according to the modified minimal subtraction ( @xmath87 ) scheme , by setting to 0 the factor arising in the dimensional regularisation : = -_e+ , with @xmath88 the number of dimensions , @xmath89 the euler constant , and @xmath90 the renormalization scale . with these definitions
we obtain : and @xmath91 .
the expressions for the other polarisabilities appear for the first time . ]
_ e1^()= , _
m1^()= , _ 1^()= , _ 2^()= , _
3^()= , _
4^()= , _ e2^()= & + - , _ m2^()= & + - .
introducing @xmath92 and recalling that @xmath93 , the contribution of the delta(1232 ) electromagnetic excitation to the dipole , spin and quadrupole polarisabilities of the proton are , respectively : _
e1^()&=&- , + _ m1^()&= & , _ 1^ ( ) & = & ( - + ) , + _ 2^ ( ) & = & -(+ ) , + _ 3^ ( ) & = & - ( - ) , + _ 4^ ( ) & = & ( -+ ) , _
e2^()&=&-_m1^()+ ( - + ) , + _
m2^()&=&-_e1^()- ( -+ ) .
for the benefit of future workers , we here provide the details of the corrections to the delta self - energy which enter beyond the order to which we work in this paper , as discussed in section [ sec : remarks ] .
we write the delta self - energy in the form : ( ) = ( s)+(s)(-m _ ) , where the one - loop contributions with the intermediate nucleon or delta are given by @xcite @xmath94\,,\\ \tau(s)_{n,\delta}&=-\frac{c_{n,\delta}}{2(8\pi f_\pi)^2}\int\limits_0 ^ 1 dx\ , x\mathcal{m}^2_{n,\delta}\left[-\xi-1+\log\mathcal{m}^2_{n,\delta}\right]\ , , \end{split}\ ] ] with @xmath95 the constant representing the ultraviolet divergence , @xmath96 , @xmath97 , and @xmath98 for the @xmath67 axial coupling we adopt the large-@xmath99 value @xmath100 .
the renormalised functions @xmath101 and @xmath102 are obtained after two and one subtractions , respectively : _
r(s)=(s)-(m_^2)-(s - m_^2)(m_^2),_r(s)=(s)-(m_^2 ) .
the running delta mass is then given by @xmath103 , while the delta width is @xmath104 .
while @xmath101 is finite , @xmath102 still contains a divergent piece which is to be renormalized by an appropriate low - energy constant . in the absence of further information on such a constant , we neglect it and merely set @xmath105 .
note that the expressions for @xmath106 and @xmath107 above do not contain the extra factor of @xmath108 seen in ref .
this is due to the fact that each of the leading @xmath51n@xmath52 vertices in the first graph in fig .
[ fig : pidloops ] brings in an extra delta momentum giving a total factor of @xmath109 ( see ref .
when the self energy graphs are resummed in the intermediate state , the factor @xmath108 is taken out from the self energy in order to be restored in front of the total result , hence this factor is absent in the expressions we use here .
we also give here the dispersion integral that simplifies the calculation of the running couplings @xmath110 .
namely , the running due to the leading loop corrections can be calculated from the already employed imaginary parts of these corrections by the once - subtracted integral g_m , e(s ) = g_m , e(m_^2 ) + ( s - m_^2)_(m_n+m_)^2^ , where the integral is understood as the cauchy principal value ; note the explicit @xmath111 factor .
h. w. griehammer , j. a. mcgovern , d. r. phillips and g. feldman , prog . part .
phys . * 67 * , 841 ( 2012 ) .
http://arxiv.org/abs/1203.6834[arxiv:1203.6834 ] [ nucl - th ] .
b. r. holstein and s. scherer , ann . rev .
nucl . part .
sci . * 64 * , 51 ( 2014 ) .
http://arxiv.org/abs/1401.0140[arxiv:1401.0140 ] [ hep - ph ] .
f. hagelstein , r. miskimen and v. pascalutsa , http://arxiv.org/abs/1512.03765[arxiv:1512.03765 ] [ nucl - th ] , to appear in prog .
s. weinberg , physica a * 96 * , 327 ( 1979 ) .
j. gasser and h. leutwyler , annals phys .
* 158 * , 142 ( 1984 ) .
j. gasser , m. e. sainio and a. svarc , nucl .
b * 307 * , 779 ( 1988 ) .
j. a. mcgovern , phys .
c * 63 * , 064608 ( 2001 ) [ erratum - ibid .
c * 66 * , 039902 ( 2002 ) ] .
http://arxiv.org/abs/nucl-th/0101057[arxiv:nucl-th/0101057 ] .
s. r. beane , m. malheiro , j. a. mcgovern , d. r. phillips and u. van kolck , phys .
b * 567 * , 200 ( 2003 ) [ erratum - ibid .
b * 607 * , 320 ( 2005 ) ] .
http://arxiv.org/abs/nucl-th/0209002[arxiv:nucl-th/0209002 ] .
s. r. beane , m. malheiro , j. a. mcgovern , d. r. phillips and u. van kolck , nucl .
a * 747 * , 311 ( 2005 ) .
http://arxiv.org/abs/nucl-th/0403088[arxiv:nucl-th/0403088 ] .
v. pascalutsa and d. r. phillips , phys .
c * 67 * , 055202 ( 2003 ) .
http://arxiv.org/abs/nucl-th/0212024[arxiv:nucl-th/0212024 ] .
j. a. mcgovern , d. r. phillips and h. w. griehammer , eur .
phys . j. a * 49 * , 12 ( 2013 ) .
http://arxiv.org/abs/1210.4104[arxiv:1210.4104 ] [ nucl - th ]
. v. lensky and v. pascalutsa , pisma zh .
* 89 * , 127 ( 2009 ) [ jetp lett . *
89 * , 108 ( 2009 ) ] .
http://arxiv.org/abs/0803.4115[arxiv:0803.4115 ] [ nucl - th ] .
v. lensky and v. pascalutsa , eur .
j. c * 65 * , 195 ( 2010 ) .
http://arxiv.org/abs/0907.0451[arxiv:0907.0451 ] [ hep - ph ] .
v. lensky and j. a. mcgovern , phys .
c * 89 * , 032202 ( 2014 )
. http://arxiv.org/abs/1401.3320[arxiv:1401.3320 ] [ nucl - th ] .
v. lensky , j. a. mcgovern , d. r. phillips and v. pascalutsa , phys .
c * 86 * , 048201 ( 2012 ) .
http://arxiv.org/abs/1208.4559[arxiv:1208.4559 ] [ nucl - th ] .
h. r. weller , m. w. ahmed , h. gao , w. tornow , y. k. wu , m. gai and r. miskimen , prog .
* 62 * , 257 ( 2009 ) .
martel _ et al . _
[ a2 collaboration ] , phys .
lett . * 114 * , 112501 ( 2015 ) .
http://arxiv.org/abs/1408.1576[arxiv:1408.1576 ] [ nucl - ex ] .
v. bernard , n. kaiser and u .- g .
meiner , phys .
lett . * 67 * , 1515 ( 1991 ) .
v. bernard , n. kaiser , a. schmidt and u .- g .
meiner , phys .
b * 319 * , 269 ( 1993 ) .
http://arxiv.org/abs/hep-ph/9309211[arxiv:hep-ph/9309211 ] .
v. bernard , n. kaiser and u .- g .
meiner , int .
j. mod .
e * 4 * , 193 ( 1995 ) .
http://arxiv.org/abs/hep-ph/9501384[arxiv:hep-ph/9501384 ] .
t. r. hemmert , b. r. holstein , j. kambor and g. knochlein , phys .
d * 57 * , 5746 ( 1998 ) .
http://arxiv.org/abs/nucl-th/9709063[arxiv:nucl-th/9709063 ] .
b. r. holstein , d. drechsel , b. pasquini and m. vanderhaeghen , phys .
c * 61 * , 034316 ( 2000 ) .
http://arxiv.org/abs/hep-ph/9910427[arxiv:hep-ph/9910427 ] .
x. d. ji , c. w. kao and j. osborne , phys .
d * 61 * , 074003 ( 2000 ) .
http://arxiv.org/abs/hep-ph/9908526[arxiv:hep-ph/9908526 ] .
k. b. vijaya kumar , j. a. mcgovern and m. c. birse , phys .
b * 479 * , 167 ( 2000 ) .
http://arxiv.org/abs/hep-ph/0002133[arxiv:hep-ph/0002133 ] .
g. c. gellas , t. r. hemmert and u .- g .
meiner , phys .
lett . * 85 * , 14 ( 2000 ) .
http://arxiv.org/abs/nucl-th/0002027[arxiv:nucl-th/0002027 ] v. bernard , t. r. hemmert and u .- g .
meiner , phys .
d * 67 * , 076008 ( 2003 ) .
http://arxiv.org/abs/hep-ph/0212033[arxiv:hep-ph/0212033 ] .
v. bernard , e. epelbaum , h. krebs and u .- g .
meiner , phys .
d * 87 * , 054032 ( 2013 ) .
http://arxiv.org/abs/1209.2523[arxiv:1209.2523 ] [ hep - ph ] .
a. h. blin , t. gutsche , t. ledwig and v. e. lyubovitskij , phys .
d * 92 * , 096004 ( 2015 ) .
http://arxiv.org/abs/1509.00955[arxiv:1509.00955 ] [ hep - ph ] .
a. m. gasparyan , m. f. m. lutz and b. pasquini , nucl .
a * 866 * , 79 ( 2011 ) .
http://arxiv.org/abs/1102.3375[arxiv:1102.3375 ] [ hep - ph ] .
v. pascalutsa , m. vanderhaeghen and s. n. yang , phys .
rept . * 437 * , 125 ( 2007 ) .
http://arxiv.org/abs/hep-ph/0609004[arxiv:hep-ph/0609004 ] .
v. pascalutsa and m. vanderhaeghen , phys .
d * 73 * , 034003 ( 2006 ) .
http://arxiv.org/abs/hep-ph/0512244[arxiv:hep-ph/0512244 ] .
k. a. olive _ et al . _
[ particle data group collaboration ] , chin .
c * 38 * , 090001 ( 2014 ) .
d. babusci , g. giordano , a. i. lvov , g. matone and a. m. nathan , phys .
c * 58 * , 1013 ( 1998 ) .
http://arxiv.org/abs/hep-ph/9803347[arxiv:hep-ph/9803347 ] .
b. pasquini , p. pedroni and d. drechsel , phys .
b * 687 * , 160 ( 2010 ) . http://arxiv.org/abs/1001.4230[arxiv:1001.4230 ] [ hep - ph ] .
v. olmos de leon _ et al .
_ , eur .
j. a * 10 * , 207 ( 2001 ) .
r. p. hildebrandt , h. w. griehammer , t. r. hemmert and b. pasquini , eur .
j. a * 20 * , 293 ( 2004 ) .
http://arxiv.org/abs/nucl-th/0307070[arxiv:nucl-th/0307070 ] .
b. pasquini , _
private communications_. b. pasquini , d. drechsel and m. vanderhaeghen , phys .
c * 76 * , 015203 ( 2007 ) .
http://arxiv.org/abs/0705.0282[arxiv:0705.0282 ] [ hep - ph ] .
h. w. griehammer , j. a. mcgovern and d. r. phillips , http://arxiv.org/abs/1511.01952[arxiv:1511.01952 ] [ nucl - th ] .
d. drechsel , b. pasquini and m. vanderhaeghen , phys .
rept . * 378 * , 99 ( 2003 ) .
http://arxiv.org/abs/hep-ph/0212124[arxiv:hep-ph/0212124 ] .
k. kossert _ et al .
_ , eur .
j. a * 16 * , 259 ( 2003 ) .
http://arxiv.org/abs/nucl-ex/0210020[arxiv:nucl-ex/0210020 ] .
l. s. myers _ et al .
_ [ compton@max - lab collaboration ] , phys .
lett . * 113 * , 262506 ( 2014 ) .
http://arxiv.org/abs/1409.3705[arxiv:1409.3705 ] [ nucl - ex ]
. m. i. levchuk and a. i. lvov , nucl .
a * 674 * , 449 ( 2000 ) .
http://arxiv.org/abs/nucl-th/9909066[arxiv:nucl-th/9909066 ] .
m. schumacher and m. i. levchuk , nucl .
a * 858 * , 48 ( 2011 ) .
http://arxiv.org/abs/1104.3721[arxiv:1104.3721 ] [ hep - ph ] .
e. epelbaum , h. krebs and u .- g .
meiner , eur .
j. a * 51 * , 53 ( 2015 ) .
http://arxiv.org/abs/1412.0142[arxiv:1412.0142 ] [ nucl - th ] .
e. epelbaum , h. krebs and u .- g .
meiner , phys .
* 115 * , 122301 ( 2015 ) .
http://arxiv.org/abs/1412.4623[arxiv:1412.4623 ] [ nucl - th ] .
r. j. furnstahl , n. klco , d. r. phillips and s. wesolowski , http://arxiv.org/abs/1506.01343[arxiv:1506.01343 ] [ nucl - th ] .
i. guiasu and e. e. radescu , annals phys .
* 120 * , 145 ( 1979 ) .
i. guiasu and e. e. radescu , annals phys .
* 122 * , 436 ( 1979 ) .
h. w. griehammer and t. r. hemmert , phys .
c * 65 * , 045207 ( 2002 ) .
http://arxiv.org/abs/nucl-th/0110006[arxiv:nucl-th/0110006 ] .
r. p. hildebrandt , h. w. griehammer and t. r. hemmert , eur .
j. a * 20 * , 329 ( 2004 ) .
http://arxiv.org/abs/nucl-th/0308054[arxiv:nucl-th/0308054 ] .
a. aleksejevs and s. barkanova , nucl .
phys . proc .
suppl . * 245 * , 17 ( 2013 ) .
http://arxiv.org/abs/1309.3313[arxiv:1309.3313 ] [ hep - ph ] ; d. drechsel , s. s. kamalov and l. tiator , eur .
j. a * 34 * , 69 ( 2007 ) .
http://arxiv.org/abs/0710.0306[arxiv:0710.0306 ] [ nucl - th ] .
r. p. hildebrandt , _ phd thesis _ , t. u. mnchen ( 2005 ) .
http://arxiv.org/abs/nucl-th/0512064[arxiv:nucl-th/0512064 ] .
g. blanpied , m. blecher , a. caracappa , r. deininger , c. djalali , g. giordano , k. hicks and s. hoblit _ et al .
_ , phys .
c * 64 * , 025203 ( 2001 ) . c. collicott ,
_ phd thesis _ , dalhousie university ( 2015 ) .
j. a. m. vermaseren , http://arxiv.org/abs/math-ph/0010025[arxiv:math-ph/0010025 ]
. t. hahn and m. perez - victoria , comput .
commun .
* 118 * , 153 ( 1999 ) .
http://arxiv.org/abs/hep-ph/9807565[arxiv:hep-ph/9807565 ] . | we update the predictions of the su(2 ) baryon chiral perturbation theory for the dipole polarisabilities of the proton , @xmath0 @xmath1 , and obtain the corresponding predictions for the quadrupole , dispersive , and spin polarisabilities : @xmath2 @xmath3 , @xmath4 @xmath3 , and @xmath5 @xmath6 .
the results for the scalar polarisabilities are in significant disagreement with semi - empirical analyses based on dispersion relations , however the results for the spin polarisabilities agree remarkably well .
results for proton compton - scattering multipoles and polarised observables up to the delta(1232 ) resonance region are presented too . the asymmetries @xmath7 and
@xmath8 reproduce the experimental data from legs and mami .
results for @xmath9 agree with a recent sum rule evaluation in the forward kinematics .
the asymmetry @xmath10 near the pion production threshold shows a large sensitivity to chiral dynamics , but no data is available for this observable .
we also provide the predictions for the polarisabilities of the neutron , the numerical values being @xmath11 @xmath1 , @xmath12 @xmath3 , @xmath13 @xmath3 , and @xmath14 @xmath6 .
the neutron dynamical polarisabilities and multipoles are examined too .
we also discuss subtleties related to matching dynamical and static polarisabilities . |
ngc 1275 , also known as perseus a and 3c 84 , is the nearby active galaxy located at the center of the perseus cluster at redshift of @xmath3 ( ( * ? ? ? *
veron ( 1978 ) ) ) .
this object has been well studied in radio , optical / uv , x - ray and mev - gev emission ( ( * ? ? ?
* abdo ( 2009 ) ) , , ( * ? ? ?
* acciari ( 2009 ) ) ) .
recently , this source has been detected by magic telescopes with a statistical significance of @xmath4 above 100 gev in 46 hr of stereo observations carried out between august 2010 and february 2011 .
the measured differential energy spectrum between 70 gev and 500 gev can be described by a power law with a steep spectral index of @xmath5 , and the average flux above 100 gev is @xmath6 ( ) .
we propose a lepto - hadronic model to describe the whole sed as the superposition of synchrotron self - compton ( ssc ) emission and the neutral pion decay products .
fermi - accelerated electrons in the emitting region are described by a broken power - law written as : @xmath7 for @xmath8 , and @xmath9 for @xmath10 and co - accelerated protons at the same place for a simple power law as : @xmath11 , where @xmath12 is the power index for electrons ( protons ) and @xmath13 ( @xmath14 ) are the electron ( proton ) lorentz factors .
the index @xmath15 , @xmath16 and @xmath17 are for minimum , break and maximum , respectively . assuming that the emitting region is endowed by a magnetic field , electrons and protons are cooled down by electromagnetic ( synchrotron radiation and compton scattering ) and hadronic ( proton - photon interaction ) channels , respectively .
taking into account the cooling processes for electron and proton distributions , the observed spectra for electron synchrotron radiation , compton scattering and proton - photon interaction are ( ( * ? ? ? * fraija ( 2014a ) ) , ( * ? ? ?
* fraija ( 2012 ) ) ) @xmath18 @xmath19 and @xmath20 respectively , where @xmath21 ( @xmath22 ) , @xmath23 ( @xmath24 ) , @xmath25 ( @xmath26 ) and @xmath27 ( @xmath28 ) are the proportionality constant and break energies for characteristic , cut - off and maximum of electron synchrotron ( compton scattering ) spectrum and @xmath29 and @xmath30 are the proportionality constant and break photon energy for photo - pion spectrum ( ( * ? ? ? * fraija ( 2014b ) ) ) .
it is important to highlight that the neutrino counterpart is calculated through @xmath31 , where the neutrino flux is dn@xmath32 , with @xmath33 .
we have presented a lepton - hadronic model through eqs .
[ espsyn ] , [ espsyn ] and [ pgamma ] to describe the broadband sed of ngc1275 . for the hadronic interactions we evoke the p@xmath0 interactions of fermi - accelerated protons with the target photons at the second ssc peak . to find the best fit of our leptonic and hadronic models , we use the method of chi - square ( @xmath34 ) minimization .
the fit values for the leptonic model are shown in table 1 and 2 while for hadronic model are @xmath35 and @xmath36 .
the leptonic model describes the photon spectrum up to a few gevs while the hadronic model up to hundreds of gevs .
the neutrino event expected in a km@xmath2 cherenkov telescope is @xmath37 per year . | the nearby active galaxy ngc 1275 , has widely been detected from radio to gamma rays .
its spectral energy distribution ( sed ) shows a double - peak feature , which is well explained by synchrotron self - compton ( ssc ) model .
however , recent tev detections might suggest that very - high - energy @xmath0-rays ( e@xmath1100 gev ) may not have a leptonic origin .
we test a lepto - hadronic model to describe the whole sed through ssc emission and neutral pion decay resulting from p@xmath0 interactions . also , we estimate the neutrino events expected in a km@xmath2 cherenkov telescope . |
in condensed matter physics there are several models presenting spatial modulated structures of some local property , such as , for instance , the position of the particles , the magnetization and the charge density bak1982 .
such modulation can be commensurate or incommensurate in relation to the underlying lattice .
a phase is named commensurate if the ratio between the period of the modulation and the period of the lattice is a rational number . otherwise , the phase is named incommensurate .
the basic mechanism leading to modulation is the competition between interactions favoring distinct orderings @xcite .
for example , the modulated structures observed in rare - earth metals @xcite were interpreted as a consequence of the competition generated by the spatially oscillatory interaction rkky @xcite . in order to explain the spatial modulation found in erbium
, elliott introduced an ising model @xcite , twenty years later named annni ( axial - next - nearest - neighbor ising ) model @xcite .
it is one of the simplest models able to exhibit a rich phase diagram containing a complex region of spatially modulated phases selke1988,yeomans1988,selke1992 .
the model is defined by the hamiltonian @xmath14 \ \sigma _ { x , y , z}% \end{array}%\ ] ] where @xmath15 is an ising spin variable at the site @xmath16 of a simple cubic lattice , @xmath17 is the nearest neighbor interaction in the @xmath18 planes , @xmath19and @xmath20 are the nearest- and next - nearest - neighbor interactions in the @xmath1 direction .
* * * * here @xmath21 and @xmath20 compete and may have opposite or same signs .
however , when @xmath19 and @xmath20 have opposite signs , the competition between them may give rise to the modulated phases .
the mean - field phase diagram of the model displaying the main commensurate phases in the plane of the reduced temperature @xmath22 against the competition parameter @xmath23 , shown in fig .
[ phase_diagram ] , was obtained in the beginning of the eighties @xcite and is divided in three major regions : the modulated region ( m ) , the paramagnetic ( p ) and the ferromagnetic ( f ) phases . in this diagram , as well as in the whole paper , we make @xmath24 .
the paramagnetic phase is separated from the ferromagnetic phase and from the modulated region by critical lines p - f and p - m respectively . on the other hand , a first - order transition ( f - m ) takes place between the ferromagnetic and the modulated region .
the critical line p - m belongs to the universality class of the xy model , whereas the critical line p - f presents ising - like critical behavior @xcite .
the transition lines p - f , p - m and f - m meet at the lifshitz point ( lp ) , introduced theoretically by hornreich et al . @xcite .
the location of the lp in the phase diagram of the annni model was obtained for the first time from high temperature series technique @xcite . however
, the precise location of the lp is difficult to obtain because , close to it , one meets the challenging task of dealing with modulated phases of very large periods .
thus , latter attempts to locate the lp were made .
single spin - flip monte carlo simulations were performed along the p - f critical line towards the lp @xcite , thus avoiding the modulated region . despite the successful results obtained from these works , it was only recently that the lifshitz point was located with high precision @xcite by means of a new variation of the cluster wolf algorithm @xcite .
concerning the critical properties of the lp , it was shown that this multicritical point belongs neither to the universality class of the xy model nor to the one of ising model @xcite .
this fact gave rise to much interest in the study of the particular critical behavior of a lp .
thus , the critical properties of the lp found in the phase diagram of the annni model were widely studied by several approaches hornreich1980,mo1991 , among them @xmath25expansions@xcite , high temperature series technique @xcite and monte carlo simulations @xcite .
experimentally , the most complete results concerning the lp were obtained for the magnetic compound mnp , which exhibits a lp belonging to the universality class of the lp present in the annni model @xcite .
here it is important to mention that although this model has been widely explored by methods from equilibrium statistical mechanics , there are no results for the 3d annni model obtained by an approach that deals with mc simulations performed far from equilibrium @xcite that will be employed in this paper .
however we remark that several works in the literature have used previously a dynamical approach to study the lifshitz points .
for example the critical dynamics of relaxation of the model near the lifshitz point has been studied by the @xmath26expansion and the exponent @xmath1 was numerically estimated for uniaxial and biaxial cases @xcite .
alternatively , in @xcite the growth of a order parameter is studied when a system at lifshitz point is quenched from the homogeneous disordered state to the ordered state where correlation and structure factors after quench in this system are analyzed via renormalization group method .
numerical estimates of exponent @xmath1 are supplied in this same reference .
studies about dynamics of annni model were developed but in two dimensional versions of the model but such works do not explore critical properties .
for example , in @xcite , @xcite , the kinetics of domain growth of the one and two - dimensional anisotropic annni model was explored via monte carlo methods using glauber and heat - bath multispin dynamics respectively . in the next subsection we briefly present the non - equilibrium approach to be used in our analysis .
the study of critical properties originated from statistical fluctuations of spin systems became possible in non - equilibrium stage after the seminal ideas of janssen , schaub and schmittmann @xcite and huse huse1989 . by quenching systems from high temperatures to the critical one
, they have shown that universality and scaling behavior appear even in the early stages of time evolving , via renormalization group techniques and numerical calculations respectively .
hence , by using short time dynamics , one can often circumvent the well - known problem of the critical slowing down that plagues investigations of the long - time regime . here
, we briefly review finite size scaling in the dynamics relaxation of spin systems .
we present our alternative derivation of some power laws in the short time dynamics context .
readers , who want a more complete review about this topic , may want to read @xcite .
this topic is based on time dependent simulations , and it constitutes an important issue in the context of phase transitions and critical phenomena .
such methods can be applied not only to estimate the critical parameters in spin systems , but also to calculate the critical exponents ( static and dynamic ones ) through different scaling relations by setting different initial conditions .
the dynamic scaling relation obtained by janssen _
k_-th moment of the magnetization , is written as @xmath27where the arguments are : the time @xmath6 ; the reduced temperature @xmath28 , with @xmath29 being the critical one , the lattice linear size @xmath30 and initial magnetization @xmath31 .
here , the operator @xmath32 denotes averages over different configurations due to different possible time evolutions from each initial configuration compatible with a given @xmath31 . on the right - hand side of the equation
, one has : an arbitrary spatial rescaling factor @xmath33 ; an anomalous dimension @xmath34 related to @xmath31 .
the exponents @xmath12 and @xmath13 are the equilibrium critical exponents associated with the order parameter and the correlation length , respectively .
the exponent @xmath1 is the dynamic one , which characterizes the time correlations in equilibrium . after choosing @xmath35 , @xmath36 @xmath29(@xmath37 ) , and @xmath38
we obtain @xmath39 . denoting @xmath40 and @xmath41
, one has : @xmath42 .
the derivative with respect to @xmath30 is : @xmath43,\end{aligned}\]]where we have explicitly : @xmath44 and @xmath45 . in the limit @xmath46 , @xmath47 ,
one has : @xmath48 .
the separability of the variables @xmath49 and @xmath50 in @xmath51 leads to @xmath52 , where the prime means the derivative with respect to the argument .
since the left - hand - side of this equation depends only on @xmath50 and the right - hand - side depends only on @xmath53 , they must be equal to a constant @xmath54 .
thus , @xmath55 and @xmath56 , resulting in @xmath57 .
returning to the original variables , one has : @xmath58 .
on one hand , choosing @xmath59 and calculating @xmath60 , at criticality ( @xmath61 ) , we obtain @xmath62corresponding to a regime under small initial magnetization .
this can be observed by a finite time scaling @xmath63 in equation ( mainshorttime ) , at critical temperature ( @xmath61 ) which leads to @xmath64 .
defining @xmath65 , an expansion of the averaged magnetization around @xmath66 results in : @xmath67 . by construction @xmath68 , since @xmath69 and @xmath70 is a constant , discarding the quadratic terms we obtain the expected power law behavior @xmath71 .
this anomalous behavior of initial magnetization is valid only for a characteristic time scale @xmath72 @xmath73 . on the other hand
, the choice @xmath74 corresponds to the case which the system does not depend on the initial trace and @xmath75 leads to simple power law : @xmath76that similarly corresponds to decay of magnetization for @xmath77 .
the evaluation of critical exponents @xmath3 and @xmath78 via monte carlo simulations concerns taking averages over different runs . in the second case ,
simpler simulations are considered because since the system starts from the ferromagnetic ( ordered ) initial state . however ,
the first one is somewhat difficult to deal with , once it demands working with prepared initial states with a precise value of @xmath31 ( sharp preparation ) , besides the delicate limit @xmath79 .
an alternative way to determine this exponent was proposed by tom and oliveira @xcite , where it was shown that the time correlation function of the order parameter also follows a power law scale form at the short - time regime , i. e. @xmath80 the main advantage in the use of eq .
( [ correlation ] ) is that one does not need to fix precisely the initial order parameter @xmath31 .
the only requirement is that @xmath81 , where @xmath82 stands for the average of the quantity @xmath83 over different initial configurations .
now , let us look at the second moment of magnetization .
since the spin - spin correlation @xmath84 is negligible for @xmath85 , we have that for a fixed @xmath6 , @xmath86and by considering the scaling relation ( with @xmath63 ) for the second moment of magnetization , we have according to [ mainshorttime ] : @xmath87 where @xmath88 is the system dimension . using monte carlo simulations ,
many authors have obtained the dynamic exponents @xmath3 and @xmath1 as well as the static ones @xmath12 and @xmath13 , and other specific exponents for several models : baxter - wu arashiro2003 , 2 , 3 and 4-state potts @xcite , ising with multispin interactions @xcite , ising with competing interactions @xcite , models with no defined hamiltonian ( celular automata ) @xcite , models with tricritical point silvaii2002 , heisenberg @xcite , protein folding arashiro3 and so on .
the sequence to determine the static exponents from short time dynamics is : first we determine @xmath1 , performing monte carlo simulations that mixes initial conditions @xcite , and consider the power law for the cumulant @xmath89 this ratio has proven to be useful for the calculation of the exponent @xmath1 for the several spin models governed by boltzmann - gibbs statistical mechanics but its application also includes models with spin - flip based on generalized statistics @xcite . in this technique ,
graphs of @xmath90 against @xmath91 lay on the same straight line for different lattice sizes , without any re - scaling in time , yielding more precise estimates for @xmath92 .
although it seems clear from eq .
( [ z ] ) , it is worth to stress here that two independent runs are necessary in order to calculate the ratio @xmath93 : in one of them @xmath94 , while in the other one @xmath75 .
equations ( [ correlation]),([initial slip ] ) , and ( [ z ] ) solve the problem in determining the dynamic critical exponents @xmath3 and @xmath1 .
but the ability of short - time monte carlo simulations goes beyond the evaluation of dynamic critical exponents , in the sense that this technique also allows us to obtain the static critical exponents which will be discussed in sequence .
starting from @xmath75 , we have the expected power law described by eq.(decay_ferro ) , so considering @xmath95 , we must expect @xmath96which is obtained by differentiating the quantity @xmath95 in relation to reduced temperature @xmath97 .
when dealing with monte carlo simulations , the partial derivative is approximated in first order by the difference @xmath98
\label{aproximation}\]]where @xmath99 .
it is clear from eq .
( [ aproximation ] ) above that two independent simulations are necessary to obtain the exponent @xmath100 : one of them evolves at the temperature @xmath101 , whereas the other one evolves at @xmath102 .
therefore , with @xmath103 estimated from eq.([z ] ) we obtain the estimated @xmath104 according to @xmath105 obtained from eq.([derivative ] ) , by calculating the product:@xmath106 ^{-1}\]]where the uncertainty in @xmath104 is calculated through uncertainty of @xmath1 ( @xmath107 ) and the uncertainty of @xmath105 ( @xmath108 ) , previously obtained by statistics over @xmath109 different seeds , according to the error propagation equation @xmath110 finally , in order to obtain an estimate @xmath111 , we firstly estimate @xmath112 obtained from the magnetization decay ( [ decay_ferro ] ) and after we perform the product @xmath113 ^{-1}\]]and the propagated error is directly calculated as a function of the respective uncertainties:@xmath114 here , it is important to mention that the ratio @xmath115 already should have been previously calculated by the evaluated @xmath112 from eq.([decay_ferro ] ) and @xmath103 from eq.([z ] ) such that:@xmath116 not only quantities related to the moments of the magnetization can explain the non - equilibrium aspects of phase transitions and critical phenomena , but also those related to the first passage time probabilities and variations of this topic ( see e.g @xcite,@xcite ) . by considering this approach and under the same nonequilibrium conditions ,
a new exponent was initially proposed by @xcite : the global persistence exponent @xmath117 .
it is related to the probability @xmath5 that the global order parameter has not changed sign up to time @xmath6 after a quench to @xmath29 , according to expected power law behavior@xmath118as argued by majumdar _
@xcite , if the dynamics of the global order parameter is described by a markovian process , @xmath119 is not a new independent exponent and it can be related to other critical exponents , @xmath120where @xmath121 is the autocorrelation exponent in the expected power law @xcite : @xmath122 , where @xmath123 is the value of the spin variable @xmath124 at the site @xmath125 of a d - dimensional system of linear size l , assumed , at instant @xmath6 , @xmath1 is the dynamic critical exponent defined as @xmath126 , where @xmath127 and @xmath128 are time and spatial correlation lengths , respectively .
however , the time evolution of the order parameter is , in general , a non - markovian process and @xmath119 turns out to be a new independent critical exponent describing the evolving of the stochastic dynamic process toward the equilibrium . in order to evaluate the persistence probability @xmath5 , we first define @xmath129 as the fraction of runs for which the magnetization changes its sign for the first time at the instant @xmath6 .
our probability @xmath5 is numerically calculated from the cumulative distribution function of such @xmath129 .
so , @xmath5 describes the probability of magnetization does not cross the origin up to time @xmath6 , @xmath130we start our simulations with a random initial condition , where @xmath131 .
here it is important to mention that such a concept is very versatile and it has been applied to characterize several applications such as tricritical points @xcite , besides interdisciplinary applications such as : analysis of bankruptcies of players in public goods games @xcite , econophysics @xcite and many others . so it can also be interesting , for example , to study lifshitz points in spin models .
the target of this paper is to enlarge our knowledge of the of the annni model , by studying the ferromagnetic - paramagnetic phase - transition with special attention to the lifshitz point ( lp ) .
the layout of this paper is as follows : in section ( [ mcsimulations ] ) we give more details about numerical simulations that will be performed at the lifshitz point .
moreover , we describe a simple method recently developed by r. da silva et al . in @xcite that refines the critical parameters based on the best determination coefficient in linear fit @xmath132 versus @xmath133 . in this same section
we present a first part of our results , where we explicitly show the refinement of critical temperatures of second order line ( ferromagnetic - paramagnetic ) up to the lifshitz point .
finally in section ( [ results ] ) we present our estimates for the critical exponents and in this case we divide our results in two branches : in the branch ( a ) , we evaluate critical exponents ( only @xmath1 and @xmath115 ) for each temperature estimated along critical line previously estimated by the refinement process . in these first calculations ,
the aim is only the monitoring of the behavior of these two exponents ( one dynamic and the other static ) to show the pronounced difference between ising - like points and the lifshitz point . in the second branch ( b ) of this same section
we present the complete results and studies for both dynamic ( @xmath3 , @xmath119 , and @xmath1 ) and static ( @xmath12 and @xmath13 ) critical exponents specifically for the lifshitz point , comparing the later with results obtained in previous experimental and theoretical works .
finally , in section ( [ conclusions ] ) we summarize and briefly discuss the main results of this paper .
monte carlo simulations were performed on simple rectangular lattices with linear dimensions @xmath134 , @xmath135 , and @xmath136 and periodic boundary conditions .
the spin states were updated by using the one - spin - flip heat - bath algorithm . for the location of the lp , it was used as basis the result obtained by henkel and pleimling @xcite : ( @xmath137 ) . for the annni model
the order parameter corresponds to time - dependent magnetization , defined as an average over all @xmath138 spins and over the different @xmath139 runs : @xmath140 where the index @xmath141 denotes the corresponding run of a simulation .
the ordered state is ferromagnetic , with all ( or most of ) the spins pointing either up or down .
as discussed in the previous section , the lattice s initial condition depends on the scaling relation considered : a ) @xmath142 eq .
( [ correlation ] ) , b ) @xmath94 numerator of eq .
( [ z ] ) , c ) @xmath75 eqs.([decay_ferro ] ) , ( [ aproximation ] ) , and d ) @xmath31 fixed , but with random configurations considering a sharp preparation ( [ initial slip ] ) .
here we address time dependent mc simulations in the context of so - called short time dynamics . before evaluation of some critical exponents and the complete study of the lifshitz point ( next section ) ,
also in this section we apply the refinement method to estimate the critical parameters of several points along second order line including the lifshitz point itself .
the algorithm to estimate the critical temperature is divided in two stages : i ) a coarse grained location ; ii ) fine scale refinement . in stage
i ) , since the magnetization must behave as a power law @xmath143 , by fixing a specific @xmath144-value , we conjecture that changing @xmath145 from @xmath146 up to @xmath147 , the corresponding best @xmath148 is the one that leads to the best linear behavior of @xmath132 versus @xmath91
. we have considered @xmath149 realizations , with initial magnetization @xmath75 .
for each @xmath150 changing from @xmath151 up to @xmath152 , with displacement @xmath153 between the values , where @xmath24 for all of our calculations , we changed @xmath145 from @xmath154 up to @xmath155 .
these extreme values were extracted from the literature , since they correspond to best known estimates for the 3d ising model ( @xmath156 which corresponds to @xmath157 ) and lp ( @xmath158 which corresponds to @xmath159 ) respectively . just as a safeguard
, we enlarge this interval , by performing @xmath160 and @xmath161 .
so , for each input @xmath144-value , by using @xmath162 , we span our temperatures on a range described by parametrization @xmath163 @xmath164 , @xmath165 , so for each temperature , a linear fit is performed and one calculates the determination coefficient of the fit as : @xmath166with @xmath167 , where @xmath168 is the number of monte carlo sweeps . in our experiments ,
we have used @xmath169 mc steps , discarding the initial @xmath170 mcsteps for more robust estimates .
hence , @xmath171 means an exact fit , and so the closer @xmath172 is from the unity , the better is the fit . here , @xmath173 and @xmath33 are the linear coefficient and the slope in the linear fit @xmath132 versus @xmath91 , respectively . from @xmath33 , one estimates the exponent @xmath174 .
after , one finishes this part of our refinement method ( i ) , we pass to the second part of refinement , the fine scale stage ( ii ) . starting from the critical temperature @xmath175 obtained in the first stage
, we use the process considering a more refined displacement , i.e. , @xmath176 .
so , by using @xmath177 and @xmath178 , we consider a new parametrization @xmath179 @xmath180 , with @xmath181 .
hence , we determine a new best temperature @xmath182 corresponding to the maximum value of @xmath172 .
plot [ refinement ] shows the determination coefficient ( @xmath172 ) as function of temperature for two extremal cases @xmath183 and @xmath158 . ) as function of temperature for two cases @xmath184 and @xmath185 .
the upper plots correspond to coarse grained regime ( i ) of process ( @xmath186 ) while the lower plots correspond to fine scale regime ( ii ) of process ( @xmath187 ) . ]
the upper plots correspond to the coarse grained regime ( first refinement
i ) of process ( @xmath188 ) while the lower plots correspond to the fine scale regime ( second refinement
ii ) ( @xmath189 ) . the estimated temperatures after two refinement states for all studied @xmath144-values can be observed in table table : critical_line_annni .
.critical values obtained by refinement procedure after two stages ( coarse - grained and fine scale ) . here ,
the values obtained for @xmath190 and @xmath185 correspond to our estimates of the critical temperatures of 3d - ising model and lifshtz point respectively . [ cols="^,^,^",options="header " , ] + @xmath119 & @xmath191 & & & @xmath192 & & & @xmath193 +
in this work we estimated the dynamic critical exponents ( @xmath3 , @xmath194 , and @xmath1 ) at the lishshitz point of the annni model , by means of time dependent monte carlo simulations .
to our knowledge , this is the first attempt in order to obtain such exponents .
we have also obtained the static and critical exponents @xmath12 and @xmath13 by exploiting scaling relations , valid in the short - time regime ( out of equilibrium ) , involving the order parameter and its second moment .
our estimates ( see table [ our_results ] ) are in very good agreement with previous experimental and theoretical works ( see table static_critical_exponents ) moreover we applied a refinement procedure to estimate several parameters for the critical points along second order transition line from the 3d ising point ( @xmath195 ) up to lifshitz point ( @xmath196 ) and we follow the behavior of some exponents along this line .
we can observe that @xmath197 is universal when @xmath20 is small , i.e. , the interaction among second neighbors in @xmath1 direction is not pronounced resulting in @xmath198 which is exactly expected for universality class of 3d ising .
however , in the neighborhood of the lifshitz point , the exponent @xmath1 has a sensitive increase resulting in @xmath199 .
finally , we also observe power law behavior for the global persistence ( @xmath5 , probability that the sign of the magnetization does not change until the time @xmath6 starting from random configurations with small magnetization ) .
we find @xmath8 ( @xmath200 ) for the lp that is smaller than 3d - ising estimate ( @xmath156 ) @xmath201 .
the authors are partly supported by the brazilian research council cnpq .
authors also thank cesup ( super computer center of federal university of rio grande do sul ) as well as prof .
leonardo g. brunet ( if - ufrgs ) for the available computational resources and support of clustered computing ( ada.if.ufrgs.br ) and prof .
mendeli h. vainstein ( if - ufrgs ) for carefully reading our manuscript .
99 p. bak , rep .
45 , 587 - 629(1982 ) .
g. toulouse , commun .
2(4 ) , 115 - 119(1977 ) .
b. r. cooper , solid state phys .
21 , 393(1968 ) m.a.ruderman , c. kittel , phys .
96 , 99 - 102(1954 ) t. kasuya , prog .
16 , 45 - 57(1956 ) .
k.yosida , phys .
106 , 893 - 898(1957 ) r. j. elliott , phys .
124 , 346 - 353(1961 ) .
m. e. fisher , w. selke , phys .
44 , 1502 - 1505(1980 ) .
w. selke , phys .
170 , 213 - 264(1988 ) .
j. m. yeomans , in solid state physics , edited by h. ehrenreich and d. turnbull ( academic press , san diego ( ca ) , 1988 ) , vol .
151 - 200 .
w. selke , in phase transitions and critical phenomena , edited by c. domb and j. l. lebowitz ( academic press , london , 1992 ) , vol .
p. bak , j.von boehm , phys .
b 21 , 5297 - 5308(1980 ) .
t. garel , p. pfeuty , j. phys .
c 9 l245-l249(1976 ) .
r. m. hornreich , m. luban , s. shtrikman , phys .
35 ( 25 ) , 1678 - 1681(1975 ) .
s. redner , h. e. stanley , phys .
b 16 ( 11 ) 4901 - 4906(1977 ) .
w. selke , z. phys .
b 29 133 - 137(1978 ) .
k. kaski , w.selke , phys .
b 31(5 ) 3128 - 3130(1985 ) . m. henkel , m. pleimling , comput .
147 , 419 - 422(2002 ) .
u. wolf , phys .
62(4 ) , 361 - 364(1989 ) .
r. m. hornreich , j. magn .
mater.15(18 ) , 387(1980 ) .
z. mo , m. ferer , phys .
b 43 ( 13 ) 10890 - 10905(1991 ) . c. c. becerra , y.shapira
, n. f.de oliveira jr .
, t. s. chang , phys.rev.lett .
44 , 1692(1980 ) .
y. shapira , c. c. becerra , n. f. de oliveira jr , t. s. chang , phys . rev .
b 24(5 ) 2780 - 2806(1981 ) .
v. bindilatti , c. c. becerra , n. f. de oliveira jr .
, phys . rev .
b 40 9412(1989 ) .
z. b. li , l. schlke , and b. zheng , phys .
e 53 , 2940 ( 1996 ) h. k. janssen , b. schaub , b. shmittmann , z. phys .
b 73(4 ) , 539 - 549(1989 ) .
d. a. huse , phys .
b 40 , 304(1989 ) .
b.zheng , int . j. mod
b12 , 1419 - 1484(1998 ) .
e. v. albano , m. a. bab , g. baglietto , r. a. borzi , t. s. grigera , e. s. loscar , d. e. rodriguez , m. l. rubio puzzo , and g. p. saracco , rep .
74 , 026501 ( 2011 ) t. tome , m. j. de oliveira , phys . rev .
e 58(4 ) , 4242 - 4245(1998 ) .
e. arashiro and j. r. drugowich de felcio , phys .
e * 67 * , 046123 ( 2003 ) r. da silva , n. a. alves , j. r. drugowich de felcio , phys .
lett . a 298 , 325 - 329(2002 ) .
r. da silva and j. r. drugowich de felcio , phys .
a 333 , 277 ( 2004 ) ; h.a .
fernandes , e. arashiro , j.r . drugowich de felcio , a.a .
caparica physica a 366 , 255 - 264 ( 2006 ) ; e. arashiro , h.a .
fernandes , j.r .
drugowich de felcio , physica a 388 , 4379 - 4386 ( 2009 ) c. s. simes , j. r. drugowich de felcio , mod .
b 15(15 ) , 487 - 496(2001 ) .
n. alves jr . , j. r. drugowich de felcio , mod
b 17(56 ) , 209 - 218(2003 ) .
r. da silva and n. alves jr . , phys . a 350 , 263 ( 2005 ) ; t. tome , j. r. drugowich de felcio mod .
b12 ( 21 ) , 873 - 879(1998 ) .
r. da silva , n. a. alves , j. r. drugowich de felcio , phys .
e 66(2 ) , 026130(2002 ) .
h. a. fernandes , r. da silva , j. r. drugowich de felicio , j. stat .
mech . , p10002 ( 2006 ) .
e. arashiro , j. r. drugowich de felcio , u. h. e. hansmann , phys .
e , 73(4 ) , 40902 ( 2006 ) ; jcp 126 , 045107(2007 ) r. da silva , j. r. drugowich de felcio , a. s. martinez , phys .
e 85 , 066707(2012 ) s.n .
majumdar , a.j .
bray , s.j .
cornell and c. sire , phys .
lett . * 77 * , 3704 ( 1996 ) .
s.n . majumdar and c. sire , phys .
lett . * 77 * , 1420 ( 1996 ) h. k. janssen , from phase transition to chaos , in : g. gyoorgyi , i. kondor , l.sasvlari , t.tel ( eds.),topics in modern statistical physics , worldscientif , singapore,1992 .
r. da silva , n. a. alves , j. r. drugowich de felcio , phys .
e * 67 * , 057102 ( 2003 ) r. da silva , ana l. c. bazzan , a. t. baraviera , slvio r. dahmen , physica a * 371 * , 610 - 626(2006 ) r. da silva , m. zembrzuski , f. correa , l. da c. lamb , physica a * 389 * , 5460 - 5467(2010 ) w. janke and r. villanova , nucl .
b 489 , 679 ( 1997 ) . k. binder , phys . rev .
47(9 ) , 693 - 696(1981 ) .
k.binder , z. phys .
b 43(2 ) , 119 - 140(1981 ) .
m. shpot , h. w. diehl , nucl . phys .
b 612(3 ) , 340 - 372(2001 ) .
y. shapira , nato advanced study institutes , in : r. pynn , a.skjeltorp(eds . ) , multicritical phenomena , vol.106 of b , plenum , new york , pp.53 - 71(1984 ) . | in this work , we study the critical behavior of second order points and specifically of the lifshitz point ( lp ) of a three - dimensional ising model with axial competing interactions ( annni model ) , using time - dependent monte carlo simulations .
first of all , we used a recently developed technique that helps us localize the critical temperature corresponding to the best power law for magnetization decay over time : @xmath0 which is expected of simulations starting from initially ordered states .
secondly , we obtain original results for the dynamic critical exponent @xmath1 , evaluated from the behavior of the ratio @xmath2 , along the critical line up to the lp .
finally , we explore all the critical exponents of the lp in detail , including the dynamic critical exponent @xmath3 that characterizes the initial slip of magnetization and the global persistence exponent @xmath4 associated to the probability @xmath5 that magnetization keeps its signal up to time @xmath6 . our estimates for the dynamic critical exponents at the lifshitz point are @xmath7 and @xmath8 , values very different from the 3d ising model ( annni model without the next - nearest - neighbor interactions at @xmath1-axis , i.e. , @xmath9 ) @xmath10 and @xmath11 .
we also present estimates for the static critical exponents @xmath12 and @xmath13 , obtained from extended time - dependent scaling relations .
our results for static exponents are in good agreement with previous works |
efforts in overcoming the limit to optical communications imposed by fiber nonlinearity can be broadly grouped into two areas : optical and digital techniques .
optical techniques include , for example , optical phase conjugation ( opc ) using twin waves @xcite or opc devices placed mid - span @xcite .
digital techniques include transmitter- @xcite and receiver - side @xcite digital nonlinearity compensation ( nlc ) , simple nonlinear phase shifts @xcite , perturbation - based precompensation @xcite , adaptive filtering @xcite and optimum detection @xcite . with the exception of optimum detection ( a special case of receiver - side nlc for single span transmission ) the digital signal processing ( dsp ) techniques are algorithms which invert the propagation equations for the optical fiber , either exactly or with simplifying approximations .
consider the model in fig .
[ fig : config ] , which shows a transmission link with digital nlc at both the transmitter and the receiver . to date ,
the best performing experimentally demonstrated digital technique for receiver - side digital nlc is the digital backpropagation ( dbp ) algorithm .
this algorithm numerically solves the inverse of the optical fiber propagation equations to compensate the linear and nonlinear impairments introduced by the optical fiber transmission ; albeit not taking account of amplifier noise ( cf .
zero - forcing equalization ) .
the demonstrations of multi - channel nlc via receiver - side dbp @xcite which takes account of the inter - channel nonlinear distortions , have recently been repeated using similar digital techniques to predistort for nonlinearity at the transmitter digital precompensation ( dpc ) @xcite .
we note that , as might be expected due to the symmetry of the transmission link , the experimentally demonstrated performance of the pre- and post - compensation algorithms is similar ; achieving a 100% increase in transmission reach@xcite . whether applying dbp or dpc , additive noise from in - line optical amplifiers is enhanced , which limits the performance of the nlc .
one can , therefore , make a heuristic argument for dividing the dbp equally between transmitter and receiver , thus limiting the noise enhancement in the compensated waveform to the signal - noise interaction present at the center ( rather than the end ) of the transmission link .
although split nlc ( dividing digital nlc between transmitter and receiver ) has previously been considered , experimental implementation was confined to the special case of simplified dsp ( single nonlinear phase shift @xcite ) and theoretical analysis considered only residual nlc after opc @xcite . in this letter , we assess the performance of split nlc via numerical simulations , and characterize performance in terms of achievable signal - to - noise ratio ( snr ) and mutual information ( mi ) .
further , we confirm these results theoretically .
to model the effect of digital nlc we used a coherent gaussian noise ( gn ) model of nonlinear interference @xcite including the effect of signal - ase ( amplified spontaneous emission ) noise interactions .
the model treats the field propagating in the fiber as a summation of signal and noise fields , incorporating the signal - ase noise interaction as a form of cross channel interference .
similar to @xcite the symbol snr ( signal - to - noise ratio ) at the receiver is approximated as @xmath0 where @xmath1 is the signal power , @xmath2 is the number of spans , @xmath3 is the ase noise power in the signal bandwidth from a single span amplifier , @xmath4 is a single span nonlinear interference factor for the self channel interference ( i.e. , signal - signal interactions ) , @xmath5 is a single span nonlinear interference factor for signal - squared ase noise interference , @xmath6 is the coherence factor for self channel interference and @xmath7 is a factor depending on the number of spans and the method used for digital nlc .
the nonlinear interference terms with ase noise squared and cubed have been neglected as insignificant . in order to analytically calculate the snr when applying different nonlinear equalization methods , the @xmath7 parameter must be computed for dpc , dbp and split nlc ( @xmath8 , @xmath9 and @xmath10 , respectively ) . following the method in the appendix
, it is found that the difference in snr at optimum signal launch power when applying split nlc versus dbp is given by @xmath11 where @xmath9 and @xmath10 are given in the appendix by eqs . and .
choosing a 50% transmitter : receiver split ratio for nlc , and for a large number of spans , becomes @xmath12 for this work , @xmath13 . for @xmath14
varying from 0 to 0.3 ( a conservatively high value ) @xmath15 varies between 1.5 db and 1.95 db . in the large bandwidth limit
, @xmath14 tends to zero .
consider the point - to - point transmission link shown in fig .
[ fig : config ] , consisting of an idealized optical transmitter and coherent receiver separated by @xmath2 spans of standard single mode fiber ( ssmf ) , followed by erbium doped fiber amplifiers ( edfa ) .
the simulation parameters are shown in table [ tab : parameters ] .
the transmitted signal was single channel 50 gbd polarization division multiplexed 256-_ary _ quadrature amplitude modulation ( pdm-256qam ) .
this format was chosen as it is sufficiently high cardinality to demonstrate increases in mi when using nlc for all transmission distances considered .
the signal was sampled at 4 samples / symbol ( to take account of the signal broadening due to fiber nonlinearity ) and shaped using a root - raised cosine ( rrc ) filter . where dpc was considered , it was applied at this point using the split step fourier method ( ssfm ) to solve the manakov equation ( * ? ? ?
* eq . ( 12 ) ) .
the optical fiber span was again modeled by solving the manakov equation using the ssfm .
each fiber span was followed by an edfa which applied gain which exactly compensated the previous span loss . where required , the receiver dsp applied either frequency domain chromatic dispersion compensation ( linear case ) , or dbp .
subsequently , a matched rrc filter was applied to the signal , and the signal was downsampled to 1 sample / symbol . to mitigate any residual phase rotation due to uncompensated nonlinear interference ,
carrier phase recovery was performed as described in @xcite .
finally , the snr was estimated over @xmath16 symbols by comparing the transmitted and received symbols as in @xcite .
for ssfm simulations , the mi was computed using monte carlo integration and is included as a figure of merit to provide insight into the gains in throughput possible when employing digital nlc .
note that the analytical mi results are obtained using numerical integration .
.summary of system parameters used in fiber simulation [ cols="<,>,<",options="header " , ] [ tab : parameters ]
c + [ only marks , mark = triangle*,mark size=3,draw = s1,mark options = fill = white , style = solid ] table [ x index=0 , y index=1 ] ; + [ only marks , mark = x , mark size=3,draw = s3 ] table [ x index=0 , y index=1 ] ; + [ only marks , mark = square*,mark size=2,draw = s2,mark options = fill = white , style = solid ] table [ x index=0 , y index=1 ] ; + [ only marks , mark=+,mark size=3,draw = s4 ] table [ x index=0 , y index=1 ] ; + [ mark = none , draw = s1,style = dotted ] table [ x index=0 , y index=1 ] ; + [ mark = none , draw = s3,style = dashdotted ] table [ x index=0 , y index=1 ] ; + [ mark = none , draw = s2,style = solid ] table [ x index=0 , y index=1 ] ; + [ mark = none , draw = s4,style = dashed ] table [ x index=0 , y index=1 ] ; ( source ) at ( axis cs:5200,19.50 ) ; ( source ) at ( axis cs:3200,21.10 ) ; ( destination ) at ( axis cs:3200,18.00 ) ; ( source)(destination ) ; at ( axis cs:154,12 ) ; + + [ mark = none , draw = s1,style = dotted ] table [ x index=0 , y index=1 ] ; + [ mark = none , draw = s2,style = solid ] table [ x index=0 , y index=1 ] ; + [ mark = none , draw = s3,style = dashdotted ] table [ x index=0 , y index=1 ] ; + [ mark = none , draw = s4,style = dashed ] table [ x index=0 , y index=1 ] ; + [ only marks , mark = triangle*,mark size=3,draw = s1,mark options = fill = white ] table [ x index=0 , y index=1 ] ; + [ only marks , mark = square*,mark size=2,draw = s2,mark options = fill = white , style = solid ] table [ x index=0 , y index=1 ] ; + [ only marks , mark = x , mark size=3,draw = s3 ] table [ x index=0 , y index=1 ] ; + [ only marks , mark=+,mark size=3,draw = s4 ] table [ x index=0 , y index=1 ] ; ( source ) at ( axis cs:5900,12.50)@xmath171 bit / symbol ; ( source ) at ( axis cs:3200,13.10 ) ; ( destination ) at ( axis cs:3200,11.75 ) ; ( source)(destination ) ; at ( axis cs:154,6 ) ; + + [ mark = none , draw = black , style = dotted ] table [ x index=0 , y index=1 ] ; + [ mark = none , draw = black , style = dashed ] table [ x index=0 , y index=1 ] ; + [ mark = none , draw = black , style = solid ] table [ x index=0 , y index=1 ] ; coordinates ( 0,0 ) ; coordinates ( 100,0.44 ) ; coordinates ( 50,1.92 ) ; coordinates ( 0,0 ) ; coordinates ( 100,0.25 ) ; coordinates ( 50,1.83 ) ; coordinates ( 0,0 ) ; coordinates ( 100,0.06 ) ; coordinates ( 50,1.74 ) ; at ( axis cs:0.3,0.26 ) ; the analytical expression for snr , , was evaluated using the methods outlined in the appendix for calculating @xmath7 for both linear signal equalization ( cdc ) and for each nlc technique : dpc , dbp and split nlc . in simulation , transmission distances
were considered between 200 and 10000 km ( 2 - 100 spans ) , with the signal launch power varied in 1 db steps . for each transmission distance , the mi and snr were determined at the optimum launch power .
[ fig : results](a ) shows how the maximum achievable snr varies with transmission distance when applying different digital nlc techniques .
a 50% split ratio is used for split nlc .
it should be noted that there is excellent agreement between the analytical expressions and the ssfm simulations , with an snr estimation accuracy better than 0.5 db for short distances , and better than 0.2 db for distances above 1000 km , where the gn model is known to have greater accuracy due to the high accumulated chromatic dispersion .
an snr improvement for split nlc over both dbp and dpc at all distances is also observed , as predicted by .
[ fig : results](b ) shows how the mi of the received signal degrades with distance when applying different digital techniques for equalization . the modulation format considered ( dp-256qam )
encodes a maximum 16 bits of information .
the advantage of digital nlc is clear as , even at short distances , this maximum mi can not be achieved without either dpc , dbp or split nlc . in longer
reach scenarios ( @xmath181000 km ) there is a clear gain in mi when using split nlc compared with dpc or dbp , and this gain saturates for distances greater than approximately 2000 km to be approximately 1 bit / symbol .
the results in fig .
[ fig : results](c ) show the snr gain that can be achieved by dividing the nlc between transmitter and receiver with different ratios .
note that the gain of dpc over dbp rapidly diminishes with transmission distance .
further , it can be seen that a 50% nlc split ratio is optimum for all transmission distances .
it should be noted that the ssfm simulations and the theoretical analysis represent a somewhat idealized model of an optical fiber transmission system .
for example , polarization mode dispersion is known to negatively impact on the performance of digital nlc , and yet is not considered in this model .
therefore , these results should be interpreted as an optimistic estimation of performance using digital nlc .
nevertheless , this work demonstrates that the current arrangements of digital nlc ( dpc or dbp ) can be substantially improved .
we used a closed form approximation for the accumulation of the signal - ase interaction over multiple spans in order to analyse the potential snr gain when dividing nlc between transmitter and receiver .
the optimum launch power , and hence snr gain , when using split nlc will increase by 1.5 db with respect to both dpc and dbp in the limit of long distance , high bandwidth transmission .
split nlc is shown , both theoretically and by numerical simulation , to globally outperform both dpc and dbp for all transmission distances .
there is scope to use this snr gain to reduce the complexity of nlc by dividing the dsp between transmitter and receiver ; a subject for further investigation .
the following is a derivation of analytical expressions for @xmath7 in the case of both linear cdc at the receiver , and nonlinear compensation using pre- , post- or split - nlc .
the nonlinear interference factors , @xmath4 and @xmath5 , were calculated using numerical integration of the gn model reference equation ( * ? ? ?
* eq . ( 1 ) ) .
note @xmath19 but is more accurately given by numerical integration of ( * ? ? ?
* eq . ( 7 ) ) where the spectral shape , @xmath20 , is replaced by unity to represent the uniformity of the ase spectrum . coherence factors @xmath6 and @xmath14 were calculated by obtaining the nonlinear interference factors for 100 spans by numerical integration and using @xmath21 where @xmath22 is the single span nonlinear interference factor and @xmath23 is the nonlinear interference factor for 100 spans . in each case , @xmath24 is substituted by @xmath4 or @xmath5 , as appropriate .
note that the purpose of @xmath25 is to change the coherence of the interference terms , altering the accumulation of the nonlinear interference with number of spans , and that @xmath26 .
the effect of nlc on the effective received snr is modeled by assuming that nlc effectively subtracts in power the nonlinear interference generated by the forward propagating field .
this simplification is customary in the literature and can be seen as the result of two assumptions : i ) a perturbative first - order approximation , and ii ) uncorrelation of all the optical fields involved in the snr calculation . as shown in @xcite , dbp generates a first - order field , identical , but with opposite sign , to the forward - propagated field , provided that the linearly - propagated field ( zeroth - order solution ) along the fiber link is the same , hence the cancellation .
however , due to the noise accumulation over the link , there is a mismatch between the linearly forward - propagated field and the backward - propagated field ( or the precompensated field ) . as a result , residual signal - ase interaction terms are still present after the application of either dbp or dpc , representing one of the main performance limitations @xcite . further assuming a weak nonlinear interaction between ase noise contributions along the link
, the calculation of the signal - ase interaction terms can be performed by considering each ase noise contribution as separately interacting with the signal in each span . the nonlinear interference scaling coefficient , @xmath7 accounts for the noise generated due to this signal - ase interaction . in the case of linear chromatic dispersion compensation ( cdc ) , each contribution of ase noise interacts with the signal from the span following its addition , up until the end of the link . in the configuration
analysed herein ( fig .
[ fig : config ] ) , the first ase noise contribution interacts with the signal in the second span .
likewise , in the case of dpc , signal - ase noise interference accumulates from the second span onwards , since the first noise source follows the first span .
thus , for the cdc and dpc scenarios , @xmath27 for dbp , noise from the last amplifier will be backpropagated as if it were signal for @xmath2 spans .
the noise from the penultimate amplifier will have interacted with the signal for one span but will be backpropagated as if it were signal over all @xmath2 spans .
thus the signal noise interaction over the final span will be correctly compensated , leading to @xmath28 spans of excess nonlinear interference .
thus the total signal - ase noise interference is given by the following sum over all spans @xmath29 if the nlc is split between @xmath30 spans of dpc and @xmath31 spans of dbp such that the total number of spans is @xmath32 , then @xmath10 is given by @xmath33 the advantage of splitting the compensation arises since @xmath9 and @xmath8 increase superlinearly with the number of spans .
@xmath10 is minimized for @xmath34 and @xmath35 .
the snr gain due to the split nlc can be quantified using an approximated closed - form expression for the summation , in each of and .
using faulhaber s formula ( * ? ? ?
ce 332 , pg . 1 )
, such a summation can be expressed as @xmath36 where the coefficients @xmath37 are known as the bernoulli numbers .
a sufficiently accurate closed - form for @xmath38 can be derived by truncating to the first 2 terms .
these terms rapidly dominate the higher order terms as @xmath2 increases , particularly considering that @xmath39 and @xmath40 .
the snr gain , @xmath15 , for split compensation over dbp can be defined as the ratio between the snrs achieved by each compensation technique at optimum launch power .
all nlc techniques remove the cubic terms in .
thus , considering that maximizing the snr leads to the optimum launch power given by @xmath41 and that @xmath42 at the optimum power since the overall ase noise power is equal to the signal - ase interaction power , the change in snr is given by . substituting the first two terms from the approximation into , and choosing @xmath34 when calculating @xmath10 , for an asymptotically large number of spans , we obtain .
the authors wish to thank prof .
a. ellis for useful discussions and comments on an earlier draft of this letter .
12 x. liu , a. r. chraplyvy , p. j. winzer , r. w. tkach , and s. chandrasekhar , `` phase - conjugated twin waves for communication beyond the kerr nonlinearity limit , '' _ nat .
photon _ , vol
. 7 , no . 7 , pp
. 560568 , apr . 2013 .
a. d. ellis , m. e. mccarthy , m. a. z. al - khateeb , and s. sygletos , `` capacity limits of systems employing multiple optical phase conjugators , '' _ opt . express _
, vol . 23 , no . 16 , pp . 2038120393 , aug .
e. temprana , _ et al .
_ , `` two - fold transmission reach enhancement enabled by transmitter - side digital backpropagation and optical frequency comb - derived information carriers , '' _ opt . express _ , vol . 23 , no . 16 , pp .
2077420783 , aug .
2015 .
a. d. ellis , _ et al .
_ , `` the impact of phase conjugation on the nonlinear - shannon limit : the difference between optical and electrical phase conjugation , '' . _ in proc .
_ ieee summer topical meeting on nonlinear optical signal processing , nassau , bs , vol .
2 , pp . 209210 , jul . 2015 .
d. marcuse , c. r. menyuk , and p. k. a. wai , `` application of the manakov - pmd equation to studies of signal propagation in optical fibers with randomly varying birefringence , ''
_ j. lightwave technol .
_ , vol . 15 , no . 9 , pp . 17351746 , sep
. 1997 .
a. alvarado , e. agrell , d. lavery , r. maher , and p. bayvel , `` replacing the soft - decision fec limit paradigm in the design of optical communication systems , '' _ j. lightwave technol .
_ , vol .
33 , no . 20 , pp .
43384352 , oct .
d. j. ives , p. bayvel , and s. j. savory , `` adapting transmitter power and modulation format to improve optical network performance utilizing the gaussian noise model of nonlinear impairments , ''
_ j. lightwave technol .
_ , vol . 32 , no . 21 , pp . 34853494 , nov . 2014 .
m. secondini , e. forestieri , and g. prati , `` achievable information rate in nonlinear wdm fiber - optic systems with arbitrary modulation formats and dispersion maps '' _ j. lightwave technol .
_ , vol . 31 , no . 23 , pp .
38393852 , dec . | in this letter we analyze the benefit of digital compensation of fiber nonlinearity , where the digital signal processing is divided between the transmitter and receiver .
the application of the gaussian noise model indicates that , where there are two or more spans , it is always beneficial to split the nonlinearity compensation .
the theory is verified via numerical simulations , investigating transmission of single channel 50 gbd polarization division multiplexed 256-_ary _ quadrature amplitude modulation over 100 km standard single mode fiber spans , using lumped amplification . for this case , the additional increase in mutual information achieved over transmitter- or receiver - side nonlinearity compensation is approximately 1 bit for distances greater than 2000 km .
further , it is shown , theoretically , that the snr gain for long distances and high bandwidth transmission is 1.5 db versus transmitter- or receiver - based nonlinearity compensation .
coherent optical communications , quadrature amplitude modulation ( qam ) , digital backpropagation , nonlinearity compensation . |
in a metal , an external magnetic field polarizes the quasi - particles giving rise to a mismatch of the fermi wave - vectors of the different spin bands . in the presence of an attractive interaction between electrons with opposite spins ,
the ground state of the system is a homogeneous superconductor at least for sufficiently small mismatch @xmath8 . in the other limit of very large mismatches
, we expect the system to be normal ( non - superconducting ) for physically reasonable attractive interactions . an important and fruitful question in the study of superconductivity
is the nature of the zero temperature phase diagram as the mismatch is reduced in the normal phase or , alternatively , how the bcs phase disappears as @xmath0 increases . a theoretical answer to this problem
was given by fulde and ferrel @xcite and larkin and ovchinikov @xcite .
these authors have shown the appearance of an intervening inhomogeneous superconducting phase between the homogeneous bcs and the normal phases as @xmath0 increases .
this phase generally known as a fflo phase is characterized by a wave - vector dependent order parameter which oscillates in space @xcite .
different types of solutions associated with a single wave - vector can be obtained with different ground state energies @xcite . here
we consider the original fulde - ferrell ( @xmath9 ) or helicoidal solutions @xcite , such that , the order parameter is given by @xmath10 .
experimentally , fflo phases have been elusive and there is no conclusive evidence for their existence . in cold atom systems ,
two fermionic species with different densities and attractive interactions provides a controllable model system to search for fflo phases @xcite . the mismatch can be tuned by varying the number of atoms of the different species .
these are clean systems in which the strength of the interaction between the different species can also be tuned . in spite of intensive search only for nearly one - dimensional systems evidence
was found for fflo type of correlations @xcite . in transition metals , rare - earths and actinides metals , electrons arising from different orbitals ( s , p , d or f )
coexist at the fermi surface @xcite .
these electrons hybridize and the new quasi - particles have different fermi wave - vectors which among other things depend on the strength of the hybridization @xcite .
these systems and their inter - metallic compounds in the case the dominant attractive interaction is between quasi - particles in different bands are natural candidates for a fflo ground state , even in the absence of a magnetic field .
for a two - band metal , the mismatch @xmath11 is directly related to hybridization @xcite . since
this can be controlled by pressure , we have an external parameter which allows to probe the phase diagram and search for fflo phases .
inter - band pairing may be particularly important in heavy fermion systems @xcite where the dominant kondo interaction between conduction electrons and the rather local @xmath12-electrons provides also a mechanism for inter - band attraction @xcite .
notice that fflo phases induced by pressure avoid all the complications associated with orbital effects due to an external magnetic field . in this paper
we study the phase diagram of a two - dimensional ( 2d ) , two - band system as a function of the mismatch between their fermi wave - vectors controlled by the strength of their hybridization .
this is relevant for nearly two - dimensional materials with cylindrical fermi surfaces .
some of our results are in common with previous ones obtained for field induced fflo phases in @xmath13 materials at finite temperatures @xcite . as in these studies
we assume the existence of weak inter planar interactions so that the mean field bcs approach is justified at zero temperature @xcite . however , these works ignore the dynamical aspects inherent to the quantum critical phenomena .
we give here a full treatment of the quantum phase transition from the normal - to - inhomogeneous superconductor as the fermi wave - vector mismatch is reduced from the normal phase .
since this @xmath14 transition is continuous or second order , it is associated with a quantum critical point ( qcp ) at a critical value @xmath7 of the hybridization .
the instability of the normal state that we consider is that for a @xmath9 superconducting state characterized by a single wave - vector @xmath15 .
this is the first zero temperature instability as @xmath1 is reduced @xcite .
our approach yields the full thermodynamic behavior near the qcp allowing to identify the @xmath9 or helicoidal superconducting phase from its precursors effects .
the hamiltonian of the two - band system with hybridization and an inter - band interaction is given by @xcite , @xmath16 where @xmath17 ( @xmath18 ) , @xmath19 ( @xmath20 ) are creation and annihilation operators for the @xmath21 ( @xmath22 ) particles and @xmath23 are their dispersion relations , defined by @xmath24 , where @xmath25 and @xmath26 is the chemical potential of the ( non - interacting ) @xmath27-particle ( @xmath28 ) . since the inter - band interaction between particles @xmath21 and @xmath22 is attractive , we take @xmath29 . in the case of a heavy fermion system ,
the @xmath21-band is a wide band of conduction electrons and @xmath22 a narrow band of @xmath12-electrons with effective mass @xmath30 .
multi - band superconductivity was previously studied by suhl et al .
@xcite and moskalenko @xcite .
these authors differently from the present study considered intra - band interactions , with an inter - band term that transfers cooper pairs between the bands . in our case superconductivity is associated with the existence of a finite anomalous correlation function , @xmath31 .
since the conduction electrons are described by plane waves , which contain all the harmonics , this correlation function vanishes only in the normal phase .
next we calculate the zero temperature response of the two - band system to a wave - vector and frequency dependent _ fictitious field _ that couples to the superconducting order parameter @xcite of interest , @xmath31 .
the hamiltonian associated with this coupling is given by @xcite , @xmath32 where the frequency @xmath33 has a small positive imaginary part to guarantee the adiabatic switching on of the fictitious field .
the linear response of the system to this field , is obtained using perturbation theory for the retarded and advanced green s functions @xcite .
we consider that in the two - band system the hybridization is sufficiently large , such that , in this initial condition the system is in the normal phase . in this case
the superconducting order parameter is zero in the absence of the _ fictitious field
_ @xmath34 .
the superconducting response to the time and @xmath35-dependent fictitious field , within the bcs decoupling is obtained in the form @xcite , @xmath36where , @xmath37 , and the non - interacting pair susceptibility @xmath38 is given by , @xmath39 where @xmath40 and @xmath41 refer to the new hybridized bands ( see fig .
[ fig1 ] ) , @xmath42where @xmath43 is the average fermi velocity and we took , @xmath44 . for simplicity we considered a situation of perfect nesting for the unhybridized system , such that , the two bands cross the fermi surface at the same @xmath45 for @xmath46 as shown in fig . [ fig1 ] .
the mismatch of the new fermi wave - vectors of the hybridized bands is given by , @xmath47 .
in the equation for the susceptibility we can neglect spin indexes since the normal system is paramagnetic . and
@xmath48 , with a mismatch @xmath49.,width=283 ] in the normal system , at zero temperature , as the strength of the hybridization @xmath1 decreases , the condition @xmath50 in eq .
[ gsinter ] is eventually satisfied . in this case
the interacting pair susceptibility @xmath51 diverges implying that even in the absence of the fictitious field @xmath52 there is a spontaneous appearance of superconductivity in the system .
the condition @xmath50 is a generalized thouless criterion @xcite for a @xmath35-dependent superconducting instability .
this instability can be triggered , as discussed below , either by decreasing the mismatch between the fermi surfaces , i.e. reducing the hybridization , or by increasing the strength of the attractive interaction .
the quantum phase transition from the normal to the @xmath9 or helicoidal superconducting state at the critical value of hybridization @xmath7 ( or @xmath53 ) is continuous and in the three dimensional case it s universality class has been obtained @xcite .
the value of the wave - vector @xmath54 characterizing the helicoidal superconducting state for @xmath55 is that for which the condition @xmath56 is first realized .
it corresponds to that where @xmath57 determined by the condition above has a maximum ( @xmath58 is a maximum ) . due to the cylindrical symmetry of the problem any linear combination
, @xmath59 with @xmath60 has the same critical value of hybridization @xcite .
this degeneracy for @xmath3 is removed by additional interactions or non - linear terms in the gap equation @xcite .
any of these phases will have the same quantum critical behavior obtained below .
the non - interacting particle - particle , or pair dynamic susceptibility given by eq .
[ chi2d ] can be easily obtained at zero temperature .
it is given by @xmath61,\ ] ] where @xmath62 is the density of states at the fermi level of the @xmath13 system , @xmath63 and @xmath64 an energy cut - off .
the real part of the static susceptibility is given by , @xmath65 \right ] , \bar{q } \le 1 \nonumber \\ & = & \rho \left [ \ln\frac{2 \omega_c}{h } -\ln\bar{q } \right ] , \bar{q } > 1.\end{aligned}\ ] ] the thouless criterion for the appearance of superconductivity is given by , @xmath66 and coincides with the condition for the divergence of the static interacting pair susceptibility , @xmath67 in eq .
[ gsinter ] . .
this has a maximum for @xmath68 with discontinuous derivatives at this point.,width=302 ] at zero temperature , for a fixed value of the interaction and momentum , as the hybridization @xmath1 is reduced , the normal system becomes unstable to a superconducting ground state at a critical value @xmath69 determined by the above condition . using that the superconducting gap for @xmath46 can be written as , @xmath70 , we obtain for @xmath71 , @xmath72 the logarithm of this function is plotted in fig .
it is continuous and has a a sharp maximum at @xmath68 .
however , it s derivative is discontinuous and different from zero at the maximum .
the value of @xmath73 for which the instability first occurs is that for which @xmath71 has a maximum , namely , @xmath74 . at this value of @xmath73 , @xmath75
also , using that @xmath76 , we get @xmath77 , i.e. , the wave vector of the instability is exactly that connecting the two fermi surfaces at @xmath78 and @xmath48 , as shown in fig .
this is also true in @xmath79 , but not in @xmath80 @xcite . of the helicoidal superconductor
is that joining the two fermi surfaces at @xmath78 and @xmath48 , i.e. , @xmath81.,width=170 ] next we expand the denominator of the interacting dynamic pair susceptibility , eq .
[ gsinter ] , near @xmath69 , @xmath82 , and small frequencies @xmath33 . for @xmath83 and @xmath84 ,
we get : @xmath85.\ ] ] this allows us to write the gaussian part of the action describing the normal - helicoidal superconductor quantum phase transition @xcite .
it is given by , @xmath86\ !
|\delta_q(q , \omega ) |^2.\ ] ] the quantity @xmath87 measures the distance to the quantum critical point ( qcp ) associated with the normal - to helicoidal superconductor instability at @xmath88 and with @xmath89 .
since hybridization increases with pressure ( p ) , near the qcp we can write @xmath90 where @xmath91 is the critical pressure below which the @xmath35-superconductor is stable . within this gaussian approximation
, we can define a reduced correlation length @xmath92 , where @xmath21 is the lattice spacing , such that , the zero temperature correlation length exponent takes the value @xmath93 for this model .
also , we can immediately identify @xcite that the dynamic critical exponent @xmath94 that scales the energy , or frequency , at the qcp is given by @xmath95 .
even at the gaussian level this quantum phase transition is in a new universality class since the critical exponents are non - standard gaussian exponents .
for example , we find @xmath93 instead of the usual gaussian correlation length exponent @xmath96 . the free energy associated with the gaussian action in two dimensions , eq .
[ gaussian ] , is given by @xcite , @xmath97 it can be written in the scaling form , @xmath98,\ ] ] with @xmath99 , @xmath93 and @xmath95 . the characteristic temperature @xmath100 . in the non - critical side of the phase diagram in fig .
[ fig4 ] there is a crossover temperature between two different regimes . for @xmath101 ,
the free energy is given by , @xmath102.\ ] ] at low temperatures , @xmath103 , the specific heat @xmath104 has a power law behavior @xmath105 with a coefficient that increases as the qcp is approached ( @xmath106 ) .
so at low temperatures before the instability , the normal system behaves as a non - fermi liquid . for @xmath107 , the specific heat beside the @xmath105 term
has contribution @xmath108 .
the latter is the scaling contribution @xmath109 which for @xmath110 gives rise to a logarithmic term .
now let us assume that the action close to the qcp has an expansion in powers of the order parameter , @xmath111 |\delta(\vec{q } , \omega ) |^2 + \nonumber \\ & & \int \prod d\vec{q_i } \int \prod d \omega_{i } u(\{\vec{q}_i\},\{\omega_{i}\})\delta(\vec{q_1},\omega_{1})\delta(\vec{q_2},\omega_{2})\delta(\vec{q_3},\omega_{3})\delta(\vec{q_4},\omega_{4})\delta(\sum\vec{q}_i)\delta(\sum\omega_{i}),\end{aligned}\ ] ] with @xmath112 to @xmath113 and a quartic coefficient @xmath114 .
if we neglect all @xmath35 and @xmath33 dependences of this term , such that , @xmath115 , simple power counting shows that the correlation function of the order parameter scales as @xmath116 and we find the critical exponent @xmath117 .
furthermore the quartic interaction among the fluctuations scales as : @xmath118 such that these interactions are irrelevant , in the renormalization group sense , for @xmath119 .
then in the present case this implies that the gaussian action gives the correct exponents describing the quantum critical point .
however @xcite , @xmath120 is dangerously irrelevant for @xmath121 implying departures from naive scaling @xcite .
for example , the shape of the temperature dependent critical line shown in fig . [ fig4 ] is given by , @xmath122 with the shift exponent , @xmath123 , where @xmath124 is the scaling exponent of the dangerously irrelevant quartic interaction @xmath120 @xcite .
we obtain @xmath125 different from the mean - field result @xmath126 and from the naive scaling prediction @xmath127 @xcite .
the correlation length along the quantum critical trajectory ( qct ) in fig .
[ fig4 ] diverges as @xmath128 as @xmath129 @xcite , which makes manifest the dangerous irrelevant character of @xmath120 .
this is different from the naive scaling result @xmath130 which does not take into account the dangerous irrelevant nature of @xmath120 .
we have given above a complete description of the normal - to - inhomogeneous superconductor quantum phase transition at the gaussian level . as we have shown
this is the correct theory of this quantum phase transition for a constant @xmath120 quartic interaction .
however , the assumption of taking @xmath115 may not be justified , particularly in two dimensions . for the case
the frequency - dependence of @xmath131 is discarded , but its momentum dependence is considered , it has been shown that the quartic interaction obtained from an expansion of the gap equation has a diverges at the wave - vector @xmath132 of the superconducting instability @xcite . at finite temperatures
this divergence is removed and this also may occur for finite frequencies or moderate disorder .
a possible consequence of this anomalous behavior of @xmath131 is to change the order of the quantum phase transition @xcite .
finally , it is interesting to point out that the correlation length exponent of the incipient fflo phase ( @xmath133 ) being larger than that of any competing spin density wave ( @xmath134 ) @xcite , the former should override the latter instability . . the critical line @xmath135 .
qct ( dashed line ) refers to the quantum critical trajectory ( see text ) .
the line @xmath136 marks the crossover between two different non - fermi liquid behaviors , as discussed in the text .
the homogeneous superconductor part of the phase diagram for small @xmath1 and its boundary with the helicoidal or @xmath9 phase are not shown in the figure since they are not studied here.,width=302 ]
in multi - orbital metallic systems electrons from different bands coexist at the fermi surface .
these electrons have different fermi wave - vectors and this difference provides a natural scale for inhomogeneous superconductivity in the case of inter - band attraction .
the fermi wave - vectors mismatch exists even in the absence of an external magnetic field and can be tuned by external pressure which directly changes the hybridization . among the multi - orbital metals , the heavy fermion materials appear as strong candidates to exhibit the phenomenon investigated here .
the dominant interaction in these compounds is the kondo coupling between itinerant and localized electrons and this naturally leads to an attractive inter - band interaction .
many of these systems have tetragonal structures with cylindrical fermi surfaces for which the results we have obtained directly apply .
also , two - dimensions favors the inhomogeneous fflo type of phases @xcite and in addition the mechanism studied here avoids the complications due to orbital effects since it does not require an external magnetic field .
heavy fermions are in general very sensitive to external pressure and this allows to probe a large region in their phase diagrams . for @xmath137
based heavy fermion systems , in many cases , at least two superconducting phases appear as they are driven with increasing pressure from antiferromagnetic , superconducting and normal states @xcite .
the @xmath9 or helicoidal superconducting state is a candidate for the first superconducting instability as pressure is reduced from the normal state .
this transition occurs at a superconducting quantum critical point ( sqcp ) whose properties and universality class we have obtained .
the theoretical description of this quantum phase transition requires considering time and space as fundamental dimensions @xcite .
this is implied by the uncertainty principle which must remain scale invariant leading to the inextricability of these quantities .
we have characterized the thermodynamic behavior near the qcp and this should be useful for identifying the helicoidal phase from its precursor effects in actual systems . the peculiar quantum critical behavior of our model is a consequence that @xmath132 is a singular point of the pair susceptibility .
in many real systems , we can observe consequences of the structure of lindhard functions particularly in low dimensional systems , as the kohn anomaly and the peierls instability in one - dimensional conductors @xcite .
it is reasonable than to expect that our results should hold for real materials .
finally , we point out that we have not explored the phase diagram for small mismatches where a homogeneous bcs phase exists up to a critical mismatch .
this problem has been treated in refs .
@xcite .
_ note added in proof . _ recently , we became aware of the paper @xcite where they calculate the susceptibility close to the quantum normal - fflo transition .
h.c . and m a.c .
are partially supported by cnpq .
the authors also acknowledge partial support from fapemig and faperj , respectively .
m.a.c . would like to thank martin garst for useful discussions .
acknowledges the kind hospitality at cbpf where this work was done . for a review on this topic ,
see m. inguscio , w. ketterle , and c. salomon .
_ ultracold fermi gases .
proceedings of the international school of physics enrico fermi , course clxiv , varenna , 2006 . _
( ios press , amsterdam , 2008 ) .
l. mendonca ferreira , t. park , v. sidorov , m. nicklas , e. m. bittar , r. lora - serrano , e. n. hering , s. m. ramos , m. b. fontes , e. baggio - saitovich , hanoh lee , j. l. sarrao , j. d. thompson , and p. g. pagliuso , phys .
lett . 101 , 017005 ( 2008 ) . | in multi - band systems , electrons from different orbitals coexist at the fermi surface .
an attractive interaction among these quasi - particles gives rise to inter - band or hybrid pairs which eventually condense in a superconducting state .
these quasi - particles have a natural mismatch of their fermi wave - vectors , @xmath0 , which depends on the strength of the hybridization between their orbitals .
the existence of this natural scale suggests the possibility of inhomogeneous superconducting ground states in these systems , even in the absence of an applied magnetic field . furthermore , since hybridization @xmath1 depends on pressure , this provides an external parameter to control the wave - vectors mismatch at the fermi surface . in this work ,
we study the phase diagram of a two - dimensional , two - band metal with inter - band pairing .
we show that as the mismatch between the fermi wave - vectors of the two hybrid bands is reduced , the system presents a normal - to - inhomogeneous superconductor quantum phase transition at a critical value of the hybridization @xmath2 .
the superconducting ground state for @xmath3 is characterized by a wave - vector with magnitude @xmath4 . here
@xmath5 is the superconducting gap in the homogeneous state and @xmath6 the average fermi velocity .
we discuss the nature of the quantum critical point ( qcp ) at @xmath7 and obtain the associated quantum critical exponents . |
network coding ( nc ) @xcite refers to mixing different information flows at the sender or intermediate nodes in a data communication network .
it has been shown that nc can substantially improve the throughput of many wireless communication systems @xcite . as a result
, it has become a promising candidate for delivering high data rate content in future wireless communication networks .
for example , nc has been considered for delivering high data rate multimedia broadcast or multicast services ( mbms ) @xcite .
in addition to being high data rate in nature , such applications also often have strict delay requirements
. however , the higher throughput offered by nc does not necessarily translate into faster delivery of information to the application @xcite .
in general , the mixed information needs to be disentangled or network decoded first . understanding the interplay between throughput and delay and devising nc schemes that strike a balance between the two are particularly important , which has proven to be challenging @xcite .
an important example that illustrates the tension between throughput and delay is random linear network coding ( rlnc ) @xcite in broadcast erasure channels . in rlnc , the sender combines a frame or block of @xmath0 packets using random coefficients from a finite field and broadcasts different combinations until all receivers have received @xmath0 linearly independent coded packets . in this case
, rlnc achieves the best throughput ( block completion time ) among block - based nc schemes @xcite .
however , the delay performance may not be desirable , as decoding at the receivers is generally only possible after @xmath0 independent coded packets are successfully received . in order to reduce the decoding delay in nc systems ,
an attractive strategy is to employ instantly decodable nc ( idnc ) .
as the name suggests , idnc aims to provide instant packet decoding at the receivers upon successful packet reception , a property that rlnc does not guarantee .
a decoding delay occurs at a receiver when it is not targeted in an idnc transmission .
that is , it receives a packet that contains either no or more than one desired packets of that receiver . compared to rlnc , idnc in broadcast erasure channels can have a lower throughput .
in other words , idnc incurs a generally higher completion time for the broadcast of the same number of @xmath0 packets .
however , it can provide a faster delivery of uncoded packets to the application layer , as required for mbms .
therefore , similar tension between throughput and delay can also be observed in idnc .
inspired by the low - complexity xor - based encoding and decoding process of idnc and its potential application in mbms and unicast settings @xcite , in this paper we are interested in understanding the interplay between its throughput and delay over broadcast erasure channels and proposing novel idnc schemes that offer a better control of these performance metrics .
the problem of maximizing the throughput for a deadline - constrained video - streaming scenario is considered in @xcite , where each packet has a delivery deadline and has to be decoded before the deadline , otherwise it is expired . in this paper , however , we consider a block - based transmission , where all the packets in the block have to be received by all the receivers and there is no explicit packet deadline .
furthermore , in this paper , no new packet arrival is considered in the system while the transmission of a block is in progress .
in addition , this study is applicable where partial decoding is beneficial and can result in lower delays irrespective of the order in which packets are being decoded .
examples of such applications can be found in sensor or emergency networks and multiple - description source coded systems @xcite , in which every decoded packet brings new information to the destination , irrespective of its order . in this context ,
the closest works to ours are @xcite and @xcite .
in particular , the authors in @xcite aimed to improve the decoding delay of a generalized idnc scheme .
they showed that for a lower decoding delay , maximum number of receivers with the lowest packet erasure probabilities should be targeted in each idnc transmission .
in separate works @xcite , the same authors aimed to improve the completion time of idnc .
they showed that for this purpose , the receivers with the maximum number of missing packets with the highest erasure probabilities should be targeted in each idnc transmission .
a close study of @xcite reveals that trying to improve either idnc s decoding delay or completion time on its own can result in undermining the other performance metric .
in other words , while trying to improve the decoding delay , the receiver(s ) with the maximum number of missing packets may remain untargeted , which can increase the completion time . also trying to improve the completion time
may limit the total number of receivers that can be targeted in each idnc transmission , which can increase the decoding delay . to the best of our knowledge
, there is no joint control of completion time and decoding delay for idnc schemes in the literature .
thus , in this paper , our objective is to take a holistic approach , in which the completion time and decoding delay of idnc are taken into account at the same time .
in addition , we have observed that the decoding delay across various receivers in idnc schemes of @xcite and @xcite can vary significantly .
this may not be desirable in mbms or other applications which should guarantee a certain quality of service across all receivers .
these observations lead us to the following open problems : _ is there an idnc scheme that can offer a balanced performance in terms of the completion time and decoding delay and can also provide a more uniform or fair decoding delay across all receivers for the broadcast of @xmath0 packets in erasure channels ? _ to address these questions in this paper , we propose a new idnc transmission scheme which builds upon the contributions in @xcite and @xcite . at its core
, our proposed scheme recognizes that 1 ) the completion time of each individual receiver is determined not only by the number of packets it is missing , but also by the number of idnc transmissions in which it is not targeted ( while still needing a packet(s ) ) and 2 ) the overall idnc completion time is the maximum of individual completion times .
therefore , our idnc transmission scheme gives priority to the receivers that have the highest expected completion time so far .
more precisely , the priority of each receiver is the sum of two terms : the first term is its number of missing packets divided by its average packet reception probability .
this is the expected number of transmissions to serve this receiver if it is targeted in all following transmissions .
the second term is the decoding delay the receiver has experienced so far . under this scheme
, a receiver with a small number of missing packets which has remained untargeted in a number of previous transmissions may take precedence over other receivers .
hence , our scheme tends to equalize the decoding delay experience across the receivers .
furthermore , we will extend our proposed scheme to the case of broadcast erasure channels with memory @xcite , where the packet erasures occur in bursts , due to deep fading and shadowing . by following the proposed channel models in @xcite
, we model the bursts of erasures ( i.e. the memory of the channel ) by a simple two - state gilbert - elliott channel ( gec ) model and propose two algorithms that can offer an improved balance between the completion time and decoding delay of idnc for different ranges of the channel memory . with this introduction ,
we summarize the contributions and findings of our paper as follows : first , we present a holistic viewpoint of idnc .
we formulate the idnc optimal packet selection that provides an improved balance between the completion time and decoding delay for broadcast transmission over memoryless channels as an ssp problem . however , since finding the optimal packet selection in the proposed ssp scheme is computationally complex , we use the ssp formulation and its geometric structure to find some guidelines that can be used to propose a new heuristic packet selection algorithm that efficiently improves the balance between the completion time and decoding delay in idnc systems .
second , we extend the proposed packet selection algorithm to erasure channels with memory and propose two different variations of the algorithm that take into account the channel memory conditions and improve the balance between the completion time and decoding delay by selecting the packet combinations more effectively based on the channel memory conditions compared to the algorithms that are ignorant to the channel memory . finally , by taking into account both the number of missing packets and the decoding delay of the receivers , the proposed algorithm provides a more uniform decoding delay experience across all receivers .
the rest of this paper is organized as follows .
the system model is presented in section ii .
the idnc graph representation and packet generation is introduced in section iii .
section iv , presents the ssp problem formulation . in section
v , we present a geometric structure for the ssp problem that helps us to find the properties of the optimal packet selection policy . a heuristic algorithm for idnc packet selection is proposed in section vi .
the proposed heuristic algorithm is then extended to erasure channels with memory in section vii , where also a new layered algorithm is introduced .
section viii presents the simulation results .
finally , section ix concludes the paper .
the system model consists of a wireless sender that is required to deliver a block ( denoted by @xmath1 ) of @xmath0 source packets to a set ( denoted by @xmath2 ) of @xmath3 receivers .
each receiver is interested in receiving all the packets of @xmath1 .
the sender initially transmits the @xmath0 packets of the block uncoded in an _ initial transmission phase_. each sent packet is subject to erasure at receiver @xmath4 with the probability @xmath5 , which is assumed to be fixed during a block transmission period .
each receiver listens to all transmitted packets and feeds back a positive or negative acknowledgment ( ack or nak ) for each received or lost packet . at the end of the initial transmission phase ,
two `` feedback sets '' can be attributed to each receiver
@xmath4 : 1 .
the has set ( denoted by @xmath6 ) is defined as the set of packets correctly received by receiver @xmath4 .
2 . the wants set ( denoted by @xmath7 ) is defined as the set of packets that are missed at receiver @xmath4 in the initial transmission phase of the current block . in other words
@xmath8 .
the senders then stores this information in the _ state feedback matrix _ ( sfm )
@xmath9 , \forall i \in \mathcal
m , j\in \mathcal n$ ] as : @xmath10 after the initial transmission phase , a _ recovery transmission phase _ starts , in which the sender exploits the diversity of received and lost packets to transmit network coded combinations of the source packets . note that we denote the wants and has sets of receiver @xmath4 at the start of the recovery transmission phase by @xmath11 and @xmath12 , respectively . after each transmission , for each received / lost packet , the receivers send ack / nak to the sender .
this information is then used by the sender to update the sfm .
this process is repeated until all receivers obtain all packets .
similar two - phase transmission schemes have been widely considered in the literature for idnc schemes @xcite .
based on the wants and has sets information , in the recovery transmission phase , the transmitted coded packets can be one of the following options for each receiver
@xmath4 : 1 .
non - innovative packet : a packet is non - innovative for receiver @xmath4 if it contains no source packets from @xmath7 .
2 . instantly decodable packet : a packet is instantly decodable for receiver @xmath4 if it contains only one source packet from @xmath7 .
the set of receivers for which the transmitted packet is instantly decodable packet are referred to as the _
targeted receivers_. 3 .
non - instantly decodable packet : a packet is non - instantly decodable for receiver @xmath4 if it contains two or more source packets from @xmath7 .
we define the completion time and decoding delay similar to @xcite as follows : it should be noted that if receiver @xmath4 is targeted by one of its missing packets in all transmissions , in the absence of packet erasures , @xmath13 will be equal to the size of its wants set at the start of the recovery transmission phase , i.e. @xmath14 .
in this paper , we adopt idnc @xcite as our nc transmission scheme .
idnc allows the sender to transmit a coded packet that includes at most one source packet from the wants sets of the targeted receivers ( either an appropriately selected subset or if possible all receivers ) .
thus , at the targeted receivers , the packet is instantly decodable .
however , at the rest of the receivers ( referred to as _ untargeted receivers _ ) , the packet is either non - innovative or non - instantly decodable , if successfully received .
thus , the untargeted receivers will experience one unit increase of their accumulative decoding delay .
we start this section by first exploring all possible packet combinations that are instantly decodable by any subset or if possible all receivers .
all the feasible packet combinations can be represented in the form of a graph model , which was first used in the context of idnc in @xcite .
then , we will briefly review the packet selection schemes in @xcite and @xcite that were used to separately minimize idnc s oct and decoding delay , respectively . as presented in @xcite , the idnc graph @xmath15
is constructed by first inducing a vertex @xmath16 for each packet @xmath17 . in other words ,
any vertex @xmath18 represents a wanted packet @xmath19 for receiver @xmath4 .
two vertices @xmath18 and @xmath20 in @xmath21 are connected by an edge @xmath22 if any one of the following conditions is true : * c1 : * @xmath23 the two vertices are induced by the loss of the same packet @xmath19 by two different receivers @xmath4 and @xmath24 .
an edge generated by this condition does not involve any combination , but expresses the interest of the two receivers in the same packet . *
c2 : * @xmath25 the wanted packet corresponding to each vertex is in the has set of the receiver of the other vertex .
an edge generated by c2 represents a possible combination of packets @xmath19 and @xmath26 of the form @xmath27 that will be instantly decodable for receivers @xmath4 and @xmath24 .
given the graph formulation , the set of all feasible packet combinations in idnc can be expressed as the set of packet combinations defined by all maximal cliques in @xmath28 ( a _ maximal clique _ is a clique that is not a subset of any larger clique ) .
consequently , the sender can generate an idnc packet for a given transmission by xoring all the packets identified by the vertices of a selected maximal clique in @xmath28 . assuming that @xmath29 is the selected maximal clique in @xmath28 , the targeted receivers of this clique are represented by @xmath30 .
the problem of minimizing the oct of the idnc scheme for broadcast erasure channels has been studied in @xcite where it is shown that the expected ict for receiver @xmath4 , denoted by @xmath31 , if addressed in all future transmissions , can be expressed as @xmath32 .
having the expected ict of all receivers calculated , it is shown in @xcite that an efficient policy for reducing the oct should select maximal cliques that include the maximum number of vertices belonging to receivers having the largest @xmath31 . in order to simplify such maximal cliques selection in the idnc graph @xmath28 ,
the authors in @xcite proposed a maximum weight vertex search algorithm , where the weights of vertices in @xmath33 reflect the properties of their inducing receivers as follows .
let us define @xmath34 to be the adjacency indicator of vertices @xmath18 and @xmath20 in idnc graph @xmath33 such that : @xmath35 given the adjacency indicator , the weighted degree @xmath36 of vertex @xmath18 in @xcite is defined as @xmath37 .
thus , the weight of vertex @xmath38 can be defined as @xmath39 .
this expression means that a vertex has a large weight when it both belongs to a receiver with large @xmath31 value and is connected to a large number of vertices having large @xmath40 values .
the problem of minimizing the decoding delay of idnc scheme for broadcast erasure channels has been studied in @xcite where it is shown that an efficient policy for reducing decoding delay is selecting maximal cliques that include the maximum number of vertices belonging to receivers having high reception probabilities that are also connected to vertices with large reception probabilities ( i.e. low erasure probabilities ) .
thus , the weight of vertex @xmath18 , @xmath41 , in @xcite is defined as @xmath42 , where @xmath36 reflects the connection of vertex @xmath18 to vertices having large reception probabilities and is defined in @xcite as @xmath43 . unlike @xcite and @xcite , in this study ,
our goal is to propose a new packet selection policy that can provide joint control of the oct and decoding delay for idnc schemes .
in this section , we present a holistic viewpoint of idnc schemes in which the completion time and decoding delay are taken into account at the same time . by taking this viewpoint , we introduce a new idnc scheme that offers an improved balance between the oct and decoding delay performances , and at the same time provides a more uniform decoding delay experience across all receivers for the broadcast of @xmath0 packets .
the key idea here is that the ict of each receiver is not only determined by the number of its missing packets , but also the decoding delay that the respective receiver experiences .
furthermore , we note that the oct of the idnc transmission is equal to the maximum of icts
. we will use this relationship between the oct and decoding delay to design an idnc scheme that provides a balance between these two performance metrics .
the proposed scheme is then solved as an ssp problem . here , we first define @xmath44 $ ] and @xmath45 $ ] as the wants and has vectors , such that @xmath46 and @xmath47 are the cardinalities of wants and has sets at the start of recovery phase , @xmath11 and @xmath12 , respectively .
furthermore , @xmath48 $ ] is defined as the final accumulative decoding delay vector , where @xmath49 is the final accumulative decoding delay experienced by receiver @xmath4 ( i.e. the accumulative decoding delay experienced by receiver @xmath4 until it receives all its missing packets ) .
the best possible performance of idnc in terms of the oct and decoding delay can be achieved if in every single transmission all the receivers with non - empty wants sets are targeted . in this case , after each transmission , assuming that no erasure occurs , the remaining number of transmissions is reduced by one and the accumulative decoding delays experienced by the receivers are zero . under this scenario ,
the ict of each receiver is equal to the size of its initial wants set , @xmath46 , and the oct of the system is equal to the maximum ict of the receivers ( the size of the largest initial wants set , i.e. @xmath50 ) .
furthermore @xmath51 . however ,
since it is not always possible to target all the receivers with non - empty wants sets in every single transmission , due to instant decodability constraint , the receivers that are not targeted will experience a decoding delay , and thus , their icts will be increased by the value of their final accumulative decoding delay ( i.e. the total number of the time - slots that they were not targeted ) . therefore , we can write the ict of receiver @xmath4 , denoted by @xmath13 , as @xmath52 as shown in , the ict of each receiver depends on the size of its initial wants set , @xmath46 , and the final accumulative decoding delay it experiences , @xmath49 .
having defined the receivers icts , it can be easily inferred that oct of the system is equal to the maximum ict of the receivers , and can be expressed as @xmath53 it is worth noting that based on , minimizing the decoding delay of receiver @xmath4 is equivalent to minimizing its ict .
furthermore , based on , minimizing the oct is equivalent to minimizing the largest icts .
therefore , the problem of providing a balance between the decoding delay and oct can be translated into balancing between @xmath54 and @xmath55 of the receivers . in the next section
, we will show that the packet selection problem that offers such balance between the oct and decoding delay of the receivers for the idnc can be formulated in the form of an ssp problem .
the ssp problem is a special case of an infinite horizon markov decision process , which can model decision based stochastic dynamic systems with a terminating state .
ssp problem was first used in the context of idnc in @xcite in order to select the packet combinations that result in minimum completion time . in ssp problem
, different possible situations that the system could encounter are modeled as states @xmath56 ( where @xmath57 denotes the state space of the ssp problem ) . in each state @xmath56 , the system must select an action @xmath58 from an action space @xmath59 that will charge it an immediate cost @xmath60 ( @xmath61 denotes the action space of the ssp problem ) . in the general form ,
the cost of a transition from state @xmath62 to state @xmath63 is modelled as a scalar that depends on @xmath62 , the taken action @xmath58 , and @xmath63 . under this scenario , in the ssp formulation , the expected cost @xmath64 is calculated as @xmath65 , where @xmath66 represents the probability of system moving from state @xmath62 to state @xmath63 once action @xmath58 is taken .
the terminating condition of the system can be thus represented as a zero - cost _ absorbing goal state_. an ssp policy @xmath67 $ ] is a mapping from @xmath68 that associates a given action to each of the states .
the optimal policy @xmath69 of an ssp problem is the one that minimizes the cumulative mean cost until the goal state is reached .
the algorithms solving ssp problems define a value function @xmath70 as the expected cumulative cost until absorption , when the system starts at state @xmath62 and follows policy @xmath71 .
it can be recursively expressed for all @xmath56 as : @xmath72 where @xmath73 is the set of successor states to @xmath62 when action @xmath58 is taken ( i.e. @xmath74 ) consequently , the optimal policy at state @xmath62 can be defined for all @xmath56 as : @xmath75 in order to express the packet selection problem that improves the balance between the oct and decoding delay of the system for idnc in the form of an ssp problem , we need to define the following : each state @xmath62 can be characterized by its has , wants and the accumulative decoding delay vectors , @xmath76 $ ] , @xmath77 $ ] and @xmath78 $ ] , respectively .
the values of @xmath79 , @xmath80 and @xmath81 at the starting state of the recovery transmission phase , @xmath82 , are represented by @xmath83 $ ] , @xmath84 $ ] and @xmath85 $ ] , respectively , where @xmath86 and @xmath87 , @xmath88 .
furthermore , we define the absorbing state , @xmath89 , as the state in which all the receivers receive all their missing packets . in other words , the absorbing state is the final state of the recovery transmission phase in which @xmath90 $ ] .
in addition , for each state @xmath62 , we define @xmath91 to be the set of receivers who still need one or more packets .
it is worth noting that the value of the accumulative decoding delay vector at the absorbing state ; i.e. @xmath92 $ ] where @xmath93 , @xmath88 ; depends on the taken actions in all states prior to reaching the absorbing state . for each state @xmath62 ,
the action space @xmath94 consists of all possible maximal cliques in graph @xmath95 constructed from the sfm @xmath96 in state @xmath62 .
defining @xmath97 as the set of maximal cliques in @xmath95 , the cardinality of state @xmath62 action space , i.e. @xmath98 , is equal to @xmath99 .
considering the fact that each state @xmath62 can be efficiently represented by the wants sets and accumulative decoding delays of all receivers , here , we further define the _ state value _
@xmath100 for receiver @xmath4 as @xmath101 in a more general framework , this equation can be written as @xmath102 where the weight @xmath103 can be designed for more control over oct or decoding delay according to the system requirements . in the rest of this paper
, we assign equal weights to @xmath104 and @xmath105 , and consider the state value of receiver @xmath4 to be of the form @xmath106 , except stated otherwise .
furthermore , the _ state vector _ for all receivers is defined as @xmath107 $ ] .
now , the state - action transition probability @xmath66 for an action @xmath108 , can be defined based on the possibilities of the variations in @xmath100 from state @xmath62 to state @xmath63 . to define @xmath66 , here ,
we first introduce the following three sets : @xmath109 @xmath110 @xmath111 where @xmath91 denotes all the receivers with non - empty wants sets at state @xmath62 and @xmath112 represents the set of all the targeted receivers in the maximal clique @xmath113 . here , the first set consists of the receivers who have been targeted by the clique @xmath113 and their @xmath100 have been decreased from state @xmath62 to state @xmath63 .
this means that these receivers have successfully received an idnc packet , which addressed them by one of their missing packets .
thus , the size of their wants sets is reduced and their accumulative decoding delays are remained unchanged .
the second set includes the receivers who have not been targeted but have successfully received the transmitted packet . in this case
@xmath100 is increased from state @xmath62 to state @xmath63 , since the wants sets of these receivers have remained unchanged and their accumulative decoding delays have increased due to successfully receiving either a non - innovative or a non - instantly decodable packet .
the third set includes the receivers who have not received any packet due to packet erasure and as a result , their wants sets and accumulative decoding delays have remained unchanged , thus @xmath114 .
based on the definitions of these three sets , @xmath66 can be expressed as follows : @xmath115 of the example sfm in , title="fig : " ] of the example sfm in , title="fig : " ] the best possible action is the action that addresses all the receivers with non - empty wants sets at state @xmath62 , denoted by @xmath91 , by one of their missing packets . under this scenario , assuming no erasure occurs , the wants sets of all the receivers are reduced from state @xmath62 to state @xmath63 and their accumulative decoding delays remain unchanged ( i.e. @xmath116 and @xmath117 ) . in this case , for each receiver @xmath4 we will have @xmath118 .
this is the best performance that can be achieved for an idnc scheme . knowing that any transition ( due to any action ) takes one packet transmission , the cost of action @xmath58 on each receiver
@xmath4 can be defined as @xmath119 .
this results in three possible cost values , i.e. @xmath120 , associated with action @xmath58 on receiver @xmath4 that can be expressed as follows : * @xmath121 means that action @xmath58 does not incur any cost on receiver @xmath4 in terms of its wants set and accumulative decoding delay , if it successfully receives one of its missing packets . in this case ,
@xmath122 and @xmath123 .
* @xmath124 means that receiver @xmath4 ( targeted / untargeted ) did not receive the coded packet due to packet erasure . in this case
, there is no cost on the accumulative decoding delay , however , the wants set of receiver @xmath4 remains unchanged , as no missing packet was decoded . here
, at least one more time - slot ( one transmission ) is required to be able to reduce the size of receiver @xmath4 s wants set . under this scenario , @xmath125 and @xmath126 .
* @xmath127 means that receiver @xmath4 was not targeted by action @xmath58 and has successfully received either a non - instantly decodable or a non - innovative packet . in this case , there are costs on both the accumulative decoding delay and wants set of receiver @xmath4 , as it experiences an increase in its accumulative decoding delay and the size of its wants set remains unchanged .
as a result @xmath128 and @xmath129 . based on the above discussion ,
if receiver @xmath4 is targeted by action @xmath58 , i.e. @xmath130 , the cost will be @xmath131 @xmath132 thus , the expected cost given receiver @xmath4 is targeted by action @xmath58 can be calculated as @xmath133 however , if receiver @xmath4 is not targeted by action @xmath58 , i.e. @xmath134 , the cost will be @xmath135 @xmath136 thus , the expected cost given receiver @xmath4 is not targeted by action @xmath58 can be calculated as @xmath137 the total expected cost of action @xmath58 over all the receivers in @xmath91 can thus be defined as @xmath138 @xmath139 the optimal policy as presented in section [ ssp_problem ] can be expressed as @xmath140\}}\end{aligned}\ ] ] where @xmath141 is the expectation operator over different transmission probabilities when action @xmath58 is taken .
thus , the optimal action at state @xmath62 is the action that minimizes the cost as well as the expectation of the optimal value functions of the successor states .
however , solving this ssp problem is computationally complex and requires exhaustive iterative techniques @xcite .
furthermore , there is no closed - form solution to this problem .
thus , instead of solving the ssp problem formulated in , we can study its properties and structure to draw the characteristics of the optimal policy . to this end , we will study the geometric structure of the ssp solution in the context of the proposed idnc scheme .
in other words , our aim of the ssp formulation is not to use it as a solution , but to study its properties by the help of its geometric structure and find some guidelines for policies that can improve the balance between the oct and decoding delay in idnc systems .
we then use these policies to design simple yet efficient heuristic algorithms in section [ sec : heuristicalg ] .
in order to find some guidelines for the policies that can efficiently improve the balance between the oct and decoding delay in idnc systems , in this section , we study the geometric structure of the ssp problem . given the representation of the ssp problem in each state @xmath62 by the state vector of the receivers @xmath107 $ ] , we can now explain the geometric structure of the problem as follows .
first , we consider an @xmath3-dimensional cartesian space , and assign to each point @xmath142 $ ] in this space all the states that have the state vectors @xmath107 $ ] equal to the coordination of this point .
although many states can have the same state vector , these states differ from one another by their sfms .
the absorbing state is the state for which all @xmath143 . under the special scenario where the accumulative decoding delays are zero , i.e. @xmath144
, the absorbing state is located in the origin of the considered @xmath3-dimensional cartesian space . however , in general , the decoding delays experienced by the receivers until arriving at the absorbing state can be non - zero positive integers , and consequently the absorbing point will not necessarily be located in the origin of the space .
after each transmission , if the packet is successfully received at a receiver , there are two possibilities , 1 ) it is instantly decodable , and thus @xmath145 , 2 ) it is either non - instantly decodable or non - innovative , and thus @xmath146 . however , if the packet is not received at the receiver , then @xmath114 .
therefore , it can be easily concluded that the system can at most move from point @xmath147 to another point @xmath148 which is a vertex in the hypercube @xmath149 defined as : @xmath150 in other words , @xmath149 is the hypercube of side length 1 , in which @xmath151 and @xmath152 are two of the corners . here , we start with the geometric structure of the ssp problem in the erasure - free case and then extend it to the case with erasures .
in the erasure - free case , since transmitted packets are always successfully received by the receivers , depending on the received packet being instantly decodable or not , we will have @xmath145 or @xmath146 . under this scenario ,
@xmath151 and @xmath152 are always two diagonal corners in the hypercube @xmath149 , i.e. @xmath153 . under the erasure - free scenario , it can be shown that at any state @xmath62 choosing the action that transitions the system to the opposite diagonal point in the @xmath154-dimensional hypercube , for which @xmath122 and thus @xmath155 , would not adversely affect the optimality of future decisions .
this is due to the fact that all packets in this action will be received by all receivers and therefore , they would not contribute to any future cost .
[ exm : subcase1 ] however , such zero - cost actions do not always exist in most states , due to the instant decodability constraint .
consequently , we need a method to find efficient actions that provide an improved balance between the oct and decoding delay . in the absence of an action with zero total cost , in order to find efficient actions that provide an improved balance between the oct and decoding delay , we consider the geometric structure in the following example .
[ exm : subcase2 ] it is worth noting that in the above example , the @xmath156 norm ( euclidian distance ) is used to represent a state s closeness to the origin .
the above discussion can be summarized as the following remark . [
remark : guidelines ] furthermore , it can also be easily inferred that having higher priorities for receivers with larger values of @xmath106 can potentially result in a lower variance of the decoding delay experienced by the receivers in the system .
it means that when the decoding delay of a receiver increases , the value of @xmath100 also increases , and as a result of that the respective receiver will be given a higher priority .
this can also be translated into improving the decoding delay fairness among the receivers while minimizing the oct of the system .
the simulation results on the variance of the decoding delay across all receivers are represented in section [ sec : results ] . due to the nature of wireless broadcast systems and the fact that the sfm changes probabilistically after each transmission as a result of packet erasures , in this paper , we design the idnc packet dynamically according to the received feedback in each time slot . under this scenario , since the packet erasures are not known ahead of time , our approach is a greedy - based algorithm in which at each transmission based on the updated sfm , a single coded packet is designed ( guided by remark 2 above ) .
it is worth noting that this greedy scheme does not necessarily result in a globally optimal policy stated in . for erasure channels , the effect of packet erasures
should be reflected on the geometric structure of the problem .
let @xmath4 and @xmath24 be two receivers having the same wants set size , but @xmath157 .
consequently , receiver @xmath4 will require on average more targeting attempts compared to receiver @xmath24 in order to deplete its wants set .
since we assume that erasure probabilities do not change during the transmission of a block , targeting receiver @xmath24 and ignoring receiver @xmath4 is expected to result in a higher oct , especially when @xmath158 is among the largest values in @xmath159 . according to these facts and the above discussion in subcase 2 , receiver @xmath4
should be given a higher priority of service than receiver @xmath24 . in order to implement the above prioritization
, we define a channel weighted wants value as @xmath160 , and consequently @xmath161 $ ] , where @xmath162 based on this new vector definition , we can re - define our space such that the points @xmath163 are identified by the coordinates of the vectors @xmath164 instead of @xmath165 . in this case , the actions move the system within hyper - rectangles @xmath166 with sides either equal to 1 or @xmath167 in the @xmath4-th dimension .
it means if an action results in an increase in the accumulative decoding delay , then @xmath168 , however , if it addresses one of the receiver @xmath4 s missing packets , it leads to @xmath169 .
moreover , if receiver @xmath4 does not receive the packet due to erasure , then @xmath170 . in other words :
@xmath172 in the next section , by the help of the above - mentioned design guidelines , we will propose a heuristic packet selection algorithm .
in this section , we propose a greedy algorithm to select the clique according to the findings in the previous section .
we use @xmath156 norm here , but other norms are also possible .
the proposed algorithm performs clique selection , using a maximum weight vertex search approach . for this search to be efficient in finding maximal cliques
, the vertices weights must not only reflect the @xmath173 values of their inducing receivers , but also their adjacency to the vertices with high @xmath174 .
we then define the weighted degree of vertex @xmath18 , denoted by @xmath175 , as : @xmath176 where @xmath34 was defined in .
thus , a large weighted degree reflects its adjacency to a large number of vertices belonging to receivers with large values of @xmath174 .
we finally design the vertex weight @xmath177 for vertex @xmath18 as : @xmath178 consequently , a vertex has a high weight if it both belongs to a receiver with large @xmath173 , and is connected to the receivers with large @xmath174 values .
based on the above weight definition , we introduce our proposed packet selection algorithm as follows . in each state @xmath62 ,
the algorithm starts by selecting the vertex with the maximum weight , denoted by @xmath179 , and adds it to the clique @xmath180 .
note that at first , @xmath180 is an empty set . then at each following iteration
, the algorithm first recomputes the new vertices weights within the subgraph connected to all previously selected vertices in @xmath180 , denoted by @xmath181 , then adds the new vertex with the maximum weight to it .
the algorithm stops when there is no further vertex connected to all vertices in @xmath180 .
we refer to this algorithm as _ maximum weight vertex search algorithm _ ( mwvs ) .
the proposed algorithm is summarized in algorithm [ alg : mwvs ] . 1 .
* initialize * @xmath182 + construct @xmath95 based on @xmath96 .
while * @xmath183 do + compute @xmath184 using , and .
+ select @xmath185 . + set @xmath186 .
+ update subgraph @xmath187
in this section , our goal is to extend our proposed mwvs scheme to the coded transmissions in erasure channels with memory . to model erasure channels with memory , we employ the well - known gilbert - elliott channel ( gec ) @xcite which is a markov model with a _ good _ and a _ bad _ state .
when the channel is in the good state packets can be successfully received , and when the channel is in the bad state packets are lost ( e.g. , due to deep fades in the channel ) .
the probability of moving from the good state @xmath188 to the bad state @xmath189 is @xmath190 and the probability of moving from the bad state @xmath189 to the good state @xmath188 is @xmath191 .
steady - state probabilities are derived as @xmath192 and @xmath193 , where @xmath194 is the channel state of receiver @xmath4 in the previous transmission . here , without loss of generality , we assume that @xmath195 , which results in equiprobable states in the steady - state regime .
other scenarios can be considered in a similar manner .
following @xcite , we define the memory content of the gec as @xmath196 , which signifies the persistence of the channel in remaining in the same state .
a small @xmath197 means a channel with little memory and a large @xmath197 means a channel with large memory .
we assume that different receivers links are independent of each other with the same state transition probabilities . here
, the proposed mwvs algorithm in section [ sec : heuristicalg ] is modified so that it takes into account the channel memory conditions . in the modified framework , the positive or negative acknowledgment ( ack or nak )
that each receiver feeds back for each received or lost packet can be utilized to infer the channel state of that receiver in the previous transmission .
the proposed mwvs algorithm in section [ sec : heuristicalg ] can then be generalized for erasure channels with memory by defining the probability of successful reception by the receiver @xmath4 as the probability of moving to the good state @xmath188 in the current time - slot from its previous state @xmath194 , i.e. @xmath198 . so the proposed mwvs algorithm can be easily implemented in erasure channels with memory by replacing @xmath199 with @xmath198 in as @xmath200 in other words , the weight of each vertex in can now be recalculated based on the conditional reception probability of its inducing receiver , given its previous state , as @xmath201 @xmath202 ^ 2\theta_{i , j}(s)\end{aligned}\ ] ] however , for erasure channels with strong memory , the receivers have a strong tendency to stay in their previous states .
it means if they have been in state @xmath188 in the previous time - slot , they are most likely to stay in state @xmath188 in the current time - slot , and vice versa , if they have been in state @xmath189 , they are most likely to stay in state @xmath189 . under this case , for the receivers in state b , @xmath203 will be very small and as a result the term @xmath204 in will be large .
consequently , high weights will be given to the receivers that have been in state @xmath189 in the previous transmission ( also referred to as _ bad - channel receivers _
( bcr ) ) .
but it should be noted that targeting the bcrs most likely would not result in any decoding for them , as with a very high probability their channels will remain in state @xmath189 in the current transmission .
however , addressing the receivers that were in sate @xmath188 in the previous transmission ( also referred to as _ good - channel receivers _ ( gcr ) ) can potentially result in the decoding of their missing packets .
inspired by these scenarios , in the next sub - section , we will introduce a _ layered maximum weight vertex search algorithm _ ( referred to as mwvs - layered ) , which is specifically designed for erasure channels with persistent memory . here
, our goal is to extend the proposed mwvs algorithm in section [ subsec : mwvsmemory ] for erasure channels with persistent memory . in order to do so
, we follow the same approach as in @xcite .
the proposed algorithm comprises two different layers of subgraphs .
the first layer of subgraph , @xmath205 , consists of vertices of gcrs . in the first step , the mwvs algorithm is applied on the subgraph @xmath206 , and @xmath207 is obtained .
then , in the second step , the algorithm finds @xmath208 by applying the mwvs algorithm another time on the second layer of subgraph , @xmath209 , consisting of bcrs that are adjacent to all the vertices of the chosen clique @xmath207 .
thus , the final clique can be obtained by the union of the cliques from the two layers as @xmath210 .
the steps of mwvs - layered algorithm are summarized in algorithm [ alg : lmwvs ] . 1 .
* initialize * @xmath211 and @xmath212 + construct @xmath95 based on @xmath96 . + form @xmath206 and @xmath209 according to the channels previous states @xmath213 .
2 . * while * @xmath214 do + compute @xmath215 using , and .
+ select @xmath216 . + set @xmath217 .
+ update subgraphs @xmath218 and @xmath219 .
while * @xmath220 do + compute @xmath221 using .
+ select @xmath222 . + set @xmath223 .
+ update subgraph @xmath219 .
4 . @xmath224
in this section , we present the simulation results comparing the performance of our proposed mwvs and mwvs - layered algorithms and the schemes in @xcite over a wide range of channel memory conditions . furthermore , as our benchmark for the minimum oct performance
, we will compare the oct of our proposed mwvs and mwvs - layered algorithms with the rlnc scheme .
we start with our simulation results for memoryless erasure channels and compare the performance of our proposed mwvs algorithm with the schemes in @xcite and @xcite , denoted by `` min - oct '' and `` min - dd '' , respectively .
furthermore , we have simulated the proposed scheme for @xmath225 and @xmath226 , denoted by mwvs ( @xmath225 ) " and mwvs ( @xmath227 ) " , respectively .
@xmath225 corresponds to the case that the objective of the proposed scheme is to reduce the accumulative decoding delay and @xmath227 corresponds to the case where the objective of the proposed scheme is to reduce the oct of the system in each time slot .
the simulation results of the proposed mwvs algorithm when equal weights are assigned to @xmath104 and @xmath105 , as in , are denoted by mwvs " . in our simulations for the broadcast memoryless erasure channels , we assume that packet erasures of different receivers change from block to block in the range @xmath228 $ ] with an average equal to 0.15 .
the simulations are performed for different number of packets and receivers in the system .
it should be noted that the presented simulation results in this section are the mean values , i.e. the oct results show the average oct of the transmission of @xmath0 packets over 500 instances of sfm . in terms of the decoding delay ,
the mean decoding delay of different receivers are computed per block , and then these mean decoding delays are averaged over 500 instances of sfm . hence , the decoding delay results are actually the mean of mean decoding delays . figure [ fig : octvsdd](a ) depicts the oct and decoding delay tradeoff curves of different algorithms for various number of packets @xmath0 for @xmath229 receivers . moreover ,
the oct and decoding delay tradeoff curves of these algorithms for various number of receivers @xmath3 for @xmath230 packets is presented in figure [ fig : octvsdd](b ) . from these figures
, we first observe that the min - oct algorithm in @xcite that achieves the minimum oct among the idnc schemes in figures [ fig : octvsdd](a ) and [ fig : octvsdd](b ) , results in the worst decoding delay performance , and the min - dd algorithm in @xcite that achieves the minimum decoding delay performance , results in the worst oct performance .
however , in these figures it is shown that our proposed mwvs algorithm provides an improved balance between the oct and decoding delay for the whole range of number of packets and receivers . furthermore ,
as it was expected , we observe that the performance of the proposed mwvs algorithm with @xmath227 is the same as the performance of min - oct algorithm proposed in @xcite .
also , it can be seen that the performance of the proposed algorithm with @xmath225 is very close to the performance of the min - dd algorithm proposed in @xcite .
however , it is worth noting that the proposed mwvs algorithm when @xmath225 aims to reduce the accumulative decoding delay ( defined in definition 4 ) , while the min - dd algorithm in @xcite aims to reduce the decoding delay in each time - slot ( defined in definition 3 ) .
figure [ fig : varddvsk ] illustrates the variance of the decoding delay versus the number of packets @xmath0 for @xmath229 receivers . from this figure , it can be seen that our proposed mwvs algorithm significantly outperforms the other algorithms in terms of the variance of the decoding delay .
this can be translated into a better fairness in the decoding delay experienced by different receivers . for erasure channels with memory ,
the full graph search and the layered graph search algorithms proposed in @xcite are used as our reference for the minimum decoding delay performance .
these algorithms are denoted by `` min - dd '' and `` min - dd - layered '' in the figures , respectively . as our reference for the minimum oct performance for erasure channels with memory ,
we have modified the algorithm in @xcite to become channel memory aware by replacing the probability of successful reception at receiver @xmath4 with @xmath198 .
we refer to this scheme as `` min - oct '' .
furthermore , we have extended this scheme to a two - layered algorithm , where the first layer consists of gcrs and the second layer consists of bcrs . in the first step ,
the algorithm is applied on the first layer and a clique of gcrs is obtained .
then , in the second step , the algorithm is applied to the second layer and a clique of bcrs that are adjacent to all the vertices of the chosen clique of gcrs is found .
then , the final clique is obtained by the union of the cliques from the two layers . in our simulation results ,
this scheme is referred to as `` min - oct - layered '' . for the broadcast erasure channels with memory , we assume @xmath231 for all the receivers , and the channel memory , @xmath232 , ranges from 0 ( memoryless ) to 0.98 ( very persistent memory ) .
the simulation results are provided for a wide range of channel memory contents as well as different number of packets and receivers .
figures [ fig : cdvsmemory](a ) and [ fig : cdvsmemory](b ) illustrate the oct of the receivers versus channel memory for @xmath233 packets and receivers , respectively . as can be seen from these figures , for low channel memory content ( roughly ranging from 0 to 0.45 ) ,
the full graph algorithms outperform their layered graph counterparts in terms of oct .
however , when the memory content of the channel is high ( roughly ranging from 0.45 - 0.98 ) , the layered graph techniques significantly outperform their full graph counterparts .
the mean decoding delay performance versus channel memory is depicted in figure [ fig : ddvsmemory ] . from this figure
we can see that in terms of the decoding delay , the min - dd algorithm outperforms the min - dd - layered for memory content ranging from 0 to 0.5 , while the min - dd - layered outperforms min - dd for higher channel memory contents ( ranging from 0.5 to 0.98 ) .
for all the other investigated schemes , the layered graph techniques always result in lower decoding delays compared to their full graph counterparts .
this is due to the fact that in the layered graph techniques , the priority is always given to the gcrs to be addressed by one of their missing packets , and as shown in @xcite giving higher priorities to the receivers with higher probabilities of successful reception improves the decoding delay experienced by the receivers .
furthermore , as shown in these figures , the proposed mwvs - layered scheme provides a better balance between the oct and decoding delay for the whole range of channel memory content .
figure [ fig : memoryzeropointsix ] shows the oct and decoding delay tradeoff curves of the system for different number of packets @xmath0 for channel memory @xmath234 and @xmath229 .
the results show that for @xmath234 the layered graph techniques outperform their full graph counterparts .
again it can be seen that the min - oct - layered algorithm that achieves the lowest oct among the idnc schemes results in the worst mean decoding delay among the layered graph algorithms , and the min - dd - layered algorithm that achieves the lowest decoding delay results in the worst oct . however , for @xmath234 as we expected , the proposed mwvs - layered algorithm results in an improved balance between the oct and decoding delay .
in this paper , we proposed a new holistic viewpoint of instantly decodable network coding ( idnc ) schemes that simultaneously takes into account both the overall completion time ( oct ) and decoding delay and improves the balance between these two performance metrics for broadcast transmission over erasure channels with a wide range of memory conditions .
we formulated the optimal packet selection for such systems using an ssp technique .
however , since solving the ssp problem in the proposed scheme is computationally complex , we further proposed two different heuristic algorithms that each improves this balance between the oct and decoding delay for a specific range of channel memory conditions . furthermore , it was shown that the proposed scheme offers a more uniform decoding delay experience across all receivers .
extensive simulations were conducted to assess the performance of the proposed algorithms compared to the best known existing algorithms in the literature .
the simulation results show that our proposed algorithms achieve an improved balance between the oct and decoding delay .
x. li , c. wang , and x. lin , `` on the capacity of immediately - decodable coding schemes for wireless stored - video broadcast with hard deadline constraints , '' _ ieee journal on selected areas in communications _ ,
29 , no . 5 , pp .
10941105 , may 2011 .
s. sorour and s. valaee , `` an adaptive network coded retransmission scheme for single - hop wireless multicast broadcast services , '' _ ieee / acm transactions on networking _ , vol .
19 , no . 3 , pp .
869878 , jun . 2011 .
r. costa , d. munaretto , j. widmer , and j. barros , `` informed network coding for minimum decoding delay , '' in _ proc .
ieee international conference on mobile ad hoc and sensor systems ( mass ) _ , 2008 , pp .
8091 .
yeow , a. t. hoang , and c .- k .
tham , `` minimizing delay for multicast - streaming in wireless networks with network coding , '' in _ proc .
ieee conference on computer communications ( infocom ) _ , apr .
2009 , pp .
190198 .
j. k. sundararajan , p. sadeghi , and m. mdard , `` a feedback - based adaptive broadcast coding scheme for reducing in - order delivery delay , '' in _ proc .
workshop on network coding , theory , and applications ( netcod ) _ , 2009 ,
a. yazdi , s. sorour , s. valaee , and r. kim , `` optimum network coding for delay sensitive applications in wimax unicast , '' in _ proc .
ieee conference on computer communications ( infocom ) _ , apr .
2009 , pp .
25762580 .
p. sadeghi , r. shams , and d. traskov , `` an optimal adaptive network coding scheme for minimizing decoding delay in broadcast erasure channels , '' _ eurasip journal on wireless communications and networking _ , pp .
114 , jan . 2010 .
b. swapna , a. eryilmaz , and n. shroff , `` throughput - delay analysis of random linear network coding for wireless broadcasting , '' in _ proc .
international symposium on network coding ( netcod ) _ , jun .
2010 , pp . 16 .
m. nistor , d. e. lucani , t. t. v. vinhoza , r. a. costa , and j. barros , `` on the delay distribution of random linear network coding , '' _ ieee journal on selected areas in communications _ , vol .
29 , no . 5 , pp .
10841093 , may 2011 .
p. sadeghi and m. yu . instantly decodable versus random linear network coding : a comparative framework for throughput and decoding delay performance .
[ online ] .
available : \url{http://arxiv.org / abs/1208.2387}[\url\{http://arxiv.org / abs/1208.2387 } ] t. ho , m. mdard , r. koetter , d. karger , m. effros , j. shi , and b. leong , `` a random linear network coding approach to multicast , '' _ ieee transactions on information theory _
52 , no .
10 , pp . 44134430 , 2006 .
s. wang , c. gong , x. wang , and m. liang , `` instantly decodable network coding schemes for in - order progressive retransmission , '' _ ieee communications letters _
17 , no . 6 , pp . 10691072 , jun .
a. le , a. s. tehrani , a. g. dimakis , and a. markopoulou , `` instantly decodable network codes for real - time applications , '' in _ proc .
workshop on network coding , theory and applications ( netcod ) _ , 2013 .
p. sadeghi , r. a. kennedy , p. rapajic , and r. shams , `` finite - state markov modeling of fading channels : a survey of principles and applications , '' _ ieee signal processing magazine _ , vol . 25 , no . 5 , pp . 5780 , sep .
m. s. karim and p. sadeghi , `` decoding delay reduction in broadcast erasure channels with memory for network coding , '' in _ proc .
ieee international symposium on personal , indoor and mobile radio communications ( pimrc ) _ , sep .
2012 , pp .
s. sorour , n. aboutorab , p. sadeghi , m. s. karim , t. al - naffouri , and m. alouini , `` delay reduction in persistent erasure channels for generalized instantly decodable network coding , '' in _ proc .
ieee vehicular technology conference ( vtc ) _
, jun . 2013 , pp . 15 . | this paper studies the complicated interplay of the completion time ( as a measure of throughput ) and the decoding delay performance in instantly decodable network coded ( idnc ) systems over wireless broadcast erasure channels with memory , and proposes two new algorithms that improve the balance between the completion time and decoding delay of broadcasting a block of packets .
we first formulate the idnc packet selection problem that provides joint control of the completion time and decoding delay as a statistical shortest path ( ssp ) problem .
however , since finding the optimal packet selection policy using the ssp technique is computationally complex , we employ its geometric structure to find some guidelines and use them to propose two heuristic packet selection algorithms that can efficiently improve the balance between the completion time and decoding delay for broadcast erasure channels with a wide range of memory conditions .
it is shown that each one of the two proposed algorithms is superior for a specific range of memory conditions .
furthermore , we show that the proposed algorithms achieve an improved fairness in terms of the decoding delay across all receivers .
instantly decodable network coding , decoding delay , completion time , broadcast , gilbert - elliott channels . |
neutrino oscillations in the standard three - flavor scheme are described by three mixing angles , @xmath4 , @xmath5 , @xmath6 , one cp phase @xmath7 , and two independent mass - squared differences , @xmath8 and @xmath9 .
thanks to the recent progress of the experiments with solar , atmospheric , reactor and accelerator neutrinos , the three mixing angles and the two mass - squared differences have been determined .
the only oscillation parameters which are still undetermined are the value of the cp phase @xmath7 and the sign of @xmath9 ( the mass hierarchy ) . in the future neutrino long - baseline experiments with intense neutrino beams the sign of @xmath9 and @xmath7 are expected to be determined@xcite . as in the case of b factories
, such high precision measurements will enable us to search for deviation from the standard three - flavor oscillations ( see , e.g. , ref.@xcite ) . among such possibilities , in this talk
, we will discuss the effective non - standard neutral current flavor - dependent neutrino interaction with matter , given by @xmath10 where @xmath11 and @xmath12 stand for fermions ( the only relevant ones are electrons , u and d quarks ) , @xmath13 is the fermi coupling constant , and @xmath14 stands for a projection operator that is either @xmath15 or @xmath16 . if the interaction ( [ nsiop ] ) exists , then the standard matter effect is modified .
we will discuss atmospheric neutrinos which go through the earth , so we make an approximation that the number densities of electrons ( @xmath17 ) , protons , and neutrons are equal . defining @xmath18 , the hermitian @xmath19 matrix of the matter potential becomes @xmath20 where @xmath21 stands for the matter effect due to the charged current interaction in the standard case . with this matter potential ,
the dirac equation for neutrinos in matter becomes @xmath22 \left ( \begin{array}{c } \nu_e(x ) \\
\nu_{\mu}(x ) \\
\nu_{\tau}(x ) \end{array } \right ) , \label{eqn : sch}\end{aligned}\ ] ] where @xmath23 is the leptonic mixing matrix , and @xmath24 .
constraints on @xmath25 have been discussed by many authors ( see , e.g. , refs.@xcite , @xcite and references therein ) , and it is known that the bounds on @xmath0 , @xmath1 and @xmath2 are much weaker than those on @xmath26 . taking into account the various constraints , in the present
talk we take the ansatz @xmath27 and analyze the sensitivity to the parameters @xmath28 of the atmospheric neutrino experiment at superkamiokande ( sk ) and the future hyperkamiokande ( hk ) facility@xcite .
this talk is based on the work@xcite and the readers are referred to ref.@xcite for details .
first let us discuss the constraint of the sk atmospheric neutrino experiment on the non - standard interaction in propagation with the ansatz ( [ ansatz ] ) .
the independent degrees of freedom in addition to those in the standard oscillation scenario are @xmath0 , @xmath29 and @xmath30 . we have performed a @xmath31 analysis for the sk atmospheric neutrino data for 4438 days@xcite . ,
@xmath29 ) plane from the sk atmospheric neutrino data for a normal ( inverted ) mass hierarchy .
, title="fig : " ] , @xmath29 ) plane from the sk atmospheric neutrino data for a normal ( inverted ) mass hierarchy .
, title="fig : " ] the result for the superkamiokande data for 4438 days is given in fig.[fig : fig1 ] , where the allowed region is plotted in the ( @xmath0 , @xmath29 ) plane .
it is understood that , at each point in this plane , @xmath31 is marginalized with respect to the parameters @xmath6 , @xmath32 , @xmath7 and @xmath30 .
all other oscillation parameters are fixed for simplicity and the reference values used here are the following : @xmath33 , @xmath34 , @xmath35 . the best - fit point for
the normal ( inverted ) hierarchy is ( @xmath0 , @xmath29 ) = ( -1.0 , 0.0 ) ( ( 3.0 , 1.7 ) ) and the value of @xmath31 at this point is 79.0 ( 78.6 ) for 50 degrees of freedom , and goodness of fit is 2.8 ( 2.7 ) @xmath36cl , respectively .
the best - fit point is different from the standard case ( @xmath0 , @xmath29 ) = ( 0 , 0 ) , and this may be because we use only the information on the energy rate and the sensitivity to nsi is lost due to the destructive phenomena between the lower and higher energy bins ( see the discussions in subsect.[hk - std ] ) .
the difference of the value of @xmath31 for the standard case and that for the best - fit point for the normal ( inverted ) hierarchy is @xmath37 ( 3.4 ) for 2 degrees of freedom ( 1.1 @xmath36cl ( 1.3 @xmath36cl ) ) , respectively , and the standard case is certainly acceptable for the both mass hierarchies in our analysis . from the fig.[fig : fig1 ] we can read off the allowed region for @xmath38 , and we conclude that the allowed region for @xmath39 is approximately @xmath40
let us now discuss the potential sensitivity of hk to @xmath0 and @xmath29 .
since hk is a future experiment , the simulated numbers of events are used as `` the experimental data '' , and we vary @xmath0 and @xmath1 as well as the standard oscillation parameters trying to fit to `` the experimental data '' . here
we perform an analysis on the assumption that we know the mass hierarchy , because some hint on the mass hierarchy is expected to be available at some confidence level by the time hk will accumulate the atmospheric neutrino data for twenty years .
first of all , let us discuss the case where `` the experimental data '' is the one obtained with the standard oscillation scenario .
the values of the oscillation parameters which are used to obtain `` the experimental data '' are the following : @xmath41 , @xmath42 , @xmath43 , @xmath44 , @xmath45 , @xmath46 , where the bar notation stands for the reference values for `` the experimental data '' .
as in the case of the analysis of the sk data , we vary the oscillation parameters @xmath6 , @xmath47 , @xmath7 and arg(@xmath1 ) while fixing the other oscillation parameters @xmath33 , @xmath34 and @xmath48ev@xmath49 . , @xmath29 ) plane from the hk atmospheric neutrino data for a normal ( inverted ) mass hierarchy from the energy - rate ( two energy - bin ) analysis . , title="fig : " ] , @xmath29 ) plane from the hk atmospheric neutrino data for a normal ( inverted ) mass hierarchy from the energy - rate ( two energy - bin ) analysis .
, title="fig : " ] , @xmath29 ) plane from the hk atmospheric neutrino data for a normal ( inverted ) mass hierarchy from the energy - rate ( two energy - bin ) analysis .
, title="fig : " ] , @xmath29 ) plane from the hk atmospheric neutrino data for a normal ( inverted ) mass hierarchy from the energy - rate ( two energy - bin ) analysis .
, title="fig : " ] we have performed both energy rate and energy spectrum ( with two energy bins ) analyses , and the results from the energy rate ( spectrum ) analysis are given by the upper ( lower ) panel in fig.[fig : fig2 ] . from the energy rate analysis
we have @xmath50 at 2.5@xmath36cl .
on the other hand , from the energy spectrum analysis we get @xmath51 and @xmath52 at 2.5@xmath36 ( 98.8% ) cl for the normal hierarchy and to @xmath53 and @xmath54 at 2.5@xmath36 ( 98.8% ) cl for the inverted hierarchy . from fig.[fig : fig2 ] we note that the allowed regions from the energy rate analysis are much larger than those from the energy spectrum analysis for both mass hierarchies , and that the allowed regions for the inverted hierarchy are larger than those for the normal hierarchy for both rate and spectrum analyses .
this is because we lose the significance ( particularly for the high energy bin in the spectrum analysis and for the inverted hierarchy case both in rate and spectrum analyses ) .
we believe that this is also the reason why the allowed regions in the sk case are so large .
next let us discuss the case where `` the experimental data '' is the one obtained with @xmath55 .
the analysis is the same as the one in subsect.[hk - std ] , except that the `` the experimental data '' is produced assuming the presence of nsi , and here we perform only an energy spectrum analysis .
the results are shown in fig.[fig : fig4 ] , where the allowed regions at 2.5@xmath36cl ( @xmath56 for 2 degrees of freedom ) around the true points are depicted .
the straight lines @xmath57 in fig.[fig : fig4 ] stand for the approximate bound from the sk atmospheric neutrinos in fig.[fig : fig1 ] , and we have examined only the points below these straight lines . as seen from fig.[fig : fig4 ] , the errors in @xmath0 and @xmath29 are small for @xmath58 in the case of the normal hierarchy and for @xmath59 in the case of the inverted hierarchy .
the errors are large otherwise , and the reason that the errors are large is because a sensitivity is lost due to a destructive phenomenon between neutrinos and antineutrinos as was discussed in subsect.[hk - std ] .
cl around the point @xmath60 , where @xmath61 is assumed . ,
title="fig : " ] cl around the point @xmath60 , where @xmath61 is assumed .
, title="fig : " ]
in this talk we have discussed the constraint of the sk atmospheric neutrino data on the non - standard flavor - dependent interaction in neutrino propagation with the ansatz ( [ ansatz ] ) . from the sk atmospheric neutrino data for 4438 days , we have obtained the bound @xmath62 at 2.5@xmath36cl , while we have little constraint on @xmath0 . we have also discussed the sensitivity of the future hk atmospheric neutrino experiment to nsi by analyses with the energy rate and with the energy spectrum .
if nature is described by the standard oscillation scenario , then the hk atmospheric neutrino data will give us the bound @xmath63 at 2.5@xmath36cl from the energy rate analysis , and from the energy spectrum analysis it will restrict @xmath0 to @xmath51 and @xmath52 at 2.5@xmath36 ( 98.8% ) cl for the normal hierarchy and to @xmath53 and @xmath54 at 2.5@xmath36 ( 98.8% ) cl for the inverted hierarchy . on the other hand , if nature is described by nsi with the ansatz ( [ ansatz ] ) , then hk will measure the nsi parameters @xmath0 and @xmath29 relatively well for @xmath58 in the case of the normal hierarchy and for @xmath59 in the case of the inverted hierarchy .
it is important to use information on the energy spectrum to obtain strong constraint , because a sensitivity to nsi would be lost due to a destructive phenomena between the low and high energy events .
this research was partly supported by a grant - in - aid for scientific research of the ministry of education , science and culture , under grants no .
24540281 and no .
1 k. a. olive _ et al . _ [ particle data group collaboration ] , chin . phys .
c * 38 * ( 2014 ) 090001 .
k. abe _ et al . _
[ hyper - kamiokande working group collaboration ] , arxiv:1412.4673 [ physics.ins-det ] . c. adams _ et al . _
[ lbne collaboration ] , arxiv:1307.7335 [ hep - ex ] .
a. bandyopadhyay _ et al .
_ [ iss physics working group ] , rept . prog .
phys . * 72 * , 106201 ( 2009 ) [ arxiv:0710.4947 [ hep - ph ] ] . s. fukasawa and o. yasuda , adv .
high energy phys . * 2015 * , 820941 ( 2015 ) doi:10.1155/2015/820941 [ arxiv:1503.08056 [ hep - ph ] ] .
k. abe , t. abe , h. aihara , y. fukuda , y. hayato , k. huang , a. k. ichikawa and m. ikeda _ et al .
_ , arxiv:1109.3262 [ hep - ex ] . k. abe _ et al . _
[ super - kamiokande collaboration ] , arxiv:1410.2008 [ hep - ex ] . | the sensitivity of the atmospheric neutrino experiments to the non - standard flavor - dependent interaction in neutrino propagation is studied under the assumption that the only nonvanishing components of the non - standard matter effect are the electron and tau neutrino components @xmath0 , @xmath1 , @xmath2 and that the tau - tau component satisfies the constraint @xmath3 which is suggested from the high energy behavior for atmospheric neutrino data . |
in a 2009 paper @xcite , cameron introduced a new class of combinatorial designs , which simultaneously generalizes various well - known classes of designs , including @xmath0-designs , mutually orthogonal latin squares , orthogonal arrays and 1-factorizations of complete graphs .
further work on cameron s `` generalized @xmath0-designs '' has been done by soicher @xcite and others @xcite , while the earlier papers of martin @xcite and teirlinck @xcite discuss related objects . in a remark near the end of his paper
, cameron suggests that a similar definition can be made for generalizing covering designs .
the purpose of this paper is to pursue such a generalization , i.e. to define a broad class of combinatorial designs with a `` covering '' property which includes previously well - studied , and widely - applied , families of designs as special cases .
analogous to @xmath0-designs and orthogonal arrays are covering designs and covering arrays , respectively ; the designs we define in this paper form a simultaneous generalization of both of these `` covering '' objects .
the key difference when studying covering problems rather than `` ordinary '' designs is that the question is typically not whether the designs exist ( this is usually trivial to answer ) , but obtaining bounds on the minimum size , and constructing optimal ( or near - optimal ) designs .
background material on most classes of designs can be found in the _ handbook of combinatorial designs _
@xcite . however , we give some relevant definitions here .
[ defn : covdes ] let @xmath1 be positive integers with @xmath2 .
@xmath3-covering design _ is a family @xmath4 of @xmath5-subsets ( called _ blocks _ ) of a @xmath6-set @xmath7 , where any @xmath0-subset of @xmath7 is contained in at least @xmath8 members of @xmath4 .
the notation `` @xmath0-@xmath9 covering design '' is also used in the literature ( in @xcite , for example ) . in the case where each @xmath0-subset occurs _ exactly _ @xmath8 times , we have a _
@xmath0-@xmath9 design_. we remark that the definition allows for blocks to be repeated .
a survey of results on covering designs can be found in mills and mullin @xcite . usually , we are only concerned with the case @xmath10 , and omit the subscript @xmath8 .
the following is an example of an @xmath11-covering design , where @xmath12 : @xmath13 by hand , it is easy to verify that each pair chosen from @xmath7 is contained in at least one of the 5-sets given . unlike @xmath0-designs ,
it is clear that if @xmath2 , then a @xmath14-covering design always exists : we simply take all the @xmath5-subsets of @xmath7 and discard any that are unnecessary .
of course , this is not an efficient approach .
we would like our designs to have the smallest number of blocks possible : this number is called the _ covering number _ and is denoted @xmath15 .
if we have parameters @xmath16 for which no covering design can exist , then we say @xmath17 ; clearly , this can only happen if @xmath18 or @xmath19 .
there are many results on finding or bounding covering numbers , which can be found in @xcite .
the most general bound is known as the _ schnheim bound _
@xcite , and is given below .
[ thm : schonheim ] where @xmath2 , we have @xmath20 however , many more specific results are known , and exact values of @xmath15 have been determined in many cases .
( often , these are the result of sophisticated computer searches : see @xcite for an online database of the best - known covering designs for small values of @xmath6 , @xmath5 and @xmath0 . ) in particular , in the case where @xmath21 and the ratio @xmath22 , the exact values of @xmath15 , and constructions of covering designs of those sizes , are all known . for @xmath23 ,
the constructions are due to various authors , and can be found in a paper by mills @xcite . for @xmath24 ,
the constructions are much newer , and are due to greig , li and van rees @xcite .
( see ( * ? ? ?
* theorem vi.11.31 ) , for a summary . )
an important asymptotic result is due to rdl @xcite , who in 1985 proved a conjecture of erds and hanani @xcite which asserts that for fixed values of @xmath5 and @xmath0 , @xmath25 an alternative proof of rdl s theorem was subsequently obtained by spencer @xcite , while constructions for covering designs which ( asymptotically ) meet this bound were obtained by gordon _
et al . _
@xcite .
covering designs , or objects obtained from them , have been used in various applications , many of which are related to computing or communications .
these include quorum systems in distributed databases @xcite , threshold schemes in cryptography @xcite , and decoding algorithms for error - correcting codes @xcite .
[ defn : covarray ] let @xmath26 be positive integers . a _ covering array _
@xmath27 is an @xmath28 array with entries from an alphabet of size @xmath29 , with the property that in every set of @xmath0 columns , each @xmath0-tuple of symbols from the alphabet occurs in at least @xmath8 rows .
note that such an array where every @xmath0-tuple occurs in _ exactly _
@xmath8 rows is known as an _ orthogonal array _ ( see @xcite ) . as with covering designs
, we usually only treat the case where @xmath10 and omit the subscript @xmath8 .
the parameter @xmath0 is called the _ strength _ of the covering array .
usually we fix the parameters @xmath5 , @xmath29 and @xmath0 and want to find the smallest @xmath30 such that there exists a @xmath31 : this value of @xmath30 is called the _ covering array number _ , and is denoted by @xmath32 .
the following is an example of a @xmath33 where the alphabet is @xmath34 : @xmath35 in each pair of columns , each of the @xmath36 possible combinations @xmath37 appears at least once .
note the slight abuse of notation : in the case of covering designs , each row of the table is a set , whereas in the case of covering arrays , each row of the table is a @xmath5-tuple .
this distinction will be important later in the paper .
a more general object is the _ mixed covering array _
, as defined by moura _
et al . _
this is a covering array where each column has its own alphabet , and these may have different sizes .
if the alphabet sizes are @xmath38 , then such an object is denoted by @xmath39 , while the least possible size is denoted by @xmath40 .
covering arrays , and mixed covering arrays , have a number of applications , most notably in software testing , but also in other areas such as computational biology .
the survey by colbourn @xcite describes many of these applications .
throughout the remainder of this paper , we will be using @xmath41-tuples , or vectors , of both integers and finite sets . we begin with explaining our terminology and notation for this .
suppose @xmath42 and @xmath43 are @xmath41-tuples of integers .
we write @xmath44 to mean that @xmath45 for all @xmath46 .
we use the notation @xmath47 to denote the vector with @xmath41 entries , all equal to @xmath48 .
there are various operations we can perform on vectors of integers , as well as the usual operations of addition , subtraction and scalar multiplication .
first , given a vector @xmath49 and subset @xmath50 of the indices @xmath51 , the _ restriction _ of @xmath49 to @xmath50 ( denoted @xmath52 ) is the vector whose entries are those from @xmath49 taken from positions labelled by @xmath50 .
for example , if @xmath53 and @xmath54 , then @xmath55 .
let @xmath56 and @xmath57 be an @xmath41-tuple and an @xmath58-tuple respectively .
define the _ concatenation _ of @xmath59 and @xmath60 to be the @xmath61-tuple @xmath62 the _ sum _ of a vector is simply the sum of its entries , while the _ weight _ of a vector is the number of non - zero entries .
( note that if a vector s entries are either 0 or 1 , its sum and weight are equal . ) now suppose @xmath63 and @xmath64 are @xmath41-tuples of sets .
we write @xmath65 to mean that @xmath66 for all @xmath46 , and say _
@xmath67 is contained in @xmath68_. the operations of restriction and concatenation are defined for vectors of sets in the same way they were for vectors of integers . for any set @xmath7 , we use the notation @xmath69 to denote the set of all @xmath5-subsets of @xmath7 .
( thus if @xmath7 is finite and has size @xmath58 , then the size of @xmath69 is @xmath70 . )
if we have an @xmath41-tuple of sets @xmath71 and an @xmath41-tuple of integers @xmath72 , define @xmath73 so a member of @xmath74 consists of an @xmath41-tuple of finite sets , of sizes @xmath75 .
other pieces of notation will be defined as and when required .
suppose @xmath1 are integers where @xmath76 and @xmath77 .
let @xmath78 be an @xmath41-tuple of positive integers with sum @xmath6 , and let @xmath72 be an @xmath41-tuple of positive integers with sum @xmath5 , and where @xmath79 .
now let @xmath80 be an @xmath41-tuple of pairwise disjoint sets , where @xmath81 .
let @xmath82 be an @xmath41-tuple of _ non - negative _ integers .
we say @xmath83 is _
@xmath84-admissible _ if @xmath85 and @xmath86 . in a similar vein ,
if @xmath87 is an @xmath41-tuple of disjoint sets , we say that @xmath88 is _
@xmath89-admissible _ if each @xmath90 is a @xmath91-subset of @xmath92 , where @xmath93 is @xmath84-admissible .
( note that since @xmath91 is allowed to be zero , the corresponding set @xmath90 is allowed to be empty . )
[ defn : gc ] suppose @xmath94 are as above . then a _ generalized covering design _ ,
@xmath95 , is a family @xmath96 of elements of @xmath74 , called _ blocks _ , with the property that every @xmath87 which is @xmath89-admissible is contained in at least @xmath8 blocks in @xmath96 .
we call @xmath97 the _ point set _ of the generalized covering design ; one can think of @xmath98 as being a partition of the point set @xmath7 .
however , by an abuse of notation , we usually label the elements of each @xmath92 as @xmath99 .
we remark that our definition of a generalized covering design is identical to cameron s definition of a generalized @xmath0-design , except his definition requires `` exactly @xmath8 '' , where ours requires `` at least @xmath8 '' . as with covering arrays
, we call the parameter @xmath0 the _ strength _ of the design . now , cameron s generalized @xmath0-designs are a common generalization of @xmath0-designs and orthogonal arrays , whereas our designs are a common generalization of covering designs and covering arrays , as we show below .
[ prop : covdes ] suppose @xmath100 and @xmath101 .
then a @xmath95 is a @xmath3-covering design . [
prop : covarr ] suppose @xmath102 and @xmath103 .
then a @xmath95 ( with @xmath30 blocks ) is equivalent to a covering array @xmath104 . since @xmath105
, we can suppose @xmath106 are disjoint copies of an @xmath29-set @xmath107 , which we regard as our alphabet . also , since @xmath108 , each block consists of exactly one element from each copy of @xmath107 , so we can regard this as a @xmath5-tuple of elements of @xmath107 .
put these @xmath5-tuples as the rows of an array .
now , each @xmath84-admissible @xmath5-tuple @xmath83 will be a 0/1 vector of weight @xmath0 , so represents a @xmath0-subset of columns of the array .
thus , for a given @xmath83 , the corresponding @xmath89-admissible @xmath5-tuples @xmath109 will represent all possible combinations of entries in those @xmath0 columns .
consequently , each such combination must appear at least once , and so the array is a @xmath110 .
carrying out the reverse of the process described above shows that a covering array @xmath110 gives a @xmath95 ( with @xmath30 blocks ) , as required .
of course , there are many examples which are neither covering designs nor covering arrays .
the following is a very basic example of such an object .
[ example : gcdnew ] let @xmath111 , @xmath112 and @xmath21 .
then the following is a @xmath113 : @xmath114 the possible admissible vectors @xmath83 are @xmath115 , @xmath116 , @xmath117 and @xmath118 . for @xmath119 , we are required to cover all possible pairs from @xmath120 in the first column . for @xmath121 ,
each symbol from @xmath120 must appear in the first part of a block with each possible symbol from @xmath122 in the second part . the case @xmath123 works similarly .
for @xmath124 each of the ordered pairs @xmath125 , @xmath126 , @xmath127 , @xmath128 must occur in the final two parts of some block .
it is a straightforward exercise to verify that all the possibilities are covered .
as is common in the study of covering problems , from now on we will consider only the case @xmath10 , and drop the subscript @xmath8 from our notation .
also we note that , by the same argument as proposition [ prop : covarr ] , a mixed covering array @xmath39 is equivalent to a @xmath95 where @xmath129 ( and with @xmath30 blocks ) . as was the case with ordinary covering designs ,
it is trivial to show that generalized covering designs @xmath130 always exist , provided that @xmath131 , simply by taking the collection of all possible blocks and discarding any that are unnecessary .
so the interesting questions , as with all covering problems , are to find bounds on the minimal size of a @xmath130 , and find constructions which meet ( or come close to ) these bounds .
borrowing notation from the study of ( ordinary ) covering designs , we have the following . the _ covering number _
@xmath132 denotes the smallest possible size of a generalized covering design @xmath130 . as with ordinary covering designs
, we say @xmath133 if no design exists , or if there are no possible @xmath89-admissible vectors of sets .
we conclude this section by giving some basic results about generalized covering designs , and the corresponding covering numbers .
our first result can be used to obtain a recursive bound on covering numbers .
[ prop : reducet ] for @xmath134 , any @xmath135 is a @xmath136 , and so @xmath137 let @xmath88 be a @xmath89-admissible vector .
then for any @xmath135 there is a block in the design that contains @xmath88 .
clearly , any @xmath41-tuple @xmath138 with @xmath139 is also contained in this block . since any @xmath140-admissible vector of sets is contained in some @xmath89-admissible vector , the result follows .
our next result shows that the case where @xmath141 is very easy . in this case , all we require is enough blocks so that each symbol from each part appears at least once in the corresponding part of a block .
[ prop : tis1 ] where @xmath142 and @xmath72 , we have @xmath143
a _ clique covering _ of a graph @xmath144 is a family of complete subgraphs @xmath145 of @xmath144 , called _ cliques _ , with the property that every edge of @xmath144 appears in at least one of @xmath145 .
we remark that the general clique covering problem allows cliques of different sizes .
if we require that all the cliques are _ @xmath5-cliques _ ( i.e. they all have exactly @xmath5 vertices ) , then we have a _ @xmath5-uniform _ clique covering . as with covering designs ,
the question of the existence of a clique covering of a graph is trivial : the set of all edges forms a 2-uniform clique covering . on the other hand , for @xmath146
, @xmath5-uniform clique coverings do not exist in general , and their existence depends on the structure of the graph .
ideally , we would like a clique covering of @xmath144 to contain the smallest number of cliques possible ; this number is called the _ clique covering number _ of @xmath144 , denoted @xmath147 .
the first result on clique coverings is attributed to hall @xcite , who showed that if a graph has @xmath58 vertices , then @xmath148 .
this was later extended by erds _
et al . _
@xcite who showed the same bound holds when every edge of the graph appears in _ exactly _ one clique of the covering .
other results on clique coverings can be found in a survey paper by monson _
et al . _
@xcite .
many techniques and concepts used to construct clique coverings and generate bounds on the clique covering number can be extended to generalized covering designs of strength @xmath149 and bounds on @xmath150 .
these ideas will be pursued in this section of the paper .
for example , one such concept is that of _ equivalent vertices _ in a graph explored by gyrfs @xcite , which we apply to generalized covering designs in section @xmath151 . in the case where @xmath21 , both covering designs and covering arrays are equivalent to clique coverings of particular graphs , as we will now explain . [ prop : covdesclique ] a @xmath152-covering design is a covering of the complete graph @xmath153 by @xmath5-cliques @xmath145 , such that every edge of @xmath153 appears in at least one of @xmath145 .
for instance , a steiner triple system on @xmath6 points , which is an example of a @xmath154-covering design , is a partition of @xmath153 into 3-cliques ( so each edge appears exactly once ) . in order to explain how covering arrays give clique coverings , we need the following definition . [ defn : multipartite ]
suppose @xmath155 is a sequence of positive integers , and let @xmath156 be disjoint sets of sizes @xmath157 respectively . then the _ complete multipartite graph
_ @xmath158 has vertex set @xmath159 , and two vertices @xmath160 , @xmath161 are adjacent if and only if @xmath162 . alternatively , the complete multipartite graph @xmath158 is the complement of the disjoint union of complete graphs with vertex sets @xmath156 . in particular : if @xmath163 , we have a empty graph ; if @xmath164 we have a complete bipartite graph ; and if @xmath165 ( with @xmath41 entries ) we have the complete graph @xmath166 . [
prop : caclique ] a covering array @xmath167 is equivalent to a covering of the complete multipartite graph @xmath168 by @xmath5-cliques @xmath169 , where @xmath170 has @xmath5 entries , all equal to @xmath29 .
each row of a @xmath167 defines a clique , where the symbol in position @xmath171 specifies the vertex chosen from the @xmath172 part of the graph @xmath168 . in any pair of columns of the array
, every pair of symbols must be used at least once ; consequently , every edge of @xmath168 appears in at least one clique . on the other hand , given a clique covering of this type , we can construct a covering array by reversing this process . for further details , see danziger
_ et al . _
@xcite , maltais @xcite , or ronneseth and colbourn @xcite .
we remark that mixed covering arrays ( with @xmath21 and @xmath10 ) can also be interpreted as clique coverings of complete multipartite graphs in exactly the same way , but where the parts may have different sizes . in order to describe generalized covering designs @xmath113 in terms of clique coverings
, we need to define a suitable graph . to do this , we need the following graph - theoretical idea . [
defn : join ] let @xmath173 and @xmath174 be graphs with @xmath175
. then the _ join _ of @xmath176 and @xmath177 , denoted @xmath178 , is the graph with vertex set @xmath179 , and whose edge set is @xmath180 .
for example , the join of two complete graphs is also complete , and the join of two empty graphs is a complete bipartite graph .
we note that this can be extended to a join of any number of graphs , and that this operation is associative .
now suppose that @xmath142 and @xmath72 are vectors of positive integers with @xmath181 .
let @xmath182 where @xmath183 represents the complement of @xmath184 ( that is , the empty graph ) . form the graph @xmath185 consisting of the join of the graphs @xmath186 such that @xmath187 has vertex set @xmath188 , where @xmath81 and each @xmath92 is the set of vertices of the corresponding @xmath186 .
there are two important special cases .
first , if all @xmath189 , then @xmath187 is precisely the complete multipartite graph @xmath168 .
second , if all @xmath190 , then @xmath187 is isomorphic to the complete graph @xmath153 ( where @xmath191 ) , but where the vertex set has a specific partition into parts of sizes @xmath192 .
[ thm : gcgraph2 ] let @xmath187 be the graph described above .
then a generalized covering design @xmath113 is equivalent to an edge covering of @xmath187 using a collection of cliques of size @xmath5 , with the property that for each clique in the covering , @xmath193 vertices of the clique come from the set @xmath92 ( for each @xmath171 ) .
suppose @xmath194 is a @xmath113 .
now , from each block in @xmath194 , we can easily construct a clique in @xmath187 .
think of the vertex set of @xmath187 as being the point set of @xmath194 , namely @xmath195 .
a block of @xmath194 contains @xmath193 points from each part @xmath92 , and the subgraph of @xmath187 induced by the corresponding vertices is necessarily a clique .
now , the admissible vectors @xmath83 have two possible forms : ( i ) a vector with two entries 1 and the rest 0 ; and ( ii ) a vector with a single entry of 2 , and all other entries 0 . type ( i )
vectors ensure that this collection of cliques covers all edges between parts .
type ( ii ) vectors are only possible with the single 2 in position @xmath171 with @xmath190 ; these vectors ensure that these cliques cover all edges within each part with @xmath190 .
consequently , these cliques cover all the edges of @xmath187 , and so we have a clique covering .
a similar argument works in the reverse direction : given a clique covering of @xmath187 with the specified form , each clique gives us a block , and the way @xmath187 was constructed ensures these blocks form a @xmath113 .
[ example : cliquegc ] recall example [ example : gcdnew ] , where we saw a @xmath113 with @xmath111 and @xmath112 . by theorem [ thm : gcgraph2 ] , this can be viewed as a clique covering of the appropriate graph @xmath187 using copies of @xmath196 , as shown in figure [ figure : cliquegc ] .
we can use this interpretation to obtain various bounds on covering numbers @xmath197 , as well as to obtain methods of constructing generalized covering designs @xmath113 . in some cases
, we can use these results to obtain designs which are optimal .
the simplest bound on @xmath197 can be obtained by counting edges . when covering a graph @xmath144 ( with @xmath198 edges ) by a collection of subgraphs isomorphic to @xmath199 , each of which has @xmath200 edges , then clearly the minimum number of subgraphs required
is @xmath201 .
this observation then provides the following bound : @xmath202 using the interpretation in theorem [ thm : gcgraph2 ] , the number of edges in @xmath187 is @xmath203 where the summation is over the indices @xmath171 such that @xmath189 , and we use the convention that @xmath204 .
however , an alternative way to write @xmath205 is @xmath206 as the first summation counts the edges in each @xmath186 and the second summation counts the edges joining each @xmath186 to each @xmath207 ( for @xmath162 ) .
now , since the number of edges in a clique of size @xmath5 is @xmath208 , we have shown the following .
[ prop : naive ] using the notation above , @xmath209 recall the notion of _ restriction _ which we introduced earlier . in this subsection
, we will show how this operation can be used to construct generalized covering designs , and to obtain bounds on @xmath197 .
[ prop : restriction ] let @xmath194 be a generalized covering design @xmath113 with @xmath30 blocks , and suppose @xmath210 is non - empty .
then , provided @xmath211 , there exists a @xmath212 with @xmath30 blocks , and in particular @xmath213 .
we exclude the case @xmath214 as no @xmath215-covering design can exist ; however , since we adopted the convention that @xmath216 , the bound still holds , albeit for trivial reasons .
so we suppose @xmath217 and construct the graph @xmath187 as in theorem [ thm : gcgraph2 ] above . clearly , the graph @xmath218 is an induced subgraph of @xmath187 , obtained simply by removing the parts not indexed by @xmath50 .
now , each block of @xmath194 is a clique in @xmath187 with @xmath193 vertices in part @xmath92 , and if we restrict these cliques to those parts in @xmath218 , every edge of @xmath218 must be covered by at least one clique . consequently , these @xmath30 cliques form a clique covering of @xmath218 with the required property , and thus correspond to a @xmath212 with @xmath30 blocks .
the bound @xmath213 follows immediately .
we denote the @xmath212 obtained from @xmath194 in the proof of proposition [ prop : restriction ] by @xmath219 , and call this the _ restriction of @xmath194 to @xmath50_. now , by restricting in all possible ways , we obtain the following lower bound . [ cor : restriction ] let @xmath220 denote the collection of all non - empty subsets of @xmath221 .
then we have @xmath222 of course , the nave edge - counting bound in proposition [ prop : naive ] can then be applied to the right - hand side to obtain a lower bound in terms of the entries of @xmath170 and @xmath223 .
unfortunately , the bound in corollary [ cor : restriction ] involves checking an exponential number of cases , so a more practical bound can be obtained just by looking at subsets of size @xmath224 . in this situation
, the restricted design has only one part , and is thus an ordinary @xmath225-covering design .
consequently , we have the following corollary .
[ cor : restriction2 ] for vectors @xmath226 and @xmath227 , we have @xmath228 we were able to exclude indices where @xmath189 from the above , as @xmath216 .
of course , if @xmath229 for some @xmath171 ( i.e. our design is not a ( mixed ) covering array ) , we can then use the schnheim bound ( theorem [ thm : schonheim ] ) to obtain @xmath230 we also have another straightforward application of restriction , which allows us to safely ignore parts where @xmath231 .
[ prop : allvertexparts ] suppose @xmath226 and @xmath227 are vectors of positive integers with @xmath131 .
let @xmath50 be the set of indices where @xmath232 .
then , provided that @xmath233 , we have @xmath234 .
we require that @xmath233 in order to ensure that the graph @xmath218 is not an empty graph ( in which case , we would have @xmath235 and the result would not hold ) . from proposition [ prop : restriction ] , we know that @xmath213 ; to show that equality holds , we will show that given a @xmath212 , we can obtain a @xmath113 with the same number of blocks .
suppose without loss of generality that @xmath236 , and let @xmath237 be a @xmath212 , which we treat as a clique covering of @xmath218 .
now , to each clique in @xmath237 , we add @xmath238 new parts of sizes @xmath239 . in doing
so we obtain a @xmath113 : ( i ) every edge we need to cover in parts @xmath240 is already covered in @xmath237 ; ( ii ) any edge between parts @xmath241 appears in every block , as does any edge within a part @xmath242 ( where @xmath243 ) ; ( iii ) every edge joining a part @xmath242 to a vertex @xmath244 is covered , as the vertex @xmath6 must appear in some block of @xmath237 .
we remark that proposition [ prop : allvertexparts ] works for arbitrary values of @xmath134 , although the argument using clique coverings ca nt be applied .
the notion of _ equivalence _ is a very useful one , when it comes to both bounds and constructions .
it is inspired by the notion of _ equivalent vertices _ for clique coverings , as studied by gyrfs @xcite . as usual , we have vectors @xmath226 and @xmath227 with @xmath131 .
we define an equivalence relation @xmath245 on the set of indices @xmath221 by @xmath246 if and only if @xmath247 and @xmath248 .
( sometimes , it is convenient to talk about @xmath249 being equivalent to @xmath250 . ) in order to demonstrate the usefulness of this idea , we have the following lemma .
[ lemma : equiv ] where @xmath170 and @xmath223 are as above , let @xmath194 be a @xmath113 of size @xmath30 .
define vectors @xmath251 and @xmath252 where @xmath253 is equivalent to some fixed @xmath249 where @xmath254 . then there exists a @xmath255 of size @xmath30 .
we think of @xmath194 in terms of a clique covering of @xmath187 ( cf .
theorem [ thm : gcgraph2 ] ) with vertex set @xmath256 . since @xmath254 , we observe that @xmath257 is precisely the graph obtained from @xmath187 by joining a complete graph @xmath258 , i.e. @xmath259 , and where @xmath257 has vertex set @xmath260 ( where @xmath261 ) . since @xmath262 , we have a one - to - one correspondence between the vertices in @xmath92 and in @xmath263 ; label these as @xmath264 and @xmath265 ( of course , @xmath266 ) .
choose some @xmath267 ( i.e. a clique in @xmath187 ) : without loss of generality , assume this clique contains vertices @xmath268 from @xmath92 .
now form a new clique @xmath269 by taking all vertices of @xmath270 , together with @xmath271 .
we claim that the set @xmath272 is a clique covering of @xmath257 corresponding to a @xmath255 .
to show this , we need to consider any pair of adjacent vertices of @xmath257 , and show that the edge joining them is covered by at least one member of @xmath273 .
now , any edge which lies in the induced subgraph @xmath187 is automatically covered ( because we had a clique covering of @xmath187 to begin with ) .
also , any edge @xmath274 ( where @xmath275 is a vertex of @xmath187 not in @xmath92 ) is covered , as the corresponding edge @xmath276 is in some clique in @xmath194 . finally , since @xmath190 , we know that each edge @xmath277 is covered by some clique @xmath267
. thus both the vertices @xmath278 and @xmath279 are in @xmath280 , and so it follows that @xmath269 covers the edges @xmath281 , @xmath282 and @xmath283 .
thus all types of edge in @xmath257 are covered by at least one clique in @xmath273 , and we are done .
the proof of lemma [ lemma : equiv ] demonstrates the reason for the requirement that @xmath190 in the statement of the lemma : if @xmath284 , we would not be able to cover edges of the form @xmath282 using this construction . with that in mind , without loss of generality
we assume that @xmath285 , where all entries of @xmath286 are at least 2 .
now write @xmath287 , where @xmath288 for @xmath289 , and where @xmath290 for @xmath291
. then we have the following theorem .
[ thm : equiv ] suppose we have vectors of positive integers @xmath292 as above , where @xmath293 for all @xmath171 .
let @xmath294 be a set of equivalence class representatives for @xmath295 and @xmath286 under the relation @xmath245 .
then we have @xmath296 the inequality @xmath297 follows from proposition [ prop : restriction ] , by restricting to @xmath294 .
the reverse inequality follows by repeatedly applying lemma [ lemma : equiv ] to a minimal @xmath298 , to obtain a @xmath299 of size @xmath300 . in the special case where all entries of @xmath223 are at least 2
, we have the following straightforward corollary .
[ cor : equiv ] suppose @xmath190 for all @xmath171 , and that @xmath294 is a set of equivalence class representatives for @xmath170 and @xmath223 under the relation @xmath245 .
then we have @xmath301 the operations of point deletion and block expansion give us another way to construct new generalized covering designs from existing ones .
suppose that @xmath170 and @xmath223 are our usual vectors of integers , where @xmath226 lists the sizes of the sets @xmath302 respectively .
now , by _ point deletion _ we mean the operation of removing some points from each part @xmath92 , to obtain @xmath303 , with sizes @xmath304 respectively ( so that @xmath305 ) .
a similar operation is _ block expansion _
, where we keep @xmath170 fixed but increase the sizes of our blocks from @xmath223 to @xmath306 , where @xmath307 . in the following discussion , we once again think of our @xmath113 in terms of a clique covering of @xmath187 ( cf .
theorem [ thm : gcgraph2 ] ) with vertex set @xmath195 .
our first lemma concerns point deletion .
[ lemma : deletion ] let @xmath194 be a @xmath113 of size @xmath30 , and suppose @xmath308 satisfies @xmath309 .
then there exists a @xmath310 of size at most @xmath30 .
suppose we have deleted the vertex @xmath311 .
for each clique in @xmath194 which includes the vertex @xmath48 , we remove @xmath48 and replace it with another vertex @xmath312 , which is not already in that clique ; since @xmath313 , we know that such a vertex must exist .
we repeat this procedure until we are left only with the vertices in @xmath314 .
it is straightforward to see that the cliques obtained will cover all edges in @xmath315 .
note that it is possible that this procedure may introduce some repeated blocks , so we remove any duplicates .
thus we can only be sure our @xmath310 has an upper bound of at most @xmath30 blocks , rather than exactly @xmath30 .
a counterpart to lemma [ lemma : deletion ] is the following , which is concerned with block expansion .
[ lemma : expansion ] let @xmath194 be a @xmath113 of size @xmath30 , where @xmath190 for all @xmath171 , and suppose @xmath306 satisfies @xmath316 .
then there exists a @xmath317 of size at most @xmath30 . since @xmath190 for all @xmath171
, it follows that @xmath318 .
thus every edge of @xmath319 is already covered by the cliques in @xmath194 , which each have @xmath193 vertices chosen from @xmath92 ( for every @xmath171 ) .
adding extra vertices to each clique , in parts where @xmath320 ( so that there are now @xmath321 vertices chosen from each @xmath92 ) , does not affect this .
again , we note that it is possible that this procedure may introduce some repeated blocks , so we remove any duplicates and thus only have an upper bound of @xmath30 rather than equality . of course , in both lemma [ lemma : deletion ] and lemma [ lemma : expansion ] ,
if the generalized design @xmath194 that we begin with is optimal , there is no guarantee that the resulting design will be optimal .
however , the two constructions do yield the following important bound .
[ thm : delexp ] suppose we have vectors of positive integers @xmath322 , each of length @xmath41 , where @xmath190 for all @xmath171 , and which satisfy @xmath323 and @xmath324
. then we have @xmath325 and @xmath326 we can combine the operations of point deletion and block expansion , and the notion of equivalence , to obtain our next bound , which is perhaps the most useful so far .
suppose we have our usual vectors @xmath226 and @xmath227 .
define @xmath327 , and let @xmath328 denote the vector of @xmath41 entries , all equal to @xmath329 ; in a similar fashion , define @xmath330 to be the vector with @xmath41 entries , all equal to @xmath331 . clearly , @xmath332 and @xmath333 .
[ thm : minimax ] suppose we have vectors @xmath170 and @xmath223 , where each @xmath190 . then @xmath334 using our results above ,
we obtain : @xmath335 as required .
[ cor : minimaxpair ] suppose we have vectors @xmath170 and @xmath223 , with an index @xmath171 where @xmath336 and @xmath337
. then @xmath338 the inequality @xmath339 is given by theorem [ thm : minimax ] .
we obtain reverse inequality by restricting to the part @xmath171 where @xmath336 and @xmath340 , and applying proposition [ prop : restriction ] .
this last corollary is especially useful , as many of the covering numbers @xmath341 are known exactly , and thus we are able to obtain the sizes of optimal generalized covering designs in many instances .
all of our proofs in this section have been constructive . as a direct consequence
, we have an algorithm for actually constructing generalized covering designs from a single covering design . furthermore , if there is a part with @xmath342 and @xmath4 is optimal , then ( provided that @xmath343 ) corollary [ cor : minimaxpair ] ensures that designs obtained using this construction are optimal .
[ constr : construction ] suppose we are given vectors @xmath170 and @xmath223 , where @xmath190 for all @xmath171 , and we have an optimal @xmath344-covering design @xmath4 .
then we construct a @xmath113 as follows : * put a copy of @xmath4 on each part ; * in any part where @xmath345 , delete the extra points , replacing them with a `` placeholder '' symbol @xmath346 ; * in any part where @xmath347 , add @xmath348 placeholders to each block ; * in each block , replace the placeholders greedily , ensuring that no symbol is repeated in a block ; * remove any repeated blocks .
we illustrate this construction with an example .
[ example : fano ] suppose @xmath349 and @xmath350 .
so we have @xmath351 and @xmath352 : an optimal @xmath353-covering design with 7 blocks is , of course , the fano plane , whose blocks are @xmath354 so our `` pre - design '' looks like : @xmath355 having filled the placeholder positions , and using the least point available at each stage , we then obtain : @xmath356 as we have @xmath357 occurring in a part , we are guaranteed that this @xmath113 is optimal . in situations where there is no pair @xmath342 ,
then it is possible to construct pathological examples which are far from optimal , even if beginning with an optimal covering design .
for instance , if @xmath358 and @xmath359 , then this construction requires a @xmath360-covering design , which has size at least @xmath361 .
we will see later that 7 blocks can be used !
we conclude this subsection with a remark concerning the placeholder symbol @xmath346 . in many constructions of covering arrays ( see @xcite , for instance ) , a `` do nt care '' symbol is often appended to the alphabet .
this can then be replaced arbitrarily without affecting the requirement that all @xmath0-tuples of symbols be covered in every @xmath0-subset of columns , which can be particularly useful in applications .
so , in our construction , it may actually be beneficial to leave the placeholder symbols _ in situ _
, rather than filling those positions greedily .
this is yet another operation to obtain a new generalized covering design from an existing one , this time by combining two parts into one . as usual , we let @xmath226 and @xmath227 , and suppose @xmath194 is a @xmath362 .
suppose further that we have two indices where the entries of @xmath223 are at least 2 , which without loss of generality we assume are @xmath363 and @xmath364 .
now let @xmath365 and @xmath366 .
the operation of _ amalgamation _ allows us to construct a @xmath367 from @xmath194 . for each block @xmath368 , let @xmath369 .
finally , let @xmath370 .
[ prop : amalg ] the design @xmath371 defined above is a @xmath367 , and so @xmath372 .
once more , we think of the generalized covering design @xmath194 as a clique covering of the graph @xmath187 .
now , observe that the graph @xmath373 must be isomorphic to @xmath187 , except that the parts @xmath374 and @xmath375 of @xmath376 , which have sizes @xmath377 , @xmath378 respectively , can be thought of as one larger part of size @xmath379 .
each block @xmath380 corresponds to a clique in @xmath187 with @xmath363 vertices in @xmath374 and @xmath364 vertices in @xmath375 , while the corresponding block @xmath381 gives exactly the same clique in @xmath373 ( with @xmath382 vertices in @xmath383 ) .
since the two graphs are isomorphic , and @xmath194 and @xmath371 contain the same cliques , it follows that @xmath371 covers all the edges in @xmath373 .
the above result can be used to prove the following .
[ prop : amalg2 ] suppose we have vectors of positive integers @xmath226 and @xmath227 with @xmath131 , and let @xmath384 be a vector of integers satisfying @xmath385 for all @xmath171 .
then @xmath386 by repeatedly applying proposition [ prop : amalg ] above , we have @xmath387 @xmath388 then we apply proposition [ prop : allvertexparts ] to show that this is equal to @xmath389 as required .
we remark that amalgamation can sometimes give much better results than the algorithm described in subsection [ subsect : construction ] .
starting from the design in example [ example : fano ] , we can use proposition [ prop : amalg ] to obtain a @xmath390 with 7 blocks , by amalgamating the first and second parts .
this must be optimal , as it meets the bound in corollary [ cor : restriction ] .
however , if we applied construction [ constr : construction ] to obtain a @xmath390 from scratch , we would need to begin with an @xmath391-covering design .
the smallest such covering design has 19 blocks ( see gordon _
et al . _
@xcite ) , meeting the schnheim bound . taking that design , applying construction [ constr : construction ] to it , and filling in the placeholders lexicographically , gives a @xmath390 with 19 blocks , which is considerably larger that that obtained using amalgamation .
proposition [ prop : amalg2 ] can be combined with the argument of theorem [ thm : minimax ] to obtain the bound @xmath392 ( provided that @xmath343),which is often a considerable improvement .
for example , @xmath393 which compares with the bound of @xmath394 from theorem [ thm : minimax ] .
( note that @xmath395 , while @xmath361 , so this is definitely an improvement ! )
there is an another interpretation of strength-2 generalized covering designs in terms of graphs , which while similar to that developed in section 3 , gives improved bounds on @xmath197 for certain parameter sets .
suppose we have vectors @xmath226 and @xmath227 , with @xmath131 , as usual . as well as the interpretation in terms of clique coverings , we can also regard a @xmath113 as an edge - covering of a complete multipartite graph , as explained below .
[ thm : gcgraph ] suppose we have @xmath142 and @xmath72 , where @xmath131 . then a generalized covering design @xmath113 is equivalent to an edge - covering of the complete multipartite graph @xmath168 by complete multipartite graphs @xmath396 , where : * for each copy of @xmath396 , the vertices contained in the part corresponding to @xmath193 are chosen from the vertices of @xmath168 corresponding to @xmath397 ; * the set of the complements of each copy of @xmath396 covers all edges of the complement of @xmath168 in parts where @xmath190 .
each block of a @xmath113 corresponds to a copy of the complete multipartite graph @xmath396 satisfying condition ( i ) .
now , we have two types of admissible vector @xmath83 .
first , we have vectors @xmath83 consisting of two @xmath224s and the rest @xmath398s correspond to pairs of vertices in distinct parts : the fact that we have an edge - covering ensures that these pairs are contained within some block .
second , we have vectors @xmath83 with exactly one @xmath149 in a position @xmath171 with @xmath190 , and @xmath398s elsewhere .
these correspond to pairs of vertices within a part , and condition ( ii ) ensures that these pairs are contained within some block .
we remark that in the case of covering arrays ( i.e. where @xmath129 ) , the two interpretations are the same : in this case , the complete multipartite graph @xmath396 is actually a complete graph , so covering with copies of @xmath396 actually gives a clique covering . using this second interpretation
, we can obtain various bounds on @xmath197 , which do not always correspond to those obtained in section 3 .
the simplest of these is the nave bound for covering a graph with isomorphic subgraphs , analogous to that in proposition [ prop : naive ] . simply by counting the number of edges in a complete multipartite graph ,
we arrive at the following result .
[ prop : naivemulti ] where @xmath142 and @xmath72 and @xmath131 , we have @xmath399 depending on the precise nature of the entries of @xmath170 and @xmath223 , it is quite possible that either of propositions [ prop : naive ] or [ prop : naivemulti ] will give a better bound . the notion of restriction , introduced in section 3.2 , can also be applied to this interpretation .
suppose @xmath50 is a subset of the index set @xmath221 .
now , we can restrict @xmath170 and @xmath223 to @xmath50 , and will still have a covering as described in theorem [ thm : gcgraph ] , provided that @xmath400 .
( if @xmath401 , then both @xmath402 and @xmath403 have no edges , so the construction is vacuous . )
this gives rise to the following bound .
[ prop : restrictmulti ] for @xmath226 and @xmath227 where @xmath131 , we have @xmath404 as with the bound in corollary [ cor : restriction ] , this bound involves checking an exponential number of cases , so a more practical bound would be useful .
we can obtain one by restricting only to pairs of indices , i.e. where @xmath405 ( for some @xmath406 ) , as shown below .
[ prop : restrictpairs ] for @xmath226 and @xmath227 where @xmath131 , we have @xmath407 however , a refinement of this bound can be obtained as a special case of a bound on @xmath132 for arbitrary values of @xmath0 , as given in the next section .
when considering generalized covering designs of strength @xmath134 , it is more difficult to obtain bounds on @xmath132 as there are a variety of possible @xmath84-admissible vectors @xmath83 .
however , one can obtain bounds by considering just one possible `` shape '' of such vectors . in this section , we consider only the @xmath84-admissible vectors consisting of @xmath0 entries of 1 and all other entries 0 , to obtain a bound which is an analogue of the schnheim bound ( theorem [ thm : schonheim ] ) .
[ prop : revised_bound ] suppose that @xmath226 and @xmath227 where @xmath131 , and that @xmath408 .
let @xmath409 , and let @xmath4 be a collection of blocks in @xmath74 which contain each @xmath0-tuple of the form @xmath410 , where @xmath411 .
then @xmath412 the proof is by induction on @xmath0 . if @xmath141 , then the result is obvious .
suppose that the result holds for any set @xmath413 .
now choose @xmath409 , and let @xmath4 be as defined in the statement of the proposition .
note that the total number of occurences ( with repetition ) of elements of @xmath414 in @xmath4 is @xmath415 .
moreover , for each element @xmath416 , @xmath4 must contain the @xmath0-tuple @xmath417 for any choice of elements @xmath411 , where @xmath418 . by the induction hypothesis
, it follows that @xmath4 must contain at least @xmath419 blocks for each element of @xmath414 . in total
, this means that the number of occurences ( with repetition ) of elements of @xmath414 in @xmath4 is at least @xmath420 hence @xmath421 and so we conclude that @xmath412 by considering all possible choices of @xmath422 , we obtain the following bound as a direct consequence .
[ cor : allposs ] if @xmath226 and @xmath227 where @xmath131 , and if @xmath423 , then @xmath424 given @xmath425 , consider the vector @xmath83 which has 1 in positions @xmath426 , and 0 elsewhere ; this vector will always be @xmath84-admissible regardless of @xmath223 . corresponding to the vector @xmath83 , we obtain that the set @xmath96 of blocks of a @xmath427 must contain each @xmath0-tuple of the form @xmath410 , where @xmath411 . by proposition [ prop : revised_bound ]
, the number of blocks must therefore be at least @xmath428 in the special case where @xmath21 , this reduces to the following .
[ cor : restrictpairs2 ] for @xmath226 and @xmath227 where @xmath131 , we have @xmath429 we remark that this second corollary gives the desired refinement of the bound in proposition [ prop : restrictpairs ] .
our first construction is based on the block - recursive construction for covering arrays that appears in poljak _ et al .
_ @xcite and also stevens and mendelsohn @xcite , and which was later extended to mixed covering arrays by colbourn _
_ @xcite .
this construction uses two strength-@xmath0 generalized covering designs to construct another strength-@xmath0 generalized covering design .
we recall the notion of concatenation of vectors from section [ section : defn ] .
[ thm : blockrec ] suppose @xmath142 and @xmath72 are @xmath58-tuples with @xmath131 and @xmath430 and @xmath431 are @xmath58-tuples with @xmath432 .
assume that @xmath433 is a generalized covering design @xmath130 with @xmath434 blocks , and that @xmath435 is a @xmath436 with @xmath437 blocks .
then there exists a @xmath438 with @xmath439 blocks .
assume that @xmath440 are the blocks of @xmath433 and @xmath441 are the blocks of @xmath435 .
since each @xmath442 is an @xmath41-tuple of sets of sizes @xmath75 , and each @xmath443 is an @xmath58-tuple of sets of sizes @xmath444 , we have that @xmath445 is the @xmath61-tuple of sets ( of sizes labelled by @xmath446 ) formed by the concatenation of @xmath442 and @xmath443 .
we claim that the set of blocks @xmath447 form a generalized covering design @xmath448 with @xmath439 blocks . to see that this claim is true , let @xmath449 be any @xmath61-tuple of sets that is @xmath450-admissible .
we will prove that there exists some block @xmath445 that contains @xmath88 .
consider the @xmath41-tuple @xmath451 formed by taking the first @xmath41 sets in @xmath88 .
clearly , @xmath452 , and since @xmath433 is a generalized covering design , by proposition [ prop : reducet ] , the @xmath58-tuple @xmath453 is contained in some block @xmath442 of @xmath96 .
similarly , if we define @xmath454 , there is a block @xmath455 of @xmath435 that contains @xmath456 .
thus @xmath457 is contained in the block @xmath445 .
this construction can be used to get an upper bound on the size of generalized covering designs .
[ cor : bcbound ] suppose we have @xmath142 and @xmath72 with @xmath131 , and @xmath458 and @xmath459 with @xmath432 .
then for all appropriate values of @xmath0 , @xmath460 unfortunately , this construction can lead to very poor upper bounds on the number of blocks in a generalized covering design .
for example , consider a @xmath113 with the condition that no entries of @xmath223 are equal to one .
then the block - recursive construction produces the bound @xmath461 whereas it follows from corollary [ cor : equiv ] the actual value of @xmath462 is equal to @xmath463 ( again , provided that all of the entries of @xmath223 are at least two ) .
however , there is a modification of this construction that in some circumstances produces a better bound than the one given in corollary [ cor : bcbound ] . again , suppose that @xmath142 and @xmath72 with @xmath131 , that @xmath430 and @xmath431 with @xmath432 , and that we have two generalized covering designs , a @xmath130 and a @xmath436 .
assume that @xmath440 are the blocks of the first design and @xmath441 are the blocks of the second design , and further assume that @xmath464 .
the key to this improvement is to consider the set of @xmath434 blocks @xmath465 ( in the last @xmath466 blocks of @xmath467 , the @xmath41-tuple @xmath468 could be replaced with any block of @xmath4 . )
we note that @xmath467 contains any @xmath450-admissible vector @xmath449 with the condition that either @xmath469 to extend the set @xmath467 to a generalized covering design , we need to add blocks that will cover all the remaining @xmath450-admissible vectors .
these vectors will all be of the form @xmath449 where @xmath470 it is possible to construct such blocks by applying the block - recursive construction to a and a @xmath471 . using this modification to the block - recursive construction
, we obtain the following bound .
[ thm : blockrecsimp ] suppose we have @xmath142 and @xmath72 with @xmath131 , and @xmath430 and @xmath431 with @xmath432 .
then for all suitable @xmath472 , @xmath473 two special cases are worth mentioning .
first , in the case @xmath141 , we obtain @xmath474 ( thanks to the convention that @xmath475 ) ; in fact , proposition [ prop : tis1 ] ensures that this bound holds with equality .
second , in the case where @xmath21 , theorem [ thm : blockrecsimp ] can be simplified by applying proposition [ prop : tis1 ] .
[ cor : blockrecsimp2 ] suppose we have @xmath142 and @xmath72 with @xmath131 , and @xmath430 and @xmath431 with @xmath432 .
then the size of a @xmath476 is no more than @xmath477 note that , because the modified block - recursive construction of theorem [ thm : blockrecsimp ] has two stages , it is possible that the design it produces may have repeated blocks ( even if the inputted designs did not ) .
we now illustrate this construction of theorem [ thm : blockrecsimp ] with an example ( and one where that possibility arises ) .
let @xmath478 below be the blocks of a @xmath479 .
@xmath480 further , let @xmath481 below be the blocks of a @xmath482 .
@xmath483 these two designs can be used to build a @xmath484 with the @xmath485 blocks @xmath486 given below . @xmath487 the blocks @xmath488 through to @xmath489 are formed as the concatenation of blocks of the original designs .
the final @xmath490 blocks are formed by applying the block - recursive construction on a @xmath491 and a @xmath492 .
note that @xmath493 , so we have a repeated block , and thus @xmath494 may be deleted leaving a @xmath495 with 15 blocks .
we note that theorem [ thm : minimax ] gives an upper bound of @xmath496 blocks for a generalized covering design with these parameters , while our various lower bounds show that the minimum number of blocks is at least 10 ( this is obtained by using corollary [ cor : restriction2 ] ) . in a 1922 paper
@xcite , macneish gave a recursive construction for mutually orthogonal latin squares of order @xmath497 from those of orders @xmath41 and @xmath58 .
this was subsequently generalized to orthogonal arrays in 1952 by bush @xcite , whose construction can also be applied to covering arrays .
our second recursive construction for generalized covering designs is based on macneish s approach .
this construction takes two strength-@xmath149 generalized covering designs and produces a third strength-@xmath149 generalized covering design . as was the case with the block - recursive construction , we need to define an operation on vectors to define this construction .
let @xmath78 and @xmath498 be two @xmath41-tuples of positive integers .
define the _ hadamard product _ of @xmath170 and @xmath295 to be @xmath499 we also need to define an operation on sets .
let @xmath500 be a subset of @xmath501 and @xmath502 a subset of @xmath503 .
then define @xmath504 noting that @xmath505 is a subset of @xmath506 of size @xmath507 .
now , if @xmath508 and @xmath509 are two @xmath41-tuples of sets and @xmath78 , we define @xmath510 [ thm : macneish ] suppose that @xmath142 and @xmath72 with @xmath131 , and that @xmath458 and @xmath459 with @xmath432 .
now suppose we have generalized covering designs @xmath113 and @xmath511 with @xmath434 and @xmath437 blocks respectively .
then there exists a with @xmath439 blocks .
let @xmath440 be the blocks of the first design and @xmath441 be the blocks of the second design .
then we claim that the collection @xmath512 of @xmath41-tuples of sets forms the blocks of a @xmath513 .
clearly , @xmath514 .
let @xmath515 be any @xmath516-admissible vector .
we will show that this is contained in some block of the form @xmath517 .
first , consider the case where @xmath88 has one entry a set of size 2 , and all other entries are the empty set . in particular , we have that @xmath518 for some @xmath519 and that @xmath520 for all @xmath521 . then the set @xmath522 is a subset of @xmath99 .
( we use the convention that if @xmath523 , then @xmath524 , rather than 0 . )
construct an @xmath41-tuple @xmath453 of sets by taking the @xmath525 entry of @xmath453 to be @xmath526 and all other entries the empty set .
then @xmath453 is a @xmath527-admissible vector .
since @xmath96 is the set of blocks of a generalized covering design , there is a block @xmath442 that contains @xmath453 . similarly , the set @xmath528 is a subset of @xmath529 ( since both @xmath530 and @xmath531 ) .
then the @xmath41-tuple @xmath456 that has @xmath532 in the entry @xmath171 and the empty set in all other entries is a @xmath533-admissible vector .
since @xmath4 is the set of blocks of a generalized covering design , there is a block @xmath443 that contains @xmath456 .
now we notice that the block @xmath517 contains @xmath88 , since @xmath534 a similar argument also works for the case where the vector @xmath88 has two entries which are singletons and all other entries are the empty set . to see this ,
assume that @xmath535 with @xmath536 and @xmath537 ( where @xmath538 ) , and that @xmath539 for all @xmath540 . consider the @xmath58-tuples @xmath541 where @xmath542 and @xmath543 where @xmath544 as before , there is a block @xmath545 that contains @xmath453 and a block @xmath443 of @xmath4 that contains @xmath456 .
consequently , the block @xmath546 contains the vector @xmath88 .
once again , we illustrate this construction with an example .
below is an example of a @xmath547 with 3 blocks : @xmath548 from this we can construct the @xmath549 with @xmath550 blocks given below : @xmath551
in this paper , we have defined a very broad class of objects . consequently , there are a considerable variety of directions for future work on this topic .
our work in this paper has primarily been focused on strength-@xmath149 generalized covering designs , principally because of the interpretation in terms of clique coverings .
two natural directions would be ( i ) to obtain further results in this case , and ( ii ) to extend our results to values of @xmath134 .
there is an improvement on the block - recursive construction for covering arrays that uses two copies of the same covering array with _
disjoint blocks _ removed from the second array @xcite . with this improvement , covering arrays can be constructed that meet the asymptotic bound due to gargano
_ et al . _
it would be interesting to see if this improvement could also be applied to the block - recursive construction for generalized covering designs in subsection [ subsect : blockrec ] .
wilson gave a recursive construction for transversal designs ( described in @xcite ) that has been applied to both covering arrays and covering arrays with mixed alphabets .
( macneish s construction is a special case of this . )
it would be worthwhile to investigate how wilson s construction can be adapted to generalized covering designs . if we were to consider generalized covering designs of higher strength ( i.e. with @xmath134 ) , the natural extension of our approach would be to define a notion of clique coverings in @xmath0-uniform hypergraphs .
thus if a clique in a @xmath0-uniform hypergraph is defined to be a set @xmath270 of vertices such that any subset of @xmath270 of size at most @xmath0 is contained in a hyperedge , then a strength-@xmath0 generalized covering design would be equivalent to a clique covering of an appropriately - constructed @xmath0-uniform hypergraph .
another direction for further work on generalized covering designs is to conduct an analysis of families of these designs with specific parameters .
for example , one family that could be examined is that of all generalized covering designs with @xmath552 .
such a design would be , in some sense , close to a covering design .
alternatively , one could consider the all generalized covering designs for which @xmath553 with @xmath554 ; such designs would not be far removed from covering arrays . for small values of @xmath5 one could consider all generalized covering designs with @xmath555 with @xmath556 . in @xcite
, cameron gives a complete description of generalized @xmath0-designs for @xmath557 and @xmath558 : such a characterization could potentially be obtained for generalized covering designs . with small values of @xmath559 , computer searches could be implemented to compare our various constructions . with sufficiently many parameters fixed ,
it may also be possible to determine the asymptotic growth of the optimal size of such a generalized covering design . in @xcite
, cameron also suggests the dual notion of generalized packing designs ; the first two authors have also been investigating this problem @xcite .
while the definition is very similar to that of generalized covering designs , as is often the case with packing problems , the theory turns out to be quite different .
various previously - known classes of designs arise as generalized packing designs ; examples include howell designs , room squares , hanani triple systems , and mutually orthogonal latin rectangles . finally , we recall that much of the motivation for work on covering designs and covering arrays was their widespread use in applications ( particularly in communications and software testing )
. it would be very interesting to discover applications for other classes of generalized covering designs .
the work in this paper was a joint project of the discrete mathematics research group at the university of regina , attended by all the authors .
they would like to thank the other members of the group ( b. ahmadi , f. alinaghipour , s. m. fallat , j. c. fisher , a. m. purdy and s. zilles ) for their participation .
the authors also thank the two anonymous referees for their helpful suggestions .
r. f. bailey is a pims postdoctoral fellow .
a. c. burgess and m. s. cavers are nserc postdoctoral fellows .
k. meagher acknowledges support from an nserc discovery grant . | inspired by the `` generalized @xmath0-designs '' defined by cameron [ p. j. cameron , a generalisation of @xmath0-designs , _ discrete math . _
* 309 * ( 2009 ) , 48354842 ] , we define a new class of combinatorial designs which simultaneously provide a generalization of both covering designs and covering arrays . we then obtain a number of bounds on the minimum sizes of these designs , and describe some methods of constructing them , which in some cases we prove are optimal . many of our results are obtained from an interpretation of these designs in terms of clique coverings of graphs .
+ * keywords : * covering design ; covering array ; generalized covering design ; clique covering . +
* msc2010 : * 05b40 ( primary ) ; 05b15 , 05c70 ( secondary ) . |
the study of graphs with forbidden subgraphs has attracted researchers from both mathematics and computing .
the popular ones are chordal graphs @xcite , chordal bipartite graphs @xcite , etc .
largely , these graphs were discovered to study the gap between np - completeness and polynomial - time solvability of a combinatorial problem . on the similar line ,
special graph classes like @xmath0-free graphs @xcite , planar graphs @xcite , interval graphs @xcite , circular - arc graphs @xcite , etc . , were discovered in the literature . in particular , @xmath0-free graphs are well studied in the literature @xcite as it contains split graphs @xcite and co - chordal graphs @xcite as its proper subgraph classes . in this paper , we revisit @xmath0-free graphs and investigate from structural and combinatorial perspectives .
this line of study has been considered important in the literature as structural observations resulting from this study may yield polynomial - time algorithms for well - known combinatorial problems such as vertex cover , coloring , etc .
+ while @xmath0-free graphs received good attention in the past , the fundamental problems like structural characterization using minimal vertex separator has not been addressed in the literature .
this line of study is fruitful as it yields efficient algorithm for testing @xmath0-free graphs . like chordal graphs @xcite
, we also give a structural characterization in terms of minimal vertex separators .
further , we show that @xmath0-free graphs have polynomially many minimal vertex separators and present an algorithm to list all of them . as a consequence ,
we show that classical problems related to constrained vertex separators such as finding minimum stable vertex separator and minimum connected vertex separator are polynomial - time solvable restricted to @xmath0-free graphs .
these problems are np - complete for general graphs .
it is important to highlight that finding minimum connected vertex separator is np - complete in chordality 5 graphs @xcite . in this paper , we identify the first non - trivial subclass of chordality 5 graphs which are @xmath0-free graphs and show that finding minimum connected vertex separator is polynomial - time solvable .
+ due to its nice structure , it is known from @xcite that @xmath0-free graphs have polynomial number of maximal independent sets but there does not exist an explicit algorithm in the literature , to list all maximal independent sets in polynomial time . in this paper , we present an algorithm for enumerating all maximal independent sets , which runs in polynomial - time .
it is important to highlight that a well - known coloring problem has a polynomial bound in terms of the maximum size of a complete subgraph for @xmath0-free graphs @xcite and @xmath1-colorability problem can be solved in polynomial - time for its super graph class @xmath2-free graphs @xcite .
although , @xmath1-colorability problem is addressed in the super graph class of @xmath0-free graphs , we present an alternative algorithm for this problem in terms of mis .
in addition to this , we also shows that a well - known combinatorial problem of finding minimum feedback vertex set can be solved in polynomial - time for the subclass of @xmath0-free graphs , whereas these problems are np - complete in general graphs and polynomial when restricted to chordal @xcite and chordal bipartite graphs @xcite .
+ we use standard graph - theoretic notations as in @xcite .
in particular , we write @xmath3 and @xmath4 to denote the vertex set and edge set of a graph @xmath5 .
the @xmath6 of a vertex @xmath7 of @xmath5 , @xmath8(@xmath7 ) , is the set of vertices adjacent to @xmath7 in @xmath5 .
the degree of the vertex @xmath7 is @xmath9 .
@xmath10 and @xmath11 denotes the minimum and maximum degree of a graph @xmath5 , respectively .
@xmath12 is a _ path _ defined on @xmath13 such that @xmath14 . for simplicity
, we use @xmath15 to refer @xmath16 .
the set @xmath17 denotes the _ internal vertices _ of the path @xmath18 .
@xmath19 denotes the path on @xmath20 vertices .
a _ cycle _
@xmath21 on @xmath20-vertices is denoted as @xmath22 , where @xmath23 and @xmath24 .
the graph @xmath5 is said to be @xmath25 if every pair of vertices in @xmath5 has a path and if the graph is not connected it can be divided into disjoint connected @xmath26 @xmath27 , @xmath28 , where @xmath29 denotes the set of vertices in the component @xmath30 . a connected component
@xmath30 is trivial if @xmath31 .
it is non - trivial if @xmath32 .
+ the graph @xmath33 is called a @xmath34 of @xmath5 if @xmath35 @xmath36 @xmath3 and @xmath37 .
the subgraph @xmath33 of a graph @xmath5 is said to be @xmath38 @xmath34 , if for every pair of vertices @xmath39 and @xmath7 of @xmath33 , \{@xmath40 } @xmath41 @xmath42 if and only if \{@xmath40 } @xmath41 @xmath4 and it is denoted by @xmath43 $ ] .
let @xmath44 be a non - empty subset of @xmath3 and let @xmath45 denotes the induced subgraph on @xmath46 vertices .
the set @xmath44 is said to be _ independent set ( stable set ) _ if every pair of vertices of @xmath44 is non - adjacent .
the set @xmath44 is said to be _ clique _ if every pair of vertices of @xmath44 is adjacent . for a non - trivial non - complete graph
, a subset @xmath47 is said to be a _
vertex separator _ if @xmath48 gives distinct connected components .
@xmath49 is minimal vertex separator if there does not exist a subset @xmath50 such that @xmath51 is a vertex separator . for a trivial graph , the graph itself a minimal vertex separator . for a non - trivial complete graph on @xmath20 vertices ,
neighborhood of every vertex is a minimal vertex separator .
a minimal vertex separator is said to be a stable separator if it forms stable set .
a minimal vertex separator is said to be a clique separator if it forms clique .
two edges @xmath52 and @xmath53 are said to induce @xmath0 if @xmath54 , @xmath55 , @xmath56 , @xmath57 @xmath58 .
a graph is @xmath0-free if it has no such @xmath59 and @xmath60 as an induced sub graph .
it is apparent that if a graph @xmath5 is @xmath0-free then the complement of @xmath5 has no induced cycle of length 4 .
let @xmath5 be a @xmath0-free graph with the vertex set @xmath3 and edge set @xmath4 such that @xmath61 and @xmath62 respectively .
we use this notation throughout this paper for analysis purpose .
in this section , we first present an alternative characterization of @xmath0-free graphs using forbidden induced subgraphs .
we use this result in the study of minimal vertex separators and testing @xmath0-free graphs .
a connected graph is @xmath0-free if and only if it forbids @xmath63 , @xmath64 and @xmath65 as an induced subgraphs .
* necessity : * on the contrary , assume that @xmath5 contains @xmath63 or @xmath64 or @xmath65 as an induced subgraph .
clearly , the edges @xmath66 and @xmath67 form @xmath0 , which is a contradiction . + * sufficiency : * on the contrary , assume that @xmath5 is not @xmath0-free i.e. , @xmath5 contains @xmath0 as an induced subgraph .
let @xmath68 and @xmath69 be any two edges in @xmath5 which induces @xmath0 .
therefore , @xmath70 .
since @xmath5 is connected , there exists a shortest path @xmath71 such that @xmath72 . ] [ 2k2 ] case 1 : : : neither @xmath39 is adjacent to an internal vertex in @xmath73 nor @xmath74 is adjacent to an internal vertex in @xmath73 .
+ since @xmath72 , @xmath5 contains induced @xmath75 , @xmath76 where @xmath75=@xmath77 .
thus , @xmath5 contains an induced @xmath78 ( _ see fig.1.(a ) _ ) .
case 2 : : : either @xmath39 is adjacent to an internal vertex in @xmath73 or @xmath74 is adjacent to an internal vertex in @xmath73 .
assume that @xmath39 is adjacent to an internal vertex in @xmath73 .
since @xmath72 , there exist an internal vertex @xmath79 such that @xmath80 .
if @xmath81 then , @xmath5 has an induced @xmath82 where @xmath83 forms an induced @xmath84 ( _ see fig.1.(b ) _ ) . if @xmath85 then , @xmath5 contains induced @xmath75 , @xmath76 where @xmath75=@xmath86 .
thus , @xmath5 contains an induced @xmath78 ( _ see fig.1.(c ) _ ) .
case 3 : : : @xmath39 is adjacent to an internal vertex in @xmath73 and @xmath74 is adjacent to an internal vertex in @xmath73 . + @xmath87 ; ; if @xmath88 then , there exist an internal vertex @xmath79 such that @xmath89 .
thus , @xmath5 contains an induced @xmath65 with @xmath90=@xmath91 where @xmath83 and @xmath92 forms an induced @xmath84 ( _ see fig.1.(d ) _ ) .
@xmath87 ; ; if @xmath93 then , there exist internal vertices @xmath94 such that @xmath95 . if @xmath81 then , @xmath5 contains an induced @xmath64 with @xmath96=@xmath97 where @xmath83 forms an induced @xmath84 ( _ see fig.1.(e ) _ ) .
the argument is similar , if @xmath98 . if neither @xmath81 nor @xmath98 then , ( @xmath99 ) forms an induced @xmath78 ( _ see fig.1.(f ) _ ) .
@xmath87 ; ; if @xmath100 then , @xmath5 contains induced @xmath75 , @xmath76 with @xmath101 .
thus , @xmath5 contains an induced @xmath78 . in all the above cases @xmath5 contains @xmath78 or @xmath82 or @xmath102 , which is a contradiction .
thus @xmath5 is a @xmath0-free graph .
@xmath103 let @xmath5 be a graph and @xmath104 .
a vertex @xmath105 is said to be a _
universal vertex _ if @xmath106 .
an edge @xmath107 is said to be a _
universal edge _ if @xmath108 , either @xmath109 or @xmath110 .
let @xmath5 be a connected graph and @xmath44 be any minimal vertex separator of @xmath5 .
let @xmath111 , @xmath112 be the connected components in @xmath45 .
@xmath5 is @xmath0-free if and only if it satisfies the following conditions : * @xmath45 contains at most one non - trivial component .
further , if @xmath45 has a non - trivial component , say @xmath113 , then the graph induced on @xmath114 does not contain @xmath78 , @xmath82 , @xmath102 as an induced subgraphs . *
every trivial component of @xmath45 is universal to @xmath44 . *
every edge in the non - trivial component of @xmath45 is universal to @xmath44 . *
the graph induced on @xmath115 is either connected or has at most one non - trivial component .
further , if the graph induced on @xmath115 has a non - trivial component , say @xmath116 , then the graph induced on @xmath117 does not contain @xmath78 , @xmath82 , @xmath102 as an induced subgraphs . *
if @xmath44 and @xmath45 has a non - trivial component , say @xmath116 and @xmath113 , respectively , then every edge @xmath118 in @xmath116 is universal to @xmath33 , where @xmath119 .
* necessity : * let @xmath5 be a @xmath0-free graph .
( i ) : : on the contrary , assume that there are at least two non - trivial components in @xmath45 . without loss of generality ,
let @xmath113 and @xmath120 be any two non - trivial components in @xmath45 , then there exists at least one edge @xmath121 and @xmath122 in @xmath113 and @xmath120 , respectively .
the edges @xmath123 and @xmath124 are not adjacent and forms @xmath0 in @xmath5 , which is a contradiction . hence , our assumption that there are at least two non - trivial components in @xmath45 is wrong .
further , since every induced subgraph of a @xmath0-free graph is @xmath0-free , the graph induced on @xmath125 is @xmath0-free . by _
lemma 1 _ , @xmath126 does not contain an induced @xmath78 , @xmath82 , @xmath102 .
( ii ) : : on the contrary , assume that there exist a trivial component , say @xmath113 , in @xmath45 such that the vertex @xmath127 is not adjacent to at least one vertex , say @xmath128 , in @xmath44 .
the set @xmath129 forms a vertex separator for @xmath5 , which is a contradiction to our assumption @xmath44 is a minimal vertex separator .
( iii ) : : if @xmath45 has only trivial components , then there is nothing to prove .
assume that there exists a non - trivial component in @xmath45 , say @xmath113 .
on the contrary , assume that there exist an edge @xmath130 , which is not universal to @xmath44 .
i.e. , there exists at least one vertex , say @xmath131 , such that both @xmath132 and @xmath133 .
since @xmath44 is a minimal vertex separator , there exist a vertex @xmath134 , @xmath135 such that @xmath136 .
this implies , the edges @xmath107 and @xmath137 forms a @xmath0 in @xmath5 , which is a contradiction .
hence , every edge in non - trivial component of @xmath45 is universal to @xmath44 .
( iv ) : : on the contrary , the graph induced on @xmath115 has at least two non - trivial components , say @xmath138 , @xmath28 .
then there exist edges @xmath139 and @xmath140 which forms @xmath0 in @xmath5 , which is a contradiction .
thus , the graph induced on @xmath115 has at most one non - trivial component . since , every induced subgraph of a @xmath0-free graph is @xmath0-free , the graph induced on @xmath117 is @xmath0-free . by _
lemma 1 _ , @xmath117 does not contain an induced @xmath78 , @xmath82 , @xmath102 .
( v ) : : on the contrary , assume that there exist an edge @xmath141 is not universal to @xmath33 .
i.e. , there exist a vertex @xmath142 and @xmath143 such that @xmath144 and @xmath145 .
thus , @xmath146 and @xmath107 form @xmath0 in @xmath5 , which is a contradiction . *
sufficiency : * our claim is to prove every pair of edges in @xmath5 do not form @xmath0 .
let @xmath121 and @xmath147 be any two edges in @xmath5 . since , @xmath44 is a minimal vertex separator for @xmath5 , @xmath5 is a graph induced on @xmath148 .
let @xmath149 be the connected components in @xmath44 .
case 1 : : : @xmath150)$ ] + then there exist a non - trivial component in @xmath44 , say @xmath116 .
thus , @xmath151 . by ( iv ) , @xmath123 and @xmath124 do not form @xmath0 .
case 2 : : : if @xmath152)$ ] , @xmath153 and @xmath154 , @xmath155 .
+ * if @xmath45 has only trivial components . by ( ii ) , @xmath156 .
thus , @xmath123 and @xmath124 do not form @xmath0 . * if @xmath45 has a non - trivial component . by ( i ) there exist exactly one non - trivial component , say @xmath113 .
if @xmath154 , @xmath135 , then there is nothing to prove . if @xmath157 , then @xmath158 .
if @xmath159 , then @xmath123 and @xmath124 do not form @xmath0 .
if @xmath160 , then , @xmath161 or @xmath162 ( by ( v ) ) , thus , @xmath123 and @xmath124 do not form @xmath0 .
case 3 : : : if @xmath163 , @xmath164 and @xmath165 , @xmath166 . by ( i ) , @xmath30 or @xmath167 is a trivial component . without loss of generality , assume that @xmath30 is a trivial component . by ( ii ) , @xmath168 , thus , @xmath123 and @xmath124 do not form @xmath0 .
case 4 : : : if @xmath113 is a non - trivial component in @xmath45 and @xmath169 , then by ( i ) , @xmath123 and @xmath124 do not form @xmath0 . case 5
: : : if @xmath113 is a non - trivial component in @xmath45 , @xmath170 and @xmath171 , then by ( iii ) , @xmath123 and @xmath124 do not form @xmath0 .
case 6 : : : if @xmath113 is a non - trivial component in @xmath45 , @xmath172 and @xmath173 , then by ( i ) , @xmath123 and @xmath124 do not form @xmath0 . in all the above cases @xmath123 and @xmath124 do not form @xmath0 .
thus , @xmath5 is @xmath0-free . hence the theorem .
@xmath103 in any @xmath0-free graph @xmath5 , for every minimal vertex separator @xmath44 , the graph @xmath45 has atleast one trivial component and every trivial component is universal to the respective minimal vertex separator . using this observation
, we can enumerate all minimal vertex separators for a given @xmath0-free graph .
now , we present the algorithm as follows : the algorithm @xmath174 enumerates all minimal vertex separator of a @xmath0-free graph , @xmath5 .
every minimal vertex separator in @xmath5 has at most one non - trivial component .
since , @xmath5 is @xmath0-free , a non - trivial component of any vertex separator , @xmath175 , is universal to @xmath175 .
moreover , every minimal vertex separator , @xmath44 , in @xmath5 has at least one trivial component and all such trivial components are universal to @xmath44 .
if there exist a trivial component which is not universal to @xmath44 then , @xmath44 is not minimal . using these observations , we list all minimal vertex separators of a @xmath0-free graphs in _ steps 5 - 15 _ of _ algorithm 4_. thus , algorithm @xmath176 outputs all possible minimal vertex separators
. @xmath103 the time complexity of the algorithm @xmath174 is @xmath177 , which is polynomial in the input size , where @xmath178 is the maximum degree for @xmath5 .
it is known from @xcite that stable separator is np - complete .
the only graph class known in the literature where stable separator is polynomial - time solvable is the class of triangle free graphs . in this section ,
we show that minimum stable separator in @xmath0-free graphs is polynomial - time solvable . as far as clique separator
is concerned , from @xcite , it is known that clique separator is polynomial - time solvable in general graphs . for the sake of completeness we also present an algorithm to find the minimum clique separator in @xmath0-free graphs . *
finding minimum clique and stable separator : * + we shall now describe an algorithm for finding minimum clique separator and minimum stable separator in @xmath0-free graphs .
we enumerate all minimal vertex separator for a @xmath0-free graph , @xmath5 , using the algorithm @xmath179 . for every minimal vertex separator
we check whether it is clique or stable , if so , we find the minimum of all such separators .
this algorithm is correct because , the addition of a vertex to a separator which is neither clique nor independent , does not create a clique(stable ) vertex separator .
note that , it is not necessary that every @xmath0-free graph has a clique as well as stable separator .
the time complexity of this algorithm is @xmath180 , which is polynomial in the input size , where @xmath178 is the maximum degree for @xmath5 .
* finding minimum connected vertex separator : * + from @xcite , it is known that minimum connected vertex separator is np - complete in chordality-5 graphs . in this paper , we show that on @xmath0-free graphs , minimum connected vertex separator is polynomial - time solvable which is a non - trivial subclass of chordality-5 graphs .
* sketch of the algorithm : * for the input @xmath0-free graph @xmath5 with @xmath20 vertices , enumerate all minimal vertex separators .
now , we group the minimal vertex separators into two lists @xmath181 and @xmath182 such that @xmath181 consists of all minimal vertex separators , whose removal from @xmath5 gives exactly two components and @xmath182 consists of remaining minimal vertex separators .
next , we sort @xmath181 and @xmath182 based on its cardinality .
now , search for the minimum connected vertex separator in @xmath181 and @xmath182 , if exists say @xmath183 and @xmath184 respectively .
finally , we compare @xmath183 ( if exists ) , @xmath184 ( if exists ) and the first minimal vertex separator in @xmath182 along with one of its trivial components .
subsequently , the one with least cardinality is returned as output .
the algorithm @xmath185 returns minimum connected vertex separator .
we first enumerate all minimal vertex separators and then we partition the minimal vertex separators into two lists , namely @xmath181 and @xmath182 , such that @xmath181 contains all the minimal vertex separators which gives exactly two components on its removal from the graph and @xmath182 contains all the minimal vertex separators which gives more than two components on its removal from the graph .
now we sort @xmath181 and @xmath182 based on the cardinality of the separators such that @xmath186 , where @xmath187 and @xmath188 , where @xmath189 .
note that minimum of all the minimal vertex separators gives the minimum vertex separator , using this fact we search for the first connected vertex separators in @xmath181 and @xmath182 separately , say @xmath190 and @xmath191 . if such @xmath190 and @xmath191 exist in @xmath181 and @xmath182 , respectively , we compare @xmath192 , @xmath193 and @xmath194 and outputs the minimum , where @xmath195 has more than two components so @xmath196 along with the trivial component forms a connected vertex separator .
if @xmath190 does not exist , then compare @xmath193 and @xmath194 and outputs the minimum .
if @xmath191 does not exist , then compare @xmath192 and @xmath194 and outputs the minimum .
if both does not exist @xmath194 will be the minimum .
thus , the algorithm @xmath185 outputs the minimum connected vertex separator .
@xmath103 the algorithm @xmath185 takes @xmath177 in _ step 3 _ , @xmath197 for _ steps 7 - 16 _ , @xmath198 for _ steps 17 - 18 _ , @xmath199 for _ steps 19 - 32 _ and constant time for the _ steps 33 - 53_. thus , the time complexity of the algorithm @xmath185 is @xmath200 , which is polynomial in the input size .
farber proves that the complement class of @xmath0-free graphs ( @xmath201-free graphs ) has polynomial number of maximal cliques @xcite . from this
we can say that there are polynomial number of maximal independent sets in @xmath0-free graphs .
we enumerate all possible maximal independent sets for a given @xmath0-free graph in the following algorithm .
let @xmath5 be a @xmath0-free graph .
the set @xmath202 , output by algorithm @xmath203 , is a maximal independent set .
first we have to prove that the algorithm outputs only the independent sets .
i.e. , to prove for any two vertices in @xmath202 , say @xmath204 , there does not exists an edge @xmath205 .
consider a minimal vertex separator @xmath44 in @xmath5 , let @xmath206 be the components in @xmath45 .
if @xmath45 has only trivial components then the algorithm returns the trivial components as one of the solution , which is clearly a maximal independent set and the remaining solutions are obtained by considering @xmath44 .
if @xmath45 has a non - trivial component , say @xmath183 , then the set @xmath202 consists of elements of @xmath207 ( which are trivial components ) and the other elements of @xmath202 are obtained recursively by considering @xmath183 .
the algorithm terminates when the non - trivial component is a clique .
let the solution returned by @xmath183 is @xmath208 .
since @xmath183 is a connected component of @xmath45 , clearly any element of @xmath208 does not have any adjacency from @xmath209 .
thus , the set output by the algorithm is a independent set .
now , our claim is to prove that every set is maximal .
let @xmath202 be any independent set , which results from the algorithm .
every vertex in @xmath210 is a vertex in some separator , thus an addition of a vertex from @xmath210 to @xmath202 will contradict the definition of independent set , by _ theorem 1.(ii)_. therefore , @xmath202 is a maximal independent set .
hence the lemma .
@xmath103 the algorithm @xmath203 enumerates all maximal independent sets for the given connected @xmath0-free graph , @xmath5 .
we shall prove the lemma by mathematical induction on the number of vertices @xmath20 of @xmath5 .
+ _ basis step : _ for all the connected @xmath0-free graphs on @xmath20 vertices , @xmath211 , the algorithm enumerates all the maximal independent sets .
+ _ hypothesis : _ assume that the lemma is true for connected @xmath0-free graphs with fewer vertices than @xmath20 , @xmath212 + _ induction step : _ let @xmath5 be a connected @xmath0-free graph with @xmath20 vertices .
our claim is to prove that the algorithm enumerates all the maximal independent sets for @xmath5 .
let @xmath213 , where @xmath7 is the minimum degree vertex of @xmath5 .
if @xmath45 has only trivial components , then @xmath45 forms one maximal independent set and by the hypothesis , all other maximal independent sets of @xmath44 will be enumerated .
if @xmath45 has a non - trivial component , then by hypothesis , the algorithm enumerates all the maximal independent sets in the non - trivial component of @xmath45 , say @xmath214 .
now , add the trivial components in @xmath45 to each of the maximal independent set @xmath215 , @xmath216 , which is a maximal independent set for @xmath5 .
thus , all the maximal independent sets of @xmath5 are enumerated .
@xmath103 + let @xmath5 be a @xmath0-free graph .
enumeration of all minimal vertex cover of @xmath5 can be done in polynomial - time and thus , the minimum vertex cover of @xmath5 .
enumeration of all minimal vertex cover can be obtained by taking the complement of each maximal independent set obtained in the _ algorithm 7_. thus , the minimum vertex cover for a given connected @xmath0-free graph can be obtained by searching for the minimal vertex cover with the least cardinality .
@xmath103 + the algorithm @xmath203 enumerates all maximal independent sets for a given connected @xmath0-free graphs .
thus , the maximum independent set for a given connected @xmath0-free graph can be obtained by searching for the maximal independent set with the largest cardinality . in algorithm
@xmath217 @xmath218 : the time complexity for enumerating all minimal vertex separators for @xmath5 in _ step 3 _ is @xmath177 , the _ steps 4 - 14 _ takes @xmath219 time .
thus , the algorithm takes @xmath219 time , which is polynomial in the input size . a well - known combinatorial problem , @xmath1-colorability in @xmath0-free graphs
is known to be solvable in polynomial time from its super graph class @xmath2-free graphs @xcite .
we provide an alternative algorithm for 3-colorability in terms of mvs and mis .
[ color ] for any connected 3-colorable graph @xmath5 , at least one of the proper coloring has an mis as one of its color class .
since , @xmath5 is @xmath1-colorable , it has three color classes , namely @xmath220 .
since , @xmath221 is a proper coloring for @xmath5 , each @xmath222 is an independent set . if any one of @xmath221 forms an mis for @xmath5 , then the lemma is true .
assume that none of the color classes , @xmath221 , forms an mis for @xmath5 , then consider the color class @xmath223 , add vertices to @xmath223 from @xmath224 and @xmath225 such that it obeys the property of independent set .
do this process , until no vertices can be added to @xmath223 from @xmath224 and @xmath225 .
thus , @xmath223 forms an mis for @xmath5 . hence proved .
we now describe an algorithm for checking whether the given @xmath0-free graph , @xmath5 , is @xmath1-colorable or not .
first , we check whether @xmath5 is bipartite or not .
if so , we say @xmath5 is 2-colorable .
if not , we enumerate all maximal independent sets and for every maximal independent set , @xmath202 , we check whether @xmath210 is bipartite or not .
if so , we say @xmath5 is 3-colorable . if not , @xmath5 is not 3-colorable .
the algorithm is correct , by _ lemma [ color]_. this algorithm is applicable for any graph class in which enumerating all maximal independent set can be done in polynomial - time .
as mentioned before , if @xmath5 satisfies all the conditions in _ theorem 1 _ , then @xmath5 is a @xmath0-free graph .
we next show an interesting claim which says that in any @xmath0-free graphs , neighborhood of a minimum degree vertex is a minimal vertex separator .
this claim helps us to choose one minimal vertex separator from the given graph and the testing can be done more efficiently .
let @xmath5 be a connected @xmath0-free graph and @xmath226 , @xmath227 .
then , @xmath228 is a minimal vertex separator of @xmath5 .
let @xmath39 be a minimum degree vertex in @xmath5 such that @xmath229 and @xmath230 , @xmath231 .
let @xmath232 be a vertex separator of size @xmath233 .
let @xmath234 @xmath112 be the connected components of @xmath45 and let @xmath235 . if @xmath236 , @xmath237 , then there is nothing to prove .
assume that there exist at least one vertex , say @xmath238 , in @xmath45 such that @xmath239 .
therefore , there exist an edge @xmath240 such that @xmath241 and belongs to one connected component in @xmath45 , say @xmath120 . our claim is to prove that @xmath44 is a minimal vertex separator . on the contrary ,
assume that @xmath44 is not minimal , then there exists a proper subset @xmath175 of @xmath44 such that @xmath175 forms a minimal vertex separator .
this implies , in @xmath242 , @xmath243 forms a non - trivial connected component .
@xmath120 is also a non - trivial connected component in @xmath242 , which is a contradiction as per _
theorem 2_. hence our assumption that @xmath44 is not a minimal vertex separator is wrong .
therefore , @xmath228 is a minimal vertex separator . @xmath103
using all the above mentioned observations , we can test whether a given graph is @xmath0-free or not .
now , we present an algorithm for testing whether the given graph is @xmath0-free or not
. + [ test ] the algorithm @xmath244 @xmath245 checks whether the given graph satisfies all the conditions of _ theorem 1 _ : the first three conditions are checked inside the loop in _ steps 6 - 16 , 27 _ , where , condition ( i ) in _ steps 6 - 7 _ and in _ step 27 _ , condition ( ii ) in _ steps 10 - 11 _ , condition ( iii ) in _ step 13 _ and the conditions ( iv ) and ( v ) are checked in _ steps 18 - 23 _ and in
_ step 25_. since , these conditions will obviously check the minimality of @xmath44 ( by condition ( ii ) and ( iii ) ) , the algorithm @xmath244 @xmath245 outputs whether the given graph is @xmath0-free or not .
+ let @xmath5 be a graph with @xmath20 vertices and @xmath246 edges .
in algorithm @xmath244 @xmath245 : the time complexity for checking vertex universal is @xmath247 , where @xmath248 is the number of vertices in @xmath44 .
the number of edges in @xmath44 is at most @xmath249 and it takes @xmath250 times for checking the edge universal from the non - trivial component to @xmath44 . hence , for _ steps 6 - 16 _ the algorithm takes @xmath250 times .
it takes @xmath251 times for checking the edge universal from the @xmath44 to non - trivial component .
hence , for _ steps 18 - 23 _ the algorithm takes @xmath251 .
overall , the algorithm @xmath244 @xmath245 takes @xmath252 , which is polynomial in the input size , where @xmath178 is the maximum degree for @xmath5 and @xmath248 is the minimum degree of @xmath5 .
+ time complexity of existing algorithms for testing @xmath0-free on @xmath20 vertices : trivial algorithm takes @xmath253 time for recognizing @xmath0-free graphs .
we also highlight that , @xmath201-free testing can be done in @xmath254 time @xcite .
note that , testing @xmath0-free is equivalent to testing @xmath201-free in the complement graph .
our algorithm takes @xmath252 time .
further , for @xmath178-bounded graphs , our algorithm runs in @xmath255 time .
although , in general our algorithm is as good as trivial algorithm , it is efficient for @xmath178-bounded graphs .
it is important to highlight that @xmath178-bounded graphs are extensively studied in the literature in the study of @xmath256-coloring in linear ( in @xmath178 ) time @xcite , 3-colorability of small diameter graphs @xcite , providing bounds for domatic and chromatic number @xcite etc .
, in this section , we trace _ algorithm [ test ] _ with two illustrations using which we highlight all cases of _ theorem 1_. + * illustration 1 : * we trace the steps of algorithm @xmath244 @xmath245 in the fig .
2 . * consider the graph @xmath5 with nine vertices as illustrated in _
2_. the vertex labelled @xmath257 is a minimum degree vertex with @xmath258 . * as shown in _ fig .
, @xmath113 and @xmath120 are the connected components in @xmath45 , where @xmath120 is a non - trivial component .
* when we call the algorithm @xmath259 , it returns @xmath257 , as all the edges in @xmath120 are universal to @xmath44 . * since @xmath260 , recursively call the algorithm @xmath244 @xmath261 .
* now , the minimum degree vertex in @xmath262 is @xmath263 and the vertex separator @xmath264 .
@xmath113 and @xmath120 are the connected components in @xmath45 , where @xmath120 is a non - trivial component .
* when we call the algorithm @xmath259 , it returns @xmath265 , as the edge @xmath266 is not universal to @xmath44 .
the edges @xmath267 and @xmath266 form @xmath0 in @xmath5 .
hence , the given graph @xmath5 is not @xmath0-free .
the algorithm is complete .
-free or not using algorithm [ test ] ] * illustration 2 : * we trace the steps of algorithm @xmath244 @xmath245 in the fig .
* consider the graph @xmath5 with eight vertices as illustrated in _ fig .
3_. the vertex labelled @xmath257 is a minimum degree vertex with @xmath268 . * as shown in _ fig .
, @xmath113 , @xmath120 , @xmath269 and @xmath270 are the connected components in @xmath45 , where @xmath269 is a non - trivial component . * when we call the algorithm @xmath271 , it returns @xmath257 , as the vertex @xmath272 in @xmath120 is universal to @xmath44 .
* when we call the algorithm @xmath273 , it returns @xmath257 , as all the edges in @xmath269 are universal to @xmath44 .
note that , @xmath274 .
* when we call the algorithm @xmath275 , it returns @xmath257 , as the vertex @xmath276 in @xmath270 is universal to @xmath44 . *
since @xmath44 has only trivial components , the algorithm is complete and outputs that @xmath5 is a @xmath0-free graph .
-free or not using algorithm [ test ] ]
note that by definition @xmath0-free forbids @xmath277 as an induced subgraph .
however , @xmath0-free graphs can allow @xmath84 , @xmath201 and @xmath278 .
we study subclass of @xmath0-free graphs by forbidding one or two cycles from @xmath84 , @xmath201 and @xmath278 .
further , we study the famous combinatorial problem which asks for the minimum number of vertex subset for the given graph whose removal creates a forest . in the literature
, this problem is known to be feedback vertex set problem .
@xmath279-free graphs form a proper subclass of @xmath0-free graphs , where every induced cycle is of length 5 .
we observed the following structural properties and concluded that it is a small graph class .
[ c3c4 ] if @xmath5 is a connected @xmath279-free graph , then for any minimal vertex separator @xmath44 of @xmath5 satisfies the following properties : * @xmath44 is an independent set . * if @xmath280 , then @xmath45 have exactly one trivial component . *
if @xmath45 has a non - trivial component , say @xmath113 , then for every edge @xmath281 , @xmath282 and @xmath283 .
i.e. , for every vertex @xmath153 , @xmath284 is an independent set . *
every vertex in a non - trivial component is adjacent to exactly one vertex in @xmath44 .
* on the contrary , assume that @xmath44 has at least one edge , say @xmath285 .
let @xmath30 be a trivial component in @xmath45 and let @xmath286 . since , @xmath5 is a @xmath0-free graph , @xmath287 ( by _ theorem 1.(ii ) _ ) .
thus , @xmath288 forms an induced @xmath84 , which is a contradiction to the definition of @xmath5 .
hence , @xmath44 is an independent set .
* on the contrary , assume that @xmath45 has at least two trivial components , say @xmath30 and @xmath167 .
let @xmath289 and @xmath290 .
let @xmath291 be any two vertices in @xmath44 .
by @xmath292 , @xmath293 and by _ theorem 1.(ii ) _ , @xmath294 , @xmath295 , @xmath296 , @xmath297 .
thus , @xmath298 forms an induced @xmath201 , which is a contradiction to the definition of @xmath5 .
hence , @xmath45 have exactly one trivial component if @xmath299 . * by _ theorem 1.(iii ) _ , every edge @xmath281 is universal to @xmath44 , thus , @xmath283 .
moreover , if @xmath300 , then every vertex in @xmath301 forms an induced @xmath84 together with @xmath39 and @xmath7 .
hence , @xmath282 . * on the contrary , assume that exist a vertex @xmath7 in a non - trivial component such that @xmath302 @xmath303 . by @xmath304 , there exist a trivial component in @xmath305 , say @xmath120 .
let @xmath306 .
therefore , @xmath307 forms an induced @xmath201 , which is a contradiction to the definition of @xmath5 . hence , the theorem .
@xmath103 from the above structural observations of minimal vertex separators in @xmath279-free graphs , @xmath5 : we can observe that the only possible structure of a non - trivial component after the removal of any minimal vertex separator from @xmath5 is @xmath308 and @xmath309 .
further , if @xmath310 , then the graph is @xmath311-free . if @xmath312 and if @xmath45 has a non - trivial component , then the graph is an induced @xmath278 .
thus , feedback vertex set problem can be solved in @xmath313 for @xmath279-free graphs .
@xmath314-free graphs are @xmath0-free graphs where every induced cycle is of length 4 .
this graphs can also be called as @xmath0-free chordal bipartite graphs .
the structural observations for this graphs are as follows : [ c3c5 ] if @xmath5 is a connected @xmath314-free graph , then for any minimal vertex separator @xmath44 of @xmath5 satisfies the following properties : * @xmath44 is an independent set . *
if @xmath45 has a non - trivial component , say @xmath113 , then for every vertex @xmath153 , @xmath284 is an independent set . * for every edge @xmath315 in a non - trivial component @xmath113 of @xmath45 , either @xmath39 is universal to @xmath44 and @xmath316 .
* the argument is similar to the proof in _ theorem [ c3c4].(i)_. * the argument is similar to the proof in _ theorem [ c3c4].(iii)_. * on the contrary , assume that there exist an edge @xmath281 such that @xmath317 , @xmath318 and @xmath319 . since , @xmath5 is @xmath0-free graph , @xmath283 and there exist a trivial component in @xmath45 , say @xmath120 .
let @xmath306 . by our assumption
, @xmath39 is adjacent to some vertex in @xmath44 , say @xmath320 and @xmath7 is adjacent to some vertex in @xmath44 , say @xmath74 , such that @xmath321 .
thus , @xmath322 forms an induced @xmath278 , which is a contradiction to the definition of @xmath5 .
@xmath103 although , it is known that the problem of finding minimum feedback vertex set in chordal bipartite graphs , a super class of @xmath0-free chordal bipartite graphs , is polynomial time solvable @xcite , using the above observation we provide a different approach for this problem in @xmath314-free graph . moreover , our approach takes linear time in terms of the input size .
[ fvsc3c5 ]
let @xmath5 be a connected @xmath314-free graph and @xmath44 be any minimal vertex separator of @xmath5 , then the cardinality of minimum feedback vertex set , @xmath323 , is equal to * @xmath324 , if @xmath45 has only trivial components , where @xmath325 is the set of all trivial components in @xmath45 .
* @xmath326 , if @xmath45 has a non - trivial component , @xmath113 , and @xmath113 is cycle - free , where @xmath327 is the set of all vertices in @xmath113 which are universal to @xmath44 . *
@xmath328 , if @xmath45 has a non - trivial component , @xmath113 and @xmath113 has at least one cycle . * if @xmath5 is a cycle - free graph , then either @xmath310 or @xmath329 .
thus , @xmath330 , which is minimum . without loss of generality , assume that @xmath5 has at least one cycle and @xmath305 has only trivial components , say @xmath234 , @xmath331 . by our assumption ,
@xmath332 and by _ theorem [ c3c5 ] _ , @xmath44 is an independent set .
let @xmath333 .
clearly , @xmath334 results in a forest , where @xmath335 - 1 vertices are chosen from @xmath44 and @xmath336 - 1 vertices are chosen from @xmath325 .
now , our claim is to prove the set @xmath323 is minimum . * * @xmath337 + on the contrary , assume that @xmath323 is not minimum , then the removal of @xmath175 vertices from @xmath5 results in a forest , where @xmath338 .
i.e. , @xmath44 has at least two vertices in @xmath334 , say @xmath339 .
therefore , @xmath340 forms an induced @xmath201 , which is a contradiction to the definition of @xmath323 . *
* @xmath341 + on the contrary , assume that @xmath323 is not minimum , then the removal of @xmath342 vertices from @xmath5 results in a forest , where @xmath343 .
i.e. , @xmath325 has at least two vertices in @xmath334 , say @xmath344 .
let @xmath320 and @xmath74 be any two vertices in @xmath44 .
therefore , @xmath340 forms an induced @xmath201 , which is a contradiction to the definition of @xmath323 .
+ hence , @xmath323 is the minimum feedback vertex set if @xmath45 has only trivial components . *
all possible structures of @xmath113 is given in _
fig.4_. from the structures of @xmath113 , it is clear that @xmath323 is a minimum feedback vertex set .
+ [ g123 ] when @xmath113 is cycle - free , title="fig : " ] * we prove this case separately for @xmath310 and @xmath299 .
* * @xmath310 and let @xmath345 .
+ it is clear that , every cycle of @xmath5 lies in @xmath113 .
thus , @xmath346 and the removal of @xmath347 vertices from @xmath327 results in a forest .
now , our claim is to prove that @xmath323 is minimum . on the contrary ,
assume that removing at least @xmath348 - 2 vertices from @xmath327 results in a forest .
i.e. , @xmath349 has at least two vertices in @xmath327 , say @xmath350 . since , @xmath5 is @xmath0-free @xmath351 .
note that , @xmath352 because every edge in @xmath113 is between an universal vertex and a non - universal vertex in @xmath113 , by _ theorem [ c3c5]_. similarly , @xmath353 .
thus , the only possibility is @xmath354 .
therefore , @xmath355 forms an induced @xmath201 , which is a contradiction to @xmath323 . *
* @xmath280 .
i.e. , @xmath44 has at least two vertices , say @xmath339 .
our claim is to prove @xmath44 is minimum . * * * @xmath356 + on the contrary , assume that there exist a proper subset @xmath33 of @xmath357 with cardinality @xmath358 , whose removal results in a forest , where @xmath359 .
suppose if we remove a vertex @xmath7 from @xmath327 to get a subset of @xmath33 , then for any @xmath360 , @xmath361 forms an induced @xmath201 , which is a contradiction .
similarly , if we remove a vertex @xmath362 from @xmath363 to get a subset of @xmath33 , then @xmath364 forms an induced @xmath201 , which is a contradiction .
thus , the set @xmath33 is not possible and @xmath323 is minimum . * * * @xmath365 + on the contrary , assume that there exist a proper subset @xmath33 of @xmath366 with cardinality @xmath358 , whose removal results in a forest , where @xmath367 .
suppose if we remove a vertex @xmath238 from @xmath327 to get a subset of @xmath33 , then for any @xmath368 , @xmath369 forms an induced @xmath201 ( by the previous subcase ) , which is a contradiction .
similarly , if we remove a vertex @xmath74 from @xmath370 to get a subset of @xmath33 , then for any @xmath359 , @xmath371 forms an induced @xmath201 , which is a contradiction .
thus , the set @xmath33 is not possible and @xmath323 is minimum .
+ hence , @xmath323 is a minimum feedback vertex set if @xmath45 has a non - trivial component , @xmath113 and if @xmath113 has at least one cycle . from all the above cases ,
it is proved that @xmath323 is a minimum feedback vertex set .
hence , the theorem .
@xmath103 _ theorem [ fvsc3c5 ] _ naturally yields an algorithm to find the minimum feedback vertex set in @xmath372 time , which is linear in the input size .
we have presented many structural observations of @xmath0-free graphs in terms of minimal vertex separator and using these results we have solved some of the unsolved combinatorial problems in the literature of @xmath0-free graphs .
the structural observations of the remaining four subclasses of @xmath0-free graphs are to be explored , to check the status of feedback vertex set problem in @xmath0-free graphs . | a connected graph is @xmath0-free if it does not contain a pair of independent edges as an induced subgraph . in this paper
, we present the structural characterization of minimal vertex separator and show that there are polynomial number of minimal vertex separators in @xmath0-free graphs .
further , using the enumeration we show that finding minimum connected vertex separator in @xmath0-free graphs is polynomial time solvable .
we highlight that finding minimum connected vertex separator is np - complete in chordality 5 graphs , which is a super graph class of @xmath0-free graphs .
other study includes , enumeration of all distinct maximal independent sets and testing @xmath0-free graphs . also , we present an polynomial time algorithm for feedback vertex set problem in the subclass of @xmath0-free graphs
. + + * keywords : * minimal vertex separator , constrained separators , @xmath0-free graphs . |
in nuclear physics , _ ab initio _ approaches seek to solve the many - body schrdinger equation in terms of constituent protons and neutrons interacting through nucleon - nucleon ( @xmath3 ) and three - nucleon ( @xmath4 ) forces that yield a high - precision fit of two- and three - body data .
their aim is twofold : firstly , to help unfold the true nature of the force among nucleons ; and secondly , to arrive at a fundamental understanding of nuclei and their role in the universe . in the three- and four - nucleon systems , where a numerically exact solution of the quantum - mechanical problem for both negative @xcite and positive energies @xcite is now possible ,
this goal has been largely achieved . for heavier systems , _ ab initio _ calculations have been mostly confined to the description of the bound - state properties of stable nuclei , but are now starting to be extended to dynamical processes between nuclei . the green s function monte carlo method has been used to describe the elastic scattering of neutrons on @xmath5he @xcite and to compute asymptotic normalization coefficients @xcite and nuclear widths @xcite . loosely bound and unbound nuclear states have been addressed within the coupled cluster technique @xcite by using a berggren basis and this method has been recently applied to compute elastic proton scattering on @xmath6ca @xcite .
an _ ab initio _ framework that promises to provide a unified treatment of a wide range of nuclear phenomena ( well - bound states , loosely - bound and unbound exotic nuclei , scattering and reaction observables ) is the no - core shell model with continuum ( ncsmc ) @xcite . here
, the nuclear many - body states are seen as superimpositions of continuous @xmath7 binary - cluster wave functions in the spirit of the resonating group method ( rgm ) @xcite and square - integrable eigenstates of the @xmath8-nucleon system , in which each cluster of nucleons and the compound nuclear states are obtained within the _ ab inito _ no - core shell model ( ncsm ) @xcite .
so far , we have laid the foundations of the ncsmc by developing the formalism to compute nucleon - nucleus collisions and applying it to the description of the unbound @xmath9he nucleus .
however , expansions on the ncsm / rgm portion of the basis @xcite have been already successfully used to describe nucleon @xcite and deuteron @xcite scattering on light nuclei and achieve the first _ ab initio _ description of the @xmath9be@xmath10b radiative capture @xcite and the @xmath11h@xmath12he and @xmath11he@xmath13he fusion rates @xcite , based on realistic @xmath3 interactions .
work is currently under way to incorporate the @xmath4 force into this binary - reaction formalism and to attain the description of deuteron - nucleus scattering and transfer reactions within the ncsmc approach .
achieving an _ ab initio _ treatment of three - cluster dynamics is another important stepping stone towards gaining a basic understanding of nuclei and their reactions . to cite a few instances , important nuclear fusion processes such as the @xmath11h@xmath14h@xmath15he or @xmath11he@xmath14he@xmath16he reactions
are characterized by three - body final states .
in addition , only with an approach capable of accounting for three - cluster configurations can one obtain an accurate description of borromean nuclei , ternary systems of two nucleons orbiting around a tightly bound core whose components are not bound in pairs .
finally , three - body configurations can be necessary even at very low energy to achieve a proper treatment of polarization and virtual excitations of breakup channels in reactions with weakly - bound projectiles such as the deuteron .
microscopic three - cluster models , where all nucleons are taken into account and the pauli principle is treated exactly , have been used for some time , particularly in combination with the hyperspherical formalism for the solution of the dynamic equations @xcite .
however , they have two main limitations : the use of central @xmath3 potentials with state - dependent parameters adjusted to reproduce the binding energy of the system under study , occasionally complemented with a spin - orbit interaction ; and a simplified description of the internal structure of the clusters , which are in most cases described by @xmath17-shell wave functions . in this paper , we report on an extension of the ncsm / rgm formalism to treat the dynamics among three nuclei made of fully antisymmetrized interacting nucleons . the solution of the three - cluster schrdinger equation is obtained by means of hyperspherical harmonic ( hh ) expansions on a lagrange mesh @xcite .
in addition , we present the first @xmath5he+@xmath1+@xmath1 investigation of the ground state ( g.s . ) of the @xmath18he nucleus based on a @xmath3 potential that yields a high - precision fit of the @xmath3 phase shifts and _ ab initio _ four - body wave functions for the @xmath5he cluster obtained consistently from the same hamiltonian . in particular , we employ a similarity - renormalization - group ( srg ) @xcite evolved chiral n@xmath11lo
@xmath3 @xcite potential .
for this first application , we include only the g.s . of the @xmath5he cluster and
estimate the importance of the core polarization by comparing the results obtained with six - body ncsm diagonalizations of the adopted hamiltonian .
the inclusion of excited states of @xmath5he to describe such effects is hard and not very efficient within the ncsm / rgm approach . on the other hand ,
core - polarization effects will be easily accounted for once the present formalism will be embedded within the ncsmc framework .
the paper is organized as follows . in section [ sec : formalism ] we define the microscopic three - cluster problem , present a brief overview of the hh functions and their application within the @xmath19-matrix method on lagrange mesh for the solution of the three - body bound - state problem , and introduce in detail the three - cluster ncsm / rgm formalism . in particular , in sec .
[ subsec : rgmkernels ] we present algebraic expressions for systems of two single nucleons plus a nucleus .
results for the g.s . of the @xmath18he borromean nucleus are presented in sec .
[ sec : applications ] , where we discuss calculations performed by solving the @xmath5he(g.s.)+@xmath1+@xmath1 ncsm / rgm equations and compare them with a diagonalization of the hamiltonian in the six - body ncsm model space .
conclusions and outlook are given in sec .
[ sec : conclusions ] .
finally , additional details on the formalism are presented in the appendix .
the intrinsic motion of a system of @xmath8 nucleons arranged into three clusters respectively of mass number @xmath20 , @xmath21 , and @xmath22 ( @xmath23 ) , can be described by the many - body wave function @xmath24 where @xmath25 are continuous variational amplitudes of the integration variables @xmath26 and @xmath27 , @xmath28 is an appropriate intercluster antisymmetrizer introduced to guarantee the exact preservation of the pauli exclusion principle , and @xmath29^{(j^{\pi}t ) } \nonumber \\ & \times \frac{\delta(x-\eta_{a_2-a_3})}{x\eta_{a_2-a_3 } } \frac{\delta(y-\eta_{a - a_{23}})}{y\eta_{a - a_{23}}}\ , , \label{eq:3bchannel } \end{aligned}\ ] ] are three - body cluster channels of total angular momentum @xmath30 , parity @xmath31 and isospin @xmath32 . here ,
@xmath33 , @xmath34 and @xmath35 denote the microscopic ( antisymmetric ) wave functions of the three nuclear fragments , which are labelled by the spin - parity , isospin and energy quantum numbers @xmath36 , @xmath37 , and @xmath38 , respectively , with @xmath39 .
additional quantum numbers characterizing the basis states ( [ eq:3bchannel ] ) are the spins @xmath40 and @xmath41 , the orbital angular momenta @xmath42 , @xmath43 and @xmath44 , and the isospin @xmath45 . in our notation , all these quantum numbers are grouped under the cumulative index @xmath46 . besides the translationally invariant coordinates ( see e.g. ref .
@xcite sec .
ii.c ) used to describe the internal dynamics of clusters 1 , 2 and 3 , respectively , in eq .
( [ eq:3bchannel ] ) we have introduced the jacobi coordinates @xmath47 and @xmath48 where @xmath49 is the relative vector proportional to the displacement between the center of mass ( c.m . ) of the first cluster and that of the residual two fragments , and @xmath50 is the relative coordinate proportional to the distance between the centers of mass of cluster 2 and 3 ( see figure [ figcoor ] ) . here
, @xmath51 denotes the position vector of the @xmath52-th nucleon .
( proportional to the vector between the c.m . of the first cluster and that of the residual two fragments ) and
@xmath48 ( proportional to the vector between the c.m . of clusters 2 and 3 ) .
in the figure , a case with three clusters of four , two and one nucleons are shown , however the formalism is completely general and can be used to describe any three cluster configuration.,width=226 ] using the expansion ( [ eq : trialwf ] ) for the wave function and projecting the microscopic @xmath8-body schrdinger equation onto the basis states @xmath53 , the many - body problem can be mapped onto the system of coupled - channel integral - differential equations @xmath54
g_{\nu}^{j^\pi t}(x , y ) = 0\nonumber\end{aligned}\ ] ] for the unknown variational amplitudes @xmath25 . here
, @xmath55 is the total energy of the system in the c.m .
frame and @xmath56 are integration kernels given respectively by the hamiltonian and overlap ( or norm ) matrix elements over the antisymmetrized basis states of eq .
( [ eq:3bchannel ] ) .
finally , @xmath57 is the intrinsic @xmath8-body hamiltonian .
denoting with @xmath58 the sum of the pairwise average coulomb interactions among the three clusters in channel @xmath59 of charge numbers @xmath60 , @xmath61 and @xmath62 , this can be separated into relative - motion and clusters intrinsic hamiltonians according to @xmath63 with @xmath64 the relative kinetic energy operator for the three - body system and @xmath65 the inter - cluster potential given by @xmath66 & \phantom{{\mathcal v}_{\rm rel } = } + { \mathcal v}^{3n}_{(a - a_{23},a_2,a_3 ) } - \bar v_c\ , .
\label{eq : interaction}\end{aligned}\ ] ] here , @xmath67 encompasses the portion of inter - cluster interactions due to the three - nucleon force , which , in general , is part of a realistic hamiltonian , and @xmath68 is the ( nuclear plus point - coulomb ) interaction between nucleons @xmath52 and @xmath69 . in the present paper
we will consider only the nucleon - nucleon ( @xmath3 ) component of the inter - cluster interaction and disregard , for the time being , the term @xmath67 .
the inclusion of the three - nucleon force into the formalism , although computationally much more involved , is straightforward and will be the matter of future investigations . in the remainder of the paper
, we will also omit the average coulomb potential @xmath58 , which is null for neutral systems such as the @xmath5he+@xmath1+@xmath1 investigated here .
the treatment of charged system is nevertheless possible and can be implemented along the same lines of ref .
@xcite . owing to the presence of the norm kernel ,
the three - cluster equations ( [ eq:3beq1 ] ) contain energy - dependent coupling terms . in alternative
, one can introduce the orthogonalized hamiltonian kernel @xmath70 of eq .
( [ eq : orthogh ] ) and solve the more familiar system of multi - channel schrdinger equations @xmath71
\chi_{\nu}^{j^\pi t}(x , y ) = 0\ , . \nonumber\end{aligned}\ ] ] the amplitudes @xmath72 of eq .
( [ eq : trialwf ] ) can then be recovered from the schrdinger wave functions @xmath73 through eq .
( [ eq : orthogg ] ) .
more details on the orthogonalization procedure can be found in appendix [ app : ortog ] .
the three - cluster schrdinger equations ( [ eq:3beq2 ] ) can be conveniently solved within the hh basis .
this basis is broadly used @xcite to treat few - body problems as its elements are eigenfunctions of the angular part of the kinetic operator written in hyperspherical coordinates @xcite .
the first step is to move to hyperspherical coordinates , i.e. @xmath74 where @xmath75 and @xmath76 are hyperradii and @xmath77 and @xmath78 hyperangles . in these coordinates ,
the relative kinetic energy operator for the three - cluster system can be written as , @xmath79 where the notation @xmath80 represents the hyperangle @xmath77 and the four angles @xmath81 and @xmath82 ( the direction angles of the jacobi coordinates @xmath83 and @xmath84 , respectively ) , @xmath85 is the grand - angular kinetic operator and @xmath86 is the mass of the nucleon .
as anticipated , the elements of the hh basis are the eigenfunctions of @xmath85 @xmath87 with eigenvalues @xmath88 . here
, @xmath89 is the hypermomentum quantum number defined as @xmath90 with @xmath91 , and the complete set of functions @xmath92 is given by @xmath93 where @xmath94 are jacobi polynomials , and @xmath95 normalization constants . as shown in appendix [ app : hh ]
, the hh functions ( [ hhbasis ] ) form a natural basis for the description of the three - cluster wave function of eq .
( [ eq : trialwf ] ) and for the solution of the three - cluster dynamical equations . indeed , by @xmath96 using the expansion @xmath97 for the relative motion wave functions , where @xmath98 are hyperradial functions analogous to those of eq .
( [ eq : chihh ] ) , and @xmath99 projecting from the left on the basis states @xmath100 , eq . ( [ eq:3beq2 ] ) can be written as a set of non - local integral - differential equations in the hyperradial coordinates : @xmath101 here , the orthogonalized hamiltonian kernel in the hyperradial variables @xmath76 and @xmath102 is given by @xmath103 the solution of eq .
( [ rgmrho ] ) for the case in which the three clusters form a bound state can be conveniently achieved within the @xmath19-matrix method as discussed in the next section .
we calculate the relative motion wave function by solving eq .
( [ rgmrho ] ) with the calculable r - matrix method @xcite .
in particular , we use a lagrange mesh which simplifies the problem as shown in many previous works for the two - cluster case @xcite and has been generalized to the three - cluster problem in ref .
@xcite . within this method ,
the configuration space is divided into two regions by assuming that the coulomb interaction ( if present ) is the only interaction experienced by the clusters beyond a finite separation @xmath104 . in the external region ( @xmath105 ) , where the schrdinger equation can be solved exactly , the hyperradial wave function is approximated by its known asymptotic form for large @xmath76 . for bound states of neutral systems ( as the one investigated in this paper ) , such asymptotic solution is given by : @xmath106 where @xmath107 is a modified bessel function of the second kind , @xmath108 is the wave number , and @xmath109 is a constant . in the internal region ( @xmath110 ) , where also the mutual nuclear interaction among the clusters is present , the wave function is written as a variational expansion on a lagrange basis of @xmath111 functions @xmath112 ( see appendix [ lagrange ] for definition ) @xmath113 where @xmath114 are the coefficients of the expansion .
the radial wave functions are then obtained by solving in the internal region the following set of bloch - schrdinger equations @xmath115 supplemented by the continuity condition @xmath116 . here , we have used the asymptotic expression of eq .
( [ external ] ) in the right - hand and the expansion of eq .
( [ internal ] ) in the left - hand side of the equation , respectively .
further , the elements of the bloch operator ( @xmath117 being arbitrary constants ) @xcite @xmath118 have the dual function of restoring the hermiticity of the hamiltonian in the internal region and enforcing the continuity of the derivative of the wave function at @xmath104 @xcite . owing to the dirac s delta in the bloch operator , the system of non - local equations ( [ bloch_eq ] ) is equivalent to that of eq .
( [ rgmrho ] ) in the internal region .
projecting eq .
( [ bloch_eq ] ) over a basis element @xmath119 and choosing the logarithmic derivative evaluated in @xmath120 @xmath121 for the constants appearing in the definition of the bloch operator ( [ bloch ] ) , the system ( [ bloch_eq ] ) reduces to @xmath122\beta_{k'\nu ' i'}=0\ , , \label{eigen}\ ] ] where the elements of the matrix @xmath123 are given by the integrals over the internal region @xmath124 the choice of lagrange functions as square - integrable basis states for the expansion of the wave function in the internal region ( [ internal ] ) greatly simplifies the evaluation of these integrals . indeed , within the gauss quadrature approximation , the lagrange functions are orthogonal to each other ( see appendix [ lagrange ] ) , the matrix elements of nonlocal potentials
are proportional to the values of the nonlocal potentials at the mesh points , and the analytical expression for the matrix elements of the kinetic energy operator is straightforward to obtain .
note that the matrix @xmath123 depends on the energy , owing to the choice ( [ constants ] ) for the boundary conditions in the bloch operator , which are functions of the wave number @xmath125 [ see eq .
( [ external ] ) ] . in practice ,
the solution of eq .
( [ eq_c ] ) is obtained recursively .
one can start from @xmath126 and iterate the solution of the eigenvalue equation ( [ eigen ] ) until the convergence in @xmath55 is reached , which typically occurs in a few iterations .
the coefficients of the expansion ( [ internal ] ) , @xmath114 , are then obtained from the corresponding eigenvector and the relative motion wave functions can be constructed using eqs .
( [ expansion ] ) and ( [ internal ] ) . the norm and hamiltonian integration kernels presented in sec .
[ subsec : microscopic ] are calculated within the ncsm / rgm approach as follows . first , the clusters eigenstates appearing in eq .
( [ eq:3bchannel ] ) are obtained by diagonalizing the @xmath127 , @xmath128 and @xmath129 intrinsic hamiltonians within the model spaces spanned by the @xmath130- , @xmath21- and @xmath22-nucleon ncsm bases , respectively .
these are complete sets of many - body ho basis states , the size of which is defined by the maximum number @xmath131 of ho quanta above the lowest configuration shared by the nucleons .
the same ho frequency @xmath132 is used for all three clusters , and the model - space size @xmath131 is identical ( differs by one ) for states of the same ( opposite ) parity .
second , for those components that are localized , the matrix elements of the translational invariant operators @xmath133 entering the expression of the integration kernels are evaluated within an ho model space using the expansion @xmath134
\ell_x & z & j_{23}\\[2 mm ] \end{array } \right\ } \left\ { \begin{array}{ccc } s & \ell_x & z\\[2 mm ] \ell_y & j & l\\[2 mm ] \end{array } \right\}\nonumber\\[2 mm ] & \times r_{n_x\ell_x}(x)r_{n_y\ell_y}(y ) \ , |\phi^{j^\pi t}_{\gamma n_x n_y } \rangle\,,\label{eq : transf}\end{aligned}\ ] ] where @xmath135 , @xmath136 etc . , and @xmath137 are the ho channel states defined by @xmath138^{(j^{\pi}t ) } \nonumber\\[2 mm ] & \times r_{n_x\ell_x}(\eta_{a_2-a_3})r_{n_y\ell_y}(\eta_{a - a_{23}})\ , , \label{eq:3bchannelho}\end{aligned}\ ] ] and labeled by the channel index @xmath139 @xmath140@xmath141 . besides the representation of the dirac s @xmath142 functions of eq .
( [ eq:3bchannel ] ) in terms of ho radial wave functions @xmath143 and @xmath144 , the transformation of eq .
( [ eq : transf ] ) reflects a different coupling scheme of the ho channels ( [ eq:3bchannelho ] ) with respect to the original basis , with @xmath145 the total ( orbital plus spin ) angular momentum quantum number of the system formed by the second and third clusters and @xmath146 the new channel spin . while the configuration of eq .
( [ eq:3bchannel ] ) is dictated by the use of the hh as basis for the solution of the three - cluster problem ( see sec .
[ subsec : hh ] and appendix [ app : hh ] ) , the binary - cluster - like coupling scheme of eq .
( [ eq:3bchannelho ] ) is more convenient for the derivation of the kernels in the ho basis , as it will become clear in a moment .
the frequency @xmath132 and the model - space size ( @xmath147 for even / odd parity states ) used to expand the relative motion are the same as those adopted for the calculation of the clusters eigenstates . finally ,
although the integration kernels are translational invariant quantities , it is computationally convenient to work within a slater determinant ( sd ) channel basis @xmath148 defined as in eq .
( [ eq:3bchannelho ] ) but with @xmath149 in place of the relative vector @xmath150 and the eigenstates of the heaviest cluster obtained in the sd basis , i.e. @xmath151 where @xmath152 and @xmath149 are respectively the coordinates proportional to the c.m .
of the first and last two clusters @xmath153 indeed , the translational invariant matrix elements can be extracted from those calculated in the sd basis , which contain the spurious motion of the c.m . , by inverting the following linear transformation : @xmath154 & \qquad = \sum_{n^{r\prime}_y \ell^{r\prime}_y , n^r_y\ell^r_y , j_r } \left\langle\phi^{j_r^{\pi_r } t}_{\gamma^\prime_r n_x^\prime n_y^{r\prime } } \right|\ , \hat { \mathcal o}_{\rm t.i . }
\left | \phi^{j_r^{\pi_r } t}_{\gamma_r n_x n^r_y } \right\rangle\nonumber\\[2 mm ] & \qquad\phantom{= } \times \sum_{\mathcal {
n l } } \hat\ell_y\hat\ell^\prime_y\hat j^2_r \ , ( -1)^{z+\ell_y - z^\prime-\ell^\prime_y}\nonumber\\ & \qquad\phantom{= } \times \left\ { \begin{array}{ccc } z & \ell^r_y & j_r\\[2 mm ] { \mathcal l } & j & \ell_y\\[2 mm ] \end{array } \right\ } \left\ { \begin{array}{ccc } z^\prime & \ell^{r\prime}_y & j_r\\[2 mm ] { \mathcal l } & j & \ell^\prime_y\\[2 mm ] \end{array } \right\}\nonumber\\[2 mm ] & \qquad\phantom{= } \times \langle n^r_y\ell^r_y \,{\mathcal { n l}}\ , \ell_y|00\,n_y\ell_y\,\ell_y\rangle_{\frac{a_{23}}{a - a_{23 } } } \nonumber\\[2 mm ] & \qquad\phantom{= } \times \langle n^{r\prime}_y\ell^{r\prime}_y \,{\mathcal { n l}}\ , \ell^\prime_y|00\,n^\prime_y\ell^\prime_y\,\ell^\prime_y\rangle_{\frac{a^\prime_{23}}{a - a^\prime_{23 } } } \ , .
\label{eq : sdtransf } \end{aligned}\ ] ] here , @xmath155 denotes a channel index identical to @xmath156 except for the replacement of the quantum number @xmath43 with @xmath157 ( the same applies for the primed indexes ) and @xmath158 is any scalar and parity - conserving translational invariant operator .
further , the transformation ( [ eq : sdtransf ] ) is diagonal in all quantum numbers but @xmath159 , and @xmath160 .
although formally not strictly necessary , with the new angular momentum coupling scheme of eq .
( [ eq:3bchannelho ] ) , the present conversion from sd to translational invariant matrix elements represents a straightforward generalization of the analogous binary - cluster transformation discussed in sec.ii.c.2 of ref .
@xcite , and the most advantageous choice from a computational point of view .
the theoretical framework presented so far is general and can in principle be applied to any three - cluster system . in the following ,
we discuss the derivation of the integration kernels for the more specialized instance of a target nucleus plus two single nucleons ( @xmath161 ) , such as the @xmath5he+@xmath1+@xmath1 system investigated here .
specifically , we will consider the case of identical @xmath162 mass partitions in both the initial and final states . in this case
, the second and third clusters are point - like nucleons with quantum numbers @xmath163 , and the inter - cluster antisymmetrizer is simply given by the product of the antisymmetrization operators for a @xmath164 mass partition and that of a two - body system @xmath165 & = \sqrt{\tfrac{2}{(a-1)a}}\left [ 1\!-\!\sum_{i=1}^{a-2}(\hat p_{ia-1}\!+\!\hat p_{ia})+\!\!\sum_{i <
j=1}^{a-2}\hat p_{ia-1}\hat p_{ja}\right]\nonumber\\ & \phantom{=}\times \tfrac{1}{\sqrt2}\left(1-\hat p_{a-1\,a}\right).\nonumber\end{aligned}\ ] ] although other factorizations of this operator are of course possible , with the present choice the antisymmetrization of nucleons @xmath166 and @xmath8 is trivial and can be included in the definition of the channel basis , i.e. @xmath167 the integration kernels for the @xmath162 mass partition are then obtained by evaluating the matrix elements of the operators @xmath168 and @xmath169 @xmath170 on the basis ( [ eq : antisym - basis ] ) .
for the norm kernel of eq .
( [ eq : nkernel ] ) , this yields the following sum of a direct and an exchange term @xmath171 & \;\;= \left ( 1-(-1)^{\ell_x+s_{23}+t_{23}}\right)\delta_{\nu^\prime\nu}\frac{\delta(x^\prime - x)}{x^\prime x}\frac{\delta(y^\prime - y)}{y^\prime y } \nonumber\\[2 mm ] & \;\;\phantom{= } + \ , { \mathcal n}^{\rm ex}_{\nu^\prime\nu}(x^\prime , y^\prime , x , y ) \,.\label{eq : nam211 } \end{aligned}\ ] ] here , the direct term arising from the identical permutation in the antisymmetrization operator is calculated in the full space , whereas the non - local exchange term is evaluated within the ho model space . as explained in the previous section , this is achieved by using the expansion ( [ eq : transf ] ) , with the translational invariant matrix elements on the ho channel basis of eq .
( [ eq:3bchannelho ] ) ( antisymmetrized for the exchange of nucleons @xmath166 and @xmath8 ) @xmath172 & \phantom{= } + \frac{(a-2)(a-3)}{2 } \left\langle\tilde\phi^{j^\pi t}_{\gamma^\prime n_x^\prime n_y^\prime } \right| \hat p_{a-2\,a } \hat p_{a-3\,a-1}\left | \tilde\phi^{j^ \pi t}_{\gamma n_x n_y } \right\rangle\!. \nonumber\end{aligned}\ ] ] obtained from the corresponding sd ones by inverting eq .
( [ eq : sdtransf ] ) . at the same time , the calculation of the matrix elements over the sd channels @xmath173 of eq .
( [ eq:3bchannelho ] ) is achieved by first performing a transformation to a fully single - particle basis , i.e. @xmath174 & \phantom{= } \times \left\ { \begin{array}{ccc } i_1 & j_{23 } & z\\[2 mm ] \ell_y & j & i\\[2 mm ] \end{array } \right\ } \left\ { \begin{array}{ccc } \ell_y & l & \ell_x\\[2 mm ] s_{23 } & j_{23 } & i\\[2 mm ] \end{array } \right\}\nonumber\\ & \phantom{= } \times \left\ { \begin{array}{ccc } \ell_a & \ell_b & l \\[2 mm ] \tfrac12 & \tfrac12 & s_{23 } \\[2 mm ] j_a & j_b & i \end{array } \right\ } \ , \left| \phi^{j^\pi t}_{\kappa_{ab}}\right\rangle_{\rm sd } \,.\end{aligned}\ ] ] here , @xmath120 and @xmath175 stand for the collections of ho single - particle quantum numbers @xmath176 and @xmath177 , respectively , @xmath178 indicates an ho bracket for two particles of equal mass , and @xmath179 @xmath180 is the index labeling the new sd channel basis @xmath181^{(j^\pi t)}. \label{sd - basis - ab}\end{aligned}\ ] ] we note that , except for a difference in the notation used for the total isospin of nucleons @xmath182 and @xmath8 , this basis is identical to that introduced for the treatment of binary - cluster channels with a di - nucleon projectile in eq . ( 18 ) of ref .
@xcite , where the interested reader can also find the algebraic expressions of the matrix elements @xmath183 and @xmath184 in eq .
( 19 ) and ( 20 ) , respectively .
the calculation of the hamiltonian kernel of eq .
( [ eq : hkernel ] ) is achieved along the same lines . in this case
, the kernel can be divided into a term proportional to the norm kernel discussed above , plus a term which resembles the expression of the potential kernel for binary - cluster channels with a di - nucleon projectile ( see ref .
@xcite , sec .
ii.b ) : @xmath185 & \qquad= \left [ \hat t_{\rm rel}(x^\prime , y^\prime ) + \hat v(x^\prime ) + e^{i^\prime_1t^\prime_1}_{\alpha^\prime}\right]\,{\mathcal n}^{j^\pi t}_{\nu^\prime\nu}(x^\prime , y^\prime , x , y ) \nonumber\\[2 mm ] & \qquad\phantom{= } + \left\langle\tilde\phi^{j^\pi t}_{\nu^\prime x^\prime y^\prime } \right| { \mathcal v}^{(a-2,2)}_{\rm rel}\hat{\mathcal a}^2_{(a-2,2)}\left | \tilde\phi^{j^\pi t}_{\nu x y } \right\rangle,\nonumber \end{aligned}\ ] ] with an analogous expression for the matrix elements of the hermitian conjugate operator @xmath186 . here , @xmath187 is the potential between nucleons @xmath8 and @xmath166 , @xmath188 is the energy of the @xmath189-nucleon eigenstate in the final channel , and @xmath190 is the sum of pairwise interactions corresponding to the first term in the right - hand side of eq .
( [ eq : interaction ] ) . as for the exchange operators of the norm ,
the matrix elements of @xmath191 are calculated within the ho model space .
this involves the evaluation on the sd channel basis of eq .
( [ sd - basis - ab ] ) of five potential terms : @xmath192 @xmath193 , @xmath194 @xmath195 , @xmath196 @xmath197 , @xmath198 @xmath199 , and @xmath200 @xmath201 .
algebraic expressions for these matrix elements can be found in eqs .
( a1-a4 ) and ( 24 ) of ref .
@xcite .
different from the deuteron - nucleus formalism of ref .
@xcite , where this interaction is already taken into account in the calculation of the ( bound ) projectile eigenstate , here the hamiltonian kernel contains the additional contribution coming from the action of the operator @xmath187 on the norm kernel . in the absence of coulomb interaction between the last two nucleons ( which , if present , can be treated separately as explained in sec .
[ subsec : microscopic ] ) , this term is localized in the variables @xmath202 and can be calculated as @xmath203 & \qquad= \hat s \hat s^\prime \hat l \hat l^\prime \ , ( -1)^{\ell_x+s+l-\ell^\prime_x - s^\prime - l^\prime}\nonumber\\ & \qquad\phantom{= } \times \sum_{j_{23 } } \hat j_{23}^2 \left\ { \begin{array}{cccc } - & i_1 & s^\prime & s_{23 } \\[2 mm ] \ell_y & - & l^\prime & \ell^\prime_x \\[2 mm ] \ell_x & s_{23 } & - & j_{23 } \\[2 mm ] l & s & j & - \end{array } \right\ } \nonumber\\[2 mm ] & \qquad \phantom{=}\times \sum_{l_x}\sum_{n^\prime_x n_x}r_{n^\prime_x\ell^\prime_x}(x^\prime ) r_{n_x l_x}(x)\nonumber\\[2 mm ] & \qquad\phantom{= } \times \langle n^\prime_x \ell^\prime_x s_{23 } j_{23 } t_{23 } | v | n_x l_x s_{23 } j_{23 } t_{23}\rangle\nonumber\\[2 mm ] & \qquad\phantom{= } \times \left(1-(-1)^{\ell_x+s_{23}+t_{23}}\right)\delta_{\tilde\gamma^\prime \gamma}\frac{\delta(y^\prime - y)}{y^\prime y } \nonumber\\[2 mm ] & \qquad\phantom{=}+ \hat v(x^\prime)\ , { \mathcal
n } ^{{\rm ex}}_{\nu^\prime\nu}(x^\prime , y^\prime , x , y ) \nonumber\,.\end{aligned}\ ] ] here , the expression between curly brackets represents a 12-j symbol of the second kind ( see appendix [ app:12j ] ) , @xmath204 are two - body matrix elements of the nuclear interaction on the translational - invariant ho basis , and @xmath205 is an index associated with the ho channel states of eq .
( [ eq:3bchannelho ] ) and identical to @xmath206 except for the replacement of the quantum number @xmath207 with @xmath208 . in the present work , the dirac s delta function in the @xmath27 variables of eq .
( [ eq : vx ] ) is approximated by an extended - size expansion in ho radial wave functions that goes well beyond the adopted ho model space ( @xmath209 ) .
the influence of such an approximation on the calculated binding energy of @xmath18he is small and will be discussed in sec .
[ sec : applications ] . finally , with the exception of the terms proportional to the exchange part of the norm kernel , the action of the relative kinetic energy operator @xmath210 and that of the eigenvalues @xmath188 are both calculated in the full space .
it is well known that @xmath18he is the lightest borromean nucleus @xcite , formed by an @xmath5he core and two halo neutrons . owing to its small mass number and the fact that its constituents do not form bound subsystems , it is an ideal first candidate to be studied within the present approach .
the ground state of this nucleus has been the subject of many investigations .
some of them are based on a three - body non - microscopic cluster formalism , representing it as a system of three inert particles @xcite .
this type of three - body methods can lead to the appropriate asymptotic behavior of the wave function , but do not allow for the exact treatment of the pauli principle , which plays a fundamental role for light nuclei , and make use of effective nucleon - nucleus potentials .
there have also been _ ab initio _ six - body calculations focused on the ground state of @xmath18he @xcite .
these are based on realistic hamiltonians and fulfill the pauli principle exactly .
however , not taking explicitly into account the three - body cluster configuration of this nucleus leads to an incorrect description of the asymptotic properties of the system . in between these two approaches
are microscopic calculations which take into account both the three - cluster configuration of the system and the internal structure of its constituents , giving better description of the asymptotic behavior of the nuclear wave function while also preserving the pauli principle @xcite .
nevertheless , so far these type of calculations have been based on semi - realistic interactions , often without spin - orbit force , and on a simplified description of the internal structure of the clusters . in this work we present for the first time an _ ab initio _ calculation
which not only uses realistic interactions but also takes into account the three - body configuration of this nucleus .
in particular , we apply the formalism presented in sec .
[ sec : formalism ] to study the ground state of @xmath18he within a @xmath5he(g.s.)+@xmath1+@xmath1 cluster basis . as stated in sec .
[ subsec : rgmkernels ] , the @xmath5he wave function is calculated within the ncsm formalism . in the present calculations , we describe the @xmath5he core only by its g.s .
wave function , ignoring its excited states .
the inclusion of excited states leads to big technical difficulties within the ncsm / rgm formalism because it increases notably the number of channels and the calculation becomes unbearable for current computational resources .
however , this is a minor setback which , once the method is established as presented in this work , can be overcome by coupling the present three - cluster wave functions with ncsm eigenstates of the six - body system within the ncsmc @xcite approach . for the time
being , we estimate core polarization effects , by comparing the computed @xmath211 @xmath5he(g.s.)+@xmath1+@xmath1 g.s . energy with that obtained from a ncsm diagonalization of the six - body hamiltonian . in both calculations
, we use the same two - body interaction , namely the srg evolved @xcite potential obtained from the chiral n@xmath11lo @xmath3 interaction @xcite with the evolution parameter @xmath212=1.5 @xmath213 .
he within the ncsm formalism .
, width=359 ] we performed ncsm calculations for @xmath5he that generate eigenstates needed as input for the subsequent three - body cluster ncsm / rgm investigations of @xmath18he .
further , we also calculated the g.s .
energy of @xmath18he within the ncsm in order to make a comparison with the @xmath5he+@xmath1+@xmath1 ncsm / rgm results .
the computed @xmath18he g.s .
energies for a range of ho frequencies and various basis sizes ( @xmath131 values ) are presented in fig .
as stated earlier , we are employing a soft srg - evolved chiral n@xmath11lo @xmath3 interaction with evolution parameter @xmath212=1.5 @xmath213 .
we intentionally adopt such a soft interaction , for which our calculations reach convergence in the ho basis expansion already at the computationally accessible @xmath214 .
we can subsequently concentrate on the exploration of the validity of other approximations in the three - cluster ncsm / rgm formalism .
we note that the same @xmath3 interaction was used in previous binary - cluster ncsm / rgm calculations of the @xmath215-@xmath5he scattering @xcite and the @xmath215-@xmath11h fusion @xcite .
the variational ncsm calculations converge rapidly and can be easily extrapolated to @xmath216 using , e.g. , an exponential function of the type @xmath217 . as shown in fig .
[ ncsm ] , at @xmath218 the dependence of the @xmath18he g.s .
energy on the ho frequency is flat in the range of @xmath219 mev .
the variational minimum is close to @xmath220 mev that we then choose for the calculation of the @xmath5he eigenstates used as input for the @xmath5he+@xmath1+@xmath1 ncsm / rgm investigations of the @xmath18he nucleus . .computed ncsm @xmath5he ( second column ) and ncsm / rgm @xmath18he [ as @xmath5he(g.s.)+@xmath1+@xmath1 ] ( third column ) g.s .
energies in mev as a function of the ho model space size @xmath131 .
the last column shows the @xmath18he g.s .
energies in mev for model space sizes of @xmath222 .
the last two rows show the extrapolated values for the calculations with their uncertainties , and the experimental values . [
cols="^,^,^,^",options="header " , ] [ energy ] in table [ energy ] , the energy of @xmath5he and @xmath18he ground states calculated with the ncsm formalism are shown ( in mev ) in the second and last column , respectively .
the extrapolated values of the ncsm calculation with their uncertainties and the experimental values @xcite are also given .
it can be observed that , with the present soft srg interaction , the g.s .
energy of @xmath5he is close to the experimental value , while the @xmath18he ground state is overbound by about 0.5 mev .
we performed calculations for the @xmath5he+@xmath1+@xmath1 three - cluster system by using the ncsm / rgm formalism described in section [ sec : formalism ] . in this first application , we neglect core polarization effects and limit the description of @xmath5he to just the @xmath223 g.s .
eigenstate in the ncsm / rgm coupled channel equations .
this is the only limitation of the model space introduced .
none of the remaining quantum numbers contained in the cumulative index @xmath59 have any restriction other than those dictated by the model space size itself . in particular , we calculated hamiltonian and norm kernels of eq .
( [ eq : hkernel ] ) and ( [ eq : nkernel ] ) for all possible @xmath160 channels up to @xmath224 , the maximum value of the total angular momentum for our largest model space of @xmath225 , in which both the @xmath42 and @xmath43 orbital angular momentum quantum numbers can vary from @xmath226 to @xmath227 .
although , for the present paper we were exclusively interested in the @xmath228 g.s . of @xmath18he
, this was a necessary step to correctly extract the translational invariant matrix elements from our sd calculations through eq .
( [ eq : sdtransf ] ) .
illustrative examples of the norm kernel can be found in sec .
[ subsubsec : norm ] . the g.s .
energy of @xmath18he is then obtained by solving the @xmath5he+@xmath1+@xmath1 non - local hyperradial equations ( [ rgmrho ] ) for the @xmath229 channel with bound - state boundary conditions , as explained in sec .
[ subsec:3eq ] .
the dimension of the hh model space used for this part of the calculation is related to the maximum value of the hypermomentum @xmath230 .
overall , the number of \{@xmath59,@xmath89 } channels is very large ( around 200 for the @xmath229 state alone in our largest model space with @xmath225 and @xmath231 ) .
the dependence of our results for the @xmath18he g.s .
with respect to both the ho and hh expansions are discussed in sec .
[ subsubsec : he6 ] .
particularly interesting are the elements of the exchange part of the norm kernel , @xmath232 , defined in eq .
( [ eq : nam211 ] ) , which give a measure of the influence of the pauli exclusion principle . in fig .
[ norm ] , we present just a few of the most relevant examples .
in addition , for visual purposes , we set the value of the primed coordinates @xmath233 and @xmath234 to @xmath235 fm . in the present calculation , where the first cluster is given by the g.s . of @xmath5he ,
i.e. a @xmath236 state , and the second and third clusters are single nucleons , the various channels can be simply labeled by the spin and orbital angular momentum quantum numbers @xmath237 , and @xmath43 .
as one would expect , for the largely @xmath17-shell @xmath5he core the antisymmetrization makes its largest contribution in the @xmath238 channel of the @xmath228 state , of which we show the diagonal exchange " norm in the top left panel of fig . [ norm ] .
large negative values of the exchange part of the norm kernel generally correspond to the presence of pauli forbidden components , in this case the @xmath239 component due to the @xmath17-wave relative motion in both @xmath26 and @xmath27 coordinates .
the norm is positive and much smaller in channels where the antisymmetrization plays a minor role , such as the @xmath240 displayed in the bottom left panel of the figure .
the fairly symmetric appearance of these norm kernels is due to the equal value of the two orbital angular momenta @xmath42 and @xmath43 .
however , the kernels are in general asymmetric in @xmath241 ( @xmath242 ) as it is particularly evident in the @xmath243 case ( top right panel ) . in this component , which appears in the @xmath244 , and @xmath245 states
, one can observe once again the repulsion due to the pauli principle in the @xmath27 coordinate , while the @xmath246-wave motion forces the norm kernel to be null for @xmath247 ( @xmath248 ) . finally , in the bottom right panel of fig .
[ norm ] we present the @xmath249 diagonal element of the @xmath250 , and @xmath251 exchange norm . in this channel , where one could naively expect a positive norm
, we find a non negligible negative contribution of the antisymmetrization , which suggests the presence of pauli - forbidden components .
these examples show that a correct treatment of the antisymmetrization is not only important to describe the g.s . of the @xmath18he nucleus , but also plays a role in important excited states such as the @xmath252 or @xmath253 resonances . he g.s . energy at @xmath254 as a function of the maximum value of the hypermomentum @xmath230 used in the hh expansion . for these calculations we used a matching radius of @xmath255 fm ,
@xmath256 lagrange mesh points , and an extended ho model space of @xmath257.,width=321 ] he g.s .
energy at @xmath254 as a function of the size of the extended ho model space @xmath258 used for the calculation of the interaction kernel of eq .
[ eq : vx ] . for these calculations we used a hyper momentum of @xmath259 , a matching radius of @xmath255 fm , and @xmath256 ( @xmath260 ) lagrange mesh points for @xmath261 ( @xmath262).,width=321 ] to calculate the g.s . of @xmath18he , we first orthogonalize the ncsm / rgm equations ( [ eq:3beq1 ] ) as explained in sec .
[ subsec : orthog ] and appendix [ app : ortog ] . during such procedure
the eigenvalues and eigenvectors of the norm kernel in the ho model space are calculated . for this @xmath263 state
, we observed the appearance of pauli forbidden norm eigenstates recognizable for their very small eigenvalues and negligible overlap with the physical g.s .
spurious states can appear and admix with the low - lying physical eigenstates of the system if such pauli forbidden norm eigenstates are not eliminated . for the present calculation ,
we have removed all norm eigenstates with eigenvalues smaller than @xmath264 .
the unprecedented large number of @xmath265 channels ( @xmath266 ) is the likely responsible for the occurrence of such unphysical eigenstates of the norm , which had been never observed in our previous binary - cluster ncsm / rgm calculations .
we then expanded the orthogonalized ncsm / rgm equation ( [ eq:3beq2 ] ) in hh functions and solved the non - local hyperradial equations ( [ rgmrho ] ) for the @xmath5he+@xmath1+@xmath1 relative motion imposing bound - state boundary conditions , by using the @xmath19-matrix method on a lagrange mesh of sec .
[ subsec:3eq ] .
we found a single bound state in the @xmath263 channel and proceeded to study the behavior of our results at fixed @xmath131 with respect to the remaining parameters of the calculation .
given the large scale of this computation , we performed this study at @xmath254 .
the rate of convergence of the bound state with respect to the size of the adopted hh model space can be judged by examining fig .
[ fig : kmax ] , where we present a study of the calculated g.s .
energy as a function of the maximum value of the hypermomentum @xmath230 .
the results start to stabilize around @xmath267 and are fully converged already at @xmath259 . at a given @xmath131 ,
the calculation is variational in @xmath230 .
then , we studied the stability of the g.s .
energy with respect to the selection of the matching radius @xmath120 , and we found that it was good as long as we were choosing values larger than 20 fm .
the number @xmath111 of mesh points required for a good convergence of the lagrange expansion depends on the value of the matching radius . for @xmath268 fm
about 30 mesh points are enough , while a larger number is needed if the matching radius is increased .
wave functions @xmath269 for @xmath131=6,8,10 , and 12 .
, width=321 ] the choice of the @xmath111 value also depends somewhat on the size of the extended ho model space @xmath258 used to represent the dirac s delta function in the @xmath27 ( @xmath234 ) coordinate ( proportional to the distance between the centers of mass of the @xmath5he and the two neutrons ) while calculating the interaction kernel of eq .
( [ eq : vx ] ) .
larger @xmath258 values correspond to a larger @xmath27-range for this potential kernel , which is localized only in the @xmath26 ( @xmath233 ) coordinate . about @xmath270 ( @xmath271 )
mesh points are sufficient to reach convergence up to @xmath272 ( @xmath273 ) . the behavior of the g.s .
energy as a function of @xmath258 is presented in fig .
[ fig : next ] . as it can be observed , an extended ho basis size of at least @xmath257 is needed to accommodate the long range of this interaction kernel . disregarding this effect by computing eq .
( [ eq : vx ] ) within the adopted ho model space ( i.e. with @xmath274 ) leads to about @xmath275 kev underbinding in the @xmath18he g.s .
energy . finally , a stable result for the integrations in the hyperangles @xmath78 and @xmath276 of eq .
( [ eq : hrho ] ) , which we perform numerically using a chebyshev - gauss quadrature ( for chebyshev polynomials of the second kind ) , was obtained with 20 mesh points .
based on this analysis , we adopted a matching radius of @xmath277 fm with @xmath278 mesh points , a hyper momentum @xmath231 , and an extended ho model space of @xmath279 for our larger @xmath131 calculations ( including the largest with @xmath280 presented in the following .
he+@xmath1+@xmath1 relative motion wave function for the @xmath211 ground state .
the quantum numbers corresponding to this component are @xmath238 . here
@xmath281 and @xmath282 are respectively the distance between the two neutrons and the distance between the c.m . of @xmath5he and that of the two neutrons .
, width=302 ] in fig .
[ f_rho ] the main components of the radial part of the relative motion wave function @xmath283 of the @xmath228 g.s.of @xmath18he are shown for different values of the ho basis size @xmath131 used for the expansions of the @xmath5he wave function and localized elements of the integration kernels . in the present calculation ,
each component is uniquely identified by the quantum numbers shown in the figure .
as it can be seen , convergence is almost reached at @xmath131=10 and a @xmath284 calculation , which is currently out of computational reach , is not expected to substantially change the present results .
this is confirmed also by the @xmath131 dependence of the related g.s .
energy , presented in the third column of table [ energy ] .
contrary to the ncsm , which gives rise to a gaussian asymptotic behavior of the wave function owing to the use of expansions over six - body ho basis states , in the ncsm / rgm the @xmath5he(g.s.)+@xmath1+@xmath1 wave functions present the asymptotic behavior of eq .
( [ external ] ) , which is included by construction when using the @xmath19-matrix method . as it can be seen in fig .
[ f_rho ] , the tails can extend up to about @xmath285 fm
. this feature will be of great importance when studying @xmath18he excited states , using scattering asymptotic conditions in the solution of the three - cluster equations . .
, width=302 ] information about the three - cluster structure of the @xmath18he g.s . can be obtained by studying the probability distribution arising from the main component of the @xmath5he+@xmath1+@xmath1 relative motion wave function , presented as a surface plot in fig .
[ probability ] , and as contour plot in fig .
[ contour ] .
this component , characterized by the quantum numbers @xmath238 , presents the well known @xcite two - peak shape distribution .
one peak corresponds to a di - neutron " configuration in which the neutrons are close together ( about 2 fm apart from each other ) while the @xmath5he core is separated from their c.m . at a distance of
about @xmath286 fm .
whereas the second peak , corresponding to the cigar " configuration , represents an almost linear structure in which the two neutrons are far from each other ( about @xmath287 fm apart ) and the alpha particle lies almost in between them at @xmath288 fm from their center of mass . the position maxima of the two peaks can be more easily seen from the contour diagram of fig .
[ contour ] . , but for the second most important component .
the quantum numbers corresponding to this component are @xmath240.,width=302 ] the second most important contribution to the wave function comes from the component with quantum numbers @xmath240 , and the probability distribution arising from it is shown in fig .
[ probability2 ] . from the amplitude of the plot ,
it can be concluded that this component contributes very little to the complete wave function , and does not significantly change the characteristic two peak picture of fig .
[ probability ] . the obtained @xmath18he g.s .
energy ( in units of mev ) for different sizes of the ncsm / rgm model space @xmath131 are presented in the third column of table [ energy ] .
the results of ncsm calculations of the @xmath5he and @xmath18he systems are shown in the second and last column , respectively . at @xmath218 , the ncsm / rgm calculation
is basically converged within its uncertainty of about @xmath270 kev , quoted in the last row .
the effect of the exclusion of the spurious eigenstates of the norm , discussed at the beginning of this section , is included in this uncertainty . unlike the ncsm case
, the present ncsm / rgm calculations are not variational in the ho model space size as at each @xmath131 value the three - cluster basis contains a different @xmath5he eigenstate . it can be observed that in the ncsm / rgm the @xmath18he ground state is under bound . in particular , the energy is about 1 mev larger than the one obtained within the ncsm .
this difference is due to excitations of the @xmath5he core , which , for technical reasons , are included only in the ncsm calculation at present . in this sense ,
the difference between the two calculations provides a measure of core polarization effects in @xmath18he .
in this work , we have extended the ncsm / rgm method @xcite to the treatment of three - cluster dynamics .
this new feature permits to study a new range of systems that present three - body configurations .
in particular , in this work we show that it can be used to study the structure of two - neutron halo nuclei such as @xmath18he , and , contrary to other _
ab initio _ methods such as the ncsm , to obtain the appropriate asymptotic behavior of its wave functions .
moreover , the present formalism combined with the appropriate scattering boundary conditions gives access to the _ ab initio _ study of resonant states of two - neutron halo nuclei ( such as the excited states of @xmath18he ) as well as to scattering problems involving channels with three fragments .
three - cluster ncsm / rgm @xmath5he+@xmath1+@xmath1 scattering calculations with the aim to study the @xmath18he low - lying resonances are currently under way and will be reported in a subsequent paper . for the present study of @xmath18he within the @xmath5he+@xmath1+@xmath1 cluster basis
, we used only the ground state wave function to described the @xmath5he core .
this leads to an underbinding of the @xmath18he ground state due to the missing treatment of core polarization .
the difference with respect to the energy obtained from a diagonalization of the hamiltonian in a ncsm six - body model space indicates that , with the present soft @xmath3 interaction , core - polarization effects amount to less than @xmath289 of the binding .
the inclusion of excited states of @xmath5he would significantly increase the number of channels in the calculation , making it computationally unbearable .
core polarization effects can be more efficiently taken into account by coupling the present three - cluster wave functions with six - body ncsm eigenstates within the ncsmc framework .
while the results of this approach will be presented in a forthcoming publication , the main difficulty of such a calculation resides in obtaining the three - cluster ncsm / rgm wave functions and was addressed in the present work .
we thank g. hupin many useful discussions .
computing support for this work came from the llnl institutional computing grand challenge program and from an incite award on the titan supercomputer of the oak ridge leadership computing facility ( olcf ) at ornl .
prepared in part by llnl under contract de - ac52 - 07na27344 .
support from the u.s .
doe / sc / np ( work proposal no .
scw1158 ) and nserc grant no . 401945 - 2011 is acknowledged .
triumf receives funding via a contribution through the canadian national research council .
the appearance of the norm kernel @xmath290 in eq .
( [ eq:3beq1 ] ) reflects the fact that the many - body wave function @xmath291 is expanded in terms of a nonorthogonal basis .
therefore , eq . ( [ eq:3beq1 ] )
does not represent a system of multichannel schrdinger equations , and @xmath292 do not represent schrdinger wave functions .
however , one can solve the equivalent set of orthogonalized equations introduced in eq .
( [ eq:3beq2 ] ) where @xmath293 is the orthogonalized hamiltonian kernel and the schrdinger wave functions @xmath73 are the new unknowns of the problem , related to @xmath294 through @xmath295 here , @xmath296 and @xmath297 represent the square root and the inverse - square root of the norm kernel , respectively , which are obtained as follows .
first , we add and subtract from the norm kernel the identity in the ho model space , @xmath298\nonumber \\[2 mm ] & \qquad\qquad+ \lambda_{\nu^\prime\nu}^{j^{\pi}t}(x^\prime , y^\prime , x , y)\,.\end{aligned}\ ] ] the norm kernel within the truncated model space , @xmath299 , is obtained by using the expansion ( [ eq : transf ] ) with the matrix elements on the ho jacobi - channel states of eq .
( [ eq:3bchannelho ] ) given by @xmath300 and @xmath301 as defined in eq .
( [ eq : nexam211 ] ) .
then , the square and the inverse - square root of the full - space norm are obtained by @xmath96 finding the eigenvalues @xmath302 and eigenvectors @xmath303 of the matrix @xmath304 of eq .
( [ eq : honorm ] ) ; @xmath99 calculating @xmath305 and the corresponding integration kernels through the expansion ( [ eq : transf ] ) , and finally @xmath306 replacing the model - space norm @xmath307 in eq .
( [ eq : norm_mod ] ) with @xmath308 , that is , @xmath309\nonumber \\[1 mm ] & \qquad\qquad+ \lambda^{\pm 1/2}_{\nu'\nu}(x',y',x , y)\ , .
\label{eq : invsqn}\end{aligned}\ ] ] for the inverse operation to be permissible in eq . ( [ eq : lambda_invers ] ) one has to exclude the subspace of ( fully ) pauli - forbidden states for which @xmath302=0 .
using dirac delta s properties and completeness relation of the set of @xmath310 functions , we have @xmath311 then the three - cluster channel states of eq .
( [ eq:3bchannel ] ) can be written as @xmath312 where @xmath313 are the channel states in the hh basis @xmath314^{(j^{\pi}t ) } \frac{\delta(\rho - \rho_\eta)}{\rho^{5/2}\,\rho_\eta^{5/2}}\ , , \label{eq:3bchannelhh } \end{aligned}\ ] ] with @xmath315 and @xmath316 the hh basis elements defined in eq .
( [ hhbasis ] ) . at the same time , inserting this expansion in eq .
( [ eq : trialwf ] ) , changing from @xmath26 and @xmath27 to the hh coordinates @xmath76 and @xmath78 and integrating over the hyperangle @xmath78 , one can demonstrate that the many - body wave function is also given by @xmath317 where the hyperradial functions @xmath318 are obtained from the projection of the variational amplitudes @xmath319 over the functions @xmath320 : @xmath321
we use a lagrange basis which is a set of @xmath111 functions @xmath322 ( see @xcite and references therein ) , given by @xmath323 where @xmath324 are legendre polynomials , and @xmath325 satisfy @xmath326 the lagrange mesh associated with this basis consists in @xmath111 points @xmath327 on the interval @xmath328 $ ] and satisfy the lagrange conditions @xmath329 where the coefficients @xmath330 are the weights corresponding to a gauss - legendre quadrature approximation for the @xmath331 $ ] interval , i.e. @xmath332 using the lagrange conditions of eq .
( [ lagcond ] ) , it is straightforward to see that within the gauss approximation the lagrange functions are orthogonal , i.e. @xmath333
the 12-j symbol of the second kind @xcite is defined by @xmath334 b_1 & - & b_3 & b_4 \\[2 mm ] c_1 & c_2 & - & c_4 \\[2 mm ] d_1 & d_2 & d_3 & - \end{array } \right\ } & = ( -1)^{b_3-a_4-d_1+c_2}\sum_x ( 2x+1 ) \left\ { \begin{array}{ccc } a_3 & b_4 & x \\[2 mm ] b_1 & d_3 & b_3 \\[2 mm ] \end{array } \right\ } \left\ { \begin{array}{ccc } a_3 & b_4 & x \\[2 mm ] c_4 & a_2 & a_4 \\[2 mm ] \end{array } \right\ } \left\ { \begin{array}{ccc } b_1 & d_3 & x \\[2 mm ] d_2 & c_1 & d_1 \\[2 mm ] \end{array } \right\ } \left\ { \begin{array}{ccc } c_4 & a_2 & x \\[2 mm ] d_2 & c_1 & c_2 \\[2 mm ] \end{array } \right\ } \nonumber \\[4 mm ] & = ( -1)^{b_3-a_4-d_1+c_2}\sum_x ( 2x+1 ) \left\ { \begin{array}{ccc } a_3 & b_3 & d_3 \\[2 mm ] a_4 & b_4 & c_4 \\[2 mm ] a_2 & b_1 & x \\[2 mm ] \end{array } \right\ } \left\ { \begin{array}{ccc } d_2 & d_1 & d_3 \\[2 mm ] c_2 & c_1 & c_4 \\[2 mm ] a_2 & b_1 & x \\[2 mm ] \end{array } \right\}\ , .
\nonumber \\\end{aligned}\ ] ] 53ifxundefined [ 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/physrevc.64.044001 [ * * , ( ) ] link:\doibase 10.1103/physrevc.84.054010 [ * * , ( ) ] @noop *
* , ( ) link:\doibase 10.1103/physrevc.83.041001 [ * * , ( ) ] link:\doibase 10.1103/physrevc.86.044330 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.physletb.2007.07.072 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevc.86.021602 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.022505 [ * * , ( ) ] link:\doibase 10.1103/physrevc.87.034326 [ * * , ( ) ] @noop _ _ ( , , ) @noop * * , ( ) \doibase http://dx.doi.org/10.1016/0375-9474(82)90322-0 [ * * , ( ) ] @noop _ _ ( , , ) \doibase http://dx.doi.org/10.1016/s0370-1573(97)00049-5 [ * * , ( ) ] link:\doibase 10.1103/physrevc.77.044002 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevc.62.054311 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1143/ptp.96.575 [ * * , ( ) ] link:\doibase 10.1103/physrevc.63.034606 [ * * , ( ) ] link:\doibase 10.1103/physrevc.63.034607 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.nuclphysa.2004.05.013 [ * * , ( ) ] link:\doibase 10.1103/physrevc.80.044310 [ * * , ( ) ] link:\doibase 10.1103/physrevc.85.034318 [ * * , ( ) ] link:\doibase 10.1103/physrevc.67.044309 [ * * , ( ) ] link:\doibase 10.1016/j.nuclphysa.2005.11.010 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1088/0954 - 3899/35/6/063101 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreva.65.052710 [ * * , ( ) ] http://stacks.iop.org/0953-4075/31/i=15/a=015 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1088/0954 - 3899/22/2/004 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.55.2676 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1016/0370 - 1573(93)90141-y [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1140/epja / i2009 - 10815 - 5 [ * * , ( ) ] link:\doibase 10.1016/j.nuclphysa.2010.06.012 [ * * , ( ) ] link:\doibase 10.1103/physrevc.68.034305 [ * * , ( ) ] link:\doibase 10.1146/annurev.nucl.51.101701.132506 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.052504 [ * * , ( ) ] @noop _ _ ( , ) | we introduce a fully antisymmetrized treatment of three - cluster dynamics within the _ ab initio _ framework of the no - core shell model / resonating - group method ( ncsm / rgm ) .
energy - independent non - local interactions among the three nuclear fragments are obtained from realistic nucleon - nucleon interactions and consistent _ ab initio _ many - body wave functions of the clusters .
the three - cluster schrdinger equation is solved with bound - state boundary conditions by means of the hyperspherical - harmonic method on a lagrange mesh .
we discuss the formalism in detail and give algebraic expressions for systems of two single nucleons plus a nucleus .
using a soft similarity - renormalization - group evolved chiral nucleon - nucleon potential , we apply the method to an @xmath0+@xmath1+@xmath1 description of @xmath2 and compare the results to experiment and to a six - body diagonalization of the hamiltonian performed within the harmonic - oscillator expansions of the ncsm .
differences between the two calculations provide a measure of core ( @xmath0 ) polarization effects . |
once upon a time , several blind men from a hindustan village touched different parts of an elephant s body and , judging from their perceptions , quarreled about what the whole creature might look like . in a similar way
, we explore color confinement by isolating different aspects of the phenomenon , and build our limited experience into various model schemes .
the moral from the ancient parable teaches us that it is vitally important to unify all different views before we can appreciate the whole beauty of the beast .
the aspirations of this contribution are more modest .
i will concentrate on common points of two seemingly unrelated pictures of color confinement .
the former assumes that confinement arises due to the condensation of a particular type of topological excitations , so called center vortices ( for a review see @xcite ) , in the qcd vacuum ; while the latter is the gribov - horizon scenario in coulomb gauge , which has been advocated by gribov @xcite and zwanziger @xcite .
i will first briefly introduce the idea of the gribov - horizon scenario , then formulate a simple criterion of confinement for static color charges in coulomb gauge , discuss how the fulfillment of this criterion depends on presence / absence of center vortices , and finally present results for the coulomb energy of a pair of static charges .
i will present a subset of our numerical results , details , as well as some analytic insights on the connections between center vortices and the gribov horizon , can be found in recent publications @xcite .
in coulomb gauge , the hamiltonian of qcd has the following form @xcite : @xmath0 a prominent role in this expression is played by the faddeev popov operator @xmath1 which enters both the interaction kernel @xmath2 and the jacobian factor @xmath3 @xmath4^{a , b}_{x , y } , \qquad { \cal{j}}\equiv\det\left[m(a)\right].\ ] ] it is well - known since the seminal paper of gribov @xcite ,
that coulomb - gauge fixing in a non - abelian theory is a difficult problem .
the transversality condition @xmath5 does not fix the gauge completely .
gribov @xcite suggested to restrict integration over configurations to the so - called gribov region ( gr ) , defined as the subspace of transverse gauge fields for which the faddeev popov operator is positive , and which are therefore local minima of the functional @xmath6=\int dx\left[^g \mathbf{a}^a(x)\right]^2,\qquad \mbox{where}\qquad{}^ga_i = g^{-1 } a_i g + g^{-1}\partial_i g.\ ] ] the boundary of this region is called the gribov horizon .
however , there exist gauge orbits which intersect the gribov region more than once ; the next step @xcite then is to restrict fields to the fundamental modular region ( fmr ) , the set of absolute minima of the functional ( [ functional ] ) .
both the gr and the fmr are bounded in every direction and convex .
the essence of the gribov - horizon confinement scenario can be phrased in a simple way : the dimension of gauge - field configuration space is huge , so it is reasonable to expect that most configurations are located close to its boundary ( the horizon ; in a similar way , the volume measure @xmath7 of a @xmath8-dimensional sphere is peaked at its surface ) .
the interaction kernel @xmath2 , which determines the interaction energy of static color sources , contains the inverse of the faddeev
popov operator , which is strictly zero on the horizon and near - zero close to the horizon .
a high density of configurations near the horizon can thus lead to a strong enhancement of the coulomb interaction energy , and hopefully cause color confinement .
let us consider a single static color charge , which can be written in coulomb gauge as @xmath9=\psi^\alpha(x)\psi_0[a],\ ] ] where @xmath10 is the color index for a point charge in color group representation @xmath11 , and @xmath12 is the coulomb - gauge ground state .
the excitation energy of this state , above the ground state , is given by @xmath13 and is proportional to @xmath14 the quantity @xmath15 is the color coulomb self - energy of unscreened color charge and is expected to be both ultraviolet and infrared divergent in a confining theory .
the uv divergence can be regulated by a lattice cut - off , however , the quantity must still be divergent at infinite volume , even after lattice regularization , due to ir effects . on a lattice , one can express ( [ e ] ) through eigenstates of the faddeev
popov operator @xmath16 simply as ( assuming that @xmath17 is invertible , i.e. excluding its zero modes ) @xmath18 and @xmath19 is the lattice 3-volume . in su(2 )
lattice gauge theory , the link variables can be expressed as @xmath20 and the lattice faddeev
popov operator @xmath21 - \delta_{x , y-\hat{k } } b_k(x ) - \delta_{y , x-\hat{k } } b_k(y ) \right\}\\ & - & \epsilon^{abc } \sum_{k } \left\ { \delta_{x , y-\hat{k } } a^c_k(x ) - \delta_{y , x-\hat{k } } a^c_k(y ) \right\}\end{aligned}\ ] ] ( @xmath22 denote lattice sites at fixed time ) is a @xmath23 sparse matrix with @xmath24 linearly independent eigenstates .
if we denote @xmath25 the number of eigenvalues in the range between @xmath26 and @xmath27 , we can introduce , on a large lattice , the density of states @xmath28 . then , as @xmath29 , @xmath30 a ( necessary )
condition for confinement can now be formulated : _ the excitation energy @xmath15 of a static , unscreened color charge is divergent if , at infinite volume ,
it is interesting to check whether the above condition is fulfilled in various ensembles of lattice configurations .
first , at zero - th order in the gauge coupling , the faddeev popov operator is simply a lattice laplacian and its eigenstates are just plane waves .
one can easily verify that , in this case , @xmath32 the excitation energy is ir finite and the confinement criterion is not met .
one needs some mechanism of enhancement of @xmath33 and @xmath34 in the region of small @xmath26 values .
i will argue that such an enhancement exists in full lattice configurations , and is provided by center vortices .
center vortices are identified by fixing to an adjoint gauge , and then projecting link variables to the @xmath35 subgroup of su(@xmath36 ) @xcite .
the excitations of the projected theory are known as p - vortices . in the direct maximal center gauge ( dmcg ) in su(2 ) @xcite one fixes to the maximum of @xmath37 \right\vert^2,\ ] ] and center projects by @xmath38.\ ] ] a lot of evidence has been accumulated that center vortices alone reproduce much of confinement physics , for a review see @xcite . we have determined the density of states @xmath33 and the mean value of the lattice laplacian @xmath34 in three ensembles of lattice configurations : 1 .
full configurations , @xmath39 , 2 .
`` vortex - only '' configurations , @xmath40\rbrace$ ] , and 3 .
`` vortex - removed '' configurations , @xmath41 . [ cols="^,<,^ " , ] the effect of vortex removal is illustrated in lower lines of data points of figure [ fig6 ] .
removing center vortices also removes the confining property of the color coulomb potential . since the potential , in coulomb gauge , is sensitive to the interaction kernel @xmath2 in the hamiltonian ( [ hamiltonian ] ) , and the kernel in turn to the density of near - zero eigenvalues of the faddeev
popov operator , this result again confirms the observation of the previous section : removal of center vortices alters the density quite drastically .
we also note that the confining property of the color coulomb potential is tied to the unbroken realization of a remnant global gauge symmetry in coulomb gauge .
this connection was studied in detail in ref .
it was demonstrated there on a few examples ( deconfined phase in pure gauge theory , pseudo - confinement phase of the gauge fundamental - higgs theory ) that confinement in the color coulomb potential is not identical to confinement in the static interquark potential , and center symmetry breaking , spontaneous or explicit , does not necessarily imply remnant symmetry breaking .
results of our numerical simulations suggest an appealing picture : the low - lying eigenvalues of the faddeev popov operator in coulomb gauge tend towards zero as the lattice volume increases .
the density of the eigenvalues goes as a small power of @xmath26 , and this , together with a similar behavior of the average laplacian , @xmath34 , assures the infrared divergence of the energy of an unscreened color charge . also , due to the enhancement of near - zero modes of the faddeev
popov operator , the coulomb energy of a pair of color charges rises linearly with their separation .
both facts support the ideas of the gribov - horizon confinement scenario .
the constant density of low - lying eigenvalues can be attributed to the vortex component of gauge - field configurations . a thermalized configuration in a pure gauge theory factors into a confining piece ( the vortex - only part ) , and a piece which closely resembles the lattice of a gauge higgs theory in the higgs phase ( the vortex - removed configuration ) .
this establishes a firm connection between the center - vortex picture and the gribov - horizon scenario .
this connection is exemplified also by the fact that vortex removal removes the color coulomb string tension of the color coulomb potential .
it is also consistent with recent investigations of gattnar et al .
@xcite in landau gauge . in this talk , i covered results of our numerical investigations .
related analytical developments were omitted : ( i ) thin center vortices lie on the gribov horizon ; ( ii ) the gribov horizon is a convex manifold in lattice configuration space and thin center vortices are conical singularities on that manifold ; ( iii ) the coulomb gauge is an attractive fixed point of a more general gauge condition , interpolating between the coulomb and landau gauges .
interested readers are invited to find them in our recent publications @xcite .
our research is supported in part by the u.s .
department of energy under grant no .
de - fg03 - 92er40711 ( j.g . ) , the slovak grant agency for science , grant no .
2/3106/2003 ( .o . ) , and the national science foundation , grant no .
phy-0099393 ( d.z . ) .
is grateful to the organizers of the conference for invitation to present this talk and for creating a stimulating , yet relaxed atmosphere . | we review numerical evidence on connections between the center - vortex and gribov - horizon confinement scenarios .
address = physics and astronomy dept . ,
san francisco state university , san francisco , ca 94117 , usa , [email protected] address = institute of physics , slovak academy of sciences , sk845 11 bratislava , slovakia , [email protected] address = physics department , new york university , new york , ny 10003 , usa , [email protected] _ so oft in theologic wars , + the disputants , i ween , + rail on in utter ignorance + of what each other mean , + and prate about an elephant + not one of them has seen ! _ + _ * john godfrey saxe * _ ( 1816 - 1887 ) @xcite |
ngc 1399 , the central galaxy of the fornax group , is one of the brightest elliptical galaxies in x - rays and has been the subject of numerous x - ray studies . although previous _ rosat _ and
_ asca _ studies demonstrated that the hot gas within @xmath10 kpc is not isothermal ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the data could not distinguish between single - phase and multiphase models .
moreover , we showed that an `` fe bias '' for systems like ngc 1399 occurs if an isothermal gas is assumed when in fact the spectrum consists of multiple temperature components with values near 1 kev ( e.g. , * ? ? ?
* ; * ? ? ?
we argued that this bias is primarily responsible for the very sub - solar fe abundances found in the central regions of the ( x - ray ) brightest ellipticals and groups .
the combined spatial and spectral resolution of the _ xmm _ ccds allows for unprecedented mapping of the temperatures and elemental abundances of the hot gas in galaxies , groups , and clusters .
these capabilities complement the high spatial resolution ( @xmath11 ) of _ chandra_which has already provided interesting constraints on the emission from discrete sources near the center of ngc 1399 @xcite . the higher energy resolution ,
sensitivity , and larger field - of - view of the _ xmm _ epic ccds are better suited for constraining the spatial and spectral properties of the diffuse hot gas out to radii well past the optical extent of the galaxy ( i.e. , out to @xmath7 kpc assuming a distance of 21 mpc using the results of @xcite for @xmath12 km s@xmath13 mpc@xmath13 ) .
we present initial results for the temperature and metal abundances of the hot gas obtained from a new _ xmm _ observation of ngc 1399 .
detailed discussions of the data reduction , spectral fitting , and results for the gravitating mass will appear in @xcite and @xcite .
ngc 1399 was observed with the epic pn and mos ccd cameras for approximately 25 ks and 30 ks respectively during ao-1 as part of the _ xmm _ guest observer program .
we generated calibrated events lists for the data using the standard sas v5.3.0 software .
our data preparation closely follows standard procedures and will be described fully in a follow - up paper @xcite .
we extracted spectra in concentric circular annuli located at the x - ray centroid of each detector .
the widths of the annuli were the same for each detector and were defined such that the width of each annulus contained approximately 6000 background - subtracted counts in the 0.3 - 5 kev band in the mos1 , and the minimum width was set to @xmath14 for psf considerations .
obvious point sources were masked out before the extraction . in each annulus
we performed a simultaneous fit to the mos1 , mos2 , and pn spectral data over 0.3 - 5 kev .
our baseline model ( 1 t ) consists of a single thermal plasma component using the apec code modified by foreground galactic absorption ( @xmath15 @xmath16 ) using the phabs model in xspec v11.1.0v .
the free parameters are all associated with the plasma component : temperature ( @xmath17 ) , normalization , and fe , o , mg , and si abundances all other elements tied to fe in their solar ratios except for s which is tied to si in their solar ratio .
( the normalizations for the mos1 and mos2 are tied together while that of the pn is varied separately because of residual calibration and background differences . )
we obtained results for the 1 t model fitted directly to the data projected on the sky ( i.e. , 2d model ) and also fitted to the deprojected data ( i.e. , 3d model ) ( e.g. , * ? ? ? * ) .
-0.4 cm ccccccc + & @xmath18 & @xmath19 & & + annulus & ( arcmin ) & ( arcmin ) & @xmath20 & dof & @xmath20 & dof + + 1 & 0.0 & 1.0 & 1061.3 & 487 & 558.2 & 484 + 2 & 1.0 & 2.2 & 700.1 & 520 & 625.5 & 517 + 3 & 2.2 & 3.4 & 691.8 & 593 & 657.8 & 590 + 4 & 3.4 & 4.7 & 824.4 & 645 & & + 5 & 4.7 & 6.5 & 925.3 & 776 & & + 6 & 6.5 & 8.2 & 821.1 & 826 & & + 7 & 8.2 & 10.1 & 1042.8 & 907 & & + + the temperature profile for the 1 t ( 2d ) model is displayed in figure [ fig ] .
the temperature rises from @xmath21 kev at the center to @xmath22 kev at large radii consistent with previous 2d _ rosat_determinations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this model is a good fit in the outer annuli but the fit degrades at progressively smaller radii ( see table [ tab ] ) .
the worst fit is in the central annulus and , as shown in figure [ fig.spec ] , yields fit residuals near 1 kev that are fully characteristic of those obtained when trying to force a single - temperature model to fit a multi - temperature spectrum ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the deviations above 2 kev also suggest the presence of another ( higher ) temperature component .
multitemperature spectra in the inner regions , especially the central bin , can arise simply from the projection of hotter gas at larger radius . in this case fitting the 1 t model to the deprojected spectra should substantially improve the fits in the central regions .
however , we find that this 1 t ( 3d ) model does not improve the fits much except at the center , where noticeable though relatively modest improvement ( @xmath23 ) is achieved . we have examined a variety of multitemperature models ( 2d and 3d ) and have found that simple two - temperature models ( 2 t ) provide the best fits to the central annuli ( table [ tab ] ) .
it is also found that the @xmath20 values are very similar and the derived parameters consistent within their @xmath24 errors for both 2d and 3d versions of the 2 t models . and
since the parameters of the 3d models are not as well constrained and require regularization of the temperatures and abundances ( cf .
* ) , in the interest of brevity we will emphasize in this letter the results for the 2d models .
detailed results for the 3d versions of 1 t , 2 t , and other models will appear in @xcite . with respect to the 1 t ( 2d ) model , not only does the 2 t ( 2d ) model improve the fit substantially within the central bin ( @xmath25 ) , but it also significantly improves the fits in annuli # 2 ( @xmath26 ) and # 3 ( @xmath27 ) .
the dramatic improvement in the central bin is seen in figure [ fig.spec ] where the large residuals in the fe l region near 1 kev have been mostly eliminated . in 3d
the improvement of 2 t over 1 t is @xmath28 in the central bin .
the other multitemperature models we examined only improved fits to the deprojected spectra in the central bins ; e.g. , a ( constant pressure ) cooling flow model consisting of ambient and cooling gas provides a modest improvement to the fit in the central bin ( @xmath29 ) over the 1 t ( 3d ) model . in figure [ fig ]
are shown the temperatures obtained for the 2 t ( 2d ) model .
the cool component is consistent with an isothermal radial profile while the hot component appears to rise slightly at the center . within the @xmath24 errors
this central rise for the hot component is not especially significant , but if real the slight excess over the isothermal extrapolation from the adjacent annuli probably represents the contribution from hard emission from discrete sources unresolved by _
xmm _ but resolved in the _ chandra _ image of ngc 1399 @xcite .
we mention that adding an additional bremsstrahlung component to account for the effect of these unresolved sources ( i.e. , 2t+brem ) in the central bin does improve the fit slightly and gives a reasonable best - fitting temperature , @xmath30 kev . adding this component to the cooling flow model instead gives @xmath31 kev indicating an attempt to account for both @xmath32 kev hot gas and @xmath33 kev unresolved sources .
_ we believe that in addition to the @xmath20 values the derived parameter values also argue in favor of the 2 t model . _
the temperature of the hot component , @xmath34 kev , is consistent with the virial temperature of a surrounding group of mass @xmath35 , whereas the temperature of the cool component , @xmath36 kev , is similar to the kinetic temperature of the stars ( @xmath37 kev ) with velocity dispersion , @xmath38 km s@xmath13 ( e.g. , * ? ? ?
the volume - averaged electron density of the cool gas ( ) declines with radius more rapidly than that of the hot component ( ): for the best - fitting models the ratio @xmath39 is 1.0 , 0.62 , and 0.27 respectively for shells # 1 , # 2 , and # 3 .
( here we have used the 3d model so that the quoted densities refer to the 3d spherical shells bounded by the same radii as the annuli listed in table [ tab ] . )
the cool component is only detected out to annulus # 3 ( i.e. , for @xmath4 kpc ) and therefore must contribute even less at larger radius .
the radius where the cool ( galaxy ) component gives way to the hot ( group ) component is similar to that inferred from the recent analysis of the _ rosat _ surface brightness profile of ngc 1399 by @xcite .
interior to shell # 3 we compute a gravitating mass of @xmath40 from each component consistent with previous x - ray ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and optical ( e.g. , * ? ? ?
* ) determinations .
the consistency of the masses determined from each phase , and the agreement with optical determinations , indicates that each phase is in approximate hydrostatic equilibrium tracing the same gravitational potential .
although hydrostatic equilibrium implies subsonic gas motions for each phase , the phases are not in pressure equilibrium .
the ratio of the gas pressure in the hot phase to that of the cool phase is @xmath41 in shell # 1 which increases to @xmath42 in shell # 3 .
the lack of pressure equilibrium implies that the low - pressure cool phase must be continuously replenished .
this requirement is consistent with the scenario where the cool phase continuously receives gas from stellar ejecta and loses gas by eventually mixing with the extended hot phase .
detailed physical models for the two phases ( including , e.g. , appropriate filling factors ) will be considered in a future study .
we take the solar abundances in xspec to be those given by the @xcite table except for fe where we use the correct `` meteoritic '' value , @xmath43 by number , as suggested by @xcite and @xcite .
this value is 1.44 times smaller than the photospheric value which is usually employed in x - ray spectral fitting .
consequently , we shall account for this factor when comparing our results to other studies .
the fe abundances ( ) obtained for the 1 t and 2 t models ( both 2d ) are displayed in figure [ fig ] .
( the values of obtained from 2d and 3d 2 t models are consistent within 1 - 2 @xmath44 errors . ) at the largest radii the abundance is sub - solar , @xmath45 .
it increases for smaller @xmath46 so that within @xmath47 kpc , @xmath48 , for the 1 t model similar to previous results for 1 t ( 2d ) models fitted to the _ rosat _ data ( e.g. , * ? ? ?
* ; * ? ? ?
compared to the 1 t model , the 2 t model gives that is 40%-50% larger in annuli # 2 and # 3 and twice as large in annulus # 1 .
( a cooling flow model , when fitted to the deprojected spectra , produces similarly larger values . ) this substantial increase in the value of along with the elimination of the characteristic @xmath32 kev residual pattern ( figure [ fig.spec ] ) is a manifestation of the `` fe bias '' we have discussed previously for ngc 1399 @xcite and related systems ( e.g. , * ? ? ?
* ; * ? ? ?
we note that the abundances in each temperature component are tied together in the fits since they can not be distinguished when fitted separately .
also , the values of deduced in the central bins for both 1 t and 2 t models differ slightly if intrinsic absorption is considered ( which is not clearly required by the data as seen in figure [ fig.spec ] ) , but the relative differences between the 1 t and 2 t models and the general finding of super - solar values are preserved . next to the broad feature of fe l lines near 1 kev , the most notable spectral lines in the epic spectra of ngc 1399 ( figure [ fig.spec ] ) are the k@xmath49 lines of si and s ; i.e. , ( 1.85 kev ) , ( 2.0 kev ) , and ( 2.45 kev ) . we find that is generally better constrained than , but when these abundances are fitted separately they inevitably yield values that are consistent within their 1 - 2 @xmath44 errors . to reduce the number of free parameters in our fits we tie and together in their solar ratio .
henceforth , we shall refer only to the si abundance .
the si abundance is plotted in figure [ fig ] for both 1 t and 2 t models .
for the 1 t model @xmath50 for @xmath51 kpc , rises to values consistent with solar for @xmath52 kpc , and is slightly larger than solar in the central bin . for the 1 t model @xmath53 at large radii , @xmath54 for @xmath52 kpc , and @xmath55 in the central bin .
the 2 t model gives slightly larger values , @xmath56 for @xmath52 kpc ( which are consistent with the 1 t values within the @xmath24 errors ) .
but in the central bin the 2 t value is @xmath57 larger ; i.e. , about a @xmath58 discrepancy . in contast to the 1 t case , for the 2 t model , @xmath8 for @xmath59 kpc .
( the overestimate of @xmath60 using the 1 t model is a `` si bias '' related to the fe bias we discussed previously for such systems @xcite . )
( unlike fe and si , the o and mg abundances are quite sensitive to intrinsic absorption
. we will discuss models with intrinsic absorption in @xcite . )
we have performed a spatially resolved spectral analysis of the _ xmm_epic data of ngc 1399 which suggests the presence of two dominant phases of hot gas within the central @xmath61 kpc .
this two - phase description is consistent with 2 t models we fitted previously to the _ asca _ data @xcite .
however , the _ asca _ data lacked the spatial resolution of _ xmm _ and could not distinguish between a radially varying single - phase medium , a two - phase medium , or a cooling flow within @xmath62 kpc .
these results for ngc 1399 are supported by previous analyses of its x - ray surface brightness profile which argues for separate galaxy and group components ( e.g. , * ? ? ?
* ; * ? ? ?
* ) similar to other centrally e - dominated groups @xcite . since the temperature structure of ngc 1399 obtained by _
asca _ is characteristic of other central e - dominated galaxy groups @xcite , the two - phase structure revealed by _ xmm _ for ngc 1399
likely applies to these other systems .
moreover , since the centers of several clusters show similar multitemperature structure with _ asca _
( e.g. , * ? ? ?
* ; * ? ? ?
* ) and similar multi - component surface brightness structure ( see * ? ? ?
* and references therein ) , we might expect two - phase structures in the centers of these systems .
a recent _ xmm _ analysis of m87 is consistent with this suggestion @xcite .
the two - phase model provides a physical explanation for the gas density and temperature profiles seen in `` cooling flows '' .
the hotter component is associated with the ambient group or cluster gas while the cooler component is associated with the stellar ejecta of the dominant central galaxy .
gasdynamical models of groups and clusters constructed by @xcite demonstrate that relatively cool ejecta from stellar mass loss combined with the hotter ambient gas can reproduce observed single - phase temperature profiles in groups and clusters .
but our results suggest that these two phases are not fully mixed .
understanding the details of the interaction of these phases may shed light on the mystery of cooling flows . using the results of @xcite , the rate of energy input by stellar ejecta is , @xmath63 , where @xmath64 s@xmath13 is the specific stellar mass - loss rate .
if within @xmath65 kpc we take @xmath66 for the stellar mass and @xmath38 km s@xmath13 @xcite , we infer the rate of energy input into the central @xmath33 kpc of ngc 1399 from stellar ejecta to be @xmath67 erg s@xmath13 . this crude estimate is only a factor of two less than the x - ray luminosity of the hot component ( and a factor of four less than the cool component ) .
if the energy input from stellar ejecta does not completely suppress mass drop - out , it may at least delay the episodes of agn feedback .
problems with such heating models are discussed by @xcite . the two - phase model ( as well as other multitemperature models )
removes the `` fe bias '' within the central regions ( @xmath4 kpc ) and gives @xmath5 .
these super - solar values exceed the stellar values in typical elliptical galaxies ( e.g. , * ? ? ?
* ) and therefore allow for significant enrichment of the hot gas from type ia supernovae ( snia ) ( cf .
for example , using the results of @xcite and our measured value of @xmath68 solar , we infer that snia have contributed @xmath9 of the iron mass within @xmath7 kpc in ngc 1399 .
this snia fraction is similar to that inferred for the sun and therefore suggests a stellar initial mass function similar to the milky way as advocated by renzini and others ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we close by noting that although correcting for the `` fe bias '' partially removes the `` iron discrepancy '' noted by @xcite , chemical models of elliptical galaxies without cooling flows predict central iron abundances even larger than we have measured for ngc 1399 ( e.g. , * ? ? ? | we present an initial analysis of a new _ xmm _ observation of ngc 1399 , the central elliptical galaxy of the fornax group .
spectral fitting of the spatially resolved spectral data of the epic mos and pn ccds reveals that a two - temperature model ( 2 t ) of the hot gas is favored over single - phase and cooling flow models within the central @xmath0 kpc .
the preference for the 2 t model applies whether or not the data are deprojected .
the cooler component has a temperature ( @xmath1 kev ) similar to the kinetic temperature of the stars while the hotter component has a temperature ( @xmath2 kev ) characteristic of the virial temperature of a @xmath3 halo . the two - phase model ( and other multitemperature models )
removes the `` fe bias '' within @xmath4 kpc and gives @xmath5 . at larger
radii the iron abundance decreases until @xmath6 for @xmath7 kpc .
the si abundance is super - solar ( 1.2 - 1.7 solar ) within the central regions while @xmath8 over the entire region studied .
these fe and si abundances imply that @xmath9 of the fe mass within @xmath7 kpc originates from type ia supernovae ( snia ) .
this snia fraction is similar to that inferred for the sun and therefore suggests a stellar initial mass function similar to the milky way . |
the _ eddington _ mission is a space telescope designed for two primary goals : asteroseismic studies and extrasolar planet finding .
both goals will be achieved through the acquisition of high - precision wide - field photometry , with a temporal stability that is possible only from space .
the basic design is an f/3 triple - reflecting telescope with a 1.2 m diameter aperture , 0.6m@xmath0 effective area , and a 3@xmath1 diameter field of view imaged by a 20-ccd mosaic camera .
the telescope will be launched into an orbit around the l2 earth - sun libration point , from which the entire sky is accessible for observations .
the first 2 years of the mission will be dedicated primarily to asteroseismic studies , with pointings lasting up to two months to a variety of targets .
the mission s second part will be dedicated to planet finding by the transit method , where the telescope will survey a single stellar field for 3 years . in total
, about 500,000 stars will be surveyed by _
eddington _ as potential hosts for planetary systems . in october 2000 ,
_ eddington _ was selected by esa as a reserve mission for the f2/f3 launch window ( 2009 - 2013 ) .
more details about the mission can be found in the _ eddington _ assessment study ( favata et al . , 2000 ) and in the _ eddington _ web - site ( http://astro.esa.int/sa-general/projects/eddington/ ) .
here we will give a short overview about its planet detection capabilities .
the evaluation of _ eddington _ s planet detection performance consisted of two parts : first , we determined the kinds of planets this mission can detect , in terms of the planet s size , orbital period , and magnitude and type of central star ( deeg et al .
second , the numbers and characteristics of the detected planets were estimated .
this employed monte - carlo simulations to fold _ eddington _
s ` detection space ' with models of the galactic stellar distribution , density and scale height ( robin & crz , 1986 ) .
with plausible exo - planet distributions , some 10,000 of planets of _ any _ size and orbital distance may be detected ( fig 1 , left ) . of the detected planets the vast majority are both large and in close orbits around their central stars .
the most extreme ones , the ` hot giants ' , will be detected not only by transits , but also by the reflected light of their central star ( even if they do not produce transits ) yielding a large and nearly complete sample of these planets . for smaller planets
there are fewer sufficiently bright host stars within the observational range , and they will be detected only by transits .
the habitable zone corresponds roughly to orbital distances at which planetary surface temperatures of 250 - 350k are expected . under the assumption that every main - sequence star has one planet in the habitable zone , a few hundred planets in that zone should be detected , among them a few tens of habitable planets with sizes of less then 2.5 @xmath2 , which should have solid surfaces ( fig 1 , right ) .
around larger planets within the habitable zone there exists the intriguing possibility that massive moons may be detected through the perturbations they cause on the arrival time of transit minima ( sarotetti & schneider , 1999 ) .
these moons may themselves be harboring sites for life .
as fig . 2 shows , _ eddington _ is the first mission that will survey the habitable zone , and deliver a broad overview of the abundance of habitable planets in the galaxy .
this is an important step beyond missions like corot and sim , which may detect a few special cases of habitable planets .
_ eddington _ will deliver a large well - defined statistical sample of planets with known planet radii , orbital distances , and host star surface temperature and type .
this is critical input for the design and target selection of future missions like darwin and tpf that aim for detailed characterizations of earth - like planets .
_ eddington _ moves us one step closer to resolving the age - old question : are there other worlds like ours ? | _ eddington _ is a space mission for extrasolar planet finding and for asteroseismic observations .
it has been selected by esa as an f2/f3 reserve mission with a potential implementation in 2008 - 13 . here
we describe _ eddington _ s capabilities to detect extrasolar planets , with an emphasis on the detection of habitable planets .
simulations covering the instrumental capabilities of _ eddington _ and the stellar distributions in potential target fields lead to predictions of about 10,000 planets of all sizes and temperatures , and a few tens of terrestrial planets that are potentially habitable . implications of _ eddington _ for future larger scale missions are briefly discussed . |
as discussed in the manuscript , when we are in the tonks - girardeau limit @xmath27 , the system may be described in terms of an effective xxz hamiltonian , @xmath36 , \label{eq : spinhsupp}\end{aligned}\ ] ] where the parameters are defined as in the main text . for even particle number @xmath14 ,
as is the case when the spin populations are balanced ( @xmath109 ) , one can show that this spin - chain hamiltonian satisfies @xmath110 , with the unitary operator @xmath111 .
thus , with the exception of parity , we expect the properties of the eigenstates to be invariant under @xmath112 . the parity operator @xmath113 , so that for even @xmath94 we have @xmath114=0 $ ] and the ground state always has the same parity . for odd @xmath94 , instead @xmath115 , and consequently the ground state changes parity at @xmath116 and @xmath68 .
on the other hand , the total energy is not invariant under the unitary transformation due to the presence of the constant @xmath117 in the strong - coupling hamiltonian , eq . .
the parameters @xmath45 which yield the values of the spin - exchange coefficients of the spin - chain hamiltonian were computed in refs .
@xcite . for completeness , we list these in table [ tab : eta ] up to @xmath118 . .distinct parameters of the spin - chain hamiltonian . the full set can be obtained using @xmath119 . for @xmath120
the coefficients are exact . for @xmath121
the coefficients may be obtained analytically @xcite , however the resulting expressions are cumbersome and are not repeated here . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] 0 pt in the manuscript , instead of using the numerically exact coefficients in table [ tab : eta ] we use an approximate model derived in ref .
@xcite , where @xmath122 .
note that @xmath123 only sets the absolute value of the energy and does not affect the eigenstates or the energy ratios for each @xmath14 .
we fix @xmath123 as following : at @xmath98 , corresponding to infinite interspecies interactions , we know the exact ground state which is fully phase separated .
thus , we may evaluate the ground state energy to find @xmath124.\end{aligned}\ ] ] requiring that we reproduce the exact ground state energy at @xmath98 thus results in the simple expression for @xmath123 : @xmath125 with @xmath45 those of table [ tab : eta ]
. this procedure yields @xmath126 , @xmath127 , @xmath128 , and @xmath129 . for larger @xmath14 the exact coefficients
are not known , and we instead extrapolate @xmath130 towards the thermodynamic limit @xmath63 . in this limit , the domain wall may be ignored , and the energy to leading non - trivial order is that of a lieb - liniger gas , eq . in the main text : @xmath131 this result was obtained in ref .
@xcite by applying the local density approximation to the uniform space bethe ansatz result .
hence , in the limit @xmath63 , the function @xmath123 is , to leading order , @xmath132 to go to finite @xmath14 , we choose the simplest possible extrapolation , requiring that the @xmath133 correction to eq
. produces the correct @xmath134 . the result for @xmath135 is @xmath136 which is shown in fig .
[ fig : etabar ] .
the high accuracy of the strong - coupling ansatz developed in ref .
@xcite and used here to determine the spin - exchange coefficients beyond the few - body limit may be confirmed by computing the overlap between the eigenstates @xmath137 of the approximate model and the eigenstates @xmath138 obtained in the exact solution of the problem .
this we have computed for balanced systems containing at most 8 particles , analogously to what was done in our earlier work @xcite . as may be seen in fig.[fig : overlaps ] ,
the wavefunction overlaps for both the ground and first excited state ( in the upper branch ) remain extremely close to 1 for all ratios of intra- and interspecies interaction strengths . the largest deviations ( of at most a few parts in @xmath139 )
are found near the boundaries of the fm region .
0 pt , 3 , and 4 particles .
, title="fig : " ]
the stability of the system in the presence of a small inter - conversion coupling may be analyzed in terms of the spin gap " , i.e. , the energy difference between the ground state of a system of @xmath14 particles , and the configuration in which one of the spins is flipped . in the fm phase ( @xmath140 ) , the ground state corresponds to a fully polarized gas of @xmath14 identical bosons , and we thus define the fm spin gap to be @xmath141 , where @xmath142 is the ground state energy of a gas composed of @xmath143 spin - up particles and @xmath144 spin - down ones . for all other values of @xmath29 , including those in the afm phase , the ground state corresponds to a balanced system with @xmath109 .
thus the spin gap in this region is instead defined as @xmath145 . at the classical ising points @xmath146 , we know the ground states exactly and thus we also know the spin gap , which at both points is @xmath147 .
this follows from the observation that in both cases the energy cost associated with flipping a spin is smallest at the trap edge .
we then apply perturbation theory to estimate the behaviour of the spin gap around these points . in the afm regime
, we have for @xmath148 , @xmath149\\ & \stackrel{n\to\infty}{\longrightarrow } & \frac{2n}{g}\bar\eta_{\rm ll}(n)(1-\phi^2 ) , \label{afmspingap2}\end{aligned}\ ] ] where in eq .
we have kept the leading order correction in @xmath14 .
we display the analytic result , eq .
, in fig .
[ fig : degeneracyandcorrelationsandspingaps](c ) to demonstrate the manner in which the afm spin gap approaches the thermodynamic limit .
we may perform a similar perturbative expansion around the fm ising point , but in this case we find that we can describe the whole fm region for large @xmath14 with the following analytic form : @xmath150 as shown in fig .
[ fig : degeneracyandcorrelationsandspingaps](c ) . | we consider two species of bosons in one dimension near the tonks - girardeau limit of infinite interactions . for the case of equal masses and equal intraspecies interactions ,
the system can be mapped to a @xmath0 xxz heisenberg spin chain , thus allowing one to access different magnetic phases . using a powerful ansatz developed for the two - component fermi system , we elucidate the evolution from few to many particles for the experimentally relevant case of an external harmonic confinement . in the few - body limit
, we already find clear evidence of both ferromagnetic and antiferromagnetic spin correlations as the ratio of intraspecies and interspecies interactions is varied .
furthermore , we observe the rapid emergence of symmetry - broken magnetic ground states as the particle number is increased .
we therefore demonstrate that systems containing only a few bosons are an ideal setting in which to realize the highly sought - after itinerant ferromagnetic phase .
quantum magnetism is ubiquitous in nature and plays a central role in important phenomena such as high - temperature superconductivity @xcite .
furthermore , it underpins the technological advances in data storage @xcite , and it promises a new generation of spintronic devices , where the spin of the electron rather than its charge is used to transfer information .
however , despite its ubiquity , magnetic phenomena are often difficult to characterise and treat theoretically ; for instance , itinerant ferromagnetism of delocalized fermions requires strong interactions and is thus not completely understood @xcite .
one can gain insight into magnetic phases by considering cleaner , more idealised versions of the phenomena .
in particular , an important question is whether or not ferromagnetism can exist _ without an underlying lattice _ @xcite .
this possibility was investigated experimentally in atomic fermi gases @xcite , but the system proved to be unstable towards fermion pairing rather than magnetism .
more recently , it has been proposed that itinerant ferromagnetism can be realised in a strongly interacting one - dimensional ( 1d ) fermi gas @xcite .
however , in this case , one can not access the ferromagnetic phase without explicitly breaking the spin symmetry with an external field @xcite , and this potentially complicates the observation of ferromagnetism . in this letter , we show that both itinerant ferromagnetism and antiferromagnetism can be investigated with a 1d two - component mixture of bosons . for the case of equal masses and equal intraspecies interactions ( @xmath1 ) ,
the bose - bose mixture may be regarded as a pseudo - spin @xmath2 system , and it can be mapped to a 1d spin chain in the limit of strong interactions @xcite . in particular , for infinite intraspecies interactions , the system is formally equivalent to a 1d fermi gas . however , in contrast to the fermi case , one can break the @xmath3 symmetry and access different magnetic phases by varying the ratio of intraspecies and interspecies interactions ( see fig .
[ fig : sketch ] ) .
moreover , the ferromagnetic state should be stable , unlike in higher dimensions @xcite .
0 pt , where we focus on the `` upper - branch '' ground state and ignore any two - body bound states .
( b ) the system may be mapped onto an xxz heisenberg spin model , eq . , with couplings @xmath4 and @xmath5 in the @xmath6-@xmath7 plane and @xmath8 direction , respectively .
equivalent points in the phase diagrams are marked by identical symbols , e.g. , @xmath9 corresponds to identical inter- and intraspecies interactions .
the thick ( red ) lines correspond to the xxx heisenberg model that is realized by the two - component fermi gas . in the thermodynamic limit , these lines bound regions of ising ferromagnetism ( fm ) and ising antiferromagnetism ( afm ) , separated by disordered spin liquid " phases . in the hatched region
all interactions are repulsive , and the spin - chain ground state is the actual ground state of the system .
, title="fig : " ] experimentally , 1d quantum gases have been successfully realized with strongly interacting bosons @xcite , two species of fermions @xcite , and , more recently , fermions with @xmath10 symmetry , where the number of spin components can be tuned from @xmath11 to 6 @xcite . all these 1d experiments
feature an underlying harmonic trapping potential , which has enabled the study of the evolution from few to many particles @xcite .
however , it is theoretically challenging to treat 1d interacting particles in a harmonic trap , since the problem can not be solved exactly in general @xcite .
only a small number of trapped particles can be treated exactly numerically @xcite .
recently , we developed a highly accurate ansatz for the strongly interacting 1d fermi gas in a harmonic potential that allowed us to analytically determine the wavefunction for a single spin-@xmath12 impurity with any number of majority @xmath13 fermions @xcite .
here , we demonstrate that this approach can be extended to provide an extremely accurate description of the bose mixture , providing a benchmark for other numerical techniques .
we use this method to address the question of how the ground state of the few - body system is connected to magnetic phases in the many - body limit . as the particle number @xmath14
is increased , we observe the rapid emergence of a doubly degenerate ground state , corresponding to the @xmath15 symmetry that is ultimately broken within the ising magnetic phases in the thermodynamic limit .
we furthermore argue that stable ferromagnetism is best observed in a 1d _ few - body _ system . _
effective spin model. _ we start by considering a 1d two - component ( @xmath16 ) gas of bosons with mass @xmath17 , trapped in a harmonic potential with frequency @xmath18 .
inter- and intraspecies interactions are assumed to be short - ranged , and parametrized respectively by @xmath19 and @xmath20 .
the hamiltonian is @xmath21 \psi_\nu(x ) \\ & + \frac{1}{2}\sum_{\nu,\nu ' } g_{\nu\nu ' } \int d x \ \psi^\dag_\nu(x ) \psi^\dag_{\nu'}(x ) \psi_{\nu'}(x ) \psi_\nu(x ) , \label{eq : h}\end{aligned}\ ] ] where @xmath22 and @xmath23 , respectively , create and annihilate a boson with pseudo spin @xmath24 . in the following we use harmonic oscillator units , where @xmath25 .
moreover , we consider the symmetric case where the two intraspecies interactions are of equal strength and the number of particles in each spin state is the same , i.e. , @xmath1 and @xmath26 . in the tonks - girardeau limit @xmath27 , the particles become impenetrable and retain their particular ordering .
thus , even in the absence of an underlying lattice , one can consider the system as a discrete chain of length @xmath14 , where the site index defines the position along the ordered chain .
perturbing away from this limit allows nearest neighbors to exchange position and one can describe the system using a spin - chain model @xcite . to make this connection evident , it is convenient to replace the inter- and intraspecies interactions @xmath19 and @xmath1 by effective parameters @xmath28 and @xmath29 , such that the entire phase diagram is traversed by changing the `` angle '' @xmath29 , as depicted in fig .
[ fig : sketch ] .
these are related to the original interactions by @xmath30 and @xmath31 . to linear order in @xmath32 , the original hamiltonian
may then be written as @xcite : @xmath33 here , @xmath34 is the energy of @xmath14 identical fermions , and @xmath35 is an effective heisenberg xxz hamiltonian : @xmath36 , \label{eq : spinh}\end{aligned}\ ] ] where @xmath37 , @xmath38 , and @xmath39 are the pauli spin matrices at position @xmath40 .
the constant @xmath41 ensures that only identical ( distinguishable ) bosons interact via @xmath42 ( @xmath19 ) .
note that , apart from parity , the properties of the eigenstates of @xmath35 are invariant under the transformation @xmath43 @xcite .
the energy cost of exchanging opposite spins at positions @xmath40 and @xmath44 depends on the underlying external confinement , an effect encapsulated in the coefficients @xmath45 of eq . .
these must in general be calculated numerically , but a key simplification is that they are independent of the interactions and thus one need only determine them for a single @xmath29 . in particular , for infinite intraspecies interaction ( @xmath46 , @xmath47 ) , the system is equivalent to a two - component fermi gas , where the @xmath45 have already been investigated for @xmath48 @xcite .
we have recently shown for the harmonically trapped fermi gas that @xmath45 is well approximated by the analytic expression @xmath49 @xcite . referring to the supplemental material @xcite
, we find that this approach also yields near exact wavefunctions for _ any _ @xmath29 .
we will , thus , use this ansatz to consider larger numbers of particles .
_ phase diagram. _ to gain insight into the behavior of the trapped boson system , it is instructive to first consider the uniform xxz model ( @xmath50 ) , where the magnetic phases are well characterized @xcite . in the thermodynamic limit , the local density approximation ( lda ) shows that the topology of the uniform phase diagram is the same as that of the trapped system . as displayed in fig .
[ fig : sketch ] , we expect to have fully saturated ferromagnetism ( fm ) of the ising type when @xmath51 , i.e. , when @xmath52 @xcite
. indeed , the first - order transition at @xmath53 corresponds to @xmath54 , the criterion for phase separation in a weakly interacting two - component bose gas @xcite .
however , as we discuss below , the interface between @xmath13 and @xmath12 bosons is sharper in the strong - coupling limit , being of order the interparticle spacing .
the other first - order fm transition , @xmath55 , has no weak - coupling analogue since it occurs on the metastable `` upper branch '' for attractive interactions ( or super - tonks " regime ) .
when @xmath56 , @xmath35 becomes equivalent to the xxx heisenberg model that can be realized with fermions , e.g. , @xmath57 can be mapped to the upper branch of the attractive fermi gas @xcite , which is the regime where itinerant ferromagnetism exists @xcite .
once @xmath58 , we will also have ising antiferromagnetism ( afm ) for @xmath59 , but the phase transition to this state is of the kosterlitz - thouless type in the uniform case @xcite . 0 pt for increasing particle number , from @xmath60 to @xmath61 ( solid lines ) .
the dash - dotted black line is an extrapolation to the limit @xmath62 @xcite . the black symbols and shaded regions correspond to the interaction strengths and magnetic phases illustrated in fig .
[ fig : sketch ] .
diamonds indicate analytical bethe ansatz results within lda , where @xmath46 corresponds to a two - component fermi gas , and @xmath57 is equivalent to identical bosons .
the dotted green line corresponds to the result for spin - polarized bosons for @xmath63 .
, title="fig : " ] _ ground - state energy. _ we now turn to the question of how the few - body system approaches the many - body limit .
we first consider the ground - state energy of the trapped gas : @xmath64 @xcite .
since @xmath34 is a simple energy shift , we take @xmath65 as a function of @xmath29 , as shown in fig .
[ fig : gscontact ] .
the result for @xmath66 exhibits discontinuities in the slope at @xmath67 , @xmath68 , corresponding to level crossings in the energy spectrum , where the total spin of the ground state is @xmath69 for @xmath70 and @xmath71 for @xmath72 . in this case
, there is no intraspecies interaction to couple states of different total spin , similarly to the two - component fermi gas , and one can not adiabatically evolve the ground state .
however , for @xmath73 , the total spin is no longer conserved and thus one can access different magnetic phases by varying @xmath29 adiabatically .
figure [ fig : gscontact ] shows that the results rapidly approach a universal curve with increasing @xmath14 .
in particular , we see that certain @xmath29 extrapolate to known bethe ansatz results within lda @xcite . for @xmath46 , where our system corresponds to a two - component fermi gas
, we recover the lda ground - state energy of the repulsive gaudin - yang model @xcite : @xmath74 for @xmath57 , we have @xmath75 and the ground state is equivalent to that of a strongly repulsive gas of @xmath14 identical bosons .
thus , we see that our results converge to the lda ground - state energy of the lieb - liniger model at strong coupling @xcite : @xmath76 within the entire region @xmath77 , we find that the ground - state energy in the limit @xmath62 corresponds to that of spin - polarized bosons , where it follows from eq .
that @xmath78 . for fixed spin populations
@xmath79 , this result physically corresponds to phase separation of @xmath13 and @xmath12 particles , where we can neglect the energy of the interface in the thermodynamic limit since it scales like @xmath80 compared to the total energy @xmath81 .
the formation of spin - polarized domains is a signature of ferromagnetism , and the emergence of kinks in the energy at @xmath9 and @xmath47 clearly indicates a first - order transition to the fm phase ( fig .
[ fig : gscontact ] ) . while there are no sharp features in the energy at the afm phase boundaries , the fact that @xmath82 at @xmath67 is a direct consequence of classical nel ordering , where identical spins are staggered and thus experience no intraspecies interaction .
0 pt , compared to the energy of the second excited state , @xmath83 .
( b ) trap - averaged magnetic correlation function evaluated in the ground state .
the dash - dotted line is an extrapolation to the thermodynamic limit @xcite , and the stars are analytic results for a uniform system .
the insets show the spatially - resolved correlation function for @xmath84 and two different values of @xmath29 .
( c ) spin gap @xmath85 for the spin polarised ( dashed ) and unpolarised ( solid ) ground states .
the green dotted lines are the analytical results for large @xmath14 see supplemental material @xcite .
, title="fig : " ] _ magnetic phases. _ a clear signature of magnetic order can be extracted from the degeneracy of the ground state .
a two - fold degenerate ground state is a _ necessary _ condition for ising magnetic order , and thus its emergence in the few - body system is a precursor for symmetry - broken magnetic states as @xmath62 . as shown in fig.[fig : degeneracyandcorrelationsandspingaps](a ) , the two lowest energy states become rapidly degenerate with increasing @xmath14 in the fm regime . in the few - body system , these two states have opposite parity , but as they become degenerate they can be combined to form states that break spin symmetry and parity .
physically , this degeneracy emerges once the tunnelling between different symmetry broken states is suppressed for large @xmath14 .
a similar situation occurs in the afm region , but the degeneracy manifests more slowly .
therefore , very few particles are necessary to observe strong magnetic correlations and a large breaking of the @xmath15 symmetry in the fm region , while larger systems are needed to observe similar effects in the afm region .
we can further characterize the different phases using the short - range correlation function @xmath86 in the ground state . referring to the insets of fig .
[ fig : degeneracyandcorrelationsandspingaps](b ) , we see that @xmath87 is always close to @xmath88 around @xmath67 , signifying afm correlations . likewise ,
for the fm phase , we obtain @xmath89 throughout the spin chain , excluding the interface between the spin - polarized domains . for strong interactions ,
this domain wall is sharp , decaying exponentially with distance even in the few - body system . to obtain a more global description of magnetic correlations , we take the trap - averaged function @xmath90 , normalized such that @xmath91 and @xmath92 . at the fm boundaries ,
the ground state is a heisenberg ferromagnet with total spin @xmath93 oriented in the @xmath6-@xmath7 plane . here
, the wavefunction is obtained by applying the total spin lowering operator @xmath94 times to the spin - polarized system @xmath95 .
thus , by symmetry , @xmath96 . with increasing @xmath14 , we find that the averaged short - range correlations approach a rectangular function in the fm region , where @xmath97 around @xmath98 for @xmath99 .
this is consistent with the appearance of classical ising fm along the @xmath8 direction .
by contrast , we see that quantum fluctuations reduce the ising afm correlations around @xmath67 , with @xmath100 in the limit @xmath62 .
furthermore , @xmath101 evolves smoothly with @xmath29 from the afm phase to the disordered `` spin liquid '' occupying the region @xmath102 .
the disordered liquid also features @xmath103 since its spin correlations resemble those in a fermi system .
indeed , for @xmath104 , the spin model can be formally mapped onto free fermions using the jordan - wigner transformation @xcite . for the uniform case ,
the free - fermion problem can be solved analytically and we find @xmath105 , which compares well with the large @xmath14 limit of the trapped system [ see fig .
[ fig : degeneracyandcorrelationsandspingaps](b ) ] .
_ spin imbalance. _ when @xmath106 , magnetic correlations strongly influence the behavior of the trapped system . in the limit of a single impurity @xmath107
, we find that the type of correlations determines the position of the @xmath12 impurity : in the afm regime , the impurity sits at the trap center @xcite , while fm correlations confine it to the edge .
more generally , we find in the few - body system that the fully spin - polarised state always has the lowest energy within the fm regime , as expected , while the ground state for all other @xmath29 is unpolarised . in practice , one can observe this by adding a small spin - interconversion coupling @xcite . to quantify this behavior , we determine the energy cost of flipping a spin in the ground state [ fig . [
fig : degeneracyandcorrelationsandspingaps](c ) ] . for the fully polarized fm ground state
, we see that the `` spin gap '' @xmath85 always vanishes at the fm phase boundaries and rapidly converges to a universal curve with increasing @xmath14 @xcite .
this further supports the idea that fm can be realized in a few - boson system .
on the other hand , for the unpolarized system we find that the afm spin gap remains finite even in the disordered phase , illustrating the different nature of the spin excitations . in both cases ,
we find that the behavior in the trap is qualitatively different from that in the uniform system , since the flipped spins predominantly reside at the trap edges and , hence , the spin gaps may not be obtained via the lda . _
concluding remarks. _ the physics described in this letter may be realized in a two - component bose gas close to overlapping 1d inter- and intraspecies resonances , which obviously necessitates some fine tuning .
however , we emphasize that the 1d bose gas is substantially more tunable than its 3d counterpart : firstly , the magnetic field need not be set exactly to any of the 3d resonances , since part of the fine tuning can be done by varying the strength of the transverse confinement @xcite , and this may even be done in a species - selective manner .
secondly , we propose to employ _ any _ two hyperfine spin states , as spin - changing collisions will be suppressed in the strongly interacting `` fermionized '' limit .
thus we anticipate that the strongly - coupled two - component bose gas could be investigated in current experimental setups .
we have argued that clear signatures of a first - order quantum phase transition to a ferromagnetic state
the analogue of itinerant electron stoner ferromagnetism appear already in the few - body limit .
furthermore , we emphasize that the few - body scenario is likely to be ideal for observing magnetic phases since the condition of staying in the strongly interacting regime , @xmath108 , becomes increasingly difficult to satisfy as @xmath14 increases .
smaller values of @xmath14 also enlarge the relative energy spacing between eigenstates , thus making it easier to tune interactions adiabatically and access stable magnetic ground states that are robust with respect to thermal fluctuations .
hence , the harmonically trapped and strongly interacting few - body bose gas appears ideally suited for the realisation of quantum magnetic phases .
* acknowledgments : * we gratefully acknowledge fruitful discussions with victor galitski , xi - wen guan , leticia tarruell , lincoln turner , and artem volosniev .
p.m. acknowledges support from erc adg osyris , eu equam , plan nacional foqus , 2014 sgr 874 , spanish mineco ( severo ochoa grant sev-2015 - 0522 ) , and the ramn y cajal programme .
49ifxundefined [ 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 10.1038/nature06480 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.61.2472 [ * * , ( ) ] @noop _ _ , , vol .
( , , ) @noop * * , ( ) link:\doibase 10.1126/science.1177112 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.240404 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.165302 [ * * , ( ) ] link:\doibase 10.1103/physreva.89.023611 [ * * , ( ) ] link:\doibase 10.1103/physreva.76.043617 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.170403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.050402 [ * * , ( ) ] http://stacks.iop.org/0034-4885/77/i=3/a=034401 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/98/30005 [ * * , ( ) ] link:\doibase 10.1038/nature02530 [ * * , ( ) ] link:\doibase 10.1126/science.1100700 [ * * , ( ) ] link:\doibase 10.1038/nature09393 [ * * , ( ) ] link:\doibase 10.1126/science.1201351 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.075303 [ * * , ( ) ] link:\doibase 10.1126/science.1240516 [ * * , ( ) ] link:\doibase 10.1038/nphys2878 [ * * , ( ) ] link:\doibase 10.1103/physreva.92.061601 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.85.1633 [ * * , ( ) ] \doibase 10.1023/a:1018705520999 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.111.045302 [ * * , ( ) ] link:\doibase 10.1103/physreva.88.033607 [ * * , ( ) ] link:\doibase 10.1103/physreva.88.021602 [ * * , ( ) ] link:\doibase 10.1103/physreva.90.013611 [ * * , ( ) ] link:\doibase 10.1038/ncomms6300 [ * * , ( ) ] http://stacks.iop.org/1367-2630/16/i=10/a=103004 [ * * , ( ) ] link:\doibase 10.1103/physreva.91.013620 [ * * , ( ) ] http://dx.doi.org/10.1038/srep10675 [ * * , ( ) ] link:\doibase 10.1126/sciadv.1500197 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.106801 [ * * , ( ) ] link:\doibase 10.1103/physreva.91.023620 [ * * , ( ) ] link:\doibase 10.1103/physreva.90.023630 [ * * , ( ) ] @noop in link:\doibase 10.1007/bfb0119591 [ _ _ ] , , vol . , ( , ) pp . link:\doibase 10.1103/revmodphys.84.1253 [ * * , ( ) ] @noop @noop link:\doibase 10.1103/physreva.72.063620 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.050402 [ * * , ( ) ] link:\doibase 10.1103/physreva.78.033612 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.81.938 [ * * , ( ) ] link:\doibase 10.1016/j.aop.2008.03.004 [ * * , ( ) ] link:\doibase 10.1016/j.aop.2008.03.005 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.185302 [ * * , ( ) ] * supplemental material : magnetism in strongly interacting 1d quantum mixtures * + + |
it is widely accepted that our milky way galaxy and m31 are larger than most spirals ( see , for example , van der kruit , 1990 ) .
this belief is largely a historical accident , resulting from early high estimates of the hubble parameter , @xmath1 . until recently ,
there has been no direct way to test this assumption ; analysis of volume limited surveys of galaxies in the local group certainly reveals that the milky way is among the largest galaxies in our immediate vicinity when we consider galaxies of all hubble types , but since the local group is dominated in number by dwarf galaxies this is hardly surprising .
we have found , on the other hand , from an analysis of nearby spirals of similar hubble type to our galaxy , and with well - determined distances , that the milky way is very much average in size ( goodwin et al . , 1997 ) .
this is strong evidence in support of the principle of terrestrial mediocrity " ( vilenkin , 1995 ) , that we live in an ordinary galaxy in an ordinary part of the universe .
if the linear sizes of enough nearby galaxies can be determined , a comparison of these sizes with the inferred sizes of galaxies at higher redshift gives an indication of the value of @xmath1 . even without knowing whether the sizes of the milky way and m31 are typical for spirals , some researchers have attempted to use this kind of argument to find a value for @xmath1 .
notably , sandage ( 1993a , b ; 1996 ) chose m101 as a typical " spiral , and argued that if the linear diameter of m101 is equal to the mean of field galaxies with similar appearance then the most probable value for @xmath1 is 43 km s@xmath0 mpc@xmath0 . a key step in his chain of inference
was the assumption that m101 not be the largest in a distance - limited sample " .
unfortunately , we have found that m101 is indeed one of the largest galaxies in our part of the universe . a similar analysis in sandage ( 1996 ) using m31 as the `` typical '' spiral yielded @xmath2 km s@xmath0 mpc@xmath0 . in this paper
we adopt a similar method to estimate @xmath1 , based on the principle that a sample of nearby spirals of specific hubble type represents a `` fair '' sample of the intrinsic population .
we extend the range of hubble types considered , however , in order to increase the size of both the calibrating sample of nearby objects and the distant population of several thousand field spiral galaxies .
( sandage s analysis used 60 galaxies ) .
we , therefore , aim to find in this way a more reliable value for @xmath1 . in this report
we summarise our preliminary results .
the corrected 25 b - mag arcsec@xmath3 isophotal diameters for all of our galaxies were taken from the rc3 catalogue ( de vaucoulers 1991 ) .
the observed major axis of this isophote was corrected for galactic extinction but not for inclination as spiral galaxies are found by those authors to be substantially optically thick " at this isophotal level .
this provides us with a uniform sample using a surface brightness diameter that is hence independent of distance ( the redshifts involved are too low to require any @xmath4 corrections ) allowing us to directly compare our local sample with the more distant sample .
a sample of 11 spiral galaxies with independent cepheid distances was chosen to act as the local calibrators .
the cepheid distances have been tabulated in giovanelli ( 1996 ) and freedman ( 1996 ) ; the figure for ngc3351 comes from graham ( 1997 ) .
table 1 summarises the data , converting the corrected isophotal diameters in arcmin to a linear diameter in kpc using the cepheid distances . in our earlier paper ( goodwin 1997 )
we also make a calculation for the galactic diameter at this isophote using data from van der kruit ( 1987 , 1990 ) which yields a 25 b - mag arsec@xmath3 diameter of @xmath5 leading to the conclusion , from comparison with the data in table 1 , that the milky way is a very average - sized galaxy , especially for its hubble type , @xmath6 ( van der kruit 1987 ) .
this small sample of galaxies also seems approximately to confirm the idea that linear diameter is relatively independent of hubble type within this range ( de jong 1996 ) .
.the hubble type @xmath7 , face - on angular 25 b - mag arcsec@xmath3 isophotal diameters @xmath8 , cepheid distances , @xmath9 , and actual linear diameters @xmath10 corresponding to this isophote , for 11 local spiral galaxies of hubble types 2 to 6 . [ cols="^,^,^,^,^,^ " , ] consider now the distribution of ( natural ) log linear diameters for this local sample of table 1 , supplemented by the milky way .
we will use log linear diameters throughout this paper since these have been shown to be well - modelled by a gaussian distribution ( c.f .
paturel 1979 ; lynden - bell 1988 ) . calculating the sample mean value , @xmath11 , and dispersion , @xmath12 , of the ( natural ) log linear diameter distribution for this local sample ,
we obtain @xmath13 fig .
[ figure : local ] shows the sample cumulative distribution of log linear diameter , compared with the cumulative distribution function of a gaussian with mean and dispersion equal to the sample values .
it can be seen that the local data give a good fit to a gaussian distribution , as borne out by the significance level ( or @xmath14 value ) of the kolmogorov smirnov ( ks ) statistic .
if the nearby galaxy sample is typical of the universal intrinsic population of galaxies in the same range of hubble types , we can use this information to obtain an estimate of the hubble parameter .
we have used angular diameters of more than 3827 spiral galaxies out to a redshift of over 20000 km s@xmath15 ( although the vast majority of galaxies in the sample have redshifts of 1000 - 5000 km s@xmath16 ) taken from the rc3 catalogue .
all these galaxies have known redshifts , so the assignment of their true linear diameters depends only on the assumed value of the hubble parameter to give their distances , together with the assumption that the hubble flow is uniform .
the sample of galaxies is spread across the sky including both cluster and field galaxies and we assume no bulk local motion that may change distances . unlike sandage , we have included spirals of all hubble types 2 through 6 .
the justification for this is that the linear diameters for a chosen value of @xmath1 are the same for all of these hubble types ( de jong 1996 ) .
the sample of distant galaxies is biased strongly toward large galaxies as the rc3 catalogue has an angular diameter cut - off at @xmath17 arcmin . at a redshift distance of 5000 km
s@xmath0 this would correspond to a galaxy with diameter @xmath18 kpc .
this result shows that galaxies even the size of the milky way will not appear in 1 arcmin angular diameter limited catalogues further away than @xmath19 km s@xmath0 . at even larger distances
only the most superlarge galaxies ( similar to ngc 1365 or m 101 ) would appear in such catalogues .
we address the question of completeness in two different ways , which we now summarise .
a comprehensive discussion of our statistical analysis will follow shortly in a full report . 1 . under the assumption that the spatial distribution of galaxies is uniform and the intrinsic distribution of log linear diameters is gaussian
, we determine the probability density function of log linear diameter for _ observable _ galaxies subject to a selection function characterised by a sharp apparent angular diameter limit at 1 arcmin , _ and _ an upper and lower redshift cutoff .
the primary function of the redshift cutoff was to exclude significant deviations from uniform hubble flow due to galaxies at low redshift with substantial peculiar motions .
as discussed by e.g. gould ( 1993 ) , the presence of a redshift ( or equivalently distance ) selection function also introduces a bias in the distribution of observable objects .
our analysis sets out to both compute and remove this bias in conjunction with the more familiar ` malmquist ' bias directly associated with the angular diameter limit .
we thus estimate from the distribution of observable galaxies the parameters of the intrinsic , gaussian , distribution of log linear diameters .
2 . as an alternative to the above approach , instead of attempting to remove the selection bias to which the distant sample is subject , we subject the distant sample to a new selection limit , dependent only on the log linear diameter .
since the selection is now only a function of an intrinsic quantity we can apply the same selection criterion to the local sample .
we discuss quantitative aspects of these two treatments of selection effects in the results section below .
the extreme range of possible values for the hubble parameter determined by different techniques and published recently is from about 40 to 80 km s@xmath0 mpc@xmath0 .
if the value of @xmath1 were as low as 40 km s@xmath0 mpc@xmath0 then the galaxy ( and most other galaxies from the local calibrating sample ) would not appear above the angular diameter cut - off at a redshift distance of 5000 km s@xmath0 mpc@xmath0 or greater .
this would seriously undermine several other recent determinations of @xmath1 , as our local sample comprises e.g. those galaxies used to calibrate the i band tully - fisher relation ( c.f .
giovanelli , 1996 ) .
if , however , @xmath1 were higher than 80 km s@xmath0 mpc@xmath0 , then the milky way and m31 ( which seem to be typical representatives of nearby galaxies over the considered range of hubble types , in the centre of our calibrating sample ) would be large spirals and ngc 1365 and m 101 would be among the largest spirals in the observable universe . we have used three methods in concert to examine the size distribution of the distant galaxies and estimate @xmath1 , quantifying the above two general conclusions .
first a subset of distant galaxies between 1500 and 5000 km s@xmath0 was selected . as mentioned in the preceding section ,
the lower limit of this sample was chosen to eliminate the worst of the contamination of the sample by peculiar motions , whilst the upper limit was chosen so as to provide an approximately distance - limited sample .
we found that the extension of the upper limit to 40000 km s@xmath0
( i.e. effectively no upper redshift limit ! ) yielded an observable distribution of log linear diameters which deviated significantly from the gaussian distribution , with a shifted mean , predicted by our model of uniform spatial density and a sharp angular diameter limit .
we considered this to be due to the breakdown of our model assumptions ; in particular the selection function at high redshift was unlikely to be depend only on angular diameter , but also on inclination and would probably begin to display a sensitivity to morphological type .
when we adopted the limit of 5000 km s@xmath0 , however while this introduced a new selection bias similar to that discussed by gould ( 1993 ) and thus rendered the predicted distribution of log linear diameters formally non - gaussian we found that the computed observable distribution with this new selection function showed deviations from gaussianity of only a few percent .
we thus continued to model the predicted observable distribution as a gaussian .
moreover , the distant sample now matched this predicted observable distribution very well .
figure [ figure : l04 ] shows the sample cumulative distribution of log linear diameter for the selected galaxies and , for comparison , the best - fit gaussian to the predicted distribution of observable galaxies with this new selection function .
note that application of the ks test suggests a good fit of the data to this approximately gaussian model .
the observed mean of the best - fit gaussian distribution lies at @xmath20 , if we assume that @xmath21 km s@xmath0 mpc@xmath0 . correcting for the effects of bias introduced by the angular diameter and redshift selection limits
, we deduce that the best - fit mean value of the _ intrinsic _ distribution of log linear diameter is @xmath22 .
the mean of the calibrating sample , however , lies at @xmath23 , where the quoted error is the standard error on the mean derived from the sample dispersion .
these two samples would , therefore , have the same mean if @xmath24 as a further means of placing limits on the likely value of @xmath1 it is useful to consider the _ order statistics _ of the modelled intrinsic distribution of log linear diameters ( c.f .
hendry , odell & collier - cameron 1993 ) .
for a sample of 12 calibrating galaxies , which we are assuming to be drawn from the same intrinsic population as the distant sample , one can therefore pose the question of how likely it is , for a given value of @xmath1 , that one would obtain e.g. as small a galaxy as the galaxy with smallest linear diameter , or conversely as large a galaxy as that with largest linear diameter .
there are many different ways to combine the order statistics and apply statistical tests to their values , and we will describe such an analysis in detail in our follow - up paper . here
we present one plot , which succinctly illustrates the same principal features as a more sophisticated analysis .
figure [ figure : order ] shows the lower tail of the third order statistic ( ie .
the probability that the third smallest galaxy in a sample of size 12 drawn from the distant sample would be as small as , or smaller than , the smallest observed member of the local calibrating sample ) and the upper tail of the tenth order statistic ( ie .
the probability that the third largest galaxy in a sample of size 12 drawn from the distant sample would be as large as , or larger than , the largest observed member of the local calibrating sample ) for different assumed values of @xmath1 in the range 30 - 100 km s@xmath0 mpc@xmath0 .
we could , of course , choose any of the order statistics for this illustration but the third and tenth are suitable examples .
as can be seen from this figure , the value of @xmath1 where these two probability curves cross is found at @xmath25 km s@xmath0 mpc@xmath0 .
in addition it can be seen that values of @xmath26 km s@xmath0 mpc@xmath0 are not favoured ( as the largest galaxies in the local sample would then be unreasonably large compared with the distant galaxies ) nor are values of @xmath27 km s@xmath0 mpc@xmath0 ( as then the smallest galaxies in the local sample would then be unreasonably small compared with the distant galaxies ) .
the above analysis assumes that we have correctly accounted for , and corrected for , the observational selection effects within the distant sample , and that the local sample is a `` fair '' sample of the intrinsic distribution of log linear diameters .
a further statistical analysis of the data was made by , instead , imposing a sharp linear diameter limit on both the distant and nearby samples .
we adopt a limit of @xmath18 kpc since at the redshift limit of 5000 km s@xmath0 ) this corresponds to an angular size of 1 arcmin above which we assume that the entire sample is reasonably complete .
provided again that the local galaxies are assumed to be a fair sample of the intrinsic population , the distribution of log linear diameter in the _ truncated _ local and distant samples should be identical .
this analysis is less powerful than that using the entire sample ( especially at low @xmath1 where the number of calibrators above the cut - off diameter becomes small ) , but has the advantage of applying the same selection function to both samples instead of essentially reconstructing the ` missing ' galaxies from the observed distribution .
thus , what is lost in statistical power is partially gained in robustness , and we consider this approach a useful consistency check .
[ figure : limit ] shows a comparison of the means of the truncated distant and local samples , together with @xmath28 errors on the means , as a function of @xmath1 ( which , of course , determines the truncation limit ) .
as can be seen from this figure , good agreement between the truncated means of the local and distant data are found at @xmath29 and @xmath30 km s@xmath0 mpc@xmath0 .
this weaker analysis therefore suggests that @xmath31 gives a reasonable bound . as above , however , one can place some further constraints on @xmath1 by considering the order statistics of the largest galaxies in the local sample , and on this basis we find that the lower end of this range is favoured by the data .
assuming the local calibrators to be a fair sample of the intrinsic distribution , we carried out a series of monte carlo simulations in order to determine the probability of a truncated random sample of 12 galaxies larger than @xmath18 kpc in diameter containing two galaxies larger than 59 kpc in diameter , for a range of different values of @xmath1 .
this probability is less than 10% for all @xmath32 km s@xmath0 mpc@xmath0 and drops below 5% for @xmath26 km
s@xmath0 mpc@xmath0 .
the results presented in this report strongly suggest that values of @xmath1 higher than 75 km s@xmath0 mpc@xmath0 or below 40 km s@xmath0 mpc@xmath0 are ruled out . clearly this is not a dramatically new conclusion , but should serve to underline that the simple idea of comparing the linear size of local and distant galaxies is certainly a plausible method for estimating @xmath1 . a value of @xmath1 in the range 50 - 55 km s@xmath0 mpc@xmath0 gives a good fit between the mean log diameter of the local and distant samples both with and without a lower diameter limit and renders the statistical properties of the largest and smallest local galaxies easily compatible with those predicted from the distant sample .
this result is also consistent with a number of independent recent estimates from other distance indicators .
our analysis has , however , used distance moduli to the local galaxies derived via the the traditional cepheid calibration of madore & freedman ( 1991 ) , adopting a true distance modulus of 18.5 for the large magellanic clouds .
recently the lmc distance modulus has been revised , using hipparcos data ( reid 1997 ; gratton 1997 ) to calibrate the cepheid period - luminosity relation in galactic open clusters .
the simplest interpretation of the new results is that our estimate for @xmath1 should be reduced by about 10 per cent . taking this into account
would imply that the true value of the hubble parameter lies in the range 45 - 50 km s@xmath0 mpc@xmath0 .
on the other hand , kochanek ( 1997 ) has embarked on a comprehensive re - analysis of the cepheid distance scale and has concluded that relative distance moduli between the lmc and a number of local galaxies , as determined from hst cepheid observations , have been significantly _ under _ estimated .
the effect of the kochanek re - calibration of the cepheid distance scale may well largely cancel out any reduction in the hubble parameter resulting from the hipparcos results . for the moment , then , we will consider the range @xmath33 km s@xmath0 mpc@xmath0 as the range most likely from this analysis . in a simple einstein - de sitter universe with @xmath34 and @xmath35 this range corresponds to an age of approximately 12 - 13 gyr since the big bang .
for the value of the hubble parameter to lie significantly outside of this range , not only the milky way but most of the spirals in our local sample must be collectively either significantly larger or significantly smaller than the population of spirals in the visible universe . if that were the case , it would mean that we live in an atypical part of the universe , and that would remove the justification for any extrapolation from the nearby region to the universe at large . given that , on the contrary , we find the milky way to be a very average - sized spiral galaxy compared with other local galaxies of similar hubble type , it seems unlikely to us that the ` principle of terrestrial mediocrity ' should hold on the scale of the local supercluster and yet break down on larger scales .
electronic rc3 data was taken from the cds internet site and we acknowledge the use of the starlink computers at sussex and glasgow .
spg would also like a postdoc , please .
freedmanw.l . 1996 , to appear in critical dialoges in cosmology , ed .
n. turok , lanl preprint no
. astro - ph/9612024 | measurements of the linear diameters of 12 nearby spiral galaxies with distances determined from primary indicators suggest that both the milky way galaxy and m31 are in the middle of the range of sizes for such galaxies . by comparing the measured linear diameters of these nearby systems with the inferred diameters of a sample of more than 3000 spirals with known redshifts , we conclude that the most likely value of the hubble parameter lies in the range 50 - 55 km s@xmath0 mpc@xmath0 . |
observations of pulsar signals provide a valuable tool for investigating electron density distribution in the interstellar medium .
the pulsar intrinsic signal is narrow in time , being about few percent of the pulsar period .
the observed signal ( averaged over many periods of pulsation ) is broad and asymmetric with a sharp rise and a slow decay .
the pulse broadening is attributed to the random refraction the waves experience while propagating in the interstellar medium .
the shapes of the signals , and their scalings with wavelength @xmath0 and with the pulsar distance have been investigated observationally and analytically for 30 years .
the standard theory of interstellar scintillations ( described in detail below ) assumes that the propagating wave is refracted by random gaussian gradients , thought to be a good approximation due to the central limit theorem @xcite . the symmetric gaussian distribution is fully characterized by its second moment , and the standard theory of scintillations aimed at reconstructing this moment from observations .
the contradictions of this theory with observations were noted in the early 1970 s , although not many pulsars were investigated at that time to make a definitive conclusion .
the signal shape is characterized by its time - width , @xmath1 , estimated at the @xmath2 level . in 1971 ,
sutton analyzed scalings of pulse broadening times , @xmath1 , with the radio wavelength , @xmath0 , and with dispersion measures , @xmath3 , corresponding to the pulsar s distance along the line of sight , @xmath4 .
observational scaling for large dispersion measures , @xmath5 , is close to @xmath6 , while the theory gave @xmath7 . the recent observational results are shown in fig .
( [ elbow ] ) . to overcome the difficulty with the anomalous dm - scaling , sutton suggested that the interstellar turbulence was not statistically uniform along the lines of sight , so that the lines of sight for more distant pulsars intersected regions with stronger turbulence .
the second paradox was encountered by @xcite , who compared the observed shapes of the pulses with the shapes predicted by the standard theory .
he obtained a surprising result that the model of continuous turbulent medium was not consistent with observations .
rather , the best agreement was given by the models where all the scattering material was concentrated either in a thin screen or in a slab covering only @xmath8 of the line of sight between the pulsar and the earth , with nearly equal quality of fit to the data [ see the discussion and fig . 4 in @xcite ] .
these two assumptions that are necessary to reconcile the standard theory with the observations may have physical grounds since the interstellar medium is not uniform and denser regions ( with stronger turbulence ) are encountered closer to the galactic center . however , in this paper we discuss a simple and perhaps more plausible physical explanation for the sutton and williamson paradoxes . according to this explanation ,
the sutton and williamson paradoxes are not the consequence of the non - uniform , large - scale galactic distribution of the electron density responsible for scattering .
rather , they reflect the universal properties of the microscopic structures of the density fluctuations . in our recent work
@xcite we proposed that the sutton paradox is the evidence of statistically uniform but _ non - gaussian _ electron distribution in the interstellar medium .
we noted that the time - broadened pulses belong to those observational objects that depend not on the moments of the electron - density distribution function , but on its full shape .
the standard gaussian approach did not recognize this fact because in the gaussian case , knowledge of the second moment is equivalent to knowing the distribution function itself .
however , for a general non - gaussian electron - density distribution the second moment may formally diverge , and the electron - density correlation function may not characterize the pulse shape . physically , this divergence means that the second moment of ray - angle deviation is dominated not by the bulk of its distribution function , but by the extremely far cut - offs at the tails of this function [ as we will see in the example of section [ boundaries ] ] .
in contrast , the observed signal shape is determined by the bulk of the distribution function and is not sensitive to its far - tail cut - offs . in other words , the electron - density correlation function and the shape of the observed pulsar signal
provide different , complementary descriptions of the electron distribution in the ism .
we also note that in the lvy model , the pulse shapes may provide a unique opportunity to diagnose plasma fluctuations at very small scales , @xmath9 cm . these scales can be smaller than the coulomb mean - free path in hii regions , and can be close to the ion gyro radius
. observations of pulse shapes may thus provide diagnostics for magnetized plasma fluctuations , which may help to elucidate the role of magnetic fields in interstellar turbulence , see also @xcite . in the theory of scintillation
, the quantity of interest is the transverse electron - density difference , averaged along the line of sight , @xmath10 $ ] .
a wave propagating in the interstellar medium is scattered ( refracted ) by density inhomogeneities , and the statistics of the refraction angle is related to the statistics of density differences across the line of sight . in the approach proposed in @xcite we assumed that the distribution function of @xmath11 was stationary and uniform , but non - gaussian , and it had a power - law decay as @xmath12 .
if this distribution does not have a second moment , then the sum of many density differences along the line of sight does not have the gaussian distribution , i.e. the central limit theorem does not hold .
instead , the limiting distribution , if it exists , must be the so - called lvy distribution , which is common in various random systems @xcite . by construction ,
such a distribution is stable under convolution , i.e. the appropriately rescaled sum of two independent variables drawn from the same lvy distribution , has again the lvy distribution . the gaussian distribution is a particular case of this more general stable distribution . the fourier transform ( the characteristic function ) of an isotropic lvy probability density function , @xmath13 , has the simple form , @xmath14 where the lvy index @xmath15 is a free parameter , @xmath16 .
this form is valid for an arbitrary dimensionality of the fluctuating vector @xmath17 .
as we explain below , in our case this vector is two - dimensional , @xmath18 $ ] , where @xmath19 is the coordinate in the plane perpendicular to the line of sight , @xmath20 , and @xmath21 is the unit vector in the direction connecting the two points in this plane , @xmath22 , where @xmath23 . equation ( [ levy ] ) can be taken as the definition of the _ isotropic _ lvy distribution , and in the present paper we will consider only isotropic distributions .
however , it is important to note that a stable distribution need not be isotropic [ see , e.g. , @xcite ] . in the simplest case ,
the anisotropic distribution is a copy of the isotropic distribution , rescaled along one axis ; but far less symmetric lvy distributions are possible .
non - isotropic stable distributions might arise in the case of the magnetized interstellar medium , when wave scattering has one or several preferred directions .
for example , anisotropic scattering may be characteristic of interstellar mhd turbulence , as described by @xcite , where the preferred direction is given by the large - scale magnetic field .
the sum of @xmath24 lvy distributed variables scales as @xmath25 , which becomes diffusion in the gaussian limit @xmath26 . for @xmath27 ,
the probability distribution function has algebraic tails , @xmath28 , for large @xmath29 , and its moments , @xmath30 , of the order @xmath31 diverge . in @xcite
, we considered smooth density fluctuations , @xmath32 , and the proposed lvy statistics of ray - angle deviation lead to the pulse - broadening scaling @xmath33 .
this agrees with observations when @xmath34 , and provides a natural resolution to the sutton paradox . in this paper
we consider scintillations produced by a general , non - smooth density fluctuations , @xmath35 .
more precisely , exponent @xmath36 denotes the scaling of the density - difference pdf defined in formula ( [ levy ] ) , i.e. this pdf should have the form @xmath37 . in the non - smooth case , the effective lvy index that enters the expressions for angle deviations is different from @xmath15 , and
is given by @xmath38 , as is explained in formulae ( [ levyscreen ] ) and ( [ tau ] ) below . as we noted above , in the lvy case the electron - density correlation function ( or second order structure function )
is contributed to by the far - tail cut - off of the electron - difference distribution function .
we will denote the correlation - function scaling as @xmath39 , where the scaling exponent , @xmath40 , is in general different and independent of the lvy distribution scaling exponents @xmath36 and @xmath15 .
the exponent @xmath40 is related to the fourier spectrum of electron density fluctuations , @xmath41 . in the present paper
we demonstrate that the proposed lvy statistics of ray - angle deviations provide a resolution for the williamson paradox as well . for this purpose
we , first , develop a theory of wave propagation in a lvy random medium by approximating the medium by a uniform series of scattering screens . then , we provide a general method for constructing the pulse - broadening function for an arbitrary lvy index @xmath42 . and finally , we compare our results with the observational signals of large - dispersion - measure pulsars , recently published by @xcite and by @xcite .
we obtain that the observational shapes agree well with the predictions of our theory for a uniform medium with the lvy index @xmath43 to @xmath44 , while they are inconsistent with the gaussian theory corresponding to the lvy index @xmath45 .
we also discuss the effect of `` over - scattering '' which is inherent for the lvy scintillations , because the time - signals have long algebraic tails that do not decay to zero during pulse periods .
this effect is crucial for comparing analytical and observational data , as we show in section [ observation ] . for illustrative purposes
, we also present a particular model of the density distribution that produces the lvy statistics of scintillations .
the proposed density distribution is strongly spatially intermittent , and it can be visualized as follows .
imagine that the electron density is concentrated in separated regions that have sharp irregular boundaries .
we make the simplest assumption that these boundaries are random , similar to randomly folded two - dimensional sheets or shocks .
if one further assumes that the line of sight intersects the boundary at a random angle whose distribution is uniform in all directions , such a picture corresponds to the particular case of a lvy distribution of @xmath46 , with @xmath47 , @xmath48 , the so - called cauchy distribution .
this distribution is rather distinct from the gaussian one , and is close to the distribution that has been predicted in our theory .
the importance of randomly oriented discontinuous objects that can be encountered across the line of sight has been emphasized in earlier theoretical work on scintillation , see e.g. , @xcite and @xcite .
as noted by @xcite , sharp density discontinuities may characterize stellar wind boundaries , supernova shock fronts , boundaries of hii regions at the strmgren radius , etc .
they may also arise from supersonic turbulent motion .
@xcite proposed that similar highly intermittent density structures may be responsible for observed low - frequency modulations ( fringes ) of dynamic pulsar spectra .
the approach of @xcite utilized the non - kolmogorov spectra of discontinuous density fluctuations .
@xcite considered wave scattering by confined or heavily modulated screens ( such as disks , filaments , etc . ) , when the statistics of angular deviations were generally assumed gaussian , with parameters varying along the screens .
both papers discussed important aspects of non - kolmogorov and spatially intermittent electron - density distribution in the ism , however , they implied the existence of the second moments of the scattering - angle distributions . therefore , in the earlier considerations , the effects that we discuss in the present paper could not be discovered . as we demonstrate in section [ boundaries ] , sharp density discontinuities can , in fact , produce a non - gaussian distribution of the scattering angle , whose second moment diverges . in our approach
we exploit such intermittent density statistics in their full generality . to elucidate the universal role of lvy distribution , we keep our consideration as simple as possible , assuming statistically uniform and isotropic scattering screens ( although generalizations for the non - isotropic and non - uniform cases are possible ) .
the presence of strongly - scattering structures is naturally represented in our model by slowly decaying , power - law tails of the scattering - angle distributions .
our model has a simple physical interpretation and provides a practical way of calculating pulse shapes and pulse scalings .
most importantly , it naturally resolves both the sutton and the williamson observational paradoxes , in a manner that is simpler and complementary to the standard gaussian picture .
the lvy - flight model may also be relevant to observations of extreme scattering events , such as those reported by @xcite .
such events have been investigated theoretically , as by @xcite .
we note that these theoretical studies invoke uncommon incidents of scattering much larger than typical values to explain these events . in this sense , these events are consistent with the lvy model we develop here , which includes rare , large events in a statistically - stationary way .
however , we do not yet know whether these phenomena might fit into a single model for interstellar scattering along with the common phenomena we discuss in this work .
the answer might possibly depend on the details of the picture for turbulence in which the lvy model is realized .
accordingly , we do not discuss the extreme scattering events in the present work .
to address the puzzles mentioned above we need to review the standard theory of interstellar scintillations .
first , we note that in the interstellar plasma with typical electron density @xmath49 , the electron plasma frequency is @xmath50 .
this frequency is much smaller than the typical observational frequency of @xmath51 , and , therefore , the propagating wave scatters only by a small angle on the scale of density inhomogeneities .
the fourier amplitude of electric field , @xmath52 , obeys the following equation : @xmath53 see , e.g. , @xcite .
we are not interested in polarization effects that are small in the considered approximation by a factor @xmath54 , where @xmath55 is the electron - cyclotron frequency , therefore , we consider the scalar wave amplitude , @xmath56 . equation ( [ waveeq ] ) can be reduced further , using the so - called parabolic approximation @xcite . assuming that the wave propagates in the line - of - sight direction , @xmath20 , we can separate the quickly changing phase of the wave from the slowly changing wave amplitude , @xmath57 , where @xmath19 is a coordinate perpendicular to @xmath20 .
the equation for the wave amplitude @xmath58 reads : @xmath59\phi_{\omega}(z,{\bf x})=0 , \label{paraboliceq}\end{aligned}\ ] ] where @xmath60 is a two - dimensional laplacian in the * x * plane , and @xmath61 is the classical radius of the electron .
following the approach of @xcite we introduce the two - point function @xmath62 , whose fourier transform with respect to time is @xmath63 . for coinciding coordinates , @xmath64 ,
this function is the intensity of the radiation whose variation with time we seek . to find this function
we need first to solve the equation for @xmath65 , which can be derived from eq .
( [ paraboliceq ] ) . assuming that @xmath66
, we obtain the following equation , @xmath67 where we denote @xmath68 , @xmath69 , and @xmath70 .
equation ( [ vequation ] ) is hard to solve since @xmath71 is an unknown random function .
the standard theory uses the fact that it takes many refraction events to appreciably deviate the ray trajectory , as we discussed in the introduction .
therefore , due to the central limit theorem , the ray deviation angle exhibits a gaussian random walk .
one can therefore assume that the density fluctuations are gaussian with the specified second - order correlator @xmath72 ^ 2 \rangle = 2 \kappa({\bf x}_1-{\bf x}_2)\delta(z_1-z_2)$ ] , where short correlation length in @xmath20 direction is the mathematical expression of the fact that the ray becomes appreciably deviated only when it travels the distance much larger than the the density correlation length @xcite .
the correlation function @xmath73 is the two - dimensional fourier transform of the spectrum of the density fluctuations .
more precisely , if the three - dimensional spectrum of the density fluctuations is @xmath74 , then @xmath73 is the fourier transform of @xmath75 with respect to @xmath76 . when the density has a power - law spectrum , @xmath77 , with @xmath78 , then @xmath79 . as we mentioned in the introduction , the statistically uniform gaussian model fails to reproduce the observational scaling of the the broadening time , @xmath80 .
numerous attempts to reproduce this scaling by using different spectra of density fluctuations within this model have not been successful either .
the purely shock - dominated density distribution has the spectrum @xmath81 , while the kolmogorov turbulence has the spectrum @xmath82 ; the difference in the spectral exponents is rather small to have a considerable consequence .
for various important aspects of the standard gaussian scintillation theory we refer the reader to @xcite .
in other contexts , the theory of wave propagation in gaussian random media was developed in @xcite . in the next section
we formulate the problem by assuming that the refraction occurs in a series of discrete structures , refracting ` screens ' , that are statistically identical , independent , and placed uniformly along the line of sight .
this setting is physically appealing since , as we have mentioned in the introduction , the refraction in the interstellar medium is consistent with the presence of spatially intermittent scattering structures , see e.g. , @xcite .
moreover , this is the simplest setting when the problem admits an exact analytic solution that allows us to treat both gaussian and lvy cases on the same footing .
let us assume that wave scattering occurs in a series of thin flat screens uniformly placed along the line of sight .
also , for simplicity assume that the incident wave is planar , although an analogous consideration can be made for spherical geometry as well .
each screen gives a contribution to the signal phase ; if we denote the wave function just before the screen by @xmath83 , then the wave function right after the screen will be given by @xmath84 .
the phase function , @xmath85 , can be found from eq .
( [ vequation ] ) , @xmath86 \right ) , \label{screen}\end{aligned}\ ] ] where the integration is done over the thickness of the phase screen , @xmath87 . between the phase screens , the wave propagation is free . effecting the change of variables , @xmath23 , and @xmath88 , we rewrite the eq .
( [ vequation ] ) in this region , @xmath89 in what follows we will be interested in the wave function , @xmath90 , averaged over different realizations of the electron density in the phase screens . due to space homogeneity ,
the averaged transfer function , @xmath91 , should depend only on the coordinate difference @xmath92 .
since refraction affects only the @xmath92 dependence of the wave function , we can assume that the averaged wave function is independent of @xmath19 , @xmath93 .
taking into account the @xmath19 dependence would lead to more cumbersome formulas , although it would not qualitatively change the results .
we therefore model the free propagation between the screens by @xmath94 to describe the scattering we need to specify the averaged phase function , @xmath95 . in the gaussian case
, one can assume that the screen width , @xmath87 , is of the order of the characteristic length of the density fluctuations , @xmath96 , and therefore is not necessary .
the only assumption required for our consideration is @xmath97 , and the final results can be generalized for this case .
this generalization would however require further assumptions about the scattering structures , which at the present level of understanding can hardly be justified .
] , @xmath98 , \label{gaussianscreen}\end{aligned}\ ] ] where @xmath99 .
assuming that the main contribution to the scattering comes from the scales , @xmath100 , much smaller than the density correlation length , we can expand @xmath101 , where @xmath102 , and @xmath103 is a typical amplitude ( say , rms value ) of density fluctuations .
for example , if density fluctuations arise due to passive advection by the kolmogorov turbulence , then @xmath104 .
the smooth density field corresponds to @xmath47 .
the situation is completely different in the lvy case , when the second - order moment of the density integral in ( [ freeequation ] ) diverges .
we however can assume that the density difference has some scaling form , @xmath105dz \sim l_0 \delta n_0 ( y / l_0)^{\alpha/2}$ ] , and use the lvy formula ( [ levy ] ) to write the most general expression for the averaged phase function @xmath95 , @xmath106 in the gaussian limit , @xmath26 , this formula reduces to the previous result ( [ gaussianscreen ] ) .
similar to the gaussian case , the exponent @xmath107 has the meaning of the density - difference scaling with the point separation , however , it characterizes not the second moment ( that does not exist ) , but the density - difference distribution function itself .
more precisely , for a chosen point separation @xmath92 , the projected density difference in the screen function ( [ screen ] ) , @xmath108=\lambda r_0 \delta n_{\perp } , \label{deltaphi}\end{aligned}\ ] ] should be drawn from a distribution function that depends only on the combination @xmath109 .
we also note that contrary to the gaussian case , the quantity @xmath103 entering equation ( [ levyscreen ] ) , does not correspond to the rms value of density fluctuations , since @xmath110 does not exist in the lvy case .
rather , it denotes the typical width of the density - difference pdf , say the density difference at which this pdf decays by the factor @xmath2 compared to its maximum value . on the contrary
, the rms value of @xmath11 would be given by far , non - lvy tails of this pdf , and would be much larger than @xmath103 . physically , the lvy screen function in the form ( [ levyscreen ] ) could correspond to the distribution of the phase integral , @xmath111 , when the integration path intersects a random density discontinuity , say a sharp boundary of an ionized region .
such a random boundary may naturally arise as a result of turbulent advection in an ionized region , and numerical simulations could test this intriguing possibility .
our suggestion is based on a result that a smooth , randomly oriented boundary reproduces the particular form of the lvy screen function ( [ levyscreen ] ) corresponding to @xmath48 , @xmath47 .
we will derive this result in the next section .
in general , particular values of @xmath36 and @xmath15 depend on the physical realization of scattering screens , which is not completely understood and can not be specified in this work . for our present purposes , we simply assume the general form of the screen function ( [ levyscreen ] ) and provide a method for calculating the shape and the scaling of the received signal .
let us consider a series of such screens placed along the line of sight , @xmath20 , perpendicular to it , such that the distance between two adjacent screens is @xmath112 .
number the screens from the source to the observer , and consider the @xmath113th screen .
denote as @xmath114 the wave function just before the wave passes the screen , and @xmath115 this function just after the screen . between the screens @xmath116 and @xmath113 , the propagation is free ,
therefore , we obtain from equations ( [ vequation ] ) and ( [ freeequation1 ] ) , @xmath117 we can write the wave function after the wave has passed @xmath24 screens , by iterating this formula @xmath24 times , and by using the expression for the screen function , ( [ levyscreen ] ) , @xmath118^{-n } \int \exp\left([ia\delta k]^{-1 } \sum\limits_{m=1}^{n}(\delta { \bf y}_m)^2 -b^{\alpha \beta/2}\sum\limits_{m=1}^{n-1 } \vert y_{m } \vert^{\alpha\beta/2 } \right)\times \nonumber \\ & u_{0}(y_{0})\ , d^2y_{n-1 } , \dots , d^2y_1\,d^2y_{0 } \label{unscreens}\end{aligned}\ ] ] where we introduced the short - hand notation : @xmath119 , @xmath120 , and @xmath121 .
to do the integrals , we need first to make a simple transformation of formula ( [ unscreens ] ) .
we will substitute the following identity , @xmath122= -i\pi a\delta k \int \exp\left(i\bxi_m\cdot \delta { \bf y}_m - i a \delta k \xi_m^2\right)d^2\xi_m , \label{identity}\end{aligned}\ ] ] for each @xmath123 , into formula ( [ unscreens ] ) . our ultimate goal is to find the time dependence of the pulse intensity at the @xmath24th screen , i.e. , @xmath124 .
we therefore change the order of integration and do the @xmath125-integral first . as the result
we get : @xmath126 introducing the non - dimensional variables @xmath127 and @xmath128 , we rewrite formula ( [ intensity ] ) as follows : @xmath129 in the rest of this section we will use only the non - dimensional variables , and will omit the tilde signs .
now we are ready to do the @xmath100-part of the integral .
for this we make a simple rearrangement in the sum @xmath130 . recalling that @xmath131 should be set to zero , we are left with @xmath132 , where @xmath133 . to complete the @xmath100-integration
, we simply use the lvy formula ( [ levy ] ) , and get : @xmath134 where @xmath135 is the fourier transform of @xmath136 , @xmath137 is the lvy distribution function introduced in ( [ levy ] ) .
the last step is to note that @xmath138 , which allows us to give a simple interpretation to formula ( [ levyintensity ] ) . before discussing this interpretation ,
we note the quantum mechanical analogy of our approach .
if we could assume that @xmath139 , and @xmath140 in the limit when the number of screens increases , @xmath141 , but the line - of - sight length is constant ( which would be natural for the gaussian case ) , the integral that we calculated in ( [ unscreens ] ) would formally become the feynman - kac path integral for the solution of the schrdinger equation with the potential @xmath142 .
the transition from ( [ unscreens ] ) to ( [ levyintensity ] ) given by formula ( [ identity ] ) is the transition from the coordinate to the momentum representation of this integral .
naturally , in the gaussian case one can find the shape of @xmath143 either by doing the multiple integral in ( [ levyintensity ] ) or by solving the corresponding schrdinger equation .
the first approach was essentially adapted in @xcite , the second one in @xcite . in our lvy model
we can not assume that @xmath144 , therefore , we have to work with the general expressions ( [ unscreens])-([intensity ] ) .
although the general integral ( [ unscreens ] ) is quite complicated , the quantity of interest , @xmath143 , which is calculated with the aid of this integral , has a quite simple meaning , and its probabilistic interpretation ( [ levyintensity ] ) can be easily understood .
the @xmath40-function in formula ( [ levyintensity ] ) means that the signal intensity @xmath143 is the probability density function of the time delay , @xmath145 .
the variable @xmath146 is proportional to the deviation angle of the ray path from the @xmath20-axis , @xmath147 .
[ this is clear from ( [ identity ] ) since @xmath148 , and this will also be the case in the geometric - optics analysis of section [ boundaries ] . ]
this angle is the sum of elementary angle deviations caused by each phase screen , viz @xmath149 .
the propagation time between two neighboring screens exceeds the propagation time along the @xmath20-axis by the amount @xmath150 \propto \xi_m^2 $ ] , and to get the total time delay , @xmath1 , we need to sum up these individual time delays .
formula ( [ levyintensity ] ) teaches us that each angle increment , @xmath151 , has a lvy distribution with the index @xmath38 , which provides a practical way of constructing the pulse broadening function @xmath143 without doing the multiple integrals in ( [ levyintensity ] ) .
this is a monte - carlo - type method for multiple integration . indeed , to find the effect caused by the interstellar medium
, we can neglect the intrinsic pulse shape , i.e. , we can assume that the initial angle of the planar wave is zero , @xmath152 .
then the function @xmath143 is the probability density function of the variable @xmath153^{4/\alpha}}{2l_0 ^ 2\pi^2c}\lambda^{2 + 4/\alpha } \sum\limits_{m=2}^n\left[\sum\limits_{s=2}^m\delta \bxi_s \right]^2 , \label{tau}\end{aligned}\ ] ] where all the variables @xmath151 are distributed independently and identically according to the lvy law with the index @xmath38 .
the same expression was obtained in our previous work for the case @xmath47 , although without a detailed derivation @xcite .
we remind that the scaling exponent @xmath36 corresponds to the spatial scaling of the density field , as is defined in eq .
( [ levyscreen ] ) . the gaussian case corresponding to @xmath47 , @xmath26 was originally considered by @xcite .
the gaussian case corresponding to the kolmogorov density fluctuations , @xmath104 , @xmath26 was considered by @xcite ; however , the central part of the signal @xmath143 predicted by their model did not differ much from the williamson case .
we do not know the analytic expression for the probability density function of @xmath1 .
however , it can be calculated numerically , by the method described in @xcite .
the generated pulse shapes for different numbers of screens , @xmath24 , seem to be universal when appropriately rescaled .
we will construct such shapes and will compare them with observations in section [ observation ] .
the scaling of the time delay @xmath1 , i.e. the approximate scaling of its probability density function , can be found from the following consideration . using the expression for the scaling of the sum of lvy distributed variables , @xmath154 ( recall that the central limit theorem does not hold )
, we derive @xmath155 , where the number of screens is proportional to the distance to the pulsar , @xmath156 . the observational scaling , @xmath6 ,
is reproduced for @xmath157 , and @xmath158 , which shows that ray - deviating density fluctuations do not have a gaussian distribution but rather a lvy distribution with the index @xmath159 .
this is the important result of the lvy model of scintillations , and it resolves the sutton paradox . in the section [ observation ] , we will demonstrate that the williamson paradox is resolved naturally in the lvy picture of scintillations as well .
but before that , in the next section we would like to present a physical model of scintillations that can be easily visualized and that demonstrates the essence of our approach .
the intensity function @xmath143 can be calculated for the phase - screen function ( [ levyscreen ] ) , once the parameters @xmath36 and @xmath15 have been specified from the physics of the problem . as an important example ,
let us consider a particular case of lvy phase screens , which we shall call cauchy screens , when the values of these parameters can be derived directly .
let us assume that the interstellar medium is filled with separated ionized regions such that all of them have the same electron - density contrast with the ambient medium , @xmath103 .
we also assume that these regions are randomly shaped and have sharp boundaries .
a propagating ray is then refracted by these random boundaries due to the snell s law , @xmath160 , where @xmath161 is the angle between the ray and the normal to the boundary , and @xmath162 is the refraction index . as we explained in the introduction ,
the interstellar parameters ensure that @xmath163 , and each refraction event results in a very small angle deviation , @xmath164 .
expanding @xmath165 through the first order in @xmath166 , we get from the snell s law @xmath167 where @xmath168 .
the distribution of @xmath169 can be found assuming ( somewhat artificially ) that the boundary normal is uniformly distributed over all directions in a half - space , at the point where it intersects the line of sight .
denote @xmath170 , where @xmath171 is a unit vector indicating the direction of the refraction in the plane perpendicular to the ray propagation .
then the distribution of @xmath172 is given in the polar coordinates by : @xmath173 quite remarkably , this distribution is the two - dimensional lvy distribution with the index @xmath48 , also known as the cauchy distribution .
its fourier transform follows from formula ( [ levy ] ) , @xmath174 , we leave the derivation of this result to the reader as an instructive exercise . to apply this result we need to calculate the screen function ( [ screen ] ) .
it can be done if we note that for the plane - like density discontinuity , the screen phase ( [ deltaphi ] ) is easily calculated , @xmath175= \lambda r_0 \delta n_0 \ , ( \bsigma\cdot { \bf y})$ ] .
since @xmath176 has the cauchy distribution ( [ cauchy ] ) , the averaged screen function can be found with the aid of formula ( [ levy ] ) , @xmath177 and we recover formula ( [ levyscreen ] ) with @xmath178 , and with @xmath47 , @xmath48 ; the parameters are close to those that we have proposed for the interstellar scintillation on the observational grounds , i.e. , @xmath157 , @xmath34 .
we expect that the observational scaling @xmath34 would correspond to more complicated , realistic structure of the boundaries , possibly , with the fractal dimension larger than 2 .
interstellar turbulence can indeed corrugate two - dimensional structures to make them have higher - than - two fractal dimensions , as is seen in numerical results and in observations of molecular clouds @xcite ; see also @xcite .
we expect this effect to have implications for interstellar scintillations , and we plan to consider it in the future work . an attentive reader has probably noticed that eq .
( [ theta ] ) for the angle deviation can not be valid for the arbitrarily large @xmath169 . indeed , the refraction angle @xmath179 can not exceed the critical angle , given by @xmath180 , which is obtained when @xmath181 , where we assume @xmath182 . to obtain the general formula for the angle deviation we need to expand @xmath183 in the snell s law up to the second order in @xmath166 , which gives @xmath184.\end{aligned}\ ] ]
the distribution of @xmath185 thus has a cut - off at @xmath186 .
for example , if the electron density fluctuations are @xmath187 , we have @xmath188 , which implies that formula ( [ theta ] ) holds in the broad range of scales .
indeed , the distribution of @xmath189 follows the cauchy law ( [ cauchy ] ) , therefore the distribution function of @xmath166 is strongly peaked at small @xmath166 .
this function has a maximum at @xmath190 , and a long asymptotic power - law tail , @xmath191 , that spans about six orders of magnitude before it reaches the cut - off at @xmath192 .
this consideration provides an illustration to the main idea of our approach .
the second moment of the distribution of @xmath166 depends not on the shape of the distribution function ( [ cauchy ] ) , but on the cut - off value of its asymptotic power - law tail , @xmath193 . on the contrary , the shape of the pulsar signal is determined by the full shape of the distribution function ( [ cauchy ] ) , and is practically independent of the large cut - off value @xmath193 .
in fig . ( [ levy_cauchy ] )
we present the pulse - broadening function , @xmath194 which is the distribution function of the time delay , @xmath1 , given by formula ( [ tau ] ) , for the standard gaussian model , @xmath195 , for the lvy - cauchy model , @xmath196 , and for the lvy model with @xmath47 , @xmath197 .
note that , in comparison with the pulse through a gaussian medium , the lvy models predict relatively greater intensity at short delays from the arrival time for the original pulse , as well as power - law tails at long delays .
these tails result from rare , large occurrences of scattering .
we generated the two - dimensional isotropic lvy distribution of @xmath198 using the methods of @xcite .
let @xmath199 be a gaussian random vector , i.e. both components @xmath200 and @xmath201 are independent , identically distributed , one - dimensional gaussian variables .
let @xmath202 be a completely skewed ( i.e. positive ) , independent lvy distributed variable corresponding to the lvy index @xmath203 .
this variable can be numerically generated using the method by @xcite , see also @xcite .
then the variable @xmath204 has an isotropic two - dimensional lvy distribution with the index @xmath42 ; the proof can be found , e.g. , in @xcite . to generate an anisotropic lvy distribution with uniform index @xmath42 ( as may be relevant for the magnetized interstellar medium ) one may choose an anisotropic gaussian vector * x * , and then proceed as above . in our present work
we consider only isotropic distributions .
we compared the predicted pulse shapes with the data recently published in @xcite and in @xcite .
out of the 76 pulsars analyzed by @xcite , we considered those with broad signals , in order to minimize the intrinsic - pulse effects .
good examples are provided by the four pulsars , p1849 + 0127 at 430 mhz ( dm=214.4 ) , j1852 + 0031 at 1175 mhz ( dm=680.0 ) , j1905 + 0709 at 430 mhz ( dm=269.0 ) , and p1916 + 0844 at 430 mhz ( dm=339.0 ) , where higher - frequency observations indicate that the intrinsic signal may be narrow .
the dispersion measures are given in the units of @xmath205 .
shapes of all of the pulses have a characteristic sharp rise and a narrow , pointed apex , which is inconsistent with the gaussian model , in exact agreement with the observations by @xcite [ see fig .
( 4 ) in williamson s paper ] . on the contrary
, our lvy model provides a good fit to such shapes . for another similar comparison we used the pulse shape of j1848 - 0123 ( dm=159.1 ) observed at 327 mhz by @xcite
; and we present this comparison in figs .
[ shapes2],[alpack ] as the illustrative example .
when making the comparisons , we tried to match the central parts of the observational and analytical curves .
however , this was possible only when we shifted the zero level of the analytical curves down , as is shown by the thin solid lines in fig .
( [ shapes2 ] ) .
this may be natural for `` over - scattered '' profiles , i.e. , when the signal has a long tail that does not approach zero during the pulse period .
this is exactly the case for our lvy model . besides
, other observational effects , including noise from the sky and the telescope itself , usually prevent one from determining the baseline unambiguously .
we note that for a precise comparison , we must convolve the calculated pulse - broadening function with the narrow intrinsic shape of the pulse , and with the response function of the receiving system @xcite .
this will lead to slight broadening of the predicted pulse profile , making the gaussian model even less consistent with the observations .
the effects of over - scattering on baseline have not been considered observationally .
this effect is important for comparison of our theory with observations , and we plan to investigate it in future work .
interferometry can distinguish between sky and telescope noise , and the overscattered tail of the pulse ; comparison of the correlated flux density at the peak of the pulse , and between pulses , thus provides a measure of the parameter @xmath42 in figures [ levy_cauchy ] and [ compare ] .
we are proceeding with additional observational tests that can distinguish among models for scattering .
as figures [ levy_cauchy],[compare ] suggest , different values of @xmath42 yield different delays between the arrival time for the unscattered pulse and the peak of the pulse , as well as different behavior in the tail .
observations at short wavelengths yield the arrival time for the unscattered pulse , with corrections for dispersion in the uniform interstellar plasma .
thus , multi - frequency observations of pulse broadening , with timing information , can distinguish among the different models .
shapes of scattered images provide another possible observational test of our theory .
as equation ( [ levyscreen ] ) suggests , the distribution of deflections of waves , at a given point , is drawn from a lvy distribution . because the lvy distribution is a stable attractor , iteration of the process yields a lvy distribution of directions , at any distance from the source .
compared with a gaussian distribution of similar width , a lvy distribution has excesses at both large and small angles : it produces large deflections more commonly , and compensates with a large population with small deflection .
thus , the scattered image for a lvy flight should have a halo at large angles , and a cusp at small angles , compared with a gaussian image .
interestingly , observations of some heavily - scattered extragalactic sources show such core - halo excesses @xcite . intrinsic structure of the source could possibly explain the observed halo structure ; but can not explain the relatively less - scattered cusp .
in the present paper , we have investigated the observational pulse shape .
another object directly related to the electron distribution function is the visibility function , @xmath206 , measured at the locations @xmath207 and @xmath208 at the earth .
however , the electron fluctuations probed by this function are limited to small scales , @xmath209 cm , which can be comparable to the inner scale of the turbulence @xcite where the statistics of the ism are different from the statistics responsible for pulse broadening .
the visibility function can be easily constructed in our multi - screen model by solving eqs .
( [ vequation],[freeequation ] ) for @xmath210 , however , we leave its analysis for future communication . here
, we estimate what restrictions would be imposed on our theory by an inner scale of the order @xmath211 . due to the formulas of section [ multi - screen ] ,
large - angle refraction is provided by small - scale fluctuations of electron density . in the case of lvy statistics , a sum of random variables can be dominated by a single term , therefore , from formula ( [ tau ] ) we can estimate @xmath212 . from formula ( [ intensity - non - dim ] ) , the electron - density fluctuations producing such an angle deviation should be strong at the scale @xmath213^{1/2}$ ] , or @xmath214 where we introduced the dispersion measure @xmath215 , and the signal frequency , @xmath216 .
quite conveniently , formula ( [ yin ] ) includes only the quantities averaged along the line of sight , but not the particular parameters @xmath87 , @xmath96 , @xmath103 , and @xmath36 of the scattering screens . as an example , let us consider the parameters for the large - dispersion - measure pulsar j1852 + 0031 ( dm=680.0 ) observed at 1175 mhz in @xcite . substituting @xmath217 , @xmath218 , @xmath219 , and assuming that @xmath220 , we obtain @xmath221 .
therefore , the condition @xmath222 is marginally satisfied . due to the scaling @xmath6
, smaller - dispersion - measure pulsars satisfy this condition even better . for the cases when this condition is violated ,
a more appropriate description will be given by `` truncated '' lvy distributions , see e.g. , @xcite .
in conclusion we would like to mention the observational correlation functions that are expressed through the higher - than - second moments of the wave amplitude , @xmath223 .
these observational quantities require a special consideration that at present is not available to us .
the examples of such quantities are given by the frequency correlation function that is used to characterize pulse broadening for @xmath224 , see , e.g. , @xcite , or by the intensity correlation functions that are used to characterize the density statistics on scales larger than the earth size @xcite .
the scaling of such functions may be easily found in the gaussian theory when these functions can be related to the second - order moment , @xmath225 .
however , the meaning of these correlation functions in the lvy case has yet to be determined .
for instance , it seems possible that the deviation of the pulse - width scaling from @xmath226 for the lower part of the elbow diagram , in fig .
( [ elbow ] ) , is due to the different method of reconstructing @xmath1 in this region .
indeed , below @xmath227 , @xmath1 is not measured directly , but rather is defined as the inverse frequency decorrelation bandwidth @xcite . to summarize , in the present paper we proposed a multi - screen model of the uniform , non - gaussian interstellar medium , and we found its analytic solution .
we demonstrate that this model explains both the shapes and the scalings of the observational pulse profiles , and it is free of the sutton and williamson paradoxes that could hardly be resolved in the standard gaussian theory of scintillations .
also , we provided a physical model of density distribution having random discontinuities , which produces lvy scintillations and which may serve as a good approximation for the analysis of the observational data .
we have demonstrated that the observational pulse shapes depend on the entire shape of the distribution of electron density , rather than simply on the second moment .
pulse profiles may therefore serve as a valuable tool for reconstructing this function from observations , and we have developed the corresponding method in the present paper .
we are very grateful to ramachandran rajagopalan for allowing us to use the observational results published in @xcite .
we would like to thank john nolan for his valuable advice on the method of numerical generation of symmetric stable distributions , and fausto cattaneo , arieh knigl , and robert rosner for important discussions .
s.b . would like to thank the aspen center for physics where a part of this work was done .
the work of s.b . was supported by the nsf center for magnetic self - organization in astrophysical and laboratory plasmas , at the university of chicago .
armstrong , j. w. , rickett , b. j. , & spangler , s. r. ( 1995 ) apj * 443 * , 209 .
bhat , n. d. r. , cordes , j. m. , camilo , f. , nice , d. j. , & lorimer , d. r. ( 2004 ) apj * 605 * , 759 .
bhat , n. d. r. , cordes , j. m. , & chatterjee , s. ( 2003 ) apj * 584 * , 782 .
blandford , r. & narayan , r. ( 1985 ) mnras * 213 * 591 .
boldyrev , s. & gwinn , c. r. ( 2003a ) apj * 584 * , 791 .
boldyrev , s. & gwinn , c. r. ( 2003b ) phys .
91 * , 131101 .
boldyrev , s. & gwinn , c. r. ( 2004 ) astrophysics and space science * 292 * , 147 .
boldyrev , s. ( 2002 ) apj * 569 * , 841 .
boldyrev , s. , nordlund , . , & padoan , p. ( 2002 ) apj * 573 * , 678 .
chambers , j. , mallows , c. , & stuck , b. , ( 1976 ) journal of american statistical association , _ theory and methods _ * 71*(354 ) , 340 - 344 .
chandran , b. d. g. & backer , d. c. , ( 2002 ) apj * 576 * , 176 .
cho , j. , lazarian , a. , & vishniac , e. ( 2000 ) apj * 564 * , 291 .
constantin , p. , procaccia , i. , & sreenivasan , k. r. ( 1991 ) phys .
* 67 * , 1739 .
cordes , j. m. , & lazio , t. joseph w. ( 2001 ) apj * 549 * , 997 .
desai , k.m .
, & fey , a.l . ( 2001 ) , apjs , * 133 * , 395 .
elmegreen , b. & falgarone , e. ( 1996 ) apj * 471 * , 816 .
feller , w. ( 1971 ) _ an introduction to probability theory and its applications _
( john wiley & sons , inc .
, new york ) , ch .
fiedler , r.l . ,
dennison , b.j . ,
johnston , k.j .
, & hewish , a. ( 1987 ) , nature , * 326 * , 675 .
goldreich , p. & sridhar , s. ( 1997 ) , apj * 485 * , 680 .
goodman , j. & narayan , r. ( 1985 ) mnras * 214 * 519 .
gupta , y. , rickett , b. j. , & coles , w. a. ( 1993 ) apj * 403 * , 183 .
gwinn , c. r. _ et al _ ( 1998 ) , apj * 505 * , 928 .
hirano , c. , gwinn , c. r. , & boldyrev , s. , ( 2002 ) , astro - ph/0204367 .
jayannavar , a. m. & kumar , n. ( 1982 ) phys .
lett . * 48 * , 553 .
kaspi , v. m. & stinebring , d. r. ( 1992 ) apj * 392 * , 530 .
klafter , j. , zumofen , g. , & shlesinger , m. s. ( 1995 ) , in _ lvy flights and related topics in physics _
( springer , berlin ) .
kraichnan , r. h. ( 1991 ) , phys .
67 * , 3634 .
kritsuk , a. g. & norman , m. l. ( 2004 ) apj * 601 * , l55-l58 .
lambert , h. c. & rickett , b. j. ( 2000 ) apj * 531 * , 883 .
lee , l. c. & jokipii , j. r. ( 1975a ) apj * 196 * , 695 .
lee , l. c. & jokipii , j. r. ( 1975b ) apj * 201 * , 532 .
lee , l. c. & jokipii , j. r. ( 1975c ) apj * 202 * , 439 .
lithwick , y. & goldreich , p. ( 2001 ) apj * 562 * , 279 .
meneveau c. & sreenivasan , k. r. ( 1990 ) phys .
rev . a * 41 * , 2246 .
nakao , h. , ( 2000 ) , cond - mat/0002027 .
nolan , j. p. , ( 2002 ) `` stable distributions .
models for heavy tailed data '' ( http://academic2.american.edu/ jpnolan / stable / stable.html ) .
phillips , j. a. & clegg , a. w. , ( 1992 ) , nature * 360 * , 137 .
ramachandran , r. , mitra , d. , deshpande , a. a. , mcconnell , d. m. , & ables , j. g. , ( 1997 ) mnras * 290 * , 260 .
rickett , b. j. ( 1977 ) ann .
. astrophys . * 15 * , 479 .
rickett , b. j. ( 1990 ) ann .
. astrophys .
* 28 * , 561 .
rickett , b. j. , lyne , a. g. , & gupta , y. ( 1997 ) mnras * 297 * , 739 .
romani , r. w. , blandford , r. d. , & cordes , j. m. ( 1987 ) nature * 328 * , 324 .
saul , l. , kardar , m. , & read , n. ( 1992 ) phys .
a * 45 * , 8859 .
scalo , j. & elmegreen , b. ( 2004 ) astro - ph/0404452 ; ann .
astrophys . , to appear .
shishov v. i. , _ et al _ ( 2003 ) a&a , * 404 * 557s .
spangler , s. r. & gwinn , c. r. ( 1990 ) apj * 353 * , l29-l32 .
stinebring , d.r . , mclaughlin , m.a . ,
cordes , j.m . , becker , k.m . ,
goodman , j.e.e . , kramer , m.a . ,
scheckard , j.l , & smith , c.t .
( 2001 ) , apj * 549 * , l97 .
sutton , j. m. ( 1971 ) mnras , * 155 * , 51 .
tatarskii , v. i. ( 1961 ) _ wave propagation in a turbulent medium _
( new york , mcgraw - hill ) .
uscinski , b. j. ( 1974 ) , proc .
london , a * 336 * , 379 .
williamson , i. p. ( 1972 ) mnras , * 157 * , 55 .
williamson , i. p. ( 1973 ) mnras , * 163 * , 345 .
williamson , i. p. ( 1974 ) mnras , * 166 * , 499 .
williamson , i. p. ( 1975 ) proc .
london , a * 342 * , 131 .
wolszczan , a. , & cordes , j.m .
( 1987 ) , apj , * 320 * , l35 . | radio waves propagating from distant pulsars in the interstellar medium ( ism ) , are refracted by electron density inhomogeneities , so that the intensity of observed pulses fluctuates with time .
the observed pulse shapes are used to diagnose electron - density distribution in the ism . the theory relating the observed pulse time - shapes to the electron - density correlation function
has developed for 30 years , however , two puzzles have remained .
first , observational scaling of pulse broadening with the pulsar distance is anomalously strong ; it is consistent with the standard model only when non - uniform statistics of electron fluctuations along the line of sight are assumed .
second , the observed pulse shapes are consistent with the standard model only when the scattering material is concentrated in a narrow slab between the pulsar and the earth .
we propose that both paradoxes are resolved at once if one assumes stationary and uniform , but _ non - gaussian _ statistics of the electron - density distribution in the interstellar medium
. such statistics must be of lvy type , and the propagating ray should exhibit a lvy flight rather than the gaussian random walk implied by the standard model .
we propose that a natural realization of such statistics may be provided by the interstellar medium with random electron - density discontinuities .
a lvy distribution has a divergent second moment , therefore , the standard approach based on the electron - density correlation function does not apply .
we develop a theory of wave propagation in such a non - gaussian random medium , and demonstrate its good agreement with observations .
the qualitative introduction of the approach and the resolution of the anomalous - scaling paradox was presented earlier in [ prl * 91 * , 131101 ( 2003 ) ; apj * 584 * , 791 ( 2003 ) ] . |
in recent years , remarkable progress has been made in finding the _ brightest _ ( @xmath9 mjy or @xmath10 mjy ) population of submillimeter ( submm ) galaxies ( smgs , e.g. , * ? ? ? * ; * ? ? ? * ) via square - degree scale surveys using far - infrared ( fir ) to millimeter ( mm ) single dish telescopes in space and on the ground .
thanks to their apparent high luminosity , usually with the aid of gravitational magnification , the brightest smgs offer a unique opportunity to investigate physical properties of the ism ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , resolved star - forming activity ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and gas dynamics ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) at the peak of the star - formation history of galaxies
. furthermore , the brightest smgs located in the _ high - redshift tail _ of the smg redshift distribution ( median @xmath113 , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) at @xmath126 provide a severe challenge to theories of galaxy formation and evolution @xcite .
the surface density of the brightest smgs holds integrated information on when galaxies undergo intense starburst and how frequently strong gravitational lensing occurs , on which galaxy formation models depend .
but , the source counts of brightest smgs in the high-@xmath3 tail are highly uncertain because of poor statistics , and it is indeed hard to find them ; only a single smg with @xmath10 mjy is expected within @xmath1310 deg@xmath14 from limited knowledge of current studies @xcite .
large - area surveys using _
herschel_/spire ( 250 , 350 , and 500 @xmath15 m ) have recently revealed strongly lensed galaxies bright at 250500 @xmath1 ( e.g. , * ? ? ?
* ; * ? ? ?
these surveys bring about a great change in our understanding of the statistical properties of the smg population .
the spire bands , however , preferentially select smgs at modest redshifts ( @xmath16 ) , and the study of smgs in the high-@xmath3 tail ( @xmath126 ) is still far from being complete .
the mm to long - submm wavelength cameras such as scuba , scuba-2 ( 850 @xmath1 , * ? ? ?
* ; * ? ? ?
* ) , mambo-2 ( 1.2 mm , * ? ? ?
* ) , aztec ( 1.1 mm , * ? ? ?
* ) , and laboca ( 870 @xmath1 , * ? ? ?
* ) are complementary to those fir to short - submm surveys ; for example , observations at longer wavelengths with the mambo and aztec cameras and the south pole telescope ( spt , * ? ? ?
* ) may exploit a uniform selection function in redshift space , allowing the efficient detection of the brightest smgs out to @xmath7 ( e.g. , * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we have used the aztec camera to carry out large - area surveys toward galactic star - forming regions , which cover @xmath17 deg@xmath14 of the sky in total with @xmath18 sensitivities of 530 mjy beam@xmath19 . from mm
/ submm number counts @xcite , several detections of ultra - bright ( @xmath10 mjy ) extragalactic sources at cosmological distances are expected within the survey area . this is complementary to earlier attempts to search for extremely - bright smgs and to constrain the brightest end of extragalactic number counts by exploiting submm maps from galactic surveys ( e.g. , * ? ? ?
* ) . in this paper
, we report the aztec detections and multi - wavelength analyses of two 1.1-mm bright sources , mm j154506.4@xmath0344318 and mm j154132.7@xmath0350320 ( hereafter mm - j1545 and mm - j1541 , respectively ; see figure [ fig : azteconspire ] ) .
these sources are found toward the lupus - i star - forming region , a local ( @xmath20 ) molecular cloud , and indeed the close proximity of mm - j1545 to the molecular cloud misled us to classify it as a starless core when it was initially identified with the aztec 1.1-mm camera .
multi - wavelength data collected by subsequent follow - up observations , however , strongly suggest that they are galaxies at cosmological distances as presented in this paper , illustrating the importance of multi - wavelength imaging and spectroscopy of such objects .
the paper is organized as follows .
section [ sect : obsdata ] describes observations and archival data obtained toward mm - j1545 and mm - j1541 . in section [
sect : results ] , we present the results from the observations and photometric redshift estimates .
section [ sect : discussions ] discusses the gravitational lensing effect on both objects , the brightest end of the 1.1 mm number counts , and the fir - to - mm colors of the mm - selected sources . in section [
sect : discussions ] , we also explore a possibility that mm - j1545 would be a galactic starless core .
finally , section [ sect : conclusions ] summarizes our conclusions . throughout this paper , we assume a concordance cosmology with @xmath21 , @xmath22 , and @xmath23 km s@xmath19 mpc@xmath19 . far - infrared to mm images of mm - j1545 ( left ) and mm - j1541 ( right ) .
the contours show the aztec 1.1-mm image , starting from @xmath24 with a separation of @xmath24 .
the background images show the _
herschel_/spire 250 , 350 , 500-@xmath1 pseudo - color images . ]
llcccc vla & 20 cm & ... & @xmath25 & mjy & 1 + vla & 6 cm & @xmath26 & ... & @xmath15jy & 2 + atca & 7 mm & @xmath27 & ... & @xmath15jy & 2 + nma & 2.7 mm & @xmath28 & ... & mjy & 2 + sma & 1.3 mm & @xmath29 & ... & mjy & 2 + aztec & 1.1 mm & @xmath30 & @xmath31 & mjy & 2 , 3 + sma & 890 @xmath1 & @xmath32 & ... & mjy & 2 + spire & 500 @xmath1 & @xmath33 & @xmath34 & mjy & 4 + & 350 @xmath1 & @xmath35 & @xmath36 & mjy & 4 + & 250 @xmath1 & @xmath37 & @xmath38 & mjy & 4 + mips & 24 @xmath1 & @xmath39 & @xmath40 & mjy & 5 + irac & 5.8 @xmath1 & @xmath41 & ... & @xmath15jy & 5 + & 3.6 @xmath1 & @xmath42 & ... & @xmath15jy & 5 + moircs & 2.15 @xmath1 & @xmath43 & ... & @xmath15jy & 2 + wfcam & 1.64 @xmath1 & @xmath44 & ... & @xmath15jy & 2 + & 1.26 @xmath1 & @xmath45 & ... & @xmath15jy & 2 + 2mass & @xmath46 & & @xmath15jy & 5 + & @xmath47 & & @xmath15jy & 5 + & @xmath48 & & @xmath15jy & 5 + dss & @xmath49 & & @xmath15jy & 5 + & @xmath50 & & @xmath15jy & 5 + & @xmath51 & & @xmath15jy & 5 llccc irac & 5.8 @xmath1 & @xmath52 & ... & @xmath15jy + & 3.6 @xmath1 & @xmath53 & ... & @xmath15jy + wise & 4.6 @xmath1 & @xmath54 & @xmath55 & @xmath15jy + & 3.2 @xmath1 & @xmath56 & @xmath57 & @xmath15jy + moircs & 2.15 @xmath1 & @xmath58 & ... & @xmath15jy + wfcam & 1.64 @xmath1 & @xmath59 & ... & @xmath15jy + & 1.26 @xmath1 & @xmath60 & ... & @xmath15jy + 2mass & @xmath46 & ... & @xmath61 & @xmath15jy + & @xmath47 & ... & @xmath62 & @xmath15jy + & @xmath48 & ... & @xmath63 & @xmath15jy + dss & @xmath49 & @xmath64 & @xmath65 & @xmath15jy + & @xmath50 & @xmath66 & @xmath67 & @xmath15jy + & @xmath51 & @xmath68 & @xmath69 & @xmath15jy + +
the 1.1-mm data were taken with the aztec bolometer camera @xcite installed on the atacama submillimeter experiment 10 m telescope ( aste , * ? ? ?
* ; * ? ? ?
* ) located at atacama ( altitude of 4860 m ) in the chilean andes , during august 2007 to december 2008 .
the weather conditions during the runs were excellent ; the typical 220-ghz zenith opacities were in the range of 0.010.08 .
aztec / aste provides an angular resolution of @xmath70 in full - width at half maximum ( fwhm ) .
the complete description of the lupus - i starless core survey will be reported elsewhere ( tsukagoshi et al . , in preparation ) .
the reduction procedure is described by @xcite and @xcite . to extract point - like sources ,
the time - stream data were intensively cleaned using a principal component analysis ( pca ) algorithm , and then mapped .
the pca cleaning works as a high - pass filter in the spatial frequency domain , and thus emission extended significantly compared with the beam solid angle is fully filtered out .
the fwhm of the point response function is @xmath71 after an optimal filter is applied to surpress high spatial frequency noises .
the pointing was checked every 1 hr using nearby radio quasars , resulting in the astrometric accuracy better than @xmath72 .
uranus and neptune were used for flux density calibration , yielding an absolute accuracy better than 10% .
the resulting root - mean - square ( r.m.s . ) noise over the mapped 4-deg@xmath73 region of the lupus - i cloud is 5 mjy beam@xmath19 .
we retrieve level-2 images of _ herschel_/spire from the _ herschel _ science archive ( hsa ) .
the spire data were obtained in 2011 january in parallel mode ( observation i d : 1342213182 ) and are processed through a pipeline software with a standard processing generation ( spg ) version of v8.2.1 .
the resulting 250500 @xmath1 images cover a @xmath74 region of lupus - i with typical @xmath18 noise levels of 515 mjy beam@xmath19 .
the details of the data are given by @xcite .
for near to mid - ir photometry , we refer to the public source catalogs obtained with the infrared array camera ( irac , * ? ? ? * ) and multiband imaging photometer for _ spitzer _ ( mips , * ? ? ? * ) onboard _ spitzer _ , wise @xcite and the two micron all sky survey ( 2mass , * ? ? ? * ) , which are available at the nasa infrared science archive ( irsa ) .
the _ spitzer _
photometric data are provided by the _ spitzer _
space telescope `` from molecular cores to planet - forming disks '' ( c2d ) legacy program @xcite .
in addition , we also retrieve the basic calibrated data ( bcd ) from the _ spitzer _ heritage archive ( sha ) in order to estimate the noise levels of the irac 3.6 and 5.8 @xmath1 and mips 24 @xmath1 photometry .
the bcd of irac and mips are processed through masking , flat fielding , background matching , and mosaicing using a standard single frame pipeline on the mopex software .
the resulting r.m.s .
noise levels at 3.6 , 5.8 and 24 @xmath1 are 0.2 , 1.5 @xmath15jy beam@xmath19 and 0.1 mjy beam@xmath19 , respectively .
the details on the _ spitzer _ observations and catalogs are reported in @xcite .
the nrao very large array sky survey ( nvss , * ? ? ? * ) and the digitized sky survey ( dss ) data are also available at the positions of mm - j1545 and mm - j1541 . for mm - j1545
, we used the submillimeter array ( sma , * ? ? ? * ) to measure its precise position and spatial extent , and then subsequently obtained multi - wavelength ancillary data that we describe below .
mm - j1545 was observed with the sma at 890 @xmath1 in 2010 january and at 1.3 mm in 2011 april .
the 890-@xmath1 observations were performed in the extended configuration with eight antennas , which provided projected baselines ranging from 15 to 171 meters .
the observing conditions were good ( 225-ghz zenith opacity of 0.05 ) .
the double side - band ( dsb ) receivers were tuned to a local oscillator ( lo ) frequency of 335.15 ghz , and the correlator provided 4-ghz band width in each sideband .
the 1.3 mm observations were carried out in the compact configuration .
the receivers were tuned so that @xmath75co , @xmath76co and c@xmath77o @xmath48 = 21 emission lines from the lupus - i cloud ( i.e. , @xmath78 ) , as well as 1.3 mm continuum emission of mm - j1545 , can be imaged .
the lo frequency was set to 224.86 ghz .
the atmospheric transparency was again good . in both observing runs , two quasars j1626@xmath0298 ( @xmath79 jy ) and j1454@xmath0377 ( 0.33 jy ) were used for complex gain calibration , while the passband response was calibrated using 3c273 and 3c279 .
the absolute flux density was scaled from the primary calibrator titan .
the accuracy of the flux calibration is estimated to be 15% .
all of the data editing and calibration were performed using idl - based standard routines in the mir software package .
the calibrated visibility data were imaged ( fourier - transformed ) and deconvolved using miriad @xcite tasks , ` invert ` and ` clean ` , respectively . in continuum
imaging , the upper ( usb ) and lower sideband ( lsb ) data , eliminating channels where the local molecular lines are expected , were combined before imaging .
the natural - weighted synthesized beam sizes at 335 ghz and 225 ghz are @xmath80 ( the position angle , pa = @xmath81 ) and @xmath82 ( pa = @xmath83 ) , respectively .
the resulting r.m.s .
noise levels at 335 and 225 ghz are 3.0 and 0.76 mjy beam@xmath19 , respectively .
the karl g. jansky very large array ( vla ) 6 cm data were taken in c - configuration in 2010 november ( project i d : 10c-226 ) .
the correlator was configured to provide @xmath84 mhz subbands covering from 4.2 to 6.1 ghz .
the data were calibrated using the vla calibration pipeline , which is based on the casa data reduction package @xcite .
imaging was also carried out using casa employing the multi - frequency synthesis algorithm ( spectral taylor expansion ) with ` nterms ` = 2 during the deconvolution , to take into account the spectral index of the sources within the field @xcite , along with briggs weighting ( ` robust ` = 0.5 ) .
the resulting synthesized beam is @xmath85 at position angle @xmath86 , and the r.m.s .
noise is 5.8 @xmath15jy beam@xmath19 .
the australian telescope compact array ( atca ) was used to take 7 mm continuum data in the h214 array configuration in 2013 october and in the h168 configuration in 2014 april ( project i d : c2910 ) .
we used five tunings to fully cover the 33.450.4 ghz band to search for a redshifted @xmath75co ( 21 ) line ( taniguchi et al .
, in preparation ) .
the center frequencies were set to 35.25 , 38.75 , 42.20 , 45.40 , 48.60 ghz , and two 2 ghz spectral windows of the cabb correlator were configured adjacently , resulting in an instantaneous frequency coverage of 3.9 ghz for each tuning .
the complex gain was monitored using a radio source b1541@xmath0375 ( @xmath87 jy , @xmath88 away from mm - j1545 ) .
bandpass and delay calibrations were performed using 3c 279 once per night before the observations started .
mars was used for absolute flux density calibration .
the absolute flux density uncertainty is estimated to be @xmath89 .
the data were calibrated using miriad and imaged using casa with briggs weighting ( robust parameter of 0.5 ) .
we did not use the five longest baselines including the antenna at the w392 station because of poor phase stability .
the resulting synthesized beam size and r.m.s .
noise level were @xmath90 ( pa = @xmath91 ) and 35 @xmath15jy beam@xmath19 , respectively .
we also use the nobeyama millimeter array ( nma ) at nobeyama radio observatory ( nro ) to constrain the 3 mm photometry .
the observations were done during 2010 april and may in the d configuration , where only five antennas were operational .
the dsb receivers were tuned at 98.20 ghz ( lsb , @xmath92 mm ) and 110.20 ghz ( usb ) , and the uwbc correlator @xcite with a 1-ghz bandwidth was used .
we did not use the usb data because of poor quality .
the visibility data were calibrated with uvproc - ii @xcite , and then imaged using the aips @xcite task , ` imagr ` . the resulting beam size and r.m.s .
noise level are @xmath93 ( pa = @xmath94 ) and 2.0 mjy beam@xmath19 , respectively .
since mm - j1545 is toward the edge region of a galactic molecular cloud , it is necessary to carefully investigate whether the source is indeed extragalactic .
we used the nro 45-m telescope to observe galactic @xmath75co ( 10 ) , @xmath76co ( 10 ) and c@xmath77o ( 10 ) emission lines .
the 45 m observations were performed during 2010 january to april .
the @xmath75co and @xmath76co observations were carried out with the on - the - fly ( otf ) mode of the multi - beam bears receiver @xcite and with the position switching mode of the t100 single - beam receiver @xcite , respectively .
the @xmath75co map covered a @xmath95 region including mm - j1545 and a part of the lupus - i cloud located north - east of mm - j1545 .
the @xmath76co otf observations covered a @xmath96 region centered on mm - j1545 . in both of the observations ,
the ac45 digital spectrometer was used . for the c@xmath77o observations
, we used the t100 single - beam receiver and the acousto - optical spectrometers with a high - dispersion mode ( aos - h ) in the standard position - switching mode , providing a spectral resolution of 0.10 km s@xmath19 at 110 ghz .
intensity calibration was done using the single - temperature chopper - wheel method , and the accuracy of intensity calibration is estimated to be 20% .
we carried out near - infrared ( nir ) imaging observations of mm - j1545 with the subaru telescope equipped with the moircs instrument @xcite on 2010 april 23 .
the observations were made with the @xmath46-band filter at @xmath97 = 2.15 @xmath1 , with a pixel scale of @xmath98 pix@xmath19 .
the total integration time was 2.9 ksec .
the position and magnitude were calibrated with several 2mass point sources with @xmath99 within @xmath100 .
the systematic uncertainties of the astrometry and magnitude are estimated to be @xmath101 and 0.07 mag , respectively .
_ jh_-band imaging observations were performed with the wide field camera ( wfcam , * ? ? ?
* ) attached to the united kingdom infrared telescope ( ukirt ) on mauna kea , during 2010 march 1519 .
these observations were complemented with the moircs @xmath46 and _ spitzer _ data .
the sky was photometric throughout these nights and seeing sizes were mostly @xmath102@xmath103 . a five - point dithering and a four - point micro - stepping
were used in all observations to compensate for bad pixels and to recover full point spread function sampling .
this results in proper sampling of the seeing size with the @xmath104 wfcam pixels .
each exposure time was 10 sec , yielding total integration times of 2.8 ksec and 2.4 ksec at @xmath48 and @xmath47 , respectively .
all of the data were reduced in a standard manner .
the astrometric uncertainties are less than @xmath101 , and the photometric errors at @xmath48 and @xmath47 are 0.18 and 0.32 mag , respectively .
mm - j1545 and mm - j1541 are detected at 1.1 mm at signal - to - noise ratios ( snr ) of 7.8 and 5.4 , respectively .
the flux densities of mm - j1545 and mm - j1541 , which are measured for the pca - cleaned map , are @xmath30 mjy and @xmath31 mjy .
we do not correct for possible flux boosting due to underlying fainter sources since it is difficult to estimate the bayesian prior from uncertain number counts at @xmath105 mjy .
they are , however , unlikely flux - boosted because such bright sources are extremely rare and , as we shall discuss in
[ sect : counts ] , the shape of the number counts are appear to be much flatter than those at @xmath106 mjy , where flux boosting is significant .
these two are amongst the four brightest ever discovered in the aztec / aste campaign , one of which is a @xmath107 mjy smg at @xmath108 reported by @xcite ( known as _ orochi _ or hxmm02 ,
* ) and the other is a 43-mjy source toward the peripheral field of the small magellanic cloud reported elsewhere @xcite . both of the sources toward the lupus - i cloud are also detected in the spire 250 , 350 , and 500-@xmath1 bands , most of which have flux densities of @xmath109100-mjy . in this respect , both are very similar in fir - to - mm brightness to the extremely - luminous smgs detected in the _
herschel_-atlas @xcite and hermes surveys @xcite .
the coordinates and results of photometry are listed in table [ tab1_multiwavelength ] .
the postage stamp images are given in figure [ fig : stamp ] .
an important point that we must note is that the sources are found toward a galactic ( @xmath20 ) molecular cloud .
multi - wavelength data require an extremely cold ( @xmath110 k ) and compact ( @xmath111 ) nature of the sources , which is too rare among galactic star - forming objects and thus unlikely associated with the molecular clouds , but are located at cosmological distances .
the spatial profiles are all consistent with a point - like source with the aztec ( @xmath112 ) and spire ( 2030@xmath113 ) beams , unlike starless cores found in galactic molecular clouds , which are typically found to be @xmath114 pc corresponding to @xmath115 at the distance to the lupus - i cloud ( e.g. , * ? ? ?
the spire photometry places a constraint on the peak positions of dust spectral energy distributions ( seds ) at @xmath116400 @xmath1 , which indicates the effective dust temperatures of @xmath117 k. given the dust temperatures found in starburst galaxies ( @xmath118 k ) , this suggests that mm - j1545 and mm - j1541 are situated at @xmath119 .
we will perform photometric redshift estimates in
[ sect : photoz ] . below is a summary of the individual sources .
+ the sma interferometric observations at 890 @xmath1 and 1.3 mm confirms the exact position at the j2000 equatorial coordinate of ( @xmath120 ) = ( @xmath121 ) with the uncertainties of @xmath122 and @xmath123 at 890 @xmath1 and 1.3 mm , respectively ( see below for estimation of positional uncertainties ) . the flux densities at 890 @xmath1 and 1.3 mm are @xmath32 mjy and @xmath29 mjy , respectively .
the 890 @xmath1 image is resolved with its @xmath124 beam , and the beam - deconvolved source size fitted with a single 2-dimensional gaussian using a miriad task ` uvfit ` is @xmath125 ( pa = @xmath126 ) .
figure [ fig : uvplot ] shows the visibility amplitudes versus projected baseline length ( i.e. , the fourier transform of a radial profile as a function of spatial frequency ) for mm - j1545 .
the fourier components are well expressed by a single gaussian ( @xmath127 in fwhm ) with a constant offset ( @xmath128 mjy ) , implying a cusp - like compact structure embedded in a extended ( @xmath129 ) component . at the sma position ,
vla 6 cm and atca 7 mm emission is also detected at @xmath130jy and @xmath131jy , respectively , but it is not resolved with the vla and atca beams .
we do not detect 2.7 mm emission in the nma image , which provides a @xmath132 upper limit of 5.9 mjy .
the counterpart was not detected in the mips 24 @xmath1 down to the @xmath132 limiting flux density of 0.3 mjy . in figure
[ fig : photoz ] , we fit the sed from the infrared to the radio with a single - component modified blackbody ( or graybody ) , @xmath133 , where @xmath134 is the planck function and @xmath135 is the dust absorption coefficient which follows a power - law function of frequency @xmath136 . the fir - to - mm part of the sed is well described with a single modified blackbody with an effective temperature of @xmath137 k and the emissivity index of @xmath138 .
however , the 6-cm flux clearly exceeds the gray - body function .
the spectral index over the 6 cm band is @xmath139 ( the error is the 1@xmath140 confidence interval ) , where @xmath141 .
this is rather flat compared with the rayleigh - jeans slope of @xmath142 , suggesting that the 6-cm flux arises from synchrotron and/or free - free emission . in the moircs ,
wfcam and irac images , a faint ( @xmath143 in the vega magnitude system , @xmath144jy ) source is detected at @xmath145 east of the sma 890-@xmath1 peak .
hereafter we refer to this nir object as j1545b .
table [ tab2_lens ] lists the optical to nir photometry of j1545b .
figure [ fig : astrometry ] shows the centroids and uncertainties of multi - wavelength counterparts detected in the nir ( 2.15 @xmath1 , 3.6 @xmath1 ) and the submm to centimeter ( 890 @xmath1 , 1.3 mm , 7 mm , 6 cm ) .
the positional uncertainty of each counterpart is estimated by adding statistical and systematic errors in quadrature .
the statistical error is obtained from @xmath146 , where @xmath147 and snr are the beam fwhm and signal - to - noise ratio .
given the small positional uncertainty of the sma and moircs ( @xmath148 ) , this offset is significant and it is unlikely that j1545b is a counterpart to mm - j1545 .
the fwhm of the moircs image of j1545b is the @xmath149 , whereas that of a nearby star ( point spread function , psf ) is @xmath150 , suggesting that j1545b is intrinsically extended ( a psf - deconvolved size of @xmath151 ) and likely a low-@xmath3 galaxy along the line of sight toward mm - j1545 .
this close association of a foreground galaxy may gravitationally magnify the background object , which can naturally explain the extreme flux density of mm - j1545 .
the possible gravitational lensing will be discussed in
[ sect : gravlens ] furthermore , we do not detect the @xmath48 = 21 and 10 transitions of c@xmath77o toward mm - j1545 , as shown in figure [ fig:45obs ] . from this we put a meaningful constraint on a molecular mass of a possible galactic dense gas core , suggesting that mm - j1545 is not of galactic origin .
the @xmath48 = 21 and 10 transitions of c@xmath77o trace molecular gas with @xmath152 and @xmath153 @xmath154 , respectively , and universally seen associated with galactic starless cores .
the @xmath132 upper limits on the main - beam temperature with a velocity resolution of 0.5 km s@xmath19 are @xmath155 k ( c@xmath77o @xmath1561 ) and @xmath157 k ( c@xmath77o @xmath1580 ) , yielding an upper limit to the dense gas mass under the local thermodynamical equilibrium ( lte ) of @xmath159 from the c@xmath77o ( 10 ) constraint . this lte mass is much smaller than those found in galactic starless cores
. no compact @xmath75co nor @xmath76co emission is significantly detected at the sma continuum position although ambient molecular gases are contaminated across the field of view of the 45-m and sma maps .
the absence of high - velocity components in @xmath75co spectra , which trace molecular outflows from an accreting protostellar system , rules out any protostellar phases .
the 6 cm emission is also critical to judge if the object is extragalactic ; starless cores have neither synchrotron nor free - free emission unlike galaxies .
the clear excess to the gray - body defined at 250 @xmath1 to 7 mm and the rather flat spectral index in the 6 cm band excludes the possibility that the 6 cm signal is dominated by dust emission from a galactic starless core .
thus , it is natural to suppose the object to be extragalactic .
we will further discuss on the galactic possibility in
[ sect : fc ] .
mm - j1541 is also unlikely a galactic source , because it is well isolated from the main clouds ( @xmath160 ) of lupus - i and meets the `` off - cloud '' criterion defined by @xcite .
it is significantly detected at 24 @xmath1 with _ spitzer _
( snr @xmath161 6 ) at @xmath162 = ( @xmath163 ) , all consistent with a @xmath164 smg , although no irac data are available .
a small enhancement ( @xmath165 , 1.3 mjy ) at 20 cm is seen in the nvss 20-cm image .
as shown in figure [ fig : photoz ] , the fir - to - mm part of the sed is well described by a single component gray - body with an effective dust temperature of @xmath166 k , where we assume a dust emissivity index of @xmath167 because of the limited number of photometric data points .
a bright nir / optical source is detected in 2mass @xmath168 and dss @xmath169 images at @xmath162 = ( @xmath170 , @xmath171 ) , which is @xmath72 south - west of the 24 @xmath1 centroid .
we hereafter refer to this 2mass / dss object as j1541b .
we also find a 3.2 and 4.6 @xmath1 object in wise data at this position .
unfortunately , we have no higher resolution images at 24 @xmath1 or longer wavelengths , at which the emission likely comes from mm - j1541 . but
given that the 24 @xmath1 source is detected at high snr and is likely a counterpart to mm - j1541 , we can place a constraint on its position ; the statistical uncertainty is estimated to be @xmath172 while the systematic error is @xmath173 , yielding a total positional uncertainty of 24 @xmath1 image of approximately @xmath174 .
the positions of j1541b measured in six 2mass / dss bands coincide with each other , and the astrometric accuracy is estimated to be better than @xmath175 for j1541b .
therefore , the offset between the 24 @xmath1 peak and the 2mass / dss position is significant at a @xmath176 level .
furthermore the shape of the sed is consistent with that of a low-@xmath3 passive elliptical as discussed later in
[ sect : gravlens ] , suggesting that j1541b may be a galaxy which lenses the background smg mm - j1541 at @xmath177 .
we list flux densities of j1541b in table [ tab2_lens ] and will further discuss the possible lensing in
[ sect : gravlens ] .
we fit the sed models to the photometric data points at @xmath178 to constrain their photometric redshifts and fir luminosities .
we use sed templates of well - characterized starburst galaxies ; arp 220 , m 82 , ( grasil , * ? ? ?
* ) , a composite of 76 radio - identified smgs with spectroscopic redshifts @xcite , and smm j2135@xmath00201 ( the cosmic eyelash , * ? ? ?
* ) to search for minimum @xmath179 by simply varying the redshifts and luminosities of the sed templates .
we consider 20% of an absolute flux density uncertainty for all photometric points in addition to the statistical error .
we also take into account the @xmath18 confusion noise @xcite for photometric errors of the spire bands .
figure [ fig : photoz ] shows the results of photometric redshift estimates and the best - fit seds .
the inferred redshift of mm - j1545 is @xmath1805 , although the derived photometric redshifts depends on the templates .
overall , the measured sed is in good agreement with the arp 220 and smm j2135 templates although the fit with m 82 is poorer than the others .
the best - fit redshifts and 68% confidence intervals are @xmath181 ( arp 220 ) , @xmath182 ( m 82 ) , @xmath183 ( average - smg ) and @xmath184 ( smm j2135 ) .
the inferred fir luminosities for the best - fit redshifts are then @xmath185 ( arp 220 ) , @xmath186 ( m 82 ) , @xmath187 ( average - smg ) , and @xmath188 ( smm j2135 ) .
the dust temperature is @xmath18940 k if @xmath1904.7 , which is similar to those found in smgs ( e.g. , * ? ? ?
the redshift of mm - j1541 is estimated to be @xmath191 , although the available photometric data are very limited . among the templates ,
arp 220 better reproduces the actual sed than the others , and the 24 @xmath1 detection is well accounted for by the 7.7-@xmath1 feature of polycyclic aromatic hydrocarbons ( pahs ) .
the photometric redshifts obtained the templates are @xmath192 ( arp 220 ) , @xmath193 ( m 82 ) , @xmath194 ( average - smg ) and @xmath195 ( smm j2135 ) .
the inferred fir luminosities for the best - fit redshifts are then @xmath196 ( arp 220 ) , @xmath197 ( m 82 ) , @xmath188 ( average - smg ) , and @xmath198 ( smm j2135 ) .
if we consider @xmath199 , then the dust temperature is approximately 34 k , again consistent with those found in smgs .
the extreme luminosities of mm - j1545 and mm - j1541 are likely attributed to galaxy - galaxy lensing ; in many cases , an elliptical galaxy seen in the optical to nir is associated with brightest ( @xmath200 ) smgs .
interferometric imaging of these bright sources often reveals multiply split images or extended structures that are well explained by strong gravitational lensing models ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in addition , detailed modeling of the mm / submm source number counts suggests that the excess of the counts at high flux densities ( @xmath105 mjy for example ) are dominated by strongly lensed smgs as well as nearby galaxies @xcite .
the nir sources , j1545b and j1541b , may be lensing objects .
while the lack of multiband optical photometry in j1545b makes it difficult to determine the lensing properties , the clear 2mass / dss detections of j1541b allow us to estimate the magnification factor and thus the intrinsic nature of mm - j1541 .
hence , we focus on the lensing property of the mm - j1541j1541b system first , and then discuss the mm - j1545j1545b system . to characterize the properties of j1541b , we perform sed fits to the optical - to - nir photometric data of j1541b using the hyperz code developed by @xcite . in sed fits , we used sed templates of @xcite . from the spire @xmath201 map ,
the visual extinction toward mm - j1541 is approximately 1.5 , but the stray light from the moon in the spire observations @xcite can bias against low visual extinction .
we assume a conservative value of @xmath202 .
note that this assumption does not dramatically affect the result because the total extinction is considered by combining the galactic reddening and the intrinsic extinction , which eventually compensates the uncertainty in the galactic extinction .
figure [ fig : sed1541b ] shows the result of sed fits .
the photometric redshift ( photo-@xmath3 ) of j1541b is @xmath203 ( @xmath204 , the error bar is from the 68% confidence interval ) .
the optical to nir sed is well described by a 33 myr elliptical model with a stellar mass of @xmath205 and an intrinsic extinction of @xmath206 .
the @xmath51-band decrement can be accounted for by the 4000 break at @xmath207 , which is well within the photo-@xmath3 range .
the best - fit stellar age is young ( 33 myr ) but maturer stellar seds can reasonably match the actual sed as well .
then we make use of the faber - jackson relation ( fjr , * ? ? ?
* ) to constrain the velocity dispersion of the lensing galaxy , which allows us to model a singular isothermal sphere ( sis ) as a lensing dark halo .
fjr has originally been proposed as an empirical relation between @xmath51-band absolute magnitude and velocity dispersion of galaxies .
similar relations are now confirmed in many filter bands out to the nir @xcite . in our case ,
the reddening can be significant because of the foreground galactic molecular cloud , so we use the nir version of fjr presented in eq . 5 of @xcite .
the einstein radius of a sis mass distribution is expressed as @xmath208 where @xmath209 is the velocity dispersion , @xmath210 and @xmath211 are the angular diameter distances from the observer to the background source and from the foreground lens to the background source , respectively . in the redshift range of @xmath212 , which covers the 68% redshift confidence interval of j1541b , we find the velocity dispersion ranging @xmath213 km s@xmath19 from the @xmath46 magnitude and fjr .
when we consider the background smg at @xmath214 ( [ sect : photoz ] ) , then we find @xmath215@xmath216 , which is smaller than the actual separation between j1541b and the 24 @xmath1 peak .
we estimate the magnification factor @xmath217 at the 24 @xmath1 position ( @xmath72 apart from the lens centroid ) using a gravitational lensing model glafic @xcite , in which we employ a circular - symmetric sis with the velocity dispersion @xmath209 derived above . for @xmath218 and @xmath219
, we find @xmath220 and the magnification factor at the position of mm - j1541 , @xmath221 .
if we go to @xmath222 that is the edge of the 68% confidence interval , then we have @xmath223 and @xmath224 . to get the einstein radius closer to the 3@xmath113 separation angle
, it would be necessary to double the angular diameter distance to the lens and/or increase the luminosity of the lens galaxy by a factor of @xmath225 , which is very difficult to achieve within the uncertainties in the measured quantities .
we have another likelihood peak at @xmath226 ( @xmath227 , see figure [ fig : sed1541b ] ) , but this is unlikely because the inferred stellar mass is too high ( @xmath228 ) . even if this would be the case , the magnification still remains a moderate value of @xmath229 .
consequently , the magnification factor of mm - j1541 is not likely as high as @xmath230 but rather moderate ( @xmath231 ) , suggesting that mm - j1541 might be an intrinsically hyper - luminous star - forming galaxy with a demagnified 1.1-mm flux density of @xmath23220 mjy or the intrinsic fir luminosity of @xmath23313.9 .
on the other hand , amplification for mm - j1545 at @xmath1805 may be larger although measurement of its magnification using existing data is difficult .
the fourier analysis of the sma visibility data of mm - j1545 clearly shows its extended morphology ( fwhm @xmath234@xmath216 , see figure [ fig : uvplot ] ) , suggesting that the source could be magnified by a foreground galaxy seen as j1545b .
unfortunately , we only have the limited number of photometric data points for j1545b , which likely suffer from relatively large galactic reddening of @xmath235 .
we find no apparent spectral break at @xmath236 , suggesting the lens redshift of @xmath237 .
so we assume several lens redshifts and a source redshift ( @xmath238 , from [ sect : photoz ] ) to estimate the einstein radius in the same manner as mm - j1541 .
note that a rest - frame @xmath46-band extinction of 0.2 mag is used to correct the galactic reddening .
we find @xmath239 at @xmath240 , which is inconsistent with our sma image showing the smaller separation ( @xmath145 ) and no counter - images split by a gravitational lens .
this suggest the lens redshift of @xmath241 .
actually , if the lensing galaxy is at @xmath242 , we have the absolute magnitude in rest - frame @xmath46-band of @xmath243 and @xmath244170 km s@xmath19 from the fjr .
this yields the einstein radius of @xmath245 , consistent with the observed situation .
unfortunately , we can not exactly predict the magnification factor of mm - j1545 ; the sma 890-@xmath1 image exhibits neither multiple counter - images nor a large arc / ring , and the spatial extent of the source is unknown .
furthermore , the accurate redshifts of the lens and source are not available .
all of the facts prevent us from precisely constructing a lens model .
however , given that no lensed source with a magnification of @xmath246 showing a single image with a @xmath124-beam has been reported thus far ( e.g. , * ? ? ?
* ) , it should be reasonable to assume @xmath247 as an upper limit for the magnification factor of mm - j1545 .
in this case , the intrinsic fir luminosity is still very high ( @xmath248 ) , even after correcting for magnification .
such a starburst galaxy in the hyperluminous regime at @xmath1805 is still rare compared to existing studies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and is a unique laboratory to investigate properties of star - formation in the early universe .
the detections of ultra - bright sources allow us to constrain the brightest end of the 1.1 mm number counts , which is complementary to the deep number counts at @xmath24920 mjy obtained from our own 1.1-mm surveys of smgs @xcite .
the area where the @xmath18 sensitivities of the lupus - i aztec map are below 7 mjy beam@xmath19 ( typically 5 mjy beam@xmath19 ) is 3.65 deg@xmath73 .
we eliminate the high column region where @xmath250 mag on the 2mass extinction map @xcite , which leaves 2.88 deg@xmath73 .
we detect three @xmath251 point sources over the 2.88 deg@xmath73 area ( tsukagoshi et al . , in preparation ) .
we carefully cross - identify known starless and protostellar cores @xcite , which leaves only two extragalactic sources , mm - j1545 and mm - j1541 .
the inferred cumulative number counts @xmath252 mjy ) for lupus - i are @xmath253 deg@xmath6 ( the error is taken from the @xmath18 confidence interval computed by * ? ? ?
* ) . furthermore , two additional bright sources ( @xmath254 mjy and @xmath255 mjy , * ? ? ?
* ; * ? ? ?
* ) have been found to date over the aztec survey fields ( 1.60 deg@xmath73 , * ? ? ? * ) and the smc peripheral field ( 1.21 deg@xmath73 , * ? ? ?
* ) , respectively .
both of them are also indicative of strongly - lensed magnification .
taking into account all of these four brightest extragalactic sources with @xmath256 mjy , we estimate the cumulative number counts at @xmath257 , 32.5 , 40 mjy are @xmath5 , @xmath258 , and @xmath259 deg@xmath6 , respectively , where the @xmath18 errors are again taken from @xcite .
figure [ fig : counts ] shows the 1.1-mm cumulative number counts of the brightest aztec sources , as well as the deeper 1.1-mm number counts obtained from blank fields toward the great observatories origins deep survey ( goods ) north and south fields , the lockman hole , the cosmic evolution survey ( cosmos ) , the subaru/_xmm newton _ deep field , the akari deep field - south @xcite and from the aztec cluster environmental survey ( aces ) @xcite .
these counts intersect at @xmath260 mjy , but the slope at @xmath261 mjy is shallower than the deeper part of the source counts .
we also plot the 1.4-mm counts obtained by the south pole telescope ( spt ) survey @xcite scaled to equivalent 1.1-mm flux densities using a scaling factor of @xmath262 @xcite .
the spt counts do not include any nearby galaxies discovered by the infrared astronomical satellite ( iras ) and those with synchrotron - dominated seds .
alternatively , these are thought to be dominated by strongly - lensed smgs .
the amplitude and slope at the brightest end of the 1.1-mm counts are consistent with those of the spt counts given the uncertainty in the scaling factor .
furthermore , a model prediction by @xcite , where flux magnification due to strong gravitational lensing is accounted for , reproduces the 1.1-mm counts at @xmath256 mjy , providing more supporting evidence that the brightest sources found at 1.1 mm are mostly attributed to the strong - lensing effect .
it is increasingly becoming clear that mm to long - submm observations are more likely to select higher redshift smgs than fir and short - submm as predicted earlier @xcite , and recent studies of ultra - bright sources selected in the mm to long submm actually reveal a lot of smgs at redshift @xmath263 out to @xmath264 @xcite .
it has been also suggested that selecting red objects in spire bands whose seds are peaked at the 500-@xmath1 band ( i.e. , objects that follows @xmath265 ; so - called 500-@xmath1 peakers ) is also a useful way to pick up high-@xmath3 candidates even at @xmath266 @xcite .
mm - j1545 is estimated to be at @xmath1805 and is formally consistent with a 500-@xmath1 peaker , implying that 1.1-mm selection and ( sub)mm color investigation are quite useful for isolating @xmath267 smgs . in the left side of figure [ fig : ccd ]
, we plot the @xmath268 and @xmath269 flux density ratios of mm - j1545 and mm - j1541 . for comparison , we also show the same color color plots of 48 aztec - selected sources with s / n @xmath270 from the aztec / aste goods - south survey @xcite , twelve of which have robust counterparts with spectroscopic redshifts ranging from @xmath271 to 4.76 @xcite .
herschel _ data are retrieved from the hsa and the @xmath18 noise levels are @xmath2720.6 mjy beam@xmath19 in all the spire bands .
the detection thresholds are set to @xmath273 = 7.6 , 9.2 , 10.4 mjy beam@xmath19 at 250 , 350 , and 500 @xmath1 , respectively , where @xmath274 is an underlying confusion limit after removing bright spire sources @xcite .
forty aztec sources are detected at 500 @xmath1 while 8 aztec sources do not have a significant counterpart in any of the spire bands .
the redshift tracks of a modified black - body with @xmath275 k and typical starburst galaxies are also overlaid .
mm - j1545 is situated in a region of the plot where @xmath2765 galaxies are actually observed or are expected from redshift tracks of sed models .
similarly , the 1.1-mm to 350-@xmath1 color of mm - j1541 is consistent with those found in @xmath277 galaxies . in the right panel of figure [ fig : ccd ] , we show the @xmath278 and @xmath279 plots for mm - j1545 and mm - j1541 , as well as the 40 aztec - selected sources in goods - south which are detected at 500 @xmath1 . despite large uncertainties in flux density ratios , at least 16 of the goods - south sources ( 40% )
are consistent with 500-@xmath1 peakers that are detected at least at both 350 and 500 @xmath1 , and up to 21 sources ( 53% ) may be 500-@xmath1 peakers if we include sources only detected at 500 @xmath1 .
at least 3 sources meet the criterion , @xmath280 , which is used to select @xmath281 candidates by @xcite .
the fir - to - mm color of mm - j1545 is overall consistent with those of the 500-@xmath1 peakers but slightly bluer than the @xmath281 criterion , while the color of mm - j1541 is in good agreement with those of @xmath177 sources , supporting the redshift estimates discussed in
[ sect : photoz ] . as demonstrated by these color
color plots , 1.1-mm selected sources that are very red in the ( sub)mm will offer a unique opportunity to investigate how frequently massive starbursts are triggered in the @xmath263 universe , which places constraints on galaxy formation models for massive dusty starbursts .
future follow - up studies using the atacama large millimeter / submillimeter array ( alma ) are needed to investigate this further .
how to securely distinguish between smgs and prestellar cores is always an issue in identifying smgs behind galactic molecular clouds , because flux densities and fir - to - submm colors of the brightest smgs looks similar to those found in the low - luminosity end of prestellar cores .
while mm - j1541 is well isolated from the lupus - i molecular cloud ( @xmath282 ) and thus likely an extragalactic source , mm - j1545 is closer to the molecular cloud ( @xmath235 , see also figure [ fig : azteconspire ] ) and is worthy of assessing the possibility that mm - j1545 is a galactic source .
we hereafter assume a distance to the lupus - i cloud of @xmath283 pc @xcite , where a @xmath124 size corresponds to the physical scale of @xmath284 pc or 150 au .
the concordance scenario of low - mass star formation ( * ? ? ? * for a review ) begins with a collapse of a sub - pc - sized gravitationally - bound molecular core , a so - called prestellar core .
prestellar cores are initially optically - thin to the thermal dust emission and isothermally collapses , and thus it is represented as a cold ( @xmath285 k , e.g. , * ? ? ?
* ) and moderately dense ( @xmath286(h@xmath287 ) @xmath288 @xmath154 ) isothermal system .
most of them are found in regions where high h@xmath287 column densities of @xmath289 cm@xmath6 are observed ( e.g. * ? ? ?
* ) . the prestellar phase is followed by protostellar phases ( a.k.a .
, classes 0 , i , ii , and iii , * ? ? ?
* ; * ? ? ?
* ) , in which mid - ir and/or nir emission powered by a central protostar is always visible .
a hypothetical small adiabatic region that is extremely - dense ( @xmath29110@xmath76 @xmath154 ) and compact ( @xmath1310@xmath73 au ) , a so - called first hydrostatic core @xcite , may occur at the center of a prestellar core in its final stage just before protostar formation , although the first core is expected to be short - lived ( @xmath292 yr ; e.g. , * ? ? ?
* ; * ? ? ?
* ) and the observational nature is still controversial ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
mm - j1545 is a very unusual source if this would be associated with the lupus - i cloud .
the stark constraint on 24-@xmath1 flux density ( @xmath39 mjy ) , the absence of a bright compact nir source _ at the position of _ the sma source or extended reflection nebulosity ( as seen in edge - on herbig - halo objects like hh30 ) , the low ( effective ) dust temperature of @xmath293 k ( [ sect : mmj1545 ] ) , and the non - detection of @xmath75co outflows virtually rule out the protostellar class-0 and advanced phases , suggesting that mm - j1545 might be in the prestellar phase and thus the mass should be dominated by dense molecular gas .
c@xmath77o @xmath48 = 10 and 21 emission lines toward this source ( i.e. , @xmath20 ) were , however , not detected with the sma and the nro 45 m telescope ( figure [ fig:45obs ] ) , which places a constraint on molecular gas mass of @xmath294 ( [ sect : mmj1545 ] ) .
in contrast , the 1.1 mm flux density of 44 mjy yields @xmath295 if this is a starless core [ @xmath296 k , @xmath297 g@xmath19
cm@xmath73 @xcite with @xmath299 , and the dust - to - gas mass ratio of 100 are assumed ] .
this discrepancy between gas and dust masses indicates that mm - j1545 is not a prestellar core .
furthermore , the sma continuum observations at 890 @xmath1 and 1.3 mm revealed a very compact source , compared to usual prestellar cores ( [ sect : individual ] ) .
as shown in figure [ fig : uvplot ] , its visibility distribution does not show any evidence for an extended structure like a dense gas envelope , which is at least @xmath2320.01 pc and typically @xmath114 pc in size , corresponding to an angular size of @xmath115 ( e.g. * ? ? ?
* ) . in figure
[ fig : uvplot ] , we show fourier transform of 890-@xmath1 brightness distribution predicted for a critical bonnor
ebert sphere @xcite as a realization of a starless core in the lupus - i cloud .
the gas temperature and central h@xmath287 density is assumed to be 7.1 k and @xmath300 @xmath154 , respectively .
the total gas mass of @xmath301 is chosen so that the total 890-@xmath1 flux density matches the observed one .
the overall visibility slope and amplitude at higher spatial frequencies are , however , largely deviated from the prediction of the bonnor ebert model .
a possible galactic interpretation of mm - j1545 might be a first hydrostatic core @xcite in its very early phase , because its compact nature and the very cold ( @xmath302 k ) and low - luminosity ( @xmath303 ) sed are consistent with those expected in a mostly - naked first core @xcite . in this case ,
c@xmath77o emission line is optically thick and the non - detection of c@xmath77o is explained if a significant fraction of gas mass , which is inferred from the dust sed ( @xmath304 ) , might be stored behind the c@xmath77o photosphere . in figure
[ fig : uvplot ] , we show the fourier components of a 890-@xmath1 brightness distribution predicted for a first hydrostatic core , which is computed by radiation hydrodynamic simulations @xcite .
the first hydrostatic core is produced by gravitational collapse of a @xmath305 bonnor ebert sphere , and the mass and inclination angle are chosen so that the overall visibility amplitudes match the observed data .
consequently , the unresolved cusp - like structure appeared at @xmath306 k@xmath97 and the extended envelope predicted from the first core model are consistent with the actual sma visibility amplitudes .
note that the adjacent compact nir source , j1545b , and the observed jvla 6 cm flux density at the sma position can not be explained consistently with the other data even in the first hydrostatic core models .
we also note that first hydrostatic cores should be extremely rare because of the short lifetime ; only one first core out of @xmath2321001000 dense starless cores is expected , if comparing the lifetime with the dynamical time scale of starless cores ( @xmath2320.11 myr ) .
we therefore conclude that the @xmath1805 lensed smg is the most likely and naturally explained by the sed and geometry / spatial extent of the multi - wavelength counterpart to mm - j1545 .
we report serendipitous detections of two 1.1-mm bright extragalactic sources , mm j154506.4@xmath0344318 ( mm - j1545 ) and mm j154132.7@xmath0350320 ( mm - j1541 ) , in the aztec / aste survey of the lupus - i star - forming region .
these sources are likely located at cosmological distances at @xmath1805 ( mm - j1545 ) and @xmath191 ( mm - j1541 ) .
the main results are as follows : 1 .
mm - j1545 is the brightest ( @xmath307 mjy ) of @xmath308 smgs identified through the whole aztec 1.1-mm galaxy surveys .
sma ( 890 @xmath1 and 1.3 mm ) interferometry confirms the exact position , and photometry from vla , atca , nma , and _ herschel _ , in addition to aztec / aste and sma , constrains the sed well , which is in good agreement with a single gray - body with @xmath309 k and @xmath138 ( [ sect : mmj1545 ] ) .
the sed fits to the photometry at @xmath310 indicate a redshift of @xmath3115.6 ( a combination of 68% confidence intervals of sed fits using the arp 220 , average smg , and smm j2135 templates ) ( [ sect : photoz ] ) .
this is also supported by the ( sub)mm color analysis , in which we show that mm - j1545 is situated in a region where @xmath1805 dusty galaxies are expected on the @xmath312@xmath313 diagram ( [ sect : color ] ) . a faint nir object , j1545b ,
is identified @xmath145 east of the sma 890-@xmath1 peak , which is likely a foreground lensing object that amplifies mm - j1545 .
although it is difficult to constrain the sed of j1545b , the einstein radius would be @xmath245 if the lens redshift is @xmath314 .
even if this is the case and the magnification factor might be as high as 10 , the demagnified fir luminosity is still extreme ( @xmath315 ) , suggesting that mm - j1545 is intrinsically a hyper - luminous galaxy at @xmath1805 ( [ sect : gravlens ] ) .
mm - j1541 is identified as the fourth brightest 1.1-mm source ( @xmath316 mjy ) of the whole aztec surveys .
this object is also seen in the spire 250500 @xmath1 and mips 24 @xmath1 bands ( [ sect : mmj1541 ] ) .
the inferred photometric redshift ranges @xmath3174.1 ( a combination of 68% confidence intervals of sed fits using the arp 220 , m 82 , average smg , and smm j2135 templates ) ( [ sect : photoz ] ) .
again , this is also supported by the aztec
spire color analysis ( [ sect : color ] ) . an optical / nir image ( j1541b )
is clearly offset from the 24-@xmath1 centroid by @xmath318 , and the sed of j1541b is consistent with a bright galaxy at @xmath319 , suggesting that j1541b gravitationally magnifies the background smg , mm - j1541 .
gravitational lens modeling using the faber - jackson relation and a singular isothermal sphere shows the magnification factor of mm - j1541 is moderate ( @xmath320 ) , suggesting that mm - j1541 might be an intrinsically hyper - luminous star - forming galaxy with @xmath23313.9 ( [ sect : gravlens ] ) .
the brightest - end ( @xmath321 mjy ) of the 1.1-mm cumulative number counts is constrained by mm - j1545 and mm - j1541 , in addition to another two sources from the literature ; @xmath322 deg@xmath6 , @xmath323 deg@xmath6 , and @xmath324 deg@xmath6 .
the slope at @xmath325 mjy is shallower than the deeper part of the source counts obtained from general deep fields such as goods and cosmos .
the amplitude and slope at the brightest - end is consistent not only with that properly scaled from the 1.4-mm counts obtained by the south pole telescope survey , which are thought to be dominated by strongly - lensed smgs , but also with a model prediction where flux magnification due to strong gravitational lensing is accounted for .
this suggests that a substantial fraction of @xmath321 mjy sources may be gravitationally amplified .
the overall sed from the optical to the radio and the spatial structure of mm - j1545 are explained neither by a local prestellar core nor a protostellar object , although it is found toward a relatively high h@xmath287 column region of the local molecular cloud . a possible explanation for a galactic object might be a first hydrostatic core with no envelope structure . even in this case , however ,
neither the 6-cm continuum emission nor the nir compact object , j1545b , @xmath145 away from mm - j1545 can be explained by any models for a first hydrostatic core . on the other hand ,
a @xmath1805 smg strongly lensed by j1545b naturally accounts for all of the observed properties .
hence we conclude that mm - j1545 is not a first hydrostatic core but a galaxy at a cosmological distance .
unexpectedly , an extremely - bright smg at @xmath7 and a galactic low - mass dense starless core ( e.g. , an exposed first hydrostatic core ) could be similar in the mid - infrared to millimeter spectral energy distributions and spatial structures at least at @xmath8 .
this indicates that it is necessary to distinguish the two possibilities by means of broad band photometry from the optical to centimeter , when a compact object is identified toward galactic star - forming regions .
( sub)millimeter spectroscopy and/or sub - arcsec imaging of the object will help to determine the redshift and the presence of gravitational magnification , which will be able to carried out using alma .
we are grateful to the aste team for making the observations possible .
we thank m. bthermin and i. shimizu for making the model number counts available .
the study is partially supported by kakenhi ( no .
25103503 ) .
ys and kt are supported by jsps research fellowship for young scientists ( no . 09j05537 and
09j00159 , respectively ) .
this work is based on observations and archival data made with the following telescopes and facilities : the aste telescope is operated by national astronomical observatory of japan . the australia telescope compact array is part of the australia telescope national facility which is funded by the commonwealth of australia for operation as a national facility managed by csiro . _
herschel _ is an esa space observatory with science instruments provided by european - led principal investigator consortia and with important participation from nasa .
nrao is a facility of the nsf operated under cooperative agreement by associated universities , inc .
the 45-m radio telescope and the nobeyama millimeter array are operated by nobeyama radio observatory , a branch of national astronomical observatory of japan .
the submillimeter array is a joint project between the smithsonian astrophysical observatory and the academia sinica institute of astronomy and astrophysics and is funded by the smithsonian institution and the academia sinica .
the _ spitzer _
space telescope is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
subaru telescope is operated by national astronomical observatory of japan .
the united kingdom infrared telescope is operated by the joint astronomy centre on behalf of the science and technology facilities council of the u.k .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
andre , p. , ward - thompson , d. , & barsony , m. 1993 , , 406 , 122 barnard , v. e. , vielva , p. , pierce - price , d. p. i. , et al .
2004 , , 352 , 961 baugh , c. m. , lacey , c. g. , frenk , c. s. , et al .
2005 , , 356 , 1191 bthermin , m. , daddi , e. , magdis , g. , et al .
2012 , , 757 , l23 blain , a. w. , & longair , m. s. 1993 , , 264 , 509 blain , a. w. , smail , i. , ivison , r. j. , kneib , j .-
p . , & frayer , d. t. 2002 , , 369 , 111 bolzonella , m. , miralles , j .- m . , & pell , r. 2000 , , 363 , 476 bonnor , w. b. 1956 , , 116 , 351 boone , f. , clment , b. , richard , j. , et al .
2013 , , 559 , l1 bothwell , m. s. , aguirre , j. e. , chapman , s. c. , et al .
2013 , , 779 , 67 bruzual , g. , & charlot , s. 2003 , , 344 , 1000 bussmann , r. s. , prez - fournon , i. , amber , s. , et al .
2013 , , 779 , 25 carlstrom , j. e. , ade , p. a. r. , aird , k. a. , et al .
2011 , , 123 , 568 casali , m. , adamson , a. , alves de oliveira , c. , et al .
2007 , , 467 , 777 casey , c. m. , narayanan , d. , & cooray , a. 2014 , , 541 , 45 chapman , s. c. , blain , a. w. , smail , i. , & ivison , r. j. 2005 , , 622 , 772 chen , x. , arce , h. g. , zhang , q. , et al .
2010 , , 715 , 1344 combes , f. , rex , m. , rawle , t. d. , et al .
2012 , , 538 , l4 comern , f. 2008 , handbook of star forming regions , volume ii , 295 condon , j. j. , cotton , w. d. , greisen , e. w. , et al .
1998 , , 115 , 1693 cox , p. , krips , m. , neri , r. , et al .
2011 , , 740 , 63 danielson , a. l. r. , swinbank , a. m. , smail , i. , et al .
2011 , , 410 , 1687 danielson , a. l. r. , swinbank , a. m. , smail , i. , et al .
2013 , , 436 , 2793 decarli , r. , walter , f. , neri , r. , et al .
2012 , , 752 , 2 dobashi , k. 2011 , , 63 , 1 dowell , c. d. , conley , a. , glenn , j. , et al .
2014 , , 780 , 75 downes , t. p. , welch , d. , scott , k. s. , et al .
2012 , , 423 , 529 ebert , r. 1955 , , 36 , 222 enoch , m. l. , lee , j .- e . ,
harvey , p. , dunham , m. m. , & schnee , s. 2010 , , 722 , l33 evans , n. j. , ii , allen , l. e. , blake , g. a. , et al .
2003 , , 115 , 965 evans , n. j. , ii , dunham , m. m. , jrgensen , j. k. , et al . 2009 , , 181 , 321 ezawa , h. , kawabe , r. , kohno , k. , & yamamoto , s. 2004 , , 5489 , 763 ezawa , h. , kohno , k. , kawabe , r. , et al .
2008 , , 7012 , faber , s. m. , & jackson , r. e. 1976 , , 204 , 668 fazio , g. g. , hora , j. l. , allen , l. e. , et al .
2004 , , 154 , 10 fu , h. , jullo , e. , cooray , a. , et al .
2012 , , 753 , 134 fu , h. , cooray , a. , feruglio , c. , et al .
2013 , , 498 , 338 gehrels , n. 1986 , , 303 , 336 gonzlez - nuevo , j. , lapi , a. , fleuren , s. , et al .
2012 , , 749 , 65 granato , g. l. , de zotti , g. , silva , l. , bressan , a. , & danese , l. 2004 , , 600 , 580 granato , g. l. , silva , l. , lapi , a. , et al . 2006 , , 368 , l72 greisen , e. w. 2003 , information handling in astronomy - historical vistas , 285 , 109 harris , a. i. , baker , a. j. , zonak , s. g. , et al .
2010 , , 723 , 1139 hatsukade , b. , kohno , k. , aretxaga , i. , et al .
2011 , , 411 , 102 hatsukade , b. , ohta , k. , seko , a. , yabe , k. , & akiyama , m. 2013 , , 769 , l27 hildebrand , r. h. 1983 , , 24 , 267 ho , p. t. p. , moran , j. m. , & lo , k. y. 2004 , , 616 , l1 holland , w. s. , robson , e. i. , gear , w. k. , et al . 1999 , , 303 , 659 holland , w. s. , bintley , d. , chapin , e. l. , et al .
2013 , , 430 , 2513 ichikawa , t. , suzuki , r. , tokoku , c. , et al .
2006 , , 6269 , ikarashi , s. , kohno , k. , aguirre , j. e. , et al .
2011 , , 415 , 3081 iono , d. , hatsukade , b. , kohno , k. , et al .
2012 , , 64 , l2 ivison , r. j. , swinbank , a. m. , swinyard , b. , et al .
2010 , , 518 , l35 johansson , d. , sigurdarson , h. , & horellou , c. 2011 , , 527 , a117 karim , a. , swinbank , a. m. , hodge , j. a. , et al .
2013 , , 432 , 2 knudsen , k. k. , van der werf , p.
p. , & kneib , j .-
2008 , , 384 , 1611 kovcs , a. , chapman , s. c. , dowell , c. d. , et al .
2006 , , 650 , 592 kreysa , e. , gemuend , h .-
p . , gromke , j. , et al .
1998 , , 3357 , 319 la barbera , f. , de carvalho , r. r. , de la rosa , i. g. , & lopes , p. a. a. 2010 , , 408 , 1335 larson , r. b. 1969 , , 145 , 271 lestrade , j .- f . , wyatt , m. c. , bertoldi , f. , menten , k. m. , & labaigt , g. 2009 , , 506 , 1455 lestrade , j .- f . , combes , f. , salom , p. , et al .
2010 , , 522 , l4 lupu , r. e. , scott , k. s. , aguirre , j. e. , et al .
2012 , , 757 , 135 marsh , k. a. , griffin , m. j. , palmeirim , p. , et al .
2014 , , 439 , 3683 masunaga , h. , miyama , s. m. , & inutsuka , s .- i .
1998 , , 495 , 346 masunaga , h. , & inutsuka , s .-
2000 , , 531 , 350 mcmullin , j. p. , waters , b. , schiebel , d. , young , w. , & golap , k. 2007 , astronomical data analysis software and systems xvi , 376 , 127 messias , h. , afonso , j. m. , salvato , m. , mobasher , b. , & hopkins , a. m. 2014 , , 562 , a144 michaowski , m. , hjorth , j. , & watson , d. 2010 , , 514 , a67 mocanu , l. m. , crawford , t. m. , vieira , j. d. , et al . 2013 , , 779 , 61 monet , d. g. , levine , s. e. , canzian , b. , et al .
2003 , , 125 , 984 nakajima , t. , sakai , t. , asayama , s. , et al .
2008 , , 60 , 435 negrello , m. , hopwood , r. , de zotti , g. , et al .
2010 , science , 330 , 800 nguyen , h. t. , schulz , b. , levenson , l. , et al .
2010 , , 518 , l5 oguri , m. 2010 , , 62 , 1017 okumura , s. k. , momose , m. , kawaguchi , n. , et al .
2000 , , 52 , 393 omont , a. , yang , c. , cox , p. , et al .
2013 , , 551 , a115 onishi , t. , mizuno , a. , kawamura , a. , tachihara , k. , & fukui , y. 2002 , , 575 , 950 ott , s. 2010 , astronomical data analysis software and systems xix , 434 , 139 pahre , m. a. , djorgovski , s. g. , & de carvalho , r. r. 1998 , , 116 , 1591 pineda , j. e. , arce , h. g. , schnee , s. , et al .
2011 , , 743 , 201 rau , u. , & cornwell , t. j. 2011 , , 532 , a71 rawle , t. d. , lucey , j. r. , smith , r. j. , & head , j. t. c. g. 2013 , , 433 , 2667 riechers , d. a. , capak , p. l. , carilli , c. l. , et al .
2010 , , 720 , l131 riechers , d. a. , cooray , a. , omont , a. , et al .
2011 , , 733 , l12 riechers , d. a. , bradford , c. m. , clements , d. l. , et al .
2013 , , 496 , 329 rieke , g. h. , young , e. t. , engelbracht , c. w. , et al .
2004 , , 154 , 25 rygl , k. l. j. , benedettini , m. , schisano , e. , et al .
2013 , , 549 , l1 saigo , k. , & tomisaka , k. 2011 , , 728 , 78 sault , r. j. , teuben , p. j. , & wright , m. c. h. 1995 , astronomical data analysis software and systems iv , 77 , 433 scott , k. s. , et al .
2008 , , 385 , 2225 scott , k. s. , yun , m. s. , wilson , g. w. , et al .
2010 , , 405 , 2260 scott , k. s. , lupu , r. e. , aguirre , j. e. , et al .
2011 , , 733 , 29 scott , k. s. , wilson , g. w. , aretxaga , i. , et al .
2012 , , 423 , 575 shimizu , i. , yoshida , n. , & okamoto , t. 2012 , , 427 , 2866 shu , f. h. , adams , f. c. , & lizano , s. 1987 , , 25 , 23 silva , l. , granato , g. l. , bressan , a. , & danese , l. 1998 , , 509 , 103 siringo , g. , kreysa , e. , kovcs , a. , et al .
2009 , , 497 , 945 skrutskie , m. f. , cutri , r. m. , stiening , r. , et al .
2006 , , 131 , 1163 sunada , k. , yamaguchi , c. , nakai , n. , et al .
2000 , , 4015 , 237 suzuki , r. , tokoku , c. , ichikawa , t. , et al .
2008 , , 60 , 1347 swinbank , a. m. , et al .
2010 , , 464 , 733 swinbank , a. m. , simpson , j. m. , smail , i. , et al .
2014 , , 438 , 1267 takekoshi , t. , tamura , y. , minamidani , t. , et al .
2013 , , 774 , l30 tomida , k. , machida , m. n. , saigo , k. , tomisaka , k. , & matsumoto , t. 2010 , , 725 , l239 tomida , k. , tomisaka , k. , matsumoto , t. , et al .
2013 , , 763 , 6 tsutsumi , t. , morita , k .-
i . , & umeyama , s. 1997 , astronomical data analysis software and systems vi , 125 , 50 valtchanov , i. , virdee , j. , ivison , r. j. , et al .
2011 , , 415 , 3473 vieira , j. d. , crawford , t. m. , switzer , e. r. , et al .
2010 , , 719 , 763 vieira , j. d. , marrone , d. p. , chapman , s. c. , et al .
2013 , , 495 , 344 walter , f. , decarli , r. , carilli , c. , et al .
2012 , , 486 , 233 wardlow , j. l. , cooray , a. , de bernardis , f. , et al .
2013 , , 762 , 59 wei , a. , de breuck , c. , marrone , d. p. , et al .
2013 , , 767 , 88 wilson , g. w. , austermann , j. e. , perera , t. a. , et al . 2008a , , 386 , 807 wilson , g. w. , hughes , d. h. , aretxaga , i. , et al . 2008b , , 390 , 1061 wright , e. l. , eisenhardt , p. r. m. , mainzer , a. k. , et al .
2010 , , 140 , 1868 yun , m. s. , scott , k. s. , guo , y. , et al .
2012 , , 420 , 957 | we report detections of two candidate distant submillimeter galaxies ( smgs ) , mm j154506.4@xmath0344318 and mm j154132.7@xmath0350320 , which are discovered in the aztec / aste 1.1 mm survey toward the lupus - i star - forming region .
the two objects have 1.1 mm flux densities of 43.9 and 27.1 mjy , and have _
herschel_/spire counterparts as well .
the submillimeter array counterpart to the former smg is identified at 890 @xmath1 and 1.3 mm .
photometric redshift estimates using all available data from the mid - infrared to the radio suggest that the redshifts of the two smgs are @xmath25 and 3 , respectively .
near - infrared objects are found very close to the smgs and they are consistent with low-@xmath3 ellipticals , suggesting that the high apparent luminosities can be attributed to gravitational magnification .
the cumulative number counts at @xmath4 mjy , combined with other two 1.1-mm brightest sources , are @xmath5 deg@xmath6 , which is consistent with a model prediction that accounts for flux magnification due to strong gravitational lensing .
unexpectedly , a @xmath7 smg and a galactic dense starless core ( e.g. , a first hydrostatic core ) could be similar in the mid - infrared to millimeter spectral energy distributions and spatial structures at least at @xmath8 .
this indicates that it is necessary to distinguish the two possibilities by means of broad band photometry from the optical to centimeter and spectroscopy to determine the redshift , when a compact object is identified toward galactic star - forming regions . |
non - thermal images of snrs are rich sources of information about the properties of interstellar magnetic field ( ismf ) behavior and kinetics of charged particles in vicinity of the strong non - relativistic shocks . despite of their importance
, images of snrs in contrast to broad - band spectra are not well studied .
synchrotron x - ray brightness profiles were used as diagnostic tools for the estimate of the post - shock magnetic field in some snrs ( e.g. * ? ? ? * ) .
radio azimuthal profiles were used for determination of some properties of sn 1006 ( * ? ? ? *
hereafter paper i ) and x - ray radial profiles were used to detect the shock precursor in sn 1006 and prove particle acceleration @xcite .
a detailed approach to modeling the synchrotron images of adiabatic snrs in uniform ismf and uniform interstellar medium ( ism ) is developed by @xcite .
@xcite use modeled synchrotron maps of snrs to put constraints on properties of accelerated particles .
properties of the inverse - compton ( ic ) @xmath0-ray maps are investigated and compared to radio images in ( * ? ? ? *
hereafter paper ii ) .
the influence of nonuniform ism and/or nonuniform ismf on the thermal x - ray morphology of adiabatic snrs are studied in , on the radio maps in @xcite and on the synchrotron x - ray and ic @xmath0-ray images in @xcite .
all studies of snr maps assume classic mhd and test - particle theory of acceleration . though they neglect effects of the back - reaction of the efficiently accelerated particles , they are able to explain general properties of the distribution of the surface brightness in radio , x - rays and @xmath0-rays .
this is because the classic theory , in contrast to the non - linear one , is able to deal with oblique shocks , that is vital for synthesis of snr images . at present , the theory which considers effects of accelerated particles on the shock and on acceleration itself is developed for the initially quasi - parallel shocks only .
one may therefore model the only radial profiles of brightness , in the rather narrow region close to the shock ( in order to be certain that obliquity does not introduce prominent modifications ) .
effects of non - linear acceleration on the radial profiles of brightness are considered in @xcite .
future studies on snr morphology should take into account the nla effects . nevertheless , the classic approach is able to reveal the general properties of snr maps determined by mf behavior and particle acceleration . beside that , it is important to know the properties of the ` classic ' images because they create the reference base for investigation of the efficiency of nla effects in the surface brightness distribution of snrs . in paper
i , we introduced a method to derive an aspect angle of ismf from the radio brightness of snr . in paper ii , we synthesized radio and ic @xmath0-ray maps and concluded that coincidence of the position of the @xmath0-ray and radio limbs is not a common case in theoretical models , because different parameters are dominant in determintion of the radio and @xmath0-ray brightness variations . on the other hand , radio , x - ray and @xmath0-ray observations @xcite
show that radio , x - ray and @xmath0-ray limbs coincide in sn 1006 .
as discussed in paper ii , such coincidence might be attributed to a combination of obliquity dependences of magnetic field and properties of emitting particles , as well as orientation versus the observer . in the present paper
, we make a step forward extending our model of non - thermal leptonic emission of sedov snrs in uniform medium to the x - ray band .
we also derive brightness profiles for representative parameters which are suited for the comparison with adiabatic snrs .
moreover , we derive analytical approximations of the azimuthal and radial profiles of radio , x - ray and @xmath0-ray brightness which are extremely useful to demonstrate their dependence on the acceleration parameters , magnetic field orientation and the viewing geometry ; they can also be directly and very easily fitted to snr images to derive estimations for the best - fit quantities . while the analytical approximation can not substitute the more accurate numerical simulations ,
we show that they retain enough accuracy to represent an effective diagnostic tool for the study of non - thermal snr shells .
the present paper is continuation of the study in paper ii .
the model and numerical realisation are therefore similar to those used in the paper ii ; the reader is referred to this paper for more details . in short , in order to investigate the properties of the leptonic emission of sedov snr , we use the @xcite solutions for dynamics of fluid as well as the @xcite description of the mf behavior downstream of the shock .
the use of analytical solutions allows us to reduce the computational time considerably .
we follow @xcite also in calculation of the evolution of the distribution function @xmath1 of relativistic electrons downstream of the shock ( see appendix [ xmaps : app2 ] for details ) .
the ismf orientation versus observer is described by the aspect angle @xmath2 , an angle between ismf and the line of sight .
let us define also the obliquity angle @xmath3 as the angle between ismf and the normal to the shock , and the azimuth angle @xmath4 in the projection plane which is measured from the direction of the component of ismf in the plane of the sky ( see fig . 1 in paper ii ) .
the compression factor for ismf at the shock front @xmath5 is modulated from unity at parallel shock to @xmath6 at perpendicular one ( where @xmath0 is the adiabatic index ) , in agreement with @xcite . at the shock ,
the energy spectrum of relativistic electrons is taken to be @xmath7 , where @xmath8 is the maximum energy of electrons , @xmath9 the spectral index , @xmath10 the normalization , free parameter @xmath11 regulates the rate of the spectrum decrease around the high - energy end .
such broadening should be attributed to the physics of acceleration @xcite , rather than to the effect of superposition of spectra in different conditions along the line of sight as it was suggested initially by @xcite . from other side , theoretical model of @xcite
demonstrates that it should be @xmath12 for the loss - limited models , whereas in practice , taking the time evolution into account it could be between 1 and 2 @xcite . also @xcite suggest that @xmath13 . ] .
evolution of the ` electron injection ability ' of the shock is represented as @xmath14 where @xmath15 is the shock velocity , and @xmath16 is a constant .
the variation of the distribution @xmath1 over the surface of the snr and its evolution downstream of the shock are calculated as described by @xcite .
following @xcite , we synthesize the synchrotron and ic emission , considering each of the three cases of variation of electron injection efficiency with the shock obliquity ( quasi - perpendicular , isotropic , and quasi - parallel particle injection ) and each of the three alternatives for time and spatial dependence of @xmath8 ( time - limited , loss - limited and escape - limited with a gyrofactor @xmath17 as a parameter ; it is a ratio of the mean free path of a particle to its gyroradius ) .
both the synchrotron and the ic losses are included as channels for the radiative losses of relativistic electrons , though the ic losses are quite small comparing to synchrotron due to large mf considered .
the surface brightness is calculated by integrating emissivities along the line of sight within the snr .
the emissivity of electrons is given by @xmath18 where @xmath19 is the photon energy and @xmath20 is the spectral distribution of synchrotron or ic radiation power of electron .
we assume that information about orientation of @xmath21 inside snr is lost because of turbulence , in practice , we use an average aspect angle downstream .
the distribution @xmath20 is calculated with the use of the analytical approximation developed by ( * ? ? ?
* see also paper ii ) .
the resulting synchrotron radio images and the ic @xmath0-ray images synthesized by our model have been already presented in paper i and paper ii .
therefore , we present here only the x - ray images ( see fig .
[ xmaps : fig3 ] ) , adding appropriate references to the previous work to let the reader do a quick comparison .
the pattern of synchrotron x - ray brightness of snr is in general similar to the radio one . in most cases , the bright x - ray limbs or other features
are located in the same azimuth as in the radio images .
the only differences appear due to radiative losses which modify downstream distribution of the electrons emitting in x - rays ( thus the features of brightness are radially thinner ) and due to variation of @xmath8 over the snr surface . in the radio
( see fig . 4 in paper ii ) as also in the x - ray band , the remnant shows two symmetric bright lobes ( for @xmath22 ) in all the injection models with the maxima in surface brightness coincident in the two bands .
the maxima are located at perpendicular shocks in the quasi - perpendicular and isotropic models ( i.e. where @xmath21 is higher ) , and at parallel shocks in the quasi - parallel model ( i.e. where emitting electrons are only presented ) .
the lobes are much radially thinner in x - rays than in radio because of the large radiative losses at the highest energies that make the x - ray emission dominated by radii closest to the shock .
the x - ray morphology of snr is different for different aspect angles ( fig .
[ xmaps : fig3 ] , cf . with figs . 4 , 5 and 6 of paper ii for radio and @xmath0-ray images , respectively ) .
in the case of quasi - perpendicular injection , the morphology is bilateral ( two lobes ) for large aspect angles ( @xmath23 , i.e. the component of ismf in the plane of the sky is larger than that along the line of sight ) and almost ring - like for low aspect angles ( @xmath24 ; see fig .
[ xmaps : fig3 ] ) with intermediate morphology between @xmath25 and @xmath26 . in the case of quasi - parallel injection
, the remnant morphology in the radio band is known to be bilateral for large aspect angles and characterized by one or two eyes for low aspect angles @xcite . on the other hand ,
it is worth noting that the remnant morphology in x - rays is in general bilateral for aspect angles @xmath27 and centrally bright for very low angles , indeed a rather limited set of possible cases ( lower panels in fig .
[ xmaps : fig3 ] ) .
this happens because the non - thermal x - ray emission originates from a very thin shell behind the shock making the effect of limb brightening in x - rays more important than in the radio band .
in addition , we note that centrally bright x - ray ( and radio ) snrs are expected to be much fainter than bilateral snrs ( see lower panels in fig . [ xmaps : fig3 ] ) and consequently much more difficult to be observed . the above considerations may affect the statistical arguments generally invoked against the quasi - parallel injection ( i.e. the fact that this model produces morphology which is not observed ; e.g. @xcite ) .
images on fig .
[ xmaps : fig3 ] are calculated assuming a dependency of @xmath8 on the obliquity angle which corresponds to the time - limited model with @xmath28 as introduced by @xcite , namely @xmath29 , where @xmath30 is a function describing smooth variations of @xmath8 versus obliquity , and @xmath31 is the pre - shock ismf strength ( in such a model no dependency on the shock velocity is present ; * ? ? ?
* ) . with this particular choice
, @xmath8 is quite similar for different obliquities , namely @xmath32 .
larger values of @xmath17 always provide @xmath33 in this model , thus the character of azimuthal variation of brightness would be similar .
the efficiency of variation of @xmath8 with obliquity in modification of the azimuthal distribution of x - ray synchrotron brightness depends obviously on the photon energy : if the maximum contribution to the emission at given photon energy is from electrons with energy much less than @xmath8 then this effect is negligible .
it is useful to introduce the reduced photon energy , as @xmath34 where @xmath35 is the synchrotron characteristic frequency , @xmath36 , or @xmath37 where @xmath38 is the photon energy in kev . if @xmath39 then most of the contribution to the synchrotron x - ray emission is from electrons with energy @xmath40 .
is from electrons with energies 72 tev ; the maximum contribution to ic @xmath0-ray emission at 1 tev originates from electrons with energies 17 tev . ]
[ xmaps : fig3 ] is calculated for @xmath41 , i.e. images shown are mainly due to emission of electrons with energy few times higher than @xmath42 ; the role of variation of @xmath8 is therefore clearly visible in the images .
( marked near respective lines ) .
calculations are done for @xmath43 , @xmath44 , @xmath45 , @xmath46 . ] ) on the radial profiles of x - ray surface brightness , at two azimuth .
the reduced fiducial energy @xmath47 ( from below ) ; @xmath48 , @xmath49 , @xmath50 . ]
in the present subsection , @xmath8 is assumed to be constant in time and the same for any obliquity ; in addition , isotropic injection , @xmath51 and @xmath52 in the energy spectrum of relativistic electrons are assumed .
the radial thickness of features in the x - ray images is sensitive to the photon energy : the larger the energy the thinner the limbs ( fig .
[ xmaps : fig_photon_energy ] ) .
this is because radiative losses @xmath53 of electrons with energy @xmath54 is efficient for more energetic electrons , @xmath55 . if @xmath56 then most of the contribution to the synchrotron x - ray emission is from electrons with energies @xmath57 where the radiative losses are of the main importance .
an important factor for emission of highly energetic electrons is the fiducial energy , which reflects the importance of radiative losses in modification of the electron distribution .
it is defined as @xmath58 @xcite , or @xmath59 radiative losses are important for @xmath60 and minor for @xmath61 .
[ xmaps : fig_fiducial ] demonstrates how the value of @xmath62 affects the radial profiles of x - ray brightness : the smaller @xmath62 the thinner the rim . ,
@xmath63 and @xmath64 .
calculations are done for @xmath43 , @xmath44 , @xmath50 , @xmath65 , @xmath66 . ]
our model does not include consistently the effects on shock dynamics due to back - reaction of accelerated crs .
however , we may approach the effect of shock modification by considering different values of the adiabatic index @xmath0 which is expected to drop from the value of an ideal monoatomic gas .
in particular , fig .
[ xmaps : fig_gamma ] considers the cases of @xmath67 ( for an ideal monatomic gas ) , @xmath68 ( for a gas dominated by relativistic particles ) , and @xmath69 ( for large energy drain from the shock region due to the escape of high energy crs ) .
the shock modification results in more compressible plasma and , therefore , in the radially - thinner features of the nonthermal images of snrs . a small distance between the forward shock and contact discontinuity @xcite
could also be attributed to @xmath70 .
effect of the index @xmath0 on the profiles of hydrodynamical parameters downstream of the adiabatic shock is widely studied ( e.g. * ? ? ?
* ) : smaller @xmath0 makes the shock compression factor higher , @xmath71 and the gradient of density downstream stronger ( e.g. appendix [ xmaps : app3 ] ) , @xmath72 ( where @xmath73 , @xmath74 is the density immediately postshock and @xmath75 ) .
in addition to that , the x - ray ( and also tev @xmath0-ray ) brightness is modified by increased radiative losses of emitting electrons . really , the larger compression leads to the higher post - shock mf and thus to increased losses , @xmath76 , which results in turn in the thinner radial profiles of brightness .
it is unknown how the injection efficiency ( the fraction of nonthermal particles ) depends on the properties of the shock .
we parameterized its evolution as @xmath77 where @xmath16 is a constant .
effect of the parameter @xmath16 on the radial profiles of the surface brightness is demonstrated on fig .
[ xmaps : fig_b ] .
the smaller @xmath16 the thicker the profiles , because there are more emitting electrons in deeper layers , which were injected at previous times .
this property affects the nonthermal emission in all bands .
however , the effect is less prominent in x - rays ( and in tev @xmath0-rays ) if radiative losses are quite effective to dominate it ( see fig . [ xmaps : fig_b ] , lines for different photon energies ) .
this is in agreement with finding of @xcite and @xcite who showed that , for @xmath57 , the width of the synchrotron limbs does not depend on the shock velocity in the loss limited case .
instead , profiles of the radio brightness may be used to put limitations on the value of @xmath16 . in a similar fashion
, the x - ray and @xmath0-ray radial profiles are affected also by the time evolution of the maximum energy , @xmath78 .
however , it seems impossible to determine @xmath79 from such profiles because contribution of other factors is often dominant .
an interesting feature of the synchrotron images of snrs is apparent from fig .
[ xmaps : fig_azimuth ] .
the maxima of the radial profiles of brightness for different azimuth are located almost at the same distance @xmath80 from the center of projection , for radio and x - rays .
thus , the best way to analyze the azimuthal profiles of the surface brightness is to find the position @xmath80 of the maximum for one azimuth and then to trace the azimuthal profile of brightness @xmath81 for fixed @xmath80 .
the rigorous model discussed in the previous section is able to predict the non - thermal emission of sedov snr under a variety of conditions .
however , in practical applications , it may be rather time - consuming to perform an extensive parameter space exploration , and the crucial dependence on the relevant parameters of the acceleration processes may be hidden . in order to understand how the properties of mf , electron injection and acceleration influence the brightness distribution
, we derived analytic approximate formulae for the azimuthal and radial profiles of the surface brightness of adiabatic snr . in this way
, we can easily see what are the main factors which determine the pattern of the nonthermal images of snrs , and which of them are mostly responsible for the azimuthal variation of the surface brightness and which for the radial one .
this turns to be extremely useful in guiding the comparison with real observations .
the analytical formulae are valid close to the shock only , but are adequate to describe azimuthal and radial variations of brightness around maxima which are located close to the edge of snr shells .
let the evolution and obliquity variation of the electron injection efficiency be denoted as @xmath82 and of the obliquity variation of mf compression / amplification as @xmath83 ; for the sake of generality we assume @xmath84 and @xmath83 to be some arbitrary smooth functions .
properties of the azimuthal and radial profiles of the radio brightness is determined mostly by ( appendix [ xmaps : app5 ] ) @xmath85 where @xmath86 @xmath87 , @xmath88 is the shock compression ratio , @xmath89 @xmath90 is close to unity for @xmath91 ( appendix [ xmaps : app3 ] ) .
the effective obliquity angle @xmath92 is related to azimuth @xmath4 and aspect @xmath2 as @xmath93 the azimuth angle is measured from the direction of ismf in the plane of the sky .
( [ xmaps : radio_brightness ] ) is a generalisation of the approximate formula derived in paper i. eq .
( [ xmaps : radio_brightness ] ) shows that the azimuthal variation of the radio surface brightness @xmath94 at a fixed radius @xmath95 of projection , is mostly determined by the variations of the magnetic field compression ( and amplification , if any ) @xmath5 and by the variation of the electron injection efficiency @xmath96 .
the radial profile @xmath97 is determined mostly by @xmath88 , @xmath16 and @xmath9 .
adiabatic index @xmath0 affects the radial and azimuthal profiles mostly through the compression factor @xmath6 because @xmath90 weakly depends on @xmath0 .
let us assume that the maximum energy is expressed as @xmath98 , where @xmath79 is a constant and @xmath30 , for the sake of generality , is some arbitrary function describing the smooth variation of @xmath8 versus obliquity .
the synchrotron x - ray brightness close to the forward shock is approximately ( appendix [ xmaps : app1 ] ) @xmath99 i{_\mathrm{rx}}(\varphi,\bar\varrho ) \label{xmaps : xray_brightness}\ ] ] where @xmath100 with @xmath101 .
\label{xmaps : ixray}\ ] ] the parameter @xmath102 is responsible for the adiabatic ( the first term ) and radiative ( the second term ) losses of emitting electrons and the time evolution of @xmath8 on the shock ( the third term ) .
the reduced electron energy which gives the maximum contribution to emission of photons with energy @xmath103 is @xmath104 it varies with obliquity ( since mf varies ; electrons with different energies contribute to the synchrotron emission at @xmath103 ) .
parameters @xmath105 , @xmath106 , @xmath107 depend on @xmath92 and , therefore , on the aspect angle @xmath2 and the azimuth angle @xmath4 .
if @xmath108 then eq . ( [ xmaps : xray_brightness ] ) for the x - ray brightness transforms to eq .
( [ xmaps : radio_brightness ] ) for the radio brightness .
the thickness of the hard x - ray radial profile is used to estimate the post - shock strength of mf in a number of snrs ( e.g. * ? ? ?
the absolute value of mf is present in eq .
( [ xmaps : xray_brightness ] ) through @xmath103 and @xmath62 , eqs .
( [ xmaps : tilde_e_def ] ) , ( [ xmaps : etaf ] ) which appear in @xmath105 and @xmath106 , eqs .
( [ xmaps : psi ] ) , ( [ xmaps : epmx ] ) . in both cases
, @xmath109 is in combination with @xmath8 ( thus , the value of the electron maximum energy may affect the estimations ) .
the idea of the method bases on the increased role of losses in x - rays due to larger mf , i.e. on the role of the second term in @xmath105 , eq .
( [ xmaps : psi ] ) . really , the influence of @xmath110 ( i.e. of @xmath9 and @xmath16 ) is minor in x - rays ( middle panel on fig .
[ xmaps : fig_b ] ) , if radiative losses affect the electron evolution downstream of the shock ( i.e. for @xmath111 , @xmath112 ) .
the role of the first and the third terms in @xmath105 are also minor in most cases ( @xmath113 for the time - limited and escape - limited models and unity for the loss - limited one ) because the second term @xmath114 .
however , the adiabatic index makes an important effect on the thickness of the profile , mostly through @xmath88 which appears in @xmath5 and in @xmath115 .
being smaller than @xmath116 ( that is reasonable especially in the case of efficient acceleration , which is actually believed to be responsible for the large mf ) , the index may compete to some extend the role of losses , used in the method for estimation of mf ( see e.g. fig .
[ xmaps : fig_gamma ] ) that might lead to smaller estimates of mf strength .
the ic @xmath0-ray brightness may approximately be described as ( appendix [ xmaps : app4 ] ) @xmath117 i{_\mathrm{ic}}(\varphi,\bar\varrho ) \label{xmaps : ic_brightness}\ ] ] where @xmath118 , \label{xmaps : iic_main } \end{array } \label{xmaps : igtext}\ ] ] @xmath119 the expression for @xmath105 is the same as ( [ xmaps : psi ] ) but @xmath106 is different : @xmath120 where @xmath19 is the @xmath0-ray photon energy , @xmath121 the temperature of the seed photon field , @xmath122 the lorentz factor of electrons with energy @xmath8 , or @xmath123 eq .
( [ xmaps : ic_brightness ] ) is a generalisation of the approximate formula derived in paper ii .
the azimuthal variation of the ic @xmath0-ray brightness depends mostly on the injection efficiency .
the role of variation of @xmath8 is prominent only if obliquity dependence of injection is not strong .
parameter @xmath11 , being smaller than unity , results in smaller azimuthal contrasts of synchrotron x - ray or ic @xmath0-ray brightness comparing to model with purely exponential cut - off in @xmath1 .
the radial distribution of ic brightness is determined mostly by @xmath88 , @xmath106 , @xmath62 , @xmath16 and , to the smaller extend , by @xmath9 and @xmath79 .
the approximations presented above do not require long and complicate numerical simulations but restore all the properties of nonthermal images discussed in the previous sections as well as in papers i and ii , including dependence on the aspect angle .
therefore , they may be used as a simple diagnistic tool for non - thermal maps of snrs .
the formulae are rather accurate in description of the brightness distribution close to the shock .
they do not represent centrally - brightened snrs . instead
, they may be used in snr shells for those azimuth @xmath4 where @xmath124 and @xmath125 , in the range of @xmath126 from @xmath127 to 1 , where @xmath128 , @xmath129 is the radius ( close to the shock ) where the maximum in the radial profile of brightness happens .
approximations are compared with numerical calculations in respective appendices and their applicability is discussed in details on example of the ic emission in appendix [ xmaps : app4_accuracy ] .
analysis of azimuthal profiles of brightness in different bands allows one to put limitations on models of injection , mf , @xmath8 . in most cases , the best way to estimate
the azimuthal variation @xmath81 of the surface brightness is the following .
an approximate radial profile @xmath130 of the brightness should be produced for azimuth where the largest losses occur ( i.e. where @xmath131 is smaller ; e.g. at @xmath132 ) .
this allows us to find @xmath129 which should be used in @xmath81 in order to estimate the azimuthal variation of brightness .
@xmath129 keeps us at maxima in the radial brightness profiles for different azimuth ( fig .
[ xmaps : fig_azimuth ] ) .
energy of electrons evolve downstream of the shock as @xmath133 , where @xmath134 is initial energy , @xmath135 , @xmath136 the lagrangian coordinate .
adiabatic and radiative losses of electrons in a given fluid element @xmath136 are represented by functions @xmath137 , @xmath138 respectively ( appendix [ xmaps : app2 ] ) .
close to the shock , they are approximately @xmath139 ( @xmath90 depends on the adiabatic index @xmath0 only and is close to unity for @xmath91 ) , @xmath140 ( appendix [ xmaps : app3 ] ) .
the latter expression clearly shows that the fiducial energy @xmath62 is important parameter reflecting the ` sensitivity ' of the model to the radiative losses , as it is shown by @xcite : the larger the fiducial energy the smaller the radiative losses .
in fact , @xmath141 means no radiative losses at all .
another fact directly visible from this approximation is that radiative losses are much more important at the perpendicular shock ( where @xmath142 ) than at the parallel one ( where @xmath143 ) .
in addition , the radiative losses depends rather strongly on the index @xmath0 : @xmath144 for @xmath67 but @xmath145 for @xmath146 .
note , that in the analysis above , the difference between the parallel and perpendicular shocks are only due to the compression factor @xmath5 which may be treated as `` compression - plus - amplification '' factor .
therefore , our consideration may also be applied in case of the very turbulent and amplified pre - shock field ( when information about obliquity is lost ) once this factor is known .
our approximations reflects also the general ` rule ' for ic emission : there is less ic emission where mf is stronger .
namely , the azimuthal variation is @xmath147 with @xmath148 : emitting electrons disappears toward the shock with larger @xmath5 because mf strength is a reason of higher losses there .
similar dependence on @xmath5 is for x - rays , eq .
( [ xmaps : ixray ] ) ; it is however dominated by the increased term @xmath149 . tev @xmath0-ray image of sn 1006 demonstrates good correlations with x - ray image smoothed to the h.e.s.s .
resolution @xcite .
we mean here both the location and sizes of the bright limbs .
let us consider the polar - caps model of sn 1006 .
the shock is quasi - parallel around the limbs , mf azimuthally increases ( in @xmath150 times ) .
the injection efficiency decreases ( in @xmath151 times ) out of the limbs ; the number of electrons emitting in x - rays and tev @xmath0-rays is therefore dramatically low at perpendicular shock comparing to the parallel , that is in agreement with no tev emission at nw and se regions of sn 1006 .
however , the azimuthal sizes of the limbs in x - rays and @xmath0-rays is expected to be different , in the polar - caps model : they should be larger in x - rays .
really , azimuthal variation of @xmath107 is smaller than variations of @xmath96 and @xmath152 ; therefore , from eqs .
( [ xmaps : xray_brightness ] ) and ( [ xmaps : ic_brightness ] ) , azimuthal variations of brightness are mostly @xmath153 while @xmath154 ( see fig .
[ xmaps : limb_sizes ] for a comparison of remnant morphologies in the radio , x - ray and @xmath0-ray bands ) .
we hope that future observations allow us to see if there is a difference in azimuthal sizes of the limbs in various bands .
how back - reaction of accelerated particles may modify nonthermal images of snrs ?
our formulae can restore some of these effects . in our approximations ,
@xmath9 , in general , is allowed to vary with @xmath54 , e.g. to be @xmath155 .
the index @xmath9 reflects the local slope of the electron spectrum around @xmath106 .
therefore , if @xmath156 , the index @xmath9 has to vary with azimuth because @xmath106 varies , eq .
( [ xmaps : epmx ] ) . generally speaking
, such approach allows one to estimate the role of the nonlinear ` concavity ' of the electron spectrum in modification of the nonthermal images .
however , we expect that this effect is almost negligible because @xmath157 is very slow function , at least within interval of electron energies contributing to images .
another effect of efficient acceleration consists in the adiabatic index @xmath0 smaller than @xmath116 .
our approximations are written for general @xmath0 .
namely , the index @xmath0 affects @xmath158 through @xmath88 .
cosmic rays may also cause the amplification of the seed ismf . in our formulae
, @xmath83 represents the obliquity variation of the ratio of the downstream mf to ismf strength , @xmath159 .
therefore , it may account for both the compression and amplification of ismf ; for such purpose , @xmath5 should be expressed in a way to be unity at parallel shock .
the main reason for the derivation of the explicit approximations of sect . [ xmaps : approximation_section ] is to highlight the factors which are the most efficient in the formation of the pattern of surface brightness .
however , sometimes the analytical formulae may help in estimates some of the remnant parameters .
we would like to present two examples .
let us consider sn 1006 , @xmath51 @xcite , @xmath67 .
the injection efficiency is isotropic if one assumes that sn 1006 evolves in the uniform ismf and uniform ism ( paper i ) . the best - fit value of the aspect angle found from the approximate eq .
( [ xmaps : radio_brightness ] ) is @xmath160 ( fig .
[ xmaps : fig_radio_ap_sn1006 ] ) while the detailed numerical calculations give @xmath161 ( paper i ) .
the same approximate formula shows that the radial profile of radio brightness depends only on the value of @xmath16 ( which shows how the injection efficiency evolve with the shock velocity ) .
unfortunately , the differences between profiles for @xmath162 and @xmath163 are comparable with accuracy of the approximate formulae and of the experimantal data ; therefore , the approximation may not be used for estimations of @xmath16 . the sharpest radial profile of x - ray brightness ( from fig . 4a in * ? ? ?
* ) was used to estimate the strength of the post - shock mf in @xmath164 @xcite .
[ xmaps : fig_xray_ap_sn1006 ] shows approximate profiles for three values of mf in comparison with the chandra profile and detailed numerical simulations ( petruk et al .
, in preparation ) .
one can also see from the approximation that @xmath165 is the most appropriate value in agreement with the value found in the literature . ) . ]
the present paper extends analysis of properties of the surface brightness distribution of spherical adiabatic snrs started in paper i ( radio band ) and paper ii ( ic @xmath0-rays ) to the nonthermal x - rays .
it also generalizes the method of approximate analytical description of the azimuthal and radial profiles of brightness introduced in these papers .
synchrotron images of adiabatic snr in x - rays are synthesized for different assumptions about obliquity variations of the injection efficiency , mf and maximum energy of accelerated electrons .
we analyze properties of these images .
different models of electron injection ( quasi - parallel , isotropic and quasi - perpendicular ) as well as models of the electron maximum energy ( time - limited , loss - limited and escape - limited ) are considered .
the azimuthal variation of the synchrotron x - ray and ic @xmath0-ray brightness is mostly determined by variations of @xmath96 , @xmath5 and @xmath8 , of the radio brightness by @xmath96 and @xmath5 only . in general , higher @xmath21 increases x - ray and decreases ic @xmath0-ray brightness .
really , higher mf is a reason of larger losses of emitting electrons ( i.e. decrease of their number ) and thus of the smaller brightness due to ic process .
in contrast , x - rays are more efficient there because @xmath166 .
the radial profiles of brightness depend on a number of factors .
it is quite sensitive to the adiabatic index : @xmath70 makes plasma more compressible .
therefore , the brightness profile is thinner due to larger compression factor , larger gradient of density and mf downstream of the shock and larger radiative losses .
the role and importance of various factors on the surface brightness in radio , synchrotron x - rays and ic @xmath0-rays are demonstrated by the approximate analytical formulae .
they accurately represent numerical simulations close to the shock and are able to account for some non - linear effects of acceleration if necessary .
this makes the approximations a powerful tool for quick analysis of the surface brightness distribution due to emission of accelerated electrons around snr shells .
the application of the approximate formulae to the case of sn1006 yields measures of the aspect angle and the post - shock mf in good agreement with more accurate analysis found in the literature . ) for @xmath167 are shown by blue , green and red lines respectively .
other parameters : @xmath168 and @xmath169 , @xmath170 , @xmath171 , @xmath172 , @xmath113 , @xmath50 , @xmath52 . ] 99 acero f. , et al . , 2010 , a&a , 516 , a62 berezhko , e. g. , ksenofontov , l. t. & vlk , h. j. 2003 , a&a , 412 , l11 berezhko e. g. & vlk h. j. , 2004 , a&a , 419 , l27 cassam - chena g. , decourchelle a. , ballet j. , ellison d. c. , 2005 , a&a , 443 , 955 g. , hughes j. p. , reynoso e. m. , badenes c. , & moffett d. , 2008 , apj , 680 , 1180 ellison d. & cassam - chena g. , 2005 , apj , 632 , 920 b. & petruk o. , 1999 , a&a , 344 , 295 fryxell b. , olson k. , ricker p. et al . , 2000 , apjs , 131 , 273 m. s. & reynolds s. p. , 1990 ,
apj , 357 , 591 long k. s. , reynolds s. p. , raymond j. c. , et al . 2003 , apj , 586 , 1162 miceli m. , bocchino f. , iakubovskyi d. , orlando s. , telezhinsky i. , kirsch m. g. f. , petruk o. , dubner g. , castelletti g. , 2009 , a&a , 501 , 239 lee s .- h . , kamae t. , ellison d. c. , 2008 , apj , 686 , 325 morlino g.,amato e. , blasi p. , caprioli d. , 2010 , mnras , 405 , l21 s. , bocchino f. , reale f. , peres g. , & petruk o. , 2007 , a&a , 470 , 927 petruk o. , 2009 , a&a , 499 , 643 petruk , o. , beshley , v. , bocchino , f. , & orlando , s. , 2009b , mnras , 395 , 1467 ( paper ii ) petruk o. , bocchino f. , miceli m. , dubner g. , castelletti g. , orlando s. , iakubovskyi d. , telezhinsky i. , 2009b , mnras , 399 , 157 petruk o. , dubner g. , castelletti g. , iakubovskyi d. , kirsch m. , miceli m. , orlando s. , telezhinsky i. , 2009 , mnras , 393 , 1034 ( paper i ) reynolds s.p .
1996 , apj , 459 , l13 reynolds s. p. , 1998
, apj , 493 , 375 reynolds s. p. , 2004 ,
33 , 461 sedov l.i . , 1959 ,
_ similarity and dimensional methods in mechanics _
( new york , academic press ) zirakashvili v. , aharonian f. , 2010 , apj , 708 , 965
relativistic electrons evolving downstream of the shock suffer from adiabatic expansion and radiative losses due to synchrotron and inverse compton processes .
electrons are considered to be confined to the fluid element which removed them from the acceleration site @xcite . and diffusion @xmath173 are comparable : @xmath174 @xcite . therefore
, the results of the present paper are robust . ]
let the fluid element with lagrangian coordinate @xmath136 was shocked at time @xmath175 .
if energy of electrons was @xmath134 at @xmath175 , it becomes later @xcite @xmath176 where the first summand in the denominator reflects adiabatic losses , the second one is due to radiative losses , @xmath73 , index s means ` immediately post - shock ' , the fiducial energy for parallel shock @xmath177 .
the effective magnetic field is @xmath178 , @xmath179 takes into account its evolution downstream @xcite , @xmath180 strength of magnetic field with energy density equal to energy density of cmbr .
@xmath181 is introduced in order to account for the inverse - compton losses @xcite , therefore it is constant everywhere .
the synchrotron channel dominates inverse - compton losses if @xmath182 .
the dimensionless function @xmath183 accounts for evolution of fluid during time from @xmath175 to @xmath184 ; it was initially defined as integral over time @xcite . in case of sedov shock ,
@xmath183 may be written in terms of spatial coordinate that is more convenient for simulations than original representation in terms of time .
namely , for uniform ism : @xmath185 eq .
( [ energylosses ] ) results in relations @xmath186 where the adiabatic and radiative losses are represented by @xmath187 shocks of different strength are able to accelerate electrons to different @xmath8 .
let @xmath188 where @xmath15 is the shock velocity , @xmath189 for loss - limited , time - limited and escape - limited models respectively @xcite .
if shock accelerates electrons to @xmath8 at present time @xmath184 , then , at some previous time @xmath175 when fluid element @xmath190 was shocked , the shock was able to accelerate electrons to @xmath191 where @xmath192 and we used sedov solutions .
the obliquity variation of the maximum energy is given by @xmath193 with @xmath30 independent of time .
let us assume that , at time @xmath175 , an electron distribution has been produced at the shock @xmath194 \label{sp - ini}\ ] ] where @xmath11 is constant .
the obliquity variation of @xmath195 is given by @xmath196 with @xmath197 also independent of time .
conservation equation @xmath198 ( where @xmath199 ) and continuity equation @xmath200 shows that downstream @xmath201 \end{array } \label{n - downstr}\ ] ] with @xmath202 . if @xmath203 , then evolution of @xmath10 is self - similar downstream @xmath204 therefore , in general ,
@xmath205 where the profile @xmath206 is independent of obliquity .
note , that @xmath207 is not affected by the radiative losses , therefore it behaves in the same way also for radio emitting electrons .
once @xmath9 is close to 2 , the radiative losses influence the shape of @xmath1 mostly through the exponential term in eq .
( [ n - downstr ] ) . in other words
, they are effective only around the high - energy end of the electron spectrum as it is shown by @xcite . the above formulae are also valid if the spectral index @xmath9 depends on @xmath54 , e.g. @xmath155 like it would be in case of the nonlinear acceleration .
in addition , no specific value of the adiabatic index @xmath0 is assumed here .
it influences the downstream evolution of relativistic electrons through @xmath208 which depends on @xmath0 @xcite .
let us find approximations for dependence of some parameter @xmath209 on the lagrangian coordinate @xmath210 downstream close to the adiabatic shock .
we are interested in approximations of the form @xmath211 where , by definition , @xmath212_{a = r } = \left[-{\partial \ln { \cal x}{_\mathrm{*}}(a)\over \partial \ln a}\right]_{a = r } \label{kappa - def}\ ] ] and star marks the dependence given by the sedov solution .
this approach yields for density @xmath213 for the relation between eulerian and lagrangian coordinates @xmath214 where the shock compression factor is @xmath215 note that the density distribution in eulerian coordinates is much more sensitive to @xmath0 ( table [ xmaps : table_param_gamma ] ) : @xmath216 magnetic field is approximately @xmath217 @xmath218 @xmath219 approximation for normalization @xmath10 follows from the definition ( appendix [ xmaps : app2 ] ) @xmath220 and approximation for @xmath221 .
adiabatic losses are accounted with @xmath222 which is defined by ( [ cale ] ) .
its approximation is therefore @xmath223 it is valid for @xmath224 with error less than few per cent .
the value of @xmath90 is close to unity for @xmath225 ( table [ xmaps : table_param_gamma ] ) . in order to approximate @xmath226 defined by ( [ cale ] )
, we substitute ( [ int ] ) with approximations for @xmath221 and @xmath227 .
then we use the property @xmath228 for function of the form @xmath229 . in this way , @xmath230 this expression is good for @xmath231 , with error of few per cent .
it depends on @xmath0 through @xmath88 in @xmath5 which is @xcite @xmath232 note that dependence on the absolute value of the magnetic field strength @xmath233 is present in ( [ calerad - approx ] ) : @xmath234 .
this approximation clearly shows that the fiducial energy @xmath235 is important parameter reflecting the ` sensitivity ' to the radiative losses , as it shown by @xcite : the larger the fiducial energy the smaller the radiative losses .
in fact , @xmath141 means no radiative losses at all , see ( [ cale ] ) .
another fact directly visible from eq .
( [ calerad - approx ] ) is that radiative losses are much more important at the perpendicular shock ( @xmath142 ) than at the parallel one ( @xmath143 ) .
radiative losses depends rather strongly on the index @xmath0 : @xmath144 for @xmath67 but @xmath145 for @xmath146 .
the values of parameters in approximations for different adiabatic index @xmath0 are presented in table [ xmaps : table_param_gamma ] .
.parameters in approximations [ cols="<,^,^,^",options="header " , ] [ xmaps : table_param_gamma ]
a formula obtained here may be useful in situations when an approximate quantitative estimation for the azimuthal variation of the synchrotron x - ray surface brightness is sufficient .
the emissivity due to synchrotron emission is @xmath236 spectral distribution of the synchrotron radiation power of electrons with energy @xmath54 in magnetic field of the strength @xmath21 is @xmath237 where @xmath238 is frequency , @xmath239 the characteristic frequency .
most of this radiation is in photons with energy @xmath240 . in the delta - function approximation ,
the special function @xmath241 is substituted with @xmath242 where @xmath243 with this approximation , ( [ xmaps : eq5 ] ) becomes @xmath244 where @xmath245 is the energy of electrons which give maximum contribution to synchrotron emission at photons with energy @xmath19 : @xmath246 .
let the energy of relativistic electrons is @xmath54 in a given fluid element at present time .
their energy was @xmath134 at the time this element was shocked .
these two energies are related as @xmath247 where @xmath248 accounts for the adiabatic losses and @xmath226 for the radiative losses ( appendix [ xmaps : app2 ] ) .
there are approximations valid close to the shock ( appendix [ xmaps : app3 ] ) : @xmath249 where @xmath135 , @xmath136 is lagrangian coordinate of the fluid element , @xmath250 is the fiducial energy for parallel shock , @xmath90 depends on @xmath0 and is given by ( [ calead - approx ] ) ; @xmath251 for @xmath67 ( for other @xmath0 see table [ xmaps : table_param_gamma ] ) .
the factor @xmath5 represents compression in the classical mhd @xcite but may be interpreted also as amplification - plus - compression factor . in the latter case , it should be written in a way to be unity at parallel shock .
the downstream evolution of @xmath10 in a sedov snr is ( appendix [ xmaps : app2 ] ) @xmath252 where @xmath96 is injection efficiency . with the approximations ( [ approxe ] ) and @xmath9 close to 2
, the distribution @xmath1 may be written from ( [ n - downstr ] ) as @xmath253 \label{xmaps : n_e_s}\ ] ] where @xmath254 and @xmath9 is allowed to vary with @xmath54 .
let us consider the azimuthal profile of the synchrotron x - ray brightness @xmath255 at a given radius @xmath95 from the center of the snr projection . like in paper ii , we consider the ` effective ' obliquity angle @xmath92 which , for a given azimuth , equals to the obliquity angle for a sector with the same azimuth in the plane of the sky ( see details in paper ii ) .
the relation between the azimuthal angle @xmath4 , the obliquity angle @xmath92 and the aspect angle @xmath2 is as simple as @xmath256 for the azimuth angle @xmath4 measured from the direction of ismf in the plane of the sky .
the surface brightness of snr projection at distance @xmath95 from the center and at azimuth @xmath4 is @xmath257 where @xmath258 is the derivative of @xmath259 in respect to @xmath260 .
the azimuthal variation of the synchrotron x - ray brightness is approximately @xmath261 \end{array } \label{xazimuth : eq2}\ ] ] where @xmath262 d\bar a , \end{array } \label{xazimuth : eq2:int}\ ] ] reflects the dependence on @xmath80 , @xmath263 is @xmath245 for @xmath264 : @xmath265 note , that @xmath266 , i.e. @xmath255 depends in our approximation on the energy @xmath19 of observed x - ray photons .
let us approximate @xmath267 .
first , we use the approximations @xmath268 , @xmath269 , which are valid close to the shock ( appendix [ xmaps : app3 ] ) , @xmath88 is the shock compression ratio .
next , we expand @xmath270 in powers of the small parameter @xmath271 and consider the only first term of the decomposition : @xmath272 the exponential term in the integral expands in powers of the small parameter @xmath273 : @xmath274 in addition , @xmath263 is used instead of @xmath245 . close to the shock , the integral of interest is therefore @xmath275 where @xmath276 @xmath277.\ ] ] the parameter @xmath278 is responsible for the losses of emitting electrons and the time evolution of @xmath8 on the shock .
the value of @xmath90 is rather close to unity for @xmath225 ( table [ xmaps : table_param_gamma ] ) ; unless radiative losses ( the second term in @xmath105 ) are negligible , one may use @xmath279 for any @xmath0 .
other parameters are @xmath280 @xmath281 is given by eq .
( [ xmaps : appe - beta ] ) , @xmath282 parameters @xmath105 , @xmath106 , @xmath5 , @xmath107 and @xmath281 depend on @xmath92 and therefore on the aspect angle @xmath2 and the azimuth angle @xmath4 .
the parameter @xmath281 reflects differences between mf distribution downstream the shock of the different obliquity .
it varies from @xmath283 at parallel shock to @xmath284 at perpendicular one , eq .
( [ xmaps : appe - beta ] ) . in the approximate formulae
, it appears in the combination @xmath285 ; the role of @xmath286 $ ] is minor in modification of the approximate azimuthal and radial profiles .
therefore , in order to simplify the approximation , we may take @xmath287 . the index @xmath9 in ( [ xmaps : n_e_s ] ) , in general ,
is allowed to vary with @xmath54 , e.g. to be @xmath155 . in our approximation , due to ( [ f - delta ] ) ,
@xmath9 reflects the local slope of the electron spectrum appropriate to @xmath106 .
therefore , if one assumes @xmath156 , the index @xmath288 may vary with azimuth because @xmath106 varies , eq .
( [ xmaps : ep_m_x ] ) .
the final formula is @xmath99 i{_\mathrm{rx}}(\varphi,\bar\varrho;\epsilon{_\mathrm{f\| } } ) \label{xazimuth : eq2_fin}\ ] ] where only @xmath115 depends on @xmath126 and @xmath289 .
the formula eq .
( [ xazimuth : eq2_fin ] ) gives us the possibility to approximate both the azimuthal and the radial brightness profiles of x - ray brightness for @xmath290 close to unity .
it may be used ( with a bit larger errors compared to the case of ic emission ; fig .
[ xmaps : app - d - x ] , cf . fig .
[ xmaps : app - d - ic ] ) , for those azimuth @xmath4 where @xmath124 and @xmath125 , in the range of @xmath126 from @xmath291 to 1 , where @xmath128 , @xmath129 is the radius where the maximum in the radial profile of brightness happens .
we have in mind the maximum which is close to the shock , say @xmath292 ; therefore , in order to determine @xmath129 , one should look for the azimuth with the largest radiative losses .
this is discussed in details on example of the ic emission in sect . [
xmaps : app4 ] .
adiabatic index @xmath0 affects the approximation through @xmath88 , @xmath110 , @xmath90 .
in paper ii , we have developed an analytic approximation for the azimuthal variation of the surface brightness of sedov snr in @xmath0-rays due to the inverse - compton process , for regions close to the forward shock .
the approximation , eq .
( 11 ) , in the cited paper accounts to zeroth order .
however , like in the case of the x - ray brightness , the fall of the @xmath0-ray emissivity downstream of the shock is quite strong in case of the efficient radiative losses of electrons .
therefore , in such cases of the efficient losses we need to consider the next order of approximation . adopting the approach from the appendix [ xmaps : app1 ] to ic emission ( see also some details in paper ii ) , we come to the approximation @xmath293 i{_\mathrm{ic}}(\theta{_\mathrm{o , eff } } , \bar\rho ) \label{xmaps : icazimuth0}\ ] ] where the energy @xmath245 of electrons which gives maximum contribution to ic emission at photons with energy @xmath19 is ( e.g. * ? ? ?
* ) @xmath294 @xmath121 is the temperature of the black - body photons .
the factor @xmath295\ ] ] is approximately @xmath296 . \end{array } \label{xmaps : ig}\ ] ] where @xmath297 and @xmath88 comes from the approximations @xmath268 , @xmath298 , @xmath299 @xmath300 @xmath301 the final formula is @xmath117 i{_\mathrm{ic}}(\varphi,\bar\varrho;\epsilon{_\mathrm{f\| } } ) \label{xmaps : icazimuth}\ ] ] it gives us the possibility to approximate both the azimuthal and the radial brightness profiles for @xmath290 close to unity .
[ xmaps : app - d - ic ] demonstrates accuracy of the approximation ( [ xmaps : icazimuth ] ) ( left and middle panels show in fact the variation of @xmath302 because both @xmath96 and @xmath107 are constant there ) .
our calculations may be summarized as follows : this approximation may be used , with errors less than @xmath303 , for those azimuth @xmath4 where @xmath124 and @xmath125 , in the range of @xmath126 from @xmath291 to 1 , where @xmath128 , @xmath129 is the radius ( close to the shock ) where the maximum in the radial profile of brightness happens ; in addition , approximation may not be used for @xmath304 . if for some azimuth , the above conditions on @xmath106 and @xmath62 do not hold , the accuracy of approximation gradually decreases because the role of the exponent in @xmath1 and of the radiative losses may not be described by the first terms in the decompositions used for derivation of the formula .
let s consider fig .
[ xmaps : app - d - ic ] .
the photon energy @xmath106 does not change with azimuth for ic process .
on the left panels , the reduced fiducial energy @xmath305 for any azimuth : @xmath65 at the parallel shock and @xmath306 at the perpendicular shock . the approximation is accurate for any azimuth , for @xmath307 at @xmath132 and for a wider range of @xmath95 at @xmath308 .
middle panels on fig .
[ xmaps : app - d - ic ] show the same case except of @xmath46 . at parallel shock ( i.e. @xmath132 ) , the range for @xmath126 is smaller , @xmath309 ( lower panel ) .
therefore , the approximation of the azimuthal profile for @xmath310 is inaccurate ( upper panel , blue line ) , especially for @xmath311 where @xmath62 decreases ; it is @xmath312 .
the azimuthal profile is however accurate for @xmath313 ( black line ) .
similar situation is for variable @xmath8 ( right panels on fig .
[ xmaps : app - d - ic ] ) .
@xmath314 for considered model , therefore @xmath315 . therefore , in order to obtain a representative approximation , the lowest possible @xmath289 should be about @xmath316 .
we see from the figure that accuracy decreases toward smaller @xmath62 ( i.e. where the role of radiative losses are very efficient in modification of the electron distribution ) and for smaller @xmath126 . in general ,
the accuracy of the approximation is better for larger @xmath62 and smaller @xmath106 . with decreasing of the aspect angle @xmath2 , the accuracy of the approximations for the azimuthal profile increases at the beginning ( because contrasts in @xmath5 , @xmath62 and @xmath107 are lower ) and then decreases again , for the case of the quasi - parallel injection , because snr becomes centrally - brightened while our approximation is developed for regions close to the edge of snr .
( solid lines ) and its approximation ( [ xmaps : app - d : i ] ) ( dashed lines ) for azimuth @xmath308 ( black lines ) and @xmath132 ( blue lines ) , @xmath317 . for smaller aspects , @xmath318 , the approximation agree better with the numerical profiles . ]
an analytic approximation for the azimuthal variation of the radio surface brightness of sedov snr ( paper i ) may be extended to allow also for a description of the radial variation close to the forward shock .
namely , the correction consists in a factor @xmath319 : @xmath320 where @xmath319 is the same as for the x - ray approximation ( [ xmaps : ix ] ) .
accuracy of this approximation for the radial profile of brightness is demonstrated on fig .
[ xmaps : app - d - r ] and on fig .
2 in paper i for the azimuthal profiles .
@xmath319 varies with azimuth less than @xmath321 ( cf . e.g. black and blue dashed lines on fig .
[ xmaps : app - d - r ] ) .
this variation is due only to @xmath322 .
thus , @xmath281 may be taken constant with a good choice @xmath323 ( see also appendix [ xmaps : app1 ] ) . | distributions of nonthermal surface brightness of supernova remnants ( snrs ) contain important information about the properties of magnetic field and acceleration of charged particles . in the present paper ,
the synchrotron radio , x - ray , and inverse - compton ( ic ) @xmath0-ray maps of adiabatic snrs in uniform interstellar medium and interstellar magnetic field are modeled and their morphology is analyzed , with particular emphasis to comparison of azimuthal and radial variations of brightness in radio , x - rays , and @xmath0-rays .
approximate analytical formulae for the azimuthal and radial profiles of the synchrotron radio and x - ray as well as the ic @xmath0-ray brightness are derived .
they reveal the main factors which influence the pattern of the surface brightness distribution due to leptonic emission processes in shells of snrs and can account for some non - linear effects of acceleration if necessary .
these approximations provide observers and theorists with a set of simple diagnostic tools for quick analysis of the non - thermal maps of snrs .
[ firstpage ] ism : supernova remnants shock waves
ism : cosmic rays radiation mechanisms : non - thermal acceleration of particles |
in general , integrable one - dimensional models show an ideally conducting behavior in contrast to most real three - dimensional materials .
@xcite there has been intensive investigation of the influence of the integrability on the conductivity for various model systems.@xcite the conductance of conducting , nearly one - dimensional devices is , on the other hand , of substantial experimental interest . over the last years it has become possible to fabricate mesoscopic devices,@xcite such as carbon nanotubes which can be viewed as a realization of systems with ballistic transport properties .
therefore , the computation of dynamical transport quantities has received considerable interest .
a basic approach for the study of the conductance has been in the past via the bosonization of appropriate model systems,@xcite valid in the low - temperature limit .
numerical studies have so far involved the density - matrix - renormalization - group ( dmrg ) technique @xcite and monte carlo ( mc).@xcite in the case of ref .
a reduced set of states was used in order to evaluate the dynamics , in ref . , a phenomenological formula by sushkov@xcite was used to compute the conductance ( see also ref . ) . the simulations by refs .
use an effective bosonized hamiltonian as a starting point . here
we will discuss how to obtain the conductance with quantum - monte - carlo ( qmc ) on the original lattice hamiltonian .
for this purpose the conductance will be calculated on the imaginary frequency axis .
we will show that a reliable extrapolation to zero frequency can be performed at finite but low temperatures .
we will thus obtain approximation - free results for the dynamics of inhomogeneous quantum - spin systems at low but finite temperatures , within a well defined numerical accuracy defined by the statistics of the mc - sampling and the accuracy of the zero - frequency extrapolation . by the jordan - wigner transform a one - dimensional spinless fermionic system can be mapped to a hard - core boson model .
hence , it is possible to calculate the conductance for a fermionic system in a bosonic one .
this is vitally important since boson models can be easily analyzed by monte carlo simulations where the sign problem is absent .
however , for an evaluation of the conductance one requires a highly efficient simulation method which performs well at low temperatures .
recently,@xcite such a powerful method has been proposed : the stochastic series expansion ( sse ) . in this paper
we will compute the conductance in a hard - core boson lattice model by the aid of this new method .
we consider the anisotropic @xmath1-hamiltonian @xmath2 where the @xmath3 are the raising / lowering operators for spin-1/2 heisenberg spins .
the spin - current operator @xmath4 at a given site @xmath5 follows from the continuity equation and is given by ( see e.g. ref . ) @xmath6 as a perturbation we will use a local `` voltage drop '' . in the hard - core boson notation this corresponds to a step in chemical potential at site
@xmath7 which is equivalent to @xmath8 in terms of the heisenberg - spin operators .
the conductance is then defined as the dynamical response of the current operator at site @xmath9 to the voltage drop at site @xmath10 : @xmath11\rangle \,dt.\ ] ] for open boundary conditions ( obc ) the relation @xmath12=\hbar\ , j_x$ ] holds and a partial integration of ( [ defg ] ) yields : @xmath13\rangle % \frac{e^{izt}}{iz}\bigr|_0^\infty\!\!\!+{\rm re}\frac{i}{\hbar } % \int_0^\infty\ ! \frac{e^{izt}}{iz}\langle[j_x(t),j_y]\rangle \\ & = & \!{\rm re}\left[(-iz)^{-1}\frac{i}{\hbar}\bigl\{\langle [ j_x , p_y]\rangle - \int_0^\infty\!\ !
e^{izt}\langle[j_x(t),j_y]\rangle dt\bigr\}\right ] .
\\\end{aligned}\ ] ] using @xmath14 in the above equation gives two contributions to the conductance . with the definition of the generalized drude weight for two operators @xmath15 and @xmath16 : @xmath17 where @xmath18 one can show using the lehmann representation@xcite that the first contribution ( @xmath19 ) reads @xmath20 the second factor in the expression ( [ the_first_term ] ) [ namely , @xmath21 gives rise to a delta - function , such that ( [ the_first_term ] ) should be the dominating contribution to @xmath22 .
the first factor of expression ( [ the_first_term ] ) ( namely , @xmath23 ) is closely related to the drude peak @xmath24 where @xmath25 is the total current operator . from the discussion of the drude peak@xcite we know that under obc s the drude peak is zero [ because @xmath26 can be written as the response to a static twist which can be removed by a gauge transformation of the form @xmath27 ( see ref . ) ] whereas it is non - zero for periodic boundary conditions ( pbc s ) . in our case
@xmath23 is zero under obc s [ use a gauge transform @xmath28 . under pbc
s we find , because of translational invariance and because of the continuity equation , that @xmath29 does depend neither on @xmath9 nor @xmath10 .
this implies @xmath30 .
since the drude peak is finite ( at least for the models we are interested in ) we conclude that expression ( [ the_first_term ] ) vanishes even under pbc s in the thermodynamic limit .
so we obtain @xmath31{\rm re}\int_0^\infty e^{izt } \langle[j_x(t),j_y]\rangle dt . \label{intbypart1}\ ] ] restarting from eq .
( [ defg ] ) we may again by partial integration
arrive at another formula .
@xmath13\rangle % \frac{e^{izt}}{iz}\bigr|_0^\infty\!\!\!+{\rm re}\frac{i}{\hbar } % \int_0^\infty\ ! \frac{e^{izt}}{iz}\langle[j_x(t),j_y]\rangle \\ & = & \!{\rm re}\left[\frac{i}{\hbar}\bigl\{-\langle [ p_x , p_y]\rangle -(iz ) \int_0^\infty\!\ !
e^{izt}\langle[p_x(t),p_y]\rangle dt\bigr\}\right ] .
\\\end{aligned}\ ] ] the first term in the square brackets does not contribute as the potentials @xmath32 and @xmath33 commute and if we restrict ourselves to @xmath34 we obtain : @xmath35\rangle dt\right ) .
\label{intbypart2}\ ] ] the latter formula is especially useful for mc - simulations as it allows the computation of the conductance in terms of the diagonal @xmath36-@xmath36-correlators ( under obc s ) .
according to its definition as it is given by eq .
( [ defg ] ) the conductance might in principle depend on the actual choice of the positions of the voltage drop @xmath10 and the current measurement @xmath9 . here , we point out that in the limit @xmath37 this is not the case . in a rather general situation
one can show ( using the continuity equation ) that the right hand side of eq .
( [ defg ] ) gives the same result for any choice of @xmath9 and @xmath10 .
( see appendix [ proofxy ] . )
it is instructive to consider the free fermion case _
en dtail_. we denote the formal dependence on @xmath9 and @xmath10 by corresponding subscripts .
of course , in a translational invariant system @xmath38 depends only on the difference @xmath39 .
the conductance @xmath22 as a function of @xmath40 ( here and in the sequel @xmath34 ) is plotted for the free fermion case in fig .
one sees that the conductance in the limit @xmath41 approaches the universal value @xmath42 .
here we emphasize that a spatial separation of voltage drop and current measurement leads to an exponential decrease in @xmath43 at small @xmath44 .
therefore , we will restrict our attention to the cases @xmath45 for the rest of the paper .
we now turn to some technical details of our simulations . the second formula for the conductance , eq .
( [ intbypart2 ] ) , lends itself to a study with monte carlo simulations ( it requires obc s ) . at the matsubara frequencies @xmath46 we may use the equivalent expression @xmath47 we employ a standard qmc method ( sse ) to compute the conductance .
since @xmath32 is diagonal in the @xmath36-basis , the simulation of @xmath48 can be easily performed with the help of the sse.@xcite here , @xmath49 is the approximation order .
one may simply obtain @xmath50 as a linear combination of the terms @xmath48 with binomial weight factors @xmath51 .
we found it convenient to assume a gaussian distribution for the @xmath51 instead of a binomial one , because the former is easier to evaluate .
the error that we introduce by this replacement is smaller than the statistical error if @xmath52 .
( note that in our simulations @xmath49 is typically of the order of @xmath53 . )
what remains to be done in order to get @xmath54 is an integration in the final step .
we performed it with the simpson rule and a grid of 800 @xmath55-values .
we are now left with the standard problem of extrapolating @xmath43 from the matsubara frequencies @xmath56 to @xmath57 .
unfortunately , the spacing of the matsubara frequencies is linear in @xmath0 , so our method becomes unstable when we increase @xmath0 . to see that an application of our mc - method makes only sense at low temperatures we compare it to a simpler method : exact diagonalization .
[ cond ] shows @xmath43 for various system sizes @xmath58 at @xmath59 .
one sees that convergence with @xmath58 is rather fast ( at high temperatures ) .
hence , one can determine @xmath43 for @xmath60 with exact diagonalization .
to compute @xmath43 at some @xmath61 with mc - methods one needs to work at a temperature @xmath62 , and even then exact diagonalization is preferable as it yields @xmath43 in a continuous interval rather than on a discrete set of points .
hence , at high temperatures the mc - method is inferior to a simple exact diagonalization . in our simulation
we make one `` mc - sweep '' between two measurements which consists of one diagonal update and several loop - updates@xcite between two measurements .
we are able to run approximately @xmath63 sweeps . if @xmath64 then @xmath43 can be exactly evaluated for arbitrary system size not only at @xmath57 ( see below ) .
we can exploit this fact in two regards : firstly , we test our mc - method by comparing it with the exact curve obtained by jordan - wigner . the result can be seen in fig .
[ condmc ] .
secondly , we can test which frequencies ( and hence which temperatures ) we need such that a linear extrapolation can be carried out without introducing a larger error than the statistical one .
one sees also which system sizes are needed to determine @xmath43 without detectable finite size error .
our conclusion is that our method works for @xmath65 . at @xmath66
a system size of @xmath67 is appropriate .
one may derive a simple analytical result for the conductance in the free fermion case . here , we consider the slightly more difficult case but also more interesting as the system has a gap@xcite of a dimerized chain with magnetic field @xmath16 , i.e. , @xmath64 and the hopping parameter alternates : @xmath68 and @xmath69 .
the energy dispersion is given by @xmath70 and we assume a positive dispersion @xmath71 if @xmath72\cup[3\pi/2,2\pi]$ ] and @xmath73 else .
the gap @xmath74 is @xmath75 at @xmath76 .
the evaluation of the formula for the conductance eq .
( [ defg ] ) is in principle straightforward .
( see appendix [ apniceres ] . )
one obtains the compact result : @xmath77 .
\label{niceresult}\end{aligned}\ ] ] in the case of zero magnetic field this reduces to : @xmath78}.\ ] ] the @xmath79-value of the conductance as a function of magnetic field is quantized : it is zero if @xmath80 is smaller than the zero field gap @xmath81 or larger than the zero field band width @xmath82 , and it is @xmath83 between these values , and precisely at these values it is @xmath84 ( all values in units of @xmath42 ) . at low temperatures
the @xmath1-chain can be described by a luttinger liquid . for this model
the conductance was first obtained by apel and rice in the eighties:@xcite @xmath85 where @xmath86 .
this formula may be derived from eq .
( [ intbypart2 ] ) if we use concrete expressions for @xmath87 which are available from conformal field theory@xcite .
[ qmcgjz ] shows our qmc - results for @xmath43 on the imaginary axis for the @xmath1-model , in fig .
[ apelrice ] we display a comparison , as a function of @xmath88 , between the @xmath89 extrapolated from fig . [ qmcgjz ] , and the exact bosonization result , eq .
( [ apelrice ] ) .
we note , that the statistical error of the qmc - results presented in fig .
[ apelrice ] does not increase much with the parameter @xmath88 .
this is due to our using the `` directed loops '' as described in ref .. the choice for transition probabilities which was proposed there makes the sse - algorithm more effective . here
the improvement is remarkable .
now , we consider the hamiltonian @xmath90 i.e. , we add an impurity in the middle of the system .
this model was first studied in a paper by kane and fisher.@xcite later , this kind of model received considerable attention by other authors.@xcite by an rg approach kane and fisher found that the perturbation @xmath91 is relevant for repulsive interactions ( i.e. , @xmath92 ) .
this means that at zero @xmath0 the chemical potential anomaly `` cuts '' the system into two halves , such that the conductance is zero .
this result can not be directly confirmed by monte carlo methods since these are necessarily finite - temperature methods .
fortunately , the scaling behavior ( with temperature ) of the conductance is also known . for @xmath93
one may derive an exact formula for the conductance by a refermionization technique:@xcite @xmath94\ ] ] where @xmath95 is the digamma function .
so we can compare our mc - data once again with an exact result . in fig .
[ kfg2 ] we present two different qmc - results for the conductance on the imaginary axis of the heisenberg - chain with one impurity , for different impurity strengths . for the upper set of curves in fig .
[ kfg2 ] the position of voltage drop and the position of the current measurement are @xmath96 ; and for the lower set of curves they are @xmath97 . the curves are not as smooth as the ones in fig .
[ qmcgjz ] , so we use a quadratic fit from the first three matsubara frequencies instead of a linear extrapolation to estimate @xmath89 .
we also note that the curves with @xmath98 are better suited for extrapolation than those with @xmath99 because the slope at @xmath57 is smaller .
the statistical error is less than one percent .
the result from the extrapolation is given in fig .
[ sassetti ] along with the exact curves from eq .
( [ wes ] ) .
we used system sizes of @xmath100 for @xmath101 and @xmath102 for @xmath103 .
we performed @xmath104 mc - sweeps .
the error bars are smaller than the symbol size , so the error that we see in the figure is mainly due to our extrapolation method [ and to possible logarithmic finite - temperature - corrections to eq .
( [ wes ] ) ] .
we see that the quadratic fit tends to underestimate the correct value . several years after the publication of the eq .
( [ apelrice ] ) by apel and rice it was generally agreed upon@xcite that it does not reflect the ( correct ) physical behavior one would encounter in an experimental realization .
experiments are never performed on a closed system but on one coupled to reservoirs which make it possible for the particles to leave and enter the system .
these reservoirs can be modeled by attaching two leads consisting of infinite non - interacting spin half - chains to our model .
the complete hamiltonian reads then : @xmath105 i.e. , the interaction is confined to a small region in the middle consisting of @xmath106 sites .
this approach has been followed by many authors , e.g. refs .. generally , the presence of leads yields a conductance which is independent of @xmath88,@xcite namely , @xmath107 in sharp contrast to eq .
( [ apelrice ] ) .
the non - interacting semi - chains which we call leads play the rle of reservoirs .
we note that the behavior of @xmath43 depends on the parity of @xmath106 a fact which was already reported in ref . .
( a similar effect was also found in the hubbard model.@xcite ) in the following we will only consider the case of @xmath106 odd . a detailed discussion of @xmath106 even / odd and a comparison with ref .
will be presented elsewhere .
the `` natural '' choice for current measurement and voltage drop would be at the two ends of the interacting region . but
here is caution advised . from fig .
[ gxy]which is again for the free fermion case
one learns two things : firstly , @xmath22 does not depend on the choice of @xmath9 and @xmath10 . secondly ,
if @xmath9 and @xmath10 are some distance apart , convergence with @xmath44 becomes slow , hence one needs to go to lower @xmath0 and larger system sizes if one wants to extract @xmath89 reliably .
a simple phenomenological explanation for this is the following : if the place of the measurement is far from the voltage drop the particles have to travel a long distance and hence one has to wait a long time , before one can determine @xmath22 .
we can now place both the voltage drop and our current measurement at the middle of the system , but this should not help much .
the problem is that the particles still have to travel a long distance until they see the leads .
hence @xmath43 will be unaffected by the introduction of leads if @xmath44 is sufficiently large .
this expectation is confirmed by the qmc - data presented in fig .
[ leads ] . in this figure the exponential decay of @xmath43 at small @xmath44 is apparent , and the curves illustrate clearly that this decay is induced by the length scale @xmath106because it becomes stronger with increasing @xmath106 .
it is this decay that prevents us from discussing larger @xmath106 .
if we increase @xmath106 the decay becomes stronger , hence we need to evaluate @xmath43 for more ( and smaller ) frequencies in order to extract @xmath89 reliably .
but smaller frequencies are only available at smaller temperatures .
as we can not decrease @xmath0 much below @xmath108 we restrict ourselves to @xmath109 . if we considered a system with @xmath110 ( at @xmath111 ) we would not see any difference from a system without leads , because the difference occurs at small @xmath44 . the data presented in fig .
[ leads ] clearly indicates an upturn of @xmath43 for @xmath112 , indicating that @xmath89 is unaffected by the interaction in the low - temperature limit .
our result may , however , not be totally convincing , since we can only analyze relatively small interacting regions .
one might argue that the enhanced conductance is not due to the leads but simply to the reduced `` mean '' interaction which is close to zero as only few sites interact . to invalidate this argument we considered another model .
we have performed qmc - simulation of a system where we attach a lead only at _
one _ side such that we obtain a chain which is non - interacting in one half and interacting in the other .
for this system we found no deviation at all imaginary frequencies from the situation where the interacting region extends over the whole chain , even though there are as many interacting as interaction - free bonds .
monte carlo simulations allow the inclusion of long - range hopping , and thus breaking the integrability of the pure @xmath1-hamiltonian , as long as the resulting system is not frustrated .
we thus consider a hamiltonian with a third - nearest - neighbor interaction : @xmath113 .
\label{hamilhop3}\end{aligned}\ ] ] for simplicity we assumed that the anisotropy is independent of the hopping range ( i.e. , @xmath114 ) . in this context we emphasize , that the long - range hopping in the spin system does not transform under jordan - wigner to a long - range hopping in a fermionic system but to a more complicated four - sites operator . in general ,
adding a new term to the hamiltonian changes the current operator which is defined via the continuity - equation , @xmath115/\hbar$ ] . in a one - dimensional system with obc s the continuity - equation
is solved however by the relation @xmath116/\hbar$ ] ( see sec . [ subsec_def_g ] ) such that eq .
( [ intbypart2 ] ) still applies .
nonetheless , it is useful to look at the current operator for this case .
it reads : @xmath117 where @xmath118 if we compare it with the current operator of the @xmath1 chain we see that the long range hopping @xmath119 gives rise to three additional terms which are analogous to the first one .
we compute the conductance as a function of the hopping amplitudes @xmath120 and present the result in fig .
if @xmath121 the conductance is of course given by the apel - rice - result eq .
( [ apelrice ] ) .
however , if @xmath122 we may eventually neglect the nearest - neighbor - hopping - term such that we end up with three uncoupled chains .
thus , we conclude that the conductance will grow towards three times the apel - rice - result when we increase @xmath119 .
> from the figure we see that the crossover between these two values is shifted to smaller values of @xmath119 when the anisotropy is increased .
( in fact , at the isotropic point the increasing of @xmath22 is barely visible . ) in conclusion we have developed a qmc - technique which allows the evaluation of the dc - conductance for a wide range of non - frustrating quantum - spin chains at low but finite temperatures .
we have presented several stringent tests for this technique , like the kane - fisher scaling for the conductance through an impurity in a luttinger - liquid .
here we provide a general argument which relies on the ( physical ) assumption that some linear response functions are finite in the thermodynamic limit .
* theorem : * * proof : * first we consider @xmath123\rangle dt.$ ] this expression corresponds to a plateau value of the response function @xmath124\rangle$ ] , i.e. , @xmath125\right\rangle dt=\lim_{t\to \infty}\varphi(t).\ ] ] the response function may be written in the following way ( by kubo s identity)@xcite @xmath126 here @xmath127 is the mori scalar product [ for operators @xmath15 and @xmath16 : @xmath128 and @xmath49 is the liouville operator . to prove that @xmath129 as @xmath130 , it is sufficient to show that @xmath131 as a function of @xmath132 is bounded .
but this is just the assumption that we made in the statement of the theorem because @xmath133 represents the linear response of the operator @xmath134 to the perturbation @xmath135 .
finally , we can prove our main assertion .
we want to show @xmath136 .
this follows easily from @xmath137 and @xmath138 .
we will only consider only the second equality [ the proof of the first equality is analogous by the symmetric structure of eq .
( [ intbypart2 ] ) ] . using eq .
( [ intbypart2 ] ) and our previous result we see : @xmath139\rangle dt=0.\ ] ] q.e.d .
for eigenvalues @xmath140 and @xmath141 the respective one - particle - eigenstates will be denoted by @xmath142 and @xmath143 , the annihilation operators by @xmath144 and @xmath145 , and the occupation numbers by @xmath146 and @xmath147 .
cazalilla and j.b .
marston , `` time - dependent density - matrix renormalization group '' phys .
lett . * 88 * , 256403 ( 2002 ) ; _ ibid . _ * 91 * , 049702 , 2003 reply to comment h. g. luo , t. xiang , and x.q .
_ , * 91 * 049701 , ( 2003 ) . | recently , the stochastic series expansion ( sse ) has been proposed as a powerful mc - method , which allows simulations at low @xmath0 for quantum - spin systems .
we show that the sse allows to compute the magnetic conductance for various one - dimensional spin systems without further approximations .
we consider various modifications of the anisotropic heisenberg chain .
we recover the kane - fisher scaling for one impurity in a luttinger - liquid and study the influence of non - interacting leads for the conductance of an interacting system . |
this work was supported by the grants from the china postdoctoral science foundation ( 20060400104 ) , the k. c. wong education foundation ( hong kong ) , the national natural science foundation of china ( 10573003,10705041 ) , and the national basic research program of china ( 2003cb716300 ) .
a. g. riess _ et al .
_ [ supernova search team collaboration ] , _ observational evidence from supernovae for an accelerating universe and a cosmological constant , _ astron .
j. * 116 * , 1009 ( 1998 ) [ astro - ph/9805201 ] ; + s. perlmutter _ et al .
_ [ supernova cosmology project collaboration ] , _ measurements of omega and lambda from 42 high - redshift supernovae , _ astrophys .
j. * 517 * , 565 ( 1999 ) [ astro - ph/9812133 ] ; + p. astier _ et al .
_ , _ the supernova legacy survey : measurement of @xmath109 , @xmath110 and @xmath28 from the first year data set , _ astron .
astrophys . * 447 * , 31 ( 2006 ) [ astro - ph/0510447 ] .
m. tegmark _ et al . _
[ sdss collaboration ] , _ cosmological parameters from sdss and wmap , _ phys .
d * 69 * , 103501 ( 2004 ) [ astro - ph/0310723 ] ; + k. abazajian _ et al .
_ [ sdss collaboration ] , _ the second data release of the sloan digital sky survey , _ astron . j. * 128 * , 502 ( 2004 ) [ astro - ph/0403325 ] ; + k. abazajian _ et al . _
[ sdss collaboration ] , _ the third data release of the sloan digital sky survey , _ astron .
j. * 129 * , 1755 ( 2005 ) [ astro - ph/0410239 ] .
d. n. spergel _ et al . _
[ wmap collaboration ] , _ first year wilkinson microwave anisotropy probe ( wmap ) observations : determination of cosmological parameters , _ astrophys .
j. suppl .
* 148 * , 175 ( 2003 ) [ astro - ph/0302209 ] ; + d. n. spergel _ et al . _ ,
_ wilkinson microwave anisotropy probe ( wmap ) three year results : implications for cosmology , _ astro - ph/0603449 .
a. einstein , _ cosmological considerations in the general theory of relativity , _ sitzungsber .
berlin ( math .
* 1917 * ( 1917 ) 142 .
s. weinberg , _ the cosmological constant problem , _ rev .
phys . * 61 * 1 ( 1989 ) ; + v. sahni and a. a. starobinsky , _ the case for a positive cosmological lambda - term , _ int . j. mod .
d * 9 * , 373 ( 2000 ) [ astro - ph/9904398 ] ; + s. m. carroll , _ the cosmological constant , _ living rev .
rel . * 4 * 1 ( 2001 ) [ astro - ph/0004075 ] ; + p. j. e. peebles and b. ratra , _ the cosmological constant and dark energy , _ rev .
phys . * 75 * 559 ( 2003 ) [ astro - ph/0207347 ] ; + t. padmanabhan , _ cosmological constant : the weight of the vacuum , _ phys .
* 380 * 235 ( 2003 ) [ hep - th/0212290 ] ; + e. j. copeland , m. sami and s. tsujikawa , _ dynamics of dark energy , _ int .
d 15 ( 2006 ) 1753 [ hep - th/0603057 ] .
p. j. steinhardt , in _ critical problems in physics _ , edited by v. l. fitch and d. r. marlow ( princeton university press , princeton , nj , 1997 ) .
p. j. e. peebles and b. ratra , _ cosmology with a time variable cosmological constant , _ astrophys .
j. * 325 * l17 ( 1988 ) ; + b. ratra and p. j. e. peebles , _ cosmological consequences of a rolling homogeneous scalar field , _ phys . rev .
d * 37 * 3406 ( 1988 ) ; + c. wetterich , _ cosmology and the fate of dilatation symmetry , _ nucl .
b * 302 * 668 ( 1988 ) ; + j. a. frieman , c. t. hill , a. stebbins and i. waga , _ cosmology with ultralight pseudo nambu - goldstone bosons , _ phys .
* 75 * , 2077 ( 1995 ) [ astro - ph/9505060 ] ; + m. s. turner and m. j. white , _ cdm models with a smooth component , _ phys .
d * 56 * , 4439 ( 1997 ) [ astro - ph/9701138 ] ; + r. r. caldwell , r. dave and p. j. steinhardt , _ cosmological imprint of an energy component with general equation - of - state , _ phys .
lett . * 80 * , 1582 ( 1998 ) [ astro - ph/9708069 ] ; + a. r. liddle and r. j. scherrer , _ a classification of scalar field potentials with cosmological scaling solutions , _ phys .
d * 59 * , 023509 ( 1999 ) [ astro - ph/9809272 ] ; + i. zlatev , l. m. wang and p. j. steinhardt , _ quintessence , cosmic coincidence , and the cosmological constant , _ phys .
lett . * 82 * , 896 ( 1999 ) [ astro - ph/9807002 ] ; + p. j. steinhardt , l. m. wang and i. zlatev , _ cosmological tracking solutions , _ phys .
d * 59 * , 123504 ( 1999 ) [ astro - ph/9812313 ] . c. armendariz - picon , v. f. mukhanov and p. j. steinhardt , _ a dynamical solution to the problem of a small cosmological constant and late - time cosmic acceleration , _ phys .
lett . * 85 * , 4438 ( 2000 ) [ astro - ph/0004134 ] ; + c. armendariz - picon , v. f. mukhanov and p. j. steinhardt , _ essentials of k - essence , _ phys . rev .
d * 63 * , 103510 ( 2001 ) [ astro - ph/0006373 ] .
a. sen , _ tachyon matter , _ jhep * 0207 * , 065 ( 2002 ) [ hep - th/0203265 ] ; + t. padmanabhan , _ accelerated expansion of the universe driven by tachyonic matter , _ phys .
d * 66 * , 021301 ( 2002 ) [ hep - th/0204150 ] .
r. r. caldwell , _ a phantom menace ?
, _ phys .
b * 545 * , 23 ( 2002 ) [ astro - ph/9908168 ] ; + r. r. caldwell , m. kamionkowski and n. n. weinberg , _ phantom energy and cosmic doomsday , _ phys .
lett . *
91 * , 071301 ( 2003 ) [ astro - ph/0302506 ]
. n. arkani - hamed , h. c. cheng , m. a. luty and s. mukohyama , _ ghost condensation and a consistent infrared modification of gravity , _ jhep * 0405 * , 074 ( 2004 ) [ hep - th/0312099 ] ; + f. piazza and s. tsujikawa , _ dilatonic ghost condensate as dark energy , _ jcap * 0407 * , 004 ( 2004 ) [ hep - th/0405054 ] ; + j. f. zhang , x. zhang and h. y. liu , _ reconstructing generalized ghost condensate model with dynamical dark energy parametrizations and observational datasets , _ astro - ph/0612642 .
b. feng , x. l. wang and x. m. zhang , _ dark energy constraints from the cosmic age and supernova , _ phys .
b * 607 * , 35 ( 2005 ) [ astro - ph/0404224 ] ; + z. k. guo , y. s. piao , x. m. zhang and y. z. zhang , _ cosmological evolution of a quintom model of dark energy , _ phys .
b * 608 * , 177 ( 2005 ) [ astro - ph/0410654 ] ; + x. zhang , _ an interacting two - fluid scenario for quintom dark energy , _ commun . theor
. phys .
* 44 * , 762 ( 2005 ) ; + h. wei , r. g. cai and d. f. zeng , _ hessence : a new view of quintom dark energy , _ class .
* 22 * , 3189 ( 2005 ) [ hep - th/0501160 ] .
l. amendola , _ coupled quintessence , _ phys .
d * 62 * , 043511 ( 2000 ) [ astro - ph/9908023 ] ; + d. comelli , m. pietroni and a. riotto , _ dark energy and dark matter , _ phys .
b * 571 * , 115 ( 2003 ) [ hep - ph/0302080 ] ; + x. zhang , _ coupled quintessence in a power - law case and the cosmic coincidence problem , _ mod .
a * 20 * , 2575 ( 2005 ) [ astro - ph/0503072 ] . c. deffayet , g. r. dvali and g. gabadadze , _ accelerated universe from gravity leaking to extra dimensions , _ phys
. rev .
d * 65 * , 044023 ( 2002 ) [ astro - ph/0105068 ] ; + v. sahni and y. shtanov , _ braneworld models of dark energy , _ jcap * 0311 * , 014 ( 2003 ) [ astro - ph/0202346 ] .
a. y. kamenshchik , u. moschella and v. pasquier , _ an alternative to quintessence , _ phys .
b * 511 * , 265 ( 2001 ) [ gr - qc/0103004 ] .
e. witten , _ the cosmological constant from the viewpoint of string theory , _ hep - ph/0002297 .
g. t hooft , _ dimensional reduction in quantum gravity , _ gr - qc/9310026 ; + l. susskind , _ the world as a hologram , _ j. math .
* 36 * , 6377 ( 1995 ) [ hep - th/9409089 ] .
a. g. cohen , d. b. kaplan and a. e. nelson , _ effective field theory , black holes , and the cosmological constant , _ phys .
lett . * 82 * , 4971 ( 1999 ) [ hep - th/9803132 ] .
s. d. h. hsu , _ entropy bounds and dark energy , _ phys .
b * 594 * , 13 ( 2004 ) [ hep - th/0403052 ] .
m. li , _ a model of holographic dark energy , _ phys .
b * 603 * , 1 ( 2004 ) [ hep - th/0403127 ] .
w. fischler and l. susskind , _ holography and cosmology , _ hep - th/9806039 ; + b. wang , e. abdalla and t. osada , _ entropy and holography constraints for inhomogeneous universes , _ phys .
* 85 * , 5507 ( 2000 ) [ astro - ph/0006395 ] ; + p.
horava and d. minic , _ probable values of the cosmological constant in a holographic theory , _ phys .
* 85 * , 1610 ( 2000 ) [ hep - th/0001145 ] ; + r. g. cai , _ holography , the cosmological constant and the upper limit of the number of e - foldings , _ jcap * 0402 * , 007 ( 2004 ) [ hep - th/0312014 ] ; + d. a. lowe and d. marolf , _ holography and eternal inflation , _ phys . rev .
d * 70 * , 026001 ( 2004 ) [ hep - th/0402162 ] .
k. enqvist and m. s. sloth , _ possible connection between the location of the cutoff in the cosmic microwave background spectrum and the equation of state of dark energy , _ phys .
lett . *
93 * , 221302 ( 2004 ) [ hep - th/0406019 ] ; + k. ke and m. li , _ cardy - verlinde formula and holographic dark energy , _ phys .
b * 606 * , 173 ( 2005 ) [ hep - th/0407056 ] ; + q. g. huang and m. li , _ the holographic dark energy in a non - flat universe , _ jcap * 0408 * , 013 ( 2004 ) [ astro - ph/0404229 ] ; + q. g. huang and m. li , _ anthropic principle favors the holographic dark energy , _ jcap * 0503 * , 001 ( 2005 ) [ hep - th/0410095 ] ; + e. elizalde , s. nojiri , s. d. odintsov and p. wang , _ dark energy : vacuum fluctuations , the effective phantom phase , and holography , _ phys .
d * 71 * , 103504 ( 2005 ) [ hep - th/0502082 ] ; + h. li , z. k. guo and y. z. zhang , _ a tracker solution for a holographic dark energy model , _ int .
j. mod .
d * 15 * , 869 ( 2006 ) [ astro - ph/0602521 ] ; + x. zhang , _ reconstructing holographic quintessence , _ phys .
b * 648 * , 1 ( 2007 ) [ astro - ph/0604484 ] ; + x. zhang , _ dynamical vacuum energy , holographic quintom , and the reconstruction of scalar - field dark energy , _ phys .
d * 74 * , 103505 ( 2006 ) [ astro - ph/0609699 ] ; + j. zhang , x. zhang and h. liu , _ holographic tachyon model , _ phys .
b * 651 * , 84 ( 2007 ) [ arxiv:0706.1185 [ astro - ph ] ] ; + j. zhang , x. zhang and h. liu , _ holographic dark energy in a cyclic universe , _ eur .
j. c ( 2007 ) , in press , arxiv:0708.3121 [ hep - th ] ; + y. z. ma and x. zhang , _ theoretical limits of holographic quintessence , _ arxiv:0709.1517 [ astro - ph ] .
x. zhang and f. q. wu , _
constraints on holographic dark energy from type ia supernova observations , _ phys .
d * 72 * , 043524 ( 2005 ) [ astro - ph/0506310 ] ; + z. chang , f. q. wu and x. zhang , _ constraints on holographic dark energy from x - ray gas mass fraction of galaxy clusters , _ phys .
b * 633 * , 14 ( 2006 ) [ astro - ph/0509531 ] ; + q. g. huang and y. g. gong , _ supernova constraints on a holographic dark energy model , _ jcap * 0408 * , 006 ( 2004 ) [ astro - ph/0403590 ] ; + k. enqvist , s. hannestad and m. s. sloth , _ searching for a holographic connection between dark energy and the low - l cmb multipoles , _ jcap * 0502 * 004 ( 2005 ) [ astro - ph/0409275 ] ; + j. shen , b. wang , e. abdalla and r. k. su , _
constraints on the dark energy from the holographic connection to the small l cmb suppression , _ phys .
b * 609 * 200 ( 2005 ) [ hep - th/0412227 ] ; + h. c. kao , w. l. lee and f. l. lin , _ cmb constraints on the holographic dark energy model , _ phys .
d * 71 * 123518 ( 2005 ) [ astro - ph/0501487 ] ; + x. zhang and f. q. wu , _
constraints on holographic dark energy from latest supernovae , galaxy clustering , and cosmic microwave background anisotropy observations , _ phys .
d * 76 * , 023502 ( 2007 ) [ astro - ph/0701405 ] .
b. wang , y. gong and e. abdalla , _ transition of the dark energy equation of state in an interacting holographic dark energy model , _ phys .
b * 624 * , 141 ( 2005 ) [ hep - th/0506069 ] ; + h. kim , h. w. lee and y. s. myung , _ equation of state for an interacting holographic dark energy model , _ phys .
b * 632 * , 605 ( 2006 ) [ gr - qc/0509040 ] ; + b. hu and y. ling , _ interacting dark energy , holographic principle and coincidence problem , _ phys .
d * 73 * , 123510 ( 2006 ) [ hep - th/0601093 ] ; + z. k. guo , n. ohta and s. tsujikawa , _ probing the coupling between dark components of the universe , _ astro - ph/0702015 .
v. sahni , t. d. saini , a. a. starobinsky and u. alam , _ statefinder a new geometrical diagnostic of dark energy , _ jetp lett .
* 77 * , 201 ( 2003 ) [ astro - ph/0201498 ] .
u. alam , v. sahni , t. d. saini and a. a. starobinsky , _ exploring the expanding universe and dark energy using the statefinder diagnostic , _ mon . not .
soc . * 344 * , 1057 ( 2003 ) [ astro - ph/0303009 ] .
l. amendola , _ acceleration at @xmath111 ? , _ mon . not .
soc . * 342 * , 221 ( 2003 ) [ astro - ph/0209494 ] ; + l. p. chimento , a. s. jakubi , d. pavon and w. zimdahl , _ interacting quintessence solution to the coincidence problem , _ phys .
d * 67 * , 083513 ( 2003 ) [ astro - ph/0303145 ] ; + s. das , p. s. corasaniti and j. khoury , _ super - acceleration as signature of dark sector interaction , _ phys .
d * 73 * , 083509 ( 2006 ) [ astro - ph/0510628 ] .
w. zimdahl and d. pavon , _ statefinder parameters for interacting dark energy , _ gen .
* 36 * , 1483 ( 2004 ) [ gr - qc/0311067 ] .
v. gorini , a. kamenshchik and u. moschella , _ can the chaplygin gas be a plausible model for dark energy ?
, _ phys .
d * 67 * , 063509 ( 2003 ) [ astro - ph/0209395 ] .
v. sahni , _ dark matter and dark energy , _ lect .
notes phys .
* 653 * , 141 ( 2004 ) [ astro - ph/0403324 ] .
x. zhang , f. q. wu and j. zhang , _ new generalized chaplygin gas as a scheme for unification of dark energy and dark matter , _ jcap * 0601 * , 003 ( 2006 ) [ astro - ph/0411221 ] .
x. zhang , _ statefinder parameters for coupled quintessence scenario in a power law case , _ commun .
phys . * 44 * , 573 ( 2005 ) .
x. zhang , _ statefinder diagnostic for coupled quintessence , _ phys .
b * 611 * , 1 ( 2005 ) [ astro - ph/0503075 ] .
x. zhang , _ statefinder diagnostic for holographic dark energy model , _ int .
j. mod .
d * 14 * , 1597 ( 2005 ) [ astro - ph/0504586 ] .
m. p. dabrowski , _
statefinders , higher - order energy conditions and sudden future singularities , _ phys .
b * 625 * , 184 ( 2005 ) [ gr - qc/0505069 ] .
p. x. wu and h. w. yu , _ statefinder parameters for quintom dark energy model , _ int .
j. mod .
d * 14 * , 1873 ( 2005 ) [ gr - qc/0509036 ] .
m. g. hu and x. h. meng , _ bulk viscous cosmology : statefinder and entropy , _ phys .
b * 635 * , 186 ( 2006 ) [ astro - ph/0511615 ] . v. sahni and a. starobinsky , _ reconstructing dark energy , _ astro - ph/0610026 .
m. r. setare , j. zhang and x. zhang , _ statefinder diagnosis in non - flat universe and the holographic model of dark energy , _ jcap * 0703 * , 007 ( 2007 ) [ gr - qc/0611084 ] .
z. l. yi and t. j. zhang , _ statefinder diagnostic for the modified polytropic cardassian universe , _ phys .
d * 75 * , 083515 ( 2007 ) [ astro - ph/0703630 ] .
b. r. chang , h. y. liu , l. x. xu , c. w. zhang and y. l. ping , _ statefinder parameters for interacting phantom energy with dark matter , _ jcap * 0701 * , 016 ( 2007 ) [ astro - ph/0612616 ] .
b. chang , h. liu , l. xu and c. zhang , _ statefinder parameters for five - dimensional cosmology , _ arxiv:0704.3670 [ astro - ph ] . c. cattoen and m. visser , _ cosmography : extracting the hubble series from the supernova data , _ gr - qc/0703122 .
r. r. caldwell and e. v. linder , _ the limits of quintessence , _ phys .
lett . * 95 * , 141301 ( 2005 ) [ astro - ph/0505494 ] .
r. j. scherrer , _ dark energy models in the @xmath112 plane , _ phys .
d * 73 * , 043502 ( 2006 ) [ astro - ph/0509890 ] .
t. chiba , _
@xmath28 and @xmath94 of scalar field models of dark energy in the large , _ phys .
d * 73 * , 063501 ( 2006 ) [ astro - ph/0510598 ] .
z. k. guo , y. s. piao , x. zhang and y. z. zhang , _ two - field quintom models in the @xmath1 plane , _ phys .
d * 74 * , 127304 ( 2006 ) [ astro - ph/0608165 ] .
d. bak and s. j. rey , _ cosmic holography , _ class .
* 17 * , l83 ( 2000 ) [ hep - th/9902173 ] .
e. e. flanagan , d. marolf and r. m. wald , _ proof of classical versions of the bousso entropy bound and of the generalized second law , _ phys .
d * 62 * , 084035 ( 2000 ) [ hep - th/9908070 ] . | in this paper , we investigate the holographic dark energy model with interaction between dark energy and dark matter , from the statefinder viewpoint .
we plot the trajectories of the interacting holographic dark energy model for different interaction cases as well as for different values of the parameter @xmath0 in the statefinder - plane .
the statefinder diagrams characterize the properties of the holographic dark energy and show the discrimination between the two cases with and without interaction . as a result
, we show the influence of the interaction on the evolution of the universe in the statefinder diagrams . moreover , as a complement to the statefinder diagnosis
, we study the interacting holographic dark energy model in the @xmath1 plane , which can provide us with a dynamical diagnosis .
* statefinder diagnosis for the interacting model of holographic dark energy * jingfei zhang,@xmath2 xin zhang,@xmath3 and hongya liu@xmath2 today it has been confirmed that our universe is undergoing an accelerating expansion through numerous cosmological observations , such as type ia supernovae ( snia ) @xcite , large scale structure ( lss ) @xcite and cosmic microwave background ( cmb ) @xcite .
this cosmic acceleration is attributed to a mysterious dominant component , dark energy , with negative pressure .
the combined analysis of cosmological observations suggests that the universe is spatially flat , and consists of about @xmath4 dark energy , @xmath5 dust matter , and negligible radiation .
many candidates have been proposed to interpret or describe the properties of dark energy , though its nature still remains enigmatic .
the most obvious theoretical candidate of dark energy is the cosmological constant @xmath6 @xcite which has the equation of state @xmath7 .
however , as is well known , there are two difficulties arise from the cosmological constant scenario , namely the two famous cosmological constant problems the `` fine - tuning '' problem and the `` cosmic coincidence '' problem @xcite .
theorists have made lots of efforts to try to resolve the cosmological constant problem but all these efforts were turned out to be unsuccessful . also , there is an alternative proposal to dark energy the dynamical dark energy scenario .
the dynamical dark energy scenario is often realized by some scalar field mechanism which suggests that the energy form with negative pressure is provided by a scalar field evolving down a proper potential .
a lot of scalar - field dark energy models have been studied , including quintessence @xcite , k - essence @xcite , tachyon @xcite , phantom @xcite , ghost condensate @xcite and quintom @xcite etc .. in addition , other proposals on dark energy include scenarios of interacting dark energy @xcite , braneworld @xcite , chaplygin gas @xcite , and so forth . by far , obviously , it is not yet clear if dark energy is a cosmological constant or a dynamical field .
generally , theorists believe that we can not entirely understand the nature of dark energy before a complete theory of quantum gravity is established @xcite . however
, in this circumstance , we still can make some efforts to probe the properties of dark energy according to some principle of quantum gravity .
the holographic dark energy model is an example of such effort , which stems from the holographic principle and can provide us with an intriguing way to interpret the dynamics of dark energy .
the holographic principle is an important result of the recent researches of exploring the quantum gravity and is enlightened by investigations of the quantum property of black holes @xcite . according to the holographic principle ,
the number of degrees of freedom for a system within a finite region should be finite and should be bounded roughly by the area of its boundary . in the cosmological context , the holographic principle will set an upper bound on the entropy of the universe .
motivated by the bekenstein entropy bound , it seems plausible that one may require that for an effective quantum field theory in a box of size @xmath8 with uv cutoff @xmath9 , the total entropy should satisfy @xmath10 , where @xmath11 is the entropy of a black hole with the same size @xmath8 .
however , cohen et al .
@xcite pointed out that to saturate this inequality some states with schwartzschild radius much larger than the box size have to be counted in . as a result ,
a more restrictive bound , the energy bound , has been proposed to constrain the degrees of freedom of the system , requiring the total energy of a system with size @xmath8 not to exceed the mass of a black hole with the same size , namely , @xmath12 .
this means that the maximum entropy is in the order of @xmath13 .
when we take the whole universe into account , the vacuum energy related to this holographic principle is viewed as dark energy , usually dubbed holographic dark energy . the largest ir cut - off @xmath8
is chosen by saturating the inequality , so that we get the holographic dark energy density @xmath14 where @xmath0 is a numerical constant is introduced to parameterize some uncertainties , such as the species of quantum fields in the universe , the effect of curved spacetime , and so forth . ]
( note that @xmath15 is assumed ) , and as usual @xmath16 is the reduced planck mass .
if we take @xmath8 as the size of the current universe , for instance the hubble scale @xmath17 , then the dark energy density will be close to the observed value .
however , hsu @xcite pointed out that this yields a wrong equation of state for dark energy .
li @xcite subsequently proposed that the ir cutoff @xmath8 should be given by the future event horizon of the universe , @xmath18 such a holographic dark energy looks reasonable , since it may provide simultaneously natural solutions to both dark energy problems , as demonstrated in ref . @xcite .
meanwhile , other applications of the holographic principle in cosmology @xcite show that holography is an effective way to investigate cosmology .
for other extensive studies , see e.g. @xcite@xcite . besides
, some interacting models are discussed in many works because these models can help to understand or alleviate the coincidence problem by considering the possible interaction between dark energy and cold dark matter due to the unknown nature of dark energy and dark matter .
in addition , the proposal of interacting dark energy is compatible with the current observations such as the snia and cmb data @xcite . for the interacting model of holographic dark energy see @xcite . on the other hand ,
since more and more dark energy models have been constructed for interpreting or describing the cosmic acceleration , the problem of discriminating between the various contenders is becoming emergent . in order to be capable of differentiating between those competing cosmological scenarios involving dark energy ,
a sensitive and robust diagnosis for dark energy models is a must .
in addition , for some geometrical models arising from modifications to the gravitational sector of the theory , the equation of state no longer plays the role of a fundamental physical quantity , so it would be very useful if we could supplement it with a diagnosis which could unambiguously probe the properties of all classes of dark energy models . for this purpose a diagnostic proposal that makes use of parameter pair @xmath19 , the so - called statefinder " ,
was introduced by sahni et al . @xcite .
the statefinder probes the expansion dynamics of the universe through higher derivatives of the scale factor @xmath20 and is a `` geometrical '' diagnosis in the sense that it depends on the scale factor and hence on the metric describing space - time . since different cosmological models involving dark energy exhibit different evolution trajectories in the @xmath21 plane
, the statefinder can be used to diagnose different dark energy models @xcite . in this paper , we focus on a model of holographic dark energy with interaction between dark energy and dark matter and study the influence of the interaction to the cosmic evolution .
moreover , we use the statefinder to diagnose various cases with different interaction strength and different parameter @xmath0 in the holographic model .
let us start with a spatially flat friedmann - robertson - walker ( frw ) universe with dust matter and holographic dark energy .
the friedmann equation reads @xmath22 where @xmath23 is the energy density of matter and @xmath24 is the dark energy density .
the total energy density satisfies a conservation law , @xmath25 where @xmath26 is the total energy density of the universe , and @xmath27 is the total pressure ( @xmath28 denotes the equation of state of dark energy ) .
note that since the matter component is mainly contributed by the cold dark matter , we ignore the contribution of the baryon matter here for simplicity . by introducing @xmath29 and @xmath30
, the friedmann equation can also be written as @xmath31 .
furthermore , if we proceed to consider a scenario of interacting dark energy , @xmath23 and @xmath32 do not satisfy independent conservation laws , they instead satisfy @xmath33 and @xmath34 where @xmath35 describes the interaction between dark energy and dark matter .
it is obvious that the interaction term @xmath35 could not be introduced by considering some micro - process currently , so a phenomenological way is the must .
one possible choice for the interaction term is setting @xmath36 where @xmath37 is a constant describing the coupling strength .
this expression for the interaction term was first introduced in the study of the suitable coupling between a quintessence scalar field and a pressureless cold dark matter component , in order to get a scaling solution to the coincidence problem @xcite .
taking the ratio of energy densities as @xmath38 and using the friedmann equation @xmath39 , we have @xmath40 and @xmath41 .
furthermore , from ( [ a1 ] ) , ( [ a2 ] ) and ( [ q ] ) , we obtain @xmath42 combining these results , we easily get the equation of state of dark energy @xmath43 where prime denotes the derivative with respect to @xmath44 . using the definition of the holographic dark energy
( [ de ] ) and the friedmann equation , the future event horizon ( [ eh ] ) can be expressed as @xmath45 .
then , for this expression , taking the derivative with respect to @xmath46 and reducing the result , we get @xmath47 . \label{5}\ ] ] it is notable that this differential equation governs the whole dynamics of the interacting model of holographic dark energy .
substituting ( [ 5 ] ) to ( [ 4 ] ) yields @xmath48 then we can compute the deceleration parameter @xmath49 with a fixed parameter @xmath0 . in this plot , we take @xmath50 , @xmath51 , and vary @xmath52 as 0 , 0.02 , 0.06 , and 0.10 , respectively.,width=377 ] with a fixed coupling @xmath52 . in this plot , we take @xmath53 , @xmath54 , and vary @xmath0 as 0.9 , 1.0 and 1.1 , respectively.,width=377 ] in order to show the influence of interaction to the cosmic evolution , the cases with dependence of the parameter @xmath52 for the deceleration parameter @xmath55 are shown in fig .
1 . in fig . 1 , we fix @xmath50 and take the coupling constant @xmath52 as 0 , 0.02 , 0.06 , and 0.10 , respectively . besides , the cases with a fixed @xmath52 and various values of @xmath0 are also interesting . in fig .
2 , fixing the coupling constant @xmath53 , we plot the evolution diagram of the deceleration parameter @xmath55 with different values of parameter @xmath0 ( here we take the values of @xmath0 as 0.9 , 1.0 , and 1.1 , respectively ) . from figs . 1 and 2 we learn that the universe experienced an early deceleration and a late time acceleration .
fig .
1 shows that , for a fixed parameter @xmath0 , the cosmic acceleration starts earlier for the cases with interaction than the ones without coupling ( for this point see also , e.g. , @xcite ) .
moreover , the stronger the coupling between dark energy and dark matter is the earlier the acceleration of universe began .
however , the cases with smaller coupling will get bigger acceleration finally in the far future .
in addition , fig .
2 shows that the acceleration starts earlier when @xmath0 is larger for the same coupling @xmath52 , but finally a smaller @xmath0 will lead to a bigger acceleration .
it should be pointed out that , in the interacting holographic dark energy model , the interaction strength has an upper limit because of the evolutionary behavior of the holographic dark energy . for detailed discussions about correlation of the coupling @xmath52 and the parameter @xmath0 ,
see @xcite .
it is remarkable that , with the interaction between dark energy and dark matter , the case of @xmath50 could not enter a de sitter phase in the infinite future . in short , the influence of the interaction between dark energy and dark matter to the cosmic evolution is obvious , as manifested by figs . 1 and 2 . on the other hand , nevertheless , as eq .
( [ 6 ] ) shows , though the deceleration parameter @xmath55 carries the information of the equation of state of dark energy @xmath28 , the property of dynamical evolution for @xmath28 can not be read out from @xmath55 . for diagnosing properties and evolutionary behaviors of dark energy models exquisitely ,
more powerful diagnostic tool is a must .
now we turn to the statefinder diagnosis . for characterizing the expansion history of the universe
, one defines the geometric parameters @xmath56 and @xmath57 , namely the hubble parameter and the deceleration parameter .
it is clear that @xmath58 means the universe is undergoing an expansion and @xmath59 means the universe is experiencing an accelerated expansion . from the cosmic acceleration , @xmath60 ,
one infers that there may exist dark energy with negative equation of state , @xmath61 and likely @xmath62 , but it is hard to deduce the information of the dynamical property of @xmath28 ( namely the time evolution of @xmath28 ) from the value of @xmath55 . in order to extract the information on the dynamical evolution of @xmath28 , it seems that we need the higher time derivative of the scale factor , @xmath63 .
another motivation for proposing the statefinder parameters stems from the merit that they can provide us with a diagnosis which could unambiguously probe the properties of all classes of dark energy models including the cosmological models without dark energy describing the cosmic acceleration .
though at present we can not extract sufficiently accurate information of @xmath64 and @xmath63 from the observational data , we can expect , however , the high - precision observations of next decade may be capable of doing this .
since different cosmological models exhibit different evolution trajectories in the @xmath21 plane , the statefinder parameters can thus be used to diagnose the evolutionary behaviors of various dark energy models and discriminate them from each other . in this paper , we apply the statefinder diagnosis to the interacting holographic dark energy model .
the expansion rate of the universe is described by the hubble parameter @xmath65 , and the rate of acceleration / deceleration of the expanding universe is characterized by the deceleration parameter @xmath55 .
furthermore , in order to find a more sensitive discriminator of the expansion rate , let us consider the general expansion form for the scale factor of the universe @xmath66 generically , various dark energy models give rise to families of curves @xmath67 having vastly different properties . in principle , we can confine our attention to small value of @xmath68 in ( [ aexp ] ) because the acceleration of the universe is a fairly recent phenomenon .
then , we see , following @xcite , that a new diagnostic of dark energy dubbed statefinder can be constructed using both second and third derivatives of the scale factor . the second derivative is encoded in the deceleration parameter @xmath55 , and the third derivative is contained in the statefinder parameters @xmath19 .
the statefinder parameters @xmath19 are defined as @xmath69 note that the parameter @xmath70 is also called cosmic jerk .
thus the set of quantities describing the geometry is extended to include @xmath71 .
trajectories in the @xmath21 plane corresponding to different cosmological models exhibit qualitatively different behaviors , so the statefinder can be used to discriminate different cosmological models . the spatially flat lcdm ( cosmological constant @xmath6 with cold dark matter ) scenario corresponds to a fixed point in the diagram @xmath72 departure of a given dark energy model from this fixed point provides a good way of establishing the `` distance '' of this model from spatially flat lcdm @xcite . as demonstrated in refs .
@xcite@xcite , the statefinder can successfully differentiate between a wide variety of dark energy models including the cosmological constant , quintessence , phantom , quintom , the chaplygin gas , braneworld models and interacting dark energy models , etc .. we can clearly identify the `` distance '' from a given dark energy model to the lcdm scenario by using the @xmath73 evolution diagram .
the current location of the parameters @xmath74 and @xmath70 in these diagrams can be calculated in models .
the current values of @xmath74 and @xmath70 are evidently valuable since we expect that they can be extracted from data coming from snap ( supernovae acceleration probe ) type experiments .
therefore , the statefinder diagnosis combined with future snap observations may possibly be used to discriminate between different dark energy models .. the statefinder parameter - pair also can be expressed as @xmath75 where @xmath76 is the total density and @xmath77 is the total pressure .
then , by using the friedmann equation , we can obtain the following concrete expressions @xmath78 directly , from eq .
( [ w ] ) , we have @xmath79,\label{w1}\end{aligned}\ ] ] where the prime denotes the derivative with respect to @xmath44 .
note that the whole dynamics of the universe in the interacting holographic dark energy model is governed by the differential equation ( [ 5 ] ) .
so by solving eq .
( [ 5 ] ) we can get the evolution solution of @xmath80 and then hold all the cosmological quantities of interest and the whole dynamics of the universe . for the interacting holographic dark energy with a fixed parameter @xmath0 and different coupling @xmath52
. selected curves of @xmath73 are plotted by fixing @xmath50 , @xmath51 and varying @xmath52 as 0 , 0.02 , 0.06 and 0.10 , respectively .
a star denotes the lcdm fixed point @xmath81 .
the dots show today s values for the statefinder parameters @xmath82.,width=377 ] @xmath83{rsb0.eps } & \includegraphics[scale=0.8]{rsb01.eps } \\ \end{array}$ ] in what follows we shall diagnose the interacting holographic dark energy model employing the statefinder method .
we shall analyze the cases with fixed coupling constant @xmath52 and with fixed parameter @xmath0 , respectively .
as demonstrated above , the information of this model can be acquired by solving the differential equation ( [ 5 ] ) . making the redshift @xmath84 vary in a large enough range involving far future and far past
, one can solve the differential equation ( [ 5 ] ) numerically and then get the evolution trajectories in the statefinder @xmath21 planes for this model .
for instance , we plot the statefinder diagram in fig . 3 for the cases of @xmath50 with various values of coupling such as @xmath85 , @xmath86 , @xmath87 and @xmath88 , meanwhile the present density parameter of dark energy
is taken to be @xmath51 .
the case @xmath85 corresponds to the holographic dark energy model without interaction between dark energy and dark matter .
the arrows in the diagram denote the evolution directions of the statefinder trajectories and the star corresponds to @xmath89 representing the lcdm model .
this diagram shows that the evolution trajectories with different interaction strengths exhibit different features in the statefinder plane .
when the interaction is absent , the @xmath73 curve for holographic dark energy ends at the lcdm fixed point , i.e. , the universe of this case will evolve to the de sitter phase in the far future .
however , taking the interaction into account , the endpoints of the @xmath73 curves could not arrive at the lcdm fixed point @xmath81 , though all of the evolution trajectories tend to approach this point .
it should be mentioned that the statefinder diagnosis for holographic dark energy model without interaction has been investigated in detail in @xcite , where the focus is put on the diagnosis of the different values of parameter @xmath0 .
the statefinder analysis on the holographic dark energy in a non - flat universe see @xcite . in @xcite
, it has been demonstrated that from the statefinder viewpoint @xmath0 plays a significant role in this model and it leads to the values of @xmath90 in today and future tremendously different . in this paper , by far , we have clearly seen that the interaction between holographic dark energy and dark matter makes the statefinder evolutionary trajectories with the same value of @xmath0 tremendously different also .
if the accurate information of @xmath91 can be extracted from the future high - precision observational data in a model - independent manner , these different features in this model can be discriminated explicitly by experiments , one thus can use this method to test the holographic dark energy model as well as other dark energy models .
hence , today s values of @xmath92 play a significant role in the statefinder diagnosis .
we thus calculate the present values of the statefinder parameters for different cases in the interacting holographic dark energy model and mark them on evolution curves with dots .
it can be seen that stronger interaction results in longer distance to the lcdm fixed point .
the interaction between holographic dark energy and dark matter prevents the holographic dark energy from behaving as a cosmological constant @xmath6 ultimately in the far future .
we also plotted the statefinder diagram in the @xmath21 plane for different values of parameter @xmath0 with @xmath85 and @xmath88 in fig .
4 .
the left panel is for the holographic dark energy without interaction while the right one is for the case involving the interaction .
the star in the figure also corresponds to the lcdm fixed point and the dots marked on the curves represent the present values of the statefinder parameters .
note that the true values of @xmath82 of the universe should be determined in a model - independent way , we can only pin our hope on the future experiments to achieve this .
we strongly expect that the future high - precision experiments ( e.g. snap ) may provide sufficiently large amount of precise data to release the information of statefinders @xmath93 in a model - independent manner so as to supply a way of discriminating different cosmological models with or without dark energy . from fig .
4 , we can learn that the @xmath73 evolutions have the similar behavior , i.e. the curves almost start from a fixed point for both cases in the @xmath21 plane .
evidently , the interaction between dark components makes the value of @xmath70 smaller and the value of @xmath74 bigger .
also , obviously , the parameter @xmath0 plays a crucial role in the holographic model .
@xmath83{ww1b0.eps } & \includegraphics[scale=0.8]{ww1b01.eps } \end{array}$ ] as a complement to statefinder diagnosis , we investigate the dynamical property of the interacting holographic dark energy in the @xmath1 phase plane , where @xmath94 represents the derivative of @xmath28 with respect to @xmath95 .
recently , this method became somewhat popular for analyzing dark energy models .
caldwell and linder @xcite proposed to explore the evolving behavior of quintessence dark energy models and test the limits of quintessence in the @xmath1 plane , and they showed that the area occupied by quintessence models in the phase plane can be divided into thawing and freezing regions .
then , the method was used to analyze the dynamical property of other dark energy models including more general quintessence models @xcite , phantom models @xcite and quintom models @xcite , etc .. the @xmath1 analysis undoubtedly provides us with an alternative way of classifying dark energy models using the quantities describing the dynamical property of dark energy .
but , it is obviously that the @xmath96 pair is related to statefinder pair @xmath97 in a definite way , see eqs .
( [ r ] ) and ( [ s ] ) .
the merit of the statefinder diagnosis method is that the statefinder parameters are constructed from the scale factor @xmath98 and its derivatives , and they are expected to be extracted in a model - independent way from observational data , although it seems hard to achieve this at present . while the advantage of the @xmath1 analysis is that it is a direct dynamical diagnosis for dark energy .
hence , the statefinder @xmath21 geometrical diagnosis and the @xmath1 dynamical diagnosis can be viewed as complementarity in some sense .
now let us investigate the interacting model of holographic dark energy in the @xmath1 plane . in fig . 5 , we plot the evolutionary trajectories of the holographic dark energy in the @xmath1 plane where the selected curves correspond to @xmath99 , @xmath100 and @xmath101 , respectively .
the left graph is an illustrative example without interaction to which we can compare the evolution of the interacting holographic dark energy in the right diagram .
fig .
5 shows clearly that the parameter @xmath0 and the interaction @xmath52 both play important roles in the evolution history of the universe .
the left graph tells us : @xmath102 makes the holographic dark energy behave as quintessence - type dark energy with @xmath103 and @xmath104 makes the holographic dark energy behave as quintom - type dark energy with @xmath28 crossing @xmath105 during the evolution history .
however , when the interaction between dark components is present , the situation becomes somewhat ambiguous because that the equation of state @xmath28 loses the ability of classifing dark energies definitely , due to the fact that the interaction makes dark energy and dark matter be entangled in each other . in this circumstance , the conceptions such as quintessence , phantom and quintom are not so clear as usual .
but , anyway , we can still use these conceptions in an undemanding sense .
it should be noted that when we refer to these conceptions the only thing of interest is the equation of state @xmath28 .
the right panel of fig .
5 tells us : with the interaction ( a case of strong coupling , @xmath53 ) , @xmath106 makes the holographic dark energy behave as phantom - type dark energy with @xmath107 and @xmath108 makes the holographic dark energy behave as quintom - type dark energy with @xmath28 crossing @xmath105 during the evolution history . in this diagram ,
the effect of the interaction is shown again .
when the coupling between the two components is absent , the value of @xmath94 first decreases from zero to a minimum then increases again to zero meanwhile the value of @xmath28 decreases monotonically .
nevertheless , for the case involving the interaction , @xmath28 increases first to a maximum and then decreases meanwhile @xmath94 decreases from a maximum to a negative minimum first and then increases to zero again .
therefore , we see that the @xmath1 dynamical diagnosis can provide us with a useful complement to the statefinder geometrical diagnosis . in summary , we have studied the interacting holographic dark energy model from the statefinder viewpoint in this paper . since the accelerated expansion of the universe was found by astronomical observations ,
many cosmological models involving dark energy component or modifying gravity have been proposed to interpret this cosmic acceleration .
this leads to a problem of how to discriminate between these various contenders .
the statefinder diagnosis provides a useful tool to break the possible degeneracy of different cosmological models by constructing the parameters @xmath92 using the higher derivative of the scale factor .
thus the method of plotting the evolutionary trajectories of dark energy models in the statefinder plane can be used to as a diagnostic tool to discriminate between different models .
furthermore , the values of @xmath92 of today , if can be extracted from precise observational data in a model - independent way , can be viewed as a discriminator for testing various cosmological models . on the other hand , though we are lacking an underlying theory of the dark energy , this theory is presumed to possess some features of a quantum gravity theory , which can be explored speculatively by taking the holographic principle of quantum gravity theory into account .
so the holographic dark energy model provides us with an attempt to explore the essence of dark energy within a framework of fundamental theory .
in addition , some physicists believe that the involving of interaction between dark energy and dark matter leads to some alleviation and more understanding to the coincidence problem .
it is thus worthwhile to investigate the interacting model of holographic dark energy .
we analyzed the interacting holographic dark energy model employing the statefinder parameters as a diagnostic tool .
the statefinder diagrams show that the interaction between dark sectors can significantly affect the evolution of the universe and the contributions of the interaction can be diagnosed out explicitly in this method . at
last , as the complement to the statefinder geometrical diagnosis , a dynamical diagnosis was also studied , which diagnoses the dynamical property of the interacting holographic dark energy in the @xmath1 phase plane .
we hope that the future high - precision observations can offer more and more accurate data to determine these parameters precisely and consequently shed light on the essence of dark energy . |
in a pioneering work fermi , pasta and ulam @xcite showed that in a hamiltonian system with _ short - range _ interactions the relaxation time to thermodynamical equilibrium can be extremely long if the energy is small enough .
this striking feature triggered an intense research activity and , nowadays , the relation between macroscopic thermodynamical properties and microscopic dynamics is actively investigated in a large variety of physical systems @xcite .
the common belief that microcanonical time averages coincide with equilibrium ensemble averages after a possibly quite long transient is confirmed by several numerical experiments ( see refs .
@xcite for reviews ) . therefore , although the rigorous proof of _ ergodicity _ is restricted to a few rather abstract dynamical systems ( e.g. the sinai billiard ) , it is nevertheless believed , and confirmed numerically @xcite , that time averages of thermodynamic functions ( temperature , internal energy , specific heat ) converge to their equilibrium values , at least in the regime of strong chaos .
the case of _ long - range _ forces is quite less explored from this point of view .
the study of relaxation to thermal equilibrium is here made more complex by the consequences of the _ inequivalence of statistical ensembles_. indeed , it is well known that the impossibility to isolate a small sub - system from a large thermal bath , due to the presence of long - range forces , prevents the derivation of _ canonical _ from _ microcanonical _ ensemble .
exactly solvable toy models @xcite with weak long - range forces explicitly display ensemble inequivalence and the presence of _ negative specific heat _ in the microcanonical ensemble .
similar phenomena were also obtained for gravitational models @xcite .
these toy models show a _ first order phase - transition _ in the canonical ensemble , which corresponds to the presence of a concave region of the entropy - energy curve in the microcanonical ensemble .
this feature is present also in microcanonical molecular dynamics simulations of _ short - range _ models without hard - core @xcite , where one observes , in the phase - transition region , a high - density _ clustered _ phase coexisting with a low - density _ gaseous _ phase .
more recently , microcanonical monte - carlo simulations have shown similar properties for the diluted three - state potts model @xcite .
all the previous remarks concern _ equilibrium _ properties , but it is known that long - range interacting systems show also extremely slow relaxation to equilibrium .
a remarkable example is the one - dimensional self - gravitating potential , where the study of long transients and quasi - equilibria has been undertaken by many groups @xcite .
more recently , a scaling law for the relaxation to equilibrium has been proposed for gravitational systems based on diffusion in phase - space @xcite .
the hamiltonian mean field ( hmf ) model @xcite has been introduced with the aim of studying _ clustering _ phenomena in @xmath0-body systems in one dimension .
the original motivation was to consider a truncated fourier series development of the one - dimensional gravitational and charged sheet models .
its canonical ensemble exact solution predicts a _ second - order _ phase transition from a clustered phase to a gaseous one .
hmf is an _ infinite range _ interacting system and
one would naturally expect unusual thermodynamical behaviour .
however , microcanonical simulations of hmf show that carefully " prepared initial states lead to a good agreement with canonical analytical results for the energy - temperature relation @xcite . on the other hand ,
water bag " initial states , _ ie . _ states with momentum distribution on a finite support , are _ metastable _ and show an extremely slow relaxation to equilibrium near the critical energy @xcite . in this paper
, we derive analytically the entropy - energy relation of the hmf model in the microcanonical ensemble .
we further study the properties of metastable states corresponding to the low energy extension of the high energy gaseous phase .
the relaxation time of these states to canonical equilibrium is also investigated , showing that it increases linearly with the number of particles @xmath0 near the critical energy .
the hamiltonian mean field model describes the dynamics of @xmath0 classical and identical particles confined to move on the unit circle @xcite .
its hamiltonian writes @xmath1 where @xmath2
are canonically conjugated coordinate and momentum variables with @xmath3 .
@xmath4 is the kinetic energy and the potential energy @xmath5 corresponds to the first harmonic of the one - dimensional self - gravitating potential @xmath6 ( up to scaling constants ) .
total momentum @xmath7 is also conserved and always set to zero without loss of generality . in
the _ canonical _ ensemble , model ( [ ham ] ) undergoes a second - order phase transition at the inverse temperature value @xmath8 @xcite , corresponding to a critical specific energy @xmath9 .
the free energy @xmath10~ , \label{free}\ ] ] where @xmath11 has been obtained using saddle - point techniques @xcite , giving @xmath12-\frac{\beta}{2 } + \max_x \biggl[x(x,\beta)\biggr]~ , \label{solution}\ ] ] where @xmath13 @xmath14 being the zero order modified bessel function .
the maximization of @xmath15 is performed by solving the consistency equation @xmath16 whose unique solution at small @xmath17 is @xmath18 , while at large @xmath17 , _ ie .
_ low temperature , a pair of symmetric non vanishing solutions is present .
the bifurcation ( of pitchfork type ) occurs at @xmath19 and at this temperature the value of the free energy f has a discontinuity in the second order derivative , which is the signature of a second - order phase transition .
this is also confirmed by the calculation of the order parameter @xmath20 which vanishes continuously at @xmath21 , remaining zero in the whole high temperature phase . at low temperatures
, @xmath22 measures the degree of clustering of the particles .
the equations of motion derived from model ( [ ham ] ) , written in terms of @xmath23 , show that each particle has a pendulum like motion @xmath24 which makes explicit the self - consistent nature of the dynamics , since each particle moves in the m - field , which is itself determined by the position @xmath25 of all the particles .
this has led @xcite to the interpretation of the phase transition as a dynamical process of particle evaporation from the cluster . moreover
, this approach follows the idea that the study of self - consistent @xmath0-body dynamics is simplified by an effective " reduction of the number of degrees of freedom @xcite .
the free energy @xmath26 in the canonical ensemble is readily computed once one knows the dependence of entropy @xmath27 on specific energy @xmath28 in the microcanonical ensemble .
one begins from the relation @xmath29 where @xmath30 is the microcanonical phase - space density at energy @xmath31 .
the lower limit of the integral in ( [ part ] ) is given by the energy of the ground state , which vanishes in our model ( in this case all of the particles are at the same position with zero momentum ) .
after rescaling by @xmath0 and exponentiating @xmath32 eq . ( [ part ] ) becomes @xmath33 this integral is then solved by the saddle - point technique in the @xmath34 limit , which implies taking the maximum of the argument of the exponential .
using the definition of entropy @xmath35 \label{ent}\ ] ] and recalling the definition of free energy ( [ free ] ) , one finally gets @xmath36~. \label{legendre}\ ] ] entropy @xmath27 and free energy @xmath37 are thus related by a legendre transform . _ assuming _ now that @xmath27 is concave ( _ i.e. _ downward bended ) , one can invert ( [ legendre ] ) and obtain @xmath38 .
\label{inv_legendre}\ ] ] this assumption is not innocent and the invertion ( [ inv_legendre ] ) can be safely used only for second order phase - transitions , as it is for our model . on the contrary
, it is well known that for first - order phase transitions entropy has a concave region @xcite , which is responsible for the negative specific heat in the microcanonical ensemble .
a generalization of model ( [ ham ] ) to particle motion on the two - dimensional torus @xcite does have a first - order phase transition , making the invertion impossible in this case .
the calculation of entropy in the microcanonical ensemble is then reduced to the min - max procedure @xmath39~ , \label{solution2}\ ] ] with @xmath15 given by ( [ var ] ) .
for @xmath40 the _ max _ is at @xmath18 and the _ min _ at @xmath41 , which gives the entropy function @xmath42 this is plotted in fig .
[ fig1 ] with full line above @xmath43 and with dashed line for @xmath44 ( below @xmath45 the temperature @xmath46 becomes negative and the solution meaningless ) .
this solution corresponds to the high temperature homogeneous phase ( @xmath47 ) with @xmath48 for @xmath49 and to the _ metastable _ ( the meaning is clarified below ) homogeneous phase @xmath50 below @xmath51 .
both phases are present also in the canonical solution and while in this latter context the @xmath50 corresponds to a local , but not global , minimum of the free energy , in the microcanonical context the entropy of the @xmath50 is not maximal .
indeed , the solution which maximizes the entropy bifurcates continuously from this one at @xmath51 . below this energy the equilibrium solution is related to the two symmetric solutions with @xmath52 of the consistency equation ( [ consistency ] ) .
they give the new maximum value of @xmath15 in ( [ solution2 ] ) and the numerical search of the minimum in @xmath17 , which is unique , allows to compute the stable branch of @xmath53 below @xmath51 .
the latter is plotted in fig .
[ fig1 ] with the continuous line and labelled @xmath54 , because it corresponds to the clustered phase , which is characterized by a non vanishing value of @xmath22 .
the temperature @xmath55 can also be derived from the standard relation @xmath56 and is reported in fig .
the stable solution @xmath53 ( full line ) is continuous with its derivative .
the discontinuity is in the second derivative , which gives a jump in the specific heat @xmath57 . _ to summarize , our results suggest that the microcanonical solution is fully consistent with the one in the canonical ensemble , and we can therefore conclude that the two ensembles lead to equivalent results for model ( [ ham ] ) . _
however , the dynamics reveals a quite peculiar behavior of the microcanonical _ metastable _ branch . we have performed numerical simulations in microcanonical fixed - energy conditions by integrating the equations of motion ( [ ham2 ] ) , starting from appropriately chosen initial states . we have used a fourth order symplectic algorithm @xcite with a time step @xmath58 , allowing relative energy conservation as good as @xmath59 .
we have checked the energy - temperature curve , definining temperature as twice the time average of the kinetic energy : @xmath60 . the equilibrium energy - temperature solution is reached from generic initial states over most part of the energy range . however , for @xmath44 relaxation to equilibrium
becomes very slow .
we have started the runs in this range with the particles homogeneously distributed on the circle ( @xmath48 ) and with a `` water bag '' momentum distribution : the momenta are uniformly distributed in an interval around zero such that the variance of the distribution coincides initially with the temperature of the _ metastable _ branch @xmath61 . the relaxation of the initial `` water bag '' to the equilibrium state proceeds through the slow development of maxwellian tails in the velocity distribution . in the literature one can already find indications that the relaxation time to equilibrium grows linearly with @xmath0 @xcite ( the authors used as a criterion the convergence to the boltzmann entropy of the maxwellian ) .
we have decided to consider directly the time evolution of the temperature of the @xmath50 states .
the initial temperature lies below the equilibrium one ( see dashed line in fig .
[ fig1 ] ) and relaxes to a fixed threshold temperature @xmath62 after a time @xmath63 , where @xmath64 is the threshold amplitude and @xmath65 is the microcanonical equilibrium temperature represented by the full line in fig .
considering different initial states , we have observed that the fluctuations of @xmath63 are poissonian and that a well defined average exists .
this average relaxation time @xmath66 is plotted in fig .
[ fig2 ] for @xmath67 as a function of @xmath0 for different values of @xmath68 . for @xmath69
, @xmath66 is almost independent of @xmath0 whereas for larger values of @xmath68 , @xmath66 increases approximately as @xmath0 .
nothing substantially changes using different thresholds by changing the value of @xmath70 .
we are therefore able to confirm , for a different observable and a better statistics ( we have considered up to 3600 initial conditions for the points in fig . [ fig2 ] ) , the results of ref .
@xcite on the linear divergence of the relaxation time with @xmath0 , at least in a restricted energy interval just below @xmath51 .
the existence of such divergence implies that performing the @xmath34 limit before the @xmath71 limit , the system will stay indefinitely out of equilibrium and the momentum distribution will never develop maxwellian tails .
this has attracted the attention of tsallis @xcite , who argues that his entropy definition should be appropriate to describe the thermodynamical properties of such _
metastable _ states in the thermodynamic limit ( in particular it would possibly describe the power law tails of the momentum distribution ) .
it should however be remarked that what we have tested here is not the divergence of the _ lifetime _ of such states , but that of the relaxation time . however , recent numerical simulations performed by rapisarda @xcite seem to indicate also the divergence of the lifetime with @xmath0 , encouraging tsallis interpretation .
furthermore , it is possible to prepare states on the metastable branch which have initially a maxwellian momentum distribution and those states are observed to relax to equilibrium pretty fast .
it seems that the water bag " condition , which corresponds to an initial truncation of the tails , is a necessary one for obtaining a divergence of the relaxation time .
let us summarize what we have found .
the _ microcanonical _ solution of the hamiltonian mean field model is here obtained with legendre transform techniques , which avoid state counting .
the entropy is computed using a min - max procedure and the energy - temperature relation is derived analytically .
the equilibrium states are shown to coincide with the _ canonical _ ones in the thermodynamic limit .
this result is not obvious since , due to the long - range interparticle coupling , ensemble equivalence might have been violated . besides the equilibrium solutions ,
other states are found , corresponding to the _ metastable _ homogeneous configurations lying below the critical energy .
the study of their relaxation time to equilibrium exihibits a divergence proportional to the number of particles , meaning that in the thermodynamic limit the trajectory in the phase - space remains indefinitely trapped in a non - equilibrium region ( something similar to so - called mechanical " initial states for the fermi - pasta - ulam model in ref .
further investigations are necessary to understand the relevance of the metastable states for other more realistic long - range forces , beginning with the very interesting extension of the hmf proposed by tsallis and co - workers @xcite , whose canonical solution has been recently obtained in ref .
these results could be relevant for the physics of self - gravitating systems and for nuclear dynamics , where long - range forces play a distinctive role .
we thank v. latora , a. rapisarda and a. torcini for fruitful discussions and ongoing collaboration on this subject .
m. a. acknowledges d. noack for important logistics support .
m.a . and s.r .
thank the institute for scientific interchange , torino ( italy ) and the laboratoire de physique of the ens - lyon ( france ) for hospitality and financial support .
the max - planck institute for physics of complex systems in dresden ( germany ) is also acknowledged for allowing us to use its powerful computer center .
e. fermi , j. pasta and s. ulam , los alamos report la-1940 ( 1955 ) , in d.c .
mattis , _ the many - body problem _ , world scientific , ( 1993 ) .
chaos , _ focus issue on chaos and irreversibility _ , t. tl , p. gaspard and g. nicolis , eds .
s. ruffo , in _ symmetry and perturbation theory _ , d. bambusi and g. gaeta eds .
, gnfm - cnr ( 1998 ) ; p. poggi and s. ruffo , physica d , * 103 * , 251 ( 1997 ) .
r. livi , m. pettini , s. ruffo , a. vulpiani , j. stat .
phys . , * 48 * , 539 ( 1987 ) ; d. escande , h. kantz , r. livi and s. ruffo , j. stat .
* 76 * , 605 ( 1994 ) . c. giardin and r. livi , j. stat .
phys . , * 91 * , 1027 ( 1998 ) .
p. hertel and w. thirring , ann . of physics ,
* 63 * , 520 ( 1970 ) .
d. lynden - bell and r. wood , mon . not .
soc , * 138 * , 495 ( 1968 ) .
a. compagner , c. bruin and a. roelse , phys .
a , * 39 * , 5989 ( 1989 ) ; h.a .
posch , h. narnhofer and w. thirring , phys .
rev . a , * 42 * , 1880 ( 1990 ) .
gross , a. ecker and x.z .
zhang , ann .
physik , * 5 * , 446 ( 1996 ) and contribution of d.h.e .
gross to this volume .
wright , b.n .
miller and w.e .
stein , astrophys .
space sci .
* 84 * , 421 ( 1982 ) ; b.n .
miller and c.j .
reidl jr . , astrophys .
j. * 348 * , 203 ( 1990 ) ; t. tsuchiya , n. gouda and t. konishi , phys .
e , * 53 * , 2210 ( 1996 ) ; t. tsuchiya , n. gouda and t. konishi , astrophys .
space sci . , * 257 * , 319 ( 1997 ) . m .-
chabanol , f. corson and y. pomeau , europhys . lett .
* 50 * , 148 ( 2000 ) .
s. ruffo , in _ transport and plasma physics _ , s. benkadda , y. elskens and f. doveil eds .
, world scientific , singapore , 114 ( 1994 ) ; m. antoni and s. ruffo , phys .
e * 52 * , 2361 , ( 1995 ) .
v. latora , a. rapisarda and s. ruffo , phys .
* 80 * , 692 ( 1998 ) ; physica d , * 131 * , 38 ( 1999 ) .
tennyson , j.d .
meiss and p.j .
morrison , physica d * 71 * , 1 ( 1994 ) ; d. del - castillo - negrete , phys . lett .
a * 241 * , 99 ( 1998 ) .
m. antoni and a. torcini , phys .
e * 57 * , r6233 ( 1998 ) , a. torcini and m. antoni , phys .
e * 59 * , ( 1999 ) .
i. mclachlan and p. atela , nonlinearity * 5 * , 541 ( 1992 ) . c. tsallis , braz .
* 29 * , 1 ( 1999 ) , cond - mat/9903356 and refs . therein .
a. rapisarda , contribution to this conference .
j. de luca , a.j . lichtenberg and s. ruffo , phys .
e , * 60 * , 3781 ( 1999 ) c. anteneodo and c. tsallis , phys .
* 80 * , 5313 ( 1998 ) .
f. tamarit and c. anteneodo , phys .
lett . * 84 * , 208 ( 2000 ) .
a. campa , a. giansanti and d. moroni , preprint cond - mat/0002168 and contribution of a. giansanti to this conference . | we study the hamiltonian mean field ( hmf ) model , a system of @xmath0 fully coupled particles , in the microcanonical ensemble .
we use the previously obtained free energy in the canonical ensemble to derive entropy as a function of energy , using legendre transform techniques .
the temperature - energy relation is found to coincide with the one obtained in the canonical ensemble and includes a _
metastable _ branch which represents spatially homogeneous states below the critical energy .
water bag " states , with removed tails momentum distribution , lying on this branch , are shown to relax to equilibrium on a time which diverges linearly with @xmath0 in an energy region just below the phase transition . |
according to latest observations @xcite , around 69% of the universe is constituted by the dark energy ( de ) sector , responsible for its current accelerating phase , and almost 26% by the cold dark matter ( cdm ) sector in the form of dust .
these two sectors are considered not to interact with each other , and the resulting scenario has been believed to be the best description of the present observable universe , due to its agreement with a large number of independent observations . however , some potential problems such as the coincidence one @xcite , namely why are the dark energy and dark matter energy densities of the same order although they follow completely different evolutions , have led to a large amount of research towards scenarios in which the above two dark sectors exhibit a direct interaction .
such a mutual direct interaction between dark energy and dark matter can not be excluded from the field theoretical point of view , independently of the specific nature of the former , i.e. whether it arises through a universe - content modification @xcite or through a gravitational modification @xcite . hence , many different scenarios of interacting dark energy have been constructed ( see @xcite for reviews and references therein ) .
interestingly enough , recently it has been shown that the current data might favor the late - time interaction between cdm and the vacuum energy @xcite or the dynamical dark energy sector @xcite . in the present work we are interested in providing updated constraints on the scenarios of interacting dark energy , using the very recent observational data , along with the new local value of the hubble parameter with a 2.4@xmath0 determination , released in @xcite .
in particular , considering a direct interaction between dark energy and dark matter in a friedmann - robertson - walker ( frw ) background , and assuming that both ( effective ) fluids obey a barotropic equation of state , the cosmological dynamics becomes richer and thus the evolution equations for cdm and dark energy become different from their standard forms .
hence , one can use observational data in order to fit the deviation from non - interaction , without the need to specify a specific interaction form , i.e. keeping the analysis in a general ground .
the manuscript is organized as follows : in section [ sec : ide ] we provide a brief description of the expansion history of the universe in the background of an interacting scenario . in section [ data ]
we describe the latest data sets for our analysis , while in section [ results ] we use them and we extract the constraints on the various quantities .
finally , section [ discu ] is devoted to summary and conclusions .
let us briefly describe the scenario of interacting dark energy . in a background of a frw universe with
metric @xmath3 $ ] , where @xmath4 is the scale factor and @xmath5 the spatial curvature ( with @xmath6 for flat , open and closed universe respectively ) , the first friedmann equation can be written as @xmath7 where @xmath8 is the hubble rate , with dots denoting derivatives with respect to the cosmic time , and with @xmath9 , @xmath10 , @xmath11 , and @xmath12 denoting the energy densities of radiation , baryons , cold dark matter and dark energy , respectively .
since the physics of the radiation and baryonic sector is known quite well , we assume that these sectors are independently conserved . on the other hand ,
if the dark energy and dark matter sectors interact directly , the energy conservation law for the interacting cdm - de components reads as @xmath13 , where the total energy - momentum tensor is @xmath14 , or equivalently @xcite @xmath15 where @xmath16 is a specific interaction form that has to be chosen by hand . as usual
, one can quantify the effect of the above interaction on the dark matter evolution through a deviation from the standard dust - matter evolution , namely considering that the dark matter energy density evolves as @xcite @xmath17 with @xmath18 the present value , and where the parameter @xmath19 becomes zero in the non - interacting case .
hence , the parameter @xmath19 quantifies the deviation from non - interacting case ( @xmath20 corresponds to energy flow from dark matter to dark energy ) , without the need to consider a specific interaction form @xmath16 . combining ( [ dotrho ] ) , ( [ rho0a3 ] ) and considering that the dark energy sector is described by an equation - of - state parameter of the form @xmath21
, we find that the energy density of the dark energy component evolves as @xmath22,\end{aligned}\ ] ] where @xmath23 is the present value of @xmath12 .
clearly , in the absence of a coupling with the cdm component , the conventional dynamical dark energy scenario is recovered .
note that for @xmath24 and @xmath25 , we may identify @xmath26 and the above expressions reduce to the vacuum decaying scenario @xcite .
in the present work we confront the constraints on the cosmological parameters that can be obtained by probes that map the late - time universe ( @xmath27 ) expansion history .
the baseline of our analysis is the hubble parameter measurements obtained with the cosmic chronometers ( cc ) technique , however we additionally consider standard probes such as supernovae type ia ( sne ia ) , local hubble parameter value @xmath28 ones , and baryon acoustic oscillation distance measurements ( bao ) , to diminish the degeneracy between the free parameters of the models . in the following subsections , we present the data sets considered in our analysis .
the cc approach to measure @xmath29 was first introduced in @xcite , and it uses relative ages of the most massive and passively evolving galaxies to measure @xmath30 , from which @xmath29 is inferred . the latest implementation has been explained in detail in @xcite , where the possible sources of uncertainty and related issues are also discussed .
we consider the compilation of hubble parameter measurements provided by @xcite .
it contains the latest updated list of @xmath29 measurements @xcite obtained with the cosmic chronometers approach , comprising of 30 measurements spanning the redshift range @xmath31 .
this sample covers roughly 10 gyr of cosmic time .
furthermore , in our analysis we also include the new local value of @xmath28 as measured by @xcite with a 2.4 @xmath0 determination , which yields @xmath32 km / s / mpc .
sne ia were the main tool to discover late - time acceleration @xcite , and they still provide the best constraints on dark energy sector .
sne ia are very bright `` standard candles '' and thus are used to measure cosmic distances .
we consider here the latest
joint light curves " ( jla ) sample @xcite , comprised of 740 sne ia in the redshift range @xmath33 $ ] . from the observational point of view
, the distance modulus of a sne ia can be abstracted from its light curve assuming that supernovae with identical color , shape and galactic environment have on average the same intrinsic luminosity for all redshifts .
this hypothesis is quantified by a empirical linear relation , yielding a standardized distance modulus @xmath34 of the form @xmath35 where @xmath36 corresponds to the observed peak magnitude in rest frame b band and @xmath37 , @xmath38 , and @xmath39 are nuisance parameters in the distance estimate .
the absolute magnitude is related to the host stellar mass ( @xmath40 ) by a simple step function : @xmath41 if @xmath42 , otherwise @xmath43 .
the light - curve parameters ( @xmath36 , @xmath44 , and @xmath45 ) result from the fit of a model of the sne ia spectral sequence to the photometric data . in our analysis
we assume @xmath39 , @xmath46 , @xmath37 and @xmath38 as nuisance parameters .
another important cosmological probe are the baryon acoustic oscillations ( bao ) .
the bao can be traced to pressure waves at the recombination epoch , generated by cosmological perturbations in the primeval baryon - photon plasma , appearing as distinct peaks in large angular scales .
we use the following bao data to constrain the expansion history of the scenario at hand : the measurement from the six degree field galaxy survey ( 6df ) @xcite , the main galaxy sample of data release 7 of sloan digital sky survey ( sdss - mgs ) @xcite , the lowz and cmass galaxy samples of the baryon oscillation spectroscopic survey ( boss - lowz and boss - cmass , respectively ) @xcite , and the distribution of the lymanforest in boss ( boss - ly ) @xcite .
these measurements and their corresponding effective redshift @xmath47 are summarized in table [ tab1 ] .
.baryon acoustic oscillation ( bao ) data measurements included in our analysis . [ cols="^,^,^,^,^",options="header " , ] as we can see , when analyzing the scenario with @xmath48 + @xmath28 data , we observe that @xmath25 combined with @xmath49 is preferred ( see table [ tab2 ] ) .
on the other hand , when we break the degeneracy of the parameters , combining the analysis with sneia / jla + bao + cmb observations , namely using @xmath48 + @xmath28 + sne ia @xmath50 bao + cmb data , we observe that @xmath51 ( with a minor tendency for @xmath52 , i.e. for an energy flow from dark energy to dark matter ) , however with @xmath53 up to 3@xmath2 confidence level .
more specifically we note that @xmath54 at 3@xmath2 .
the 2@xmath2 confidence - level bounds can be seen in table [ tab3 ] .
hence , we deduce that the global fits from the map of the universe expansion history have a slight preference towards a cosmological scenario of interacting dark energy .
this is the main result of the present investigation , and it is in agreement with the results of other observational works @xcite , however it has been arisen through the novel use of the recently released cosmic chronometers data .
in the present work we have extracted observational constraints on the scenario of interacting dark energy , without the need to consider a specific interaction form , using the recently released cosmic chronometers data . in particular , considering a general direct interaction between dark matter and dark energy sectors has the general effect of altering the evolution of the former to @xmath55 , where @xmath19 is the parameter that quantifies the deviation from the non - interacting case .
hence , one can use observational data in order to fit @xmath19 as well as the other cosmological parameters .
in our analysis we used ( i ) the very recently released cosmic chronometer data sets along with the very latest measured value of the local hubble parameter , @xmath32 km / s / mpc @xcite , ( ii ) `` joint light curves '' ( jla ) sample containing 740 latest supernovae type ia data points , ( iii ) baryon acoustic oscillation data points , ( iv ) cmb distance priors adopting 2015 planck tt , te , ee + lowp data , marginalized over the amplitude of the lensing power spectrum @xcite .
we presented two different sets of values of the model parameters , one arising using the combined set of @xmath48 + @xmath28 data ( summarized in fig .
[ qm1 ] and table [ tab2 ] ) , and one arising using the @xmath48 + @xmath28 + sneia / jla + bao + cmb data ( summarized in fig .
[ qm2 ] and table [ tab3 ] ) .
we found that the combined analysis @xmath48 + @xmath28 slightly favors a non - zero value for the interacting parameter @xmath19 , while the dark energy equation - of - state parameter lies below @xmath56 .
additionally , for the combined analysis , using @xmath48 + @xmath28+sneia / jla @xmath50 bao @xmath50 cmb data , we found that @xmath19 is close to zero , nevertheless the possibility of a small interaction is not ruled out , and moreover we find that the dark energy equation - of - state parameter lies below @xmath56 at 3@xmath2 confidence level .
in summary , using the latest observational data , we have found that an interaction between dark matter and dark energy sectors is mildly favored .
s.p . acknowledges science and engineering research board ( serb ) , govt . of india , for awarding the national post - doctoral fellowship ( file
no : pdf/2015/000640 ) .
this article is based upon work from cost action `` cosmology and astrophysics network for theoretical advances and training actions '' , supported by cost ( european cooperation in science and technology ) .
the authors thank n. tamanini for useful comments .
finally , the authors gratefully acknowledge an anonymous referee for his / her valuable comments .
y. f. cai , e. n. saridakis , m. r. setare and j. q. xia , phys .
rept . * 493 * , 1 ( 2010 ) .
s. nojiri and s. d. odintsov , phys .
rept . * 505 * , 59 ( 2011 ) .
s. capozziello and m. de laurentis , phys .
rept . * 509 * , 167 ( 2011 ) .
a. g. riess et al . , ( supernoave search team collaboration ) , astron .
j. * 116 * , 1009 ( 1998 ) .
s. perlmutter et al . ,
( supernoave cosmology project ) astrophys . j. * 517 * , 565 ( 1999 ) .
m. betoule et al . , ( sdss collaboration ) , astron . astrophys . * 568 * , a22 ( 2014 ) .
f. beutler , c. blake , m. colless , d. h. jones , l. staveley - smith , l. campbell , q. parker , w. saunders and f. watson , mnras * 416 * , 3017 ( 2011 ) .
a. j. ross , l. samushia , c. howlett , w. j. percival , a. burden and m. manera , mnras * 449 * , 835 ( 2015 ) .
j. r. bond , g. efstathiou and m. tegmark , mon . not .
. soc . * 291 * , l33 ( 1997 ) .
g. efstathiou and j. r. bond , mon . not .
. soc . * 304 * , 75 ( 1999 ) .
y. wang and p. mukherjee , phys .
d * 76 * , 103533 ( 2007 ) .
d. blas , j. lesgourgues , and t. tram , jcap * 1107 * , 034 ( 2011 ) .
b. audren , j. lesgourgues , k. benabed and s. prunet , jcap * 1302 * , 001 ( 2013 ) .
y. wang , d. wands , l. xu , j. d. santiago and a. hojjati , phys .
d * 87 * , 083503 ( 2013 ) . | we use the latest compilation of observational hubble parameter measurements estimated with the differential evolution of cosmic chronometers , in combination with the local value of the hubble constant recently measured with 2.4@xmath0 precision , to constrain the cosmological scenario where dark energy interacts directly with the dark matter sector . to diminish the degeneracy between the parameters we additionally consider standard probes , such as supernovae type ia from joint light curves ( jla ) sample , baryon acoustic oscillation distance measurements ( bao ) , and cosmic microwave background data from planck 2015 estimations .
our analysis shows that the direct interaction between dark energy and dark matter is mildly flavored , while the dark energy equation - of - state parameter is @xmath1 at 3@xmath2 confidence level . |
during his remarkable bakerian lecture at june 19 , 1941 , the published version of which dates back to 1942 @xcite , p.a.m .
dirac summarized in his decisive manner a by now very influential conception of the physical world to which he had contributed more than significantly ( see also e.g. @xcite ! ) .. as bosonic particles do not follow pauli s exclusion principle , dirac s concept of filling up negative energy states could not be applied to bosonic systems .
hence , dirac positioned the bosonic vacuum at @xmath0 , in order to avoid bose - einstein condensation of bosonic particles towards @xmath1 . admitting
hereby bosonic particles of only positive energy dirac found nevertheless a very elegant way to make use of the negative energy eigenstates of a bosonic system .
dirac noticed that bosonic states of negative energy have simultaneously also _ negative norm _ and _ negative probability _
( see also @xcite ! ) . therefore dirac @xcite declared bosonic negative energy states to be _
responsible for the annihilation ( i.e. absorption ) of positive energy particles _ ( the negative energy states discussed in @xcite reflect and support nicely this interpretation of dirac ( even for fermions ) , even though completely differently interpreted by the respective author ! ) . strictly speaking , dirac associated even bosonic positive energy states with the _ emission _ of bosonic particles and _ not _ with their _ existence _ ! ] from a modern point of view the conception of dirac has besides a rather asymmetric treatment of fermions and bosons its particular drawbacks in two points causing serious difficulties in their theoretical handling : the mere existence of a `` dirac sea '' with its infinite negative energy ( causing in particular problems to cosmologists ) and the missing description of bosonic antiparticles . in vol . 1 of his book `` the quantum theory of fields '' @xcite s. weinberg seems to suggest that both problems have been surpassed by formalisms developed 1934 by w. pauli and v. weisskopf ( for spin 0 bosons ) @xcite and w.h .
furry and j.r .
oppenheimer ( for spin 1/2 fermions ) @xcite .
furry and oppenheimer picked up dirac s idea that the positron is the absence of a negative - energy electron ; the anticommutation relations are symmetric between creation and annihilation operators , so they defined the positron creation and annihilation operators as the corresponding annihilation and creation operators for negative - energy electrons @xmath2 , @xmath3 ( for @xmath4 ) where the label @xmath5 on @xmath6 denotes a positive - energy positron mode with momenta and spin opposite to those of the electron mode @xmath5 . @xmath7 it is necessary also to specify that the physical vacuum is a state @xmath8 containing no positive - energy electrons or positrons : @xmath9 ( @xmath10 ) , @xmath11 ( @xmath12 ) @xmath7''_. then , in the context of the work of w. pauli and v. weisskopf , he remarks on p. 26
ff in @xcite : _
`` @xmath7 the existence of two different kinds of operators @xmath13 and @xmath6 , which appear in precisely the same way in the hamiltonian , shows that this is a theory with two kinds of particles with the same mass . as emphasized by pauli and weisskopf , these two varieties can be identified as particles and the corresponding antiparticles , and if charged have opposite charges .
thus as we stressed above , bosons of spin zero as well as fermions of spin 1/2 can have distinct antiparticles , which for bosons can not be identified as holes in a sea of negative energy particles @xmath7''_. in the course of time even dirac changed according to s. weinberg his point of view as is stated on p. 13
ff in @xcite : _ `` @xmath7 and
if the hole theory does not work for bosonic antiparticles , why should we believe it for fermions ?
i asked dirac in 1972 how he then felt about this point ; he told me that he did not regard bosons like the pion or @xmath14 as ` important ' . in a lecture
a few years later , dirac referred to the fact that for bosons ` we no longer have the picture of a vacuum with negative energy states filled up ' , and remarked that in this case ` the whole theory becomes more complicated ' .
@xmath7''_. ] _ if _ one characterizes stable particles serving as asymptotic states by _ real _ energies , then one might be tempted to agree _ for such real energy particles _ to the view of pauli , weisskopf , furry and oppenheimer ( pwfo ) ( adverted by weinberg and finally also by dirac ) , as the introduction of two operators @xmath15 and @xmath16 for particles and antiparticles , respectively , being related by @xmath2 and @xmath3 for @xmath4 indeed admits the vacuum to be positioned at @xmath0 , while the operator @xmath16 reflects even without a dirac sea all properties of holes in dirac s sea of negative _ real _ energy states .
furthermore the concept has the advantage to be applicable not only to fermions , but also to bosons . yet
as will be explained in more detail below the view of pwfo _ can not _ be kept as it stands without inconsistency , when one tries to formulate a local , causal , analytic and poincar - covariant quantum theory ( qt ) for systems with complex - valued selfenergies and vertex - functions .
the result of such an exercise @xcite may be called nhqt or ( a)cqt and will be described to some extend in the following text .
it will yield a `` revised '' conception of the physical world treating fermions and bosons completely symmetrically and being free of the above conjectured inconsistency arising not only in the early conception by dirac , but also in the subsequent , more sophisticated conception by pwfo adverted e.g. by weinberg ( see also @xcite ! ) .
this inconsistency , the resolution of which is a striking argument in favour of ( a)cqt , may be sketched as simple as follows : assume that @xmath17 is the annihilation operator of a _ causal _ particle with _ complex _ energy @xmath18 ( re@xmath19 > 0 $ ] , im@xmath19 < 0 $ ] ) .
then the respective creation operator of the particle with _ complex _ energy @xmath18 is @xmath20 ( @xmath21 according to pwfo ) .
the hermitian conjugation of @xmath20 leads then to @xmath22 ( @xmath23 according to pwfo ) being the annihilation operator of an _ anticausal _ particle with the _ complex conjugated _ energy @xmath24 , and not to the annihilation operator of a _ causal _ antiparticle with energy @xmath18 , as suggested by pwfo . for completeness
we note that @xmath25 is the creation operator of the _ anticausal _ particle with energy @xmath24 .
and @xmath26 mesons when performing the identification @xmath27 .
] this unexpected lesson from a non - hermitian causal approach to qt should not surprise , yet be seriously taken into account , as illustrated also by the following example . as early as 1954 t.d .
lee observed in the so - called _ normal _ `` lee - model ''
@xcite ( see also e.g. @xcite ! ) being a somehow extended version of the preceding `` friedrichs - model '' @xcite of 1948 a similar surprise due to unexpected non - hermiticities , which he comments as follows : _
`` @xmath7 it can be shown that the result of the renormalization process can not be obtained by any limiting process that involves only real values of the unrenormalized coupling constant .
this difficulty may , however , be overcome by a modification of the present rules of quantum mechanics .
@xmath7 '' _ @xcite . @xcite and _ abnormal _
@xmath28 @xcite lee - model described according to @xcite by a hamilton operator @xmath29 with @xmath30 and @xmath31 $ ] consists of a _ normal / abnormal _ fermion @xmath32 and a _ normal _ fermion @xmath33 with no kinetic energy coupling to the _ positive frequency part _ @xmath34 of a boson @xmath35 ( `` @xmath36-particle '' ) , i.e. @xmath37 ! ] a subsequent discussion by w. pauli @xcite in june , 1955 , based on a joint analysis with g. klln @xcite resulting in the so - called _ abnormal _ lee - model clarified in a very decisive and instructive way , what is going on in the lee - model : the lee - model s coupling constant renormalization @xmath38 with @xmath39 determined by the square of the renormalized coupling @xmath40 and the unrenormalized coupling @xmath41 changes sign when violating the inequality @xmath42 yielding as already observed by lee a _ purely imaginary _ unrenormalized coupling @xmath41 for a finite , _ purely real _ renormalized coupling @xmath40 . while lee @xcite was interpreting the unexpected change of the hermiticity character between the _ non - hermitian unrenormalized _ hamilton operator resulting from a _ purely imaginary _ coupling @xmath41 and the _ hermitian renormalized _ hamilton operator resulting from a _ real _ coupling @xmath40 as a _ non - unitary similarity transform _ , pauli @xcite stressed that the violation of the inequality @xmath42 _ `` @xmath7 leads to a contradiction with the concept of physical probability ( indefinite metric of the hilbert space ) @xmath7 connected with the appearance of new discrete stationary states whose contribution to the conserved sum of ` probabilities ' is negative ( ` ghosts ' ) @xmath7 '' _
according to foregoing literature pauli called states with positive norm ( and probability ) _ normal _ and states with negative norm ( and probability ) _ abnormal _ ( see also @xcite ! ) .
pauli s and klln s observation about the existence of an indefinite metric and abnormal states in the lee - model allowed them not only to understand that the hamilton operator of their _ abnormal _ lee - model is _ pseudo - hermitian _ the concept of unitarity can not be defined without specifying the metric@xmath7 '' _
( for subtile critics of this understanding from the point of view of an afterwards introduced _ new _
terminology see e.g. @xcite ! ) . ] with respect to this indefinite metric , yet also , what is actually going on in the subspace ( @xmath43 ) of the lee - model during the process of violating the inequality @xmath42 by increasing @xmath44 @xcite -particle collides at a critical value of @xmath40 , when @xmath42 is violated and @xmath45 vanishes , with a real energy eigenvalue of an additional _ abnormal _ eigenstate ; this collision point ( or _ critical point _ ) of the abnormal lee - model is an _ exceptional point _ , as the eigenspace of the hamilton operator spanned by the normal and the abnormal eigenstate reduces at the exceptional point its dimension by one and is now characterized by one state only called _ good ghost _
@xcite due to its zero norm resulting from a now degenerate metric ; fortunately there can be constructed an additional state called _ dipole ghost _ @xcite or _ bad ghost _
@xcite , if chosen such ( see p. 111 in @xcite ! ) that its norm vanishes ( in consequence the good and bad ghost form a _
biorthogonal basis _ ! ) ; this dipole ghost is required to preserve the original dimension of the space spanned by the normal and abnormal state before their collision ; a further increase of @xmath44 beyond the critical value leads then to the formation of a pair of again _ biorthogonal states _ ( called e.g. `` complex ghosts '' @xcite , `` complex roots '' @xcite , `` complex poles '' @xcite , or `` exponential ghosts '' @xcite ) of _ zero ( traditional ) norm _ with complex valued , mutually complex conjugate eigenvalues ( note that such a bifurcation from real to complex pairs of energy eigenvalues does not only occur in the lee - model .
it takes place e.g. in quantum electrodynamics ( qed ) at the landau - pole and in quantum chromodynamics ( qcd ) at the confinement - deconfinement phase - transition ! ) .
the collision scenario in the lee - model involving just a dipole ghost is just a special case of the more general situation , when there develop higher order degeneracies with non - trivial jordan - block structure @xcite at the critical point , i.e. so - called _ multipole ghosts _
the theory of parameter dependent flow , degeneracy and bifurcation of eigenvalues at or around critical points has made since the work of a.m. lyapunov in 1892 @xcite ( see also @xcite ! ) in several fields of theoretical physics ( chaos theory , renormalization group theory , @xmath7 ) great progress .
relevant references also in the context of exceptional / diabolic points may be found in the work of m.v .
berry @xcite and other research groups , e.g. @xcite . ] .
the analysis of the surprising properties of the lee - model required and hence induced similarly as the analysis of the antiparticle concept discussed above not only a much deeper understanding of formalistic aspects of qt in particular in the presence of an _ indefinite metric_-formalism '' ( presumably ) by w. pauli @xcite ( @xmath46 `` metric operator '' ) , which was already thoroughly discussed e.g. in @xcite , has recently been intensively `` rediscovered '' by a. mostafazadeh ( see e.g. @xcite and references therein ! ) .
it is interesting to observe that the mathematical literature on operators in spaces with indefinite metric seems to start approximately 1944 , i.e. shortly after p.a.m .
dirac and w. pauli , by contributions of l.s .
pontrjagin ( 1944 ) @xcite , m.g .
kren ( @xmath47 1950 ) @xcite , r. nevanlinna ( 1952,1956 ) @xcite , and i.s .
iohvidov ( 1956,1959 ) @xcite ( see also references therein ! ) .
selective early mathematical works suitable for consideration by physicists might be also @xcite . in the meantime there have appeared several
well written mathematical reviews on the huge amount of related mathematical literature , e.g. @xcite , while relevant contributions of mathematical physicists ( e.g. @xcite ) seem to have developed their own characteristic language .
following @xcite we want here just to recall some important terminology : an _ inner product space _ is a complex vector space @xmath48 together with a complex valued function @xmath49 ( @xmath50 _ inner product _ ) on @xmath51 which satisfies the axioms of _ linearity _
( @xmath52 for all @xmath53 and @xmath54 ) and _ symmetry _
( @xmath55 for all @xmath56 ) ; the _ antispace _ of an inner product space @xmath57 is the inner product space @xmath58 ; a _ hilbert space _ is a strictly positive inner product space @xmath59 over the complex numbers which is complete in its norm metric ; a _ kren space _
@xmath60 is an orthogonal direct sum of an hilbert space @xmath61 and its antispace @xmath62 , i.e. @xmath63 ; a _ pontrjagin space _
@xmath64 is a kren space with a finite negative index , i.e. @xmath65 ; note that @xmath66 . ] and _ complex energy eigenvalues _
( see e.g. @xcite ! ) , it raised also serious new questions .
one immediately arising question motivated by w. heisenberg , it was mainly heisenberg who emphasized that the existence of a unitary s - matrix for physical states will be sufficient to guarantee the usual quantum - mechanical probability interpretation .
@xmath7 '' _ ] i.e. whether the scattering matrix of a theory containing simultaneously ( complex ) ghosts and eigenstates of strictly real energy eigenvalues can be unitary , was very early answered positively by w. pauli @xcite ( see also @xcite ! ) and w. heisenberg @xcite ( see also @xcite ! ) .
this result was later confirmed by t.d .
lee & g.c .
wick @xcite and h.p .
drr & e. seiler @xcite ( see also @xcite ! ) .
as will be sketched below , another serious and unresolved question remained until very recently , how to achieve beyond unitary simultaneously also _ analyticity _ , _ causality _ , _ locality _ and _ poincar _ or _ lorentz invariance _ in theories containing ( complex ) ghosts .
as in the context of the antiparticle concept discussed above this puzzling point can be resolved within the new framework of ( a)cqt .
what concerns the issue _ analyticity _ , the situation had been summarized along to the results obtained in @xcite on p. 135 in @xcite by w. heisenberg : _
`` @xmath7 taking the dipole case as an example it is easily seen that the amplitudes ( or related functions ) can not be analytic at a possible threshold to the creation of ghost - states , because the boundary conditions must be different below and above threshold.@xmath7''_. according to @xcite ( see also @xcite ! )
he calls such an unsatisfactory situation _ piecewise analyticity_. the complications with _ causality _ , _ locality _ and _ lorentz invariance _ were approached 1968 in an instructive manner by e.c.g .
sudarshan et al .
@xcite in considering the field theory of a _ tachyonic _ klein - gordon ( kg ) field with _ purely imaginary mass_. , @xmath67 ;
2 ) @xmath68 , @xmath67 ; 3 ) @xmath69 ; 4 ) @xmath70 , @xmath71 ; 5 ) @xmath72 , @xmath71 ; 6 ) @xmath73 . case 3 ) discussed on p. 76 in @xcite is obviously tachyonic ! ] they observed that in such a seemingly poincar invariant theory with _ real _ energies and 3-momenta on one hand positive ( negative ) energy states can be transferred to negative ( positive ) energy states , respectively , by lorentz transformations , on the other hand that a consistent construction of interaction terms involves causal _ and _ anticausal propagators simultaneously yielding _ non - local interactions _ and a _ causality violating _ theory .
furthermore they note @xcite : _
`` @xmath7 any attempt at avoiding fock states for negative - energy particles violates the relativistic invariance of the theory .
@xmath7''_. it is surprising to observe that already in 1959 m. froissart @xcite considered a to our present knowledge much more well behaved and lorentz covariant model of two hermitian conjugate complex ghosts with complex mass @xmath74 described by the lagrange density @xmath75 with allows quite nicely to illustrate the concept of canonical quantization in the presence of a dipole ghost within a relativistic context and related complications in the formulation of a s - matrix , it simultaneously yields a theory lacking physical significance for the description of nature . ]
@xmath76 . in choosing
without loss of generality re@xmath77>0 $ ] and im@xmath77<0 $ ] we will call in the follwoing the fields @xmath78 _ causal _ and the fields @xmath79 _ anticausal_. 13 years later , in 1972 , n. nakanishi @xcite then studied froissart s lagrange density with the more significant phase choice @xmath80 under the name _ complex - ghost relativistic field theory _ not refering to froissart in his first submitted work @xcite , yet to work by t.d . lee & g.c . wick @xcite of 1970 who considered a similar problem ironically already for the more complicated case of vector and dirac fields with complex mass ( in the presence of a _ purely imaginary _ gauge - coupling constant ! ) . as a result of an intensive analysis of the model he emphazised on p. 68 in @xcite that _ `` @xmath7 the relativistic complex - ghost field theory is a non - trivial , divergence - free quantum field theory which is manifestly covariant at any finite time and whose physical s - matrix is unitary and macrocausal .
so far we know no other theory which has all these features . the lorentz non - invariance of the physical s - matrix is not necessarily its demerit but it can be a merit .
@xmath7''_. obviously n. nakanishi concluded his analysis of the theory apart from an outline of its many beautiful features reveiled with a general , discouraging claim about its lorentz non - invariance . and @xmath79 with complex mass @xmath74 and @xmath81 , respectively , which were in great detail discussed by a.m. gleeson et al .
@xcite at the end of 1970 and led there to the problem of _ piecewise analyticity _ mentioned above , are not lorentz - invariant . ] hence , even being not tachyonic nakanishi s complex - gost relativistic field theory seemed to suffer from similar problems as the tachynonic kg theory considered by sudarshan et al . in 1968 . yet , contrary to the tachynonic kg theory studied by sudarshan et al .
, there is fortunately as conjectured by the author a very simple way to cure the lack of lorentz - invariance , analyticity , causality and locality in the complex - ghost relativistic field theory of n. nakanishi : it is our postulate that there should be _ no interaction terms between causal fields @xmath78 and anticausal fields @xmath79 in the lagrange density_. the resulting qt called ( a)cqt being described in the following text has the nice feature to be lorentz covariant , analytic , causal and local even in the presence of fields of complex mass with finite imaginary part describing short - lived intermediate particles , which is due to a _ very tricky interplay _ between the underlying normal and abnormal states having caused so many troubles and puzzles since the lee - model . in summary contrary to hqt in ( a)cqt the causal and the anticausal sector of the theory are completely disconnected , which is the price we have to pay to achieve simultaneously lorentz covariance , causality , locality and analyticity .
it is interesting to observe present strong renewed interest ( see e.g. @xcite ! ) in quantum - theoretic questions of the lee - model by physicists and mathematicians , yet now under a different headline called _ pt - symmetry _
@xcite , i.e. symmetry under space- and time - reversal . may be real .
based on a fascinating conjecture by d. bessis ( and j. zinn - justin ) of 1992 , that the complete spectrum of this hamilton operator is real and positive , c.m .
bender and s. boettcher @xcite ( see also @xcite ! ) suggested in 1997 that a whole class of such hamilton operators possess this feature due to their _ antiunitary _ @xcite _ pt - symmetry_. these developments were accompanied by related investigations in the context of ( anharmonic ) quartic oscillators ( see e.g. @xcite ) .
a review provided in 2001 by m. znojil as a preprint math - ph/0104012 ( unfortunately published very delayed in 2004 @xcite ) containing a lot of important references to related work showed that the new research field had again , yet independently reached a similar level of understanding as the disciples of the lee - model around 1970 , who were fascinated by imaginary bare couplings required to obtain real couplings in the renormalized hamilton operator of the lee - model . in @xcite one does not only find discussed the concept of _ pseudo - hermiticity _ , _ pseudo - unitarity _ and _ indefinite metric _ ( i.e. @xmath82 ) .
it is also explained under the headline _ spontaneous broken pt symmetry _ the situation , when the hamilton operator develops complex ghosts , i.e. pairs of mutually complex conjugate complex - valued eigenvalues associated to a biorthogonal eigenbasis of states .
meanwhile there appeared several important new contributions to the field , the enumeration of which would go far beyond the scope of this work .
we just want selectively refer here to contributions e.g. to the formalism of _ pseudo - hermiticity _ ( e.g. @xcite and references therein ) , _ spontaneous breakdown of pt - symmetry _ ( e.g. @xcite ) , _ pt - symmetric potentials _
( e.g. @xcite ) , _ pt - symmetric matrix models _ @xcite , _ reality proofs for spectra of pt - symmetric hamilton operators _ ( e.g. @xcite ) , _ the c- and cpt - operator in pt - symmetric qm _
( e.g. @xcite ) , _ pt - symmetric supersymmetry _
( e.g. @xcite ) , and _ generalized creation / annihilation operators _ ( e.g. @xcite ) . ] due to its intimate relation to the lee - model the researchers investigating pt - symmetric qt will have to answer the same questions as were raised in the context of the lee - model , i.e. the questions about unitarity , analyticity , poincar - covariance , causality , locality in time - dependent pt - symmetric scattering problems .
like in the foregoing discussion the solution to the questions is provided in terms of ( a)cqt , the spacial representation of which is shown below to correspond to the _ holomorphic representation _ in complex analysis , while pt - symmetric qt appears just a subset of ( non - hermitian ) ( a)cqt .
as mentioned above , n. nakanishi @xcite investigated in 1972 a lagrangean ( see also @xcite ! ) for a kg field @xmath83 with complex mass @xmath84 ( and the hermitian conjugate field @xmath85)was@xmath861999 - 2000 independently rederived by the author ( see e.g. @xcite ) . ] . here
we want to introduce immediately isospin and to consider a set of @xmath33 equal complex mass kg fields @xmath87 ( @xmath88 ) ( i.e. a charged `` nakanishi field '' with isospin @xmath89 ) described by the larangean @xmath90 the lagrange equations of motion for the causal and anticausal `` nakanishi field '' @xmath87 and @xmath91 , i.e. @xmath92 and , are solved by a laplace - transform .
the result is : @xmath93 \big|_{p^0 = \omega(\vec{p}\ , ) } \ ; , \nonumber \\ \phi_r^+ ( x ) & = & \int \frac{d^3 p}{(2\pi ) ^3 2\omega^\ast ( \vec{p}\ , ) } \big [ c ( \vec{p } , r ) \,e^{\displaystyle - i p^\ast x } + a^+ ( \vec{p } , r ) \,e^{\displaystyle i p^\ast x } \big ] \big|_{p^0 = \omega(\vec{p})}\ ; , \end{aligned}\ ] ] where we defined @xmath94 and @xmath95 on the basis of the complex `` frequency '' @xmath96 ( @xmath97 ) . '' for complex arguments which has been illuminated by n. nakanishi @xcite . nowadays
it may be embedded in the framework of ( tempered ) ultradistributions @xcite . ]
the theory is quantized by claiming _
canonical equal - real - time commutation relations_. and @xmath91 are given by @xmath98 and @xmath99 .
] the non - vanishing commutation relations in configuration space are ( with @xmath100 ) @xmath101 = i\ , \delta^{\ , 3 } ( \vec{x } - \vec{y}\ , ) \ ; \delta_{rs}$ ] and @xmath102 = i\ , \delta^{\ , 3 } ( \vec{x } - \vec{y}\ , ) \ ; \delta_{rs}$ ] .
the resulting non - vanishing momentum - space commutation relations , which display an _ indefinite metric _ ! ] , are ( @xmath100 ) : @xmath103 } & = & ( 2\pi)^3 \ , 2 \ ; \omega \,(\vec{p}\,)\;\ , \delta^{\ , 3 } ( \vec{p } - \vec{p}^{\,\,\prime}\ , ) \ ; \delta_{rs } \ ; , \nonumber \\ { [ \ , c \ , ( \vec{p},r ) \ ; , \ ; a^+ ( \vec{p}^{\,\,\prime},s ) \ , ] } & = & ( 2\pi)^3 \ , 2 \ ; \omega^\ast(\vec{p}\,)\ ; \delta^{\ , 3 } ( \vec{p } - \vec{p}^{\,\,\prime}\ , ) \ ; \delta_{rs } \ ; .\end{aligned}\ ] ] the hamilton operator is derived by a standard legendre transform @xcite : @xmath104 the `` nakanishi - kg propagator '' @xmath105 is obtained by _ real - time ordering _ of causal kg fields @xcite , i.e. @xmath106\,\left|0\right>$ ] .. the anticausal `` nakanishi - kg propagator '' is obtained by hermitian conjugation or by a vacuum expectation value of an _ anti - real - time ordered _ product of two anticausal fields . for intermediate states with complex mass these propagators
lead to _ poincar covariant _ results , _ if _ causal and anticausal states do not interact . at each interaction
vertex coupling to _ intermediate _ complex mass fields there holds _ exact _
4-momentum conservation . only if complex mass fields with finite @xmath107 appeared as asymptotic states or causal and anticausal states could interact , then poincar covariance would be violated ! ]
a lorentz transformation @xmath108 for a given ( symmetric ) metric @xmath109 is defined by @xmath110 .
let @xmath111 be a timelike unit 4-vector @xmath112 and @xmath113 an _ arbitrary complex _
4-vector with @xmath114 .
we want to construct @xcite a lorentz transformation @xmath115 relating the 4-vector @xmath113 with its `` restframe '' , i.e. @xmath116 and @xmath117 .
the independent quantity @xmath118 being also a lorentz transformation is simultaneously admissible , as @xmath119 and @xmath120 . in a first step
we define now the _ inversion matrix _ @xmath121
. then we note : if there is a solution @xmath115 ( @xmath122 ) of the defining equation @xmath110 relating the 4-vector @xmath113 with its `` restframe '' , then @xmath123 ( @xmath124 ) will be also a such a solution , respectively . keeping these remaining solutions in mind
we find after employing some general ansatz for @xmath125 _ four _ independent solutions @xmath126 and @xmath127 being related by @xmath128 , which solve the defining equation @xmath110 and relate the 4-vector @xmath113 with its `` restframe '' \big [ n_{\,\rho } \pm \frac{\xi_{\,\rho}}{\sqrt{\xi^2 } } \big ] \ , \right\ } p^{\,\rho}_{\,\;\,\nu}$ ] . ] .
@xmath129 , @xmath130 , @xmath131 and @xmath132 represent the well known orthochronous - proper , orthochronous - improper , nonorthochronous - proper , and nonorthochronous - improper lorentz transformations , respectively ! , @xmath133 , and @xmath134 , @xmath135 for @xmath136 . ]
orthochronous - proper lorentz transformations respect @xmath137 for @xmath138 . in the following text
we use the abbreviation @xmath139 . to obtain lorentz - boost matrices for a particle with _ complex _ mass @xmath74 and 3-momentum @xmath140 being on its _ complex mass - shell _ @xmath141
, we simply have to set @xmath142 and to remember , that we live in a world with a metric @xmath143 . , if the imaginary part of the complex masses of particles appearing as asymptotic states are infinitesimally close to zero and causal and anticausal states do not interact . in ( a)cqt _
causal _ states are lorentz transformed by @xmath144 , while _ anticausal _ states are lorentz transformed by @xmath145 .
hence , in ( a)cqt states are against our original expectations @xcite - _ not _ representations of the covering group of the _ complex _ lorentz group @xmath146 ( or , more generally , the covering group of the respective poincar group ) @xcite which links continuously @xmath144 ( containing the unitity @xmath147 ) and @xmath148 ( containing the pt - transformation @xmath149 ) . ]
the causal dirac equation and its relatives obtained by hermitian conjugation / transposition are @xcite @xmath150 , @xmath151 , @xmath152 , and @xmath153 ( with @xmath84 and @xmath154 ) .
note that @xmath155 is an isospin index and @xmath156 , @xmath157 , @xmath158 , @xmath159 are grassmann fields .
the underlying lagrangean is given by @xcite ( see also @xcite ) ( @xmath160 yields neutrinos ! ) , @xmath161 ) are e.g. represented by the isospin combination @xmath162 , while the positron ( @xmath163 ) and electron ( @xmath164 ) are represented by @xmath165 .
@xmath166 , @xmath167 or @xmath168 , @xmath169 are non - hermitian fields describing , respectively , pairs of causal neutral particles with equal complex mass .
this holds for bosons and fermions .
it leads like in the bosonic case to the fact that the anti - fermions have the same intrinsic parity as the fermions , contrary to the conclusions of v.b .
berestetskii obtained in hqt and mentioned in the introduction . in spite of this feature ( a)cqt reproduces exactly the high precision results of qed . ] @xmath170 4-spinors @xmath171 in complex 4-momentum space ( @xmath172 ) are introduced by the defining equation @xmath173 .
the spinors are normalized according to @xmath174 \ ; \sum_s \ , u(p , s ) \ ; \overline{v^c}(p , s ) = \ ; p\!\!\!/ + \sqrt{p^2}$ ] for @xmath175\;\ne\;0 $ ] .
equations of motions are solved by a laplace - transformation . ,
@xmath176 , @xmath177 we obtain : @xmath178 \big|_{p^0 = \omega ( \vec{p}\ , ) } \ , , \nonumber \\ \psi_r^c ( x ) & = & \sum\limits_s \int \frac{d^3p}{(2\pi)^3 \ ; 2 \ , \omega^\ast ( \vec{p}\ , ) } \ , [ \,e^{+\ , i \ , p^\ast\cdot x } b^+_{\,r } ( p , s ) \ ,
u^c(p , s ) + e^{- \ , i \ , p^\ast\cdot x } d_{\,r } ( p , s ) \ ,
v^c(p , s)\ , ] \big|_{p^0 = \omega ( \vec{p}\ , ) } \ , , \nonumber \\
\overline{\psi_r } ( x ) & = & \sum\limits_s \int \frac{d^3p}{(2\pi)^3 \ ; 2 \ , \omega^\ast ( \vec{p}\ , ) } \ , [ \,e^{+ i \ , p^\ast \cdot x } \ , b^+_{\,r } ( p , s ) \ , \bar{u}(p , s ) \ ; + e^{- \ , i \ , p^\ast\cdot x } d_{\,r } ( p , s ) \ ,
\bar{v}(p , s)\ , ] \big|_{p^0 = \omega ( \vec{p}\ , ) } \ , , \nonumber \\
\overline{\psi^c_r } ( x ) & = & \sum\limits_s \int \frac{d^3p}{(2\pi)^3 \ ; 2 \ , \omega ( \vec{p}\ , ) } \ , \ , \ ; [ \,e^{- i \ , p \cdot x } \;\,\ , b_{\,r } ( p , s ) \ ,
\overline{u^c}(p , s ) + e^{+ \ , i \ ,
p \cdot x } d^+_{\,r } ( p , s ) \
, \overline{v^c}(p , s)\ , ] \big|_{p^0 = \omega ( \vec{p}\,)}\ , , \nonumber \end{aligned}\ ] ] and @xmath179 note that the spinors obey the analyticity property @xmath180 . the ( anti)causal dirac equation is lorentz covariant due to standard transformation properties of spinors and @xmath181-matrices , i.e. @xmath182 and @xmath183 .
the causal `` nakanishi - dirac propagator '' @xmath184 of a spin 1/2 fermion is obtained by _
standard fermionic real - time ordering _ of causal dirac fields , i.e. @xmath185 \left|0\right>$ ] .
the anticausal `` nakanishi - dirac propagator '' is obtained by hermitian conjugation or by a vacuum expectation value of a _ anti - real - time ordered _ product of two anticausal dirac fields . ] .
with @xmath186 and @xmath187 we define the polarization vectors @xmath188 . in the chosen unitary gauge they obey @xmath189 , @xmath190 , and @xmath191 . based on these polarization vectors
the bosonic field operators for causal vector fields @xmath192 ( @xmath193 isospin index ) and anticausal vector fields @xmath194 are as the nakanishi - kg field introduced by @xmath195 \big|_{p^{0 } = \omega ( \vec{p}\,)\,}$ ] .
the causal kg equation is second order in time , which gets transparent by factorizing the respective _ causal kg differential operator _ according to @xmath196 . and the respective _ adjoint
causal kg differential operator _
@xmath197 being obtained by partial integration and providing us with the left eigensolutions of the causal kg differential operator .
the anticausal kg equation @xmath198 and the adjoint anticausal kg equation @xmath199 are obtained by hermitian conjugation from the adjoint causal kg equation and the causal kg equation , respectively . ]
inspection of the `` factorized '' causal and anticausal kg equations and their adjoints .
] yields eight underlying differential equations of first order in time : @xmath200 for a non - hermitian hamilton operator @xmath201 we can construct corresponding schrdin - ger equations by replacing the spacial parts of kg differential operators by respective hamilton operators ( replace e.g. @xmath202 by @xmath201 , @xmath203 by @xmath204 , etc . ! ) .
hence , the ( anti)causal schrdinger theory is described by the following eight equations : @xmath205 causal equations ( left column ) are related to corresponding anticausal equations ( right column ) by hermitian conjugation . from the causal equations ( left column )
we can derive the continuity equations @xmath206 and @xmath207 .
the right hand side of these equations leads for standard even non - hermitian hamilton operators being quadratic in the momentum at most to ( spacial ) surface terms ! the corresponding non - vanishing in general _ complex _ densitities @xmath208 and @xmath209 being in the standard way related to in general _ complex _ conserved charges may be interpreted as _ complex
_ probability densities @xcite which replace born s @xcite suggested ansatz @xmath210 for a _ real _ probability density .
we refer here also to other interesting attempts to change or extend born s scalar product , e.g. @xcite ! the hamilton operator of the ( anti)causal harmonic oscillator in 1-dim .
qm is given in analogy to the hamilton operator eq .
( [ nakham1 ] ) of the nakanishi model by : @xmath211_\pm + \frac{1}{2 } \ , \omega^\ast \ , [ a^+ , c\ , ] _ \pm \qquad ( h_a = h^+_c)\ ] ] ( @xmath212 for bosons / fermions + \frac{1}{2 } \omega^\ast [ b^+ , d ] $ ] with @xmath213 etc .. ] ) @xcite .
the relevant ( anti)commuation relations are : @xmath214_\mp } & { [ c , a^+]_\mp } \\ { [ a , c^+]_\mp } & { [ a , a^+]_\mp } \end{array}\right ) \quad = \quad \left ( \begin{array}{cc } 0 & 1 \\ 1 & 0 \end{array}\right ) \quad = \quad \mbox{``indefinite metric '' } \ ; .\end{aligned}\ ] ] of course there also holds : @xmath215_\mp } & { [ c , a]_\mp } \\ { [ a , c]_\mp } & { [ a , a]_\mp } \end{array}\right ) \quad = \quad \left ( \begin{array}{cc } { [ c^+,c^+]_\mp } & { [ c^+,a^+]_\mp } \\ { [ a^+,c^+]_\mp } & { [ a^+,a^+]_\mp } \end{array}\right ) \quad = \quad \left ( \begin{array}{cc } 0 & 0 \\ 0 & 0 \end{array}\right ) \ ; .\end{aligned}\ ] ] obviously there holds @xmath216=0 $ ] .
one of the important problems is now to find the right and left eigensystem of the hamilton operator for given energy @xmath217 , i.e. to solve the following two ( stationary ) schrdinger equations @xmath218 and @xmath219 .
the resulting ( normalized ) _ normal _ right eigenstates @xmath220 and left eigenstates @xmath221 are given by ( bosons : @xmath222@xmath223 ; fermions : @xmath224 ) @xmath225 they are solutions of the equations @xmath226 and @xmath227 for the eigenvalues @xmath228 .
the eigenvalues @xmath229 are obviously _ real _ , while the eigenvalues @xmath230 and @xmath231 form a
_ complex conjugate _
pair for @xmath232 which arises typically for the case of broken `` pt''-symmetry .
( bi)orthogonal _ ( here _ normal _ ) eigenstates are complete : @xmath233 we note that one can construct ( e.g. @xcite ) for the ( anti)causal harmonic oscillator also an
_ abnormal _ basis of eigenstates being complete with respect to negative ( complex ) energy eigenvalues by applying annihilation ( creation ) operators on a newly defined _ dual vacuum
_ state @xmath234 ( @xmath235 ) annihilating creation ( annihilation ) operators and creating annihilation ( creation ) operators , respectively .
as in the traditional formulation of qm it is now natural to look for a spacial representation of the `` representation free '' oscillator introduced above . the particular complication induced by the existence of _ two _ types of annihilation operators and creation operators ( @xmath236 , @xmath237 ) indicating a _ doubling _ of the degrees of freedom compared to the traditional harmonic oscillator
is overcome by replacing the originally _ real _ spacial variable @xmath238 by a _ complex _ variable @xmath239 and its complex conjugate @xmath240 . by an overall shift @xmath241 .
] this replacement of functions of one real argument @xmath242 ( @xmath243 ) by respective functions @xmath244 of complex arguments @xmath245 is well known and used in complex analysis @xcite under the terminology `` holomorphic representation '' ( e.g. @xcite ) .
replacing the right and left eigenstates @xmath246 and @xmath247 of the position operator by respective states in holomorphic representation @xmath248 and @xmath249 , we can analogously to traditional qm denote the schrdinger equations and their adjoints given e.g. in the first column of eq .
( [ schreq1 ] ) in their holomorphic representation by : @xmath250 the spacial integration contours @xmath251 are to be performed such that there holds the generalized completeness relation @xmath252 . inversely , by the used notation it is understood that there holds a generalized orthogonality relation @xmath253 , in which the @xmath254-distributions for complex arguments are assumed to exist with respect to the chosen integration contours mentioned above .
the holomorphic representation of the ( anti)causal _ bosonic _ harmonic oscillator with @xmath255 is provided by a translational invariant hamilton operators @xmath256 and @xmath257 with @xmath258 momentum operators @xmath259 and @xmath260 are introduced by @xmath261 and @xmath262 .
obviously there holds the following correspondence between annihilation / creation operators in the representation free case and in the holomorphic representation : @xmath263 the inverse correspondence is : @xmath264 by this correspondence it is straight forward to construct the _ normal _ eigensolutions of the stationary schrdinger equations @xmath265 and @xmath266 of the ( anti)causal bosonic harmonic oscillator in holomorphic representation .
they are given by : @xmath267 & = & i^{\,n+m}\ , \sqrt{\frac{|m\omega|}{2^{\,n+m } \
, n!\ , m ! \ , \pi } } \;\ , \exp\left(-\,\frac{1}{2}\ , ( \xi^2 + \xi^{\ast\,2}\,)\right ) \ ;
h_n(\xi ) \,h_m(\xi^\ast ) \ ; , \nonumber \\[1 mm ] \lefteqn{\big<\!\big < n , m\big|z , z^\ast \big > \ ; = \ ; \frac{1}{\sqrt{m!\,n ! } } \,\ , \big<\!\big < 0\big| \ ; c^m \ ; a^n \big|z , z^\ast \big > } \nonumber \\[1 mm ] & = & ( -i)^{\,m+n}\ , \sqrt{\frac{|m\omega|}{2^{\,m+n } \ , m!\ , n ! \ , \pi } } \;\ , \exp\left(-\,\frac{1}{2}\ , ( \xi^{\ast\,2}+\xi^2\,)\right ) \ ; h_m(\xi^\ast ) \ , h_n(\xi ) \ ; , \end{aligned}\ ] ] with @xmath268 and @xmath269 .
the inverse oscillator length @xmath270 is here _ complex _ valued . and
@xmath271 are defined by : @xmath272 ] and @xmath273 are quite different from the ones of glauber coherent states @xcite : @xmath274 with @xmath275 . ]
based on the previous considerations it is now straight forward to introduce operations @xmath82 , @xmath276 , @xmath277 , @xmath278 , and @xmath279 involving reflections of space and/or time , whose effect on @xmath239 , @xmath259 , @xmath240 , @xmath260 , @xmath13 , @xmath280 , @xmath237 , @xmath236 , and the imaginary unit @xmath281 is specified in table [ tabop1 ] . the standard space reflection @xmath82 transforms the holomorphic coordinates @xmath239 and @xmath240 and the respective momenta @xmath259 and @xmath260 in the following way : @xmath282 , @xmath283 , @xmath284 , @xmath285
. these operations result in a transformation of _ bosonic _ annihilation and creation operators according to the rules @xmath286 , @xmath287 , @xmath288 , @xmath289 , while the imaginary unit @xmath281 remains invariant ( @xmath290 ) . ]
.effect of the operations @xmath82 , @xmath276 , @xmath277 , @xmath278 , and @xmath279 within _ bosonic _ systems . [ cols="<,<,<,<,<,<",options="header " , ] the time - reversal operation is not unique within ( anti)causal qm . in table [ tabop1 ] we suggest within the framework of a bosonic system two important time - reversal operations @xmath276 and @xmath277 . is supposed to exchange in- and out - states , while the anti - linear operation @xmath277 is even for fermionic systems to be identified as usual with the hermitian conjugation , which changes causal systems into anticausal systems , and anticausal systems into causal systems .
amusingly , the author of @xcite working within a hermitian framework comes to the opposite conclusion , i.e. that the anti - unitary time - reversal interchanges initial and final states , while the unitary time - reversal does not .
the existence of two distinct time - reversal operations explains nicely , why the tcp - operations in `` schwinger - pauli convention '' and `` wightman convention '' yield different results ( see e.g. appendix a.5 . in @xcite ! ) .
it is tempting to define the time - reversal operation @xmath276 for fermionic systems like in the case of bosonic systems . yet
to our understanding one should additionally take into account the anticommutation properties of fermionic operators , when interchanging initial and final states . ] as shown in table [ tabop1 ] it is straight forward to construct from the two suggested time - reversal operations two distinct `` pt''-transformations , i.e. @xmath278 and @xmath279 .
everybody talking about `` pt''-symmetry should be well aware which of the two operations he / she is addressing !
we observe also that the hamilton operator of the bosonic 1-dimensional ( anti)cau - sal harmonic oscillator remains invariant under all introduced transformations , i.e. @xmath291.-transformation is just the hermitian conjugation , the @xmath277-transformation symmetry of the hamilton operator reflects just the hermiticity of the overall hamilton operator . yet , as already mentioned above , the hamilton operator should be more accurately called _ pseudo - hermitian _ , rather than _
, as the commutation relations of the respective creation and annihilation operators are governed by an _ indefinite metric_. ] ( and respective lagrangeans ) with @xmath292 . in this case
there appear only real energy eigenvalues and complex - valued mutually complex - conjugate pairs of energy eigenvalues .
this is what we observe traditionally for pt - symmetric hamilton operators in the phase of pt - symmetry or spontaneously broken pt - symmetry . yet
nhqt allows also models for which @xmath293 .
then we are in an interesting domain , where complex energy eigenvalues do _ not necessarily _ appear as complex conjugate pairs , while analyticity , causality , poincar - invariance & locality still persists , if there is at least one eigenstate serving as an asymptotic state .
we shall devote future work to such models . ]
the causal ( anticausal ) part of the hamilton operator @xmath294 ( @xmath295 ) of the bosonic ( anti)causal harmonic oscillator are only invariant under @xmath82- , @xmath276- , and @xmath278-transformations ( @xmath296 , @xmath297 ) , while they are not invariant under @xmath277-/@xmath279-transformations ( @xmath298 , @xmath299 , @xmath300 , @xmath301 ) .
as we mentioned in the introduction ( a)cqt has to obey the postulate of non - interaction of ( non - hermitian ) causal and anticausal degrees of freedom to avoid analyticity violations .
this postulate defines a rather tricky interplay between the underlying hermitian degrees of freedom , which we shall here call _ shadow fields _ according to h.p .
stapp @xcite .
relevant for the dynamical description but do not contribute to probability @xmath7 '' _ ! ] we want to illustrate this decomposition into hermitian shadow fields for the lagrangeans of the free ( anti)causal kg and dirac theory and for convenience for the hamilton operators of the 1-dimensional ( anti)causal bosonic and fermionic harmonic oscillator , i.e. respectively for : @xmath302 \ ; + \frac{1}{2 } \ ; \omega^\ast \ , [ \ , b^+ , d\ , ] \ , .
\qquad \end{aligned}\ ] ] @xmath78 , @xmath85 , @xmath303 , @xmath304 are decomposed in hermitian shadow fields @xmath305 , @xmath306 , @xmath307 , @xmath308 by @xmath309 , @xmath310 , and @xmath311 , @xmath312 , while the respective replacement rules for the bosonic / fermionic harmonic oscillator are listed in eq .
( [ reprules1 ] ) . as a result of the replacements we obtain : @xmath313 ( \phi_{(1 ) } ( x))^2 \big ) \nonumber \\[1 mm ] & - & \frac{1}{2 } \,\big ( ( \partial \phi_{(2 ) } ( x ) ) ^2 - \mbox{re}[m^2 ] ( \phi_{(2 ) } ( x))^2 \big ) + \mbox{im}[m^2 ] \ ; \phi_{(1 ) } ( x ) \,\phi_{(2 ) } ( x ) \ ; , \nonumber \\ { \cal l } ( x ) & = & \frac{1}{2 } \ , \overline{\psi}_{(1 ) } ( x ) \ , \big ( \frac{1}{2 } \ ,
\stackrel{\;\,\leftrightarrow}{\not\!\partial } \ ! - \,\mbox{re}[m ] \big ) \psi_{(1 ) } ( x ) - \frac{1}{2 } \ ; \overline{\psi}_{(2 ) } ( x ) \ , \big ( \frac{1}{2 } \,i \!\ ! \stackrel{\;\,\leftrightarrow}{\not\!\partial } \ ! - \ , \mbox{re}[m ] \big ) \psi_{(2 ) } ( x ) \nonumber \\ & + & \frac{1}{2 } \ ; \mbox{im}[m ] \ ; \big ( \overline{\psi}_{(2 ) } ( x ) \;\psi_{(1 ) } ( x ) + \overline{\psi}_{(1 ) } ( x ) \ ; \psi_{(2 ) } ( x ) \big ) \ ; , \nonumber \\
h & = & \frac{1}{2}\,\mbox{re}[\omega ] \
, \big(\ { a^+_{{}_{(1 ) } } , a_{{}_{(1)}}\ } - \{a^+_{{}_{(2 ) } } , a_{{}_{(2)}}\}\big ) - \mbox{im}[\omega ] \ , \big(a^+_{{}_{(1 ) } } a_{{}_{(2 ) } } + a^+_{{}_{(2 ) } } a_{{}_{(1)}}\big ) \ ; , \nonumber \\ h & = & \frac{1}{2}\,\mbox{re}[\omega ] \ , \big(\ ; [ \ , b^+_{{}_{(1 ) } } , b_{{}_{(1)}}\ , ] - [ \ , b^+_{{}_{(2 ) } } , b_{{}_{(2)}}\,]\ ; \big ) - \mbox{im}[\omega ] \ , \big(\,b^+_{{}_{(1 ) } } b_{{}_{(2 ) } } + b^+_{{}_{(2 ) } } b_{{}_{(1)}}\big ) \ ; . \end{aligned}\ ] ] first we observe that in all cases one shadow field has _ positive _ norm , the other has _ negative _ norm displaying the underlying _
indefinite metric_. secondly we recall that shadow fields are not described by causal or anticausal propagators , but by _ acausal _ linear combinations which reduce for quasi - real masses to _ principal value propagators _ or _
@xmath254-distributions_. finally we see that in all cases the lagrangeans / hamilton operators are not diagonal in the shadow fields , while the interaction terms are proportional to the imaginary part of @xmath314 , @xmath74 or @xmath315 being necessary to make the theory ( anti)causal and analytic . in the kg theory , or @xmath316 in the holomorphic representation of the bosonic harmonic oscillator ) leading not only to a _ violation of causality _ , yet also as described in the introduction in the context of the lorentz non - invariance of n. nakanishi s complex - ghost relativistic field theory to the _ loss of analyticity _ and _ lorentz covariance_. ] according to m. znojil @xcite pt - symmetric qt is characterized by _ quasi - parities _
@xmath317 , while @xmath318 is introduced by the orthogonality and completeness relations @xmath319 and @xmath320 on the right eigenstates of the hamilton operator ( see also @xcite ! ) . by decomposing the eigenstates @xmath321 of the ( anti)causal bosonic harmonic oscillator given in eq .
( [ hostates1 ] ) into hermitian shadow states we observe that in ( a)cqt the quasi - parity in the space of shadow states can take the values @xmath322 and @xmath323 . hence , ( a)cqt appears here clearly as a more general framework than pt - symmetric qt . for complex mass fields several important continuity equations do not hold in their traditional form . , are @xmath324 = 2\ , i \ , \varepsilon \,\phi^+ ( x ) \ , \phi ( x)$ ] , @xmath325 = -\ , 2\ , i \ , \varepsilon \,\bar{\psi } ( x ) \ , \psi ( x)$ ] , and @xmath326 + \frac{1}{2\ , m } \!\stackrel{\rightarrow}{\nabla } \!\cdot \ , [ \ , \psi^+ ( x ) \stackrel{\rightarrow}{\nabla } \psi ( x ) - \psi ( x ) \stackrel{\rightarrow}{\nabla } \psi^+ ( x ) \ , ] = \psi^+ ( x ) \ , [ \ , v(x ) - v^+(x ) \ , ] \ , \psi ( x ) - \frac{i\,\varepsilon}{2\ , m } \ , [ \ , \psi^+ ( x ) \stackrel{\rightarrow}{\nabla } { } ^2 \ , \psi ( x ) + \psi ( x ) \stackrel{\rightarrow}{\nabla } { } ^2 \ , \psi^+ ( x ) \ , ] $ ] , respectively .
note that all these currents are not conserved due to the finite imaginary part of the mass ( here @xmath327 ) or non - hermitian _ causal potentials _
being laplace - transforms of _ causal _ propagators ! the breakdown of the traditional probability concept of max born was sketched already in section [ maxb1 ] .
note that even the `` massless '' causal dirac equation @xmath329 is not chiral invariant ( see the discussion in @xcite ! ) ! ] fortunately there can be derived @xcite ( see also section [ maxb1 ] ! ) new exact conservation laws respected even by fields of arbitrary complex mass .
+ _ norm or probability conservation in ( a)cqt : _ ( anti)causal kg / dirac fields are decomposed into positive / negative _ complex _ frequency parts ( i.e. @xmath330 , @xmath331 )
. subtraction of equations of motion for @xmath332 , @xmath333 , @xmath334 and respective adjoints yields for the kg , dirac and schrdinger theory continuity equations @xmath335 = 0 $ ] , @xmath336 = 0 $ ] , and @xmath337 + \frac{1}{2\ , m } \!\stackrel{\rightarrow}{\nabla } \!\cdot \ , [ \ , \tilde{\psi } ( x ) \stackrel{\rightarrow}{\nabla } \psi ( x ) - ( \stackrel{\rightarrow}{\nabla } \tilde{\psi } ( x ) ) \,\psi ( x ) ] = 0 $ ] , respectively .
note that all currents are conserved and in general non - zero , even for the neutral kg field !
we conclude that the schrdinger norm is @xmath338 , and _ not _ @xmath339 ! + _ charge conservation in ( a)cqt : _ simply charged ( anti)causal systems are introduced according to the isospin concept .
we define for the kg theory @xmath340 , for the dirac theory @xmath341 , and for the schrdinger theory @xmath342 and @xmath343 .
subtraction of causal equations of motion and the respective adjoints for these fields leads for the kg , dirac and schrdinger theory to the ( charge conserving ) continuity equations @xmath344 = 0 $ ] , @xmath345 = 0 $ ] , and @xmath346 + \frac{1}{2\ , m } \stackrel{\rightarrow}{\nabla } \!\cdot \ , [ \ , \tilde{\psi}_\mp ( x ) \stackrel{\rightarrow}{\nabla } \psi_\pm ( x ) - ( \stackrel{\rightarrow}{\nabla } \tilde{\psi}_\mp ( x ) ) \,\psi_\pm ( x ) \ , ] = 0 $ ] , respectively .
( @xmath347 , @xmath348 ) or @xmath349 ( @xmath350 , @xmath351 ) equal to zero ! ) .
the reason is a cancellation of underlying norm currents ! in schrdinger theory
being just a theory first order in time this cancellation does not occur .
note that the neutral dirac field does _ not _ admit any _ abelian _ gauge couplings due to transposition properties @xmath352^t = \
! - \,\overline{\psi^{\,c } } ( x ) \not\!\!a(x ) \ , \psi ( x)$ ] , @xmath353^t = - \,\overline{\psi^{\,c } } ( x ) \,\sigma^{\mu\nu } \ , f_{\mu\nu}(x ) \ , \psi ( x ) $ ] .
] the concept of local ( non)abelian gauge invariance in ( a)cqt is sketched in the footnote .
( with @xmath354 ) .
it is invariant under the local gauge transformations @xmath355 $ ] , @xmath356 , and @xmath357 .
the non - abelian yields @xmath358 and @xmath359 . _
non - abelian _ gauge fields admit minimal coupling even to neutral fermions , _ if _ @xmath360^t =- a^{\mu}(x)$ ] , as @xmath361^t = + \,\overline{\psi^{\,c } } ( x ) \not\!\ !
a(x ) \ , \psi ( x)$ ] and @xmath353^t = + \,\overline{\psi^{\,c } } ( x ) \,\sigma^{\mu\nu } \ , f_{\mu\nu}(x ) \ , \psi ( x ) $ ] . ] as @xmath363 is not a probability density in ( anti)causal schrdinger theory @xcite , @xmath364 is _ not _ to be interpreted as a _ transition probability _ in ( anti)causal scattering theory ! in ( anti)causal scattering theory we have instead to consider a quantity @xmath365 , where @xmath362 ( @xmath366 ) is called the _
conjugate @xmath276-matrix_. the construction of @xmath362 has been addressed in @xcite .
first we denote the causal @xmath276-matrix @xmath367 in the interaction picture : @xmath368 & = & \left<\!\left<0\right|\right .
{ \cal a}(\vec{p}^{\,\,\prime}_{\,n_f } ) \ldots { \cal a}(\vec{p}^{\,\,\prime}_{\,1 } ) \ ; t [ \ , \exp(+\,i \ , s_{int } ) - 1 \ , ] \ ; ( { \cal c}(\vec{p}_{\,1}))^+ \ldots ( { \cal c}(\vec{p}_{\,n_i}))^+ \left|0\right>_c \nonumber \\[1 mm ] & \equiv & \left<\!\left<\psi_f\right|\right .
t [ \ , \exp(i \ , s_{int } ) - 1 \ , ] \left|\psi_i\right>\end{aligned}\ ] ] with @xmath369 and @xmath370 .
call @xmath371 the overall number of fermionic operators in the intitial and final state
. then we obtain for @xmath362 : @xmath372\ ; ( { \cal c}(\vec{p}^{\,\,\prime}_{\,1}))^+ \ldots ( { \cal c}(\vec{p}^{\,\,\prime}_{\,n_f}))^+ \big)^t \left|\bar{0}\right>_c \nonumber \\ & \stackrel{!}{= } & ( -1)^{n_f ( n_f-1)/2 } \times \nonumber \\ & & \times \left<\!\left<\bar{0}\right|\right . ( { \cal c}(\vec{p}^{\,\,\prime}_{\,n_f}))^+ \ldots ( { \cal c}(\vec{p}^{\,\,\prime}_{\,1}))^+ \ ; \big ( t \ , [ \ , \exp(-\,i \ , s_{int } ) - 1 \ , ] \big)^t { \cal a}(\vec{p}_{\,1 } ) \ldots { \cal a}(\vec{p}_{\,n_i } ) \left|\bar{0}\right>_c \nonumber \\ & = & ( -1)^{n_f ( n_f-1)/2 } \times \nonumber \\ & & \times \left<\!\left<\bar{0}\right|\right . ( { \cal c}(\vec{p}^{\,\,\prime}_{\,n_f}))^+ \ldots ( { \cal c}(\vec{p}^{\,\,\prime}_{\,1}))^+ \ ; \overline{t } \ , [ \ , \exp(-\,i \ , s^{\,t}_{int } ) - 1 \ , ] \ ; { \cal a}(\vec{p}_{\,1 } ) \ldots { \cal a}(\vec{p}_{\,n_i } ) \left|\bar{0}\right>_c \nonumber \\ & \stackrel{!}{= } & ( -1)^{n_f ( n_f-1)/2 } \times \nonumber \\ & & \times \left<\!\left<\bar{0}\right|\right .
( { \cal c}(\vec{p}^{\,\,\prime}_{\,n_f}))^+ \ldots ( { \cal c}(\vec{p}^{\,\,\prime}_{\,1}))^+ \ , \overline{t } \ , [ \ , \exp(-\,i \ , s_{int } ) - 1 \ , ] \ , { \cal a}(\vec{p}_{\,1 } ) \ldots { \cal a}(\vec{p}_{\,n_i } ) \left|\bar{0}\right>_c \label{contmat1}\end{aligned}\ ] ] with @xmath373 and @xmath374 and @xmath234 ( and @xmath375 ) being the _ dual vacuum _ annihilating creation operators and creating annihilation operators ( see also section [ repfree1 ] ! ) . above
we used the useful identity @xmath376 \big)^t \ ; = \ ; \overline{t}\ , \big [ \,({\cal o}(x_1))^t \ldots ( { \cal o}(x_n))^t \ , \big]$ ] holding for bosonic and fermionic operators .
( [ contmat1 ] ) may be rewritten in the following way : @xmath377\ ; ( { \cal c}(\vec{p}^{\,\,\prime}_{\,1}))^+ \ldots ( { \cal c}(\vec{p}^{\,\,\prime}_{\,n_f}))^+ \ , \big(\left<\!\left<\bar{0}\right|\right.\big)^t \nonumber \\ & \equiv & \left|\right.\!\overline{\psi}_i\!\left.\right>^t \ , t [ \ , \exp(- i \ , s_{int } ) - 1 \ , ] \left(\left<\!\left<\right.\right.\!\overline{\psi}_f\!\!\left.\left.\right|\right.\right)^t \ ; .\end{aligned}\ ] ] the ( eventually complex ) transition probability for a causal process is : to be in a state @xmath378 appears to be @xmath379 .
for a further short discussion see @xcite . ]
is complex for a causal process , then also the respective causal cross section will be complex . only if the underlying theory represented by a causal lagrangean or causal hamiltonian
is probability conserving , i.e. non - absorptive , @xmath365 and therefore also the causal cross section will be quasi - real , i.e. infinitesimally close to a real number .
hence , for selective ( so called `` inelastic '' ) causal processes in probability non - conserving theories ( e.g. open quantum systems ) the respective causal cross sections will develop a finite imaginary part . within particle physics
this new feature complements in a beautiful manner , what is well understood for a long time in theoretical optics , i.e. that the imaginary part of the refractive index of a material is related to its absorption coefficient .
similar arguments hold for anticausal processes and respective anticausal cross sections .
a temptative interpretation of complex cross sections was provided by t. berggren @xcite .
furthermore , he suggested to study _ experimentally observable _ effects to measurable lineshapes in inelastic scattering experiments due to complex cross sections .
inversely we suggest that a reality constraint on causal cross sections will yield important theoretical constraints on ( mass and coupling ) parameters of lagrangeans used to describe elastic scattering processes . ]
@xmath380 \left(\left<\!\left<\right.\right.\!\overline{\psi}_f\!\!\left.\left.\right|\right.\right)^t\ ! \left<\!\left<\psi_f\right|\right .
t [ \ , \exp(i \ , s_{int } ) - 1 \ , ] \left|\psi_i\right > } \nonumber \\ & = & ( -1)^{n_f(n_f-1)/2 } \,\times \nonumber \\ & & \times \left<\!\left<\right.\right.\!\overline{\psi}_f\!\!\left.\left.\right|\right .
\!\big ( t [ \ , \exp(- i \ ,
s_{int } ) - 1 \ , ] \big)^t \ !
\left|\right.\!\overline{\psi}_i\!\left.\right > \left<\!\left<\psi_f\right|\right .
t [ \ , \exp(i \ , s_{int } ) - 1 \ , ] \left|\psi_i\right > \ ; . \end{aligned}\ ] ]
in 1997 c.m . bender & s. boettcher @xcite ( see also m. znojil @xcite ) used the non - hermitian hamilton operator @xmath381 obtained from a harmonic oscillator shifted to a complex space point @xmath382 as an example to show that its spectrum @xmath383 can be indeed real due to the underlying `` pt - symmetry '' . noting that in this work there is seemingly not made any distinction between the real variable @xmath238 and the complex variables @xmath239 and @xmath240 appearing in the holomorphic representation , we want to illustrate here the explicit formalism for a respective shift of the ( anti)causal bosonic harmonic oscillator within ( a)cqt .
the hamilton operator of the `` unshifted '' ( anti)causal bosonic harmonic oscillator in holomorphic representation and in `` representation free '' form are given by : @xmath384 without changing its spectrum the hamilton operator is shifted from @xmath385 to by applying the following operator , which is denoted in the holomorphic representation as @xmath386 and in the representation free case as @xmath387 $ ] : @xmath388 & = & \exp\big ( i\,\alpha \ , \sqrt{\frac{m\,\omega}{2 } } \ ; ( a+c^+ ) + i\ , \beta^\ast \ , \sqrt{\frac{m^\ast\,\omega^\ast}{2 } } \ ; ( c+a^+ ) \big ) \ ; .\end{aligned}\ ] ] the respective inverse operators are given by @xmath389 and @xmath390\equiv u[-\alpha,-\beta^\ast]$ ] .
the `` shifted '' hamilton operator obtained by a standard equivalence transform @xmath391 and @xmath392= u[\alpha,\beta^\ast ] \ ; h \ ; u^{-1}[\alpha,\beta^\ast]$ ] is respectively : @xmath393 & = & h + m \ ,
\omega^2 \ , \alpha \left(\frac{i}{\sqrt{2\,m\ , \omega } } \ ; ( a - c^+)+\frac{\alpha}{2}\right ) \nonumber \\ & & + \ ,
m^\ast \ , \omega^{\ast \,2 } \ , \beta^\ast \left(\frac{i}{\sqrt{2\,m^\ast\ , \omega^\ast } } \ ; ( c - a^+ ) + \frac{\beta^\ast}{2}\right ) \ , .\end{aligned}\
] ] the `` shifted '' hamilton operator will be @xmath279-symmetric ( i.e. @xmath394 ) for @xmath395
. it will be hermitian/@xmath277-symmetric ( i.e. @xmath396 ) for @xmath397 .
bender s example `` @xmath398 '' is essentially obtained by @xmath399 with @xmath400 . for @xmath401
even being strictly @xmath279-symmetric only some part of its spectrum @xmath402 with @xmath222@xmath223 is real , namely if @xmath403 .
non - hermitian hamilton supersymmetric operators with pt - symmetry have been investigated already for quite some time ( see e.g. @xcite ! ) .
we want here to go one step further and consider the non - pt - symmetric hamilton operator of an ( anti)causal supersymmetric harmonic oscillator , being the sum of a causal bosonic and fermionic harmonic oscillator with equal _ complex _ frequency @xmath404 , and an anticausal bosonic and fermionic harmonic oscillator with equal _ complex _ frequency @xmath405 : @xmath406 + \frac{1}{2 } \ ; \omega_a^\ast \ , \{a^+ , c\ , \ } + \frac{1}{2 } \ ; \omega_a^\ast \ , [ \ , b^+ , d\ , ] \nonumber \\ & = & \omega_c \ , \big(\,c^+ a+\frac{1}{2}\ , \big ) + \omega_c \ , \big(\,d^+ b\,-\frac{1}{2 } \,\big ) + \omega_a^\ast \big(\,a^+ c+\frac{1}{2}\ , \big ) + \omega_a^\ast \big(\,b^+ d\,-\frac{1}{2 } \,\big ) \nonumber \\[1 mm ] & = & \omega_c \ ; ( c^+ a+ d^+ b ) + \omega_a^\ast \ , ( a^+ c+ b^+ d ) \ ; .\end{aligned}\ ] ] as usual in a supersymmetric system the positive and negative contributions to the bosonic and fermionic vacuum energy , respectively , cancel . the hamilton operator is easily diagonalized by the _
right eigenstates @xmath407 or _
normal _ left eigenstates @xmath408 given by @xmath409 ( with @xmath410 , @xmath411@xmath223 and @xmath412 , @xmath413 ) yielding the respective energy eigenvalues @xmath414 .
now we define supercharges @xmath415 and respective hermitian conjugate supercharges @xmath416 by : @xmath417 the supercharges @xmath418 and @xmath419 ( and @xmath420 and @xmath421 ) are nilpotent , yet _ not _ related by hermitian conjugation , i.e. : @xmath422 the hamilton operator can be expressed in terms of supercharges as follows : @xmath423 application of supercharges interrelates as usual distinct eigenstates of the hamilton operator belonging to the same eigenvalue . for _ normal _ right eigenvectors
we have : @xmath424 in order to compare this non - hermitian supersymmetric harmonic oscillator with its hermitian counterpart one has to decompose creation and annihilation operators of this oscillator into their hermitian ( `` shadow '' ) components . ]
the creation and annihilation operators of the hermitian shadow fields obey the following ( anti)commutation relations : @xmath425 } & { [ a_{{}_{(1)}},a^+_{{}_{(2 ) } } ] } \\ { [ a_{{}_{(2)}},a^+_{{}_{(1 ) } } ] } & { [ a_{{}_{(2)}},a^+_{{}_{(2 ) } } ] } \end{array}\right ) & = & \left ( \begin{array}{cc } { \{b_{{}_{(1)}},b^+_{{}_{(1)}}\ } } & { \{b_{{}_{(1)}},b^+_{{}_{(2)}}\ } } \\ { \{b_{{}_{(2)}},b^+_{{}_{(1)}}\ } } & { \{b_{{}_{(2)}},b^+_{{}_{(2)}}\ } } \end{array}\right ) \ ; = \ ; \left ( \begin{array}{cc } 1 & 0 \\ 0 & -1 \end{array}\right ) \ ; , \nonumber \\[1 mm ] \left ( \begin{array}{cc } { [ a_{{}_{(1)}},a_{{}_{(1 ) } } ] } & { [ a_{{}_{(1)}},a_{{}_{(2 ) } } ] } \\ { [ a_{{}_{(2)}},a_{{}_{(1 ) } } ] } & { [ a_{{}_{(2)}},a_{{}_{(2 ) } } ] } \end{array}\right ) & = & \left ( \begin{array}{cc } { [ a^+_{{}_{(1)}},a^+_{{}_{(1 ) } } ] } & { [ a^+_{{}_{(1)}},a^+_{{}_{(2 ) } } ] } \\ { [ a^+_{{}_{(2)}},a^+_{{}_{(1 ) } } ] } & { [ a^+_{{}_{(2)}},a^+_{{}_{(2 ) } } ] } \end{array}\right ) \;\;\ , = \ ; \left ( \begin{array}{cc } 0 & 0 \\ 0 & 0 \end{array}\right ) \ ; , \nonumber \\[1 mm ] \left ( \begin{array}{cc } { \{b_{{}_{(1)}},b_{{}_{(1)}}\ } } & { \{b_{{}_{(1)}},b_{{}_{(2)}}\ } } \\ { \{b_{{}_{(2)}},b_{{}_{(1)}}\ } } & { \{b_{{}_{(2)}},b_{{}_{(2)}}\ } } \end{array}\right ) & = & \left ( \begin{array}{cc } { \{b^+_{{}_{(1)}},b^+_{{}_{(1)}}\ } } & { \{b^+_{{}_{(1)}},b^+_{{}_{(2)}}\ } } \\ { \{b^+_{{}_{(2)}},b^+_{{}_{(1)}}\ } } & { \{b^+_{{}_{(2)}},b^+_{{}_{(2)}}\ } } \end{array}\right ) \ ; = \ ; \left ( \begin{array}{cc } 0 & 0 \\ 0 & 0 \end{array}\right ) .
\qquad\end{aligned}\ ] ] as usually fermionic and bosonic field operators commute . ]
the hamilton operator of the ( anti)causal supersymmetric harmonic oscillator in terms of creation and annihilation operators of the hermitian shadow fields is given by : @xmath426 this expression reflects most transparently the tricky interplay of normal / abnormal bosonic / fermionic shadow fields , when providing us with supersymmetry even for arbitrary complex - valued oscillator frequencies @xmath404 and @xmath405 .
the pt - symmetric hamilton operator is obtained by setting @xmath427 yielding @xmath428re@xmath429 $ ] and @xmath430im@xmath429 $ ] . the limit im@xmath429=0 $ ] was considered recently in @xcite , where the abnormal fermions @xmath431 and @xmath432 are called _ phermions _ rather than @xmath433 , as he believes that @xmath434 should be a positive operator .
instead , we follow here @xcite who understands as discussed in the introduction that hermitian conjugation can not be defined without specifying the metric ! ] .
is is common belief that the theory of strong interaction is described by qcd @xcite no renormalizable field theory without non - abelian gauge fields can be asymptotically free @xmath7''_. the absence of asymptotic freedom in qed had been proved earlier by m. gell - mann & f. low @xcite and `` shown '' for ( pseudo)scalar - fermion theories involving one coupling constant by a. zee @xcite . ] , i.e. gluons and ( anti)quarks . the clear experimental evidence allowing such a belief ( in particular with respect an eventual existence of gluons ) is still lacking
. on the theoretical side there seem to be now even strong arguments against the existence of gluons due to calculations by r. alkofer et al .
@xcite ( see also @xcite ! ) who observe that _ spectral functions of gluons violate positity _ , while the spectral functions of quarks do _ not _ ! at the first sight there does not seem to be a substitute for gluons at hand , which allows to achieve asymptotic freedom without getting in conflict with the statements by coleman , gross , and zee , as asymptotically free theories with e.g. scalars and fermions are required to be rather more @xcite than less @xcite _ exotic_. mysteriously , the quark - level linear sigma model @xcite ( qll@xmath435 m ) has been a rather successful theory to describe various experimental facts involving hadronic physics at low and intermediate energy ( see e.g. @xcite and references therein ! ) .
guided by this observation we `` mapped '' @xcite in 2002 by a simplistic argument the lagrangean of qcd ( strong coupling @xmath40 ! ) into a lagrangean of a qll@xmath435 m which is supposed to describe quark - quark scattering equally well _ at high energies_. in _ unitary gauge _ we obtained the following qll@xmath435 m lagrangean : @xmath436 \big ) \
, q_- \nonumber \\ & & + \ , \frac{1}{2 }
\,\mbox{tr } \big [ \big ( \partial^\mu s - i\,g \ , \sqrt{\frac{n_f}{2\ , n_c } } \big(s_v [ v^\mu , s ] - i\ ,
s_s s_p s_y \{y^\mu , p\ } \big ) \big)^2 \big ] - \frac{1}{2 } \ , m^2 \,\mbox{tr } \big [ s^2
\big ] \nonumber \\ & & + \ ,
\frac{1}{2 } \,\mbox{tr } \big [ \big ( \partial^\mu p - i\,g \ , \sqrt{\frac{n_f}{2\ , n_c } } \big(s_v [ v^\mu , p ] + i\ , s_s s_p s_y \{y^\mu ,
s\ } \big ) \big)^2 \big ] - \frac{1}{2 } \ , m^2 \,\mbox{tr } \big [ p^2
\big ] \nonumber \\ & & - \ , \frac{1}{8 } \ ; \mbox{tr } \big [ ( v^{\mu\nu}_+)^2 \big ] + \frac{1}{4 } \ ; m^2 \ , \mbox{tr } \big [ \big(v^\mu_+\big)^2 \big ] \ ; - \ , \frac{1}{8 } \ ; \mbox{tr } \big [ ( v^{\mu\nu}_-)^2 \big ] + \frac{1}{4 } \ ; m^2 \ , \mbox{tr } \big [ \big(v^\mu_-\big)^2 \big ] \nonumber \\ & & -\ , \frac{\lambda}{2 } \ ; \mbox{tr } \big [ \big((s+ i\ , p)(s- i\ , p ) \big)^2 \big ] + \mbox{gauge fixing terms / ghosts } + \mbox{h.c . } , \label{linsig1}\end{aligned}\ ] ] with @xmath437 , @xmath438 , @xmath439 , @xmath440 , @xmath441 , @xmath442/(-i\ , g\,\sqrt{n_f/(2\,n_c)}\,)$ ] , @xmath443 , tr@xmath444``flavour - trace '' , and @xmath445 , @xmath82 , @xmath446 , @xmath447 are scalar , pseudoscalar , vector and axial vector @xmath448 meson field matrices in flavour space , respectively , while @xmath449 .
this unexpected result is a _ non - hermitian _
qll@xmath435 m lagrangean , which is _ asymptotically free _ due to a purely _ imaginary _ yukawa coupling , the pt - symmetry of which suggests in correspondence to c.m .
bender s `` physical '' @xmath450-theory @xcite a _ real spectrum _ of the respective hamilton operator ! additionally we observe that all vertices are accompanied by weightfactors completely in accordance with the @xmath451 expansion @xcite , while the @xmath32 and @xmath452 mesons couple to the @xmath445 and @xmath82 mesons in a similar manner as in an 1-loop effective action of the gauged linear sigma model in @xcite or the extended chiral quark model in @xcite obtained by a tedious bosonization from the qcd lagrangean .
furthermore we observe that the obtained lagrangean provides strictly sakurai s _ vector meson universality _ required in @xcite by renormalization constraints .
hence , the obtained lagrangean is a theoretically very admissable lagrangean for the description of strong interactions , which may be used to replace the lagrangean of qcd at high energies . after including photons
the lagrangean can be even spontaneously broken down to low energies by allowing the scalar mesons to develop a complex - valued vacuum expectation value , which will provide the quarks with complex masses explaining their invisibility at low energies . based on experience with the dynamical generation @xcite of linear sigma models
we finally feel the need to perform a slight important modification of the value of the quartic coupling @xmath453 having resulted from our simplistic `` mapping '' for a reason , which we will illustrate here on the basis of a much simpler `` wess - zumino - like '' linear sigma model lagrangean involving only one scalar @xmath435 , one pseudoscalar @xmath454 , and @xmath455 fermions @xmath456 : @xmath457 the quadratically divergent tadpole appearing in this theory can be absorbed by a suitable counterterm . as in the supersymmetric wess - zumino model @xcite there can be extracted a relation between the yukawa - coupling @xmath40 and the quartic coupling @xmath453 by requiring an exact cancellation of quadratic divergencies on the level of @xmath435- and @xmath454-selfenergies .
a cancellation of quadratic divergencies up to two loops yields the important relation @xmath458 .
the well known one loop relation @xcite @xmath459 being obtained by skipping the term @xmath460 suggests , that a linear sigma model with purely _ imaginary _ yukawa coupling @xmath40 should be accompanied up to one loop by a _ negative _
quartic coupling @xmath453 .
this explains partially the intimate relation between c.m .
bender s @xcite `` physical '' @xmath450-theory and k. symanzik s & c.m .
bender s @xcite(see g. t hooft @xcite ! ) `` physical '' @xmath461-theory . moreover suggests the ( non - perturbative ) two - loop relation between @xmath40 and @xmath453 that at least one of these couplings have to be complex - valued for @xmath453 finite .- theory possesses a _
non - trivial _ phase when treated non - perturbatively @xcite . ] a corresponding non - perturbative two - loop relation between @xmath453 and @xmath40 for the lagrangean eq .
( [ linsig1 ] ) is in preparation .
recall , that it is ( a)cqt allowing us to work _ consistently _ in the presence of _ complex _ masses and couplings ! due to these our observations we suggest to look for supersymmetric partners of standard model particles _ not at high energies _ , where we do nt find them , yet _ in the complex energy plane _
, where they have escaped our observation by now ! similarly also the scalar higgs - boson may hide from present experiments not only due to adler - like zeros of reaction amplitudes like in hadronic physics @xcite , yet also due to the feature of being a broad resonance whose pole is situated in correspondence to a complex - valued scalar vacuum condensate somewhere in the complex energy plane .
we are fully aware that this introductory presentation and mentioned literature is far from being exhaustive .
many important and interesting issues and aspects of the comprehensive formalism have been omitted to keep the present size of the manusscript under control , yet may be certainly subject to and make part of a more detailed forthcoming outline . as an example we mention that the sketched complex probability concept going hand in hand with the presence of _ non - hermtitian density matrices _ will have important implications not only to modern information theory and statistical physics , yet might also contribute e.g. to the resolution of certain quantum - theoretic paradoxa @xcite .
even though it should have become clear that renormalization in ( a)cqt is performed even for fields with finite complex mass in complete analogy to traditional hqt , the ( observable ) consequences of imaginary parts of mass and coupling parameters to resulting finite quantities deserve certainly a more thorough investigation .
the present discussion on pt - symmetric qm being a cousin of the lee - model is reminding us again that we still have not made use of all possibilities which qt is offering to us .
the existence of physically acceptable non - hermitian hamilton operators with a real spectrum opens interesting new fields in physics which have not been explored yet and which seem to allow simple solutions to tradionally complicated problems like e.g. asymptotic freedom and confinement in strong interactions having been unnecessarily constraint by the influential and impressive `` proofs '' e.g. by a. zee , s.r .
coleman , d.r .
gross , f. wilczek performed under too restraining assumptions .
the formalism of ( a)cqt presented here may be used conveniently to approach such problems and even problems with complex spectra without running into quantum - theoretic inconsistencies . as in the formalism of _ open quantum systems _
( oqs ) @xcite , where one needs to close the system by a `` heat bath '' ( `` reservoir '' , `` environment '' ) before quantizing , we had to `` close '' a _ causal _ system in ( a)cqt by adding the respective hermitian conjugate _
anticausal _ system . to make ( a)cqt causal , poincar invariant , analytic and local we had to apply the postulate of _ non - interaction _ of the _ causal _ and the _ anticausal _ system .
such an uncorrelated situation between system and heat bath would be called in the formalism of oqs _ complete positivity _
it is quite attractive to draw more correspondences between the formalisms of oqs and ( a)cqt . as in thermal field theory ( tft ) @xcite , but for different reasons and already at zero temperature , we observe also in ( a)cqt a _ doubling of degrees of freedom _ due to the underlying indefinite metric . differently from tft or the respective real time formulation ( rtf ) @xcite the position of the vacuum at zero energy and the new antiparticle concept in ( a)cqt does not allow to excite particle - antiparticle pairs by distribution functions having an overlap with negative energies due to finite temperature .
hence , in ( a)cqt the excitation of particle - antiparticle pairs ( even with complex masses ) at finite temperature has to be resolved by distribution functions making use of positive energies only , requiring a modification of tft and its rtf , which will simultaneously lead to the concept of a _ causal heat bath_. a similar argument rules out gribov s theory for confinement @xcite within ( a)cqt .
our considerations showed that the way along which scientific progress proceeds is far from geodesic .
dirac tried to explain with his hole theory for electrons _ protons _
@xcite , while experimentalists disovered shortly afterwards @xcite the _ positron _ , i.e. the electron s antiparticle .
since then the positron has served as the example to explain the existence of fermionic antiparticles on the basis of the negative energy states of the dirac equation , while dirac had parallelly developed for bosons the orthogonal idea , that negative energy states are responsible for particle absorption .
now , after investigating the implications of causality to the antiparticle concept even for complex mass particles we had to notice , that even for fermions negative energy states are responsible for annihilating positive energy particles and their positive energy antiparticles , and not for their existence .
surprisingly , in the resulting antiparticle concept fermions and anti - fermions being isospin partner show up to have without harm and contrary to the `` proof '' of v.b .
berestetskii @xcite the _ same _ intrinsic parity , like traditionally bosons and anti - bosons !
even though the original ingenious construction of dirac is nearly replaced , the ideas of dirac persist through the reveiled existence of antiparticles and a lot of important insights in the theoretical formalism of qt and particle physics .
unfortunately it seems that very influential still active authors of important ideas related to indefinite metric , negative norm states and complex ghosts are still denying the possibility of a meaningful use of complex poles in qt . stipulating that such negative norm states ( negative hilbert space metric ! ) can not be present in any physical theory
will eliminate the possibility of obtaining an inverted spin - statistics connection @xmath462 '' _ and shifting some of the discussion on non - hermtitian fields to an appendix , even if we have demonstrated above , that causal states being non - hermitian necessarily consist of underlying normal _ and _ abnormal shadow states ; n. nakanishi still defending his discouraging conclusions with respect to his complex - ghost relativistic field theory states : _
`` @xmath7 since the appearance of negative probability is the greatest problem in the indefinite metric theory and since , historically , many great physicists proposed wrong resolution of it , i can not believe that you could resolve this problem correctly .
but since it seems to me that you understand the indefinite - metric theory quite differently from the standard one , it is very difficult to discuss this problem with you .
@xmath7 '' _
( n. nakanishi , private communication , 31.10.2003 ) .
] anyway young scientists and physics students should be obliged to read their most important and `` thrilling '' papers ! this work has been supported by the _ fundao para a cincia e a tecnologia _ ( fct ) of the _ ministrio da cincia e da tecnologia ( e do ensinio superior ) _ of portugal , under grants no .
praxis xxi / bpd/20186/99 , sfrh / bdp/9480/2002 , and pocti / fnu/49555/2002 .
w. heisenberg : _ indefinite metric in state space _ , p. 129
ff in _ aspects of quantum theory _
( dedicated to p.a.m .
dirac ) , eds .
a. salam , e.p .
wigner , 1972 by cambridge university press ; isbn 0 521 08600 0 .
dirac : proc .
a * 126 * ( 1930 ) 360 .
dirac : commun .
dublin inst . adv .
studies , series a , no . 1 ( 1943 ) 1 (= p. 1077
ff in @xcite ) .
w. pauli : rev .
* 15 * ( 1943 ) 175 ( also : rev .
* 13 * ( 1941 ) 203 ) .
w. mckenheim et al .
* 133 * ( 1986 ) 337 .
f. henry - couannier : gr - qc/0404110
. s. weinberg : _ the quantum theory of fields .
vol . 1 : foundations _ , published by cambridge university press ; 1995 by s. weinberg ; isbn 0 521 55001 7 .
w. pauli , v.f .
weisskopf : helv .
acta * 7 * ( 1934 ) 709 .
furry , j.r .
oppenheimer : phys . rev .
* 45 * ( 1934 ) 245 .
vol . 1 of _ at the frontiers of particle physics
handbook of qcd _
( boris ioffe festschrift ) , ed .
m. shifman ; 2001 by world scientific publishing co. pte . ltd .
; isbn 981 - 02 - 4967 - 5 .
berestetskii , e.m .
lifshitz , l.p .
pitaevskii : _ quantum electrodynamics _ ( vol .
4 , 2nd edition ( 1982 ) in l.d .
landau , e. m. lifshitz : _ course of theoretical physics _ ) ( first edition published 1971 as _ relativistic quantum theory _ ; second edition translated from _
_ published by izdatelstvo `` nauka '' , moskow , 1980 ) , 1980 by pergamon press plc ; isbn 0 - 08 - 026503 - 0 .
f. kleefeld : hep - th/0312027 .
f. kleefeld : proc . of inst . of mathematics of nas of ukraine ,
* 50 * , part 3 ( 2004 ) 1367 [ hep - th/0310204 ] .
f. kleefeld : aip conf .
* 660 * ( 2003 ) 325 .
f. kleefeld : few - body systems suppl .
* 15 * ( 2003 ) 201 .
f. kleefeld , e. van beveren , g. rupp : nucl .
a * 694 * ( 2001 ) 470 .
f. kleefeld : acta phys .
b * 30 * ( 1999 ) 981 .
f. kleefeld : _ consistent effective description of nucleonic resonances in an unitary relativistic field - theoretic way _ , in proc .
xiv ishepp 98 ( 17 - 22 august , 1998 , dubna ) , eds .
a.m. baldin , v.v .
burov , jinr , dubna , 2000 , part 1 , 69 - 77 [ nucl - th/9811032 ] .
f. kleefeld : doctoral thesis ( university of erlangen - nrnberg , germany , 1999 ) .
b. machet , v.a .
novikov , m.i .
vysotsky : hep - ph/0407268 .
lee : phys . rev .
* 95 * ( 1954 ) 1329 .
g. klln , w. pauli : kong .
. vid .
sel . mat .
fys . med .
* 30 * ( 1955 ) 407 .
w. pauli : nuovo cim . suppl .
* 4 * ( 1956 ) 70 (= proc . of `` int .
conf . on elementary particles '' and
`` xli national congress of physics '' , pisa , 12 - 18.6.1955 ) . w. glaser , g. klln : nucl .
* 2 * ( 1956 ) 706 .
w. heisenberg : nucl .
phys . * 4 * ( 1957 ) 532 .
w. heisenberg : _ einfhrung in die theorie der elementarteilchen _ , lecture given at university of munich in 1961 , lecture notes prepared by h. rechenberg , k. lagally ( ictp trieste library number 539.12 h473 ) .
n. nakanishi : prog .
* 19 * ( 1958 ) 607 .
r. ascoli , e. minardi : nuovo cim .
* 14 * ( 1959 ) 1254 .
b. ferretti : nuovo cim .
* 12 * ( 1959 ) 393 .
nagy : nuovo cim . * 17 * ( 1960 ) 385 .
nagy : nuovo cim
. suppl .
* 17 * ( 1960 ) 92 .
sudarshan : phys .
* 123 * ( 1961 ) 2183 .
schnitzer , e.c.g .
sudarshan : phys .
* 123 * ( 1961 ) 2193 .
vaughn : nuovo cim . * 40 a * ( 1965 ) 803 .
nagy : _ state vector spaces with indefinite metric in quantum field theory _
, akadmiai kiad , budapest 1966 ( russian translation by `` mir '' , moscow , 1966 ) .
lee , g.c .
wick : nucl .
b * 10 * ( 1969 ) 1 .
n. nakanishi : prog .
* 51 * ( 1972 ) 1 .
d. hittner : nuovo cim . * 15 a * ( 1973 ) 401 . f. casagrande , l.a .
lugiato : nuovo cim . * 17 a * ( 1973 ) 464 .
bender and c. nash : phys .
d * 10 * ( 1974 ) 1753 .
t. keszthelyi , k.l .
nagy : acta phys
. hung . *
37 * ( 1974 ) 259 .
i. rabuffo , g. vitiello : nuovo cim .
* 44 a * ( 1978 ) 401 .
friedrichs : commun .
pure appl . math .
* 1 * ( 1948 ) 361 .
streater : _ lost causes in theoretical physics _ ( see _ non - self - adjoint observables _ ! ) , url : http:/@xmath463/www.mth.kcl.ac.uk/@xmath464streater / lostcauses.html t.d .
lee , g.c .
wick : phys .
d * 2 * ( 1970 ) 1033 .
a. mostafazadeh : quant - ph/0404025 .
w. pauli : _ the indefinite metric with complex roots _ , p. 127 in proc . of the 1958 annual int .
conf . on high energy physics at cern , geneva , 30.6 - 5.7.1958 , ed . b. ferretti , 1958 by cern .
pandit : nuovo cim . * 11 * ( 1959 ) 157 .
drr , e. rudolph : nuovo cim .
a * 62 * ( 1969 ) 411 .
a. mostafazadeh : czech .
j. phys .
* 53 * ( 2003 ) 1079 [ quant - ph/0308028 ] .
r. ascoli , e. minardi : nuovo cim .
* 8 * ( 1958 ) 951 .
lee : _ relativistic complex pole model with indefinite metric _ , p. 260 in `` quanta '' , [email protected].@xmath465freund,@xmath465c.j.@xmath465goebel , y.@xmath465nambu,@xmath465@xmath4651970,@xmath465the@xmath465univ.@xmath465of@xmath465chicago .
c. puntmann : int .
j. theor .
* 36 * ( 1997 ) 2295 [ quant - ph/9705013 ] .
a. bohm , m. loewe , p. patuleanu , c. puntmann : quant - ph/9705010 .
lyapunov ( 1892 ) : _ problme gnral de stabilit du mouvement _ ( french translation , 1949 ) , ann . of math .
studies * 17 * , princeton univ . press , princeton .
j. okoowicz , m. poszajczak , i. rotter : phys .
* 374 * ( 2004 ) 271 .
heiss , m. mller , i. rotter : phys .
e * 58 * ( 1998 ) 2894 ; w.d .
heiss : quant - ph/9901023 ; e. hernndez et al . : rev .
* 49 ( s4 ) * ( 2003 ) 17 ; e. hernndez , a. juregui , a. mondragn : phys . rev . a * 67 * ( 2003 ) 022721 ; a. mondragon , e. hernandez , quant - ph/9710018 ; c.m
. bender , t.t .
wu : phys .
* 184 * ( 1969 ) 1231 , phys .
d * 7 * ( 1973 ) 1620 ; c.m .
bender , h.j .
happ , b. svetitsky : phys .
d * 9 * ( 1974 ) 2324 ; c. dembowski et al .
: nlin.cd/0402015 ; c. dembowski : doctoral thesis ( technical university of darmstadt , germany , 2003 ) ; c. dembowski et al . : phys . rev .
* 86 * ( 2001 ) 787 ; u. gnther , f. stefani , g. gerbeth : math - ph/0407015 .
w. pauli , f. villars : rev .
* 21 * ( 1949 ) 434 .
gupta : proc .
. lond . a * 63 * ( 1950 ) 681
. k. bleuler : helv .
acta * 23 * ( 1950 ) 567 .
gupta : can .
j. phys .
* 35 * ( 1957 ) 961 .
r. ascoli , e. minardi : nucl .
* 9 * ( 1958 ) 242 .
g. konisi , t. ogimoto : prog .
phys . * 20 * ( 1958 ) 868 .
g. konisi , t. ogimoto : prog .
* 21 * ( 1959 ) 727 .
s. schlieder : z. naturforsch .
, teil a , * 15 * ( 1961 ) 448 . b. gazdy : phys .
a * 78 * ( 1980 ) 309 .
pontrjagin : _ hermitian operators in spaces with indefinite metric _ , izv .
nauk sssr ser . mat .
* 8 * ( 1944 ) 243 .
r. nevanlinna : ann .
a * i. * ( 1952 ) 108,113,115 ; ann .
a * i. * ( 1954 ) 163 ; ann .
a * i. * ( 1956 ) 222 .
louhivaara : ann .
a * i. * ( 1958 ) 252 .
e. scheibe : ann .
sci . fenn .
a * i. * ( 1960 ) 294 .
h. langer : math .
* 146 * ( 1962 ) 60 .
dritschel , j. rovnyak : _ operators on indefinite inner product spaces _ , lectures on operator theory and its applications , fields institute monographs , vol . 3 , pp . 141 - 232 , amer . math . soc . ,
providence , ri , 1996 .
l. jakobczyk : annals phys .
* 161 * ( 1985 ) 314 .
l. jakobczyk : rept .
* 22 * ( 1985 ) 71 .
broadbridge : j. phys .
a * 16 * ( 1983 ) 3271 .
o. civitarese , m. gadella : phys .
* 396 * ( 2004 ) 41 ; a. bohm , m. loewe , b. van de ven : fortsch .
phys . * 51 * ( 2003 ) 551 [ quant - ph/0212130 ] ; a. bohm , h. kaldass , s. wickramasekara : fortsch .
phys . * 51 * ( 2003 ) 569 [ hep - th/0212280 ] ; fortsch .
* 51 * ( 2003 ) 604 [ hep - th/0212282 ] .
n. moiseyev : phys .
rept . * 302 * ( 1998 ) 211 .
g. hagen , m. hjorth - jensen , j.s .
vaagen : nucl - th/0303039 ; n. michel , w. nazarewicz , m. poszajczak , j. okoowicz : phys .
c * 67 * ( 2003 ) 054311 [ nucl - th/0302060 ] ; d. espriu , j. manzano , p. talavera : phys .
d * 66 * ( 2002 ) 076002 ; b.a .
kniehl , a. sirlin : phys .
b * 530 * ( 2002 ) 129 ; v.i .
kukulin , v.m .
krasnopolsky , j. horek : _ theory of resonances : principles and applications _ , czechoslovak academy of sciences , prague , published in co - edition with kluwer academic publishers , dordrecht , the netherlands , 1989 by the authors ; l.s .
ferreira , e. maglione : chaos , solitons & fractals * 12 * ( 2001 ) 2697 ; m. baldo , l.s .
ferreira , l. streit : nucl .
a * 467 * ( 1987 ) 44 ; h.c .
baker : phys .
* 50 * ( 1983 ) 1579 ; phys .
a * 30 * ( 1984 ) 773 ; h.c .
baker , r.l .
singleton , jr .
a * 42 * ( 1990 ) 10 ; l. fonda , g.c .
ghirardi , a. rimini : rept .
* 41 * ( 1978 ) 587 ; c.a .
nicolaides , d.r .
beck : int .
j. quant .
* 14 * ( 1978 ) 457 ; y. komninos , c.a .
nicolaides : chem .
* 78 * ( 1981 ) 347 ; c.a .
nicolaides et al .
j. quant .
* s15 * ( 1981 ) 355 ; c.a .
nicolaides , s. themelis : phys .
a * 45 * ( 1992 ) 349 ; c.a .
nicolaides : int .
j. quant .
* 89 * ( 2002 ) 94 ; g. garca - caldern , a. rubio : nucl .
a * 258 * ( 1986 ) 560 ; g. garca - caldern : nucl .
a * 261 * ( 1976 ) 130 ; g. garca - caldern , r. peierls : nucl .
a * 265 * ( 1976 ) 443 ; t. berggren : nucl .
a * 109 * ( 1968 ) 265 ; phys .
b * 37 * ( 1971 ) 240 ; nucl .
a * 389 * ( 1982 ) 261 ; u. fano : phys .
* 124 * ( 1961 ) 1866 ; g. gamow : z. phys .
* 51 * ( 1928 ) 204 ; nature * 122 * ( 1928 ) 805 ; _ constitution of atomic nuclei and radioactivity _ , oxford : clarendon press , london , 1931 ; _ resonances models and phenomena _ , proceedings of a workshop held at the centre for interdisciplinary research , bielefeld university , germany , 9.-14.4.1984 , eds .
s. albeverio , l.s .
ferreira , l. streit , published by springer - verlag ( berlin ) in lect .
notes phys .
* 211 * ( 1984 ) 1 - 359 .
drr , e. seiler : nuovo cim .
a * 66 * ( 1970 ) 734 .
m. demuth : math .
nachr . * 102 * ( 1981 ) 107 .
lee , g.c . wick : nucl .
b * 9 * ( 1969 ) 209 .
a.m. gleeson , r.j .
moore , h. rechenberg , e.c.g .
sudarshan : phys .
d * 4 * ( 1971 ) 2242 .
sudarshan : aip conf .
* 16 * ( 1973 ) 115 .
arons , e.c.g .
sudarshan : phys .
* 173 * ( 1968 ) 1622 .
j. dhar , e.c.g .
sudarshan : phys .
* 174 * ( 1968 ) 1808 .
wigner : _ invariant quantum mechanical equations of motion _ , p. 59 in `` theoretical physics : lectures presented at the iaea seminar on theoretical physics , 16.7.-25.8.1962 , palazzino miramare , trieste '' , directed by a. salam , published by the international atomic energy agency , vienna , 1963 ; sti / pub/61 . o.w .
greenberg : hep - th/0405211 .
n. nakanishi : phys .
d * 3 * ( 1971 ) 811 .
lee , g.c .
wick : phys .
d * 3 * ( 1971 ) 1046 .
t. berggren : phys .
b * 33 * ( 1970 ) 547 ; nucl .
a * 169 * ( 1971 ) 353 ; phys .
b * 73 * ( 1978 ) 389 .
workshop on _ pseudo - hermitian hamiltonians in quantum physics _
, 16.-17.6.2003 , villa lanna , prague , czech republic , czech .
j. phys .
* 54 * ( 2004 ) 1 - 156 .
bender , s. boettcher : phys .
lett . 80 ( 1998 ) 5243 [ physics/9712001 ] .
m. znojil : rend .
circ . mat .
palermo , serie ii , suppl .
* 72 * ( 2004 ) 211 [ math - ph/0104012 ] .
m. znojil : czech .
j. phys .
* 54 * ( 2004 ) 151 [ quant - ph/0309100 ] .
e. caliceti , s. graffi , m. maioli : commun .
* 75 * ( 1980 ) 51 .
e. delabaere , f. pham : phys .
lett . a 250 ( 1998 ) 25 , 29 . c. bender , s. boettcher , p. meisinger : j. math .
* 40 * ( 1999 ) 2201 .
m. robnik , m.v .
berry : j. phys . a : math .
gen . 19 ( 1986 ) 669 .
v. buslaev , v. grecci : j. phys . a : math .
gen . 26 ( 1993 ) 5541 .
fernndez , r. guardiola , j. ros , m. znojil : j. phys .
a : math . gen . 31 ( 1998 ) 10105
; see also quant - ph/9812026 .
g. scolarici : czech .
j. phys .
* 54 * ( 2004 ) 119 ; a. blasi , g. scolarici , l. solombrino : quant - ph/0407158 , quant - ph/0310106 .
s. weigert : czech .
j. phys .
* 54 * ( 2004 ) 147 [ quant - ph/0310050 ] .
s. weigert : j. optics b * 5 * ( 2003 ) s416 [ quant - ph/0209054 ] .
g. lvai : czech .
j. phys .
* 54 * ( 2004 ) 77 .
p. dorey , c. dunning , r. tateo : j. phys .
a * 34 * ( 2001 ) l391 [ hep - th/0104119 ] .
g. lvai , m. znojil : j. phys .
a * 33 * ( 2000 ) 7165 .
z. ahmed : quant - ph/0302141 .
bender , p.n .
meisinger , q. wang : j. phys .
a * 36 * ( 2003 ) 6791 ; q. wang : czech .
j. phys .
* 54 * ( 2004 ) 143 .
bender , d.c .
brody , h.f .
jones : phys .
* 89 * ( 2002 ) 270401 [ erratum - ibid .
* 92 * ( 2004 ) 119902 ] [ quant - ph/0208076 ] .
bender : czech .
j. phys .
* 54 * ( 2004 ) 13 .
p. dorey , c. dunning , r. tateo : czech .
j. phys .
* 54 * ( 2004 ) 35 [ hep - th/0309209 ] .
e. caliceti , s. graffi , j. sjoestrand : math - ph/0407052 .
bender , h.f .
jones : hep - th/0405113 ; c.m .
bender , d.c .
brody , h.f .
jones : phys .
d * 70 * ( 2004 ) 025001 ( see also hep - th/0402011 ! ) .
e. caliceti , f. cannata , m. znojil , a. ventura : math - ph/0406031 . m. znojil : hep - th/0012002 .
m. znojil , f. cannata , b. bagchi , r. roychoudhury : phys .
b * 483 * ( 2000 ) 284 .
b. bagchi , c. quesne : mod .
lett . a * 16 * ( 2001 ) 2449 [ quant - ph/0110133 ] .
n. nakanishi : prog . theor
. phys .
* 19 * ( 1958 ) 607 .
bollini , t. escobar , m.c .
rocca : int .
j. theor .
* 38 * ( 1999 ) 2315 ; + c.g .
bollini , m.c .
rocca : int .
j. theor .
* 43 * ( 2004 ) 59 ; hep - th/0312214 ; i.m .
guelfand , n.y .
vilenkin ( and translator g. rideau ) : _ applications de lanalyse harmonique _ (= part 4 of i.m .
guelfand et al . : _ les distributions _ ) , dunod , paris , 1967 .
bollini , m.c .
rocca : int .
j. theor .
* 37 * ( 1998 ) 2877 .
greenberg : hep - ph/0309309 .
streater , a.s .
wightman : _ pct , spin and statistics , and all that _ , w.a .
benjamin , new york , 1964 ; isbn 0 - 8053 - 9252 - 1 .
jun - chen su : hep - th/9805195 , 9805192 , 9805193 , 9805194 ; nuovo cim . * 117 b * ( 2002 ) 203 .
a. pais : _ max born and the statistical interpretation of quantum mechanics _ , presented to annual mtg .
of optical soc . of america , tucson ,
arizona , oct . 21 , 1982 , on occasion of the centenary of born s birth [ kek lib . preprint 198301341 ]
. c.g .
bollini , o. civitarese , a.l .
de paoli , m.c .
rocca : phys .
b * 382 * ( 1996 ) 205 ; w.j .
romo : nucl .
a * 116 * ( 1968 ) 617 ; j. math .
* 21 * ( 1980 ) 311 ; t. berggren : phys .
b * 373 * ( 1996 ) 1 .
a. kossakowski : open sys . &
information dyn .
* 9 * ( 2002 ) 1 .
y. habara , h. b. nielsen , m. ninomiya : hep - th/0312302 .
ablowitz , a.s .
fokas : _ complex variables : introduction and applications _ ( 2nd edition ) ; 2003 by cambridge university press ; isbn 0 521 53429 1 .
faddeev , a.a .
slavnov : _ gauge fields .
introduction to quantum theory _ , front .
* 83 * ( 1990 ) 1 .
d. chruciski : open sys . & information dyn .
* 9 * ( 2002 ) 207 [ math - ph/0206009 ] .
glauber : phys .
* 131 * ( 1963 ) 2766 .
stapp : nuovo cim . a * 23 * ( 1974 ) 357 ; phys . rev .
d * 13 * ( 1976 ) 947 .
m. znojil : phys .
a * 259 * ( 1999 ) 220 [ quant - ph/9905020 ] .
m. znojil : hep - th/0209062 .
m. znojil : j. phys .
a * 35 * ( 2002 ) 2341 [ hep - th/0201056 ] .
gross , f. wilczek : phys .
d * 8 * ( 1973 ) 3633 ; phys .
d * 9 * ( 1974 ) 980 .
gross , f. wilczek : phys .
lett . * 30 * ( 1973 ) 1343 ; h.d .
politzer : phys .
lett . * 30 * ( 1973 ) 1346 ; g. t hooft , unpublished ( see also : g. t hooft : hep - th/9812203 ; d.j .
gross : _ the discovery of asymptotic freedom and the emergence of qcd _ , p. 89
ff in @xcite ! ) .
coleman , d.j .
gross : phys .
* 31 * ( 1973 ) 851 .
m. gell - mann , f.e .
low : phys .
* 95 * ( 1954 ) 1300 .
a. zee : phys .
d * 7 * ( 1973 ) 3630 .
r. alkofer , w. detmold , c. s. fischer , p. maris : hep - ph/0309078 ; hep - ph/0309077 .
a. cucchieri , t. mendes , a. r. taurines : hep - lat/0406020 .
vaughn : z. phys .
c * 13 * ( 1982 ) 139 ; see also p. 244
f in g. t hooft : _ under the spell of the gauge principle _ ,
1994 by world scientific publishing co. pte . ltd .
, isbn 981 - 02 - 1308 - 5 .
hari dass , v. soni : phys .
d * 65 * ( 2002 ) 095005 .
r. delbourgo , m.d .
scadron : mod .
lett . a * 10 * ( 1995 ) 251 ; int .
j. mod .
a * 13 * ( 1998 ) 657 .
f. kleefeld , e. van beveren , g. rupp , m.d .
scadron : phys .
d * 66 * ( 2002 ) 034007 ; e. van beveren , f. kleefeld , g. rupp , m.d .
scadron : mod .
a * 17 * ( 2002 ) 1673 ; m. d. scadron , f. kleefeld , g. rupp , e. van beveren : nucl .
a * 724 * ( 2003 ) 391 ; aip conf .
* 660 * ( 2003 ) 311 ; m.d .
scadron , g. rupp , f. kleefeld , e. van beveren : phys .
d * 69 * ( 2004 ) 014010 [ erratum - ibid .
d * 69 * ( 2004 ) 059901 ] ; j. meyer , k. schwenzer , h.j .
pirner , a. deandrea : phys .
b * 526 * ( 2002 ) 79 ; h.j .
pirner : hep - ph/0209221 . g. t hooft : nucl .
b * 75 * ( 1974 ) 461 ( see also chapter 8 in @xcite ! ) a.a .
andrianov , d. espriu : jhep * 10 * ( 1999 ) 022 [ hep - ph/9906459 ] ; j. bijnens : phys .
* 265 * ( 1996 ) 369 [ hep - ph/9502335 ] .
d. djukanovic , m.r .
schindler , j. gegelia , g. japaridze , s. scherer : hep - ph/0407239 . n.g .
deshpande , r.j .
johnson , e. ma : phys .
d * 29 * ( 1984 ) 2851 .
j. wess , b. zumino : phys .
b * 49 * ( 1974 ) 52 ; see also p. 323
ff in a. das : _ finite temperature field theory _ , 1997 by world scientific publishing co. pte . ltd .
, isbn 9810228562 .
bender , k.a .
milton , v.m .
savage : phys .
d * 62 * ( 2000 ) 085001 ; k. symanzik : lett .
nuovo cim .
* 6 * ( 1973 ) 77 (= proc .
marseille conf .
, 19.-23.6.1972 , ed .
korthals altes ) .
p. cea , m. consoli , l. cosmai : hep - lat/0407024 ; m. consoli , p.m. stevenson : int .
j. mod .
phys . a * 15 * ( 2000 ) 133 ; z. phys .
c * 63 * ( 1994 ) 427 ; hep - ph/9403299 .
bugg , phys .
b * 572 * ( 2003 ) 1 [ erratum - ibid .
b * 595 * ( 2004 ) 556 ] ; phys .
* 397 * ( 2004 ) 257 .
a. afriat , f. selleri : _ the einstein , podolsky , and rosen paradox in atomic , nuclear and particle physics _ , 1999 by plenum press , new york ; isbn 0 - 306 - 45893 - 4 .
r. romano : _ doctoral thesis _
( university of trieste , 2002 ) [ hep - ph/0306164 ] ; e.b .
davies : _ quantum theory of open systems _ , academinc press , london , 1976 , isbn 0 - 12 - 206150 - 0 ; t. dittrich , p. hnggi , g .-
ingold , b. kramer , g. schn , w. zwerger : _ quantum transport and dissipation _ , wiley - vch , 1998 , isbn 3 - 527 - 29261 - 6 ; u. weiss : _ quantum dissipative systems _
( series in modern condensed matter physics , vol .
2 ) , 2nd .
enlarged edition , world scientivic , singapore , 1999 , isbn 981 - 02 - 4091 - 0 ; h.p .
breuer , f. petruccione : _ the theory of open quantum systems _
, oxford university press , 2002 , isbn 0 - 19 - 852063 - 8 ; e. celeghini , m. rasetti , g. vitiello : annals phys .
* 215 * ( 1992 ) 156 ; a. iorio , g. vitiello : annals phys .
* 241 * ( 1995 ) 496 [ hep - th/9503136 ] .
weldon : nucl .
b * 534 * ( 1998 ) 467 .
s. sarkar , b.k .
patra , v.j .
menon , s. mallik : indian j. phys .
* 76a * ( 2002 ) 385 [ hep - th/0010062 ] . , ed .
j. nyiri , 2001 by world scientific publishing co. pte . ltd .
; isbn 981 - 02 - 4709 - 5 . | present hermitian quantum theory ( hqt ) , i.e. quantum mechanics ( qm ) and quantum field theory ( qft ) , is revised and replaced by a consistent non - hermitian formalism called non - hermitian quantum theory ( nhqt ) or ( anti)causal quantum theory ( ( a)cqt ) after lining out some inherent inconsistencies and problems arising in the context of causality , which is observed to introduce an indefinite metric in canonical commutation relations . choosing some ( very selective ) historical approach to introduce necessary terminology and explain complications when quantizing non - hermitian systems in the presence of an indefinite metric we propose a way how to construct a causal , analytic , poincar invariant , and local nhqt , the spacial representation of which is related to the so - called holomorphic representation used in complex analysis . besides providing a revised antiparticle , spinor , and probability concept , a new neutrino lagrangean , two distinct time - reversal operations , and generalized non - hermitian poincar transformations , we will apply nhqt to consider three important issues : pt - symmetry and non - hermitian similarity transforms , non - hermitian supersymmetry , and the construction of an asymptotically free theory of strong interactions without gluons . |
modern portfolio theory @xcite formulates the asset allocation problem as an optimization model , with the objective of maximising the expected portfolio return subject to keeping the estimated risk below a pedefined level ( the `` risk budget '' ) . in theory
, this approach should result in a carefully diversified asset allocation across various investable assets . in practice , however , optimal asset allocations computed from portfolio models are often seen as counterintuitive and ill - diversified , as they may contain large positions in just a few assets , and a large number of very small positions @xcite . to overcome this problem
, fund managers often impose additional constraints in the form of strategic allocation targets .
we have found however , that the imposition of allocation targets is artficial and unnecessary if both the parameter estimation and the numerics that go into a portfolio model are carried out very carefully .
this paper is aimed primarily at practitioners . inspired by an essay by wilmott @xcite ,
we give a list of seven sins in porfolio optimization that should be avoided at all cost .
the best known investment model is the one - period mean - variance ( mv ) model of markowitz @xcite . for the purposes of this paper we restrict ourselves to this model , as it is both very simple and widely known among the readership and yet of fundamental interest in finance .
a very similar discussion could be held about more complex multiperiod models that involve more general risk terms . for the sake of completeness and to define the notation , we start with a brief recap of the mv model . an investor wishes to actively manage a portfolio in @xmath0 risky assets .
the investor holds fixed positions @xmath1 in asset @xmath2 over an investment period @xmath3 $ ] ( one hour , day , week , month , ... ) , at the end of which he / she is prepared to adjust the positions again .
the expected return of asset @xmath2 is @xmath4 $ ] .
@xmath5 is the random variable describing the return per unit position in asset @xmath2 over the investment period @xmath3 $ ] .
hence the expected portfolio return is @xmath6 = x^{\operatorname{t}}e[r].\ ] ] the variance of the portfolio return is @xmath7 the elements @xmath8 of the @xmath9 variance - covariance matrix @xmath10 are @xmath11\bigr)\bigl(r_j-\operatorname{e}[r_j]\bigr ) \right].\ ] ] the problem the investor faces is to decide on the positions @xmath1 .
the mean - variance approach to this problem proposes to choose the positions by solving the following optimization problem , @xmath12\label{mean variance}\\ \text{s.t . } & x^{\operatorname{t}}\operatorname{cov}(r , r)x\leq\operatorname{\sigma^2_{max}},\nonumber\end{aligned}\ ] ] where @xmath13 refers to the @xmath14 where the objective function is maximised . to make model useable in practice
, the expectations and covariances need to be replaced by estimates @xmath15,\\ \sigma_{ij}&\approx\operatorname{cov}\bigl(r_i , r_j\bigr).\end{aligned}\ ] ] this is usually done via proprietary methods that use historical prices and other data available at time @xmath16 .
the practical problem being solved is then @xmath17 where @xmath18 denotes the transpose of a vector , @xmath19 is the vector of @xmath20 and @xmath21 is the matrix of @xmath22 . for @xmath23 .
the red line marks the boundary of the ellipsoid @xmath24 .
the eigenvectors of @xmath21 define the principal directions of the ellipsoid and the inverse of the square roots of the eigenvalues are the corresponding equatorial radii .
, scaledwidth=75.0% ] the set @xmath25 of decision vectors that satisfy the constraints of an optimization problem is called the _
feasible domain_. for the model problem this is the interior of an ellipsoid induced by the inequality constraint @xmath26 . if @xmath25 is a convex set ( that is , for any pair of points @xmath27 the line segment @xmath28\}$ ] between @xmath29 and @xmath30 lies in @xmath25 ) , and if @xmath31 is a convex function ( that is , for any @xmath27 and @xmath32 $ ] , @xmath33 ) , then the problem is called a _
convex optimization problem _ be a convex set . ] .
if @xmath21 is symmetric positive definite and @xmath34 , the analytic solution of this problem is given as @xmath35 note that @xmath36 , e.g. the optimum lies at the boundary of the aforementioned ellipsoid .
we introduce the sharpe ratio @xcite for @xmath37 in this context as @xmath38 where @xmath39 is a risk - free interest rate over the considered investment period .
we will henceforth assume short investment periods on the order of days and set @xmath40 , but all the calculations easily extend to the case where @xmath41 .
note that the sharpe ratio @xmath42 of the optimal portfolio does not depend on @xmath43 .
we identified the model formulation , the estimation of parameters and the algorithmic solution as potential mine fields of portfolio optimization .
another mine field concerns parameter uncertainty . in order to keep the technical difficulty of this paper to a minimum , we chose not to deal with this issue here , but we refer our readers to the extensive recent literature on robust optimization @xcite . among the countless ways to produce unsatisfactory results through portfolio modelling , we chose to discuss just seven of the most common mistakes that we have seen committed .
this is a true classic that deserves the top spot here . a single negative eigenvalue ( even if close to zero ) can spoil it all . in practice
the covariance matrix @xmath21 is estimated using historic data .
it is tempting to estimate the elements of the matrix using independent processes for each entry , e.g. moving averages with distinct update rates .
this is a recipe for disaster .
we illustrate the consequences : let @xmath21 be a symmetric real matrix .
then there exists a real orthonormal matrix @xmath44 such that @xmath45 is a diagonal matrix .
the diagonal entries of @xmath46 are the eigenvalues of @xmath21 .
the columns of @xmath44 are the eigenvectors .
let us assume that there exists an eigenvector @xmath47 associated with a negative eigenvalue @xmath48 .
then @xmath49 .
hence we can find a portfolio with a negative variance , that is , @xmath47 corresponds to a portfolio of negative risk .
taking weights @xmath50 , where @xmath51 when @xmath52 and @xmath53 otherwise , and where @xmath54 , we obtain a portfolio with expected return @xmath55 that satisfies the risk budget , since @xmath56 this might entice one to take arbitrarily large positions in the erroneous belief that the risk budget will not be exceeded .
some argue that the introduction of extra constraints such as the self - financing condition @xmath57 and non - shortselling conditions @xmath58 avoid the problem of large positions .
however , the positions will still be completely nonsensical , and the extra constraints may not be applicable to all asset classes . although there exist some approaches to correct estimated and polluted covariances , see higham@xcite , we found that a careful analysis of the estimation process itself adds far more value to the trading strategy .
the second sin is of more subtle nature and yet more dramatic . for almost rank deficient matrices the ellipsoid in fig .
[ fig : volcon1 ] becomes very cigar - shaped .
the eigenvectors associated with small eigenvalues will change strongly under even very small perturbations of the matrix . in a typical backtest
the entries of such a matrix are updated and hence the ellipsoid will rotate
. even if @xmath21 has only positive eigenvalues they might be still too small .
the inverse of @xmath21 is @xmath59 where @xmath44 and @xmath46 are the aforementioned matrices of eigenvectors and eigenvalues .
this decomposition can be represented as sum of outer products , that is @xmath60 we assume @xmath61 .
inserting into yields @xmath62 the position @xmath63 is a linear combination of @xmath0 _ principal portfolios _ @xmath64 .
this term only gives the explicit solution for the standard model .
the effect of ill - conditioning is equally important in problems with more constraints .
this reveals the fundamental problem of the markowitz model : the weights scale as @xmath65 , that is , the optimizer will try to align the position with principal portfolios linked with small eigenvalues .
michaud@xcite calls this effect the maximization of the estimation error as the eigenvectors associated with small eigenvalues are most sensitive to noise . in practice
, it has often been observed that this results in rather extreme positions .
there are numerous approaches to overcoming this behavior .
a popular approach is to ignore eigenvalues smaller than a threshold in the sum for @xmath63 .
often the threshold is induced by the wigner semicircle distribution .
another powerful approach is to use shrinkage estimators @xcite , the most simple of which is the convex combination @xmath66 where @xmath67 is the identity matrix .
so the eigenvalues of @xmath68 are @xmath69 .
although matrix shrinkage towards @xmath67 does not change the eigenvectors , and hence eigenvectors are still equally sensitive to noise , the eigenvalues change ( and are in particular larger than @xmath70 ) .
this weakens the impact of the last few eigenvectors , especially as @xmath19 tends to be close to orthogonal to them . a third approach that shows excellent practial results
was developed by zuev @xcite , who used semi - definite programming to maximize the smallest eigenvalue among the covariance matrices @xmath21 that lie in a certain subset of the set of symmetric positive definite matrices .
we have observed that some find it simply too tempting to bypass all the maths by using a more trivial intermediate stage in the optimization process . rather than solving the toy problem
, some practitioners first identify the unit - vector @xmath30 that results in the maximal expected return .
this vector is @xmath71 . in a second stage
they then identify the scaling factor @xmath72 that achieves to achieve @xmath73 . using this approach
the optimal vector is @xmath74 although this looks close to the analytic solution , dramatic differences exist : in the numerator of @xmath75 all information contained in @xmath21 is ignored . as a result ,
this approach does not exploit diversification , one of the uncontested benefits of investing in a portfolio rather than a single asset .
worse even , there is some danger lurking in the denominator .
as soon as @xmath19 is aligned with an eigenvector corresponding to a small eigenvalue the denominator can be extremely small and blow up the position size and hence the portfolio .
a modern solver may iterate through the feasible domain by taking dozens of intermediate steps that converge towards the unique global maximum of .
however , the choice of those steps should be left to the solver .
most portfolio optimization models are convex programming problems , a class of well - studied optimization problems with rich mathematical structure @xcite .
non - convex optimization problems may have multiple local extrema and convergence to a global extrema ( it may be not unique ) can not be guaranteed as the solver may get attracted to any of the local extrema @xcite . when multiple optimization problems are solved sequentially , the solution may get attracted to a different local extremum each time , even when the model parameters change only slightly , and this can result in high trading costs .
it is therefore often preferrable to use a convex portfolio model that avoids these costs or at least do extensive backtesting to estimate the impact of these costs on performance .
a related problem is the failure to recognize convexity when it is present implicitly in a model , as this prevents one from using convex optimization software that would be able to solve the problem much efficiently and robustly than most nonconvex solvers . as an example , consider a portfolio manager who aims to maximize the sharpe ratio @xcite @xmath76 we note that the sharpe ratio is not defined for @xmath37 and the limit of @xmath77 for @xmath78 does not exist .
further , the sharpe ratio function has no unique maximizer , since @xmath79 for all @xmath80 and @xmath37 .
using model in conjunction with a nonconvex solver may lead to the evaluation of @xmath77 for points close to the origin , which causes numerical inaccuracy . to avoid this problem
, we wish to avoid the origin and a reasonably large region around it . exploiting the fact that @xmath79 , we may rescale @xmath29 to satisfy @xmath81 . hence
, the problem becomes @xmath82 the numerator of @xmath77 now being constrained to a fixed number , it plays no role in the maximization of @xmath77 and can be neglected , that is , the optimal decision vector @xmath63 can be found by solving the simpler problem @xmath83 this problem is still not convex , as the feasible domain is the boundary of the risk ellipsoid , but it provably has the same solution as the convex problem @xmath84 for generically we have @xmath85 , and if @xmath86 , then @xmath87 is an improved feasible solution for @xmath88 .
this contradicts the optimality of @xmath63 and shows that the optimal @xmath63 must satisfy the constraint of the nonconvex problem automatically . in this fashion , solving the convex problem yields a maximizer @xmath63 of the nonconvex problem whilst avoiding the associated numerical problems .
most portfolio managers solve a variant of with additional constraints for which the optimal solution is no longer given in closed form .
a solution has to be computed algorithmically in this case , and the choice of a good solver is crucial .
the most robust and powerful solvers require the problem to be reformulated in a specific standard form , which usually requires a lift - and - project approach that will be further described below .
for example , matlab s ` quadprog ` is unable to solve problem directly without lifting the risk constraint into to the utility function . bypassing such steps by writing proprietary optimization software
is not recommended , as the numerical challenges faced in such attempts are easily underestimated .
a particularly popular route in this context is to use _ simulated annealing _ , because it is intuitive to understand and trivial to code .
however , whilst simulated annealing has its place in highly nonconvex and unstructured global optimization problems , it is not designed to exploit any of the strong mathematical structures that underly portfolio problems , in particular convexity , and as a result it converges far less robustly and exceedingly more slowly than modern convex optimization solvers . to demonstrate the fallacy of simulated annealing we tried to implement a simple scheme . however , even the most basic convex and constrained problems are hard to approximate with such algorithms and require careful tuning . it remains a black art . in a multi - period setting in which the portfolio positions are frequently updated , it is of utmost importance to choose a robust solver that yields reproducible results free of random fluctuations that cause unnecessary transactions costs when rebalancing the portfolio .
the issue of transaction costs is so important that is wise to reduce the fluctuations that occur as a function of the changing model parameters even further , by ways of regularization terms . the use of a fast , robust , reliable , deterministic algorithm is also hugely important in any back - testing framework for quantitative trading in which one wishes to study the statistical behavior of the trading strategy and avoid all avoidable sources of artificial randomness .
most portfolio problems have reformulations as second order cone programming problems ( socps ) , or , occasionally semidefinite programming problems ( sdps ) .
such problems can be solved in polynomial time via iterative schemes known as _ interior - point methods_. practical implementations of such schemes are currently the leading codes for solving portfolio problems in terms of robustness , speed and reliability .
leading codes include the following : * sdpt3 @xcite , see http://www.math.nus.edu.sg/@xmath89mattohkc/sdpt3.html , is freely available , only available for matlab .
* sedumi @xcite , see http://sedumi.ie.lehigh.edu/ , is very similar to sdpt3 . * mosek @xcite , see http://www.mosek.com is the leading commercial conic programming solver and has interfaces for the most common programming languages .
lifting is a powerful technique to make hard problems look simple by introducing additional dimensions . using a classic example we demonstrate what it is all about :
often problems of type are solved in a large loop over thousands of investment periods . for each period some of the estimates may change and the portfolio is rebalanced .
it is certainly a good idea to take into account some costs when changing the position .
most quants use quadratic costs as the resulting cost term is differentiable .
however , such terms tend to overestimate costs for large trades .
let us assume that the system currently holds the position @xmath90 and a new investment periods starts . the formulation above is not taking into account the position @xmath90 at all .
this may result in large costs and spurious oscillations in the position.to address these problems we penalize deviations from @xmath90 , that is @xmath91 the positive parameters @xmath92 reflect the estimate costs per unit position .
introducing auxiliary variables @xmath93 the problem can be reformulated as @xmath94 \label{lttf2}\\ \text{s.t . } & x^{\operatorname{t}}c x\leq\operatorname{\sigma^2_{max } } , \nonumber \\ & x_i - x^0_i \leq t_i,\nonumber \\ & x^0_i - x_i \leq t_i.\nonumber\end{aligned}\ ] ] problem is called a _ lifting _ of problem . through the introduction of extra variables ,
liftings involve an inflation of the problem dimension .
this seeming disadvantage is offset when the following occurs : * the lifted problem belongs to a problem class with lower complexity than the original problem .
for example , in the above case , a nonsmooth problem turned into a smooth one ( the nondifferentiable absolute value terms in the constraint have disappeared ) .
in other cases , nonconvex problems can be convexified through liftings . *
the lifted problem belongs to a problem class for which efficient standard software already exists , avoiding the need to implement a custom designed algorithm . in the example above , a quadratic programm resulted , one of the most efficiently solved class of optimization problems .
as long as @xmath97 problem has a non - empty feasible domain as it contains the trivial portfolio @xmath98 . in equity portfolios
the position @xmath29 is often interpreted as a fraction of the investors s capital .
being _ fully invested _ is then reflected by the constraint @xmath99 .
the enhanced problem @xmath12\label{equity}\\ \text{s.t . } & x^{\operatorname{t}}\operatorname{cov}(r , r)x\leq\operatorname{\sigma^2_{max}},\nonumber\\ & e^{\operatorname{t } } x = 1 \nonumber\end{aligned}\ ] ] has no solution if @xmath43 is smaller than the minimum variance @xmath100 of all fully invested portfolios , e.g. @xmath101 c. stein .
inadmissibility of the usual estimator for the mean of a multivariate distribution . in j. neyman ( ed . ) , _ proc .
berkeley symp .
statist .
_ , vol . 1 , pp .
california press , 1956 .
todd , k.c .
toh and r.h .
ttnc . on the implementation and usage of sdpt3
a matlab software package for semidenite - quadratic - linear programming , version 4.0 .
_ http://www.math.nus.edu.sg/
mattohkc / sdpt3.html_. p. wilmott .
the commonest mistakes in quantitative finance : a dozen basic lessons in commonsense for quants and risk managers and the traders who rely on them . _ frequently asked questions in quantitative finance , 2nd edition _ , wiley , 2009 , pp . | although modern portfolio theory has been in existence for over 60 years , fund managers often struggle to get its models to produce reliable portfolio allocations without strongly constraining the decision vector by tight bands of strategic allocation targets .
the two main root causes to this problem are inadequate parameter estimation and numerical artifacts .
when both obstacles are overcome , portfolio models yield excellent allocations . in this paper , which is primarily aimed at practitioners
, we discuss the most common mistakes in setting up portfolio models and in solving them algorithmically .
primary 91g10 .
secondary 90c25 , 90c90 .
portfolio theory , mean - variance optimization , conic optimization . |
it is well known that many of the stars in the solar neighbourhood exist in multiple systems . as the number of planetary surveys increases ,
planets are regularly being found not only in single star systems , but binaries and triples as well .
for example , recently a hot jupiter has been claimed in the triple system hd 188753 ( konacki 2005 ; but see also eggenberger et al .
2007 ) , and @xcite lists 33 binary systems and 2 other hierarchical triples known to harbour exoplanets . as the majority of work on planetary dynamics has been for single star systems , the dynamics of bodies in these multiple stellar systems is of great interest . at first sight , it might not seem likely to expect long term stable planetary systems to exist in binary star systems , let alone triples , but numerical and observational work is starting to show otherwise . in recent years
, much study has been devoted to the stability of planets and planetesimal discs in binary star systems .
there are several investigations of individual systems ( e.g. work on @xmath0 cephei by @xcite , @xcite and @xcite ) as well as substantial amounts of work on more general limits for stability of smaller bodies in these systems .
notably , @xcite approach this problem by using numerical simulation data to empirical fit critical semimajor axes for test particle stability as functions of binary mass ratio and eccentricity .
these general studies are of great use in the investigation of observed systems and their stability properties , giving an effective and fast method of placing limits on stability in the large parameter space created by observational uncertainties .
the aim of this work is to investigate test particle stability in hierarchical triple star systems , and to see if any similar boundaries can be defined for them . to do this , a new method for numerically integrating planets in such systems
is presented , following the ideas for binary systems presented by @xcite .
although there have been empirical studies of the stability of hierarchical star systems themselves , there do not appear to have been studies of small bodies in such systems .
there is a great deal of literature concerning periodic orbits in the general three and four body problems ( see e.g. @xcite ) , but these are almost entirely devoted to considerations of planetary satellites in single star systems , for example satellite transfer in the sun - earth - moon systems .
also , while periodic orbits prove that stable solutions exist in these problems , they are of little practical use in determining general stability limits . the problem of planetary orbits in triple systems is more complex than for those in binary systems , as there are many different orbital configurations possible relative to the three stars .
these are considered in section [ sec : orbits ] , and a method for classifying them is described in order to simplify the discussion in this paper .
the derivation of a method to numerically integrate such planets is then given in section [ sec : maths ] . in section [ sec : stats ] is a brief overview of the statistical properties of known triple systems , as a basis for deciding the parameters of the systems used in the numerical simulations .
the results of numerical investigations into stability properties are then presented in section [ sec : sp ] for one of the possible orbital types .
the triple systems studied here are all hierarchical in nature .
that is , they can all be considered to be a close binary orbited by a distant companion in effect , an inner binary pair and an outer binary pair . here
, the inner binary stars are labelled as a and b , with a being the more massive , and the distant companion star as c. there are however three possible ways to pair the three stars into the hierarchy described above . @xcite
define the hierarchy by requiring that the instantaneous keplerian orbits of the inner and outer pair are bound , that the outer binary has a longer period than the inner binary and that the ratio of the two periods is the largest of the three possible pairings of the stars .
this definition is adopted here .
although other , non - hierarchical , types of motion are possible for triple stars ( see for example @xcite ) and observed ( see for example the trapezium systems listed in @xcite ) , they are not considered here .
it is an open question whether non - hierarchical systems can sustain long - term stable planets .
next some system of classifying planetary orbits is needed .
the orbital motion of small particles under the influence of two or three masses has been studied to some extent through periodic orbits of the three and four body problems , as mentioned in the introduction . as a result of this many families and classification schemes exist .
@xcite gives a good review , mentioning for example the ( a ) to ( r ) types designated for the copenhagen problem , dependent on the nature and location of the particles motion in both inertial and corotating frames .
many other features have been used to describe orbital families , for example symmetry @xcite or a parameter used to generate the orbital family . however , as mentioned by @xcite , there is no overall method , and the descriptions would not be applicable to non - periodic orbits .
orbital types in binary systems have generally been classified into three main groups , those around a single star ( circumstellar ) , those around both stars ( circumbinary ) and those in the middle ground as it were , coorbiting with one of the stars i.e. librating about the triangular lagrangian points , similar to trojan asteroids in the solar system .
a convenient labelling of these was given by @xcite , who designated planets as either s - orbits ( satellite ) , p - orbits ( planet ) , or l - orbits ( librator ) for circumstellar , circumbinary and coorbital respectively .
these ideas can be extended to hierarchical triple systems fairly simply .
it is clear that there will be three types of circumstellar motion , one for each star , and as such orbits like this can be labelled s(a ) , s(b ) and s(c ) for their primary .
the circumtriple orbit about the centre of mass of all three can be identified as a p orbit and labelled as such in analogy to dvorak s scheme .
finally , planets which orbit the centre of mass of the inner binary are circumbinary but also share characteristics with the satellite orbit , and can be labelled with the combination s(ab)-p .
these orbital types are shown in figure [ fig : types ] along with the binary cases for comparison , and listed in table [ tab : types ] .
a superscript 2 has been given to those in binary systems and a 3 for those in triple systems for clarity .
although these are the only types of motion studied here , for completeness the coorbital types are also be included
. this type of motion can occur for both the inner and outer binary , and labels ( ab ) and ( abc ) can be used to designate this . instead of the single l - orbit type
they can be broken up into t and h orbits to indicated tadpole ( about one of the triangular lagrangian points ) or horseshoe ( about both triangular points and the @xmath1 collinear point ) .
these are again illustrated in figure [ fig : types ] and included in table [ tab : types ] .
these orbits are defined to be bound relative to their focus and hierarchical in the same sense as the definition given for the stellar system . a particle that orbits outside the extent of the inner binary but has a bound orbit with respect to star a and the binary centre of mass would be an s(ab)-p orbit and not an s(a ) orbit .
this method of labelling clearly and concisely designates the exact nature of a planetary orbit , and extends a system already in use for binary stars .
it can also be easily applied to systems with other levels of hierarchy , for example if star c was replaced with another close binary pair c and d additional classes s(d ) and s(cd)-p would be possible , and the outer p type could be relabelled p(abcd ) . .
labels for the basic planetary orbital types in multiple star systems as defined in this work and compared to those of @xcite .
[ cols= " < , < , < , < " , ] [ tab : hwfit2 ] a more detailed investigation of the inner boundary can be carried out in a similar manner to the outer edge , by fixing star c at 1 @xmath2 and in a circular 50 au orbit and varying the inner binaries mass ratio and eccentricity , as discussed earlier .
the radius of the inner pair was kept at 1 au and their total mass as 2 @xmath2 .
the eccentricity of this binary was varied from 0.0 to 0.6 in steps of 0.2 again , and the mass ratio @xmath3 varied from 1.0 to 2.0 in steps of 0.1 .
figure [ fig : sp_crit_in ] shows the results from these simulations , with the locations of the first and last stable radii plotted as functions of the inner binaries eccentricity @xmath4 and mass ratio @xmath5 .
also plotted again is the fit given by @xcite to the critical radius for p type planets in binary systems .
as expected , the outer stability boundary seems unaffected by changes to the inner binary pair .
there are some unstable points between the two stability boundaries , most notably around 3.2 au for the simulations with @xmath6 .
@xcite find unstable islands appearing at the first @xmath7 beyond the critical semimajor axis in this configuration .
however , the location of these unstable radii here is well beyond the first @xmath7 mmr in the stable region .
in addition , running the @xmath8 and @xmath6 case without star c reveals that these locations are now stable .
this would indicate that this is again an effect due to the combination of all three stars .
because of this the inner edges location is simply taken as the same as the first stable radius .
the location of this boundary appears to be a very weak function of mass ratio and an approximately linear function of eccentricity , and also agrees well with the predictions of @xcite . for this boundary ,
they fit a function of the form @xmath9 where the constants are given in the third column of table [ tab : hwfit2 ] .
the terms up to fourth order are included to fit a variation in the position of the critical radius at smaller mass ratios than those considered here . fitting this function to the results here
does not produce well determined coefficients despite a low @xmath10 value of about 29 , as shown in the first column of the table .
in fact , a better solution is obtained by only including the first four terms , as shown in the middle column of the table .
the @xmath10 value for this fit is slightly higher at about 42 .
this second fit is plotted in figure [ fig : sp_i_fit ] , and seems to describe the data well , although it seems to slightly overestimate the size of the stable region . despite the smaller @xmath10 values here the fitted parameters are less well determined than those for the outer edge .
this is a reflection on the relative size of the test particle grid spacing compared to the size of the inner binaries orbit . here
the boundary is determined to within a tenth of the relative separation of the stars , while for the outer edge it was determined to within a thousandth of the size of the binary s orbit .
the parameter space investigated so far is somewhat limited .
the stars orbital longitudes have been ignored , assumed to be a minor influence on stability , the sets of simulations have always kept one star on a circular orbit , and all objects have been taken as coplanar .
brief investigations of these three extensions to the parameter space can be made .
firstly , the assumption that the initial longitudes of the stellar orbits has little effect on the stability boundaries was tested by running the inner edge simulations with mass ratios @xmath110.33 to 0.40 with a different initial longitude of the inner binary pair .
the results from these simulations were almost identical to those of the initial set , providing some evidence that this assumption is valid . next , the effect of both stars having non - circular orbits was investigated .
the semimajor axes of the inner and outer binary were kept at 1 and 50 au , and the eccentricities of both varied in steps of 0.2 from 0.0 to 0.6 .
the mass ratio of the outer binary only was varied as before .
the results of these simulations are shown in figure [ fig : eccentricity ] .
each panel shows the location of the stability radii and unstable points as for all values of @xmath4 and one given value of @xmath12 .
there is almost no change in the position of the outer stability boundary , but as both stellar eccentricities and the mass ratio increase , the inner boundary starts to move outwards .
the stellar eccentricities are still not varying to any significant extent and the additional instability must be due to the combined perturbations of all three stars .
lastly , the effect of inclination was studied .
the semimajor axes were kept as 1 and 50 au , @xmath4 was set to 0 , and the outer mass ratio and eccentricity varied as before .
sets of simulations where then run for inclinations of the outer star of @xmath13 , @xmath14 and @xmath15 .
the test particles were kept coplanar with the inner binary .
this is a rather limited investigation , as any dependence on the longitude of ascending node has been ignored , and only a small range of inclinations included .
however , higher inclinations will be subject to the kozai instability , causing large variations in the stellar orbits , which is expected to rapidly destabilise test particles . in these simulations
the stellar mutual inclination remains fairly constant and their orbits are stable .
figure [ fig : inclination ] shows the stability boundaries for these simulations , each panel comparing the different inclination results for a different value of @xmath12 .
there is little change in the inner boundary but the outer edge moves somewhat .
interestingly , for the @xmath16 case higher inclinations are more stable .
if the test particles are started instead coplanar with the outer binary the stability is similar , although not identical .
these results are consistent with the conclusions of @xcite , who show that inclination is not a significant effect on the stability of p type planets in binary system .
the main achievement of this paper is the formulation of a symplectic integrator algorithm suitable for hierarchical triple systems .
this extends the algorithm for binary systems presented by @xcite .
the positions of the stars are followed in hierarchical jacobi coordinates , whilst the planets are referenced purely to their primary .
each of the five distinct cases , namely circumtriple orbits , circumbinary orbits and circumstellar orbits around each of the stars in the hierarchical triple , requires a different splitting of the hamiltonian and hence a different formulation of the symplectic integration algorithm . here
, we have given the mathematical details for each of the five cases , and presented a working code that implements the algorithm .
as an application , a survey of the stability zones for circumbinary planets in hierarchical triples is presented . here , the planet orbits an inner binary , with a more distant companion star completing the stellar triple . using a set of numerical simulations
, we found the extent of the stable zone which can support long - lived planetary orbits and provided fits to the inner and outer edges .
the effect of low inclination on this boundary is minimal . a reasonable first approximation to a behaviour of a hierarchical triple
is to regard it as a superposition of the dynamics of the inner binary and a pseudo - binary consisting of the outer star and a point mass approximation to the inner binary .
if it is considered as two decoupled binary systems , then the earlier work of holman & wiegert ( 1999 ) on binaries is applicable to triples , except in the cases of high eccentricities and close or massive stars .
the implication here is that the addition of a stable third star does not distort the original binary stability boundaries .
as mentioned , @xcite have shown that overlapping sub - resonances are the cause of the boundary in the binary case .
it is reasonable to expect that in triples the same process is responsible , and the similarities between the binary and triple results support this theory .
it is also expected that there is a regime in which the resonances from each sub - binary start to overlap as well , further destabilising the test particles .
evidence of this is the deviation from the binary results when the stars are close , massive and very eccentric , when resonances would be both stronger and wider . since the parameter space investigated was chosen to reflect the observed systems it would seem to be a reflection on the characteristics of known triple stars that most lie in the decoupled regime .
the relatively constant nature of the stellar orbits in the simulations is however a consequence of the test particle orbits being destabilised long before the stars are close enough to interact . by extension of all these arguments
, it is expected that the binary criteria can be used to fairly accurately predict the stability zones in any hierarchical stellar system , no matter the number of stars .
the results presented here can be used to estimate the number of known hierarchical triple systems that could harbour s(ab)-p planets .
@xcite lists 54 systems with semimajor axis , eccentricity and masses for both the inner and outer components .
the mutual inclinations of most are not well known , but there are nine systems listed for which this angle can be determined . for five of these it is less than @xmath17 , two are around @xmath18 and two are retrograde .
although a small sample it suggests that there are systems that fall within the low inclination regime investigated here . using the criteria of @xcite and those found here for the position of the inner and outer critical semimajor axis the size of the coplanar stable region for each of these triples can be calculated .
this can be considered an upper limit , since it is likely that very non - coplanar systems and those with significant eccentricities for both binary components will further destabilise planets .
of the 54 systems 13 are completely unstable to circumbinary planets according to the four parameter fits ( compared to 11 using @xcite s criteria ) .
figure [ fig : zone ] shows a plot of the width of the stable region for the remaining systems .
interestingly , the majority seem to have very small stable zones , with 16 smaller than an au . whether this is a feature of triple systems , or an observation bias is not apparent
. it does indicate though that circumbinary planets are unlikely to exist in at least 50 % of observable systems . .
note that using the criteria of @xcite gives almost identical results . ]
beust , h. 2003 , a&a , 400 , 1129 broucke , r. a. 2004 , ann . new york acad .
sciences , 1019 , 408 chambers , j. e. , quintana , e. v. , duncan , m. j. , & lissauer , j. j. 2002 , aj , 123 , 2884 danby , j. m. a. 1988 , richmond , va . , u.s.a . : willmann - bell , 1988 .
2nd ed .
desidera , s. , & barbieri , m. 2007 , a&a , 462 , 345 dvorak , r. , pilat - lohinger , e. , funk , b. , & freistetter , f. 2003 , a&a , 398 , l1 dvorak , r. 1986 , a&a , 167 , 379 dvorak , r. 1984 , celestial mechanics , 34 , 369 eggenberger a. , udry s. , mazeh t. , segal y. , mayor m. , 2007 , a&a , in press eggleton , p. , & kiseleva , l. 1995 , apj , 455 , 640 haghighipour , n. 2006 , apj , 644 , 543 holman , m. j. , & wiegert , p. a. 1999 , aj , 117 , 621 konacki , m. 2005 , nature , 436 , 230 mudryk , l. r. , & wu , y. 2006 , apj , 639 , 423 murray c. d. , dermott s. f. 2000 , solar system dynamics , cambridge university press pilat - lohinger , e. , funk , b. , & dvorak , r. 2003 , a&a , 400 , 1085 press , w. h. , teukolsky , s. a. , vetterling , w. t. , & flannery , b. p. 2002
, numerical recipes in c++ : the art of scientific computing , cambridge university press szebehely , v. 1977 , revista mexicana de astronomia y astrofisica , vol . 3 , 3 , 145 szebehely , v. 1967 , theory of orbits , chapter 9 , new york : academic press tokovinin , a. a. 1997 , a&as , 124 , 75 verrier , p. e. in prep . , phd thesis verrier , p. e. , & evans , n. w. 2006 , mnras , 368 , 1599 wisdom , j. , & holman , m. 1991 , aj , 102 , 1528
in this appendix are details of the symplectic integration method derived in section [ sec : maths ] for the orbital types p , s(a ) , s(b ) and s(c ) .
the split hamiltonian in each case is presented , and the evolution under these is readily obtainable using the method given in section [ sec : maths ] . also presented are details of the testing of the implementation of this symplectic integrator .
note that throughout this appendix @xmath19 is the total system mass and @xmath20 the total inner binary mass , including planets where relevant . for s(a ) orbits , the hamiltonian is @xmath21 where @xmath22 is the mass of star a and its planets , @xmath23 is the reduced mass of the inner binary , including the planets , @xmath24 is the reduced mass of the outer binary and @xmath25 . for s(b ) type planets only , the equations are @xmath26 where @xmath27 is the mass of star b and its planets , @xmath23 is the reduced mass of the inner binary , including the planets and @xmath24 is the reduced mass of the outer binary . here
@xmath28 now , and @xmath29 and @xmath30 . the corresponding hamiltonian for s(c )
type planets is @xmath31 where @xmath23 is the reduced mass of the inner binary stars and @xmath32 is the reduced mass of the outer binary , including the planets . here now @xmath33 , @xmath34 , @xmath35 and @xmath36 is the mass of star c and its planets . these differing orbital cases are all dealt with separately at present . to test the method and its implementation comparisons of the simulations of systems of various levels of complexity were made with the literature . as the equations given above allow @xmath39 to be set to zero , the conservation of the jacobi constant of test particles in the circular restricted three body problem was calculated and compared to that given by @xcite for their binary systems . the evolution of various triple star system was compared to studies of hierarchical triples in the literature , for example those given in @xcite . simulations of discs of test particles were also compared to those given by @xcite for circumbinary discs .
more complex systems of planets and test particles were also compared to the results of a standard bulirsch - stoer integrator @xcite . in all cases
excellent agreement was found . | a symplectic integrator algorithm suitable for hierarchical triple systems is formulated and tested .
the positions of the stars are followed in hierarchical jacobi coordinates , whilst the planets are referenced purely to their primary .
the algorithm is fast , accurate and easily generalised to incorporate collisions .
there are five distinct cases circumtriple orbits , circumbinary orbits and circumstellar orbits around each of the stars in the hierarchical triple which require a different formulation of the symplectic integration algorithm . as an application ,
a survey of the stability zones for planets in hierarchical triples is presented , with the case of a single planet orbiting the inner binary considered in detail .
fits to the inner and outer edges of the stability zone are computed .
considering the hierarchical triple as two decoupled binary systems , the earlier work of holman & wiegert on binaries is shown to be applicable to triples , except in the cases of high eccentricities and close or massive stars .
application to triple stars with good data in the multiple star catalogue suggests that more than 50 % are unable to support circumbinary planets , as the stable zone is non - existent or very narrow .
[ firstpage ] celestial mechanics planetary systems binaries : general methods : _ n_-body simulations |
dynamical systems which display scale - free behaviour have attracted much interest in recent years .
equilibrium thermodynamic systems do only exhibit scale - free behaviour for a subset of the parameter space of measure zero ( the critical values of the parameters ) .
nevertheless , in nature scale - free systems can be found in abundant variety ( earthquakes @xcite , avalanches in rice - piles @xcite , infected people in epidemics @xcite , jams in internet traffic @xcite , extinction events in biology @xcite , life - times of species @xcite and many more .
see also @xcite and the references therein ) .
the origin of this abundance lies probably in the broad variety of systems far from equilibrium that can be found in nature . with the onset of non - equilibrium dynamics ,
new mechanisms come into play which seem to make scale - free behaviour a generic feature of many systems .
however , unlike equilibrium thermodynamics , where scaling is thoroughly understood @xcite , for non - equilibrium dynamical systems there does not yet exist a unified theory of scale - free phenomena ( apart from non - equilibrium phase transitions ) .
there do , however , exist several distinct classes of systems with generic scale - free dynamic .
one of the first ideas to explain scale - free behaviour in a large class of dynamical systems was the notion of self - organized criticality ( soc ) proposed by bak , tang and wiesenfeld in 1987 @xcite .
they proposed that certain systems with local interactions can , under the influence of a small , local driving force , self - organize into a state with diverging correlation length and therefore scale - free behaviour .
this state is similar to the ordinary critical state that arises at the critical point in phase transitions , although no fine - tuning of parameters is necessary to reach it .
since 1987 literally thousands of research papers have been written concerning soc ( for a review see @xcite ) , and many different dynamical systems have been called soc ( e.g. @xcite ) , mostly just because they showed some power - law distributed activity patterns .
recently @xcite it has become clear that several soc models ( sandpile models , forest - fire models ) can be understood in terms of ordinary nonequilibrium critical phenomena ( like e.g. @xcite ) . driving rate and dissipation act as critical parameters
. the critical value , however , is 0 .
therefore , it suffices to choose a small driving rate and dissipation to fine - tune the system to the critical point , and this choice is usually implicit in the definition of the model .
scale - free behaviour does not , however , depend crucially on some sort of critical phenomenon .
a simple multiplicative stochastic process ( msp ) of the form @xmath3 where @xmath4 and @xmath5 are random variables , can produce a random variable @xmath6 with a probability - density function ( pdf ) with power - law tail @xcite . in processes of this type
, the power - law appears under relatively weak conditions on the pdf s of @xmath4 and @xmath5 , thus making the intermittend behaviour a generic feature of such models .
applications of eq .
( [ eq : multstoch ] ) can be found in population dynamics with external sources @xcite , epidemics @xcite , price volatility in economics @xcite , and others .
another class of models with a very simple and robust mechanism to produce scale - free behaviour has been introduced recently by newman and sneppen @xcite .
these so called coherent - noise models consist of a large array of agents which are forced to reorganize under externally imposed stress . in their simplest form ,
coherent - noise models do not have any interactions between the agents they consist of , and hence , certainly do not display criticallity .
nevertheless , these models show a power - law distribution of the reorganization events with a wide range of different exponents @xcite , depending on the special implementation of the basic mechanism .
moreover they display power - law distributions in several other quantities , e.g. , the life - time distribution of the agents .
these models have been used to study earthquakes @xcite , rice piles @xcite , and biological extinction @xcite .
coherent - noise models have a feature that usually is not present in soc models and is never present in msp s , which is the existence of aftershocks . in most coherent - noise models
the probability for a big event to occur is very much increased immediately after a previous big event and then decays with a power - law .
this leads to a fractal pattern of events that are followed by smaller ones which themselves are followed by even smaller ones and so on . in most soc models and all msp s , on the contrary
, the state of the system is statistically identical before and after a big event .
therefore in these models no aftershocks are visible .
the existence or non - existence of aftershock events should be easily testable in natural systems .
this could provide a means to decide what mechanism is most likely to be the cause for scale - free behaviour in different situations @xcite .
but to achieve this it is important to have a deep understanding of the decay - pattern of the aftershock events .
the goal of the present paper is to investigate in detail the aftershock dynamics of coherent - noise models .
we concentrate mainly on the original model introduced by newman and sneppen because there can be obtained several analytical results .
we find a power - law decrease in time of the aftershocks probability to appear , as has been found already in @xcite .
but unlike stated there , we can show that the exponent does indeed depend on the microscopic details of the simulation .
we find a wide range of exponents , from 0 to values well above 1 , whereas in @xcite the authors report only the value 1 .
we will now describe the model introduced by newman and sneppen @xcite .
the system consists of @xmath7 units or agents. every agent @xmath8 has a threshold @xmath9 again external stress .
the thresholds are initially chosen at random from some probability distribution @xmath10 .
the dynamics of the system is as follows : 1 .
a stress @xmath11 is drawn from some distribution @xmath12 .
all agents with @xmath13 are given new random thresholds , again from the distribution @xmath14 .
a small fraction @xmath15 of the agents is selected at random and also given new thresholds .
3 . the next time - step begins with ( i ) .
step ( ii ) is necessary to prevent the model from grinding to a halt . without this random reorganization the thresholds of the agents
would after some time be well above the mean of the stress distribution and average stress could not hit any agents anymore .
the most common choices for the threshold and stress distributions are a uniform threshold distribution and some stress distribution that is falling off quickly , like the exponential or the gaussian distribution . under these conditions
( with reasonably small @xmath15 ) it is guaranteed that the distribution of reorganization events that arises through the dynamics of the system will be a power - law .
there are several possibilities to extend the model to make it more general , without loss of the basic features .
two extensions that have been studied by sneppen and newman @xcite are * a lattice version where the agents are put on a lattice and with every agent hit by stress its nearest neighbours undergo reorganization , even if their threshold is above the current stress level . * a multi - trait version where , instead of a single stress , there are @xmath16 different types of stress , i.e. the stress becomes a @xmath16-dimensional vector @xmath17 .
accordingly , every agent has a vector of thresholds @xmath18 .
an agent has to move in this model whenever at least one of the components of the threshold vector is exceeded by the corresponding component of the stress vector .
an extension that is especially important for the application of coherent noise models to biological evolution and the dynamics of mass extinctions has been studied recently by wilke and martinetz @xcite . in biology
it is not a good assumption to keep the number of agents ( in this case species ) constant in time .
rather , species which go extinct should be removed , and there should be a steady regrowth of new species . in a generalized model , the system size is allowed to vary .
agents that are hit by stress are removed from the system completely , but at the end of every time - step a number @xmath19 of new agents is introduced into the system . here @xmath19 is a function of the actual system size @xmath7 , the maximal system size @xmath20 and some growth rate @xmath21 .
wilke and martinetz have studied in detail the function @xmath22 which resembles logistic growth . in the limit @xmath23 their model reduces to the original one by newman and sneppen . in the following
we will refer to the original model as the infinite - growth version and to the model introduced by wilke and martinetz as the finite - growth version.
we base our analysis of the aftershock structure on the meassurement - procedure proposed by sneppen and newman @xcite .
they drew a histogram of all the times whenever an event of size @xmath24 some constant @xmath25 happened after an initial event of size @xmath24 some constant @xmath26 , for all events @xmath27 .
consequently , we measure the frequency of events larger than @xmath25 occuring exactly @xmath0 time - steps after an initial event larger than @xmath26 , for all times @xmath0 .
this means that we consider sequences of events in the aftermath of initial large events .
normalized appropriately , our measurement gives just the probability to find an event @xmath28 at time @xmath0 after some arbitrarily chosen event @xmath29 .
for this to make sense in the context of aftershocks we usually have @xmath30 . throughout the rest of this paper we use @xmath26 and @xmath25 as percentage of the maximal system size .
therefore a value @xmath31 for example means that we are looking for initial events which span the whole system .
we will denote the probability to find an event of size @xmath32 at time @xmath0 after a previous large event by @xmath33 . in order to keep the notation simple
we omit the constant @xmath26
. it will be clear from the context what @xmath26 we use in different situations .
note that @xmath33 is not a probability distribution , but a function of time @xmath0 .
therefore , every increase or decrease of @xmath33 in time will indicate correlations between the initial event and the subsequent aftershocks .
for @xmath34 we expect all correlations to disappear , and hence @xmath33 to tend towards a constant .
it is possible to obtain some analytical results about the probability @xmath33 if we restrict ourself to the model with infinite growth and a special choice for the threshold and stress distributions . if not indicated otherwise , throughout the rest of this section we assume @xmath10 to be uniform between 0 and 1 , and the stress distribution to be exponential : @xmath35 .
furthermore , we focus on the case @xmath31 .
that means that we are looking at the events in the aftermath of an initial event of infinite size , an event that spans the whole system .
this is a reasonable situation because we use a uniform threshold distribution . in this case
there is a finite probability to generate a stress @xmath11 which exceedes the largest threshold , thus causing the whole system to reorganize .
for some of the examples presented in section [ sec : numerical_results ] the probability to find an infinite event is even higher than @xmath36 .
this probability can be considered relatively large in a system where one has to do about @xmath37 time - steps to get a good statistics . the exact way to calculate
@xmath33 is the following .
one has to compute the distribution @xmath38 which is the distribution that arises if after the big event at time @xmath39 a sequence of stress values @xmath40 is generated during the following time steps
. then the equation @xmath41 has to be solved . that gives the quantity @xmath42 , the threshold that has to be exceeded by the stress at time @xmath0 to generate an event @xmath28 . from the stress distribution
one can then calculate the corresponding probability @xmath43 .
finally the average over all possible sequences @xmath40 has to be taken to end up with the exact solution for @xmath33 .
obviously there is no hope doing this analytically . instead of the exact solution for @xmath33 we can calculate a mean - field solution if we average over all possible sequences @xmath40 before we solve eq .
( [ exact_cal_p ] ) .
note that in this context , the notion mean - field does not stand for the average state of the system , which does not tell us anything about aftershocks , but for the average fluctuations typically found in a time - intervall @xmath44 .
these average fluctuations are a measure for the return to the average state , after a large event has caused a significant departure from it . consequently , the mean - field solution is time - dependent . for @xmath45 ,
the time - dependent mean - field threshold distribution converges to the average threshold distribution , denoted by @xmath46 in @xcite in appendix [ a : master - equation ] , we show that the averaging over all fluctuations in a time intervall of @xmath0 time - steps equals to @xmath0 times iterating the master - equation . therefore , to calculate the mean - field solution for @xmath33 we have to insert @xmath47 , the @xmath0-th iterate of the master - equation , into eq .
( [ exact_cal_p ] ) .
the details of this calculation are given in appendix [ a : master - iteration ] .
in this paragraph we will calculate the dependency of the exponent @xmath2 on @xmath25 under the assumption that the probability to find aftershocks decays indeed as a power - law , i.e. that we can assume @xmath48 .
a fairly simple argument shows that for the probability @xmath33 to decrease as a power - law the exponent @xmath2 must be proportional to @xmath49 for @xmath25 not too small .
again we concentrate on exponentially distributed stress only .
we begin with an approximation of the quantity @xmath50 , which is the average threshold at time @xmath0 above which a stress value must lie to trigger an event of size @xmath28 . in the mean - field approximation , @xmath50 is defined by the equation @xmath51 because @xmath47 and @xmath25 are normalized , we can rewrite this equation ( again we assume @xmath52 to be uniform between 0 and 1 ) : @xmath53 for the most reasonable stress distributions the distribution of the agents @xmath47 forms a plateau in the region close to @xmath54 . therefore for sufficient large @xmath25 we can approximate the integral in eq .
( [ reversed_integral ] ) by substituting @xmath47 with its value at @xmath54 , which is @xmath55 .
( [ reversed_integral ] ) then becomes @xmath56 the values @xmath55 are a function of @xmath0 .
we define @xmath57 and find for @xmath50 : @xmath58 the probability @xmath33 now becomes @xmath59 the principal idea to derive a relation between @xmath2 and @xmath25 is as follows .
the function @xmath60 is obviously independent of @xmath2 and @xmath25 .
we make the ansatz @xmath48 , rearrange eq .
( [ p_texp ] ) for @xmath60 and then get a condition on @xmath2 and @xmath25 since they should cancel exactly .
hence we have to solve the equation @xmath61 where @xmath62 is the constant of proportionality .
we take the logarithm on both sides to get @xmath63 and finally @xmath64 this is of the form @xmath65 , where @xmath66 and @xmath67 for every function of the form @xmath65 , the constants @xmath68 and @xmath69 are unique , as can be seen if we write @xmath70 a change in @xmath69 results in a rescaling of the variable @xmath0 , while a change in @xmath68 results in a rescaling of the whole function .
consequently , neither @xmath68 nor @xmath69 can depend on @xmath2 or @xmath25 .
this can be seen as follows . if , e.g. , @xmath68 depended on @xmath25 , then a change in @xmath25 would rescale the function @xmath60 .
but this function is independent of @xmath25 according to its definition ( eq . ( [ rtdef ] ) ) .
hence @xmath68 must be independent of @xmath25 in itself .
a similar argument holds for the variable @xmath69 .
therefore , @xmath25 and @xmath2 must cancel exactly in eq .
( [ c_1def ] ) .
this leads to the condition @xmath71 up to now we have not done any assumptions about the size of the first big event after which we are measuring the subsequent aftershocks . therefore the proportionality @xmath72 should hold in general , as long as @xmath25 is not too small . if we additionally assume the inital event to have infinite size ( @xmath31 ) we can easily calculate the constant @xmath62 in eq .
( [ p_texp2 ] ) .
the meaning of this constant is the probability to get an event of size @xmath28 immediately after the initial big event , as can be seen by setting @xmath73 : @xmath74 for the case @xmath31 the distribution of thresholds @xmath75 is uniform and thus @xmath76 with eqs .
( [ p_texp ] ) , ( [ p_texp2 ] ) , ( [ c_1def ] ) , and ( [ const_a ] ) we can write the probability @xmath33 as @xmath77 in section [ sec : numerical_results ] we will find numerically that @xmath78 , and therefore @xmath79 . for two limiting cases we can deduce the behaviour of the exponent @xmath2 regardless of the stress distribution .
we begin with the case @xmath31 , @xmath80 . from eq .
( [ tau_prop_s1 ] ) we find that @xmath81 as @xmath80 under the assumption of an exponential stress distribution .
but this result is more general .
for @xmath82 the probability @xmath33 reads simply @xmath83 and hence is constant in time .
consequently we have @xmath84 for any stress distribution . from continuity
we have @xmath81 as @xmath80 .
a similar argument holds when either @xmath26 or @xmath25 approaches 0 . for @xmath85 ,
the probability @xmath33 reduces to the mean probability to find an event of size @xmath32 .
hence @xmath84 . for @xmath86 ,
the probability @xmath33 becomes 1 , because an event of size at least zero can be found in every time step .
hence also in this case @xmath84 . from continuity
we have again @xmath81 as either @xmath87 or @xmath88 .
in the previous section we have focused on the behaviour of the system in the aftermath of an infinite event .
this situation is not only analytically tractable , but it also makes it simpler to obtain good numerical results . if we want to measure the probability to find aftershocks following events exceeding some finite but large @xmath26 , we have to wait a long time for every single measurement since the number of those large events vanishes with a power - law .
this makes it hard to get a good statistics within a reasonable amount of computing time .
if , on the other hand , we focus on the situation of an infinite initial event , we can simply initialize the agents with the uniform threshold distribution , take the measurement up to the time @xmath0 we are interested in and repeat this procedure until the desired accuracy is reached .
unless stated otherwise , the results reported below have been obtained in this way , and with exponentially distributed stress .
the @xmath0-th iteration of the master - equation should give exactly the average distribution of the agent s thresholds at time @xmath0 . in fig .
[ fig_rho_t ] it can be seen that this is indeed the case .
the points , which represent simulation results , lie exactly on the solid lines , which stem from the exact analytical calculation .
the mean - field approximation for @xmath33 should be valid if the agent s distribution at time @xmath0 does not fluctuate too much about the average distribution @xmath47 .
since there are many more small events than big ones the fluctuations should occur primarily in the region of small @xmath89 . consequently we expect the mean - field approximation to be valid for large @xmath25 , and to break down for small @xmath25 .
[ fig_mf1 ] shows that already for moderately large @xmath25 the mean - field approximation captures the behaviour of @xmath33 , with a slight tendency to underestimate the results of the measurement .
note that the statistics is becoming worse with increasing @xmath25 due to the rapidly decreasing probability to find any events of size @xmath28 for large @xmath25 . in fig .
[ fig_limf ] a measurement of the probability @xmath90 is presented for a number of simulations with different values of the parameter @xmath15 .
as it can be seen , the parameter @xmath15 does not affect the exponent of the power - law , but limits the region where scaling can be observed .
note the difference between the effect seen here and typical cut off effects in the theory of phase transitions .
the quantity @xmath90 is not a probability distribution , but a time dependent probability , which tends towards a constant for @xmath34 .
therefore , we do not see an exponential decrease at the cut off timescale . instead , the probability @xmath90 levels out to the time - averaged value @xmath91 , which is the average probability to find events of size @xmath32 .
in section [ sect : approx_tau ] we showed that @xmath92 , under the condition of a sufficiently large @xmath25 .
simulations indicate that the constant of proportionality is just 1 , which means the constant @xmath68 in eq .
( [ c_1def ] ) equals @xmath93 .
if this hypothesis is true , eq . ( [ p_tfinal ] ) becomes @xmath94 this means , a rescaling of the form @xmath95 should yield a functional dependency @xmath96 .
[ fig_scaledp ] shows the results of such a rescaling for different @xmath97 and @xmath25 .
all the data - points lie exactly on top of each other in the region where the statistics is good enough ( about @xmath98 ) .
we find that for @xmath97 up to 0.1 , eq .
( [ eq : conj1 ] ) is very accurate for @xmath25 between about 0.1 and 1 . with smaller @xmath97 , eq .
( [ eq : conj1 ] ) holds even for much smaller @xmath25 .
the situation becomes more complicated if we study the sequence of aftershocks caused by an initial event of finite size .
the probability to find an event of size @xmath32 after some initial event of size @xmath99 decreases with a power - law , but the exponent is not a simple function of @xmath25 .
rather , it depends on @xmath26 as well . in fig .
[ fig_finites0 ] we have displayed the results of a measurement with @xmath100 and several different @xmath26 , ranging from @xmath101 to 1 . the curve for @xmath31 has been obtained with the method described at the beginning of this section .
we find that the aftershocks decay pattern for @xmath102 continuously approaches the one for @xmath31 as @xmath103 .
this shows that it is indeed justified to study the system in the aftermath of an infinite initial event and then to extrapolate to finite but large initial events .
note that in fig .
[ fig_finites0 ] , @xmath25 is so small that eq .
( [ eq : conj1 ] ) does not hold anymore .
sneppen and newman have argued that the decay pattern of the aftershocks is independent of the respective stress distribution .
our results do not support that . despite the fact that the exponent of the power - law seems to be independent of @xmath97 in the case of exponential stress , as we could show above
, the exponent is not independent of the functional dependency of the stress distribution .
if we impose , for example , gaussian stress with mean 0 and variance @xmath97 , we find ( fig . [ fig_gauss.1 ] ) exponents larger than 1 for moderate @xmath25 .
we do event find a qualitatively new behaviour of the system .
the exponent is getting larger with increasing @xmath25 , as opposed to the exponent getting smaller for exponential stress .
of course , this can only be true for intermediate @xmath25 .
ultimately , we must have @xmath81 for @xmath80 . finally , we present some results about systems with finite growth
. in these systems , there exists some competitive dynamics between the removal of agents with small thresholds through stress and their regrowth .
aftershocks appear in the infinite growth model because the reorganization event leaves a larger proportion of agents in the region of small thresholds , thus increasing the probability for succeding large events . in the finite growth version , on the contrary , this can happen only if the regrowth is faster than the constant removal of agents with small thresholds through average stress .
if the regrowth is too slow , the probability to find large events actually is decreased in the aftermath of an initial large event .
the interplay between these two mechanisms is shown in fig .
[ fig_grow ] .
the regrowth of species is done according to eq .
( [ eq : logistic_growth ] ) .
for a small growth - rate @xmath21 , the probability to find aftershocks is reduced significantly , and it aproaches the equilibrium value after about 100 time - steps . with larger @xmath21 ,
the probability @xmath33 increases in time until a maximum well above the equilibrium level is reached , and then decreases again .
the maximum moves to the left to earlier times @xmath0 with increasing @xmath21 .
when @xmath21 is so large that the maximum coincides with the post initial event , the original power - law is restored .
note that , as in the case of infinite growth , the measurement depends on the choice of the parameters @xmath26 and @xmath25 .
consequently , with a different set of parameters the curves will look different , and the maximum will appear at a different time .
nevertheless , we find that the qualitative behaviour is not altered if we change @xmath26 or @xmath25 . instead of logistic growth
, we can also think of linear growth , i.e.@xmath104 , where @xmath21 is again the growth rate . in order to keep the system finite ,
we stop the regrowth whenever the system size @xmath7 exceedes the maximal system size @xmath20 .
in such a system , aftershocks can be seen for much smaller growth rates ( fig .
[ fig_growl ] ) .
note that apart from the growth rate , all other settings are identical in fig .
[ fig_grow ] and fig .
[ fig_growl ] .
linear growth refills the system much quicker than logistic growth with the same growth rate .
therefore the time intervall in which aftershocks are suppressed is much shorter for linear growth .
we could show in the present paper that the decay pattern of the aftershock events depends on the details of the measurement .
although the qualitative features remain the same for different parameters @xmath26 and @xmath25 ( e.g. a power - law decrease in the infinite - growth version ) , the quantitative features vary to a large extend .
the exponent @xmath2 of the power - law is significantly affected by an alteration of @xmath26 or @xmath25 .
therefore the measurement - procedure proposed by sneppen and newman can reveal the complex structure of aftershock - events only if the change of the measured decay pattern with varying @xmath26 and @xmath25 is recorded over a reasonably large intervall of different values .
this should be considered in a possible comparison between the aftershocks decay pattern from a model and from experimental data .
a more in - depth analysis could probably be achieved with the formalism of multifractality ( see e.g. @xcite ) .
we found the aftershocks decay pattern to vary with different stress distributions .
this is in clear contrast to sneppen and newman .
they reported a power - law with exponent @xmath105 for the infinite growth version , independent of the respective stress distribution they used .
the question remains why sneppen and newman measured such an exponent in all their simulations .
the answer to this lies in the fact that they did only simulations with @xmath102 .
for the reasons explained at the beginning of section [ sec : numerical_results ] , under this condition one has to choose a relatively small @xmath26 , and accordingly , a very small @xmath25 .
this causes the measurement almost inevitably to lie in the intermediate region between the limiting cases of sec .
[ sect : limiting_cases ] . in this region , for the most reasonable stress distributions and a large array of different values for @xmath26 and @xmath25
, we find indeed exponents around 1 .
the application of coherent - noise models to earthquakes has been discussed in @xcite .
two very important observations regarding earthquakes , the gutenberg - richter law @xcite and omori s law @xcite , can be explained easily with a coherent - noise model .
the gutenberg - richter law states that earthquake magnitudes are distributed according to a power - law .
omori s law , which interests us here , is a similar statement for the aftershocks decay pattern of earthquakes . in the data from real earthquakes , the number of events larger than a certain magnitude @xmath106 decreases as @xmath1 in the aftermath of a large initial earthquake .
consequently , we can only apply infinite - growth coherent - noise models to earthquakes .
but this is certainly no drawback , since we expect the thresholds against movement at various points along a fault ( with which we identify the agents of the coherent - noise model ) to reorganize almost instantaneously during an earthquake .
the exponent @xmath2 is not universal , but can vary significantly , e.g. , in @xcite from @xmath107 to @xmath108 .
this would cause problems if the statement was true that for coherent - noise models we have @xmath109 .
but as we could show above , the exponent can assume values in the observed range , depending on the stress distribution , the size of the initial event , and the lower cut - off ( which we called @xmath106 for earthquakes and @xmath25 throughout the rest of the paper ) .
for a further comparison , it should be interesting to study the dependency of the exponent @xmath2 on variation of the cut - off @xmath106 in real data .
we are only aware of a single work where that has been done @xcite .
interestingly , the authors do not find a clear dependency @xmath110 .
nevertheless , this is not a strong evidence against coherent - noise models , since the aftershock series analysed in @xcite consists mainly of very large earthquakes with magnitude between 6 and about 8 , which does not allow a sufficient variation of @xmath106 .
statistical variations in the exponent @xmath2 are probably larger for this aftershock series than the possible variations because of an assumed @xmath110 dependency .
numerical simulations of the finite - growth version have revealed a much more complex structure of aftershock events than present in the infinite - growth version .
the competition between regrowth of agents and agent removal through reorganization events leads to a pattern where the probability to find events after an initial large event is suppressed for short times , enhanced for intermediate times and then falls off to the background level for long times .
this observation can be important for the application of coherent - noise models to biological extinction .
it might be possible to identify a time of reduced and a time of enhanced extinction activity in the aftermath of a mass extinction event in the fossil record .
this would be a good indication for biological extinction to be dominated by external influences ( coherent - noise point of view ) rather than by coevolution ( soc point of view ) .
we thank mark newman for interesting conversations about coherent - noise models .
in this appendix we are interested in the average state a coherent - noise system will be found in several time - steps after some initial state with threshold distribution @xmath111 .
our calculations will lead to a rederivation of the master - equation for coherent - noise systems .
although a master - equation has already been given for the infinite - growth version and has been generalized to the finite - growth version , these master - equations have not been derived in a stringent way , but just have been written down intuitively .
our calculation will confirm the main terms of the previously used equations , but we will find an additional correcting term that becomes important for large @xmath15 . consider the case of infinite growth . at time
@xmath112 the threshold - distribution may be @xmath111 .
we construct the distribution @xmath113 , which is the distribution that arises at time @xmath114 if a stress @xmath11 is generated at time @xmath112 . a stress
@xmath11 will cause a proportion @xmath115 of the agents to move .
we have to distinguish two regions . for @xmath116 ,
all agents are removed .
then they are redistributed according to @xmath117 .
a small fraction @xmath15 of the agents is then mutated , which results in @xmath118 for @xmath119 , the redistribution of the agents gives @xmath120 . with the subsequent mutation
we obtain : @xmath121 we take the average over @xmath11 to get the distribution @xmath122 that will on average be found one time - step after @xmath111 : @xmath123 \notag\\ & \qquad + \int\limits_0^x \!d\eta\ , p_{\rm stress } ( \eta ) \rho_{t_0 } ( x)(1-f ) \notag\\ & = p_{\rm thresh}(x)\big[f+(1-f)\int\limits_0^\infty \!d\eta\ , p_{\rm stress}(\eta ) \int\limits_0^\eta \rho_{t_0}(x')dx'\big ] \notag\\ & \qquad + \rho_{t_0 } ( x)(1-f)(1-p_{\rm move}(x))\,.\end{aligned}\ ] ] here , @xmath124 is the probability for an agent with threshold @xmath89 to get hit by stress , viz .
@xmath125 . to proceed
further we have to interchange the order of integration in the remaining double integral .
note that @xmath126 , and therefore @xmath127 \notag\\ & \qquad + \rho_{t_0 } ( x)(1-f)(1-p_{\rm move}(x ) ) \notag\\ & = p_{\rm thresh}(x)\big[f+(1-f)\int\limits_0^\infty \!dx'\ , \rho_{t_0}(x')p_{\rm move}(x')\big ] \notag\\ & \qquad + \rho_{t_0 } ( x)(1-f)(1-p_{\rm move}(x ) ) \notag\\ & = p_{\rm thresh}(x)\int\limits_0^\infty \!dx'\,\big ( f+(1-f ) p_{\rm move}(x')\big)\rho_{t_0}(x ' ) \notag\\ & \qquad + \rho_{t_0 } ( x)(1-f)(1-p_{\rm move}(x ) ) \end{aligned}\ ] ] we are thus led to the master - equation @xmath128 where @xmath129 is the normalization constant @xmath130 .
we notice the appearance of the term @xmath131 which was not present in the master - equation used by sneppen and newman .
the term arises if one takes into account the fact that the agents which are hit by stress get new thresholds before the mutation takes place .
therefore every agent with threshold @xmath89 has the probability @xmath132 to get two new thresholds in one time - step .
but obviously this is exactly the same as geting only one new threshold .
consequently , the term @xmath131 has to be present to avoid double - counting of those agents which are hit both by stress and by mutation .
nevertheless , this term does not affect the scaling behaviour of the system , because the derivation of the event size distribution in @xcite has been done under the assumption @xmath133 .
( [ master_cont ] ) gives the average state of the system one time - step after the initial state @xmath111 .
if we are interested in the average state @xmath0 time - steps after the initial state , we have to repeat the calculations in eqs .
( [ rhostep])-([master_cont ] ) @xmath0 times .
since all averages in these calculations can be taken independently , this is exactly the same as @xmath0 times iterating the master - equation ( [ master - equation ] ) .
we assume that at time @xmath39 a big event takes place and produces the distribution @xmath75 .
if we apply the master - equation ( [ master - equation ] ) @xmath0 times to this distribution @xmath75 , we will end up with a distribution @xmath47 that tells us the average state of the system at time @xmath0 after the big event .
in the following we will use @xmath134 and write @xmath135 for the normalization constant that appears on the right - hand side of eq .
( [ master - equation ] ) at time @xmath0 .
the average distribution at time @xmath0 then becomes @xmath136 we integrate both sides of eq .
( [ master - t ] ) and find a recursion relation for the constants @xmath135 : @xmath137 all integrals can be calculated analytically for a special choice of the threshold and stress distributions . as threshold distribution ,
we choose the uniform distribution @xmath138 , and as stress distribution we choose the exponential distribution @xmath139 .
furthermore , we assume that the initial event was so large as to span the whole system , i.e. @xmath140 . under the above assumptions there is basically one integral that appears in eq .
( [ at - recursion ] ) , which is @xmath141 and eq .
( [ at - recursion ] ) becomes @xmath142 with the aid of the binomial expansion @xmath143 we find @xmath144 we are now in the position to calculate the probability that an event of size @xmath32 occurs at time @xmath0 after the initial big event .
the minimal stress value @xmath145 that suffices on average to generate such an event is the solution to the equation @xmath146 the corresponding probability is @xmath147 .
we invert this expression and insert it into eq .
( [ eta_mineq ] ) .
the resulting equation determines the probability @xmath33 : @xmath148 the integrals that appear after inserting @xmath47 into the above equation are very similar to the integral @xmath149 defined in eq .
( [ i_ndef ] ) .
we define @xmath150 this integral can be taken in the same fashion as the calculation of @xmath149 in eq .
( [ i_ncomp ] ) .
we find @xmath151 eq . ( [ p - eq ] ) now becomes @xmath152 all the quantities which appear in this equation are given above in analytic form .
therefore solving eq .
( [ p - eq2 ] ) is simply a problem of root - finding . with a computer - algebra program such as mathematica
, the recursion relation for the constants @xmath153 as well as the sums that appear in the quantities @xmath149 and @xmath154 can be evaluated analytically if we restrict ourselves to moderate @xmath0 .
then the only numerical calculation involved in the computation of @xmath33 is the calculation of the root of eq .
( [ p - eq2 ] ) .
99 ma zongjin , fu zhengxiang , zhang yingzhen , wang chengmin , zhang guomin , liu defu , _ earthquake prediction : nine major earthquakes in china ( 1966 - 1976 ) _ , seismological press beijing and springer - verlag berlin heidelberg , 1990 .
v. frette , k. chistensen , a. malthe - srenssen , j. feder , t. jssang , and p. meakin , nature * 379 * , 49 ( 1996 ) . c. j. rhodes and r. m. anderson , nature * 381 * , 600 - 602 ( 1996 ) .
m. takayasu , h. takayasu , and t. sato , physica a*233 * , 824 ( 1996 ) .
r. v. sol and j. bascompte , proc .
london b*263 * , 161 ( 1996 ) .
t. h. keitt and p. a. marquet , j. theor . biol . * 182 * , 161 ( 1996 ) .
p. dutta and p. m. horn , rev .
phys . * 53 * , 497 ( 1981 ) .
m. e. fischer , rev .
phys . * 46 * , 597 ( 1974 ) .
k. g. wilson , rev . of mod .
phys . * 47 * , 773 ( 1975 ) .
p. bak , c. tang , and k. wiesenfeld , phys .
* 59 * , 381 ( 1987 ) .
p. bak , c. tang , and k. wiesenfeld , phys .
a * 38 * , 364 ( 1988 ) .
p. bak and m. creutz , in _ fractals in science _ , ed . by a. bunde and s. havlin , ( springer , berlin , 1994 ) .
p. bak and k. sneppen , phys .
* 71 * , 4083 .
s. clar , b. drossel , and f. schwabl , j. phys .
: cond . mat .
* 8 * , 6803 ( 1996 ) .
b. drossel and f. schwabl , phys .
* 69 * , 1629 ( 1992 ) .
z. olami , h. j. s. feder , and k. christensen , phys .
68 , 1244 ( 1992 ) .
m. paczuski , s. maslov , and p. bak , phys .
e*53 * , 414 ( 1996 ) .
a. vespignani and s. zapperi , phys .
78 , 4793 ( 1997 ) .
p. grassberger and a. de la torre , ann
. phys . * 122 * , 373 ( 1979 ) .
e. w. montroll and m. f. schlessinger , proc .
usa 79 , 3380 - 3383 ( 1982 ) .
m. levy and s. solomon , int .
j. of mod .
c * 7 * 595 ( 1996 ) .
s. solomon and m. levy , int .
j. of mod .
c*7 * 745 ( 1996 ) .
d. sornette and r. cont , j. phys .
i france * 7 * , 431 ( 1997 ) .
d. sornette , cond - mat/9708231 .
d. sornette , physica a ( in press ) , also cond - mat/9709101 .
r. mantegna and h. e. stanley , nature * 376 * , 46 ( 1995 ) m. e. j. newman and k. sneppen , phys .
e * 54 * , 6226 ( 1996 ) . k. sneppen and m. e. j. newman , physica d*110 * , 209 ( 1997 ) .
m. e. j. newman , proc .
london b*263 * , 1605 ( 1996 ) .
m. e. j. newman , physica d*107 * , 292 ( 1997 ) . m. e. j. newman , j. theor . biol .
( in press ) , also adap - org/9702003 . c. wilke and t. martinetz , phys . rev . e*56 * , 7128 ( 1997 ) .
r. pastor - satorras , phys .
e*56 * , 5284 ( 1997 ) .
b. gutenberg and c. f. richter , _ seismicity of the earth _
, princeton university press , 1954 .
b. gutenberg and c. f. richter , ann .
di geofis .
* 9 * , 1 ( 1956 ) .
f. omori , journal of the college of science , imperial university of tokyo * 7 * , 111 ( 1894 ) . h. m. hastings , g. sugihara , _ fractals : a user s guide for the natural sciences _ , oxford university press , 1993 . | the decay pattern of aftershocks in the so - called coherent - noise models [ m. e. j. newman and k. sneppen , phys .
rev .
e*54 * , 6226 ( 1996 ) ] is studied in detail .
analytical and numerical results show that the probability to find a large event at time @xmath0 after an initial major event decreases as @xmath1 for small @xmath0 , with the exponent @xmath2 ranging from 0 to values well above 1 .
this is in contrast to sneppen und newman , who stated that the exponent is about 1 , independent of the microscopic details of the simulation .
numerical simulations of an extended model [ c. wilke , t. martinetz , phys . rev .
e*56 * , 7128 ( 1997 ) ] show that the power - law is only a generic feature of the original dynamics and does not necessarily appear in a more general context .
finally , the implications of the results to the modeling of earthquakes are discussed . , , |
the spectral dimension @xmath0 of quantum geometries has been a rather hot topic in recent quantum gravity research .
@xmath0 is one of the indicators of topological and geometric properties of quantum spacetime as described by different approaches , as well as a way to check that an effective semi - classical spacetime is obtained in appropriate sectors of the theory .
one of the main goals of any such analysis is to prove that , among all the states and histories that appear in the theory ( most of which far from describing any smooth spacetime geometry ) , configurations which do so are either dominant or the approximate result of averaging over the others .
a further goal is to study whether the effective dimension of spacetime remains the same at different scales or if it is subject to a dynamical reduction in the ultraviolet regime @xcite . in dynamical triangulations
@xcite , the spectral dimension has been important as one of the few observables available to classify phases of spacetime ensembles . in particular , in _ causal dynamical triangulations _ ( cdt ) in four dimensions it was found that , while very close to the spacetime topological dimension @xmath1 at large scales , at least in the geometric phase , it takes smaller values at short distances , approaching the value @xmath2 in the ultraviolet limit @xcite .
similar results are obtained in three dimensions @xcite .
two is , of course , an interesting value for the effective dimension in the ultraviolet , because gravity is perturbatively renormalizable in two dimensions . in fact , also in the asymptotic safety scenario @xcite one can find a reduction of the spectral dimension from the topological dimension @xmath3 to half of it under the renormalization group flow using analytical methods @xcite .
other formalisms manifest similar behaviour .
lifshitz gravity has a built - in dynamical dimensional reduction , due to the defining anisotropic scaling @xcite .
modified dispersion relations in non - commutative spacetimes @xcite , super - renormalizable non - local gravity @xcite , black - hole effective physics @xcite and other scenarios ( including string theory @xcite ) give rise to a dimensional flow as well .
apart from the intrinsic interest in this effect , its occurrence in approaches to quantum gravity which are otherwise hard to compare has raised hope that the spectral dimension might help to relate such approaches to one another , or at least to confront their results @xcite .
although the spectral dimension is a rather coarse tool in comparison with more refined geometric indicators @xcite , its determination in full - fledged quantum geometries is still a technically challenging problem .
all the above considerations gave good reasons to investigate the spectral dimension also in loop quantum gravity ( lqg ) @xcite .
using the spectrum of the area operator in lqg to guess an effective dispersion relation for propagating particles , some evidence was found for a dimensional flow of kinematical lqg states @xcite and their spacetime spin - foam dynamics @xcite . in the present work ,
we go beyond the arguments used in previous analyses and investigate the spectral dimension of lqg states in detail , at least at the level of kinematical states .
in particular , we will study the spectral dimension for states defined on large graphs or complexes .
lqg differs from all the above - mentioned approaches in that the degrees of freedom of the theory are discrete geometric data , i.e. , a combination of holonomies , fluxes or lorentz group representations @xcite associated with combinatorial structures such as graphs and cellular complexes .
this combination poses new challenges .
so far , the spectral dimension has been investigated in more or less traditional smooth settings , in terms of either modified differential structures ( multi - scale spacetimes @xcite ) or modified dispersion relations on a smooth geometric manifold ( asymptotic safety , where anomalous scaling is a general feature at the non - gaussian fixed point ; non - commutative spacetimes , where modified dispersion relations are a consequence of a deformed poincar symmetry ; hoava lifshitz gravity , where the laplacian is multi - scale by construction ; and so on ) . also previous work on the lqg case @xcite has been limited to an approximation in which the scaling of the laplacian is first extracted from the area spectrum and then effectively treated as continuous . in other approaches , the setting is purely combinatorial : such is the case of cdt ( the dynamical degrees of freedom are equilateral triangulations ) or graph models @xcite .
therefore , the first important step is a definition of the spectral dimension on the type of discrete geometries lqg is built on .
this preliminary work has already been carried out in @xcite , where we have generalized discrete exterior calculus @xcite to be applicable to such discrete geometries , and we have given explicit expressions for the laplacian ( the basic geometric operator the spectral dimension depends on ) in terms of the variables relevant for lqg . unless stated otherwise , from now on by ` spectral dimension ' we always mean the one _ of spatial ( not spacetime ) geometries_. also , by _ quantum spectral dimension _ of a quantum state of geometry we indicate the spatial spectral dimension as derived from the expectation value of the heat kernel trace on the quantum geometry . in the usual interpretation
, this heat kernel is the solution of the diffusion equation for a test pointwise particle propagating on the quantum geometry , where the differential variable is an abstract ` diffusion time ' @xmath4 .
one can employ a different and perhaps more convincing interpretation @xcite , where the diffusion equation is a running equation with respect to the resolution @xmath5 at which geometry is probed .
we will explain why , in the case of discrete quantum gravity , for a well - defined notion of quantum spectral dimension the quantum expectation value has to be taken at the level of the heat kernel , and not directly of the laplacian .
the combination of combinatorial discreteness and additional ( pre-)geometric data obviously must result in an interplay of their respective effects . to the best of our knowledge ,
the latter have never been classified in the literature .
a major aim of this paper is to understand this interplay in the case of kinematical lqg states .
we first study systematically the topological and geometric effects in the spectral dimension of smooth spheres and tori as well as the discreteness effects of combinatorial complexes with geometric realizations as ( hyper-)cubulations or triangulations of these .
this groundwork having been done , we can then compare the spectral dimension of quantum states with that of the discrete geometries they are semi - classical approximations or superpositions of . as a general tendency
, we find that the effect of the underlying combinatorics dominates and that a relatively large size of the base complexes is needed for the spectral dimension to behave as a spatial dimension at all .
to construct explicit lqg states on large complexes and calculate the corresponding expectation values of global observables is a major and seldom accepted challenge . in the case of the spectral dimension , however , this becomes inevitable . as a line of attack
, we set up a way to numerically construct large abstract simplicial complexes and define geometries thereon .
we follow @xcite to derive their combinatorial properties needed for studying the action of the laplacian operator . while the preparatory work on the classical geometries is done in arbitrary dimensions , we restrict the quantum analysis to kinematical states of ( 2 + 1)-dimensional euclidean lqg .
there are two reasons for this restriction .
the first is merely technical .
the number of degrees of freedom for given assignments of algebraic data grows with the combinatorics of the complex as @xmath6 , where @xmath7 and @xmath8 are , respectively , the topological dimension and the size of the combinatorial complex .
the laplacian operator defined thereon is an @xmath9 matrix , which makes the computational effort more severe as @xmath7 increases .
thus , for a given @xmath8 one can calculate much larger @xmath10 discrete geometries than in @xmath11 . on the other hand ,
a sufficient combinatorial size turns out to be crucial for the physical interpretation of the spectral dimension as a notion of positive - definite dimension somewhere close to @xmath7 .
this limitation could be easily overcome by using a more powerful computational environment for numerical analysis .
the second reason lies in the structure of quantum gravity itself . in 2 + 1 dimensions ,
the spin - network basis of lqg simultaneously diagonalizes all edge - length operators , permitting a straightforward definition of the laplacian and a direct identification of deviations of the quantum spectral dimension from its classical counterpart as quantum corrections . in 3 + 1 dimensions
, one would have to deal further with effects due to the non - commutativity of the full set of geometric operators needed to define the laplacian , as well as with the role of non - geometric configurations which we know are present in the standard @xmath12-based hilbert space of the theory .
its construction thus becomes much more involved and it is much harder to be implemented efficiently for numerical calculations .
while these challenges do not pose any obstacle in principle , we prefer to concentrate here on the more straightforward 2 + 1 case .
once the general influence of quantum geometric effects ( discreteness , quantum fluctuations , and so on ) on the spectral dimension is understood , one can move on to study the additional complications that are present in the four - dimensional case .
this paper is quite explorative in nature .
there are no constraints on the kind of combinatorial manifolds lqg states could be defined on .
thus , the only guidance may come from classical geometries one might want to approximate semi - classically by coherent states . furthermore , there is no unique way to parametrize the ` quantum - ness ' of states .
in fact , to search for genuine quantum effects of lqg states , it is not enough to pick one state and look only at the value of its spectral dimension at small scales , as is usually done when some kind of dimensional flow is expected .
instead , we compute the spectral dimension for those states we consider as ` more quantum ' .
these are coherent spin - network states peaked at smaller representation labels and with larger spread and , more generally , states highly randomized in the intrinsic geometry or superposed with respect to various geometric degrees of freedom .
moreover , in order to explore quantum effects of the combinatorial structure and to compare with cdt , we consider superpositions of these structures as well . in section [ sec : dsriem ] , the spectral dimension is introduced by the common definition on riemannian manifolds .
we present then a definition on discrete geometries .
this is the basis whereupon to define the quantum spectral dimension in general , and explicitly in the case of ( 2 + 1)-dimensional lqg , which we review . in section [ sec : topo ] ,
geometric and topological features are discussed , both analytically and numerically , in particular for the simple examples of spheres and tori . in section [ sec : dsdiscrete ] , we study the features of the spectral dimension on discrete structures , exemplified through various cases . these preliminary considerations are carried out also in higher dimensions .
this part of the paper classifies phenomena which can appear in all contexts where a discrete spacetime structure is employed , both in analytic and numerical applications ; hence , it may be of interest also for the general reader not specialized in quantum gravity . in section [ sec : dskinematics ] ,
the expectation value of kinematical quantum space states in 2 + 1 lqg is analyzed . here
the focus is on investigating the relevance , for the spectral dimension , of the geometric data additional to the geometry coming from purely combinatorial structures .
the spectral dimension of semi - classical states is compared with the classical one to identify quantum corrections , which turn out to be rather small .
furthermore , superpositions of coherent states peaked at different spins will be discussed ; the spectral dimension of these configurations will be found via a certain averaging procedure .
complementary to these superpositions in geometry , in section [ sec : sumdis ] we analyze the spectral dimension of superpositions of states on two classes of torus triangulations .
it turns out that the quantum spectral dimension in lqg is more sensitive to the underlying graph structure of states than to the associated geometric data .
in this section , we start by introducing the concept of spectral dimension in the usual context of riemannian manifolds .
we show how this definition can be transferred to discrete pseudo - manifolds , i.e. , abstract finite simplicial complexes with manifold properties .
it is straightforward then to define the quantum spectral dimension in terms of expectation values on quantum states . for kinematical states of lqg in 2 + 1 dimensions
, we argue that this expression is well defined under rather mild conditions .
the definition of the spectral dimension is related to the scaling of the propagator of a massless test particle field ( or , equivalently , to a diffusion process on the spatial manifold ) .
@xmath0 is defined via the heat kernel @xmath13 , which describes the diffusion of the particle in fictitious ` time ' @xmath4 from point @xmath14 to @xmath15 .
@xmath16 is the solution of the diffusion equation @xmath17 where @xmath18 is the laplacian with respect to the underlying space .
actually , @xmath4 has dimension of a squared length since we set the diffusion parameter in front of the laplacian to 1 . alternatively ,
equation ( [ eq : heatkereq ] ) can be regarded as a running equation establishing how much we can localize the point particle , placed at some point @xmath14 , if we probe the geometry with resolution @xmath5 @xcite .
the length scale @xmath19 represents the minimal detectable separation between points .
while in the diffusion interpretation @xmath16 is the probability to find the particle at point @xmath15 after diffusing for some time @xmath4 from point @xmath14 , in the ` resolution ' interpretation @xmath16 is the probability to see the particle in a neighborhood of point @xmath15 of size @xmath20 if the geometry is probed with resolution @xmath5 . for infinite resolution ( @xmath21 ) and a delta initial condition , the particle is found where it was actually placed ( at @xmath14 ) , while the smaller the resolution ( i.e. , larger diffusion time ) the wider the region where it can be seen .
there is no practical difference between the two interpretations but the latter is more suitable in diffeomorphism - invariant theories where the notion of diffusion time makes no sense ( while a lorentz - invariant scale does ) . a formal solution to the problem is provided by @xmath22 where the initial condition is incorporated as the definition of the inner product @xmath23 of particle states .
usually , the particle is taken to be concentrated at @xmath21 in the point @xmath14 , but it could start with other distributions . in the usual context of a riemannian @xmath7-manifold @xmath24 , the beltrami laplace operator is ( on 0-forms the connection laplacian and hodge laplacian agree ) @xmath25 with the initial condition @xmath26 the heat kernel is @xmath27 with @xmath28 the geodesic distance .
on @xmath24 there is a well - defined ( the so called seeley
dewitt ) expansion of the heat kernel around the solution on a flat manifold @xmath29 with the metric distance being the euclidean norm @xmath30 , @xmath31 where the @xmath32 are computable in terms of geometric invariants @xcite .
the spectral dimension probed by the particle can now be defined as the scaling of the trace of the heat kernel @xmath33 called return probability .
while in the flat case @xmath34 , where @xmath35 , the general heat - kernel expansion yields a power series @xmath36 where @xmath37 and the coefficients @xmath38 consist of geometric invariants averaged over the manifold .
the spectral dimension @xmath0 can be implicitly defined as the scaling of the heat trace in some limit of @xmath4 , @xmath39 such that @xmath40 for the flat case , while in general there are corrections related to global properties described by the geometric invariants @xmath41 .
the implicit definition ( [ eq : dsimplicit ] ) can be generalized as @xmath42 by definition ( via heat trace @xmath43 , heat kernel @xmath16 and laplace operator @xmath18 ) the spectral dimension is a functional of the geometry @xmath44 $ ] on the given manifold @xmath45 . in most of the literature on quantum gravity , equation ( [ eq : heatkereq ] )
is taken on euclideanized spacetime but , as stressed in the introduction , we will strictly confine our analysis to _ spatial _ geometries . consequently , @xmath11 is the case that fits large scale , semiclassical observations . from a quantum gravity perspective
, the spectral dimension is particularly interesting because it can be rather easily extended to discrete geometries as well .
discrete geometries constitute the degrees of freedom of lqg , spin foams and group field theory ( and of cdt in the purely combinatorial case , that is when the graph distance is chosen to specify completely the geometry ) .
let us briefly recapitulate the framework of discrete calculus needed for the definition of the spectral dimension on discrete geometries provided in @xcite .
in the most general context we deal with combinatorial complexes , in particular abstract finite simplicial @xmath7-pseudo - manifolds @xmath46 , with additional ( pre-)geometric data , i.e. , volume variables associated with simplices or cells . if @xmath46 has a geometric realization @xmath47 on some metric space , it can be thought of as the triangulation of a smooth geometry @xmath24 . but this needs not be the case . in this
setting , there is a definition of the laplace operator in terms of discrete exterior calculus , giving a precise meaning to the formal expression ( [ eq : kformal ] ) . to this end , we take the diffusing scalar test particle field @xmath48 as a field living on the @xmath7-simplices @xmath49 , or equivalently on the vertices of the combinatorially dual complex . then , the action of the laplacian on @xmath48 evaluated at some simplex @xmath49 takes the form @xmath50 where the sum runs over neighboring simplices and @xmath51 , @xmath52@xmath53 and @xmath54 are , respectively , the @xmath7-volume of @xmath49 , the @xmath55-volume of the common face and a dual length through that face . by definition , this laplacian has most of the properties of the continuum laplacian .
that is , it obeys the null condition , it is self - adjoint ( which relates to the symmetrizability of its matrix ) and it is local , i.e. , the action on a given position depends only on neighboring positions . in @xcite
we have also argued for the convergence to the continuum laplacian under refinement of triangulations .
nevertheless , there are properties which are further dependent on the definition of the volumes .
we have argued that a definition in terms of a barycentric dual interpretation of the volumes should be preferred to a circumcentric one in the field theoretic context because only the former preserves positivity of the laplacian .
for this reason , we will adhere to this convention throughout this paper .
if we further assume that the volumes associated with the finite simplicial complex @xmath56 are finite and non - degenerate , in particular non - vanishing , the matrix elements of the laplacian are finite and well defined . the laplacian is then diagonalized by its eigenfunctions @xmath57 corresponding to eigenvalues @xmath58 , @xmath59 and these form a complete orthonormal basis defining momentum space .
using the transformation to momentum space the heat trace of @xmath48 turns our to be @xcite @xmath60 since @xmath18 is symmetrizable , for real geometric volume coefficients in equation ( [ eq : dislap ] ) the spectrum spec and thus the heat trace are real valued . in the trivial case of constant volumes over the complex ,
e.g. , equilateral triangulations , the laplacian is just proportional to the combinatorial graph laplacian of the dual 1-skeleton of the complex @xmath46 . in a quantum theory of geometry ,
the metric or other classically equivalent variables are promoted to operators .
the spectral dimension as an observable is a functional of these operators and we are now interested in its expectation value for a given state @xmath61 of space geometry ( using the heat trace ( [ eq : heattraceunnormed ] ) ) .
one possibility is to take the heat trace as the primary observable to be quantized .
formally , @xmath62 suppose the quantum laplacian @xmath63 , operator on the hilbert space of geometry states , is diagonalizable with orthonormal eigenbasis @xmath64 ( later we will take the states @xmath64 to be the spin - network basis of the kinematical lqg hilbert space ) .
then every state can be expanded in this basis , @xmath65 , and the heat trace simplifies : @xmath66 an explicit expression of the spectral dimension ( [ eq : dsquantum ] ) is thus @xmath67 in the following , we will drop the superscript @xmath61 in the spectral dimension .
alternatively , one might consider the laplacian instead of the heat kernel as the fundamental observable for the spectral dimension .
this would give a different result with a product instead of the sum over the eigenbasis .
from @xmath68 there follows @xmath69 on states @xmath70 with discrete support , e.g. , on the simplicial complexes considered here , this formal expression makes sense only if all states live on the same discrete structure .
otherwise , the product of @xmath71 and @xmath72 is not defined , since it would amount to a sum of matrices of different size in the exponent . in theories of quantum gravity with discrete geometries as degrees of freedom , such as lqg
, the full hilbert space usually contains states ( including superpositions ) on various distinct combinatorial structures . for this reason ,
the possibility of defining the quantum spectral dimension as derived from the expectation value of the laplacian has to be excluded .
a similar issue arises when taking the expectation value at the level of the heat kernel @xmath16 , since it is a function of positions on the ( discrete ) manifold .
it can only be avoided by tracing over the manifold .
therefore , we will take the expression in terms of the expectation value of the heat trace , equation ( [ eq : dsquantum ] ) , as the proper definition of quantum spectral dimension .
this is the same as used in cdt @xcite , while in the smooth approaches where this issue does not arise @xcite the expectation value is taken at the level of the laplacian .
as a side remark , notice that using the ` normed ' heat trace @xmath73 instead of @xmath74 would lead to a classically equivalent definition of the spectral dimension ( [ eq : dsexplicit ] ) , as its scaling is independent of proportionality factors .
this definition is very often used for effective continuous spacetimes in quantum gravity , as it gets rid of an infinite volume prefactor in @xmath43 . however , at the quantum level and in the case of superposition states , there is a notable difference . for the volume - normed heat trace ( [ eq : heattracenormed ] ) ,
@xmath75 states are weighted by an additional inverse - volume factor @xmath76 . since for quantum geometries
this factor might have singular behaviour ( for instance , on states of degenerate geometry where the volume operator is not densely defined and has 0 as an eigenvalue ) , we prefer equation ( [ eq : heattraceunnormed ] ) to define the quantum spectral dimension .
lqg states are functions of spatial geometric variables ( holonomies , flux variables or representation labels ) smeared over graphs @xmath77 which we restrict to be dual 1-skeletons of a discrete pseudo - manifolds @xmath46 , to be able to use the discrete laplacian ( [ eq : dislap ] ) .
furthermore , as anticipated and motivated in the introduction , we will restrict here to ( 2 + 1)-dimensional lqg .
now , the elements of @xmath78 are functions of the holonomies on the links of the dual graph @xmath79 of a simplicial complex @xmath46 corresponding to a @xmath10 spatial slice by the dual graphs @xmath77 instead of the corresponding simplicial pseudo - manifolds is allowed because their relation is one - to - one .
this can be easily seen since dual 1-skeletons of simplicial pseudo - manifolds are colorable and colorable graphs are dual to simplicial pseudo - manifolds @xcite . ] . since @xmath77 is 3-regular , i.e.
, all nodes are 3-valent , the gauge - invariant spin - network basis @xmath80 is given by contraction of the link representations @xmath81 with the unique trivial intertwiners , i.e. , clebsch gordan symbols . in 2 + 1
dimensions the spin - network basis diagonalizes the length operators with a spectrum given in terms of the @xmath12 casimir operator @xcite : @xmath82|s\rangle\,,\ ] ] where the scale @xmath83 is set by the planck length and the barbero
immirzi parameter and @xmath84 is a constant dependent on the quantization map chosen for the casimir operator @xcite . if @xmath85 , the clebsch gordan conditions
@xmath86 implicit in the intertwiners on vertices of @xmath77 yield triangle inequalities on the primal complex , @xmath87 this can be seen from the inequality @xmath88 only for @xmath89 there are degenerate spin configurations where the inequality on the lengths is not strict ( e.g. , @xmath90 ) . even in this case
one could obtain the triangle inequalities by restricting to those states in @xmath91 corresponding to non - degenerate configurations .
we do not expect this to be a significant restriction in the calculation of the quantum spectral dimension . in the following we take @xmath92 , although calculations of the spectral dimension with different values indicate that @xmath0 is not so much sensitive to such choice .
in fact , we have tested the @xmath93 case with very similar results .
now , an obvious choice is to take an explicit form of the discrete laplacian in terms of edge - length variables ( see ( * ? ? ?
* equation ( 63 ) ) ) . as an operator , it acts on a spin - network state @xmath80 as @xmath94 while the form of the primal volumes in the laplacian , i.e. , the triangle areas @xmath95 and the edge lengths themselves , is straightforward , there is in principle some freedom in choosing the form of the dual edge lengths as functions of primal lengths .
we choose an expression related to a barycentric dual as argued for above and in @xcite .
let us now check whether the formal expression for the heat - trace expectation value is well defined in this context , i.e. , if the heat trace is a self - adjoint operator .
more precisely , the question is the following : is the heat trace operator @xmath96 self - adjoint on the kinematical hilbert space @xmath91 ?
that is , for @xmath97 , @xmath98 with the above laplacian , the left - hand side is @xmath99 where @xmath100 is the identity of graphs up to automorphisms ( taking the usual inner product of lqg , states defined on distinct graphs are orthogonal ) .
this is equal to the right - hand side of ( [ usef ] ) if , and only if , the spectrum of @xmath74 is real : @xmath101^{*}\delta_{\gamma,\gamma'}\delta_{j , j'}\,.\ ] ] in general , since the entries of @xmath18 consist of lengths and areas , they are real on geometric states where triangle inequalities ( [ trianineq ] ) , or equivalently closure constraints ( which are geometricity conditions ) are satisfied .
then , @xmath102 is real as well .
hence @xmath103 is a good quantum observable on the kinematical states of 2 + 1 gravity with the operator ordering of @xmath104 chosen @xcite such that @xmath85 , which we are considering here .
so far , we have given and discussed the definition of spectral dimension in the smooth , discrete and quantum cases , in particular for quantum states of geometry in 2 + 1 lqg , thereby setting the stage for its calculation .
since the quantum properties of such observable depend on their interplay with classical topological and discreteness effects , in the next two sections we will first investigate these effects separately before addressing the quantum case .
in this section , we calculate the spectral dimension of spheres and tori and discuss the role of topology and geometry in these examples . in preparation for the actual lqg calculations ( where these topologies are chosen ) , it is important to have these effects under control
. at large diffusion scales @xmath4 , the dominant effects in the behaviour of the heat trace and the spectral dimension are due to the topology of @xmath45 .
qualitatively , a compact topology leads to a fall - off of @xmath105 to zero .
the scale at which this happens is related to the geometry ( i.e. , the curvature radii ) . in the light of the heat trace interpretation in terms of random walks , this can be easily understood : the return probability approaches one after diffusion times @xmath4 larger than the circumferences ( closed geodesics ) .
the resolution interpretation is also easy to spell out : as the resolution is lowered , one can not distinguish different points but , due to the limited compact geometry of the set , the particle does not spread over an infinite manifold and it roughly appears localized in the same region . to clarify these topological and geometric features
, we discuss the case of the circle @xmath106 and the generalization to the @xmath7-torus @xmath107 and the @xmath7-sphere @xmath108 . for @xmath107
, we even find analytic solutions in closed form for the heat trace . on a circle of radius @xmath109 ,
the spectrum of the laplacian is @xmath110 , @xmath111 , yielding a heat trace @xmath112 in terms of the third theta function . while @xmath113 for @xmath114 , due to the periodicity there is a fall - off to zero for @xmath115 ( figure [ fig : dssmooth ] ) .
rescaled to volume @xmath116 , for comparison divided by @xmath7 , for @xmath117 ( @xmath118 dashed line , @xmath119 solid , @xmath120 dot - dashed and @xmath121 dotted ) .
the inverse square of the corresponding radii sets the scale of the topological effect . at the same time , it is a comparison with the case of @xmath107 with radii @xmath122 as these tori are equivalent to the case of @xmath106 .
[ fig : dssmooth],width=283 ] the geometry defined by the constant curvature radius @xmath109 governs only the scale at which the topological effect takes place . in that sense
, the fall - off of the spectral dimension can be understood as due to the combination of geometry and topology .
the torus @xmath123 with radii @xmath124 generalizes the case of the circle straightforwardly with spectrum @xmath125 , @xmath126 , such that @xmath127 where @xmath128 is the ( multi - dimensional ) riemann @xmath128-function @xcite .
the spectral dimension of the case where all radii are equal turns out to be just @xmath7 times the spectral dimension of @xmath129 , since @xmath130^{d}.\ ] ] on the other hand , with more geometric parameters given by the constant curvature radii @xmath124 , one can see that the geometry affects not only the scale at which the decay starts , but also accounts for various geometric regimes ( figure [ fig : dst2 ] ) .
if the radii are ordered as @xmath131 , the spectral dimension is constantly the topological dimension for @xmath132 and zero for @xmath133 . in intermediate regimes
, it can have plateaux at heights corresponding to intermediate integer dimensions , if the radii are of sufficiently different order .
this can be easily understood : when @xmath134 radii are much smaller than a given scale , the @xmath7-torus effectively appears as a @xmath135-torus for the diffusion process at that scale . with various geometries , @xmath136 , @xmath137 , @xmath138 ( solid
, dash - dotted and dashed curve),width=283 ] the spectrum of the laplacian acting on functions on the sphere @xmath139 of radius @xmath109 is proportional to the @xmath12 casimir , @xmath140 , for @xmath141 with @xmath142-fold degeneracy , leading to a heat trace @xcite @xmath143 more generally , for the @xmath7-sphere @xmath108 , @xmath144 , the spectrum is given by the eigenvalues @xmath145 with multiplicities such that @xmath146\,\rme^{-\frac{j(j+d-1)}{r^{2}}\tau}\,.\ ] ] the spectral dimension of @xmath108 has a slower fall - off in comparison with the torus @xmath107 , with @xmath147 .
thus , @xmath7-spheres are examples where the topological dimension is only obtained in the limit @xmath148 .
in contrast with the torus , spheres do not exhibit any multi - scale behaviour since they are governed by only one geometric parameter . in this section
, we have discussed the generic effects of spherical and toroidal topology and of the geometric curvature parameters .
all compact topologies have a spectral dimension going to zero at scales of the order of the curvature radii . below this scale
, @xmath0 agrees with the topological dimension in the case of the tori , while for spheres such an accordance holds only in the small - scale ( or infinite resolution ) limit @xmath149 ( figure [ fig : dssmooth ] ) .
furthermore , in general a @xmath7-torus is given by @xmath7 geometric parameters setting the intermediate scales at which the torus has an effective lower - dimensional behaviour ( figure [ fig : dst2 ] ) .
now we consider classical effects of space discreteness in the spectral dimension . to this aim ,
we calculate the spectral dimension of various examples of hyper - cubulations and triangulations .
this is preliminary to the quantum case in a twofold way .
first , it is important for the distinction of effects of the underlying discrete structure from additional quantum effects .
second , from a practical perspective , one can then choose those complexes with smaller and more controllable discreteness features .
they are more appropriate for analyzing quantum effects of states thereon .
we find that regular torus triangulations are the best candidates for this purpose . as a simple , purely combinatorial case to start with
, we consider finite and infinite hypercubic lattices . on a toroidal lattice @xmath150
, the spectral dimension can be calculated analytically . for the cycle graph @xmath151 , which corresponds to the @xmath152 case @xmath153 ,
the eigenvalues of the laplacian are known to be @xcite @xmath154 such that @xmath155 and @xmath156.\ ] ] using trigonometric relations , a given sum over exponentials of cosines can be further simplified to a sum over hyperbolic cosines ; e.g. , for @xmath157 @xmath158.\ ] ] as in the case of smooth tori , for the general case of arbitrary dimension @xmath7 @xmath159 yielding just a pre - factor of @xmath7 for the spectral dimension : @xmath160 this can be compared with an infinite lattice @xmath161 . for @xmath152 , because of translational symmetry one can consider only the heat kernel @xmath162 for the sites @xmath163 , satisfying @xcite @xmath164 with initial condition @xmath165 .
the solutions are @xcite @xmath166 where @xmath167 is the hyperbolic bessel function .
one can use translational symmetry ( @xmath168 ) to get the return probability @xmath169.\ ] ] therefore , only for @xmath170 the topological dimension @xmath171 is obtained , while in general @xmath172,\label{eq : ds1d}\ ] ] with a maximum of @xmath173 at @xmath174 .
we can generalize the solution of @xcite to @xmath7-dimensional lattices where the heat equation is @xmath175 for a generic lattice labeling @xmath176 and shifts @xmath177 in the @xmath178-direction .
a more useful labeling is a multi - index @xmath179 : @xmath180 the solution is @xmath181 in particular , @xmath182^{d}.\ ] ] the heat trace scales as @xmath7 times the heat trace of the one - dimensional case . the small-@xmath4 behaviour of the finite lattice is exactly the same , as one can easily see numerically ( figure [ fig : dszn ] ) . for infinite @xmath183 lattice , compared with the finite @xmath184 ( dotted curve ) , @xmath185 ( dot - dashed ) and @xmath186 ( dashed).,width=283 ] extending the discrete spectrum ( [ distor ] ) to one parametrized by a continuous parameter @xmath187 in the same interval , @xmath188 the same result
can also be shown by computing the heat trace analytically : @xmath189 the straightforward extension to @xmath7 dimensions is given by @xmath190 so that @xmath191^{d}.\end{aligned}\ ] ] incidentally , the use of a continuous parameter @xmath8 allows one to connect these results to the continuous case , @xmath192 . introducing a lattice spacing @xmath193 ( compare with lattice gauge theory , e.g. , @xcite ) compatible with this spectrum , @xmath194 where now @xmath195 ,
the usual continuum limit is @xmath196 the formul for the heat trace and spectral dimension are modified accordingly .
the type of combinatorial structures that are most relevant for the quantum - gravity applications below are simplicial complexes . in order to avoid having to deal with the difficult problem of constructing quantum - geometry states on infinite spaces , we restrict to simplicial manifolds with compact topology .
obvious triangulations of the 2-sphere are the boundaries of the three triangular platonic solids ( tetrahedron , octahedron , icosahedron ) .
these are the only regular ( i.e. , equilateral ) triangulations of the smooth 2-sphere in terms of non - degenerate simplicial complexes .
the result for @xmath0 are shown in figure [ fig : s2triang ]
. the larger the triangulation , the taller the height of the peak . in particular ,
only the spectral dimension of the icosahedral triangulation can be seen as providing a good approximation for the topological dimension @xmath10 , provided one defines it to coincide with the height of the peak ( and assuming this would become a plateau for larger complexes , as seems to be the case in feasible calculations ) . of equilateral triangulations of @xmath197 in terms of the boundary of the ( triangular ) platonic solids , icosahedron , octahedron and tetrahedron ( solid , dashed and dash - dotted curve ) and of the dipole ( dotted curve ) .
[ fig : s2triang],width=283 ] the extent to which too - small triangulations fail to capture the topological dimension can be seen even more drastically in the case of the degenerate complex , called dipole or ( super)melon , triangulating a @xmath7-sphere in terms of just two simplices .
independently of the dimension @xmath7 , its heat trace is proportional to ( see appendix c in @xcite ) @xmath198 where the only geometric factor @xmath193 can be absorbed into the diffusion parameter .
this yields a spectral dimension @xmath199 from its derivative @xmath200 it can be seen that the maximum is at @xmath201 ( where @xmath202 is the upper branch of the real lambert @xmath203-function ) and has value @xmath204 independent of the parameter @xmath193 . only the position of the maximum is rescaled by @xmath193 .
while there are no further , larger equilateral triangulations of @xmath197 , for the @xmath7-torus @xmath107 there are regular equilateral triangulations of arbitrary combinatorial size ( i.e. , number of vertices ) @xmath205 .
these are obtained from hypercubic lattices via a so called standard triangulation of each hypercube @xcite and are non - degenerate for @xmath206 . in two dimensions ,
these are triangulations of the flat torus with radii ratio @xmath207 ( figure [ fig : t2triang ] ) .
( 6,0 ) ; ( 1,1.722 )
( 7,1.722 ) ; ( 2,3.444 )
( 8,3.444 ) ; ( 3,5.166 )
( 9,5.166 ) ; ( 0,0 )
( 3,5.166 ) ; ( 2,0 )
( 5,5.166 ) ; ( 4,0 )
( 7,5.166 ) ; ( 6,0 )
( 9,5.166 ) ; ( 2,0 )
( 1,1.722 ) ; ( 4,0 )
( 2,3.444 ) ; ( 6,0 )
( 3,5.166 ) ; ( 7,1.722 )
( 5,5.166 ) ; ( 8,3.444 ) ( 7,5.166 ) ; at ( 0,0 ) [ place , label = below:(0,0 ) ] ; at ( 2,0 ) [ place , label = below:(1,0 ) ] ; at ( 4,0 ) [ place , label = below:(2,0 ) ] ; at ( 6,0 ) [ place , label = below:(0,0 ) ] ; at ( 1,1.722 ) [ place , label = left:(0,2 ) ] ; at ( 3,1.722 ) [ place , label=[label distance=1pt]5:(1,2 ) ] ; at ( 5,1.722 ) [ place , label=[label distance=1pt]5:(2,2 ) ] ; at ( 7,1.722 ) [ place , label = right:(0,2 ) ] ; at ( 2,3.444 ) [ place , label = left:(0,1 ) ] ; at ( 4,3.444 ) [ place , label=[label distance=1pt]5:(1,1 ) ] ; at ( 6,3.444 ) [ place , label=[label distance=1pt]5:(2,1 ) ] ; at ( 8,3.444 ) [ place , label = right:(0,1 ) ] ; at ( 3,5.166 ) [ place , label = above:(0,0 ) ] ; at ( 5,5.166 ) [ place , label = above:(1,0 ) ] ; at ( 7,5.166 ) [ place , label = above:(2,0 ) ] ; at ( 9,5.166 ) [ place , label = above:(0,0 ) ] ; when comparing triangulations of different combinatorial size @xmath208 , one can either ( a ) fix the edge lengths to some scale @xmath209 such that the geometric size of the triangulations is growing with the combinatorial size , or one can ( b ) rescale them to @xmath210 according to the combinatorial size to keep the overall geometric size fixed . thus , in the limit @xmath211 the former case gives a triangulated plane @xmath212 , while the latter approximates the smooth flat @xmath213 geometry .
indeed , the calculations in figure [ fig : dst2triang ] ( for various finite @xmath8 ) indicate that the spectral dimension of these triangulations capture both limits .
triangulation with @xmath214 vertices , @xmath215 from left to right ( dotted , dot - dashed , dashed thin , solid curve ) indicating a convergence to the topological dimension for large @xmath4 in the @xmath211 limit , compared with the infinite quadrangular lattice ( dashed thick ) . left plot : ( b ) rescaled triangulations ( @xmath215 from right to left ) indicating a convergence to the smooth @xmath213 ( dashed thick ) for @xmath211 .
[ fig : dst2triang],title="fig:",width=283 ] triangulation with @xmath214 vertices , @xmath215 from left to right ( dotted , dot - dashed , dashed thin , solid curve ) indicating a convergence to the topological dimension for large @xmath4 in the @xmath211 limit , compared with the infinite quadrangular lattice ( dashed thick ) . left plot : ( b ) rescaled triangulations ( @xmath215 from right to left ) indicating a convergence to the smooth @xmath213 ( dashed thick ) for @xmath211 .
[ fig : dst2triang],title="fig:",width=283 ] moreover , these calculations show that the only difference between a quadrangular ( @xmath216 hypercubic ) and a triangular lattice of the plane consists in a qualitatively different discretization effect given by the peak around @xmath217 .
the same analysis can be repeated in higher dimensions . in terms of the standard triangulation of the cube
, there is an equilateral triangulation of the 3-torus with radii ratios @xmath218 and @xmath219 .
the calculations in figure [ fig : dst3triang ] again indicate the correct behaviour in the limit of infinite triangulations .
the discretization effect here is slightly more marked , in that below the relatively small peak there is a small regime of relatively weak slope before the usual steeper fall - off at small @xmath4 sets in .
triangulation in terms of @xmath220 vertices , @xmath221 from left to right ( dotted , dot - dashed , dashed thin , solid curve ) , compared with the cubic lattice ( dashed thick ) .
right : ( b ) rescaled triangulations.[fig : dst3triang],title="fig:",width=283 ] triangulation in terms of @xmath220 vertices , @xmath221 from left to right ( dotted , dot - dashed , dashed thin , solid curve ) , compared with the cubic lattice ( dashed thick ) .
right : ( b ) rescaled triangulations.[fig : dst3triang],title="fig:",width=283 ] an alternative strategy for obtaining combinatorially larger simplicial manifolds of a given topology is to subdivide them .
a natural way to do so is the pachner 1-(@xmath7 + 1 ) move where one @xmath7-simplex is subdivided into @xmath3 simplices by inserting a vertex in the middle of the original simplex and connecting it to all its vertices .
again , concerning the geometric realization of such a combinatorial structure , one can either ( a ) consider it as an equilateral triangulation ( although this is not a triangulation of a torus with flat geometry anymore ) or ( b ) rescale the edge lengths such that it is a triangulation of the flat torus .
one can , for example , consider the vertex @xmath222 inserted by the pachner 1-(@xmath7 + 1 ) move as the barycenter of a @xmath7-simplex with vertex set @xmath223 .
the new edges @xmath224 then need to have squared lengths @xmath225 to preserve the flat geometry approximated by the triangulation ( see @xcite , appendix b.2 ; sums are running over vertices @xmath226 and edges @xmath227 of the simplex not containing the vertex @xmath178 ) .
we compare these two possibilities in two cases .
( i ) first , we subdivide the above @xmath213 triangulation ( @xmath228 applying the 1-(@xmath7 + 1 ) move simultaneously on all triangles ; ( ii ) second , we apply it randomly . 1 . in the first case of global subdivisions ( figure [ fig : dst2subdiv ] ) , considered as equilateral triangulations ,
there is a peak slightly lower than @xmath10 at the diffusion time scale of the size of the triangulation .
but at smaller intermediate scales we find small oscillations around a value of about @xmath229 , obtained after integrating over a period .
it is not surprising that there is a deviation from the topological dimension , since these equilateral triangulations have geometric realizations only in terms of a torus curved at various scales in a specific manner . +
what might be more surprising is that also the complex with barycentrically rescaled edges triangulating the flat torus has a spectral - dimension plot quite different from the equilateral triangulations and cubulations considered above .
there is a more complicated oscillatory behaviour but now at a value around @xmath230 .
furthermore , the fall - off at small @xmath4 is much less steep .
+ the rescaled pachner subdivisions are thus an example of a triangulation which substantially deviates with respect to the spectral dimension in the corresponding continuum geometry , due to the particular combinatorial structure . in particular , in the finite case
there is no plateau at the value of the topological dimension , nor is it expected that the topological dimension be recovered in the large - size limit .
an important lesson one can draw is that the spectral dimension of a triangulation of a given geometry does depend on the combinatorics of the chosen triangulation .
+ global subdivisions of the smallest regular @xmath213-triangulation , @xmath231 ( dotted , dot - dashed , dashed and solid curve ) . left : ( a ) purely combinatorial complex , with a dashed line at @xmath232 .
right : ( b ) complex with rescaled edge lengths triangulating the flat torus , with a dashed line at @xmath233.,title="fig:",width=283 ] global subdivisions of the smallest regular @xmath213-triangulation , @xmath231 ( dotted , dot - dashed , dashed and solid curve ) .
left : ( a ) purely combinatorial complex , with a dashed line at @xmath232 .
right : ( b ) complex with rescaled edge lengths triangulating the flat torus , with a dashed line at @xmath233.,title="fig:",width=283 ] 2 . in the second case
, we construct subdivisions by choosing ( with uniform random distribution ) one triangle to subdivide , performing the subdivision , and then re - iterating the process .
the results ( figure [ fig : dst2subran ] ) hint at some kind of averaging effect . in the equilateral case , the oscillations are washed away and the height of the peak at the volume - size scale seems to be dependent on the particular element chosen in the random ensemble .
the closer is the element to a global subdivision , the more pronounced the peak . + in the rescaled case ,
the regime around @xmath234 is much smaller than in the global subdivided case and the fall - off is even less steep .
this might be explained by the fact that the random subdivision effectively averages over both the regime corresponding to a plateau and the regime of low-@xmath4 fall - off .
-triangulation by @xmath235 single random subdivisions , @xmath231 ( dotted , dot - dashed , dashed and solid curve ) taken ( a ) as equilateral ( left , with dashed line at @xmath232 ) or ( b ) rescaled triangulations of the flat torus ( right , with dashed line at @xmath233 ) .
[ fig : dst2subran],title="fig:",width=283 ] -triangulation by @xmath235 single random subdivisions , @xmath231 ( dotted , dot - dashed , dashed and solid curve ) taken ( a ) as equilateral ( left , with dashed line at @xmath232 ) or ( b ) rescaled triangulations of the flat torus ( right , with dashed line at @xmath233 ) .
[ fig : dst2subran],title="fig:",width=283 ] finally , we have checked a very peculiar element in the random ensemble , namely the repeated subdivision around the same vertex of the triangulation .
this is interesting because it shows that the spectral dimension is very much dependent on the combinatorics , yielding a spectral dimension which could be conjectured to run to @xmath236 in the large - size limit , as suggested by the calculations at finite sizes ( figure [ fig : dst2frac ] ) .
since this property of the example is actually independent of the global structure of the complex , one can have the same result starting with only one triangle .
single subdivisions , @xmath215 , around the same vertex of the @xmath237 torus triangulation ( left ) and of a single triangle ( right ) , both unrescaled . in both cases , @xmath0 is independent of the global structure in the large complex size limit .
( in the right plot , the spectral dimension goes to infinity for large @xmath4 since there is no zero eigenvalue in the spectrum of the laplacian due to the boundary . )
[ fig : dst2frac],title="fig:",width=283 ] single subdivisions , @xmath215 , around the same vertex of the @xmath237 torus triangulation ( left ) and of a single triangle ( right ) , both unrescaled . in both cases , @xmath0 is independent of the global structure in the large complex size limit .
( in the right plot , the spectral dimension goes to infinity for large @xmath4 since there is no zero eigenvalue in the spectrum of the laplacian due to the boundary . )
[ fig : dst2frac],title="fig:",width=283 ] in this section , we have investigated discreteness effects in the spectral dimension for various equilateral lattices . in general , there are three characteristic features .
( a ) a zero spectral dimension at scales below the lattice spacing , @xmath238 , coming from the fact that the test particle feels the discrete spacing of the lattice : for too - short or infinitesimal times @xmath239 , the probe does not have the chance to diffuse from a given node to another .
( b ) a peak larger than the topological dimension at the lattice scale . ( c ) at larger scales
, there is agreement with the smooth spectral dimension in the case of regular toroidal hypercubulations ( figure [ fig : dszn ] ) and triangulations ( figure [ fig : dst2triang ] ) . in particular
, there is a plateau with height close to the topological dimension , the more extended the larger the complex .
this can not be found for spherical equilateral triangulations , which only exist for certain sizes too small to establish a plateau ( figure [ fig : s2triang ] ) .
triangulations obtained by subdivisions via pachner 1-(@xmath7 + 1)-moves do not reproduce the topological dimension neither in the unrescaled nor rescaled case ( figures [ fig : dst2subdiv ] and [ fig : dst2subran ] ) .
thus , already at the classical level the spectral dimension is very sensitive to the precise structure of the discrete manifolds combinatorics .
after these preliminary considerations , we can now calculate the spectral dimension of various states of 2 + 1 euclidean quantum geometry .
first , we want to consider coherent states , that is states which are peaked both on an intrinsic , two - dimensional spatial geometry as well as on its conjugate extrinsic geometry .
the aim is to identify the quantum corrections to the classical spectral dimension and find the dependence of @xmath0 on the parameters of the coherent states . in these cases ,
our results show only qualitative and small deviations from the classical case . in an attempt to uncover stronger quantum effects
, we finally investigate superpositions of coherent states .
the resulting spectral dimension turns out to be an average of the superposed cases , thereby showing a more distinct quantum behaviour . in the lqg literature ,
the starting point to get coherent lqg states is usually group coherent states in the holonomy basis .
coherent states peaked at phase space points @xmath240 can be obtained from the heat kernel @xmath241 on the group @xmath242 with peter
weyl expansion : @xmath243 here @xmath244 refers to the dimension and @xmath245 to the character of the @xmath246 representation labeled by @xmath226 , @xmath49 is a positive spread , @xmath247 .
there are two known constructions for a generalization including a non - trivial @xmath248-dependence : either by analytic continuation to @xmath249 such that @xcite @xmath250 or using the flux representation @xcite where the heat kernel appears to be a ( @xmath251-non - commutative ) gaussian @xmath252(x - y)\propto \rme_{\star}^{-\frac{1}{2\kappa^{2}}\frac{(x - y)^{2}}{2\sigma^2}}\star \rme^{\frac{\rmi}{\kappa}|p(g)|(x - y)}.\ ] ] here , @xmath253 and plane waves @xmath254 use group coordinates @xmath255 \vec{n}$ ] in the usual coordinates in which the group element is parametrized as @xmath256 , with @xmath257 the pauli matrices .
the notation @xmath258 indicates the non - commutative exponential defined as a power series expansion of @xmath259-monomials @xcite . transforming back to group space , this results in just an additional plane wave factor , @xmath260
since we are considering the laplacian operator diagonal in the spin representation , we need the spin expansion of these coherent states . in both cases , there is a limit ( for large enough spins ) in which these can be described @xcite as gaussian - type states peaked at spin representation labels of intrinsic geometry @xmath261 and angles of extrinsic geometry @xmath262 : @xmath263 with spin - network basis coefficients @xmath264 in fact , following @xcite , in the large-@xmath15 approximation , one finds that the @xmath265 can be identified ( up to a factor dependent on @xmath49 ) with the modulus @xmath15 of the flux , and the @xmath266 are angles in the representation of the group elements @xmath267 ( in the plane orthogonal to the flux @xmath15 ) . for the details we refer to @xcite and @xcite respectively , since in the following only the intrinsic curvature
as captured by the @xmath265 will be relevant .
their dependence on @xmath49 does not play a role for fixed @xmath49 or small variations of it , with respect to ( assumed large ) @xmath15 .
the heat - trace expectation value is then @xmath268\mbox{tr}\,\rme^{\tau\delta_{\gamma}(j_{l})},\label{eq : htsuperposed}\end{aligned}\ ] ] from which the spectral dimension is derived according to equation ( [ eq : qds ] ) . as the spatial laplacian does not depend on the extrinsic curvature ,
it is natural that also the phase of the coherent state drops out of the expectation value .
the reason for considering coherent states here is twofold .
first , as they are semi - classical states peaking at classical geometries , it is interesting to check if their properties are reflected in the spectral dimension .
second , if the quantum spectral dimension was comparable with the spectral dimension of the classical geometries peaked at , the difference should be understood as due to quantum corrections , which can be studied in a controlled manner in terms of the parameters of the coherent states .
these parameters are the spins @xmath269 , the classical extrinsic geometries @xmath270 and the spread @xmath49 , as well as the graph @xmath77 they are defined on .
the interpretation of the semi - classical limit in terms of these parameters is rather subtle here .
as far as the spread is concerned , with respect to the whole coherent state @xmath271 there is a value where heisenberg inequalities are minimized , which would be the ` most semi - classical ' case . on the other hand ,
since the spectral dimension on the spatial state is not dependent on the extrinsic curvature , obviously the deviation from the classical spectral dimension vanishes for @xmath272 , i.e. , states sharply peaked at the intrinsic curvature but totally random in the extrinsic one .
nevertheless , these states are highly quantum .
more interesting for our purpose are quantum effects of those states randomizing the intrinsic curvature , that is states with large spread @xmath49 . furthermore , from the physical interpretation of the geometric spectra with corresponding eigenbasis in terms of the spin representations , the limit @xmath273 is often seen as another semi - classical limit . finally , one should not forget the dependence of the states on the underlying graphs .
at least with respect to the spectral dimension , classically we have already seen the important dependence on the combinatorial structure . at the level of kinematical states ,
this dependence is still poorly understood in the literature . in the following
, we will consider the dual 1-skeletons of the previous regular torus triangulations as spin - network graphs @xmath77 .
they are parametrized by the number of nodes @xmath274 and their spectral dimension converges to the topological dimension @xmath7 if they are large or fine enough , i.e. , in the limit @xmath275 ( see section [ sub : triangulations ] ) .
therefore , in such a geometric regime we can interpret the quantum corrections as actual deviations from the topological dimension . accordingly with this regular combinatorial structure
, we will consider states peaked at all equal @xmath276 for @xmath277 .
a direct implementation of equation ( [ eq : qds ] ) is unfeasible .
this is because the number of terms in the quantum sum over representations @xmath278 even with cutoffs @xmath279 and @xmath280 @xmath281 grows exponentially fast with the number of links @xmath282 .
we have already shown that only for large classical triangulations ( e.g. , for @xmath213 of the order @xmath283 ; see figure [ fig : dst2triang ] ) a geometric regime is obtained .
furthermore , although the state amplitude for many spin configurations vanishes due to the clebsch
gordan conditions implicit in the intertwiners , the resulting effective space of spin configurations is highly non - trivial and , in general , not well understood .
alternatively , the quantum sum can be approximated by summing over some number of configuration samples @xmath278 chosen randomly according to the coefficients @xmath284 ( where the norm must be included now to give a proper probability density ) .
if the space of representations is discrete , as for the spins of @xmath12 , this measure needs to be discrete . here we can choose the binomial distribution @xmath285 .
for large enough @xmath261 and @xmath49 , this is in turn well approximated by the the normal distribution @xmath286 : @xmath287 the floor function @xmath288 is needed since the first argument must be an integer .
the variable @xmath289 is a random field dependent on @xmath290 and @xmath49 .
the probability density function associated with it is , by virtue of the approximation ( [ eq : binomdis ] ) , the gaussian profile @xmath291 . using this method ,
we have evaluated coherent spin networks on the @xmath213 triangulations discussed in section [ sub : triangulations ] .
to compare @xmath0 for various peaks @xmath290 , we choose the scale @xmath209 such that @xmath292 ( the scale of the laplacian ) is kept fixed .
that is , we include a rescaling factor @xmath293 : @xmath294\mbox{tr}\,\rme^{\tau r(j)\delta_{\gamma}(j_{l})}.\ ] ] without this rescaling , as in the classical cases , one observes a shift @xmath295 of the @xmath0 plot due to the @xmath296 scaling of the laplacian .
we will include this rescaling factor in the following calculations , mainly to allow for a more direct comparison of the spectral dimension of states .
we start with testing the dependence of coherent states with respect to the spin @xmath290 peaked at .
as expected , the quantum spectral dimension function of these states does not differ much from the classical version : for @xmath297 and @xmath298 , the deviation is at most of order @xmath299 ( figure [ fig : coherjs ] ) .
this is an important consistency check .
note , though , that all the known coherent states are peaked at discrete geometries .
therefore , strictly speaking , it is not the topological dimension but the particular spectral - dimension profile of these discrete geometries to be approximated well by the coherent states . what can be considered as quantum corrections in the spectral dimension is thus the difference between the quantum spectral dimension and the spectral dimension of discrete geometries . for coherent states peaked at @xmath300 ( solid , dashed and dot - dashed curve ) on a regular torus triangulation ( with @xmath301 triangles ) with spread @xmath302 .
right : deviation of @xmath0 from the classical case .
[ fig : coherjs],title="fig:",width=283 ] for coherent states peaked at @xmath300 ( solid , dashed and dot - dashed curve ) on a regular torus triangulation ( with @xmath301 triangles ) with spread @xmath302 .
right : deviation of @xmath0 from the classical case .
[ fig : coherjs],title="fig:",width=283 ] second , we test the dependence of states with spread @xmath49 . since the binomial distribution approximation ( [ eq : binomdis ] ) is only defined for @xmath303 , it is difficult to probe the regime of larger @xmath49 within this method . probing the @xmath49 dependence for @xmath304 , we observe increasing quantum corrections , but still of order @xmath305 ( figure [ fig : cohersigmas ] ) . for coherent states peaked at @xmath306 with spread @xmath307 ( dotted , dot - dashed , dashed and solid curve ) on a regular torus triangulation ( with @xmath308 right triangles ) .
right : deviation of @xmath0 from the classical case.[fig : cohersigmas],title="fig:",width=283 ] for coherent states peaked at @xmath306 with spread @xmath307 ( dotted , dot - dashed , dashed and solid curve ) on a regular torus triangulation ( with @xmath308 right triangles ) .
right : deviation of @xmath0 from the classical case.[fig : cohersigmas],title="fig:",width=283 ] the main challenge when extending the calculations to larger spreads is to deal with the highly non - trivial space of group representations due to the clebsch
gordan conditions , as we noted .
however , there is a very straightforward way to define pure states ( or their superpositions ) by bounding the range of spins to an interval @xmath309 $ ] such that the clebsch
gordan conditions are trivially fulfilled for any combinatorial ( i.e. , simplicial complex ) structure of the states . in terms of the difference @xmath310
, one could for instance construct states uniformly randomized over the interval @xmath311\cap\tfrac{1}{2}{\mathbb n}\,,\ ] ] on which any three elements satisfy the clebsch
gordan conditions . as an example ,
we consider uniformly distributed spins for the same @xmath290 as in the above cases of coherent states ( figure [ fig : dsiid ] ) .
again , the difference with respect to the classical triangulation randomized around is of order @xmath305 . for a sum over seven random states with spins in @xmath312 for @xmath313 ( solid , dashed and dot - dashed curve ) on a regular torus triangulation ( with @xmath314 .
right : deviation of @xmath0 from the classical case.,title="fig:",width=283 ] for a sum over seven random states with spins in @xmath312 for @xmath313 ( solid , dashed and dot - dashed curve ) on a regular torus triangulation ( with @xmath314 .
right : deviation of @xmath0 from the classical case.,title="fig:",width=283 ] it is worth noticing that , in all the above examples of quantum states , the difference with the corresponding classical state is rather marginal .
the key result of the calculations for coherent states on a given regular triangulation is that the quantum corrections are very small , since by varying spins @xmath290 and spread @xmath49 one gets deviations from the classical spectral dimension only of order @xmath315 .
a basic feature of quantum mechanics is the superposition principle .
the coherent and randomized states already are a typical example in this sense , as they are superpositions in the spin - network basis .
one can probe the quantum features of spatial geometry also by constructing other kinds of superposition states .
one obvious strategy is to superpose coherent states themselves .
there are various ways one could do so .
possible choices would be sums in the coherent - state labels @xmath290 , @xmath16 and @xmath49 , latexmath:[\[\label{sup1 } or over different complexes and their dual graphs @xmath77 : @xmath317 we will first deal with superposed states on the same complex and consider superpositions of complexes in the next section . in the case ( [ sup1 ] ) , the expectation value of the heat trace does not simplify to a single sum over the expectation values of squared coefficients . assuming trivial intertwiners , @xmath318-\rmi(k_{l}-k'_{l})j_{l}}\mbox{tr}\,\rme^{\tau\delta_{\gamma}(j_{l})}. \label{eq : htcrossterms}\end{aligned}\ ] ] assuming a fixed extrinsic curvature @xmath319 there is no phase causing interferences , and for sharply peaked coherent states the cross terms with @xmath320 are suppressed . in this case , one would expect that the ` naive ' sum disregarding the cross terms @xmath321\mbox{tr}\,\rme^{\tau\delta_{\gamma}(j_{l})}\ ] ] would still be a good approximation for @xmath322 .
when summing over coherent states peaked at different spins , it makes a crucial difference whether we take them as peaked at classical geometries of different size or at the same classical geometry given by a fixed scale @xmath323 . in both cases
, the spectral dimension turns out to be an average of the spectral dimensions of the parts summed over .
however , in the first case ( figure [ fig : supjs ] ) these individual profiles are shifted with respect to one another such that the superposition has dimension of order one only in the regime where most of them overlap . for the superposition of coherent states
@xmath324 on a regular torus triangulation ( with @xmath301 triangles ) with unrescaled laplacian compared with the states summed over ( dashed curve ; see figure [ fig : coherjs]).,width=283 ] in the second case ( figure [ fig : supjref ] ) , all profiles have features at the same scales , so that the superposition yields a @xmath0 plot close to the classical geometry peaked at ; the quantum correction is even less important , being smaller than for the individual coherent states ( figure [ fig : coherjs ] ) .
for the superposition of coherent states @xmath324 on a regular torus triangulation ( with @xmath301 triangles ) with rescaled laplacian compared with the states summed over ( dashed curve ; see figure [ fig : coherjs ] ) .
right : deviation of @xmath0 from the classical case.,title="fig:",width=283 ] for the superposition of coherent states @xmath324 on a regular torus triangulation ( with @xmath301 triangles ) with rescaled laplacian compared with the states summed over ( dashed curve ; see figure [ fig : coherjs ] ) .
right : deviation of @xmath0 from the classical case.,title="fig:",width=283 ] in this section , we have investigated the spectral dimension of coherent states and superpositions of these on toroidal complexes .
quantum corrections are small in the small - spin regime ( figure [ fig : coherjs ] ) , for large spreads ( figure [ fig : cohersigmas ] ) and for total randomization in a given spin interval ( figure [ fig : dsiid ] ) .
superpositions of coherent states peaked at the same geometry smoothen out the quantum deviations even more ( figure [ fig : supjref ] ) .
only superpositions of states peaked at different geometries show a distinct behaviour .
this happens when the spectral dimension of the individual states is noticeably different from one another ( figure [ fig : supjs ] ) .
in quantum gravity , another source of non - trivial features in effective geometry is expected to come from summing over the combinatorial structures .
thus , in this section we will investigate superpositions of states on distinct combinatorial complexes .
in the following , we are interested in superposed states of the form @xmath325 in this definition , we have included a weighting factor @xmath326 for the graphs to be superposed .
a natural choice would be , for example , @xmath327 , a symmetry factor for the automorphism group , which may or may not be included .
the calculations we will present do not include it , but later on we will briefly discuss the effect of weighting , which we checked in detail . comparing with the heat - trace expectation value of the superposition of coherent states on the same complex , equation ( [ eq : htcrossterms ] ) ,
the sum here does not contain cross terms since spin - network states on different graphs are orthogonal .
we will only consider the summation over equilateral triangulations .
these can be understood as the extreme case of minimal spread @xmath49 , that is , sharply peaked at the intrinsic curvature but totally randomized in the extrinsic one .
the reason is that we have already seen in the previous section that the effect of the quantum fluctuations in coherent states is only of order @xmath299 .
one would then expect that the effect of superposing truly quantum coherent states can be reproduced by a superposition of the discrete geometries associated with their peak values ( equation ( [ eq : supgraphs ] ) ) .
indeed , we will find that the effects of superposing graphs are of order higher than @xmath299 .
this justifies the approximation of the full sum , which in the case of superposition of graphs is considerably more challenging from a numerical point of view .
as before , it is crucial whether the superposed triangulations are taken as purely combinatorial or as refinements of the same smooth geometry through rescaling the edge lengths ( figure [ fig : sumt2 ] ) . while in both cases the result is an averaging of the spectral dimension ( depending on the weights @xmath326 ) , in the first one this effect takes place at larger diffusion scales , in contrast with the small - diffusion - scale regime of the second .
triangulations of figure [ fig : dst2triang ] ( included as dashed curves for comparison ) with trivial weights , unrescaled ( left ) and rescaled ( right ) .
[ fig : sumt2],title="fig:",width=283 ] triangulations of figure [ fig : dst2triang ] ( included as dashed curves for comparison ) with trivial weights , unrescaled ( left ) and rescaled ( right ) .
[ fig : sumt2],title="fig:",width=283 ] the most interesting case here is the second ( rescaling ) .
since all complexes summed over are triangulations of the same smooth geometry , it can be interpreted as a special case of a semi - classical state incorporating a ( kinematical ) continuum approximation ( which , obviously , can only be implemented up to some finite order in our explicit calculations ) .
furthermore , the peak in the spectral dimension is a discreteness effect which does not appear for these states as a result of averaging . for these reasons , it is interesting to calculate not only the superposition of a few rescaled triangulations but also of all regular triangulations of the type described in section [ sub : triangulations ] , of size ranging from @xmath328 up to some @xmath329 ( figure [ fig : sumt2pmax ] ) .
triangulations of size @xmath330 ( solid curve ) and , for comparison , the individual @xmath331 and @xmath332 cases ( dotted and dot - dashed curve ) .
[ fig : sumt2pmax],width=283 ] the effect is a more extended plateau .
thus , from these calculations one would expect the appearance of a plateau at the topological dimension for sums over more and larger triangulations , but without the discretization effect of a peak at the characteristic lattice scale .
finally , the same can be done for the randomly subdivided triangulations .
the effect of summing is , again , an averaging of the @xmath0 profiles .
qualitatively , these are quite different from the regular triangulations . triangulations ( dashed line , for comparison ; see figure [ fig : dst2subran ] ) , unrescaled ( left ) and rescaled ( right ) .
[ fig : dst2randomsum],title="fig:",width=283 ] triangulations ( dashed line , for comparison ; see figure [ fig : dst2subran ] ) , unrescaled ( left ) and rescaled ( right ) .
[ fig : dst2randomsum],title="fig:",width=283 ] in this final section , we have seen that the superposition of states on distinct combinatorial complexes produces an interesting behaviour of the spectral dimension . for a superposition of regular triangulations of the same smooth geometry but of different size , we find that the discreteness artifact of the the peak , appearing in classical cases , disappears ( figure [ fig : sumt2 ] ) and a plateau with the topological dimension is obtained ( figure [ fig : sumt2pmax ] ) . still , there are no hints for an effective smaller dimension at smaller scales in one superposition state .
the running of the spectral dimension in the profiles of the figures is , in fact , mainly due to discreteness artefacts ( @xmath333 at small @xmath4 ) and topological effects ( @xmath333 at large @xmath4 ) . on the other hand , superpositions of 1 - 3 pachner subdivisions
have a @xmath0 plateau at a height smaller than the topological dimension ( figure [ fig : dst2randomsum ] ) . summarizing
, the averaging effect stemming from superposition states is the only manifestation of additional geometric data in the profile of the spectral dimension , which would be otherwise reproduced by a classical triangulation .
in this paper , we have explored the spatial spectral dimension of states of quantum geometry in the case of ( 2 + 1)-dimensional euclidean lqg . to this end , we have discussed the definition of the spectral dimension for combinatorial complexes equipped with the geometric variables defining the lqg configuration space .
we have then showed that quantum states have a well - defined spectral dimension , identified with the scaling of the quantum expectation value of the heat trace .
we also checked the quantum observable properties of these lqg geometries .
we have presented for the first time a systematic classification of topological effects and discreteness artefacts in classical settings , pushing the analysis to analytic expressions whenever possible .
after that , we studied coherent states in lqg , both as superpositions of geometries on the same complex and as superpositions of complexes , in particular triangulations of the same geometry .
one of the main results of this work is the lack of any strong indication of a dimensional flow _ not _ due to topology or discreteness , as otherwise suggested in @xcite . to clarify the source of this discrepancy ,
let us discuss the differences between our work and @xcite more in detail .
the basic idea in @xcite is to consider the scaling of the ( quantum ) laplacian with a length scale @xmath334 . since the laplacian scales as the inverse metric , and the metric roughly scales as an area
, one could relate the scaling of the spectrum of the area operator in lqg to the scaling of the quantum laplacian . with the further identification , on one hand , of the length scale with the spin representation label @xmath226 as @xmath335 , and from dimensional reasonings in momentum space on the other hand , one ends up with an effective dispersion relation which is eventually used to derive the spectral dimension , as in the smooth approaches mentioned in the introduction .
our starting point , in contrast , is to consider the full structure of the laplacian as directly acting on states of quantum geometry .
the proper discrete laplacian ( [ eq : dislap ] ) indeed turns out to have a much richer structure than in the inverse - area heuristic argument .
it is not obvious whether there could be any regime of approximation where the dimensional reasoning about the inverse area would apply .
the laplacian takes necessarily the form of a matrix with non - zero entries for incident , neighboring nodes .
therefore , it is further required to take the full underlying combinatorial structure into account , which is neglected in the approximations of @xcite .
also , the validity of an approximation based on a single simplex is hard to understand from our perspective , as we have found that it is crucial to consider large enough complexes in order to reproduce the topological dimension at large scales ( small probing resolution ) .
as we have seen , calculations on triangulations with few simplexes , such as the dipole , do not show any geometric regime between the ones in which discreteness and topology effects dominate , thus forbidding any interpretation of the spectral dimension as a spatial dimension .
nevertheless , one might try to understand the approximations of @xcite in terms of regular lattices such as the ones used here with the regular triangulations .
however , in contrast with the non - trivial dispersion relation derived from the area spectrum there , in our setting the effect of the discrete ( length ) spectrum seems to be washed away in a twofold way .
first , because of the more involved manner in which the spectra are entering the expression of the laplacian and , second , because of the tracing over the dual nodes of large complexes .
these two differences seem to be the main reason for the distinct results . due to the complexity of the problem
, we have been unable to obtain the explicit dependence on such geometric data in the spin representation ( pure or coherently peaked at ) in a closed , analytic form .
yet , this can still be analyzed when comparing states labelled with different spins .
for the latter , our results clearly show only a minimal dependence ( figure [ fig : coherjs ] ) .
notwithstanding , it has to be stressed that the calculations we have presented are only a first step , so our results can only be seen as a hopefully interesting indication of how actual lqg geometries behave .
first , we have analyzed the spectral dimension only for _ kinematical _ states and in 2 + 1 euclidean lqg . in @xcite ,
the argument was for kinematical states in 3 + 1 dimensions , although the result was interpreted as the spectral dimension of the full ( 3 + 1)-dimensional spacetime and the effective dispersion relation was inserted in a heat trace over four - dimensional momentum space .
furthermore , the argument was extended to physical states on certain one - simplex spin foams in the boundary formalism , in 2 + 1 dimensions @xcite as well as in 3 + 1 dimensions @xcite .
the spectrum yielding the effective dispersion relation was taken to be the same in both dimensions there , but for the full laplacian considered in the present paper the step to 3 + 1 dimensions gives rise to essential differences , as already explained in the introduction .
it is also clear that , eventually , the quantum spectral dimension should be considered on physical states , i.e. , the quantum dynamics needs to be taken into account for the results to have a solid physical significance .
let us briefly discuss how these issues could be addressed .
the most significant differences for gravity in 3 + 1 dimensions versus 2 + 1 dimensions lie in the dynamics . in this respect
, one would not necessarily expect crucial qualitative discrepancies at the level of kinematical states in the quantum theory .
however , in lqg the polyhedral interpretation in terms of the spectra of geometric observables is different in that the length spectra in 2 + 1 dimensions already provide a classical geometric interpretation as regge geometries , while in 3 + 1 dimensions the corresponding geometric variables are not sufficient .
more variables such as dihedral angles would be needed , but their corresponding quantum operators do not commute . in principle , these uncertainties do not pose any obstacle to define the laplacian @xcite and analyze the quantum spectral dimension .
such features could be effectively captured using the flux representation of states and operators @xcite or , alternatively , the spinor representation @xcite . beside the implicit operator
ordering ambiguities , the major challenge in pursuing this step is just technical .
the explicit form of the laplacian becomes very complicated and clever ideas and approximations would be needed to cast the expectation value of the corresponding heat trace into a form feasible for actual computations .
this is going to be an interesting challenge , as the non - commutativity in 3 + 1 dimension might give rise to unexpected features in the quantum spectral dimension .
from the point of view of a complete theory of quantum gravity , the challenge to include the full dynamics is more important .
this is indeed the starting point for studying the spectral dimension of spacetime , which is the actual quantity mainly studied in the quantum gravity literature so far .
this observable could be defined in a covariant , path - integral - like formalism .
then , the expectation value of the heat trace would amount to the insertion of the heat trace function in appropriate variables into a sum over spacetime histories .
independently of the precise model of lqg dynamics chosen ( i.e. , the exact spin - foam or , equivalently , group field theory model ) , new issues arise .
a first challenge is the concept of spacetime heat trace as a path - integral observable .
the latter is not simply a polynomial in the configuration variables .
this raises several computational difficulties : the insertion of the heat trace in the path integral can not be easily split into integrals of monomials of local variables . as a result
, it seems quite hard , if not impossible , to solve the path integral with such heat - trace insertion .
one should then find some viable approximation , for example one in which earlier results @xcite on ` ultralocal ' observables ( i.e. , observables depending only on the moduli of the discrete @xmath285-field values ) in the ponzano
regge model can be used to evaluate the heat - trace expectation value .
even if a method for solving the path integral was at hand , a further issue is raised by degenerate configurations in the state sum .
already in the ( 2 + 1)-dimensional theory , the spin - foam sum generically contains degenerate geometries , in contrast with its kinematical states as shown in section [ sec : deflqg ] .
we have done preliminary calculations on small complexes showing this feature @xcite . as a consequence ,
the spacetime heat - trace insertion has a complex part .
this is not , in itself , a sign of any inconsistency .
nevertheless , it raises an interesting question about the physical relevance and interpretation of a complex spectral dimension , and it may represent an intriguing line of investigation in the analysis of the spin - foam quantum dynamics . finally , a general issue with any explicit spin - foam evaluation is divergences .
this is crucial for a global object such as the spectral dimension , since common regularization techniques may not be appropriate .
for instance , gauge - fixing geometric data along a maximal tree of the dual complex @xcite are problematic because the heat trace might trivialize on the corresponding configurations . considering coherent states on the boundary suppressing larger spin configurations ( the cause of divergences ) in the bulk works only for smaller complexes .
but , as we have discussed , larger complexes are needed in order to obtain a meaningful spectral dimension . one way to circumvent this problem would be spin - foam models with quantum groups @xcite .
an actual solution , however , would consist in the proper renormalization of spin foams in terms of their group field theory ( second - quantized ) description @xcite .
10 t hooft g 1993 dimensional reduction in quantum gravity _
salamfestschrift _ ed
a ali , j ellis and s randjbar - daemi ( singapore : world scientific ) ( arxiv : gr - qc/9310026 ) carlip s 2009 spontaneous dimensional reduction in short - distance quantum gravity ?
_ aip conf .
proc . _ * 1196 * 72 ( arxiv:0909.3329 )
calcagni g 2010 fractal universe and quantum gravity _ phys .
_ * 104 * 251301 ( arxiv:0912.3142 ) ambjrn j , jurkiewicz j and loll r 2010 quantum gravity as sum over spacetimes _ lect .
notes phys . _
* 807 * 59 ( arxiv:0906.3947 ) ambjrn j , grlich a , jurkiewicz j and loll r 2012 nonperturbative quantum gravity _ phys . rep . _
* 519 * 127 ( arxiv:1203.3591 ) ambjrn j , jurkiewicz j and loll r 2005 the spectral dimension of the universe is scale dependent _ phys .
lett . _ * 95 * 171301 ( arxiv : hep - th/0505113 ) ambjrn j , jurkiewicz j and loll r 2005 reconstructing the universe _ phys . rev .
_ d * 72 * 064014 ( arxiv : hep - th/0505154 ) benedetti d and henson j 2009 spectral geometry as a probe of quantum spacetime _ phys . rev .
_ d * 80 * 124036 ( arxiv:0911.0401 ) niedermaier m and reuter m 2006 the asymptotic safety scenario in quantum gravity _ living rev .
relativity _ * 9 * 5 reuter m and saueressig f 2012 quantum einstein gravity _ new j. phys _ * 14 * 055022 ( arxiv:1202.2274 ) lauscher o and reuter m 2005 fractal spacetime structure in asymptotically safe gravity _ j. high energy phys .
_ jhep10(2005)050 ( arxiv : hep - th/0508202 ) reuter m and saueressig f 2013 asymptotic safety , fractals , and cosmology _ lect .
notes phys . _
* 863 * 185 ( arxiv:1205.5431 ) calcagni g , eichhorn a and saueressig f 2013 probing the quantum nature of spacetime by diffusion _ phys .
_ d * 87 * 124028 ( arxiv:1304.7247 ) hoava p 2009 spectral dimension of the universe in quantum gravity at a lifshitz point _
* 102 * 161301 ( arxiv:0902.3657 ) benedetti d 2009 fractal properties of quantum spacetime
lett . _ * 102 * 111303 ( arxiv:0811.1396 ) alesci e and arzano m 2012 anomalous dimension in semiclassical gravity _ phys .
_ b * 707 * 272 ( arxiv:1108.1507 ) modesto l 2012 super - renormalizable quantum gravity _ phys .
_ d * 86 * 044005 ( arxiv:1107.2403 ) arzano
m and calcagni g 2013 black - hole entropy and minimal diffusion _ phys .
_ d * 88 * 084017 ( arxiv:1307.6122 ) calcagni g and modesto l 2013 nonlocality in string theory ( arxiv:1310.4957 ) calcagni g 2013 diffusion in multiscale spacetimes _ phys .
_ e * 87 * 012123 ( arxiv:1205.5046 ) calcagni g and nardelli g 2013 spectral dimension and diffusion in multiscale spacetimes ( arxiv:1304.2709 )
rovelli c 2004 _ quantum gravity _
( cambridge : cambridge university press ) thiemann t 2007 _ modern canonical quantum general relativity _
( cambridge : cambridge university press ) perez a 2013 the spin foam approach to quantum gravity _ living rev
. relativity _ * 16 * 3 oriti d 2013 group field theory as the 2nd quantization of loop quantum gravity ( arxiv:1310.7786 ) oriti d 2012 the microscopic dynamics of quantum space as a group field theory _ foundations of space and time _ ed j murugan , a weltman and g f r ellis ( cambridge : cambridge university press ) ( arxiv:1110.5606 ) modesto l 2009 fractal structure of loop quantum gravity _ class .
quantum grav . _
* 26 * 242002 ( arxiv:0812.2214 ) caravelli f and modesto l 2009 fractal dimension in 3d spin - foams ( arxiv:0905.2170 ) magliaro e , perini c and modesto l 2009 fractal space - time from spin - foams ( arxiv:0911.0437 ) baratin a , dittrich b , oriti d and tambornino j 2011 non - commutative flux representation for loop quantum gravity _ class . quantum grav . _ * 28 * 175011 ( arxiv:1004.3450 ) giasemidis g 2013 spectral dimension in graph models of causal quantum gravity ( arxiv:1310.8109 ) reid d 2003 manifold dimension of a causal set : tests in conformally flat spacetimes _ phys .
d * 67 * 024034 ( arxiv : gr - qc/0207103 ) calcagni g , oriti d and thrigen j 2013 laplacians on discrete and quantum geometries _ class .
quantum grav .
_ * 30 * 125006 ( arxiv:1208.0354 ) thrigen j 2013 fields and laplacians on quantum geometries _ the thirteenth marcel grossmann meeting _ ed k rosquist , r t jantzen and r ruffini ( singapore : world scientific ) ( arxiv:1302.7135 ) desbrun m , hirani a n , leok m and marsden j e 2005 discrete exterior calculus ( arxiv : math/0508341 ) calcagni g , modesto l and nardelli g 2013 quantum spectral dimension in quantum field theory in preparation bell n and hirani a n 2012 pydec : software and algorithms for discretization of exterior calculus _
acm trans . math .
software _ * 39 * 3 ( arxiv:1103.3076 ) vassilevich d v 2003 heat kernel expansion : user s manual _ phys .
rep . _ * 388 * 279 ( arxiv : hep - th/0306138 ) gurau r 2010 lost in translation : topological singularities in group field theory _ class .
quantum grav . _
* 27 * 235023 ( arxiv:1006.0714 ) ben achour j , geiller m , noui k and yu c 2014 spectra of geometric operators in three - dimensional lqg : from discrete to continuous _ phys .
_ d * 89 * 064064 ( arxiv:1306.3246 ) freidel l , livine e and rovelli c 2003 spectra of length and area in 2 + 1 lorentzian loop quantum gravity _ class .
quantum grav . _
* 20 * 1463 ( arxiv : gr - qc/0212077 ) guedes c , oriti d and raasakka m 2013 quantization maps , algebra representation and non - commutative fourier transform for lie groups _ j. math .
phys . _ * 54 * 083508 ( arxiv:1301.7750 ) olver f w j , lozier d w , boisvert r f and clark c w ( ed ) 2010 _ nist handbook of mathematical functions _
( cambridge : cambridge university press ) chung f r k 1997 _ spectral graph theory _ ( providence , ri : american mathematical society ) ambjrn j , loll r , nielsen j l and rolf j 1999 euclidean and lorentzian quantum gravity lessons from two dimensions _ chaos solitons fractals _ * 10 * 177 ( arxiv : hep - th/9806241 ) brehm u and khnel w 2011 lattice triangulations of @xmath336 and of the 3-torus _ isr .
j. math .
_ * 189 * 97 wiese u 2009 introduction to lattice field theory in
_ 15th saalburg summer school _
( http://www.itp.uni-hannover.de/~lechtenf/events/lectures/wiese.pdf[http://www.itp.uni-hannover.de/
lechtenf / events / lectures / wiese.pdf ] ) hall b c 1994 the segal
bargmann ` coherent state ' transform for compact lie groups _ j. funct .
* 122 * 103 sahlmann h , thiemann t and winkler o 2001 coherent states for canonical quantum general relativity and the infinite tensor product extension _ nucl .
_ b * 606 * 401 ( arxiv : gr - qc/0102038 ) bahr b and thiemann t 2009 gauge - invariant coherent states for loop quantum gravity : ii . non - abelian gauge groups _ class .
quantum grav . _
* 26 * 045012 ( arxiv:0709.4636 ) oriti d , pereira r and sindoni l 2012 coherent states in quantum gravity : a construction based on the flux representation of loop quantum gravity _ j. phys . a : math .
gen . _ * 45 * 244004 ( arxiv:1110.5885 ) bianchi e , magliaro e and perini c 2010 coherent spin - networks _ phys .
_ d * 82 * 024012 ( arxiv:0912.4054 ) baratin a , dittrich b , oriti d and tambornino j 2011 non - commutative flux representation for loop quantum gravity _ class .
quantum grav .
_ * 28 * 5011 ( arxiv:1004.3450 ) livine e r and tambornino j 2012 spinor representation for loop quantum gravity _ j. math .
phys . _ * 53 * 012503 ( arxiv:1105.3385 ) livine e r and ryan j p 2009 a note on @xmath285-observables in ponzano
regge @xmath337 quantum gravity _ class .
quantum grav . _
* 26 * 035013 ( arxiv:0808.0025 ) calcagni g , oriti d and thrigen j 2014 in preparation freidel l and louapre d 2004 ponzano regge model revisited .
i : gauge fixing , observables and interacting spinning particles _ class .
quantum grav . _
* 21 * 5685 ( arxiv : hep - th/0401076 ) han m 2011 4-dimensional spin - foam model with quantum lorentz group _ j. math .
phys . _ * 52 * 072501 ( arxiv:1012.4216 ) carrozza s , oriti d and rivasseau v 2014 renormalization of tensorial group field theories : abelian @xmath338 models in four dimensions _ commun .
phys . _ * 327 * 603 ( arxiv:1207.6734 ) carrozza s , oriti d and rivasseau v 2014 renormalization of an @xmath12 tensorial group field theory in three dimensions _ commun .
phys . _ * * ( arxiv:1303.6772 ) | the spectral dimension is an indicator of geometry and topology of spacetime and a tool to compare the description of quantum geometry in various approaches to quantum gravity .
this is possible because it can be defined not only on smooth geometries but also on discrete ( e.g. , simplicial ) ones . in this paper
, we consider the spectral dimension of quantum states of spatial geometry defined on combinatorial complexes endowed with additional algebraic data : the kinematical quantum states of loop quantum gravity ( lqg ) .
preliminarily , the effects of topology and discreteness of classical discrete geometries are studied in a systematic manner .
we look for states reproducing the spectral dimension of a classical space in the appropriate regime .
we also test the hypothesis that in lqg , as in other approaches , there is a scale dependence of the spectral dimension , which runs from the topological dimension at large scales to a smaller one at short distances . while our results do not give any strong support to this hypothesis
, we can however pinpoint when the topological dimension is reproduced by lqg quantum states .
overall , by exploring the interplay of combinatorial , topological and geometrical effects , and by considering various kinds of quantum states such as coherent states and their superpositions , we find that the spectral dimension of discrete quantum geometries is more sensitive to the underlying combinatorial structures than to the details of the additional data associated with them
. http://dx.doi.org/10.1088/0264-9381/31/13/135014[class
. quantum grav .
* 31 * ( 2014 ) 135014 ] http://arxiv.org/abs/1311.3340[(arxiv:1311.3340 ) ] + aei-2013 - 196 |
the statistical features of coupled systems have attracted considerable interest in many branches of physics . random matrix theory ( rmt )
has been successfully used in many of those investigations .
rmt was founded by wigner @xcite .
it is a schematic model @xcite in which the hamiltonian , or , more generally , the wave operator of the system , is replaced by a random matrix .
the necessary prerequisite is that the system be sufficiently `` complex , '' implying that the matrix elements of the hamiltonian , or wave operator , calculated in an arbitrary basis , behave like random numbers .
it has been shown that the spectral fluctuations in numerous different systems , if measured on the scale of the local mean - level spacing , are very well modeled by rmt ; see the reviews in refs .
@xcite . due to the connection with chaos
, one frequently refers to those systems as quantum chaotic which show correlations of rmt type .
similarly , systems are often referred to as regular if they lack spectral correlations .
we consider two coupled systems .
we assume that either the two systems are chaotic before they are coupled or that the coupling itself introduces chaoticity if the separate systems are regular .
this scenario is equivalent to the breaking of symmetries , if only two values of the quantum number belonging to that symmetry are taken into account .
the statistical features crucially depend on the strength of the coupling measured on the scale of the local mean - level spacing .
many studies have been devoted to this issue of chaotically coupled systems or , equivalently , to symmetry breaking .
we mention isospin breaking in nuclear physics @xcite , symmetry breaking in molecular physics @xcite , symmetry breaking in resonating quartz crystals @xcite and coupled microwave billiards @xcite . while these studies addressed the spectral correlation , several investigations in nuclear physics @xcite focused on the statistics of the wave functions and related observables in the presence of symmetry breaking or similar effects . in all these cases , rmt approaches in the spirit of the rosenzweig - porter model
@xcite were successful .
sometimes observables in the time domain such as spectral form factors are more appropriate than the eigenenergy correlations functions @xcite .
this is so , for example , in the case of the presently much discussed fidelity ; see refs .
@xcite and references therein .
another example is the study of the energy spread in chaotic systems @xcite . in the context of coupled systems ,
the time evolution of wave packets was investigated in ref .
@xcite . in the present contribution
, we study energy localization in two coupled systems in the time domain .
this problem was addressed in a recent work by weaver and lobkis @xcite who measured the time dependence of the wave intensity distribution in two coupled reverberation rooms . to this end , these authors recorded the time response to an elastic excitation of two coupled aluminum cubes .
moreover , they investigated the same problem theoretically and they numerically calculated the response in coupled two - dimensional membranes . in our study , we se tup and analyze an rmt model , based on the approaches in refs .
its general character makes our model useful for similar problems in different physics contexts .
in particular , we expect that our rmt approach also applies to coupled quantum dots .
the article is organized as follows . in sec .
[ sec : exp_and_model ] we sketch the work of weaver and lobkis @xcite . in sec .
[ sec : rmt ] we set up the rmt model and analyze it analytically and numerically .
we compare our results to those of weaver and lobkis in sec .
[ sec : rmt_result ] .
discussion and conclusions are given in sec .
[ sec : dic_concl ] .
as we aim at a comparison with their findings , we present the work of weaver and lobkis @xcite in some detail .
thereby , we also introduce the notation and conventions .
the system studied experimentally consists of two aluminum cubes coupled by a solid connection , manufactured out of a solid aluminum block .
the corners of the cubes were removed to desymmetrize the structure .
this was done to ensure `` chaotic '' motion .
elastomechanical wave modes were excited in one room , and the response was measured in the other room . in this way , 16 different curves of energy intensity versus time were recorded , each in a small region around a different frequency .
the results show that the energy does not always spread equally over the two rooms . if the coupling is weak , then the wave intensity is higher in the room where the initial excitation was performed than in the other room , regardless of how long one waits .
hence , the energy ratio never approaches unity .
this deviation from the equipartition of the energy in the two rooms is referred to as energy localization .
the resulting data are shown in fig .
[ fig : wl_exp ] , and as expected there is localization in the bins of larger mean - level spacing , but not in the bins of small mean - level spacing .
we will discuss these results further in sec .
[ sec : rmt_result ] .
two numerical studies were performed on membranes with rough boundaries @xcite .
the dynamics of the system was in both cases governed by discretized wave equations in each of the rooms , and the coupling was realized in different ways in the two numerical calculations , to which we refer as n1 and n2 in the sequel .
in the first one , n1 , the connection had the form of a window between the two membranes similar to the situation in the experimental setup . in the second numerical calculation , n2 , the rooms were separated , but springs were attached to a few different sites in rooms 1 and 2 , thereby coupling those sites . in n1 and n2 ,
a nonvanishing initial condition was given to one site in one of the rooms , and the response was calculated at different sites in the other room .
the resulting time series were cosine - bell time windowed to focus on a specific instant in time .
then , the time series were fourier transformed and integrated over a small region in frequency to accumulate data around a certain frequency . as in the experiment
, 16 different curves of intensity versus time were obtained around different frequencies . at different frequencies ,
the systems have different effective couplings . therefore
, one expects @xcite the time behavior of the different curves to differ in the degree of localization , as well as in the way in which this asymptotic saturation value is reached . due to the differences in the coupling mechanism , there will also be differences between the results of n1 and n2 .
moreover , weaver and lobkis performed an analytical model study .
the elastic wave equation for the state of the system @xmath0 , say , is of second order in time . to focus on the response in a narrow interval around a certain frequency @xmath1 , the ansatz @xmath2 is made with the assumption that @xmath3 varies slowly with time
this leads to a first - order differential equation in time for @xmath3 which has the form @xmath4 where @xmath5 is the wave operator of the original second - order equation .
the energies @xmath6 and @xmath7 in rooms 1 and 2 at time @xmath8 are defined as the total probability density of finding the system state @xmath3 in one of the states @xmath9 , which are good eigenstates in room @xmath10 : @xmath11 strictly speaking , @xmath12 is no energy .
nevertheless , we find this terminology introduced in ref .
@xcite appropriate and use it as well , because @xmath12 measures the degree of motion in room @xmath10 .
if no energy dissipates into the surrounding environment , the total energy @xmath13 is conserved
i.e. , independent of time . in the sequel
, it is always assumed that the system is excited in room 1 , and the energy is measured in room 2 .
analytical solutions for two coupled states are presented in ref .
@xcite by employing different statistical assumptions .
we set up the model in sec . [
sec : rmt_model ] .
the connection to the two - level form factor is established in sec .
[ rmtff ] , and a @xmath14 version of the model is evaluated in sec .
we discuss numerical simulations of the rmt model in sec .
[ sec : rmt_num ] . finally , we comment on chaotically coupled regular systems in sec . [ sec : rmt_pos ] .
spectral correlations in elastomechanics have been shown to be well described by rmt @xcite .
this is also true in the case of symmetry breaking @xcite , which is of direct relevance for the present study .
thus , rmt is also likely to be capable of modeling the time behavior of elastomechanical systems .
as the first - order equation has proved to be a good approximation to the experimental situation , we also base our model on this schrdinger type of equation .
thus , it is more natural to replace @xmath15 by the random matrix @xmath16 than to replace @xmath5 itself .
it turns out that this is indeed the best choice .
the appropriate rmt model is an extension of the one employed in refs .
the random matrix @xmath16 modeling the operator @xmath15 reads @xmath17 + \alpha \left [ \begin{array}{cc } 0 & v \\
v^\dagger & 0 \end{array } \right ] \ .
\label{eq : rmt_model_def}\ ] ] the two matrices @xmath18 , model the uncoupled rooms 1 and 2 .
they are real symmetric and have random entries .
we draw them from ( two independent ) gaussian orthogonal ensembles ( goe s ) . as the rooms in the experiments and the numerical calculations n1 and n2 were of the same size , the level densities were also the same .
this can be adjusted in the rmt model by giving the matrices @xmath19 the same statistical weights and the same dimension @xmath20 , such that @xmath21 is the total dimension of @xmath16 .
the matrix @xmath22 also has random entries .
the strength of the coupling is measured by the dimensionless parameter @xmath23 .
it is sufficient to always assume @xmath24 .
the statistical weight of @xmath22 is chosen such that the total @xmath16 is in the goe of @xmath25 matrices for @xmath26 .
we write the eigenvalue equation for the total hamiltonian @xmath16 in the form @xmath27 the eigenvalues @xmath28 and eigenvectors @xmath29 are functions of the coupling parameter @xmath23 .
it is convenient to introduce the notation @xmath30 \ , \label{psi12}\end{aligned}\ ] ] where @xmath31 , is the projection of @xmath32 onto the subspace @xmath10 .
we emphasize that @xmath31 , are functions of the coupling parameter @xmath23 .
the eigenvalue equations for the hamiltonians @xmath19 are written as @xmath33 where the eigenvalues @xmath34 and eigenvectors @xmath35 are not functions of @xmath23 .
this difference between @xmath36 and @xmath35 is an immediate consequence of the fact that @xmath16 depends on @xmath23 , while @xmath37 and @xmath38 do not .
we will also use the notation @xmath39 for the corresponding @xmath21-dimensional vector with zeros in the first @xmath20 components . at time @xmath40 ,
the system is excited in room one such that the state of the total system can be written as @xmath41 \ , \label{source}\end{aligned}\ ] ] where @xmath42 in room one is not specified in detail .
we refer to @xmath43 as to the source .
the time evolution of the source is then simply @xmath44 is the time evolution operator . using the eigenvectors
@xmath45 , the energy in room 2 is thus given by @xmath46 t(t ) s \ .
\label{eq : rmt_e2}\end{aligned}\ ] ] the block matrix only contains the unit matrix @xmath47 for room 2 .
we write @xmath48 instead of @xmath7 to emphasize the dependence on the coupling parameter @xmath23 in the rmt model .
we will study averages @xmath49 over the ensemble of matrices introduced in eq . .
occasionally , we will also consider averages over the direction of the source .
this average is denoted by angular brackets @xmath50 .
the total energy @xmath51 is trivially conserved in the framework of our model .
we have @xmath52 hence , we can always construct @xmath53 , once @xmath48 has been calculated .
as all correlations have to be measured @xcite on the local scale of the mean - level spacing @xmath54 , we introduce an unfolded time @xmath55 .
this is also the scale on which @xmath23 acts ; we introduce the unfolded coupling parameter @xmath56 .
the energy on the unfolded scale is then given by @xmath57 small values of @xmath23 have a large effect on the correlation functions if the mean - level spacing @xmath54 is small as well .
we now derive an estimate for the time evolution of @xmath48 which should apply to strong coupling strength i.e. , to a parameter @xmath23 which is large on the unfolded scale of the mean level spacing . if one assumes that the source @xmath43 comprises excitations into all states @xmath58 , one may average over the direction of the source .
we find , from eq . , @xmath59 t(t ) \ , \label{e2est1}\end{aligned}\
] ] where @xmath60 is the trace over the @xmath61 sector of the whole matrix
i.e. , over the upper left block .
the constant @xmath62 results from the average . expanding the time evolution operator in terms of the eigenvectors , @xmath63
we arrive after a short calculation at @xmath64 \ . \label{e2est2}\end{aligned}\ ] ] the vectors @xmath65 and @xmath66 are , according to eq . , the projections of the full eigenvector @xmath29 onto the subspaces corresponding to the two rooms .
they depend on the coupling parameter @xmath23 .
these vectors coincide with the eigenvectors for the matrices @xmath37 and @xmath38 only for @xmath67 .
hence , they are only orthogonal in this special case . for all values @xmath68 ,
the scalar products @xmath69 and @xmath70 are neither zero nor given by @xmath71 . the ensemble average over the eigenenergies @xmath72 and over the eigenfunctions @xmath29 decouples only for the parameter values @xmath67 and @xmath73 i.e. , if the system either falls into two goe s or is represented by one goe .
we now estimate the ensemble average of the energy @xmath74 by making the approximation that the averages over eigenenergies and eigenfunctions decouple . for strong coupling this should yield a reasonable result , and we will use this approximation in that limit only .
we average over the eigenfunctions . for each term in the double sum in eq .
the product of scalar products gives the same contribution which we can take out of the sums .
thus , we find @xmath75 } \ , \label{e2est3}\end{aligned}\ ] ] where we absorbed the contributions from the average over the scalar products into the constant @xmath62 .
the average over the eigenvalues is yet to be performed .
luckily , the average in eq . is recognized as the fourier transform of the two level correlation function @xcite .
this yields , on the unfolded scale , @xmath76 \ , \label{e2est4}\end{aligned}\ ] ] where the function @xmath77 is referred to as the two - level form factor .
the term @xmath78 is due to the diagonal contributions in the double sum of eq . .
we notice that the conventions used here require a rescaling of time with a factor of @xmath79 .
the expression is an estimate for the energy @xmath80 , exclusively in terms of the fourier - transformed two - level spectral correlation function for the transition from two goe s to one goe .
unfortunately , the two - level form factor is not known analytically for all values of @xmath81 . the result for one goe , corresponding to @xmath82 , reads @xcite @xmath83 we notice that the function @xmath84 has the limit properties @xmath85 the second property will be used to fix the scale in comparison with the results of ref .
the estimate should apply to large coupling
i.e. , to parameter values @xmath86 .
quite often , one obtains surprisingly good information about an rmt model by restricting it to the smallest possible matrix dimension such that the nontrivial specific characteristics of the model are still present .
this is successful in the case of the nearest - neighbor spacing distribution for the gaussian orthogonal , unitary and symplectic ensembles ( ggoe , gue and gse , respectively ) @xcite , but also for crossover transitions @xcite . here
, we proceed analogously .
we obtain a @xmath14 rmt model by setting @xmath87 in eq . .
it turns out convenient to absorb the parameter @xmath23 into the matrix element @xmath22 such that @xmath88 \label{m1}\ ] ] and to readjust the probability density function @xmath89 for the ensemble average accordingly : @xmath90 as before , the goe is recovered for @xmath73 .
we introduce eigenvalue and angle coordinates @xmath91 = \left [ \begin{array}{cc } \cos\varphi & -\sin\varphi \\
\sin\varphi & \cos\varphi \end{array } \right ] \left [ \begin{array}{cc } \omega_1 & 0 \\ 0 & \omega_2 \end{array } \right ] \left [ \begin{array}{cc } \cos\varphi & \sin\varphi \\
-\sin\varphi & \cos\varphi \end{array } \right ] \ , \label{m3}\ ] ] which implies that the time evolution can be written in the same form with eigenvalues @xmath92 and @xmath93 . for the integration measure one has @xmath94 the source in this @xmath14 model
is simply given by @xmath95 .
thus , collecting everything , we find , with eq . , @xmath96 \nonumber \\ & & \qquad \sin^22\varphi \ , \sin^2\frac{\omega_1-\omega_2}{2}t \ .
\label{m5}\end{aligned}\ ] ] we introduce @xmath97 and @xmath98 as new integration variables , perform the @xmath99 and @xmath100 integrals , and arrive at @xmath101 \nonumber\\ & & \qquad \qquad \left[{\rm i}_0 ( y^2 ( 1-\alpha^2 ) ) - { \rm i}_1(y^2 ( 1-\alpha^2))\right ] \cos 2\alpha yt \ , .
\label{eq:2d_result}\end{aligned}\ ] ] here , @xmath102 and @xmath103 denote the modified bessel functions of zeroth and first order , respectively . formula is the final result in the framework of our @xmath14 model . in ref .
@xcite , solutions for two - state models based on eq . with different statistical assumptions were given .
although the general behavior is similar to our @xmath14 rmt result , the analytical forms of these solutions in ref .
@xcite are different . for obvious reasons ,
the unfolding of formula is meaningless . nevertheless ,
experience with @xmath14 model for the spacing distribution tells that meaningful statements can be achieved if the transition parameter
i.e. , the coupling strength @xmath23 in our case is interpreted properly . here
, we can do that in the following manner . for large times @xmath8 ,
the function @xmath49 becomes constant , because it reaches its saturation limit .
the latter will depend on @xmath23 .
thus , comparing the saturation limit for the @xmath14 model with that of the @xmath25 rmt simulation or with those of the experiment and the numerical calculations n1 and n2 allows one , in principle , to interpret @xmath23 on the unfolded scale . the saturation limit is easily obtained from eq . . due to the riemann - lebesgue lemma @xcite , we have @xmath104 because the cosine function in the integrand oscillates so rapidly for large @xmath8 , that the integral gives zero in the limit @xmath105 .
we notice that there is no equipartition of the energies for @xmath106 at @xmath26 .
this may seem a bit unexpected , because @xmath16 is a @xmath14 goe matrix for that parameter value .
it simply reflects the need to properly interpret the parameters , as just discussed .
we study a two - state system and compare the probability for a transition between the two states with the probability of staying in one state .
there is no reason why these should be equal . in the @xmath107 limit , however , we expect equipartition for the full goe .
as the coupling parameter has to be measured on the scale of the local mean level spacing , the limit @xmath108 corresponds to the limit @xmath109 in the @xmath14 version , and in that limit the right - hand side of eq
. tends to @xmath110 and therefore to equipartition .
versus ( dimensionless ) time .
data from n1 of ref .
@xcite , bin 16 , plotted as dashed line .
the result from eq . for @xmath111 , @xmath62 set by property 2 of eq . and a rescaling of time to make the two curves fit for early times , is plotted as a solid line.,height=151 ] the numerical simulations of the rmt model are performed as in refs .
the unfolding of the results , however , is not done in the standard way . to compare with the numerical calculations of weaver and lobkis , it is more convenient to first unfold the level densities and then `` refold '' the spectra of our model with the level densities of ref .
thereby we ensure that the mean - level densities of the rmt model acquire a form given by the weyl formula for the billiardlike system of ref .
the appropriate weyl formula for the modal density in a square membrane with dirichlet boundary conditions reads @xcite @xmath112 depending on the frequency @xmath1 . here
, @xmath113 is the area of the membrane and @xmath114 is the side length of one room .
moreover , @xmath115 is the wave velocity and is chosen as unity here . using twice the area of a room ( @xmath116 ) and a value of @xmath114 which takes the roughness of the boundaries into account ( @xmath117 ) , we aquire an expression for the level density of the entire system . a comparison with the weyl formula of ref .
@xcite shows that the terms from the extra edges introduced to model the disorder cancel .
hence , we use the formula for a square membrane , which is precisely eq . .
this weyl formula is employed to refold our rmt model for both numerical calculations n1 and n2 . in the latter case , the system is not really a billiard , due to the springs used for the coupling
. nevertheless , the weyl formula should be a good approximation to the real level density .
for every simulation , we generate random matrices of dimension @xmath118 , unfold , refold , calculate @xmath48 , and average over the results of 800 such simulations . as the functions @xmath48 are not measured on the unfolded scale , but on that of the weyl formula
, we do not introduce the notation @xmath80 in the present context .
moreover , we notice that the numerical calculations n1 and n2 depend on the mean frequency @xmath1 in the bin under consideration . accordingly , we arrive at a two - parameter family of time - dependent curves @xmath119 .
we now associate each bin in the numerical calculation performed by weaver and lobkis with the corresponding frequency @xmath1 .
this leaves us with a one - parameter family of time - dependent curves @xmath120 , for each bin labeled by @xmath121 . in the rmt model defined by eq . and in its subsequent analysis
, we always drew the matrices @xmath18 , from ( two independent ) goe s , implying that we model the two rooms individually as chaotic systems before the coupling is considered .
one can also assume that the two rooms are regular systems before they are coupled .
in that case , the matrices @xmath18 , would be drawn from poisson ensembles @xcite .
we also did such numerical simulations .
the results are qualitatively the same if the coupling
i.e. , the matrix @xmath22 introduces enough chaos .
the main effect is an adjustment of the scales .
what matters for the qualitative behavior is the chaoticity of the total system .
+ + + + we now compare the results of our rmt model with those of ref .
@xcite . the quantity studied in ref .
@xcite is mostly the fraction of energy in room 2 , or @xmath122 . in our figures , we plot @xmath123 instead .
we first consider the estimate , eq .
, involving the goe form factor and the result of the @xmath14 rmt model and compare with the numerical calculations n1 and n2 .
as formula was derived under the assumption of strong coupling on the scale of the local mean - level spacing , it should apply to the high - frequency bins of n1 . anticipating the later extraction of the coupling parameter , we already now mention that indeed @xmath124 in that bin . in fig .
[ fig : wl_num1_vs_form_factor ] we compare eq . with data from bin 16 ( the highest frequency bin ) of n1 .
a good description is obtained , although no equipartition is reached .
this implies that the effective coupling @xmath81 is large , but not very large .
we recall that formally @xmath82 corresponds to the strongest coupling @xmath73 i.e. , to one single system .
expansion of the form factor for short times @xmath125 reveals a linear short time dependence of the ensemble - averaged energy @xmath80 .
this is in agreement with the analytical discussion of ref .
@xcite . .coupling parameters @xmath23 for the @xmath126 model used in fig . [ fig:2d_result ] .
the asymptotic saturation @xmath127 values for n2 in bins 2 , 3 , and 4 are taken from ref .
the corresponding coupling parameters are calculated from eq . .
[ cols="^,^,^",options="header " , ] [ tab : num_rmt ] we find good qualitative agreement between the rmt simulation and n1 and n2 .
importantly , fluctuations between different random matrices are quite large , and the rmt curves presented here are averages of @xmath128 over 800 random matrices . among those some show very much closer similarity to the curves of weaver and lobkis @xcite , and we believe that most of the discrepancy in the low - frequency bins of n1 are due to fluctuations resulting from the specific choice of the boundary and perhaps other parameters in the numerical calculations of ref .
the reason for this interpretation of the discrepancy is the peculiar form of the fluctuations in bin 2 of n1 ; see fig .
[ fig : rmt_n1_bin2 ] .
since these lower bins have a lower level density , they should also be more sensitive to this kind of fluctuations . in the intermediate - frequency bins
, we find that the numerical calculations n1 and n2 are well reproduced by the rmt model . in the figures for the high - frequency bin , figs .
[ fig : rmt_n1_bin16 ] and [ fig : rmt_n2_bin16 ] , the rmt results are seen to deviate downwards for large times .
it is conceivable that the short - time behavior strongly depends on the specific realization of the coupling while the long - time behavior is universal .
the result from the experiment is shown in fig .
[ fig : wl_exp ] .
the qualitative features are well described by the rmt model .
it can be seen that the system localizes in the low - frequency bins , but does not do so in the high - frequency bins . in other words ,
the system approaches equipartition of the energy .
the localization behavior depends on the strength of the coupling measured on the scale of the local mean - level spacing , and the mean - level density is significantly higher in the high - frequency bins .
it seems that the saturation value has not yet been reached on the time scale visible in fig .
[ fig : wl_exp ] .
a closer inspection shows an initial behavior similar to the one in the results from our numerical rmt calculations and a slight upwards trend of the energy ratio towards the end of the time window studied . according to ref .
@xcite this means that , first , the expected asymptotic value of @xmath129 is reached in the middle of the time window studied and that , second , there is another unknown effect acting on longer time scales adding to the energy spread which is likely to be due to the coupling of the setup to the environment .
thus , we do not attempt to extract a quantitative estimate of the coupling parameter @xmath23 for the experiment .
we have set up and analyzed an rmt model to describe the time behavior of coupled reverberation rooms .
this system shows localization effects under certain conditions . within our rmt model
, we gave an estimate of the strong coupling behavior which involved the two - level goe form factor .
moreover , we studied the @xmath14 version of our rmt model analytically for arbitrary coupling strength and performed numerical simulations for the @xmath25 version . from the comparison with the work of weaver and lobkis , we conclude that the rmt model yields a good qualitative description .
moreover , we find an interpretation of the localization effect by relating it to the universal features of rmt models for crossover transitions . formally similar rmt models have been studied in connection with symmetry breaking .
then , the parameter @xmath23 measures the degree of symmetry breaking .
this is so for isospin breaking in nuclear physics @xcite , symmetry breaking in molecular physics @xcite , and symmetry breaking in resonating quartz crystals @xcite .
the experiment that comes closest to the present situation is the study of spectral correlations in coupled microwave billiards @xcite . in this work and in the experiment of weaver and lobkis @xcite , an excitation initially in system one would stay there for all times if the coupling is zero
this is formally analogous to a conserved quantum number . in the rmt model
the parameter measuring the size of the connection has a most natural counterpart : the root - mean - square matrix element due to the coupling measured on the scale of the local mean - level spacing .
thus , there are two ways of making the effective , dimensionless coupling parameter @xmath81 small and to thereby introduce localization : either the original coupling parameter @xmath23 which always has a dimension is small or the mean - level spacing @xmath54 is made large .
it might be surprising that no equipartition of the energy is seen for all coupling strengths even if one waits very long
. this would be the expectation if one compares the system with two water basins coupled by a channel .
suppose that initially the two basins are empty and the channel is closed by a gate .
now one of the basins is filled with water and the gate is opened at @xmath40 . obviously , the water levels in the two basins will be equal after sufficiently long time .
the speed with which this equipartition is reached simply depends on the cross section of the channel . in the present case of the two coupled acoustic rooms
the situation is different , because the wave character of the excitations has to be taken into account .
thus , the crucial parameter entering is the size of the coupling connection
i.e. , its geometrical width compared to a typical wavelength .
the first waves after the excitation which come from system 1 into system 2 enter a silent territory and cause the first excitations there .
the next waves coming from system 1 , however , encounter these first excitations in system 2 and interfere with them constructively or destructively .
this process continues and , of course , after being excited in system 2 , waves also travel back into system 1 . for smaller times , this complicated dynamics certainly depends strongly on the realization of the coupling .
this is clearly so in the numerical calculation of ref .
nevertheless , the loneg - time behavior , in particular the saturation limit , shows universal characteristics , consistent with general features of quantum chaotic systems ; see ref .
@xcite . in the energy domain , this is borne out in the fact that the correlations on smaller energy scales are described by universal rmt features , while system - specific properties show up on larger scales , leading to deviations from the rmt prediction . to avoid confusion ,
we emphasize that these universal rmt features include those in the presence of symmetry breaking . by system - specific properties we mean , most importantly , the scales set by the shortest periodic orbits .
it is worthwhile to realize that the chaoticity of the individual subsystems before the coupling is not crucial , if the coupling itself introduces enough chaos .
we tested numerically that the behavior of the corresponding model for two regular subsystems coupled chaotically shows the same qualitative behavior .
the saturation value is reached slightly faster , which simply means that time is rescaled . finally , we mention that our model is not restricted to elastomechanics
. it would also apply to coupled quantum dots and other coupled systems .
this project was prompted by a conversation between r.l .
weaver and one of the present authors ( t.g . ) during a workshop at the centro internacional de ciencias ( cic / unam ) , cuernavaca , mexico .
we thank r.l .
weaver for providing us with the data for the numerical calculations n1 and n2 .
we acknowledge financial support from det svenska vetenskapsrdet .
wigner , reprinted in _ statistical theories of spectra : fluctuations _ , edited by c.e .
porter ( academic press , new york , 1965 ) , p. 199 .
m. l. mehta , _ random matrices _ , 2nd ed .
( academic press , new york , 1991 ) . t. guhr , a. mller -groeling , and h.a .
weidenmller , phys . rep . * 299 * , 189 ( 1998 ) .
stckmann , _ quantum chaos an introduction _
( cambridge university press , cambridge , england , 1999 ) .
f. haake , _ quantum signatures of chaos _ , 2nd ed .
( springer - verlag , berlin , 2001 ) .
mitchell , e.g. bilpuch , p.m. endt , and j.f .
shriner , jr .
lett . * 61 * , 1473 ( 1988 ) .
t. guhr and h.a .
weidenmller , ann .
( n.y . ) * 199 * , 412 ( 1990 ) .
leitner , phys .
e * 48 * , 2536 ( 1993 ) ; d.m .
leitner , h. kppel , and l.s .
cederbaum , phys .
lett . * 73 * , 2970 ( 1994 ) . c. ellegaard , t. guhr , k. lindemann , j. nygrd , and m. oxborrow , phys .
77 * , 4918 ( 1996 ) .
h. alt , c.i .
barbosa , h.d .
grf , t. guhr , h.l .
harney , r. hofferbert , h. rehfeld , and a. richter , phys .
* 81 * , 4847 ( 1998 ) .
adams , g.e .
mitchell , and j. f. shriner , jr . ,
phys . lett .
b * 422 * , 13 ( 1998 ) .
adams , g. e. mitchell , w.e .
ormand , and j. f. shriner , jr .
b * 392 * , 1 ( 1997 ) .
barbosa , t. guhr , and h.l .
harney , phys .
e. * 62 * , 1936 ( 2000 ) .
a. hamoudi , r.g .
nazmitdinov , e. shahaliev , and y. alhassid , phys .
c. * 65 * , 064311 ( 2002 ) .
n. rosenzweig and c.e .
porter , phys .
120 * , 1698 ( 1960 ) .
n.r . cerruti and s. tomsovic , j. phys .
a. * 36 * , 3451 ( 2003 ) . g. benenti , g. casati , and g. veble , phys .
e. * 68 * , 036212 ( 2003 ) .
t. prosen , t.h .
seligman , and m. znidaric , prog .
* 150 * , 200 ( 2003 ) .
j. vanicek , and e.j .
heller phys .
e. * 68 * , 056208 ( 2003 ) .
d. cohen , f.m .
izrailev , and t. kottos , phys .
* 84 * , 2052 ( 2000 ) .
t. kottos and d. cohen , phys .
e. * 64 * , 065202 ( r ) ( 2001 ) .
o. bohigas , s. tomsovic , and d. ullmo , phys . rep .
* 223 * , 43 ( 1993 ) .
r.l . weaver and o. lobkis , j. sound vib . * 231 * , 1111 ( 2000 ) .
r.l . weaver , j. acoust .
85 * , 1005 ( 1989 ) .
g. lenz and f. haake , phys .
lett * 65 * , 2325 ( 1990 ) .
i.s . gradshteyn and i.m .
ryzhik , _ table of integrals , series and products _ , ( academic press , new york , 2000 ) .
m.v . berry , ann.phys .
( n.y . ) * 131 * , 163 ( 1981 ) . | we set up and analyze a random matrix model to study energy localization and its time behavior in two chaotically coupled systems .
this investigation is prompted by a recent experimental and theoretical study of weaver and lobkis on coupled elastomechanical systems .
our random matrix model properly describes the main features of the findings by weaver and lobkis .
due to its general character , our model is also applicable to similar systems in other areas of physics for example , to chaotically coupled quantum dots . |
this is a continuation of a previous paper @xcite where the hermiticity properties of the quantum group invariant @xmath2 spin - chain hamiltonian alc87 @xcite @xmath3have been investigated . here
the anisotropy parameters @xmath4 are defined in terms of a single variable @xmath1,@xmath5and @xmath6 denote the pauli matrices acting on the @xmath7 site of the spin - chain represented by the state space @xmath8 with @xmath9 being isomorphic to @xmath10 .
the above hamiltonian , besides belonging to an integrable model , is distinguished by its @xmath11-invariance .
namely , one has the following representation @xmath12 in terms of the matrices @xmath13which obey the familiar @xmath11-commutation relations@xmath14=[2s^{z}]_{q}\ .
\label{qg}\]]here , as usual , we have set@xmath15_{q}:=\frac{q^{x}-q^{-x}}{q - q^{-1}}\ .\ ] ] the case when the deformation parameter @xmath1 lies on the unit circle @xmath16 is of particular interest , since then the corresponding lattice model is believed to correspond in the thermodynamic limit to a cft with central charge @xcite@xmath17however , for these values of @xmath1 one easily verifies that @xmath18 is non - hermitian with respect to the conventional scalar product , @xmath19with @xmath20 denoting the hermitian adjoint ( or conjugate transpose in the case of matrices ) .
it is therefore not clear whether ( [ h ] ) constitutes a well - defined quantum integrable system or is even diagonalizable .
nevertheless , numerical computations show that the hamiltonian possesses real spectrum for @xmath21 .
this raises the questions whether the hamiltonian ( [ h ] ) is related to a hermitian hamiltonian via introducing a different inner product or performing a similarity transformation . in our previous work we established for the two cases 1 .
@xmath22 and 2 .
@xmath23(here @xmath24 is the number of sites ) that the hamiltonian ( [ h ] ) is quasi - hermitian ( see e.g. sgh92,most04 ) .
that is , we proved by explicit construction that there exists a _ positive , hermitian and invertible _
operator @xmath25 such that@xmath26this enabled us to introduce a new inner product @xmath27with respect to which the hamilton operator becomes hermitian,@xmath28hence , we rigorously established that ( [ h ] ) does indeed give rise to a well - defined quantum system .
there is , however , an important difference between the cases ( i ) and ( ii ) : for @xmath22 one has first to carry out a reduction of the state space @xmath29 employing what is known as `` quantum group reduction '' @xcite in the literature in order to remove non - trivial jordan blocks of the hamiltonian . for @xmath30 such a reduction proved to be unnecessary . in both cases , ( i ) and ( ii ) , we found in @xcite explicit algebraic expressions for @xmath31 utilising concepts related to @xmath32-symmetry ( see e.g. @xcite for recent reviews ) and the quantum analogue of schur - weyl duality @xcite .
in fact , we established not only quasi - hermiticity of the hamiltonian but of an associated representation of the temperley - lieb algebra @xmath33 which is _ _
_ _ obtained from @xmath34generators @xmath35 satisfying the commutation relations@xmath36again there is a representation @xmath37 in terms of pauli matrices by mapping @xmath38 , where the @xmath39 s are defined in terms of the local hamiltonians@xmath40 in @xcite we constructed a quasi - hermiticity operator @xmath31 such that @xmath41thus , we obtained a self - adjoint representation of the temperley - lieb algebra @xmath33 with respect to the inner product ( [ etaprod ] ) .
we briefly review these results in section 2 of this paper .
this is one of the few examples of a non - hermitian hamiltonian where quasi - hermiticity has been proved by an _
explicit _ and _ exact _ construction of the operator @xmath31 .
the new result in the present article is an entirely independent approach ( which does neither use @xmath32-symmetry nor quasi - hermiticity ) to arrive at the same self - adjoint representation of the temperley - lieb algebra for the case ( ii ) , i.e. when @xmath30 .
namely , we will define for each fixed spin sector @xmath42 ( here @xmath43 is the number of down spins ) a positive functional @xmath44over the temperley - lieb algebra using only its graphical representation in terms of kauffman diagrams @xcite .
we then employ a well - kown tool from the representation theory of @xmath45-algebras , the gelfand - naimark - segal ( gns ) construction , which provides us with an inner product @xmath46 in terms of the functional @xmath47 and an associated hilbert space @xmath48 .
the gns construction is such that the temperley - lieb algebra generators @xmath49 act as self - adjoint operators over @xmath48 , thus giving rise to a hermitian hamiltonian @xmath50 .
the main conjecture of this article is that the two self - adjoint representations , one induced by @xmath31 and the other by the functionals @xmath51 are unitarily equivalent .
that is , there exists for each @xmath52 an isomorphism @xmath53 satisfying@xmath54and the identity@xmath55the left - hand side of the last equality can be computed by purely graphical means .
this provides us with a novel , efficient formalism to investigate the spectrum , eigenvectors and , in the long term , correlation functions of the quantum group invariant xxz hamiltonian .
moreover , this is also a new representation theoretic result for the temperley - lieb algebra which ( to the best of the author s knowledge ) appears not to be contained in the extensive literature on this subject , see for instance @xcite .
the extension of this graphical calculus to the root of unity case , where a reduction of the state space needs to be carried out first , is still an open problem .
in order that the reader can fully appreciate the main conjecture of this article we briefly review the results obtained for the quasi - hermiticty operator @xmath31 in the previous paper @xcite .
we recall the following definitions from @xcite.@xmath56 .
[ parity , time and spin - reversal ] .
@xmath57 _ then we define the linear operator _ @xmath58 _ _
on _ _ @xmath8 _ _ by setting__@xmath59_in contrast , the operator _ @xmath60 _ _ acts on the basis vectors as the identity , _ _ @xmath61__but is defined to be antilinear , such that _ _ @xmath62__thus , any matrix _ _ @xmath63 _ _ ( such as the hamiltonian _ _ @xmath64 _ _ ) is transformed into its complex conjugate under the adjoint action of _ _
@xmath60 , @xmath65 _ finally , we introduce the ( linear ) spin - reversal operator _ @xmath66 _ _ by setting__@xmath67one now easily computes the transformation properties of the quantum group and temperley - lieb generators under the adjoint action of the involutions @xmath68 .
they are summarised in table 1 . [ cols="^,^,^",options="header " , ] again an illustrative example of how to obtain the above values might be helpful .
let us consider the matrix element @xmath69 .
the kauffman diagram is obtained by flipping the one for @xmath70 at the horizontal axis and then connecting it with the one for @xmath71 from above,@xmath72{scalarproduct610.eps}\ .\]]obviously , we have @xmath73 . adding the orientation @xmath74 on the top and bottom of the diagram we see that there are no unoriented lines or arcs , @xmath75 , and that we have two anti - clockwise oriented cups , @xmath76 .
hence , @xmath77 the reader might wonder why we picked out this particular basis among other choices .
the above basis vectors transform particularly simple under the action of the temperley - lieb algebra : the only coefficients which occur in the expansion are 0 , @xmath78 and powers of @xmath79 .
it is in this basis that @xmath31 takes its simplest form and which is most suitable for the diagonalization of the hamiltonian .
in this article we have presented a conjecture for a new construction of a self - adjoint representation of the temperley - lieb algebra @xmath33 with deformation parameter @xmath80 .
clearly , in the thermodynamic limit @xmath81 this section of the unit circle shrinks to the point @xmath82 where the hamiltonian is hermitian and its algebraic structure drastically simplifies . nevertheless , this construction is significant as it provides a practical advantage in computing the quasi - hermiticity operator @xmath31 for finite lattice size and , as for now , is the simplest starting point for the computation of the square root @xmath83 .
the latter is needed to obtain the hamiltonian@xmath84which is hermitian with respect to the original scalar product on the state space .
this hamiltonian might look very different form the initial hamiltonian ( [ h ] ) and there is a priori no reason to expect that it will have only nearest neighbour bulk interaction as the similarity transformation is highly nonlocal .
preliminary computations for the case when @xmath85 , not discussed in @xcite , show that ( [ littleh ] ) indeed contains nonlocal bulk interactions .
namely consider the non - hermitian hamiltonian@xmath86 , \quad 0<g<1 , \label{hg}\]]then one can show for small @xmath87 its spectral equivalence with the hermitian hamiltonian@xmath88 , \label{hg}\]]where @xmath89 are fermionic creation and annihilation operators at lattice site @xmath90 . the hopping probability between a site @xmath90 and its @xmath7 neighbour is encoded in the real coefficients @xmath91 .
up to order @xmath92 one finds that the nonvanishing contributions are@xmath93a more detailed discussion of this model , which warrants an investigation in its own right , will be presented elsewhere @xcite .
the close investigation of the regime ( ii ) , @xmath94 might shed light on how to perform similar gns constructions for other sections of the unit circle .
such a purely algebraic formulation is desirable in order to apply the results to a wider range of temperley - lieb type models , see e.g.kulish and references therein . for instance , by showing the more restrictive identity @xmath95 for the temperley - lieb algebra generators the results immediately generalize also to transfer matrices .
namely , it easily follows from the results presented in @xcite and this article that we have constructed a unitary representation of the hecke algebra@xmath96the latter form the basic building blocks for transfer matrices , as for instance the ( non - symmetric ) double row transfer matrices considered in connection with lattice systems associated to logarithmic minimal models in @xcite . with regard to these applications
it is natural to ask if the change of the inner product will effect the description in terms of logarithmic conformal field theory when @xmath81 and how this connects to possibly non - local interactions in the bulk .
it is planned to address these questions in future work .
finally , it is worth noting that we discussed in this article the problem of a non - hermitian quantum hamiltonian in the language of @xmath97-algebras .
instead of using the concepts of quasi - hermiticity or @xmath32-symmetry we have considered an associated @xmath45-algebra of the given quantum mechanical system and investigated the existence of positive linear functionals .
the latter have then be used to construct via the gns approach an inner product with respect to which the hamiltonian in question is hermitian .
the formulation of quantum mechanics in the language of @xmath98-algebras is not new , but the novel aspect in this article is the presentation of an explicit example where this formulation can be connected with quasi - hermiticity and @xmath32-symmetry .
* acknowledgments*. the author would like to thank catharina stroppel and robert weston for many helpful discussions .
is financially supported by a university research fellowship of the royal society . | this is a continuation of a previous joint work with robert weston on the quantum group invariant xxz spin - chain ( math - ph/0703085 ) .
the previous results on quasi - hermiticity of this integrable model are briefly reviewed and then connected with a new construction of an inner product with respect to which the hamiltonian and the representation of the temperley - lieb algebra become hermitian .
the approach is purely algebraic , one starts with the definition of a positive functional over the temperley - lieb algebra whose values can be computed graphically . employing the gelfand - naimark - segal ( gns ) construction for c@xmath0-algebras a self - adjoint representation of the temperley - lieb algebra
is constructed when the deformation parameter @xmath1 lies in a special section of the unit circle .
the main conjecture of the paper is the unitary equivalence of this gns representation with the representation obtained in the previous paper employing the ideas of pt - symmetry and quasi - hermiticity .
an explicit example is presented . |
carbon nanotubes are a new form of carbon with rather unique properties @xcite .
a single wall nanotube can be considered as resulting from rolling up a single graphene sheet to form a hollow cylinder .
the chirality and diameter of a particular tube can be described in terms of a role - up vector @xmath0 , where @xmath1 and @xmath2 denote the unit vectors of the hexagonal honeycomb lattice , and n and m are integers @xcite . in a 2-d graphene sheet
the @xmath3-bonding and @xmath4-antibonding states become degenerate at the k - point of the hexagonal brillouin zone , resulting in a zero band - gap semiconductor . in a nanotube ,
quantization of the wavefunction along its circumference restricts the allowed wave - vectors to certain directions of the graphite brillouin zone so that @xmath5 , where @xmath6 is an integer . if at least one of these wavevectors passes through the k point , the tube is metallic , otherwise it is a semiconductor with a finite band - gap .
thus , @xmath7 zig - zag tubes are expected to be metallic if @xmath8 is an integer , and semiconducting otherwise @xcite .
as the role - up vector @xmath9 rotates away from the @xmath7 direction , the resulting @xmath10 tubes are chiral and are expected to be metallic if @xmath11 is an integer .
otherwise they are semiconductors with a gap @xmath12 , where @xmath13 is the tube radius . finally , when @xmath9 is rotated 30@xmath14 away from the @xmath7 direction , @xmath15 and the resulting armchair tubes are metallic .
recent scanning tunneling microscope studies have verified the basic conclusions of simple theory @xcite . + the above predictions assume a semi - infinitely long tube , and corresponding experimental electronic structure studies have utilized tubes several microns in length .
it is very important to know how an additional confinement of the electrons , along the nanotube axis , in finite length tubes will affect the tube s electronic structure .
as the tube length is diminished , there should be a transition from a 1-d structure ( quantum wire ) to a 0-d structure ( quantum dot ) .
currently , little is known about the effects of finite length on the properties of carbon nanotubes .
recently , however , venema _ et al .
_ @xcite have been able to cut nanotubes into segments of a few tens of nanometers length and , using tunneling spectroscopy , obtained evidence of increased band - gaps in these short tubes . understanding quantum size effects in carbon nanotubes , besides the basic scientific value of understanding a prototypical 1-d to 0-d transition , is essential in device applications of nanotubes .
the miniaturization offered by nanotube - based devices requires that not only their diameter but also their length should be in the nanometer range .
for example , micrometer - long nanotubes have been used as coulomb islands in single electron transistors ( set ) operating at liquid helium temperatures @xcite . by using shorter tubes ,
the operating temperature of sets can be raised drastically . in general , the electrical properties of nanotubes
may be tailored by selecting their length . + to obtain the needed insight into the electronic structure of short nanotubes and learn about their stability ( information useful in studies of growth mechanisms ) we performed electronic structure calculations on different length @xmath16 armchair nanotubes . to determine the local properties and the spatially - resolved electronic structure of finite - size nanotubes , quantum - chemical methods are more suitable than band - structure calculations .
however , the theoretical determination of electronic properties , such as the band - gap value , remains a difficult task , as has been shown in studies of low band - gap polymers @xcite . to compensate for the weaknesses of the various quantum chemical techniques
, we utilized a number of both first - principles ( density functional and hartree - fock ) and semi - empirical computational techniques ( mndo - pm3 and extended hckel ) .
we used the above techniques to calculate the band - gap values , the dos spectrum and the binding energy of carbon atoms as a function of the length of the tube segment .
while , according to the discussion in the introduction , infinitely long @xmath16 nanotubes should be metallic , all computational techniques predict a band - gap for short nanotubes whose value decreases towards zero with increasing tube length .
interestingly , the variation of the band - gap is not monotonic but shows strong oscillations as a function of tube length . while all computational techniques used predict these oscillations , the actual values of the band - gap vary strongly with the technique used .
computed dos spectra and local density of states distributions show that the transition from a 0-d to a 1-d - like spectrum takes place already in very short 5 - 10 nm tubes .
the binding energy of the carbon atoms increases with increasing tube length in a similar manner .
the behavior of finite graphene sheets is also investigated .
electronic structure calculations were performed using both ab - initio : hartree - fock ( hf ) and density functional theory ( dft ) , and semi empirical : modified neglect of differential overlap - parametrized model 3 ( mndo - pm3 ) and extended hckel ( ehmo ) computational methods . in the hf calculations , we employed a parallel version of the gamess program @xcite in which sto-3 g quality basis sets were employed for carbon and hydrogen . for the dft computations
, we used the demon - ks software @xcite .
the calculations were performed with huzinaga s minimal basis sets @xcite in conjunction with the generalized gradient approximation ( gga ) of perdew and wang for exchange @xcite and correlation @xcite .
the semi - empirical calculations were carried out with the mndo - pm3 method @xcite included in the gamess program @xcite , while for the ehmo calculations we used the program included in the yaehmop package @xcite .
+ in the carbon nanotube models , the c - c and c - h bond length were fixed at the values observed in bulk graphite @xcite which are 1.42 and 1.09 , respectively .
all calculations were performed at fixed geometries without optimization of the atomic structures .
the elementary unit used , called here a section , is defined as a single circular plane of carbon atoms that are packed along the length of the nanotube ( note : the distance between two consecutive sections in the nanotube is 1.22 ) .
the dangling bonds at the ends of the tube were saturated with hydrogen atoms .
the energy band - gap is obtained as the difference between the lumo and homo energies , while the fermi energy ( @xmath17 = 0 ev ) is the mean energy between lumo and homo .
+ density of states ( dos ) plots were generated by convoluting the computed electronic structure with a 50:50 combination of gaussian and lorentzian functions . in order to analyze the nature of the energy bands , we performed a series of projections to obtain the local density of states ( ldos ) , where each molecular orbital was weighted by the contribution obtained from a mulliken analysis of specific carbon atoms .
the general expression used to generate the dos plots is : + @xmath18 + [ \frac{2 { n_i}{\omega}}{{\pi } ( { \omega}^2 + 4({\epsilon } - { \epsilon}_i)^2)}]\bigg)\ ] ] where @xmath19 is the resolution , _
n@xmath20 _ is the population of the level _
i _ and ( @xmath21 - @xmath22 ) is the energy difference between the energy @xmath21 and the eigenvalue @xmath22 .
to evaluate the influence of finite length on the electronic properties of the nanotubes , we calculated the electronic structure of different length segments of a ( 6,6 ) nanotube .
as figure 1 shows , all computational techniques used predict that , unlike the infinite tube that is metallic , very short ( @xmath23 100 ) nanotubes have an energy band - gap ( @xmath24 ) .
furthermore , figure 1 reveals that @xmath24 shows a regular oscillatory dependence on tube length .
the oscillation of the band - gap values is particularly strong in the hf method where @xmath24 varies from 2.8 to 9.0 ev for tubes shorter than 20 .
the semi - empirical mndo - pm3 method gives results comparable to those obtained by hf . on the other hand , extended hckel molecular orbital ( ehmo ) and density functional theory - generalized gradient approximation ( dft - gga ) methods lead to quite similar values of the band - gap .
the oscillation amplitude decreases , and the band - gap value converges slowly to zero for larger nanotubes . for a 96 long nanotube
, ehmo predicts a metallic behavior while mndo - pm3 still shows a band - gap of 2.0 ev .
the comparison of the different quantum - chemical methods used to determine @xmath24 will be discussed in more detail below .
+ the inset in figure 1 gives an expanded view of the band - gap oscillations for nanotubes shorter than 25 .
all computational methods predict a regular and converging oscillation of @xmath24 .
the oscillation is related to the structure of the nanotube ; high band - gap values are invariably obtained for tubes having @xmath25 ( where @xmath26 = 1,2 , ... ) sections ( reminder : a section contains a single circular plane of carbon atoms ) .
furthermore , consecutive medium or low band - gap values are periodically observed for nanotubes having @xmath27 and @xmath28 sections .
the main difference between the oscillatory behavior of @xmath24 obtained with the different computational methods lies in the different length needed for an inversion between medium and low band - gap behavior to occur .
the different oscillations observed , as well as the low to medium band - gap inversion for small nanotubes can be explained in terms of the changing bonding characteristics of the homo and lumo orbitals .
+ figure 2 shows the nodal structure of the frontier @xmath3-orbitals for the nanotube models , while figure 3 gives the energies of ten molecular orbitals ( obtained from ehmo theory ) which lie near the band - gap region .
the relative stability of frontier orbitals ( homo and lumo ) for the first three models ( s2 , s3 and s4 with respectively , 2,3 and 4 sections ) is based on the nature of the bonding along the circumference of the tube , i.e intra - section interactions , as well as the bonding along the tube axis , i.e. inter - section interactions .
furthermore , the s2 , s3 and s4 models constitute the repeat units from which the homo / lumo orbitals of the longer tubes may be constructed . from figure 2
, we see that the homo for s2 ( two section model ) , the first member of the s@xmath28 group , has bonding character along the circumference , while the lumo has bonding character between sections , i.e. , along the tube axis .
we should then expect the energy difference between homo and lumo of the s2 structure to be small , leading to a low band - gap . on going from the s3 ( s@xmath27 group ) to the s4 ( s@xmath25 group ) structures , we observe from figure 2 that the homo acquires an additional bonding character along the tube axis and hence is stabilized more . on the other hand , the lumos of both s3 and s4 type structures show antibonding inter - section interactions which lead to a larger band - gap .
the gap is lower for s3 due to the net bonding character of the lumo along the circumference .
the structure and the stability of the frontier orbitals for longer nanotubes are directly related to the characteristics of those three first structures .
the s@xmath28 group ( i.e. s2 , s5 , s8 , ) is characterized by non - bonding inter - section interactions , as is evident by considering the orbital nodal properties in figure 2 .
the stability of the frontier orbitals in the s@xmath28 group is weakly influenced by inter - section interactions .
the slight destabilization of the frontier orbitals is due to long - range antibonding lateral interactions that reduce slowly the band - gap with increasing tube length . in the s@xmath27 group ( s3 , s6 , s9 , ) ,
the decreasing band - gap is caused by a destabilizing lateral antibonding interaction in the homo coupled with a stabilizing lateral interaction in the lumo .
the lateral interactions have a strong influence on the value of the band - gap of the s@xmath27 group so that it actually becomes lower than the band - gap value of the s@xmath28 group at @xmath26=3 .
this behavior occurs at different tube lengths depending on the computational method used to determine the energies of the molecular orbitals .
finally , the s@xmath25 group that is characterized by high band - gaps follows a similar trend as the s@xmath27 group with increasing nanotube length ; a destabilization of the homo through antibonding interactions between sections and stabilization of the lumo through partly bonding lateral interactions . + band - gap oscillations were also reported recently by yoshizawa _
@xcite for 2-d polyphenanthrenes on the basis of extended hckel band structure calculations .
since , instead of the carbon section used here , a phenanthrene - edge structure was used as the elementary building unit in that study , high band - gaps were found in systems with @xmath29 sections rather than @xmath25 sections . nevertheless , although the compositions of the frontier orbitals of the @xmath16 armchair nanotube are to a certain extent different from those of the periodic 2-d polyphenanthrenes , both structures give rise to a similar pattern of band - gap oscillations . in figure 4
we compare the variation of the band - gap value for a perfect @xmath16 nanotube to that of an equivalent 2-d graphene fragment ( whose broken c - c bonds were saturated with hydrogen atoms ) , and that of the periodic sheet of 2-d polyphenanthrenes as determined by yoshizawa @xcite . we find that the growth of the finite 2-d graphene sheet does not lead to a clear band - gap oscillation pattern and the @xmath24 rapidly converges to low values .
furthermore , the finite graphene sheet gives quite different @xmath24 values than the infinite polyphenanthrene sheet .
figure 5 shows the influence of the nanotube diameter on the magnitude of the band - gap oscillations with respect to the fluctuations observed for a periodic graphene sheet . increasing the nanotube diameter by going from the @xmath16 to the @xmath30 tube , brings its band - gap value closer to that of the periodic 2-d graphene sheet .
however , the 1/(diameter ) scaling of @xmath24 predicted for infinite tubes @xcite is not observed in very short tube segments . + crucial to the understanding of the nanotube growth process are changes in the binding energy of c atoms as a function of the increasing nanotube length .
figure 6 compares the binding energy ( be ) of carbon atoms in a finite ( 6,6 ) nanotube as determined by different computational techniques .
a converged be value indicates that all carbon atoms are nearly equivalent to each other .
major changes are observed only for very short tubes where the be found to increase sharply up to about 15 , and then changes more gradually upto 96 ( not shown ) .
the smaller changes in be observed with the ehmo method can be understood as due to the treatment of long - range interaction in ehmo theory ; carbon atoms are weakly perturbed by second and higher neighbor carbon atoms .
the highest estimate of be is obtained with mndo - pm3 ( 7.0 ev / c atom ) , and the lowest with ehmo ( 5.5 ev / c atom ) .
the be obtained with the ab - initio dft - gga ( estimated ) and hf lie between these two extremes .
the low binding energy obtained for short tubes suggests that carbon in short nanotubes would be more reactive .
+ although the energetic properties of nanotubes seem to converge rapidly , a sufficiently long nanotube model is necessary to reproduce the main characteristics of the electronic structure of a quasi-1d periodic system .
the influence of the length on the general electronic structure of a @xmath16 armchair nanotube is reported in figure 7 obtained using the ehmo method .
the nanotube electronic structure is characterized in terms of the position and the intensity of the main dos peaks , the magnitude of dos at the fermi level , and the presence of fine structure .
the spectrum of the very short 4 ( 4 carbon sections ) model is characteristic of a 0-d ( quantum dot ) system with discrete quantum levels .
however , the band position and its overall dos profile is not drastically different from those of more extended systems .
the dos profile for a 96 long model , on the other hand , clearly reveals the presence of characteristic fine structures at high binding energy ( be ) between -22 and -12 ev .
these structures are van hove singularities " characteristic of periodical one - dimensional ( 1-d ) systems @xcite .
each of these peaks is characterized by a specific number of nodes in the wavefunction along the circumference of a single nanotube section , while the @xmath31 tail reflects the free electron character along the tube axis .
in fact , these singularities are becoming evident even in the 47 nanotube that represents the minimum length nanotube that exemplifies such 1-d - like characteristics .
for nanotubes larger than 47 , the main changes in electronic structure occur around the fermi level , where the dos is found to increase slowly with increasing nanotube length .
although it gives a different estimate for the band - gap , the semi - empirical mndo - pm3 method shows similar trends as ehmo concerning the electronic structure of the growing nanotubes .
+ figure 8 gives a more detailed description of the electronic structure of the 96 nanotube @xcite .
in the general dos diagram , the van hove singularities at high be are @xmath32-states that originate from c@xmath33 atomic orbital combinations .
the @xmath32-states formed from c@xmath34 orbitals lie at lower be between -11 and -4 ev , while @xmath3-states extend from -5 to 0 ev .
these states are important as they are responsible for the electrical properties of nanotubes .
the states above the fermi level are mainly @xmath35-states , while @xmath36-states extend above 10 ev .
the inset gives an expanded view of the fine structure of the 96 nanotube near the fermi energy .
in addition , the values of the local density of states ( ldos ) near the fermi energy for sections ranging from the tube boundary ( index 1 ) to the middle of the tube ( section 40 ) are plotted . to construct these ldos plots , we summed over the density of the twelve carbon atoms contained in that section of the nanotube . from figure 8
it can be seen that the ldos of the @xmath37-states near @xmath17 hardly changes along the length of the tube , indicating that these states are indeed well delocalized within this short tube .
+ finally , we shall compare briefly the results obtained with the different computational techniques . as shown in figure 1
, both hf and mndo - pm3 methods give a large band - gap ( 9.0 and 6.5 ev , respectively ) for a short 4 nanotube , while dft and ehmo give moderate band - gap values ( 2.4 and 1.8 ev , respectively ) . due to the poor description of unoccupied orbital energy levels in hartree - fock theory ,
the band - gap values are overestimated by several electron volts for low band - gap polymers @xcite .
the problem is slightly less important in the mndo - pm3 method , where the parametrization includes a treatment of correlation through the fitting with experimental data @xcite .
previous mndo - band structure calculations have predicted a metallic behavior ( @xmath24 = 0 ev ) for periodic 2-d graphite@xcite .
this suggests that the computed band - gap should converge to zero for extended nanotube models ; we estimate from figure 1 that nanotubes larger than 450 would be metallic .
on the other hand , dft - based methods tend to systematically underestimate the band - gap value of low band - gap polymers @xcite .
however , recent developments of dft / hybrid functionals that include a weighted contribution of hartree - fock and dft exchange allow a better evaluation of the band - gap value @xcite .
given that the computed band - gaps with ehmo are similar to the dft results , we may conclude that the band - gap values are also underestimated with ehmo .
based on dft and ehmo results , nanotubes longer than 100 would be metallic . however , given the above discussion on the limitations of the computational techniques , the transition to the metallic state likely occurs at longer lengths , probably in the range of 10 - 20 nm .
we have investigated the electronic structure of finite - length armchair carbon nanotubes using a number of ab - initio and semi - empirical quantum chemistry techniques .
electrons in carbon nanotubes are confined along their circumference . in this study
we imposed an additional confinement along the tube axis by using short tube segments .
we found that armchair tubes that are metallic when infinitely long , develop a band - gap .
the value of the band - gap decreases with increasing tube length , but the decrease is not monotonic ; it shows a well defined oscillation in short tubes .
this oscillation can be accounted for in terms of the changes in the bonding characteristics of the homo and lumo orbitals of the tubes as a function of their increasing length .
finite size graphene sheets were also found to have a finite band - gap , but no clear oscillation was observed .
the dos spectrum of short nanotubes evolves with increasing length from that characteristic of a 0-d system to that characteristic of a delocalized 1-d system .
this transformation is complete already for tubes about 10 nm long .
the chemical stability of the nanotubes , expressed by the binding energy of a carbon atom , is low for short tubes but increases and saturates at tube lengths @xmath23 10 nm .
e - mail : [email protected] e - mail : [email protected] e - mail : [email protected] s. iijima , _ nature _ , * 354 * ( 1991 ) 56 .
b.i.yacobson and r.e .
smalley , _ am .
* 85 * ( 1997 ) 324 .
dresselhaus , g. dresselhaus and p.c .
eklund , _ science of fullerenes and carbon nanotubes _
( academic press , san diego , 1996 ) .
r. saito , m. fujita , g. dresselhaus and m.s .
dresselhaus , _ appl .
* 60 * ( 1992 ) 2204 .
mintmire , b.i . dunlap and c.t .
white , _ phys .
* 68 * ( 1992 ) 631 .
x. blase , l.x .
benedict , e.l .
shirley and s.g .
lett . _ * 72 * ( 1994 ) 1878 .
wilder , l.c .
venema , a.g .
rinzler , r.e .
smalley and c. dekker , _ nature _ * 391 * ( 1998 ) 59 .
odom , j .- l
. huang , p. kim and c.m .
lieber , _ nature _ * 391 * ( 1998 ) 62 .
venema , j.w.g .
wilder , h.l.j .
temminck tuinstra , c. dekker , a.g .
rinzler and r.e .
smalley , _ appl .
* 71 * ( 1997 ) 2629 .
tans , m.h .
devoret and c. dekker , _ nature _ * 386 * ( 1997 ) 474 .
m. bockrath , d.h .
cobden and r.e .
smalley , _ science _ * 275 * ( 1997 ) 1922 .
j. roncali , _ chem .
* 97 * ( 1997 ) 173 .
schmidt , k.k .
baldridge , j.a .
boatz , s.t .
elbert , m.s .
gordon , j.h .
jensen , s. koseki , n. matsunaga , k.a .
nguyen , s.j .
windus , m. dupuis and j.a .
_ j. comput .
* 14 * ( 1993 ) 1347 - 1363 .
a. st - amant and salahub , d.r .
_ chem.phys.lett .
_ * 169 * ( 1990 ) 387 .
a. st - amant , _ ph.d .
thesis _ , universit de montral , 1992 .
demon - ks version 3.4 , m.e .
casida , c. daul , a. goursot , a. koester , l. pettersson , e. proynov , a. st - amant , d.r .
salahub , h. duarte , n. godbout , j. guan , c. jamorski , m. leboeuf , v. malkin , o. malkina , f. sim , and a. vela , demon software , 1996 .
s. huzinaga and j. andzelm , _ gaussian basis sets for molecular calculations _
, physical sciences data ; vol.16 , amsterdam ; new york : elsevier , 1984 . j.p . perdew and y.
wang , _ phys .
b _ * 33 * ( 1986 ) 8800 .
perdew and y. wang , _ phys.rev .
b _ * 46 * ( 1992 ) 12947 . m.j.s .
dewar and w. thiel
* 99 * ( 1977 ) 4899 - 4907 .
j. comput .
* 10 * ( 1989 ) 209 - 220 .
g. landrum , _ yaehmop _ ( yet another extended hckel molecular orbital package , cornell university , ithaca , ny , 1995 ) .
gmelin , _ handbuch der anorganishen chemie _ , 8th ed . , verlag chemie , weinheim , 1968 , vol.14b/2 , p.413 . k. yoshizawa , k. yahara , k. tanaka and t. yamabe , _ j. phys . chem . _
* 102 * ( 1998 ) 498 .
n.w . ashcroft and n.d .
mermin , _ solid states physics _
saunders college publishing , philadelphia , ( 1976 ) .
the ehmo results obtained for the 96 nanotube model are included in a previous publication submitted to _ chem .
_ by the authors .
w.j . hunt and w.a .
goddard iii , _ chem .
lett . _ * 3 * ( 1969 ) 414 .
m. kertsz , _ adv .
quantum chem . _ * 15 * ( 1982 ) 161 .
u. salzner , p.g .
pickup , r.a .
poirier and j.b .
lagowski , _ j. phys
. chem . _ * 102 * ( 1998 ) 2572 .
d. bakowies and w. thiel _ chem .
lett . _ * 192 * ( 1992 ) 236 - 242 .
dewar , h.s .
rzepa , _ j. am .
* 100 * ( 1978 ) 784 .
w. thiel _ tetrahedron _ * 44 * ( 1988 ) 7393 .
s . lee , m. kertsz , _ j. chem
* 88 * ( 1988 ) 2609 .
d. bakowies and w. thiel
_ * 113 * ( 1991 ) 3704 - 3714 .
dreizler and e.k.u .
gross , _ density functional theory _ , springler - verlag , berlin , 1990 .
u. salzner , j.b .
lagowski , p.g . pickup and r.a .
poirier , _ j. comput
. chem . _ * 18 * ( 1997 ) 1943 . | the electronic structure of finite - length armchair carbon nanotubes has been studied using several ab - initio and semi - empirical quantum computational techniques .
the additional confinement of the electrons along the tube axis leads to the opening of a band - gap in short armchair tubes . the value of the band - gap decreases with increasing tube length , however , the decrease is not monotonic but shows a well defined oscillation in short tubes .
this oscillation can be explained in terms of periodic changes in the bonding characteristics of the homo and lumo orbitals of the tubes .
finite size graphene sheets are also found to have a finite band - gap , but no clear oscillation is observed . as the length of the tube increases the density of states ( dos ) spectrum evolves from that characteristic of a zero - dimensional ( 0-d ) system to that characteristic of a delocalized one - dimensional ( 1-d ) system .
this transformation appears to be complete already for tubes 5 - 10 nm long .
the chemical stability of the nanotubes , expressed by the binding energy of a carbon atom , increases in a similar manner . 0.8 cm 1.0 cm abstract |
the instantaneous collapse of a two - particle quantum state vector has no covariant meaning , since the time - ordering of two different measurements depends on the reference frame for spatially separated events . in order to analyze this problem in a concrete situation , it is convenient to study entanglement from a relativistic point of view @xcite .
this analysis is also important for the study of relativistic quantum infomation tasks @xcite . in this work ,
we analyze the entanglement change , after a lorentz boost , of a system composed by two spin - one massive particles in an epr - like situation @xcite .
we compare our results with similar analyses done by friis et al . , and jordan et al . , for spin-@xmath0 particles @xcite . as well as these authors ,
we consider the particle s momentum as a discrete two - level variable , and calculate the entanglement change , with respect to different decompositions of the hilbert space , caused by a lorentz boost in a fixed direction .
we address the question of the dependence of the entanglement change on the initial spin entanglement and argue that , instead of the initial entanglement , quantum superposition plays the key role in the initial entanglement - final entanglement relation .
we focus on initial spin states for which no entanglement change occurs at all , and show that , for the situation under study , such states form a complete basis of the two - particle spin hilbert space . within this basis , maximally entangled invariant states , interesting for quantum information purposes , are explicitly obtained . on the other hand , since there are elements of this basis that suffer a distinct global phase change under the boost , an arbitrary linear combination of them will _ not _ in general be lorentz - invariant .
we generalize our results to higher spins . in section [ two ]
we introduce , for the sake of competeness , the relativistic transformation of momentum and spin states , while section [ three ] gives a description of the problem under study .
the results are presented and discussed in section [ four ] , where we also show the existence of a complete set of invariant states the two - particle spin space .
conclusions are presented in section [ five ] .
in this section we present the relativistic transformation of a single - particle state with momentum and spin degrees of freedom .
we use units such that @xmath1 .
the four - momentum of a particle in its own rest frame is given by @xmath2 where @xmath3 is the rest mass of the particle .
we label the four - vector components with the set of indices @xmath4 , and use boldface letters to denote three - vectors .
the quantum state of the particle in the rest frame is denoted by @xmath5 .
this state has a well - defined linear four - momentum , @xmath6 , and a well - defined spin projection along the @xmath7-axis , @xmath8 . in our case , @xmath9 .
let @xmath10 denote the standard lorentz boost that takes the rest frame four - momentum @xmath6 to the four - momentum @xmath11 .
the four - momentum eigenstates are defined in terms of the rest - frame state and the standard boosts as follows @xcite : @xmath12 where @xmath13 is a unitary representation of the poincar group on the complete momentum - spin hilbert space .
these states form a complete set and satisfy @xmath14 where @xmath15 is the @xmath16-th component of the four - momentum operator , and @xmath17 stands for the @xmath16-th component of @xmath11 . + now consider an arbitrary lorentz transformation @xmath18 .
the action of @xmath18 on the four - momentum eigenstate @xmath19 is given by @xmath20 here we have defined the wigner rotation as @xmath21 it is clear that @xmath22 is a pure rotation , since it leaves the four - momentum @xmath6 invariant : @xmath23 . the rotation angle @xmath24 , called the wigner angle , is a function of the magnitudes of the rapidities @xmath25 and @xmath26 that define , respectively , the boosts @xmath27 and @xmath18 , and under a suitable choice of setting ( see below ) may be written as @xcite : @xmath28 thus , under an arbitrary lorentz transformation @xmath18 , the momentum is changed from @xmath11 to @xmath29 and the spin part of the state transforms under the action of the rotation group , since , by the definition of the four - momentum eigenstates ( eq.([deffinition ] ) ) , the spin label remains unchanged after the standard boost @xmath30 .
hence , we have @xmath31 where @xmath32 is a spin-@xmath33 representation of the rotation group . in our case ,
@xmath33 = 1 .
we consider an epr - like situation with two spin-1 particles propagating in opposite directions with respect to each other , and compare the entanglement of the system before and after a lorentz boost on the state .
we take the direction of propagation to be the @xmath7-axis .
one of the particles is supposed to be under control of one observer , alice , while the other is supposed to be under control of another spatially separated observer , bob .
the entanglement is calculated with respect to different partitions of the complete hilbert space , which is a tensor product of alice s and bob s hilbert spaces .
each of the single - particle spaces is composed by spin and momentum degrees of freedom , so that the complete hilbert space can be written as a tensor product of four physically distinct subspaces : @xmath34 here , the superscrips @xmath35 and @xmath36 denote alice and bob subsystems , while the subscripts @xmath11 and @xmath37 denote the momentum and spin degrees of freedom , respectively . following @xcite , we calculate the entanglement change of the state , after a lorentz boost , with respect to three different partitions of @xmath38 : _ i ) _ the @xmath39 vs. @xmath40 partition , that is , the partition formed by alice s and bob s subsystems ; _ ii ) _ the @xmath11 vs. @xmath37 partition , obtained by tracing over the momentum or spin degrees of freedom of the whole state in @xmath38 ; and _ iii ) _ the entanglement change with respect to the partition formed by one given subspace opposed to the other three ; we shall call this the @xmath41 vs. @xmath42 partition .
we choose the initial state @xmath43 to be a pure and separable state in @xmath11 and @xmath37 , i.e. @xmath44 where @xmath45 and @xmath46 .
furthemore , we consider only sharp - momentum distributions , i.e. states where the momentum of each particle is concentrated around a definite value that can be @xmath47 , for propagation along the positive direction of the z - axis , and @xmath48 , for propagation along the opposite direction . for a study of wave packets in the spin @xmath0 case ,
see @xcite , while a wave packet analysis for the photonic case may be found in @xcite . the states in this work are considered to be normalized in the sense that @xmath49 , where the subscripts @xmath50 and @xmath33 can stand for the momentum labels @xmath51 and @xmath52 , and @xmath53 is the usual kronecker delta @xcite .
if the lorentz transformation @xmath18 of eq.([transformation ] ) were a pure rotation , then the associated wigner rotation would not depend on the momentum of the particles and would therefore have no effect on the entanglement of the states .
furthermore , since every lorentz transformation can be decomposed into a pure rotation and a pure boost , and pure rotations have no effect on entanglement , we consider in what follows the transformations @xmath18 to be pure boosts . on the other hand , we see from eq.([wigner rotation ] ) that if the lorentz boost @xmath18 is along the direction of propagation of the particles , which we take to be the @xmath7-axis , then the wigner rotation becomes the identity transformation , and causes no change in entanglement .
therefore , without loss of generality , we can choose the boost @xmath18 to be along the @xmath54-axis , perpendicular to the direction of propagation of the particles . in this seting ,
the spin - one representation of the wigner rotation takes the form @xmath55 where @xmath24 is given by eq .
( [ wigner angle ] ) .
let @xmath56 denote the state obtained from the original state @xmath43 after the lorentz transformation @xmath57 , i.e. , @xmath58 is the state described by an observer that is transformed with @xmath59 with respect to the observer that describes the state with @xmath43 .
the initial entanglement calculated from @xmath60 is to be compared with that obtained from @xmath61 . as an entanglement measure
, we use the linear entropy , defined by @xmath62 where the reduced density matrix @xmath63 is obtained from the original density matrix @xmath64 by tracing over all subsystems except the @xmath50-th .
note that the way a state transforms is independent of the entanglement measure chosen , and we shall see that weakly- or not - entangled states can transform into maximally entangled ones , so that if @xmath65 is not lorentz - invariant nor will any other appropriate entanglement measure be . as @xmath65 is a simpler measure to calculate , we choose to work with this measure ( for a whole family of entanglement measures the reader may see @xcite ) .
for the initial state ( [ initial state ] ) we choose a family of states @xmath66 and @xmath67 parametrized as follows ( this reduces to the parametrization found in @xcite for the case of spin-@xmath0 particles ) : @xmath68 [ spin state ] @xmath69 the three spin parametrizations cover the eigenstates of the operator @xmath70 in the composite hilbert space for definite values of the parameters @xmath71 , @xmath72 , and @xmath73 .
in this section we analize the entanglement change for each of the partitions of the complete hilbert space . in agreement with @xcite ,
entanglement is conserved for all states with respect to the @xmath39 vs. @xmath40 partition , due to the unitarity of the transformation induced in each of the subspaces @xmath74 and @xmath75 .
that is , since we have @xmath76 , where @xmath77 acts on the @xmath39 subsystem and @xmath78 on the @xmath40 subsystem , then @xmath79 and the linear entropy defined in eq .
( [ linear entropy ] ) remains unchanged when we trace out the @xmath39 or @xmath40 degrees of freedom .
physically , the unitary character of the @xmath57 transformation and the resulting conservation of entanglement is very important , since it is a necesary condition for the results of local measurements to depend , for _ every _ reference frame , only on the information given by the partial states @xmath80 and @xmath81 .
if entanglement failed to be conserved in this case , there would exist different quantum correlations for different inertial observers , in contradiction with the principle of relativity . in this way ,
every physical process that involves quantum correlations , such as quantum telportation , is independent of the reference frame in which it is analyzed and the experimental results will be the same for every observer .
entanglement with respect to the @xmath11 vs. @xmath37 partiton _ is not _
conserved in the general case .
this is because of the dependence of the wigner rotation on the momentum of the particle .
if we have an initial momentum superposition , then we have different transformations for the spin part of the system , one for each value of the momentum .
thus , the spin part of the system transforms according to a superposition of wigner rotations , and the resulting transformed state is no longer a tensor product of spin and momentum states . on the other hand , if the initial momentum state is not a superposition , which means @xmath82 or @xmath83 in eq .
( [ momentum state ] ) , there is only one wigner rotation for the state ; therefore , in this case , the resulting transformed state differs from the original one only by a unitary transformation , which laves the entanglement of the system unchanged . the change in linear entropy depends strongly on the wigner angle , as can be seen from fig .
[ par1partspmax ] . in this figure
we plot the entanglement change @xmath84 as a function of @xmath85 and @xmath86 for different values of @xmath24 .
we have chosen values of @xmath24 that differ considerably to stress the effect that this parameter has on the shape of the surface .
when both the speed of the particles and the speed of the boost @xmath18 approach the speed of light we have @xmath87 .
interestingly , in the limit of the speed of light ( @xmath87 ) , the entanglement change of the spin state given by eq .
( [ parametrization 1 ] ) produces exactly the same function of @xmath71 and @xmath72 as that obtained by friis et al in @xcite for the spin-@xmath0 composite state @xmath88 . in both parts of fig .
[ par1partspmax ] there are minima that correspond to the maximally entangled state @xmath89 . in all cases ,
the dependence of @xmath90 on the parameter @xmath91 is merely a change of scale , which is zero for non - superposed momentum states and is maximal for the initial momentum state @xmath92 .
for this reason , in what follows we consider only the case @xmath93 without loss of generality .
it is important to point out that the entanglement change is due to the linear superposition of initial momentum states and not due to the entnglement of the initial state @xmath94 . to see this ,
let us start with the separable state @xmath95 .
this state is not of the form of eq .
( [ momentum state ] ) but we use it here for the sake of argument . with this choice of initial momentum superposition , any state of the form @xmath96 , with @xmath67 being an arbitrary initial spin state , transforms into @xmath97 , where @xmath98 is the transformation induced by the boost in the two - particle spin subspace
when the first particle has momentum @xmath11 and the second has momentum @xmath99 .
this final state is entangled in the spin and momentum degrees of freedom since , in general , the transformations on spin space @xmath100 and @xmath101 are not equal .
we now return to our discussion of the @xmath11 vs. @xmath37 partition .
[ par2partspograndemax ] ( left ) shows the entanglement change for the spin initial state given by eq .
( [ parametrization 2 ] ) and a wigner angle @xmath102 .
the minima of the surface correspond to the bell - type state @xmath103 , which undergoes no entropy change for this particular wigner angle . when the parametrization given by eq .
( [ parametrization 3 ] ) is used ( figure [ par2partspograndemax ] ( right ) ) it is possible to generate entanglement for a large number of states under a lorentz transformation with the same wigner angle of @xmath104 .
here we have used @xmath105 .
the invariant state ( minima of plot ) is @xmath106 .
+ as in the previous case , entanglement is _ not _ conserved for all states in this type of decomposition and the maxima and minima of @xmath84 vary significantly with the wigner angle @xmath24 . to illustrate this point
, we take the state given in eq .
( [ parametrization 1 ] ) . in fig .
[ par1part1 ] we show the entanglement change , as a function of @xmath71 and @xmath72 , for the values values of @xmath107 and @xmath87 .
when the wigner angle equals @xmath108 , i.e.small " velocities , the entanglement change is maximum for the state @xmath109 , while the states @xmath110 and @xmath111 correspond to local maxima . above on the right of fig .
[ par1part1 ] is shown the same function as that on the left side but turned upside down , in order to look at the minima .
such minima correspond to the maximally entangled states @xmath112 . in spite that for this partition both of these states conserve entanglement for all values of the wigner angle , only the state @xmath113 is invariant under transformations of the form @xmath114 , that are , as we will show later , the kind of maps induced on the spin space by the lorentz boost . in this notation
, @xmath115 stands for the transformation given by eq .
( [ spin 1 rotation ] ) and acts on the @xmath116 subspace .
this invariance appears more clearly when we consider the @xmath11 vs. @xmath37 partition , in which @xmath113 has zero entanglement change for all values of @xmath24 , but @xmath117 _ does _ generate a change in entanglement for some values of the wigner angle . on the other hand , for @xmath87 ( fig .
[ par1part1 ] , below ) , the states @xmath118 and @xmath119 have a maximal change in linear entropy , while the state @xmath120 stays now at the bottom of the plot .
we see , thus , how the velocity of the particles , as well as the rapidity of the lorentz boost , play an important role in the entanglement change of the state . for the second parametrization , given in eq .
( [ parametrization 2 ] ) , the linear entropy behaves in a similar way as in the first parametrization . as can be seen from fig .
[ par2part1max ] , the states @xmath121 and @xmath122 , have a smaller entanglement change than the state @xmath120 for a wigner angle of @xmath102 .
we know from the last case that @xmath120 has a maximum entanglemet change for @xmath107 . as the wigner angle approaches @xmath123 , the change in linear entropy corresponding to the initial spin state @xmath120 decreases , while that corresponding to @xmath124 and @xmath125 increases continuously .
finally , in the limit of the speed of light , @xmath126 , the change in linear entropy corresponding to the state @xmath120 vanishes ( as we already know from the previous section ) , and the states @xmath124 and @xmath125 ( both with momentum part given by eq .
( [ momentum state ] ) with @xmath93 ) become , after the lorentz boost , maximally entangled states with respect to the @xmath41 vs. @xmath42 partition . it is interesting to note that the state @xmath127 is invariant under the maps induced by the lorentz boost studied here , as is the case of the state @xmath128 studied in the last section . both of these spin states are maximally entangled and , as they remain unchanged under the lorentz transformations presented here , they are interesting for quantum information purposes in an epr - like relativistic framework . in the physical situation considered here the boost direction is kept fixed and the particles momenta are always opposite in relation to each other ( and given by sharp distributions ) .
these two facts imply the that there exists a complete set of states in the two - particle spin hilbert space such that , for each state in this set , the action of the lorentz boost is merely a multiplication by a global phase factor .
to see this , first we note that the wigner rotation for a single particle is restricted to be an element of @xmath129 , the rotation group in the plane , since a fixed boost direction implies a fixed axis of rotation .
moreover , since the momenta of the particles are always opposite , by eq .
( [ wigner angle ] ) the angle of rotation for one particle will be of the same magnitude but opposite sign as that of the other particle .
then , the transformation on the two - particle spin space induced by the lorentz transformation will be of the form @xmath130 .
now we show that the set @xmath131 forms a representation of @xmath129 .
the product of two elements @xmath132 and @xmath133 @xmath134 @xmath129 is given by the sum of the parameters that define these elements @xmath135 thus , @xmath136 which proves our claim .
since the representations of @xmath129 are completely reducible we can , by means of a change of basis , reduce the transformation @xmath137 , for any @xmath24 , in terms of the irreducible representations of @xmath129 , which are of the form @xmath138 , with @xmath139 . for spin - one systems
, the transformation @xmath137 takes the form @xmath140 therefore , each element of the basis which diagonalizes @xmath141 changes under a lorentz boost only by a global phase factor . in this sense
, such states may be called invariant under the kind of lorentz boosts considered in this work ( always perpendicular to the momenta of the particles ) .
moreover , entanglement is conserved for every linear combination of states that transform under the same irreducible representation of @xmath129 , and all the information about the entanglement change of an arbitrary state is kept in the relative phase factors @xmath138 . for particles with arbitrary spin , the transformation @xmath141 is also diagonal .
the multiplicity @xmath142 of a given irreducible representation @xmath143 , that is , the number of times this representation appears in the ( reducible ) representation element @xmath141 , can be calculated as @xcite : @xmath144 where @xmath145 is the order of the group @xmath146 , @xmath147 is the character of the element @xmath148 for an arbitrary representation , and @xmath149 is the character of the element @xmath150 for the irreducible representation labeled by @xmath3 .
the sum is taken over all group elements and becomes an integral for the case of continous groups .
for the group @xmath129 , @xmath151 .
for a two spin-@xmath33 paticle system @xmath152 where the characters for the single - particle space transformations , @xmath153 and @xmath154 , are both given by @xcite : @xmath155 putting together eqs .
( [ multiplicity ] ) , ( [ tensor character ] ) and ( [ character ] ) yields , after a short calculation , @xmath156 so that for an arbitrary spin @xmath33 we have @xmath157 finally , and based on this last result , we note that the invariance of entanglement dos not depend , in general , on the initial entanglement of the spin state , but on its transformation properties . for example , the state @xmath158 is separable and also invariant under the @xmath141 transformations .
the entanglement with respect to the @xmath39 vs. @xmath40 partition is invariant under a lorentz transformation .
this fact is a direct consequence of the unitarity of the transformation @xmath76 .
the conservation of this kind of entanglement is fundamental for the consistency between quantum - mechanical predictions and relativistic transformations .
on the other hand , entanglement is not conserved neither for the @xmath11 vs. @xmath37 partition nor for the @xmath41 vs. @xmath42 partition .
this is due to the momentum - dependence of the wigner rotation , which induces , for momentum - superposed initial states , different transformations for different particles .
the entanglement change , for a given @xmath24 , reaches a maximum for a homogeneous momentum superposition ( @xmath93 ) , while it vanishes for non - superposed momentum states .
the dependence of the linear entropy change with respect to the initial spin state and to the wigner angle @xmath24 is more interesting . generally , the states for which the entanglement change is greater are separable spin states , while this quantity remains constant ( normally ) for states with maximal entanglement .
however , the initial spin entanglement is not a crucial factor regarding the entanglement change after a lorentz boost , since we can find separable spin states which conserve entanglement with respect to all partitions and all values of the wigner angle . in the same way
, there exist maximally entangled spin states that do not conserve entanglement with respect to the @xmath11 vs. @xmath37 partition . the transformation properties of the spin states is the crucial fact that determines the change in entanglement .
more precisely , what matters is the way in which the spin states transform under operations of the form @xmath130 , that are representations of the group @xmath129 .
in particular , the spin states that are invariant under this type of transformations conserve entanglement for every partition and all values of the wigner angle , with no regard to the initial momentum state .
since the action of the group @xmath129 is reducible in the spin state space , we can find an orthonormal basis of this space such that each state in the basis transforms , under a lorentz boost , only by multiplication of a factor of the form @xmath138 . in this basis
, the wigner rotation takes a simple form and all the information about the entanglement change lies in the relative phases @xmath159 , for @xmath3 , @xmath160 @xmath161 .
all linear combinations of states that transform under the same irreducible representation of @xmath129 are invariant under the kind of lorentz boosts considered in this work and therefore conserve their entanglement .
this result can be generalized to systems of arbitrary spin @xmath33 so that , in principle , invariant subspaces of an arbitrary dimension can be constructed , considering the adequate type of particles .
this fact may be useful for quantum information processes in which relativistic considerations are relevant .
this work was partially supported by dgapa - unam under project in102811 . | we analize the entanglement change , under a lorentz transformation , of a system consisting of two spin - one particles , considering different partitions of the hilbert space , which has spin and momentum degrees of freedom .
we show that there exists a complete set of states of the spin subspace in which the entanglement change of any state in the set is zero for all partitions and all values of the wigner angle .
moreover , these states only change by a global phase factor under the lorentz boost . within this basis ,
maximally entangled invariant states , interesting for quantum information purposes , are explicitly obtained . on the other hand ,
the entanglement in the particle - particle partition _ is _ lorentz invariant , thus protecting the consistency of quantum correlations and teleportation results .
we show how our results may be generalized to arbitrary spin . |
all observable information about a many - body quantum system is encoded in its correlation functions . for interacting systems in the thermodynamic limit this
information is , in general , only accessible through various approximations . until rather recently very few examples were known , where correlation functions could be calculated exactly . among the known examples there was none with short - range interactions of finite and tunable strength . in particular , for the large and prototypical class of integrable models that are solvable by bethe ansatz , the solution was limited to the spectral properties , but except for what can be concluded from the finite size corrections to the spectrum rather little was known about their correlation functions .
this situation is about to change .
extensive studies by several groups , mainly focusing on the xxz quantum spin chain , have eventually uncovered a structure which we believe to be typical for the bethe ansatz solvable models : the one - point and the neighbour correlation functions determine everything .
we call this phenomenon factorization . based on the factorization we can calculate for the first time short - range correlation functions with arbitrary accuracy , both , for finite temperatures in the thermodynamic limit @xcite , and for arbitrary finite length in the ground state @xcite .
this has applications e.g. in the determination of esr - line shifts @xcite and is useful for estimating the errors associated with standard numerical techniques like the quantum monte carlo algorithm @xcite .
factorization appeared first @xcite as the explicit factorization of certain multiple integrals @xcite describing the density matrix elements of the isotropic heisenberg chain in the ground state with no magnetic field applied .
a similar factorization for the ground state correlation functions of the xxz chain in the massive regime , i.e. for the model considered in this article , was carried out in @xcite .
after a multiple integral representation for finite temperature correlation functions had become available @xcite , the factorization of the integrals for short distances was extended to this case @xcite . following the success with the direct factorization of the integrals a deep exploration into the algebraic structure of the static correlation functions of the xxz chain
was conducted by boos , jimbo , miwa , smirnov and takeyama , culminating in the works @xcite , where the factorization of all static correlation functions of the xxz chain was proved in a very general setting , including the finite temperature and finite magnetic field cases . in all the recent works cited above it was crucial not to deal directly with the spin chain hamiltonian , but with the associated six - vertex model @xcite .
the six vertex model naturally carries inhomogeneity parameters in horizontal and vertical directions and can be distorted by a disorder field without loosing its distinctive feature , the integrability . in this setting the density matrix of a finite segment of the xxz chain is naturally generalized to include the inhomogeneity parameters and the strength @xmath6 of the disorder field .
the vertical inhomogeneity parameters @xmath7 and the disorder field regularize the expression for the density matrix .
the horizontal spectral parameters allow one to introduce the temperature into the model @xcite and to adjust the hamiltonian @xcite .
the density matrix of the inhomogeneous six - vertex model with disorder field depends polynomially on only two complex functions @xmath8 and @xmath9 which are efficiently described in terms of the solutions of integral equations @xcite .
the calculation of the coefficients of the polynomials is a purely algebraic problem .
it is related to the construction of a special fermionic basis @xcite on the space of local operators acting on the space of states of the spin chain .
we call this the algebraic part of the problem and the calculation of the functions @xmath10 and @xmath11 the physical part of the problem , since all dependence on the physical parameters , like temperature magnetic field or boundary conditions , is in these two functions . for the time being no efficient algorithm for the algebraic part of the problem is known .
this limits the range of the correlation functions we are actually able to calculate to a few lattice sites . in @xcite
we calculated the coefficients for the two - point functions for up to four lattice site by brute force computer algebra .
then we solved the physical part of the problem in the massless regime . at a late stage of the calculation the homogeneous limit and the limit of vanishing disorder field had to be taken in order to obtain the correlation functions of the spin chain . due to the singular nature of this limit
the first derivative of @xmath11 with respect to @xmath6 and various derivatives with respect to the inhomogeneity parameters appear .
the final expressions are polynomials in three functions @xmath11 , @xmath12 ( basically the @xmath6 derivative of @xmath11 ) and @xmath10 and their derivatives .
the coefficients in these polynomials are rational functions in the deformation parameter @xmath13 , which determines the anisotropy of the xxz chain , and are algebraic and universal .
therefore the expressions derived in @xcite for the two - point functions in the massless regime also apply in the massive regime , if one replaces the physical part , @xmath11 , @xmath12 and @xmath10 appropriately .
this is the main subject of this work .
we reformulate the expressions for the functions @xmath11 , @xmath12 and @xmath10 originally obtained in @xcite in a way that is convenient for numerical calculations in the massive regime @xmath14 of the xxz chain . combining this with the general formulae and the numerical results from @xcite we obtain accurate data for the two - point functions in the full parameter plane of the infinite antiferromagnetic chain ( anisotropy @xmath15 and magnetic field @xmath2 arbitrary ) .
they exhibit a surprisingly rich non - monotonic behaviour .
we should comment on the level of mathematical rigour of our results .
in @xcite we conjectured the physical part of the problem as well as the contribution to the algebraic part connected with the one - point functions in the limit , when @xmath16 .
meanwhile we know @xcite the exact physical part even for finite @xmath6 and we could show @xcite that it reproduces our conjecture for @xmath16 .
the algebraic part for distances up to four lattice sites was checked in @xcite by several independent means ( comparison with the high temperature expansion of the multiple integrals @xcite , direct numerical computation , consideration of various limits ) .
now , as we may infer from @xcite , it is clear that the exponential formula of @xcite is valid for arbitrary distances and zero magnetic field and at least for the distance of two lattice sites at finite magnetic field . on the other hand , the work @xcite offers a future way for the rigorous calculation of the algebraic part which needs no exponential form of the density matrix and will hopefully lead to the calculation of static correlation functions for points at larger distances .
we consider the hamiltonian @xmath17 of the xxz chain in the massive antiferromagnetic regime ( @xmath18 and @xmath0 ) .
the @xmath19 , act locally as pauli matrices , @xmath20 is the anisotropy parameter and @xmath21 the exchange coupling .
the xxz hamiltonian preserves the @xmath22-component of the total spin @xmath23 thus , the magnetization in @xmath22-direction is a thermodynamic quantity , and the thermal equilibrium of the finite system is characterized by the statistical operator @xmath24 depending on the temperature @xmath25 and the external magnetic field @xmath2 . in the thermodynamic limit @xmath26 the system severely simplifies , since ,
from the six - vertex model point of view , a single eigenstate of the so - called quantum transfer matrix determines the state of thermodynamic equilibrium and hence all static correlation functions @xcite .
clearly the naive limit @xmath27 makes no sense in ( [ defrhon ] ) . to perform this limit in a sensible way we fix integers @xmath28 with @xmath29 and
introduce the density matrix @xmath30 } ( t , h)=\lim_{n\rightarrow\infty } \operatorname{tr}_{-n+1,\ldots , m-1,n+1,n+2,\ldots , n}\rho_n(t , h)\ ] ] of the segment @xmath31 $ ] of the infinite chain which is well defined .
it has the reduction properties @xmath32 } ( t , h ) = d_{[m+1,n ] } ( t , h ) { \ , , } \quad \operatorname{tr}_n d_{[m , n ] } ( t , h ) = d_{[m , n-1 ] } ( t , h ) { \ , .}\ ] ] we consider the vector space @xmath33 of operators on the infinite chain which act non - trivially only on a finite number of lattice sites ( @xmath34 for arbitrary fixed @xmath35 and @xmath36 is an example of such an operator ) . on this space
we define a map @xmath37 } : { \cal w } \rightarrow \bigl ( { \mathbb c}^2 \bigr)^{\otimes ( n - m + 1)}$ ] by @xmath38 } { \cal o } = \dots { { \textstyle \frac{1}{2}}}\operatorname{tr}_{m - 2 } { { \textstyle \frac{1}{2}}}\operatorname{tr}_{m - 1 } { { \textstyle \frac{1}{2}}}\operatorname{tr}_{n + 1 } { { \textstyle \frac{1}{2}}}\operatorname{tr}_{n + 2 } \dots { \cal o } { \ , .}\ ] ] it restricts the action of @xmath39 to the interval @xmath31 $ ] . because of ( [ densredu ] ) the definition @xmath40 , @xmath41 } ( t , h ) \operatorname{\bf trf}_{[m , n ] } { \cal o}\ ] ] makes sense and determines the thermal expectation value of @xmath39 . the functional @xmath42 has the natural interpretation of the statistical operator on the space @xmath33 and may be thought of as the thermodynamic limit of @xmath43 , equation ( [ defrhon ] ) .
the functional @xmath42 allows for a generalization within the framework of the six - vertex model with disorder field . in the seminal paper @xcite
this generalization was denoted @xmath44 ( with @xmath45 referring to the magnetic field in certain units ) .
it is this functional which depends on only two functions @xmath8 and @xmath9 and which clearly separates the physical and the algebraic part of the problem . for details
we refer the reader to @xcite . in this work we are dealing with applications .
we shall only need an understanding of the general idea of factorization and the concrete formula obtained in @xcite .
for this reason we refrain in the following from any sophisticated mathematical notation and simply write @xmath46 for thermal expectation values on the infinite chain .
as explained in @xcite the limit @xmath16 leads to three functions @xmath47 , @xmath48 and @xmath49 that determine the correlation functions of the xxz chain .
a relatively simple description of these functions in terms of an auxiliary function @xmath50 and a generalized magnetization density @xmath51 was given in @xcite .
we call it the @xmath50-formulation .
the @xmath50-formulation is useful for deriving multiple integral formulae @xcite and for studying the high - temperature expansion of the free energy and the correlation functions @xcite .
it allows for a rather uniform description of the massless and the massive xxz chain , the only difference between the two cases being a different choice for the so - called canonical integration contour in the complex plane @xcite . for numerical calculations of thermodynamic properties
@xcite or short - distance correlators from the multiple integral @xcite one has to change to a different formulation we refer to as the @xmath52-formulation .
it needs pairs of functions , but the defining integral equations involve only straight contours .
it is here where one has to distinguish between the massless and the massive case , when it comes to numerical calculations .
two separate computer programs are needed .
in @xcite we studied the massless case .
here we proceed to the massive case .
we start from the @xmath50-formulation of the physical part proposed in @xcite and switch to the @xmath52-formulation .
this is a standard procedure which ( in the massive case ) basically requires the application of fourier series and the convolution theorem .
as far as the integral equations are concerned the reader may e.g. refer to @xcite .
then it is rather obvious how to proceed in a similar way with the expressions for the functions @xmath10 , @xmath11 and @xmath12 ( equations ( 15 ) , ( 17 ) and ( 20 ) for @xmath53 in @xcite ) .
for this reason we just state the results .
let us define a pair of auxiliary functions @xmath54 , @xmath55 as the solution of the non - linear integral equations ( nlie ) [ eq : nlie ] @xmath56
\ln{\bar{\mathfrak{b}}}(x ) = & \frac{h}{2 t } - \frac{2j\operatorname{sh}(\eta)}{t}d(x ) + \kappa\ast\ln\left(1+{\bar{\mathfrak{b}}}\right ) ( x ) - \kappa^+\ast\ln\left(1+{\mathfrak{b}}\right ) ( x)\end{aligned}\ ] ] with @xmath57 denoting a convolution .
the integral equations are specified by the integration kernels @xmath45 , @xmath58 and the driving terms , which contain a single transcendental function @xmath59 .
note that the physical parameters temperature @xmath25 , coupling @xmath21 and magnetic field @xmath2 enter into the calculation of the correlation functions only through the driving terms of the nlie . the functions @xmath59 and @xmath45 , @xmath58 , as well as some other functions occurring in other integral equations below , have simple fourier series representations which arise naturally in the derivation of the non - linear equations and which are also useful for solving them numerically , @xmath60 { \ , , } \\
\kappa(x ) & = \sum_{k=-\infty}^\infty \frac{e^{-\eta|k| + 2ikx}}{2\operatorname{ch}(\eta k ) } { \ , , } \quad \kappa^\pm(x ) = \kappa(x\pm i\eta^\mp ) { \ , .}\end{aligned}\ ] ] alternatively the kernel and the function @xmath59 can be realized in terms of special functions . for the function @xmath59
we find @xmath61 where @xmath62 is one of the jacobi elliptic functions @xcite , and @xmath63 is the complete elliptic integral of the first kind .
the integration kernel @xmath45 can be expressed in terms of the normalized trigonometric gamma function @xmath25 introduced in @xcite . in order to define it we first of all recall the definition of the @xmath13-gamma function @xcite , @xmath64
then , for @xmath65 , @xmath66 and @xmath67 { \ , .}\ ] ] in addition to the auxiliary functions @xmath54 and @xmath55 we need two more pairs of functions @xmath68 and @xmath69 to express @xmath11 , @xmath12 and @xmath10 by means of auxiliary functions .
both are solutions of linear integral equations involving @xmath54 and @xmath55 , @xmath70 g^-_\mu(x ) & = -d(x-\mu ) + \kappa\ast\frac{g^-_\mu}{1+{\bar{\mathfrak{b}}}^{-1}}(x ) - \kappa^+\ast\frac{g^+_\mu}{1+{\mathfrak{b}}^{-1}}(x)\end{aligned}\ ] ] and [ eq : gprime ] @xmath71 \notag { g'}^-_\mu(x ) = & - \eta c_{-}(x-\mu ) + \eta l\ast\frac{g^-_\mu}{1+{\bar{\mathfrak{b}}}^{-1}}(x ) - \eta l^+\ast\frac{g^+_\mu}{1+{\mathfrak{b}}^{-1}}(x)\\ & \mspace{144.mu } + \kappa\ast\frac{{g'}^-_\mu}{1+{\bar{\mathfrak{b}}}^{-1}}(x ) - \kappa^+\ast\frac{{g'}^+_\mu}{1+{\mathfrak{b}}^{-1}}(x ) { \ , .}\end{aligned}\ ] ] the new integration kernel @xmath72 and the new functions @xmath73 occurring in ( [ eq : gprime ] ) are conveniently described as fourier series , @xmath74 where we used the standard convention @xmath75 , and @xmath76 the functions @xmath77 , @xmath48 and @xmath49 which determine the inhomogeneous correlation functions can be expressed as integrals over the above functions .
the function @xmath78 with .
] @xmath79 is related to the magnetization @xmath80 which is the only independent one - point function of the xxz chain .
the function @xmath81 with @xmath82 also determines the internal energy @xcite .
the last function which is necessary to determine the correlation functions is @xmath83 where @xmath84 the physical meaning of @xmath12 is less intuitive .
its derivative appears in the two - point functions and is therefore related to certain neighbour correlators .
as shown in @xcite it is necessary to perform the homogeneous limit for the calculation of correlation functions .
this limit seems rather singular , because the coefficients coming from the algebraic part in general have poles , when two inhomogeneities coincide .
still , these poles are canceled by zeros coming from certain symmetric combinations of the functions @xmath85 , @xmath86 and @xmath87 in the numerators . in order to perform the limit one has to apply lhspital s rule
this finally leads to polynomials in the functions @xmath88 , @xmath89 , @xmath90 and in derivatives of these functions with respect to the inhomogeneity parameters evaluated at zero .
we denote derivatives with respect to the first argument by subscripts @xmath91 and derivatives with respect to the second argument by subscripts @xmath92 and leave out zero arguments for simplicity . then e.g. @xmath93 etc . for the examples in the next section the non - linear integral equations for @xmath54 and @xmath55 as well as their linear counterparts for @xmath94 and @xmath95
were solved iteratively in fourier space , using the fast fourier transformation algorithm .
the derivatives of @xmath94 and @xmath95 with respect to @xmath96 , needed in the computation of the respective derivatives of @xmath11 and @xmath12 , satisfy linear integral equations as well , which were obtained as derivatives of the equations for @xmath94 and @xmath95 .
the modifications due to the derivatives are particularly simple in fourier space .
in the sequel we shall restrict ourselves to the longitudinal and transversal two - point functions @xmath97 , @xmath98 for @xmath99 . for @xmath100 .
still , since this will give more bulky expressions of the type , , we refrained from the temptation of presenting them all . ]
the algebraic part for these correlation functions was already calculated in our previous papers @xcite and @xcite . here
we merely cite those results . for the nearest neighbour correlation functions
the following equations were obtained [ corrfncts2 ] @xmath101 the expansions for next - nearest neighbours read @xmath102 the length of the formulae grows rapidly with the number of lattice sites . for @xmath103 it reads @xmath104 \biggl\ { 384 q^4 \left(1+q^2\right)^2 \left(5 - 4 q^2 + 5 q^4\right ) \eta ^2 \omega \mspace{210.mu}\\ -8 \left(1+q^4 \left(52 + 64 q^2 - 234 q^4 + 64 q^6 + 52 q^8 + q^{12}\right)\right ) \eta ^2 \omega_{xy } \\
+ 192 q^4 \left(-1+q^2\right)^2 \left(1 + 4 q^2+q^4\right ) \eta ^2 \omega_{yy } \\ + \left(-1+q^2\right)^4
\left(1+q^4\right ) \left(1 + 4 q^2+q^4\right ) \eta^2 \bigl [ -4 \omega_{xyyy } + 6 \omega_{xxyy } \bigr ] \\[1ex ] - 768 q^4 \left(-1-q^2+q^6+q^8\right ) \eta { \omega'}_y \\ + 16 \left(-1+q^2\right)^3 \left(1 + 6 q^2 + 11 q^4 +
11 q^6 + 6 q^8+q^{10}\right ) \eta { \omega'}_{xyy}\\ -2 \left(-1+q^2\right)^5 \left(1 + 2 q^2 + 2 q^4+q^6\right ) \eta { \omega'}_{xxyyy } \displaybreak[0 ] \\
+ 8 \left(-1+q^2\right)^3 \left(1+q^2\right ) \left(1 + 6 q^2 + 34 q^4 + 6 q^6+q^8\right ) \eta ^2 \bigl[\omega_y^2 - \omega \omega_{xy } \bigr ] \\
+ \left(-1 - 4 q^2 - 22 q^4 - 12 q^6 + 12 q^{10}+22 q^{12 } + 4 q^{14}+q^{16}\right ) \eta ^2 \bigl [ - 6 \omega_{yy}^2 \\
+ 12 \omega_{yy } \omega_{xy } + 4 \omega_y \omega_{yyy } -12 \omega_y \omega_{xyy } -4 \omega \omega_{xyyy } + 6 \omega \omega_{xxyy } \bigr ] \displaybreak[0 ] \\
+ 16 \left(-1+q^2\right)^4 \left(1+q^2\right)^2 \left(1+q^2+q^4\right ) \eta \bigl [ \omega_{yy } { \omega'}_y - \omega_y { \omega'}_{yy } + \omega { \omega'}_{xyy } \bigr ] \\
+ \left(-1+q^4\right)^2 \left(1 + 5 q^2 + 6 q^4 + 5 q^6+q^8\right ) \eta \bigl [ 4 \omega_{xyyy } { \omega'}_y -6 \omega_{xxyy } { \omega'}_y -2 \omega_{yyy } { \omega'}_{yy } \\ + 6 \omega_{xyy } { \omega'}_{yy } + 2 \omega_{yy } { \omega'}_{yyy } -4 \omega_{xy } { \omega'}_{yyy } -6 \omega_{yy } { \omega'}_{xyy } + 4 \omega_y { \omega'}_{xyyy } -2 \omega { \omega'}_{xxyyy}\bigr]\\ + 3 \left(-1+q^4\right)^3 \left(1+q^2+q^4\right ) \bigl [ { \omega'}_{yyy } { \omega'}_{xyy } - { \omega'}_{yy } { \omega'}_{xyyy } + { \omega'}_y { \omega'}_{xxyyy } \bigr ] \biggr\}\end{gathered}\ ] ] and in the transversal case , @xmath105\\ + 192 q^2 \left(-3-q^2-q^4+q^8+q^{10}+3 q^{12}\right ) \eta { \omega'}_y \\
+ 8 \left(-1+q^2\right)^3 \left(1 - 12 q^2 - 25 q^4 - 25 q^6 - 12 q^8+q^{10}\right ) \eta { \omega'}_{xyy}\\ + 2 \left(-1+q^2\right)^5 \left(1 + 2 q^2 + 2 q^4+q^6\right ) \eta { \omega'}_{xxyyy } \displaybreak[0 ] \\
+ 16 q^2 \left(-1+q^2\right)^3 \left(17 + 7 q^2 + 7 q^4 + 17 q^6\right ) \eta^2 \bigl [ \omega \omega_{xy } - \omega_y^2 \bigr ] \\
+ q^2 \left(-5 - 4 q^2 - 13 q^4 + 13 q^8 + 4 q^{10}+5 q^{12}\right ) \eta ^2 \bigl [ 12 \omega_{yy}^2 -24 \omega_{yy } \omega_{xy } \\ -8 \omega_y \omega_{yyy } + 24 \omega_y \omega_{xyy } + 8 \omega \omega_{xyyy } -12 \omega \omega_{xxyy } \bigr]\\ + 8 \left(-1+q^2\right)^4 \left(1 - 9 q^2 - 8 q^4 - 9 q^6+q^8\right ) \eta \bigr [ \omega_{yy } { \omega'}_y - \omega_y { \omega'}_{yy } + \omega { \omega'}_{xyy } \bigr ] \displaybreak[0 ] \\
+ \left(-1+q^4\right)^2 \left(1 + 5 q^2 + 6 q^4 + 5 q^6+q^8\right ) \eta \bigl [ -4 \omega_{xyyy } { \omega'}_y + 6 \omega_{xxyy } { \omega'}_y\\ + 2 \omega_{yyy } { \omega'}_{yy } -6 \omega_{xyy } { \omega'}_{yy } -2 \omega_{yy } { \omega'}_{yyy } + 4 \omega_{xy } { \omega'}_{yyy } + 6 \omega_{yy } { \omega'}_{xyy } -4 \omega_y { \omega'}_{xyyy } + 2 \omega { \omega'}_{xxyyy } \bigr ] \displaybreak[0 ] \\ + 3 \left(-1+q^4\right)^3 \left(1+q^2+q^4\right ) \bigl[- { \omega'}_{yyy } { \omega'}_{xyy } + { \omega'}_{yy } { \omega'}_{xyyy } - { \omega'}_y { \omega'}_{xxyyy } \bigr ] \biggl\ } { \ , , } \end{gathered}\ ] ] with @xmath106 . using the representations via linear and non - linear integral equations for the functions @xmath11 and @xmath12 of the previous section we can determine high - precision numerical values for the various two - point correlators . figures [ fig : corr_von_h_normalized]-[fig : corr_von_h_normalized_fein ] show selected examples of connected two - point functions .
the longitudinal correlation functions @xmath107 ( @xmath108 ) are shown in the left panels , while the right panels show the transversal correlation functions @xmath109 .
note that the contributions from the one - point functions are somewhat trivial .
due to the translational invariance of the hamiltonian the longitudinal one - point functions do not depend on the site index , @xmath110 , and are given by for all @xmath111 .
the conservation of the @xmath22-component of the total spin , on the other hand , implies that @xmath112 . the above notation for the connected two - point functions is merely used for systematic reasons . in figure
[ fig : corr_von_h_normalized ] we show the dependence of the correlation functions on the magnetic field for @xmath113 and for several values of the temperature . at low temperatures
one can clearly see the two critical fields of the ground state phase diagram , figure [ fig : phasendiagramm_xxz ] .
the saturation field is at @xmath114 and the critical field at which the excitation gap opens can be calculated @xcite to be located at @xmath115 .
this can be seen even better in figure [ fig : corr_von_h_normalized_fein ] , where the @xmath116 correlation is shown for different low temperatures and for magnetic fields in the vicinity of the two critical fields . here
the curves for @xmath117 and @xmath118 are clearly distinguishable , which is not the case on the larger scale of figure [ fig : corr_von_h_normalized ] . for fields above the saturation field
@xmath114 the correlators have a fixed value as saturation sets in . for @xmath2 below the lower critical field we observe a constant behaviour as well , which in this case is due to the excitation gap . in between , however , for values of the magnetic field for which the system is critical , an interesting non - monotonic behaviour can be seen .
it is most pronounced for the correlators @xmath119 for which two local extrema exist if the temperature is sufficiently small .
if the temperature is too high , thermal fluctuations dominate and all correlations die out . in figure
[ fig : corr_von_h_normalized ] this is illustrated with the curve for @xmath120 in the @xmath121 case .
for @xmath122 the effect is similar . in figure
[ fig : corr_von_t_normalized ] we show the correlation functions for several values of the magnetic field as functions of the temperature . here
we observe a non - monotonic behaviour as well for intermediate magnetic fields . regarding the low temperature behaviour we notice that the connected correlation functions depend in a non - trivial way on the distance and on the magnetic field .
there exists a field for which the connected correlation functions have maximal modulus . for nearest neighbours the figures for the transversal correlation functions on the right side ,
show the largest values for @xmath123 , but for next nearest and next - to - next nearest neighbours the correlations for @xmath124 are more pronounced at low temperature .
for the field strength @xmath125 , which is above the saturation field , the connected correlation functions deviate the lesser from zero at intermediate temperatures the larger the distance is . the excitation gap can also clearly be seen in the left panel in figure [ fig : corr_von_t_normalized_fein ] , where we show @xmath126 for low temperatures and several magnetic fields in the vicinity of the lower critical field . with increasing magnetic field
the correlations start deviating from the @xmath127 line at lower and lower temperatures , until finally the zero temperature limit is different from the zero field case .
close to the critical field the system is particularly sensitive to small changes in the field strength , which can be well observed by comparing e.g. the correlation functions for @xmath128 , slightly above the critical field , and @xmath129 , slightly below the critical field .
the right panel again shows @xmath130 , but this time as a function of @xmath5 for fixed small temperature @xmath117 and several magnetic fields . for large values of @xmath5
the connected correlation function is independent of the magnetic field due to the excitation gap .
the effect of the phase transition from the massive fully polarized state to the critical antiferromagnet is also visible in this figure . for @xmath131 , for instance
, we observe an abrupt change at @xmath132 , which , at this field strength , is the critical anisotropy at which saturation sets in .
note that for producing the data for @xmath133 in the figure we used the formulation and computer implementation of our previous work @xcite .
the curves smoothly match at @xmath3 . and @xmath134 for different values of temperature @xmath135 and fixed anisotropy @xmath113 .
the rows are for @xmath108 . ] and @xmath134 for different values of the magnetic field @xmath136 and fixed anisotropy @xmath113 .
the rows are for @xmath108 . ] for different magnetic fields @xmath136 and fixed anisotropy @xmath113 at low temperatures .
the right panel shows @xmath137 for different magnetic fields @xmath136 as a function of the anisotropy @xmath5 at @xmath117 . ] for different low temperatures @xmath135 and fixed anisotropy @xmath113 .
the values of the magnetic field are in the vicinity of the critical fields . ]
the subject of this work is the study of correlation functions of the xxz chain for @xmath14 .
this includes the two extreme cases @xmath3 and @xmath4 .
the physically interesting isotropic point @xmath3 is difficult to access analytically .
in fact , the regularization of the density matrix by a disorder field mentioned in the introduction fails exactly at this point , and it seems that the picture of factorization described above needs a slight modification in the presence of a magnetic field , when additional independent functions , called ` moments ' in @xcite , appear . on the other hand ,
the numerics discussed in the previous section remains remarkably stable , if one approaches the isotropic point from above , and smoothly matches the numerics for approaching it from below @xcite . for @xmath4 or ,
more precisely , in the limit @xmath138 for finite @xmath139 the xxz hamiltonian ( [ ham ] ) turns into the hamiltonian of the ising chain @xmath140 the ising chain may be viewed as a classical one - dimensional model of statistical mechanics with @xmath141-transfer matrix ( see ( [ tising ] ) below ) . for this reason
its correlation functions can be calculated explicitly @xcite , [ eq : ising_corr_and_mag ] @xmath142 an interesting and non - trivial test of the formulae of the previous section is to reproduce these results analytically and numerically in the limit @xmath143 . a similar exercise starting from the multiple integral formulae of @xcite
was carried out in @xcite . in our case at hand
we start with the auxiliary functions in the @xmath52-formulation which turn into [ eq : bbq_ising ] @xmath144 in the ising limit .
note that , in this special case , they do not depend on the spectral parameter .
this is also true for the functions @xmath145 which moreover do not depend on @xmath96 anymore .
the same is therefore true for @xmath146 and @xmath147 similarly , for @xmath69 , for which a rescaling is necessary to obtain a sensible limit , one obtains @xmath148 note that this rescaling is unproblematic for @xmath149 due to .
thus , we obtain @xmath150 the above leads to the correct results for @xmath151 and @xmath152 .
for the next - to - nearest neighbour and all higher correlation functions , however , some of the coefficients in the factorized form of the correlation functions may diverge in the ising limit , e.g. the term @xmath153 on the right hand side of .
all the terms where such type of divergence occurs have to be handled separately .
sticking with the example of equation we define @xmath154 h^\pm(x ) = & \lim_{\eta\rightarrow\infty}\operatorname{sh}(\eta)\partial_\mu^2 \frac{{g'}^\pm_{{\tilde{\mu}}}(x)}{\eta}\biggr|_{\mu=0}\end{aligned}\ ] ] implying @xmath155 with @xmath54 , @xmath55 according to . using the latter expressions
we conclude that @xmath156 which indeed reproduces when inserted into the equation for @xmath157 . after having shown that the ising limit is included in our representation of the correlation functions by means of solutions of nlie
, it appears natural to compare the analytic limit with numerical results for large @xmath5 . in figure
[ fig : ising ] we show the temperature dependence of the connected two - point function @xmath158 for the heisenberg chain near the ising limit , for different values of the anisotropy parameter and the magnetic field .
in general , these curves do not depend on @xmath5 for large anisotropies and match the curves for the ising limit given by .
e.g. , the curves in the left panel for @xmath159 and @xmath160 or @xmath161 , respectively , match the corresponding ising curves to the precision of the line width in the figure . in the vicinity of @xmath162
, however , we observe large deviations from the low - temperature ising curves .
this behaviour is induced by a critical point at @xmath163 , @xmath162 in the zero temperature phase diagram of the xxz chain ( see figure [ fig : phasendiagramm_ising ] ) which separates regions of different asymptotics for the two - point correlation functions of the ising chain . the latter fact is most easily understood by inspection of the @xmath141 transfer matrix of the ising chain @xcite , @xmath164 its eigenvalues @xmath165 ( @xmath166 ) and
eigenvectors @xmath167 determine the partition function of the ising chain as well as its two - point functions .
in fact , when we wrote , we used the formula @xcite @xmath168 and the explicit expression for the eigenvalues and eigenvectors of @xmath169 , which the reader may easily calculate from ( [ tising ] ) . in order to understand the different zero temperature behaviour of the ising correlation functions it suffices to consider the low temperature asymptotics of the transfer matrix . using ( [ tising ] ) , ( [ isinggenform ] )
one easily distinguishes three different asymptotic regimes .
@xmath170 \bullet \quad h<4j_i : & & t \sim e^{2j_i / t } & \left(\begin{array}{cc } 0 & 1\\ 1 & 0\end{array}\right ) { \ , , } & \quad \left\langle \sigma^z_1 \sigma_{n+1}^z \right\rangle \sim ( -1)^n { \ , , } \\[1ex ] \bullet \quad h=4j_i : & & t \sim e^{2j_i / t } & \left(\begin{array}{cc } 1 & 1\\ 1 & 0\end{array}\right ) { \ , , } & \quad \left\langle \sigma^z_1 \sigma_{n+1}^z \right\rangle \sim \frac{1}{5 } + \frac{4}{5}\left(\frac{1-\sqrt{5}}{1+\sqrt{5}}\right)^n { \ , .}\end{aligned}\ ] ] thus , in the ising limit the absolute value of the ground state correlation functions is generally independent of the spatial separation , except at the critical field @xmath171 , where we see exponential decay due to the residual entropy resulting from an exponential degeneracy of the ground state . for finite @xmath5
this degeneracy is lifted by residual quantum mechanical interactions causing algebraically decaying correlations .
the critical point of the ising chain corresponding to a first order phase transition appears as a critical point in the ground state phase diagram of the xxz chain if we draw it in the @xmath2-@xmath172 plane for fixed @xmath173 , see figure [ fig : phasendiagramm_ising ] .
contrary to the ising case , there are always two critical fields if @xmath5 is finite .
they correspond to two second order phase transition lines at the boundary of the critical phase in the @xmath2-@xmath172 diagram .
this means that , away from the ising limit , even the ground state correlation functions depend continuously on the magnetic field . hence , for sufficiently low temperatures , there must be deviations from the ising curves for values of @xmath2 and @xmath5 belonging to the critical phase . in general the zero temperature limit should depend on the anisotropy and the magnetic field .
this can be seen in the left panel in figure [ fig : ising ] for @xmath174 , @xmath175 and @xmath176 . however , as the width of the critical phase gets smaller with increasing @xmath5 , one can find for each magnetic field , except for the critical field , a sufficiently large @xmath5 such that for larger anisotropies the correlation functions behave as in the ising model . for @xmath177
, there is always a deviation in the zero temperature behaviour between the ising chain and the xxz chain for finite @xmath5 .
see figure [ fig : ising ] , where the connected correlation functions are shown for @xmath175 , 2000 , 5000 and in the ising limit .
interestingly in this case the zero temperature limit does not depend on the ( finite ) value of @xmath5 .
this is a general behaviour as we observe that the correlation functions are asymptotically constant on straight lines ending in the ising critical point .
we can see this exemplarily in the right panel of figure [ fig : ising ] , where the connected correlation functions show the same zero temperature asymptotics for three different pairs of values of the anisotropy and the magnetic field . comparing these curves with those for the ising model with the same magnetic fields , we identify four different temperature regimes .
for very high temperatures the curves are independent of @xmath5 and @xmath2 .
then , for intermediate temperatures , they are independent of the ( large ) anisotropy and only depend on the magnetic field .
this is where the curves for finite @xmath5 and the curves of the ising model match .
next comes an regime where the correlation functions depend on the anisotropy and the magnetic field , and finally , for very low temperatures , only the product @xmath178 determines the correlation functions .
in fact , this is a rare example , where the full low temperature asymptotics of thermodynamic quantities and short - range correlation functions in the vicinity of a critical point can be worked out analytically @xcite . for large values of the anisotropy , fixed @xmath173 and different values of @xmath179 .
the labels in both panels are the tupels @xmath180 where ` ising ' denotes the analytic ising curves . ]
-@xmath172 plane for fixed @xmath139 and large @xmath5 . ]
another limit , which is interesting from a technical point of view , is the paramagnet @xmath181 . at a first glance
it might look rather trivial . yet , as the coefficients in section [ sec : examples ] depend on @xmath182 , the functions @xmath11 , @xmath12 and @xmath10 depend on @xmath182 as well , but the final result must not . in this sense
the paramagnetic limit is even more intricate than the ising limit , as , in contrast to the latter , all the summands in the algebraic part contribute .
this way we have an additional test for the coefficients in - .
we note that the physical part takes the form @xmath183 in the paramagnetic limit .
we have studied the short - range correlation functions of the xxz chain in the massive phase by means of the equations ( [ corrfncts2])-([sxsx4 ] ) representing them in factorized form . for this purpose we derived in section
[ sec : physpart ] a representation of the physical part of the correlation functions which is well suited for the implementation on a computer .
we obtained high - accuracy data for the correlation functions of the spin chain in the thermodynamic limit . together with our previous results
@xcite we can now access the full parameter plane of the infinite antiferromagnetic chain ( anisotropy @xmath15 and magnetic field @xmath2 arbitrary ) at arbitrary temperatures .
the short - range correlation functions show a surprisingly rich non - monotonous behaviour at intermediate magnetic fields and temperatures . at low enough temperatures
the critical lines of the ground state phase diagram can be read off from our data .
the numerics is stable in the isotropic limit @xmath184 and in the ising limit @xmath143 .
so far our method is still limited in the accessible range of the correlation functions . in order to extend this range , an efficient algorithm for the calculation of the algebraic part of the correlation functions is needed . since this is a well defined , purely algebraic problem though , we have little doubt that it will be solved in the future . for the exact calculation of the large - distance asymptotics at finite temperatures , on the other hand
, we believe that new insight will be needed ( for recent progress on the corresponding ground state problem see @xcite ) .
the authors are grateful to m. karbach for helpful discussions .
ct likes to acknowledge support by the research program of the graduiertenkolleg 1052 funded by the deutsche forschungsgemeinschaft . | we explore short - distance static correlation functions in the infinite xxz chain using previously derived formulae which represent the correlation functions in factorized form .
we compute two - point functions ranging over 2 , 3 and 4 lattice sites as functions of the temperature and the magnetic field in the massive regime @xmath0 , extending our previous results to the full parameter plane of the antiferromagnetic chain ( @xmath1 and arbitrary field @xmath2 ) .
the factorized formulae are numerically efficient and allow for taking the isotropic limit ( @xmath3 ) and the ising limit ( @xmath4 ) . at the critical field separating the fully polarized phase from the nel phase , the ising chain possesses exponentially many ground states .
the residual entropy is lifted by quantum fluctuations for large but finite @xmath5 inducing unexpected crossover phenomena in the correlations . |
high amplification gains are required in micro - pattern gaseous detectors ( mpgd ) in order to achieve a comfortable signal to noise ratio , therefore improving the performance of detectors in terms of spatial resolution and efficiency .
the dense electron avalanches achieved at high gains with few primary electrons entail the risk of producing a discharge at the cathode of the detector when the critical electron density of @xmath1 electrons per avalanche is reached ( related to the raether limit @xcite ) .
discharges might affect the detector response in different ways ; * _ reducing its operating lifetime _ due to intense currents produced in short periods of time , heating and melting the materials at the affected regions , * _ damaging the readout electronics _ which have to support huge current loads , * _ increasing the detector dead - time _ resulting from a _ _ discharge of the cathode__.
it was first observed in rpc - type detectors that the introduction of a high impedance resistive coating at the anode limits the detector current , constraining the spark process to the streamer phase during a time interval of at least some microseconds . thus reducing the total amount of charge released @xcite .
furthermore , the limited discharge current affects the field locally , thus reducing the dead - time of the detector which remains operative in non - affected regions .
micromegas detectors were introduced in 1996 @xcite as a good candidate for high particle flux environments , and spark studies with detectors based on micromegas technology were also carried out @xcite . recently , additional efforts have pushed the development of resistive strip micromegas detectors in order to further increase its robustness in high particle flux environments by limiting spark discharges in the same way as it was done for rpcs . in particular , the mamma collaboration @xcite is developing large area micromegas detectors and introduced the resistive coating technique @xcite for the upgrade of the hl - lhc@xmath2s@xmath3 ) ] .
the existing micromegas technology @xcite allowed the mamma collaboration to investigate detector prototypes with different resistive coating topologies @xcite .
this new type of detectors will be installed in atlas at the new small wheel project .
thus , it should be proved to be long term radiation resistant .
the introduction of a new technology made of new materials adds some uncertainties of operation during long periods of time in intense particle flux environments .
the results that we report here represent the first ageing tests with this type of detectors using different types of highly ionizing radiations .
an ageing study of resistive - anode detectors is required to operate these type of detectors at the hl - lhc . for this study , two new identical micromegas prototypes ( named r17 )
were used , lent by the mamma collaboration and built at the cern workshop .
these detectors are based on a resistive strips technology @xcite with a 2-dimensional readout .
the x and y readout strips are in copper .
the top y strips have been covered by a 60@xmath4 m thick insulating coverlay .
the 35@xmath4 m thick resistive strips are placed on top of this layer parallel to the x strips . both resistive and copper strips have a pitch of 250@xmath4 m and a width of 150@xmath4 m .
the resistivity along the strips and boundary resistance value was measured during the fabrication process . the first detector , r17a , showed a linear resistivity of 45 - 50 m@xmath5cm@xmath3 , and a boundary resistance of 80 - 140 m@xmath5 .
the resistivity obtained for the second detector , r17b , was comparable with a linear resistivity of 35 - 40 m@xmath5cm@xmath3 and a boundary resistance of 60 - 100 m@xmath5 . a first characterization in ar+co@xmath7 mixtures , and ar + 7% co@xmath7 for the test beam performance .
] showed that the gain behaved as expected , and the results obtained with both detectors , r17a and r17b , were comparable .
figure [ resist_princ ] shows the gain curves in these mixtures .
ageing tests took place at different cea - saclay facilities , one prototype ( r17b ) was kept unexposed as a reference and the other ( r17a ) was exposed to different types of radiations including x - rays , cold neutrons , high energetic gammas , and alphas .
these tests will be described in the following sections .
the prototype under test was placed inside a cage with a high intensity x - ray generator , which emits in all the x - range up to few kev , with a copper fluorescence at 8kev .
a mask was prepared by using an aluminum plate with a 4 @xmath8 hole aperture .
the plate was placed on top , and it was always fixed in the same position .
the radiation exposure tests aim to accumulate an amount of charge comparable to the values that will be integrated during the lifetime of the hl - lhc .
the estimation of the total charge produced at the hl - lhc in the muon chambers of atlas is based on the energy deposit @xmath9 of a minimum ionizing particle ( mip ) in a 0.5 cm micromegas conversion gap . in our gas mixture @xmath10 kev .
considering the detector gain @xmath11 , the charge produced by each incident particle at the hl - lhc @xmath12 where @xmath13 ev is the mean ionization energy of the gas . taking into account a nominal gain of g=5,000 ,
the charge produced per mip yields @xmath14 fc . assuming the expected rate at the hl - lhc future muon chambers will be 10 khz/@xmath8 , the total charge generated in 5 years of operation ( 200 days / year )
will be 32.3 mc/@xmath8 .
the flux produced at the x - ray generator will accumulate an amount of charge well above this value for an exposure of a few days . the detector has been exposed for more than 20 days with a gain of 5,000 and a gas flow of one renewal per hour .
the current remained stable ( see figure [ xray_current ] ) , proving the detector gain was not affected on the whole irradiation period , which corresponds to 21.3 days of exposure and an integrated charge of 918 mc , that is 5 years of hl - lhc with a security factor above 7 .
the gain of both detectors was measured using an @xmath15fe source at different positions , before and after exposure .
we used a dedicated 9-hole mask which covers the full active area of the detector in one of its axis .
these measurements took place before the ageing period ( 26-sep-2011 ) , when the grounding connectors were removed ( 8-oct-2011 ) and after the exposure ( 19-oct-2011 ) ( further details can be found at references @xcite ) .
the relative gain at each position for these three sets of measurements is plotted in figure [ gain_profile ] and shows that the gain profile is compatible with previous measurements .
moreover , the exposed detector region does not show any significant difference in relative gain compared to the other , non - exposed regions .
the x - rays irradiation had therefore no effect on the detector response .
neutron irradiation took place at the orphe reactor in cea - saclay .
the reactor , which operates only for research purposes , is connected to different lines which guide different fluxes of cold neutrons produced in the reactor .
the line where the detector was installed provides a neutron beam of about 8@xmath1610@xmath17@xmath2s@xmath3 , with energies in the range of 5 - 10mev within an area of a few @xmath8 .
different irradiation periods were scheduled with an increased time exposure .
the mesh current remained stable and at the same level in each of these periods ( see figure [ neutron_current ] ) .
the expected neutron flux at the csc ( cathode strip chamber ) in atlas is about 3@xmath1610@xmath18 @xmath2s@xmath3 .
the total exposure time of the prototype r17a was more than 40 hours , accumulating a total amount of neutron flux which is equivalent to 5 years of operation of the hl - lhc with a safety factor well above 10 .
before and after the neutron tests the gain was measured using the same 9-hole mask used for the tests described in the previous section .
the gain profile is compatible and the performance of the detector shows no degradation with respect to the measurements before neutron irradiation ( see figure [ neutron_profile ] ) . after proving that the detector r17a was operating properly it was installed at the cocase @xcite gamma facility , which provides a high activity @xmath0co source , of about 500 mgy / h , emitting gammas at 1.17 mev and 1.33 mev .
the highest activity for gammas at the muon spectrometer is recorded in the forward csc region @xcite , with a flux which is below 1.8@xmath1610@xmath18@xmath2s@xmath3 . for 5 years of running hl - lhc ,
we consider a factor 5 in luminosity increase , with a safety factor 3 the integrated gamma flux results to be 2.3@xmath1610@xmath19@xmath2 .
the detector was placed at 50 cm from the source receiving an equivalent flux of 1.7@xmath1610@xmath20@xmath2s@xmath3 which should be uniformly distributed in the active area of the detector .
thus , time required to reach the expected gamma flux integrated for 5 years of hl - lhc would be of 16 days in cocase .
the detector was irradiated during 480 hours between march,22@xmath21 and april,11@xmath22 2011 .
the integrated charge during this period was 1.48c at a mean mesh current of 858 na .
figure [ gamma_current ] shows the evolution of the mesh current which fluctuates around the mean value within a 5% , variation which can be perfectly attributed to environmental effects . the gain profile before and after exposure shows a reasonable reproducibility ( see figure [ gamma_profile ] ) .
fe peak energy resolution ) . ]
after the completion of the above described exposures , the r17a and r17b prototypes were installed in the h6 cern - sps beam line .
the beam consists of positively charged pions at 120gev / c in spills of about 10s every 48s .
the beam intensity was typically 5@xmath2310@xmath18/spill over an area of 20mm@xmath1610 mm .
the daq recorded around 800 events per spill .
the intention of this last test was to determine any hint of ageing of the r17a by comparing the spatial resolution ( sr ) and efficiency obtained with respect to the reference detectors .
the r17a and r17b were installed in the existing micromegas telescope line , as shown in figure [ beam_setup ] . the micromegas telescope ( tmm2 , tmm3 , tmm5 , tmm6 ) is used as reference for the track reconstruction of the beam .
the detectors under test were placed 2 m downstream of tmm6 . .
the line is shared with other detectors which were under test at the same time and have been excluded in this drawing for simplicity .
our resistive prototypes were placed behind the telescope . ]
the data were taken within two weeks at the end of october 2012 .
three different regions were exposed to the beam in order to compare zones which received different types of radiations .
we took data at each of these regions for several values of the amplification field by varying the mesh voltage , observing no significant difference between zones .
figure [ efficiency ] ( left ) shows the averaged sr as a function of the mesh voltage pointing to an optimum sr at about 550v , which corresponds to gains slightly above 10,000 .
figure [ efficiency ] ( right ) shows the detection efficiency of both detectors as a function of the absolute gain .
the efficiencies were determined as the probability to detect one charge cluster when a track is found in each of the _ four _ telescope detectors .
both detectors reach efficiencies of about 99.5% for the highest values of the gain , proving that there is no visible degradation effect in these measurements ( further details can be found at @xcite ) .
a 2d - readout resistive strip detector was tested to different intense irradiation sources : x - rays , high energy gammas , neutrons and alphas @xcite . for each radiation type ,
the total charge or interactions produced during the ageing process was well above the expected levels at the hl - lhc for 5 years of operation with an additional safety factor .
in addition we investigate the performance of these detectors to the cern pion beam test facility .
the results obtained are comparable to those obtained with a non - irradiated detector of same construction .
we did not observe any degradation .
the intense irradiation of r17a detector did not harm seriously the performance of the detector in terms of efficiency and sr , similar to the reference telescope .
we present it as a definitive proof that irradiations of different nature , which will be present at the hl - lhc , do not affect this new technology .
the values obtained for the spatial resolution and detection efficiency are reasonably good considering that the setup was not optimized to minimize the error on the track definition .
_ , _ aging studies of micromegas prototypes for the hl - lhc _ , _ 2012 jinst 7 c01041_. f. jeanneau _
_ , _ performances and ageing study of resistive - anodes micromegas detectors for hl - lhc environment _ , _ nuclear science , ieee transactions on _ , vol.59 , no.4 , pp.1711 - 1716 , aug . 2012
_ , _ characterization of microbulk detectors in argon- and neon - based mixtures _ , _ 2012 jinst 7 p04007_. r. chipaux and o. toson , _ resistance of lead tungstate and cerium fluoride to low rate gamma irradiation or fast neutrons exposure _ , cms technical note cms tn , 1995 , p. 95 - 126 .
s. baranov _ et al .
_ , _ estimation of radiation background , impact on detectors , activation and shielding optimization in atlas_,_atl - gen-2005 - 001 .
_ j. galan _
_ , _ an ageing study of resistive micromegas for the hl - lhc environment _ , submitted to jinst [ arxiv:1301.7648 ] . | resistive - anode micromegas detectors are in development since several years , in an effort to solve the problem of sparks when working in high flux and high radiations environment like in the hl - lhc ( ten times the luminosity of the lhc ) .
they have been chosen as one of the technologies that will be part of the atlas new small wheel project ( forward muon system ) .
an ageing study is mandatory to assess their capabilities to handle the hl - lhc environment on a long - term period .
a prototype has been exposed to several types of irradiations ( x - rays , cold neutrons , @xmath0co gammas ) up to an equivalent hl - lhc time of more than five years without showing any degradation of the performances in terms of gain and energy resolution .
beam test studies took place in october 2012 to assess the tracking performances ( efficiency , spatial resolution , ... ) .
results of ageing studies and beam test performances are reported in this paper .
resistive micromegas , detector ageing , mamma , atlas |
the nuclear equation of state ( eos ) entails in a broad sweep knowledge of the diverse properties of nuclear matter : its saturation density , energy per nucleon @xmath2 ( @xmath3 is the density ) , incompressibility @xmath1 , the symmetry energy and its density content , @xmath4 the symmetry coefficient @xmath5 , the symmetry slope parameter @xmath6 , the symmetry incompressibility @xmath7 and all the higher symmetry derivatives . a cultivated attention is naturally drawn in recent times to have a refined understanding of this nuclear eos from both experimental and theoretical sides .
the binding energies of stable atomic nuclei are the most accurately known experimental entities in nuclear physics , these supplemented with knowledge of giant monopole and dipole resonances followed by theoretical analysis have yielded some of the eos parameters like the saturation density @xmath8 of symmetric nuclear matter and @xmath9 and @xmath10 in reasonably tight bounds @xcite .
the knowledge of the symmetry derivatives @xmath11 and @xmath12 is still not very certain @xcite .
analyses of the different nuclear observables do not help much in removing the uncertainty .
correlation systematics of nuclear isospin with neutron skin thickness @xcite , isospin diffusion @xcite , nucleon emission ratios @xcite or isoscaling @xcite in heavy ion collisions , all yield values of the symmetry slope parameter @xmath13 that are not much in consonance with one another .
some attempt was recently made to constrain @xmath14 in tighter bounds from nuclear masses aided by microscopic calculations @xcite on neutron skin of heavy nuclei , it was found to give @xmath14 = 59 @xmath15 13.0 mev .
other recent attempts , from analysis of the isovector giant dipole and quadrupole resonances in @xmath16pb nucleus @xcite give @xmath14 = 43@xmath17 and 37@xmath1518 mev respectively .
this underscores the still unresolved uncertainty in getting to the value of @xmath14 and asks for newer avenues to understand it .
the present state of the art on symmetry energy and related parameters can be found in the topical issue on nuclear symmetry energy @xcite .
the eos parameters so mentioned pertain to only one density , the saturation density @xmath8 . if all of them are known
precisely , it is in principle possible to construct with a suitable energy density functional ( edf ) the nuclear eos @xmath18 where @xmath19 is the isospin asymmetry .
whereas the high density end of the eos would be of immediate value in understanding the dynamical evolution of the core collapse of a massive star and the associated explosive nucleosynthesis @xcite or the radii and lower bound of the maximum mass of cold neutron stars @xcite , the low - density end helps in getting a closer estimate of the neutron skin thickness or the neutron density distribution @xcite in neutron - rich nuclei .
further information on the eos parameters at densities other than @xmath8 may put the nuclear eos on firmer grounds , unfortunately , there are only a few of them . at density
higher than @xmath8 , information from experimental data has still large uncertainty @xcite ; at sub - saturation density , from giant dipole resonance analysis , a quantitative constraint on @xmath5 could be put as 23.3 mev @xmath2024.9 mev @xcite .
further information derived from theoretical analyses at around this density may be of added significance , ( i ) the energy per nucleon of neutron matter is @xmath21 mev at @xmath3= 0.1 @xmath22 @xcite and ( ii ) the density derivative of the nuclear incompressibility @xmath23 1100@xmath1570 mev where @xmath24 is @xmath25 0.7@xmath8 @xcite .
as is evident from the previous discussion , the plethora of nuclear eos failed to effectively well - constrain the density content of the nuclear symmetry energy and the nuclear incompressibility from fits to diverse microscopic nuclear data .
the reason lies in the choice of different sets of microscopic observables to be fitted .
the isoscalar and isovector quantities associated with nuclear matter , however , have emerged to be very well - constrained .
the isoscalar entities are i ) @xmath260.1 mev ) , ii ) saturation density @xmath27(=0.155 @xmath150.008 @xmath22 ) , where the pressure @xmath28 , iii ) the incompressibility coefficient @xmath29 = 240 @xmath1520 mev ) @xcite .
all these quantities refer to symmetric nuclear matter .
the isovector quantities are iv ) @xmath10([email protected] mev)@xcite and v ) @xmath31 @xmath32(=24.1 @xmath150.8 mev)@xcite . in this article
, we have tried to find how the input of the empirical knowledge of these quantities can be used to construct an edf for nuclear matter and to predict the as yet not so well - constrained density dependence of its symmetry properties in reasonably tighter bounds .
these isoscalar and isovector nuclear parameters effectively contain condensed experimental information on the bulk nuclear properties . as opposed to direct investigation of the microscopic properties of nuclei as done , _
e.g , _ in ref .
@xcite which can lead to somewhat different predictions depending on the observables chosen to be explored , the alternate edifice for the nuclear edf built in this article on the well established nuclear bulk parameters is so structured , as we see later , that it gives an easy and transparent look to the correlations of the predicted values of the density derivatives of the symmetry energy and the nuclear incompressibility to the input parameters .
furthermore , if the values of the density derivatives @xmath33 etc could be well settled by as yet other unexplored means and differ from our predicted values , the foundation _
i.e. , _ the values of the nuclear bulk parameters or the choice of the skyrme edf then become subject of fresher scrutiny .
computationally , our method is also much less intensive . to our knowledge ,
a comprehensive study of this kind has not been done before .
henceforth , the quantities corresponding to the densities @xmath8 and @xmath340.1 @xmath22 would be denoted with the subscripts 0 and 1 , respectively ( like @xmath35 , @xmath36 etc . ) .
the value of @xmath8 is an indirectly obtained entity . from acceptable skyrme energy density functionals ,
it is @xmath37 0.16 @xmath22 @xcite where as the relativistic mean - field models give a value of @xmath38 in the vicinity of @xmath37 0.15 @xmath22 @xcite .
our choice for @xmath8 covers this range .
the paper is organized as follows .
ii contains a brief discussion of the theoretical elements .
results and discussions are presented in sec .
iv contains the concluding remarks .
the skyrme framework is chosen for the energy density functional @xcite .
the energy per nucleon for nuclear matter is then , @xmath39\rho^{2/3 } \nonumber \\ & & + ( b_1+b_2\delta^2)\rho + ( c_1+c_2\delta ^2)\rho^{{\alpha+1 } } \nonumber \\ & & + \left[d_1 \left \ { \left(\frac { 1+\delta}{2}\right)^{5/3}+ \left(\frac { 1-\delta } { 2}\right)^{5/3}\right \ } \right . \nonumber \\ & & \left .
+ d_2 \left \{\left ( \frac { 1+\delta } { 2}\right)^{8/3}+ \left(\frac { 1-\delta } { 2}\right)^{8/3 } \right \}\right ] \rho^{5/3}. \end{aligned}\ ] ] the first term on the right hand side is the free fermi gas energy ; @xmath40 = 119.14 mev @xmath41 , where @xmath42 is the nucleon mass . for the chosen values of @xmath43 and @xmath8 , values of @xmath44 ranging only from 1/6 to 1/3 allow for an acceptable set of ( @xmath45 , @xmath46 ) @xcite where @xmath47 is the nucleon effective mass .
we therefore chose @xmath48 , this allows @xmath45 to lie in the acceptable range @xmath49 @xcite .
we take the median value of @xmath500.2 .
we are left with six unknown parameters , namely , @xmath51 and @xmath52 which would completely define the edf .
as already mentioned , we have , however , five equations , three from isoscalar entities and two from isovector entities .
the isoscalar equations are @xmath53 @xmath54,\end{aligned}\ ] ] and @xmath55.\end{aligned}\ ] ] the isoscalar equations yield the values of @xmath56 and @xmath57
. the isovector equations , evaluated at @xmath8 and @xmath58 ( = 0.1 @xmath22 ) are given by @xmath59 \rho^{5/3}.\end{aligned}\ ] ] to fix the remaining parameters , we need an extra condition in conjunction with the two isovector equations .
for this , we take that the energy per particle of isospin asymmetric nuclear matter is quadratic in the asymmetry parameter @xmath60 .
this condition is found to be an excellent approximation from nearly all energy density functionals @xcite and also from microscopic calculations in the bruckner - hartree - fock ( bhf ) formalism @xcite at all densities up to @xmath8 and a little beyond .
this implies that the difference between the symmetry energy coefficients defined by eq .
( 5 ) and the one by following equation @xmath61\rho^{5/3},\end{aligned}\ ] ] should be minimal . to achieve this minimality , we take the help of the equation for the symmetry slope parameter @xmath62 . from its definition @xmath63 ,
one gets from eq .
( 5 ) , @xmath64 , \label{eq : l1}\end{aligned}\ ] ] with a given value of @xmath14 ( = @xmath11 ) , from known @xmath35 and @xmath36 , at the two densities @xmath8 and @xmath58 , one can solve for @xmath65 and @xmath66 $ ] . since @xmath67 is known from the isoscalar equations , @xmath68 and @xmath52 are now obtained .
.[tab1]parameters of the energy density functional corresponding to the central values of the isoscalar and isovector inputs . [ cols="^,^,^,^ " , ] our results for the equivalent density shown as a shaded region as a function of mass number .
the blue triangles are from ref .
@xcite , the magenta line is from ref .
] the eos for symmetric nuclear matter ( upper panel ) and for pure neutron matter ( lower panel ) .
the red , yellow and green shaded regions represent the experimental data taken from ref .
the blue shaded regions are the eos obtained in this work .
see text for details . ]
to sum up , from consensus empirical inputs for values of some of the key nuclear parameters at saturation and sub - saturation densities , we have constructed a skyrme - type energy density functional for homogeneous nuclear matter .
this is then employed to understand the density dependence of the nuclear symmetry energy and incompressibility and to predict values for the important nuclear parameters like the symmetry slope parameter @xmath14 , the symmetry incompressibility parameter @xmath69 , the incompressibility slope parameter @xmath70 .
separate estimates of these quantities have been given from different perspectives ; sizeable uncertainties remain there . the structural edifice for the energy density functional built on
a few known input bulk parameters gives a coherence in the evaluated values of the observables ; their uncertainties can be constrained better provided the input bulk entities are known with better precision .
the general agreement of our eos with the experimental one at supranormal densities is interestingly striking .
the near concordance of our calculated lower bound of the maximum mass of neutron star with the experimental observation of a neutron star of mass @xmath71 is also very noticeable .
inclusion of exotic degrees of freedom in the interior of the star , however , softens the eos and lowers the value of @xmath72 though and this needs further investigation .
jnd acknowledges support from the department of science and technology , government of india .
g.c . would like to thankfully acknowledge the nice hospitality extended to him during his visit to sinp , when this work started .
the authors gratefully acknowledge the assistance of tanuja agrawal in the preparation of the manuscript . | from empirically determined values of some of the characteristic constants associated with homogeneous nuclear matter at saturation and sub - saturation densities , within the framework of a skyrme - inspired energy density functional , we construct an equation of state ( eos ) of nuclear matter.this eos is then used to predict values of density slope parameters of symmetry energy @xmath0 , isoscalar incompressibility @xmath1 and a few related quantities .
the close consonance of our predicted values with the currently available ones for the density dependence of symmetry energy and incompressibility gleaned from diverse approaches offers the possibility that our method may help in settling their values in tighter bounds .
extrapolation of our eos at supranormal densities shows that it is in good harmony with the one extracted from experimental data . |
atomic nuclei are made of two distinct particles , the protons and the neutrons .
these nucleons constitute more than 99.95% of the mass of an atom . in order to form a stable atomic nucleus
, a subtle equilibrium between the number of protons and neutrons has to be respected .
this condition is fulfilled for 259 different combinations of protons and neutrons .
these nuclei can be found on earth .
in addition , 26 nuclei form a quasi stable configuration , i.e. they decay with a half - life comparable or longer than the age of the earth and are therefore still present on earth .
these stable or quasi - stable nuclei are characterised by an approximately equal number of neutrons and protons for light nuclei , whereas going to heavier nuclei the neutron excess increases to reach n - z=44 for the heaviest stable nucleus @xmath1pb .
in addition to these 285 stable or quasi - stable nuclei , some 4000 - 6000 unstable nuclei are predicted to exist , the exact number depending on the theoretical model used .
close to 2500 nuclei have been observed already .
these unstable radioactive nuclei can be classified into seven categories depending on their decay modes : i ) @xmath2 emitters ( about 375 ) which decay by ejecting a @xmath3he nucleus , ii ) @xmath4 emitters ( 1040 ) which transform a proton into a neutron plus a positron and a neutrino , or by capturing an electron from their atomic shells and emitting a neutrino , iii ) @xmath5 emitters ( 1020 ) which transform a neutron into a proton by emitting an electron and an anti - neutrino , iv ) fissioning nuclei ( 30 ) which split up into two roughly equal pieces , v ) one - proton ( 1p ) emitters ( about 25 ) , vi ) two - proton ( 2p ) emitters ( 2 ) and , finally , ( vii ) exotic - cluster emitters ( about 25 different nuclei emitting fragments of c , o , f , ne , mg , and si ) .
these exotic radioactivities with masses from 14 to 32 may be considered as intermediate between @xmath2 decay and fission .
however , it should be mentioned that this decay mode is never dominant .
although some radioactive nuclei can decay by two or three of these different decay modes , the main decay mode is dictated by the proton - to - neutron ratio .
for example , nuclei with a modest proton excess decay by @xmath4 disintegration .
if the proton excess increases , the nuclear forces can no longer bind all protons and one reaches the proton drip line where the spontaneous emission of protons takes place from the ground state .
heavier neutron - deficient isotopes will preferentially decay by @xmath2 emission , and for the heaviest nuclei fission is the dominant decay mode .
for some of the heaviest nuclei , exotic - cluster emission is a small decay branch .
the study of these different decay types started with the discovery of radioactivity by h. becquerel in 1896 . following the work of pierre and marie curie , e. rutherford in 1899 classified the observed radioactivity into two different decay modes ( @xmath2 and @xmath0 rays ) according to the detected particle in the decay .
villard discovered @xmath6 radiation in 1900 as a third radioactive decay mode .
finally , the nuclear fission process was observed for the first time by o. hahn and f. strassmann and correctly explained by l. meitner in 1938 .
these `` classical '' decay modes have helped largely to understand the structure of the atomic nucleus and the forces acting in the nucleus .
they have also found many applications which range from solid - state physics to astrophysics and medicine . with the discovery of a large number of new radioactive isotopes the understanding of nuclear structure advanced as well .
beginning of the 1960 s , goldanskii @xcite , zeldovich @xcite , and karnaukhov @xcite proposed new types of proton radioactivity in very proton - rich nuclei . for nuclei with an odd number of protons , one - proton radioactivity was predicted .
this decay mode was indeed observed at the beginning of the 1980 s in experiments at gsi , darmstadt @xcite .
today about 25 ground - state one - proton emitters are known @xcite .
proton emission was also observed in many cases from long - lived excited states , the first being @xmath7co @xcite .
the study of one - proton radioactivity allowed testing the nuclear mass surface , to determine the sequence of single - particle levels and the detailed structure of the wave function of the emitted proton ( `` @xmath8 content '' of the wave function ) , and to investigate nuclear deformation beyond the proton drip line .
zeldovich @xcite was probably the first to mention the possibility that nuclei may emit a pair of protons .
however , it was goldanskii @xcite and jnecke @xcite who first tried to determine candidates for 2p radioactivity .
many others followed and the latest predictions @xcite determined that @xmath9ti , @xmath10cr , @xmath11fe , and @xmath12ni should be the best candidates to discover this new decay mode .
goldanskii @xcite also coined the name `` two - proton radioactivity '' and it was galitsky and cheltsov @xcite who proposed a first theoretical attempt to describe the process of 2p emission .
more than 40 years after its theoretical proposal , 2p radioactivity was discovered @xcite in the decay of @xmath11fe . in later experiments ,
@xmath13zn @xcite and most likely @xmath14ni @xcite were also shown to decay by two - proton radioactivity .
the observation of two - proton emission from a long - lived nuclear ground state was preceded by the observation of two - proton emission from very short - lived nuclear ground state , e.g. in the cases of @xmath15be @xcite and @xmath16o @xcite , and by the emission of two protons from excited states populated by @xmath0 decay @xcite or inelastic reactions @xcite .
a particular case is the emission of two protons from an isomer in @xmath17ag@xmath18 @xcite .
the 2p emission process leads from an initial unbound state to a final state of separated fragments interacting by the coulomb force .
this transition to the final state can be realized in many ways , such as the sequential two - body decay via an intermediate resonance , the virtual sequential two - body decay via the correlated continuum of an intermediate nucleus , the direct decay into the three - particle continuum or its particular diproton ( @xmath19he ) limit of two sequential two - body emissions .
different types of theoretical approaches exist which focus on the three - body decay width : the extended r - matrix theory @xcite , the real - energy continuum shell model ( the so - called shell model embedded in the continuum ( smec ) @xcite ) describing sequential two - body emission , virtual sequential two - body emission and diproton emission via two sequential two - body emissions @xcite , the gamow shell model ( gsm ) with no separation between two- and three - body decays from the multi - particle continuum @xcite , the three - body models with outgoing flux describing direct decay into the continuum @xcite , the three - body models combined with complex coordinate scaling @xcite , or the faddeev equations combined with either outgoing flux or complex scaling @xcite .
in the 2p decay , the three - body decay width provides global information with weighted contributions of different decay paths leading from the initial to the final state .
hence , the value of the width by itself does not reveal the dominant transition mechanism . in the theoretical analysis
however , one can switch off certain decay paths and look for the three - body partial decay width resulting from a particular decay path .
this strategy is used in the smec analysis of the experimental results .
obviously , the conclusions of such partial studies of decay mechanisms are not unambiguous . in the present review
, we will summarize the state of the art of experimental and theoretical studies on two - proton radioactivity .
previous reviews , written before the actual discovery of ground - state two - proton radioactivity , can be found in @xcite .
protons are charged particles , therefore they are sensitive to the charge of other protons which create a coulomb barrier .
this barrier prevents protons from quickly leaving the atomic nucleus even if they are unbound .
the tunnelling probability depends on the available energy and the height of the coulomb barrier , which in turn depends on the nuclear charge z ( number of protons ) ( see fig . [
fig : barrier ] ) . the barrier penetration can give rise to measurable half - lives , if a certain balance between the available decay energy and the barrier height is respected . fig . [
fig : tunnel ] shows in a simple model the relation between barrier penetration half - life and decay energy for different nuclear charges z. in general , the higher is the available energy , the shorter is the tunnelling time . in turn , for higher z more energy
is needed for the same tunnelling time .
the delay associated with the tunnelling process allows for the observation of 1p and 2p radioactivity .
even if protons are unbound by e.g. 1 mev , the tunnelling of the combined coulomb and centrifugal barriers is not instantaneous , i.e. the nuclear decay is delayed by a measurable amount of time .
however , due to experimental constraints , mainly linked to the techniques used to study 1p or 2p radioactivity and due to the competition with @xmath4 decay , observation limits exist . the lower half - life limit of about 1 @xmath20s comes from the fact that often the observation of the 1p or 2p emitter is accomplished with the same detection setup which is used to detect the decay of these nuclei .
this means that a decay signal of typically 1 mev has to be observed a very short time after an identification signal ( i.e. an implantation signal in a silicon detector ) of several hundred mev .
this observation limit has been reached in 1p - emitter studies ( see sect .
3 ) . the upper limit ( see fig . [
fig : tunnel ] ) strongly depends on the structure of the decaying nucleus which governs the @xmath0-decay half - life .
it can range from a few milliseconds to a few seconds . to observe charged - particle emission , the barrier - penetration half - life should be comparable or shorter than the @xmath0-decay half - life . in the 1p decay , there are only two particles in the final state and the decay is a simple back - to - back decay , where the energy is shared between the two partners , the heavy recoil and the emitted proton , according to energy and momentum conservation . in the case of 2p decay , the situation is more complicated ( see fig . [
fig : barrier ] ) .
the decay characteristics depend sensitively on the decay pattern itself .
two schematical pictures are usually used to represent possible limiting cases : i ) the three - body decay and ii ) the diproton decay . in the former case
, the two protons do not have any correlation beyond the phase - space constraints , which means that only energy and momentum conservation have to be respected .
such a decay pattern yields an isotropic angular distribution of protons which share the total decay energy , with individual proton energies ranging from zero to the total decay energy .
a more realistic discussion of the tunnelling process changes this picture because the barrier penetration strongly favours an emission of two protons with similar energies . in the latter case of diproton emission
, one assumes that a pre - formed @xmath19he-cluster penetrates the coulomb barrier and decays outside the barrier .
the decay half - life depends sensibly on the @xmath19he resonance energy .
the diproton decay corresponds to two sequential binary decays for which the kinematics is rather easy .
these two limits of 2p decay are not very realistic , but they are easy to grasp and give at least a schematic idea about the 2p emission . a theoretical description of 2p radioactivity ( see sect .
8) is more sophisticated , although none of the models provide an unconstrained description of the nuclear structure and the decay dynamics involved in the 2p decay . up to now
, only total decay energies and half - lives have been measured in 2p radioactivity experiments .
a deeper insight into the decay mechanism can be provided by the measurements of individual proton energies and the proton - proton angle in the center - of - mass . to access these observables , specific detectors have to be developed ( see sect .
soon after the prediction of proton emission from atomic nuclei , the search for these new types of radioactivity started .
the first exotic decay mode discovered was the @xmath0-delayed emission of one proton in the decay of @xmath21si @xcite .
a similar type of proton emission from an excited state was observed in the search for @xmath0-delayed proton emission from @xmath22ni . instead of populating @xmath22ni ,
the authors @xcite observed the population and decay of @xmath22co@xmath18 .
this isomer ( t@xmath23 = 247 ms ) decays with a branching ratio of 1.5% by emission of a 1.59(3 ) mev proton .
the first case of ground - state one - proton radioactivity was reported by hofmann et al .
@xcite and by klepper et al .
both experiments were conducted at gsi , the first one at the velocity filter ship using a fusion - evaporation reaction with a @xmath24ni beam and a @xmath25ru target to produce the proton emitter @xmath26lu , the second one at the on - line mass separator with the reaction @xmath24ni + @xmath27mo to synthesize @xmath28tm .
one - proton radioactivity studies have developed soon into a powerful tool to investigate nuclear structure close to and beyond the proton drip line @xcite with now more than thirty different nuclei known to emit protons from their ground states or low - lying isomeric states .
these studies have given valuable information on the sequence of single - particle levels and their energies in the vicinity of the proton drip - line , on the @xmath8 content of the nuclear wave function , on the nuclear deformation , and they allow to test predictions of nuclear mass models in this region . most of this information can be obtained only by means of 1p radioactivity studies .
more recently proton radioactivity was also used to identify nuclei and to tag events . in experiments
e.g. at the fragment mass separator of the argonne national laboratory , the observation of a decay proton is used as a trigger to observe @xmath6 rays emitted in the formation of the proton emitter .
therefore , not only the decay of these exotic nuclei can be studied , but also their nuclear structure in terms of high - spin states and their decay via yrast cascades @xcite .
the relatively large number of known ground - state proton emitters ( for almost all odd - z elements between z=50 and z=83 , at least one proton - radioactive nucleus is experimentally observed ) is due to the fact that the 1p drip line , i.e. the limit where odd - z nuclei can no longer bind all protons , is much closer to the valley of stability than the two - proton drip line . as an example
, we mention that at z=50 the t@xmath29 isotope @xmath30sn is most likely the first 2p unbound nucleus . for z=51 , @xmath31sb with t@xmath32=1/2
is probably proton unbound .
this means that in this case the 2p emitter is three mass units more exotic than the 1p emitter , which leads to orders of magnitude lower production cross sections .
the first experimental investigation of a possible ground - state 2p emitter was performed by karnaukhov and lu hsi - tink @xcite .
these authors tried to produce @xmath33ne by bombarding a nickel target with a 150 mev @xmath34ne beam from the jinr , dubna 300 cm cyclotron .
the non - observation of any 2p event led the authors to conclude that either the production cross section is much smaller than assumed ( smaller than 10@xmath35 b ) or that the half - life of @xmath33ne is shorter than 10@xmath36s . as early as in 1963 , jnecke predicted its half - life to be of the order of 10@xmath37s @xcite and recently its half - life was given as 9@xmath3810@xmath39s @xcite .
if we omit early measurements which used @xmath41-induced reactions in bubble or spark chambers @xcite , and mass excess measurements performed by ajzenberg - selove et al .
@xcite , the first successful study of a radioactive decay with an emission of two protons was performed in experiments with @xmath15be @xcite . in these studies , @xmath15be was produced in a @xmath42li(@xmath43he , t)@xmath42be reaction .
the triton as well as the @xmath2 particle and the protons from the decay of @xmath15be were observed .
the ground state of @xmath42be is known to have a life - time of the order of 10@xmath39s .
therefore , this nucleus can only be identified via its decay products .
the authors observed an enhancement of low - energy @xmath2 particles compared to a simple phase - space distribution which could not be explained by means of final - state interactions .
however , no conclusions on the decay mechanism could be drawn from these observations .
a much more detailed study of the @xmath42be decay was performed over several years at the kurchatov institute , moscow ( see e.g. @xcite ) .
these authors measured energy and angular distributions of the emitted particles and interpreted their data using jacobi coordinates . the most important conclusion of this analysis
is that the spectra can be interpreted as a `` democratic '' three - body decay , i.e. the decay where none of the two - body subsystems of this three - body final state has a width which is small in comparison with the transition energy between the initial nucleus and the subsystem .
the next step in the investigation of short - lived 2p ground - state emitters was the study of @xmath16o at the national superconducting cyclotron laboratory ( nscl ) of michigan state university , east lansing , usa @xcite . in this experiment ,
@xmath16o was produced by one - neutron stripping from a @xmath45o secondary beam .
the complete reaction kinematics was reconstructed by measuring the momentum vectors of two protons in coincidence with the momentum of the heavy @xmath46c residue . from these measurements , the authors could produce an excitation energy spectrum of @xmath16o , determine the proton - proton energy difference and the proton - proton angle in the @xmath16o center of mass .
these experimental data were interpreted by means of monte - carlo simulations , which employed the r - matrix theory @xcite to describe the 2p decay in two extreme pictures : diproton emission with a @xmath47he resonance energy of 50 kev and three - body decay which is only restricted by the available phase space .
[ fig : o12 ] summarizes these results .
a comparison of the experimental results and the simulations clearly favours an uncorrelated three - body emission or a sequential decay , which may have roughly the same characteristics . at the time of this work , these conclusions were somewhat surprising , as it was believed that the intermediate 1p daughter state in @xmath48n was inaccessible due to energy conservation .
in addition , a very large @xmath16o decay width was needed to explain the experimental data .
however , later measurements @xcite revealed that the @xmath48n ground state lies lower in mass and opens therefore the way for a sequential decay . with these new results , a consistent description of the decay of @xmath16o could be provided @xcite .
the decay of @xmath51 mg was recently studied in an experiment at the fragment separator frs of gsi @xcite . in this experiment , @xmath51 mg
was produced by a one - neutron stripping reaction at the frs intermediate focal plane of @xmath34 mg , produced itself at the frs target in a fragmentation reaction of the @xmath52 mg primary beam .
the protons emitted in the decay of @xmath51 mg were detected and tracked by a set of three dsssd , whereas the heavy decay product , @xmath53ne , was identified and its momentum measured in the second half of the frs .
the measured data were compared to monte carlo simulations and allowed to extract the half - life of @xmath51 mg of @xmath54 = 4.0(15 ) ps and the associated two - proton q - value : @xmath55 = 0.75(5 ) mev .
this value is much smaller than standard predictions like the one of the atomic mass evaluation @xcite ( @xmath55 = 2.00(25 ) mev ) , the garvey - kelson relation @xcite ( @xmath55 = 1.02(2 ) mev ) , or the prescription from antony et al .
@xcite ( @xmath55 = 1.19(3 ) mev ) . however , the correlation between the q value and the decay half - life is well reproduced within the three - body model of grigorenko and co - workers @xcite . in the same experiment
@xcite , the decay of @xmath33ne by two - proton emission was also observed and the known decay q value @xcite of 1.41(2 ) mev was used to test the experimental procedure .
for the systems @xmath15be , @xmath16o , and @xmath33ne , the decaying states have large widths so that the 2p emitter state and the 1p daughter state overlap .
therefore , the sequential decay channel is opened and the decay will most likely take place sequentially .
the situation in the case of @xmath51 mg is less clear , as the structure and the mass of the one - proton daughter @xmath56na is not known .
however , with increasing nuclear charge , the coulomb barrier becomes stronger and the wave function is more and more confined in the nuclear interior which yields narrower states .
therefore , it might well be that in nuclei like @xmath57si , @xmath58s , or @xmath59ar the coulomb barrier is already strong enough so that , with a possibly small decay energy , narrow levels avoid overlapping and the sequential decay path is , at least in a simple picture , not open .
generally speaking , the 2p emission can take place as a sequential - in - time process or a simultaneous process . in its standard form
, the sequential process happens when levels in the 1p daughter nucleus , both resonant and non - resonant ones , are accessible for 1p emission .
this situation may be encountered when the emission takes place from highly excited levels of the 2p emitter .
these levels can be fed either by @xmath0 decay or by a nuclear reactions .
a clear signal for a sequential emission can be obtained when the intermediate level through which the emission passes is identified .
this is the case , for example , when well defined one - proton groups sum up to a fixed 2p energy .
the one - proton groups give then the energy difference between two well - defined nuclear levels .
the possibility of @xmath0-delayed 2p ( @xmath02p ) emission was first discussed by jnecke @xcite .
he concluded that there should be no `` @xmath0-delayed diproton emitters '' with z@xmath6014 .
this decay mode was also studied by goldanskii @xcite who proposed possible candidates . the first experimental observation of @xmath0-delayed 2p emission was reported by cable et al . in 1983 @xcite .
their experiments , using a @xmath43he induced reaction on a magnesium target and a helium - jet technique , led to the first observation of @xmath61al by means of its @xmath0-delayed 1p emission from the isobaric analogue state ( ias ) in @xmath61 mg to the ground and first excited states of @xmath57na @xcite .
the @xmath02p decay mode was subsequently observed with two silicon detector telescopes which allowed to measure the energy of the two individual protons with good resolution @xcite . from the observation of
well defined one - proton groups , cable et al . concluded that this decay is a sequential mode via intermediate states in the 1p daughter nucleus @xmath57na .
this conclusion was confirmed in a further experimental study @xcite where it turned out that the angular distribution of the two protons is compatible with an isotropic emission , although a small angular correlation ( less than 15% ) could not be excluded . meanwhile more @xmath02p emitters have been identified . shortly after the discovery of @xmath02p decay from @xmath61al @xcite , the same authors observed @xmath02p emission from @xmath58p @xcite and from @xmath62ca @xcite .
all in all , nine @xmath02p emitters have been observed : @xmath61al @xcite , @xmath63si @xcite , @xmath58p @xcite , @xmath64s @xcite , @xmath65ar @xcite , @xmath62ca @xcite , @xmath9ti @xcite , @xmath66cr @xcite , and @xmath67ni @xcite .
most of these studies were characterized by rather low statistics inherent to most investigations of very exotic nuclei and hence no detailed search for angular correlations could be performed .
in addition , most of the observed decays proceed via the ias in the @xmath0-decay daughter nucleus .
the 2p emission from this state is forbidden by the isospin conservation law and can only take place via a small isospin impurity of the ias .
the first high - statistics study of a @xmath02p emitter was carried out at isolde in cern @xcite .
for the first time , a rather high - efficiency , high - granularity setup which allows a high - resolution measurement in energy and angle was used .
this experiment , which studied the decay of @xmath65ar , demonstrated the occurrence of @xmath02p decay branches via states other than the ias .
indeed , many gamow - teller fed states could be shown to decay by 2p emission .
the main interest of these decays is that they are isospin allowed and cover a wider range of spins for the 2p emitter states , the possible intermediate states and the final states .
detailed studies of the proton - proton angular correlation for different 2p emission branches did not yield evidence for any angular correlation .
in fact , the results were compatible with sequential emission via intermediate states in the 1p daughter nucleus .
although there was some activity detected with the two protons sharing the available decay energy equally , the statistics of these events was too low and the angular coverage of the setup was not uniform enough to draw any conclusion on the possible observation of events , where the protons were correlated in energy and angle .
therefore , despite a factor of 10 - 100 higher statistics as compared to earlier experiments , no angular and energy correlation , a possible signature of a strong proton - proton correlation , could be evidenced .
this observation is not really astonishing . according to brown @xcite , who studied the competition between 1p emission and correlated 2p ( diproton ) emission from the ias in @xmath61 mg
, the weight of the direct 2p branch is expected to be at best of the order of a few percent and therefore rather difficult to observe experimentally .
new attempts to search for such a correlated 2p emission in @xmath65ar are on the way @xcite .
these studies take profit from a much more uniform angular coverage .
besides populating excited 2p emitter states by @xmath0 decay , these nuclear states can also be fed by nuclear reactions like pick - up , transfer , or fragmentation . in this section
, we will describe experiments which used this kind of interaction to study 2p emission from excited states . in an experiment performed at louvain - la - neuve , bain et al .
@xcite bombarded a ( ch@xmath69)@xmath70 target with a @xmath45n radioactive beam populating highly - excited states in @xmath68o .
the 7.77 mev state in @xmath68o , which is strongly populated in 2p transfer reactions @xcite , is expected to have a significant admixture of a 2p configuration outside of the @xmath16c core .
in addition , this state is 2p unbound by about 1.2 mev and , unlike states populated by a super - allowed @xmath0 decay , the 2p decay to the ground state of @xmath16c is isospin allowed .
therefore , this resonance was believed to have a small 2p decay branch , either directly to the @xmath16c ground state or sequentially via a narrow state in @xmath45n ( see fig . [
fig : bain ] ) .
the two protons from the decay of the 7.77 mev resonance were detected with the leda device @xcite .
the data showed clearly a resonant structure when the beam energy was on - resonance. the authors determined a 0.16(3)% 2p branch for the decay of this resonance .
the individual energies of the two protons could be explained by a sequential decay pattern via the 2.37 mev state in @xmath45n .
monte - carlo simulations employing r - matrix theory showed that the data could be best explained with a 100% sequential decay pattern .
the decay of excited states of @xmath53ne by 2p emission was studied several times in recent years @xcite .
the first excited state of @xmath53ne is 1p bound but unbound with respect to 2p decay by 344 kev and may therefore decay by direct 2p emission to the ground state of @xmath71o ( see fig .
[ fig : ne17_scheme ] ) .
chromik et al .
@xcite populated the first excited states of @xmath53ne by coulomb excitation and observed their decay by @xmath6 de - excitation @xcite or in a complete - kinematics experiment @xcite .
in the first experiment , the missing strength of the decay of the first excited state by @xmath6 emission , as compared to theoretical calculations , was interpreted as a possible unobserved 2p branch . however
, it was found that such an interpretation was in contradiction to barrier - penetration calculations .
in fact , a 2p partial half - life shorter by a factor 1700 than predicted was needed to account for the missing decay strength .
the second complete - kinematics experiment allowed determining all decay channels and a partial 2p half - life for the first excited state larger than 26 ps was deduced .
no 2p emission from the first excited state was observed .
the decay of the second excited state , also populated by coulomb excitation , was observed to decay by sequential 2p emission to the ground state of @xmath71o yielding an isotropic emission pattern for the two protons .
this experimental result was confirmed by zerguerras et al .
@xcite . in their experiment ,
excited states of @xmath53ne were populated by one - neutron stripping reactions from a @xmath56ne beam at 36 mev / nucleon .
states up to 10 mev excitation energy were observed to decay by 2p emission .
the decay products were detected in the must detector array @xcite and in the speg spectrometer @xcite .
these measurements allowed a reconstruction of the complete decay kinematics and therefore to determine the excitation energy of @xmath53ne and the relative proton - proton angle in the center - of - mass frame .
[ fig : zerg1 ] shows the invariant mass spectrum for @xmath71o+2p events .
the angular distribution of the two protons was measured to be isotropic for the second and third excited states in @xmath53ne ( fig .
[ fig : zerg2]a ) .
however , for higher lying states , a distinct angular correlation of the two protons could be observed for the first time ( fig .
[ fig : zerg2]b ) .
this result is rather surprising , as many intermediate states in the 1p daughter nucleus @xmath33f should be accessible and therefore a sequential decay mechanism via these states should dominate .
the interpretation of this decay pattern is still controversial .
kanungo et al .
@xcite and grigorenko et al .
@xcite suggested that some of the excited states of @xmath53ne may have a pronounced 2p halo structure and a large overlap with the @xmath71o ground state leading to much larger spectroscopic factors for a direct 2p decay than for a sequential decay .
obviously , the validity of this assertion depends on whether the configuration mixing in @xmath53ne leaves the @xmath71o core intact . on the other hand ,
as shown in smec studies @xcite the correlated 2p emission is _ always _ hindered if 1p decay channels are open .
moreover , the measured energies of the two emitted protons from @xmath53ne show large differences which can not be reconciled with a diproton scenario where roughly equal energies are expected for the two protons .
another explanation of the 2p emission pattern in @xmath53ne could be the large deformation in higher lying states .
strong anisotropy of the coulomb barrier could be a source of dynamical correlations between emitted protons even if this 2p decay is a sequence of two successive 1p emissions .
effects of this kind have been discussed in the de - excitation of heavy - ion residues ( see e.g. ref .
a similar explanation has been put forward also in connection with the anomalous decay properties of @xmath17ag@xmath18 @xcite .
obviously , for a consistent interpretation of the @xmath71o + 2p experiment , higher statistics data is mandatory . in an experiment performed at ornl oak ridge , gomez del campo et al .
@xcite used a radioactive @xmath53f beam on a ( ch@xmath69)@xmath70 target to produce excited states in @xmath56ne .
the experiment used a thick - target approach , i.e. only the protons could be detected behind the target , whereas the beam - like particles ( e.g. the heavy @xmath33o recoils from the decay of excited states in @xmath56ne ) could not penetrate the target .
therefore , only protons with a relatively large energy loss depending on their emission point in the target could be observed . using two different incident energies of the @xmath53f beam , these authors observed at higher incident energy the emission of two protons which they interpreted as a decay of the 6.15 mev resonance state in @xmath56ne .
for the lower energy , where this state could not be populated , no coincident protons were observed .
both the relative proton - proton angle and the energy difference between the two protons was in agreement with simulations for a three - body decay of this state as well as for a diproton - type decay .
this lack of sensitivity is to a large extent due to the limited statistics of the experiment , but also due to the limited angular coverage of the experimental setup .
the decay of @xmath17ag@xmath18 is unusual in several respects .
it possesses two @xmath0-decaying isomers and has the highest spin state ever observed which decays by @xmath0 decay .
in particular , the second isomeric level seems to decay by several different decay branches : @xmath0-delayed @xmath6 decay @xcite , @xmath0-delayed proton emission @xcite , direct proton decay @xcite , and direct 2p emission @xcite .
all these experiments were performed at the on - line separator of gsi @xcite .
the two protons were observed by means of silicon strip detectors which surrounded the implantation point of the on - line separator .
this setup allowed for the measurement of the individual proton energies and the relative proton - proton angle .
although the statistics of the 2p experiment is rather low @xcite , the authors identified proton - proton correlations .
they interpreted these data as due to simultaneous 2p emission from a strongly deformed ellipsoidal nucleus . in this case , the asymmetric coulomb barrier favours the emission of two protons within narrow cones around the poles of the ellipsoid , either from the same pole or on opposite sides .
one should mention , that the observed proton - proton correlations in such a scenario are largely unrelated to the proton - proton correlations inside a nucleus .
the microscopic theoretical study of the @xmath17ag@xmath18 decay is also impossible using the present - day computers , though only such studies could disentangle an internal structure of decaying states from dynamical effects of the anisotropic coulomb barrier , providing more details about the 2p decay pattern in the nucleus .
unfortunately , further experimental studies of the @xmath17ag@xmath18 decay can not be made without new technical developments to produce @xmath17ag@xmath18 , as the on - line separator of gsi has been dismantled a few years ago .
goldanskii @xcite called it a pure case of 2p radioactivity , if 1p emission is not possible because the 2p emitter level is lower in energy than the level in the 1p daughter nucleus and , moreover , the corresponding levels are sufficiently narrow and thus do not overlap
. the limitations of this picture will be discussed in chapter 8 .
goldanskii also gave an arbitrary half - life limit of 10@xmath72 s @xcite to distinguish between two - proton radioactivity and the decay of particle - unstable nuclei .
none of the examples discussed above satisfy goldanski s radioactivity criterion ; they all have rather short half - lives ( of the order of 10@xmath39 s ) and for the light ground - state emitters , the states which emit the first and the second proton overlap significantly . in this sense ,
2p radioactivity was observed for the first time in @xmath11fe experiments @xcite .
later @xmath13zn @xcite and possibly @xmath14ni @xcite were shown to be other ground - state 2p emitters .
recent theoretical predictions @xcite agree that @xmath9ti , @xmath10cr , @xmath11fe , and @xmath12ni are candidates to search for 2p radioactivity .
firstly , their predicted decay energy is close to or in the range of energies ( 1.0 - 1.5 mev ) where 2p radioactivity is most likely expected to occur , and secondly these isotopes are already or could soon be produced in projectile - fragmentation type experiments .
this production technique allows to unambiguously identify the isotopes before their decay and thus measure the decay half - life and the decay energy , the necessary ingredients to identify this new radioactivity .
the most promising candidates were @xmath11fe and @xmath14ni with decay energies of 1.1 - 1.4 mev .
the first of these isotopes , @xmath9ti , was observed at the grand acclrateur national ions lourds ( ganil ) , caen @xcite in 1990 .
this experiment was also the first attempt to observe 2p radioactivity of these medium - mass nuclei after projectile fragmentation .
however , no indication for 2p radioactivity could be found . instead , a @xmath0-decay half - life of 28@xmath73ms was measured and @xmath0-delayed protons were observed .
another attempt to search for 2p radioactivity from @xmath9ti was made at ganil in 1991 with an improved setup @xcite .
however , again no 2p events could be observed .
the nuclei @xmath10cr , @xmath11fe , and @xmath74ni were identified for the first time a few years later at gsi @xcite . in this experiment ,
no spectroscopic studies could be performed and the observation of these isotopes was the only outcome .
nonetheless , this experiment provided the first observation of one of the most promising 2p radioactivity candidates and opened the door for the search of @xmath14ni , another candidate .
this doubly - magic nucleus was then discovered in 1999 @xcite in a ganil experiment using the sissi / lise3 device @xcite . in the same experiment ,
the observation of @xmath11fe was confirmed with about a factor of 10 higher statistics .
the ganil experiment was also set up to perform spectroscopic studies of the very exotic nuclei produced @xcite .
the decay trigger in this experiment allowed mainly @xmath0-delayed events to be registered .
@xmath0-delayed decay events from e.g. @xmath11fe decay had a detection efficiency of about 30% , as they were triggered by two detectors adjacent to the implantation detector .
2p events without @xmath0 particle had a very low trigger rate ( a few percent ) due to a heavy - ion trigger for the implantation device .
this situation allowed for the observation of the decay of @xmath11fe , however , basically of only its @xmath0-delayed branch ( see sect .
6.2 ) . in any case
, no claim about the observation of 2p radioactivity was made . nevertheless , the experiment allowed one to test two other possible , although less promising , 2p candidates : @xmath10cr and @xmath74ni . as mentioned above , these two nuclei
were predicted to be possible 2p emitters , however , with decay energies most likely too small to exhibit this decay . the decay energy spectrum obtained during this experiment for @xmath10cr is shown in fig .
[ fig : cr42 ] .
the prominent peak at 1.9 mev was excluded to be of 2p nature , as such a high energy would lead to an extremely short barrier penetration time for the two protons ( see fig .
[ fig : cr42 ] , right - hand side ) , in contradiction with the @xmath10cr half - life of 13.4@xmath75ms measured in the same experiment .
cr implantation .
the 1.9 mev peak is interpreted as a @xmath0-delayed decay @xcite .
right - hand side : barrier penetration half - life for a di - proton in @xmath10cr as a function of the available decay energy @xcite .
the red line represents the 1.9 mev peak from the decay - energy spectrum yielding a barrier penetration half - life of 10@xmath72s if it were of 2p origin . ] in a similar way , @xmath74ni could be excluded from having a significant direct 2p branch .
the last 2p emitter candidate tested in this experiment was @xmath9ti .
in agreement with previous experiments , this nucleus turned out to decay predominantly by @xmath0 decay .
several @xmath0-delayed proton groups could be identified and the 2p emission q value ( q@xmath76 = 670@xmath77100 kev ) could be determined , indicating a rather high barrier - penetration half - life and therefore a rather low probability for 2p radioactivity .
the discovery of ground - state 2p radioactivity had to await the year 2002 , when two experiments , one performed in 2000 at the sissi / lise3 facility of ganil @xcite and one performed at the frs of gsi in 2001 @xcite , were published .
both experiments used the fragmentation of a @xmath24ni primary beam .
the fragments of interest were implanted in a silicon detector telescope ( see fig . [
fig : fe45_setup ] ) , where correlations in time and space between implantation events and decay events could be performed .
this technique allowed to unambiguously correlate @xmath11fe implantation events with their decays .
particles @xcite .
right - hand side : setup used in the gsi experiment .
a silicon telescope of seven elements was surrounded by a high - efficiency nai array .
implantation and decay events were correlated in different detectors depending of the implantation depth of the @xmath11fe ions .
the nai device served to search for @xmath6 radiation from positron annihilation @xcite . ] fig .
[ fig : fe45_e ] shows the decay energy spectra from the two experiments .
both find a decay energy of about 1.1 mev with a decay half - life in the 3 - 5 ms range .
an additional feature of both experiments was the capability to detect additional radiation like @xmath0 particle in the ganil experiment , or 511 kev annihilation radiation and other @xmath6 rays in the gsi experiment ( see fig .
[ fig : fe45_setup ] ) .
this capability allowed for a rejection with a high probability of a possible @xmath0-delayed origin of the decay events with an energy release of about 1.1 mev .
other pieces of evidence from the ganil experiment were the smaller width of the 1.14 mev peak as compared to @xmath0-delayed proton emission peaks in neighbouring nuclei and the observation of daughter decays consistent with the decay of @xmath66cr , the 2p daughter of @xmath11fe .
in addition , the experimentally observed decay energy is in nice agreement with theoretical predictions @xcite .
these experimental results led the authors to conclude the discovery of ground - state 2p radioactivity in the decay of @xmath11fe . of particular interest
was also the observation of decay events from @xmath11fe with an energy release different from the 1.1 mev of the main peak .
some of these additional events could be identified as being in coincidence with @xmath0 particles and were therefore attributed to @xmath0-delayed decay branches . from these first results , a 2p radioactivity branching ratio of 70 - 80% was determined for @xmath11fe .
the partial half - lives of @xmath11fe for 2p emission and for @xmath0 decay are therefore of the same order and allowed the two decay modes to compete .
these results on @xmath11fe have recently been confirmed in a new ganil experiment @xcite .
several experimental improvements allowed a better energy calibration and a smaller data acquisition dead time .
the decay energy was determined to be 1.154(16 ) mev , the half - life to be 1.6@xmath78ms , and the 2p branching ratio to be 57(10)% .
together with the results from the two earlier experiments @xcite , this experiment allowed to determine average values for the decay energy ( q@xmath76= 1.151(15 ) mev ) , the half - life ( t@xmath23= 1.75@xmath79ms ) , the 2p branching ratio ( br = 59(7)% ) , and the partial 2p half - life ( t@xmath80= 3.0@xmath81ms ) . in 2007 , a new experiment performed at the a1900 separator of michigan state university allowed the observation of 2p events in a tpc ( see below ) .
the newly determined half - life and the 2p branching ratio are included in table [ tab:2p - emitters ] to determine average values .
these average values will be used in the comparison with theoretical model of 2p radioactivity .
fe decay - energy spectrum from the ganil experiment exhibiting a peak at ( [email protected] ) mev @xcite .
right - hand side : @xmath11fe decay - energy spectrum from the gsi experiment showing four events at 1.1(1 ) mev @xcite .
both spectra also show events at higher energies . ]
another outcome from the ganil experiments @xcite on @xmath11fe was the determination of the half - life of the 2p daughter nucleus . from the observation of the second decay after @xmath11fe implantation , the half - life of the daughter decay , in coincidence with the 1.151 mev peak ,
could be determined . in fig .
[ fig : daughter_half - life ] , this half - life is compared to the half - lives of all possible @xmath11fe daughter nuclei .
only the half - life of @xmath66cr is in agreement with the experimentally observed daughter half - life , thus giving an independent proof for the observation of 2p radioactivity .
the nucleus @xmath13zn was early identified as a possible 2p emitter @xcite . more recent @xmath55 estimates yielded decay energies of 1.79(12 ) mev @xcite , 1.87(24 ) mev @xcite , and 1.33(14 ) mev @xcite .
due to rather large theoretical error bars and the uncertainty concerning the decay mechanism , it was rather unclear whether @xmath13zn would exist at all , i.e. whether its half - life would be long enough to be observed in projectile - fragmentation type experiments .
for this purpose , a life time of the order of a few hundred nano - seconds was mandatory .
the search for @xmath13zn began with the observation of two other proton - rich zinc isotopes , @xmath82zn and @xmath83zn @xcite .
they were observed for the first time in experiments at the lise3 separator @xcite via 2p pick - up reactions with a @xmath24ni primary beam .
the identification of these two isotopes allowed also for an extrapolation of the production cross section of @xmath13zn .
the nucleus @xmath13zn was synthesized and observed for the first time in 2004 in an experiment at the sissi / lise3 facility of ganil caen @xcite .
similarly to the experiment with @xmath11fe , @xmath13zn was produced with a primary @xmath24ni beam at 75 mev / nucleon impinging on a natural nickel target .
the fragments of interest were selected and separated with the lise3 separator and implanted in a silicon telescope with a double - sided silicon strip detector being the central device .
the @xmath13zn nuclei were identified by means of time - of - flight , energy - loss and residual - energy measurements .
[ fig : zn54_id ] shows the two - dimensional identification plot for the @xmath13zn experiment .
all in all , eight @xmath13zn nuclei could be unambiguously identified .
the setup allowed correlating in time and space these implantation events with subsequent decays .
seven of the eight implantation events are followed by a decay with an energy release of 1.48(2 ) mev ( see fig . [
fig : zn54_e ] ) .
none of these decay events is in coincidence with a @xmath0 particle . as for @xmath11fe ,
the 1.48 mev peak is much narrower than @xmath0-delayed proton lines in neighbouring nuclei .
although less significant as in the case of @xmath11fe , daughter decays in agreement with the decay of @xmath84ni , the 2p daughter of @xmath13zn , could be observed . finally , the experimental half - life ( t@xmath23= 3.2@xmath85ms ) is in agreement with what is expected for a 2p decay with an energy release of 1.48 mev . from these experimental observations ,
the authors concluded the observation of ground - state 2p radioactivity from @xmath13zn .
the 2p branching ratio was determined to be 87@xmath86% , which yielded a partial half - life for 2p decay of 3.7@xmath87 ms .
these results will be compared to theoretical predictions in sect .
the discovery of @xmath14ni and the observation of its decay was considered important for several reasons .
firstly , it is a doubly - magic nucleus which was predicted to be particle stable or quasi - stable only due to strong shell effects .
secondly , @xmath14ni is one of the most exotic of all doubly - magic nuclei within experimental reach .
its observation and the measurement of its properties , like e.g. the excitation energy of the first 2@xmath88 state , were expected to teach us a lot about the persistence of shell structure far from the valley of stability .
the nucleus @xmath14ni was rather early identified as a possible 2p emitter @xcite .
this prediction was confirmed and refined by new mass and @xmath55 calculations , which yielded 1.36(13 ) mev @xcite , 1.14(21 ) mev @xcite , 1.35(6 ) mev @xcite , and 1.29(33 ) mev @xcite . when compared to @xmath11fe and @xmath13zn ,
it is evident that this @xmath55 value is again in the right range to permit @xmath14ni to have a significant 2p radioactivity decay branch .
the isotope @xmath14ni was first observed in 1999 at ganil caen @xcite , where four events could be unambiguously identified .
this result was confirmed in 2004 , when again four events of @xmath14ni could be registered in a new sissi / lise3 experiment at ganil @xcite .
these data were obtained in the same run as the second @xmath11fe data set from ganil , because the acceptance of the sissi / lise3 device allows the measurement of different neighbouring nuclei at the same time .
[ fig : ni48_id ] shows the fragment identification plot .
the four @xmath14ni implantation events could be correlated in time and space with subsequent decay events . from the four decay events ,
only one event has all characteristics of a 2p emission : a decay energy of 1.35(2 ) mev in the expected energy range and no coincident @xmath0 particle .
the decay takes place after 1.66 ms and the daughter decay is also fast as expected for @xmath89fe , the 2p daughter of @xmath14ni .
these four decay events , one of possible 2p origin and three from @xmath0-delayed decays , were also used to determine the half - life of @xmath14ni ( t@xmath23= 2.1@xmath90ms ) . although this single possible 2p decay event is not considered sufficient to claim the observation of 2p radioactivity for @xmath14ni , these data were nonetheless used to determine the 2p branching ratio to be 25@xmath91% and the partial 2p half - life to be t@xmath80=8.4@xmath92ms @xcite . from all 2p emitters observed up to now , the ground - state emitter @xmath11fe is the best candidate to search for a clear signal of correlated emission and to study in detail the decay mechanism .
as just mentioned , 2p emission could only be governed by phase space or there could be more or less strong angular and energy correlations . in order to study these questions , new experimental setups had to be imagined , the basic limitation of the silicon - detector setups being that the 2p emitters are deeply implanted in a silicon detector , which then does not allow the detection of individual protons , but rather of the total decay energy and the half - life only .
the solution to this problem is the use of gas detectors capable to visualize the traces of the emitted protons .
therefore , time projection chambers ( tpc ) were set up in bordeaux @xcite and warsaw @xcite .
the bordeaux detector is based on gas electron multiplier technology @xcite and a double - sided microgroove detector @xcite with an application - specific integrated - circuit read - out , whereas the warsaw detector uses an optical read - out system by means of a digital camera and a photomultiplier tube ( pmt ) which registers the time sequence .
both tpc detectors have recently taken first data @xcite ( see figs .
[ fig:2pevent ] , [ fig:2pmsu ] ) . due to the highest production rates ,
@xmath11fe was the first 2p emitter to be studied with these devices . due to the higher statistics , the data taken at michigan state university with the warsaw tpc allowed to compare the experimental data to predicitons of the @xmath93 model of grigorenko et al .
model @xcite , for which nice agreement for the angular and energy distribution was reached @xcite indicating e.g. an equal sharing of the energy between the two protons .
the observation of ground - state 2p radioactivity for @xmath11fe , @xmath13zn and possibly for @xmath14ni allowed this type of radioactivity to be established as a new nuclear decay mode .
although only based on theoretical mass predictions , these decays are believed to be simultaneous emissions of the two protons .
a sequential decay could have a sizable contribution only if the @xmath94 values were off by a few hundred kev what is rather unlikely .
nonetheless , the assumption that the decay is indeed simultaneous still awaits experimental verification though , as will be discussed in sect .
9 , a symmetric energy distribution of individual protons emitted in the 2p decay of @xmath11fe @xcite is a strong indication for a predominance of the simultaneous 2p emission .
table [ tab:2p - emitters ] summarizes the experimental information available for @xmath11fe , @xmath14ni , and @xmath13zn . ' '' '' & 2p decay energy & half - life & branching ratio & partial half - life + & ( mev ) & ( ms ) & & ( ms ) + [ 0.5em ] ' '' ''
@xmath11fe & [email protected] & 2.5@xmath95 & [email protected] & 3.9@xmath96 + [ 0.5em ] ' '' '' @xmath14ni & [email protected] & 2.1@xmath90 & 0.25@xmath97 & 8.4@xmath98 + [ 0.5em ] ' '' '' @xmath13zn & [email protected] & 3.2@xmath85 & 0.87@xmath99 & 3.7@xmath87 + [ 0.5em ] to get a coherent picture of this new decay mode , more 2p emitters have to be studied experimentally .
however , as these nuclei are , by definition , situated beyond the proton drip line , their production is rather difficult and the rates are very low . rather high rates can be achieved today at the sissi / lise3 facility of ganil with about 10 - 15 @xmath11fe per day , about 2 @xmath13zn per day , and roughly 1 @xmath14ni per day under optimal conditions .
it was recently shown that at the coupled - cyclotron facility and the a1900 separator of the nscl even higher rates can be achieved .
other candidates like @xmath100ge , @xmath101se or @xmath102kr might be reached in the near future at the new riken radioactive ion beam facility ( ribf ) in tokyo , japan .
the nuclear many - body hamiltonian is supposed to describe all nuclei that can exist and not merely one nucleus of a given number of protons and neutrons . in this sense ,
a nucleus is never a closed , isolated quantum system but communicates with other nuclei through virtual excitations , decays and captures .
the communication is broken and the nucleus becomes artificially closed if the subspace of continuum states is excluded in the network of coupled systems . obviously , the closed quantum system idealization of a real many - body system has very different features from those observed @xcite , in particular in the neighborhood of each reaction threshold @xcite .
most of the nuclear states are embedded in the continuum of decay channels due to which they get a finite lifetime .
the initial impulse for a mathematical formulation of the shell model ( sm ) for open quantum systems goes back to feshbach and fano @xcite who introduced the two subspaces of hilbert space : ( i ) the subspace of discrete states , and ( ii ) the subspace of real - energy scattering states , respectively .
feshbach achieved a unified description of direct processes at the short - time scale and of compound nucleus processes at the long - time scale @xcite .
a unified description of both nuclear structure and nuclear reactions in a single theoretical framework became possible only in the continuum sm @xcite .
first realistic studies using the continuum sm have been presented only at the turn of the century in the framework of smec @xcite ( see also ref .
@xcite for a recent review ) .
this model is based on the completeness of the single - particle basis consisting of bound orbits and a real - energy scattering continuum .
single - particle resonances are regularized , i.e. they are included in a discrete part of the spectrum after removing scattering tails which are incorporated in the embedding continuum .
subsequently , the new set of discrete states and the non - resonant scattering states are reorthogonalized .
configuration mixing in the valence space ( the internal mixing ) is given by an effective sm interaction .
coupling of discrete states to the continuum of decay channels is calculated microscopically and leads to the so - called external mixing of sm wave functions .
the decay channels are obtained in the @xmath103-matrix ( scattering - matrix ) formalism ( see ref .
@xcite and references quoted therein ) . another approach referring to the continuum sm has been proposed recently in refs .
internal mixing in this approach is given by an effective sm hamiltonian but an external mixing is neglected .
this scheme has been applied to the description of observables in the one- and two - particle continuum @xcite .
the genuine three - body decay models for 2p radioactivity have been proposed in refs .
@xcite using the hyperspherical harmonics method .
these models combine the cluster model description of the decaying system with a treatment of three - body asymptotic states of charged particles .
rigorous treatment of a three - body kinematics in these models allows to study both angular and energy / momentum correlations between decaying fragments .
a different approach for describing many - body open quantum systems is provided by a generalization of sm in the complex - momentum plane ( the gsm ) @xcite .
this general formalism with no restriction on the number of particles in the scattering continuum has not yet been applied for a description of experimentally observed 2p decays . a general theory of the decaying many - body system is provided by the open quantum system formulation of the sm .
before we turn to the description of 2p decay , let us introduce the basic features of smec in the simpler case of the one - particle continuum .
in smec , the function space consists of two sets : the space of @xmath104-functions , as in the nuclear sm , and the space of scattering states ( decay channels ) embedding sm states .
the decay is a result of the coupling between discrete states and decay channels .
these two sets of wave functions are found by solving the schrdinger equation for discrete states : @xmath105 where @xmath106 is the sm hamiltonian , and for scattering states : @xmath107 where @xmath108 in ( [ smeq2 ] ) is the standard coupled - channel hamiltonian .
different channels are defined by coupling the motion of an unbound particle relative to the residual nucleus with @xmath109 particles in a discrete sm state @xmath110 .
@xmath111 are scattering states projected on the channel @xmath112 , and the channel numbers @xmath112 are defined by the quantum numbers of of the states @xmath8 of the residual nucleus and those of the unbound particle which are coupled to the total quantum numbers @xmath113 and @xmath114 of the system . using these two function sets : @xmath115 , @xmath116 , one defines projection operators : @xmath117 satisfying @xmath118 , @xmath119 , and projected hamiltonians : @xmath120 and @xmath121 .
the closed system hamiltonian @xmath122 is the nuclear sm hamiltonian @xmath106 , whereas @xmath123 is given by @xmath124 . to construct a complete solution in @xmath125 : @xmath126 one has to find a wave function connecting @xmath127 and @xmath128 subspaces .
this wave function @xmath129 is obtained as a solution of non - homogeneous coupled - channel equations : @xmath130 * is the source term , @xmath131^{-1}\ ] ] is the green s function in @xmath128 subspace , @xmath132 is the total energy of the nucleus @xmath133 $ ] , and @xmath134 . using the three function sets : @xmath135 , @xmath136 , and @xmath137 ,
one can find the solution in the total function space : @xmath138 @xmath139 is the energy - dependent effective hamiltonian in the function space of discrete states : @xmath140 which takes into account couplings to decay channels . in that sense
, @xmath141 is the open quantum system hamiltonian in @xmath127 subspace .
the second term on the right - hand side of eq . ( [ smeq3 ] ) provides an imaginary part of energy eigenvalues for states above the particle emission threshold .
the real part of this term contributes to the real part of energy eigenvalues both below and above the threshold .
for that reason , it would be more appropriate in phenomenological applications to fit matrix elements of @xmath141 and not of @xmath122 to nuclear spectra in a given mass region .
this ambition task has not yet been accomplished .
energies @xmath142 and widths @xmath143 of the resonance states derive from the solutions of fixed - point equations : @xmath144 where functions @xmath145 and @xmath146 follow from ( complex ) eigenvalues of @xmath141 .
the eigenvalues of @xmath141 have a physical meaning if the total function space is divided into the subspaces of the _ system _ ( the subspace @xmath127 ) and the _ environment _ ( the subspace @xmath128 ) as follows : the system contains all resonance - like phenomena while the environment describes the smooth part in the energy region considered .
this operational definition of subspaces implies that in smec one has to construct carefully the single - particle basis .
the narrow single - particle resonances in the continuum are included into the set of discrete states either as resonance anamneses @xcite or as quasi - bound single - particle states @xcite .
the new definition of discrete single - particle states implies a redefinition of the continuous spectrum , as described in ref .
@xcite .
the description of borromean ( halo ) systems and two - nucleon decays involves couplings to the two - nucleon decay channels . in the 2p decay , since two protons
do not form a bound two - particle system , one neither can uniquely specify the final state of two protons nor the transition path leading to this state .
various transition paths may interfere making distinction between different decay mechanisms difficult if not impossible . in this respect ,
the 2p decay is qualitatively different from any standard two - body decays such as the proton- , neutron- or @xmath2-particle decays .
a generalization of the smec formalism for larger number of particles in the scattering continuum is conceptually similar to the one - particle case discussed above ( cf sect .
[ smec_1p ] ) .
for the description of two - particle continuum , one has to introduce a subspace @xmath147 with two particles in the continuum , and @xmath148 the corresponding projection operator .
the completeness relation in total function space is then : @xmath149 .
similarly as in sect .
[ smec_1p ] , one can decompose the hamiltonian @xmath150 into components acting in different subspaces and the coupling terms between those different subspaces .
the @xmath127 subspace effective hamiltonian @xmath139 can be written as @xcite : @xmath151\tilde{g}_{t}^{+}(e)[h_{tq}+h_{tp}g_{p}^{+}(e)h_{pq } ] \label{h_eff_t}\end{aligned}\ ] ] the first term on the right - hand side is the closed quantum system hamiltonian , the second term describes a coupling with the one - particle continuum , and the third term corresponds to all possible couplings with the two - particle continuum .
@xmath152 is the green function in @xmath153 @xmath154^{-1}\ ] ] modified by couplings to @xmath128 .
similarly as in the case of one - particle continuum , one defines the source term @xmath155 of sm wave function @xmath156 into the two - particle continuum : @xmath157 matrix elements of the operator responsible for @xmath158 couplings ( the third term on the right - hand side of eq .
( [ h_eff_t ] ) ) can be expressed as the overlap between the source term and the function @xmath159 ( see eqs .
( [ intro_01 ] ) and ( [ intro_02 ] ) ) .
effective hamiltonian @xmath139 can be written in various equivalent forms . for our purpose , it is convenient to extract the direct coupling term between @xmath127 and @xmath153 subspaces : @xmath160\tilde{g}_{p}^{(+)}(e)[h_{pq}+h_{pt}g_{t}^{(+)}(e)h_{tq } ] ~ \ .
\label{h_eff_t2}\end{aligned}\ ] ] @xmath161 in ( [ h_eff_t2 ] ) stands for @xmath128-subspace green function modified by couplings to @xmath153 @xmath162^{-1 } ~ \ , \ ] ] where @xmath163^{-1}\ ] ] is the green function in @xmath153 .
@xmath139 ( cf eqs . ( [ h_eff_t ] ) and ( [ h_eff_t2 ] ) ) describes all emission processes involving one and two nucleons ( two protons ) .
observed 2p partial decay width is a sum of contributions of different interfering processes which can not be disentangled experimentally . in theoretical studies , one can always isolate a given processes by switching - off all others and calculate the contribution of such an isolated process to the 2p partial width .
however , relative weights of different isolated processes do not include interference effects and , hence , the resulting picture of the 2p decay is somewhat distorted . in this sense , and
contrary to the two - body decays , the interpretation of experimental 2p decay data is unavoidably model dependent .
in the observed 2p decays , certain emission scenarios may be less probable , so it is instructive and certainly less cumbersome to consider limiting cases of the 2p decay .
for example , if @xmath164 couplings are weak then the dominant process is a direct 2p emission described by @xmath165 where the second term on the right - hand side is responsible for couplings to one - nucleon decay channels .
on the contrary , if an intermediate system @xmath166 $ ] plays an important role , then @xmath167 couplings can be neglected and the effective hamiltonian reads : @xmath168 this operator describes a standard sequential 2p emission via an intermediate resonance in the @xmath166 $ ] nucleus . the sequential 2p emission process may also occur via continuum states which are correlated by either weakly bound or by weakly unbound states of @xmath166 $ ] nucleus .
this process which goes through the tail of an energetically inaccessible state , is called the _ virtual sequential _ two - body decay @xcite . for its study , it is important to separate 1p- and 2p - decays in the effective hamiltonian : @xmath169 the second term in eq .
( [ 3_express ] ) describes the 1p decay , whereas the third term describes the genuine sequential decay . the sequential 2p decay via the correlated scattering continuum of @xmath166 $ ] nucleus ( the virtual sequential emission process ) is an important component of the 2p decay mechanism .
an effective hamiltonian for this process ( cf eq .
( [ 3_express ] ) ) describes interference effects between the 1p decay and the sequential 2p decay , both for closed ( virtual @xmath127-@xmath128 excitations ) and opened ( 1p decays ) 1p decay channels .
diagonalizing @xmath170 , one obtains energies of states in @xmath133 $ ] nucleus as well as their widths associated with the emission of one and two protons . in general
, these two decay modes can not be separated one from another . however , since couplings responsible for the 2p decay are smaller than those associated with the 1p decay , one may proceed in two steps . in the first step , one neglects the term @xmath171 ( cf eq .
( [ 3_express ] ) ) responsible for the 2p decay , and diagonalizes the remaining operator in a sm basis @xmath172 .
this provides new basis states @xmath173 which are linear combinations of sm states . in the second step , using those mixed sm states one calculates the sequential 2p emission width of a decaying state @xmath174 , i.e. one calculates an imaginary part of the matrix element @xmath175 let us assume that different steps of the sequential process are uncorrelated , i.e. the first emitted proton is a spectator of the second emission .
this implies : @xmath176 where primed quantities refer to @xmath109-nucleon space . in the @xmath177 subspace ,
@xmath109 nucleons are in quasi - bound single - particle orbits , whereas in the @xmath178 subspace , one nucleon is in the continuum and @xmath179 nucleons are in quasi - bound single - particle orbits .
@xmath180 is a one - body potential describing an average effect of @xmath109-particles on the emitted proton and @xmath181 denotes a projector on one - particle continuum states . with this identification
, @xmath182 becomes a coupling between newly defined @xmath177 and @xmath178 subspaces .
the virtual sequential 2p emission appears always if @xmath183 . in that sense ,
the virtual sequential process is a component of any quantum - mechanical three - body decay . in the approximate scheme employed in smec ,
the virtual sequential 2p decay process competes with the diproton mode even for closed 1p emission channels .
the effective hamiltonian ( [ h_eff_direct ] ) describes both the three - body direct 2p decay and its limit of two sequential two - body decays . in this section
, we shall discuss the latter two - step scenario @xcite .
in the first step , protons are emitted as a @xmath19he -cluster which , in the second step , disintegrates outside of a nuclear potential due to the final state interaction @xcite .
the final state proton - proton interaction is taken into account by the density @xmath184 of proton states corresponding to the intrinsic energy @xmath185 in the proton - proton system @xcite .
the total system @xmath186}\otimes{\rm [ 2p]}$ ] is described as the two - body system in a mean - field @xmath187 : @xmath188 { \hat t}^{(cl ) } ~ \ , \label{h_tt_cluster}\end{aligned}\ ] ] where @xmath189 is the intrinsic hamiltonian of the daughter nucleus @xmath190 $ ] , and @xmath191 is the intrinsic hamiltonian of @xmath19he-cluster .
the cluster is described as a particle of mass @xmath192 ( @xmath193 denotes the proton mass ) and charge @xmath194 .
@xmath195 is the intrinsic kinetic energy of the system @xmath186}\otimes{\rm [ 2p]}$ ] , and @xmath20 is the reduced mass of the system .
@xmath196 is the projection operator on the subspace of cluster states in the continuum of the potential @xmath197 .
the 2p decay width is given by the matrix element : @xmath198}&=&-2 im\left ( \langle{\tilde \phi}^{(int)}_{i}|h_{qt}g_{t}^{(+)}(e)h_{tq}|{\tilde \phi}^{(int)}_{i}\right ) \rangle \\ \nonumber & = & -2 i m \left(\langle w_i^t|\omega_i^{t,(+)}\rangle \right ) ~ \ , \end{aligned}\ ] ] where @xmath199 is the intrinsic state corresponding to a mixed state @xmath200 in the parent nucleus , @xmath201 is the source term , and @xmath202 is a continuation of @xmath199 in @xmath153 .
the coulomb interaction is included as an average field and does not enter in @xmath203 .
the matrix element ( [ eq4 ] ) is the integral over an intrinsic energy @xmath185 of the overlap of a channel - projected source term and a channel - projected continuation in @xmath153 of an intrinsic state @xmath199 , weighted by the proton state density @xmath184 @xcite . working with intrinsic states allows to account for the recoil correction .
the decay energy is divided into an intrinsic energy of the @xmath19he-cluster and its center of mass motion energy according to the proton - state density @xcite . in this way ,
intrinsic degrees of freedom of the two protons are reduced and , therefore , the three - body decay is reduced to the two sequential two - body decays . [ [ rm - r - matrix - model - of - brown - and - barker - barker_theo ] ] @xmath204-matrix model of brown and barker [ barker_theo ] + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + this model extends the r - matrix approach for sm description of a three - body decay @xcite .
similarly as in the sm approach for two - body decays , one may assume that the three - body decay is independent of the small - distance many - body dynamics .
this opens for a definition of three - body spectroscopic factors which can be computed in the sm framework @xcite in an analogous way as the preformation factors used to describe the @xmath2-particle decay .
the r - matrix model of a diproton decay @xcite includes the @xmath205-wave proton - proton interaction as an intermediate state .
this final state interaction is crucial for a quantitative description of experimental data , changing the diproton lifetimes by several order of magnitudes .
@xmath47he preformation factor is taken into account by @xmath47he spectroscopic factors @xmath206}$ ] . to calculate them , the sm wave function
is projected onto the @xmath207 internal relative wave function for a @xmath47he-cluster @xcite : @xmath208}=\left ( \frac{a}{a-2 } \right)^{\lambda}g^2c(a , z)~.\ ] ] in @xmath209 shell model sace , @xmath210 , @xmath211 .
@xmath212 is the overlap for the @xmath47he wave function @xmath213 with @xmath214 , @xmath215 , and @xmath216 in the @xmath217 basis .
the extended r - matrix model @xcite neglects both the small - distance dynamics and the virtual scattering to continuum states .
the latter process changes the @xmath19he emission probability by modifying the wave function of a decaying state and changing the @xmath19he preformation factor @xcite .
this change is expected to be particularly strong for states close to the 2p emission threshold(s ) , e.g. the ground states of @xmath11fe , @xmath14ni and @xmath13zn .
reasons for the continuum - induced change of @xmath19he preformation probability or one - nucleon spectroscopic factors @xcite are the same as for a dominance of cluster states close to their respective decay thresholds @xcite : both phenomena are consequences of the alignment of near - threshold states with the channel state @xcite .
it is not easy to assess quantitative consequences of those simplifications in the extended r - matrix model . in the following sections
, we shall try to address this problem by ( i ) comparing smec results for diproton decays with those of an extended r - matrix approach , and ( ii ) by analyzing smec results for virtual sequential 2p emission with and without external mixing . in general
, structures in the continuum are more difficult to treat than bound - state problems .
firstly , the three - body problem for unbound systems is much less established , although studied for short and long - range interactions @xcite .
secondly , the coulomb problem with three - body asymptotics is still considered unsolved @xcite even though an important progress has been reported recently in the description of proton - deutron scattering and of three - nucleon electromagnetic reactions involving @xmath43he @xcite .
an infinite range of the coulomb interaction in channel - coupling potentials does not allow to decouple equations at infinity .
consequently , an asymptotic behaviour of @xmath218 can not be defined without a certain approximation .
one approximate way to proceed is to neglect off - diagonal channel - coupling potentials above a certain value of the hyperradius and to define an effective sommerfeld parameter from diagonal coupled - channel potentials @xcite .
this technique has been applied in the three - cluster model of the 2p emission @xcite . in order to calculate the direct 2p emission probability
, one has to evaluate the matrix element ( [ eq4 ] ) without a simplifying assumption of two sequential two - body decays .
even though this three - body problem has been conveniently formulated in jacobi coordinates @xcite , its numerical solution in smec has not yet been given .
firstly , if the residual interaction is a contact force , then the interaction of two protons in the continuum leads to the ultra - violet divergence . in principle
, this divergence could be avoided using a finite - range residual interaction .
however , such an interaction leads to non - local channel - coupling potentials and , hence , equations for channel - projected @xmath219 functions become integro - differential coupled - channel equations . [ [ the - three - body - cluster - model - the - hyperspherical - adiabatic - expansion - method - garr_theor ] ] the three - body cluster model : the hyperspherical adiabatic expansion method [ garr_theor ] + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in this approach , the dynamics of @xmath133$]-particle system is reduced to a three - body dynamics which is described by solving faddeev equations @xcite .
the three - body wave function is a sum of three faddeev components , each of them expressed in one of the three possible sets of jacobi coordinates @xmath220 .
each component is then expanded in terms of a complete set of hyperangular functions .
essential quantities in this approach are the effective radial potentials in the hyperradius variable : @xmath221 with @xmath222 , where @xmath223 is the coordinate of particle @xmath224 , @xmath225 and @xmath226 is an arbitrary normalization mass .
the set of indices @xmath227 is a permutation of @xmath228 .
@xmath229 , @xmath230 are proportional to the distance between two particles and the distance between their center of mass and the third particle , respectively . the behaviour of effective radial potentials ( the hyperspherical adiabatic potentials ) as a function of @xmath231 for different angular momentum quantum numbers determines the structure of a three - body system at both small and large distances . for short - range interactions , the lowest hyperspherical adiabatic potential usually dominates in the expansion of the wave function .
if both short and long - range interactions are present , the wave - function expansion at large distances requires more components . with the dominating adiabatic potential , the width of a given three - body resonance
can be estimated as @xmath232^{1/2}d\rho \right\ } ~ \ , \label{wkb}\end{aligned}\ ] ] where @xmath233 is the energy of the resonance . @xmath234 and @xmath235 in eq .
( [ wkb ] ) are the inner and outer classical turning points defining endpoints of the pass through the barrier . if crossings of various adiabatic potentials occur along the way through the barrier , the system is supposed to follow a path corresponding to the smoothest adiabatic potential .
the hyperspherical adiabatic expansion ( hae ) method has been extensively used to study different possible situations in three - body decay resulting from various combinations of short and long - range interactions and different binding - energy situations of all two - body subsystems @xcite . in particular
, it has been applied for the description of the borromean nucleus @xmath236 @xcite . in this application ,
a parameterized proton - proton interaction was used @xcite and the proton - core interaction was described by an @xmath237-dependent potential with central , spin - spin and spin - orbit radial potentials .
the hae approach allows to distinguish between sequential and direct three - body decays .
large separation between all pairs of particles in the final state can be achieved either by a uniform separation of all particles or by first moving one particle to infinity while the others remain at essentially the same distance from each other and second by moving the two close - lying particles apart .
the latter scenario defines the sequential process where the second step of a two - step process starts after the first particle is at a distance larger than the initial size of the three - body system . in @xmath53ne , both the ground state ( 1/2@xmath238 ) and the first excited state ( 3/2@xmath238 ) have a lower energy than the proton - unstable ground state of @xmath33f .
hence , the standard sequential decay process through the intermediate resonance(s ) in @xmath166 $ ] nucleus is not possible .
however , nothing prevents the decay from proceeding via the energetically favourable path described by the lowest adiabatic potential until at some point the energy conservation dictates that also this two - body structure must be broken . this mechanism which resembles the virtual sequential decay process discussed above ( see sect .
[ sub_seq ] ) leads to a significant reduction of the 2p partial half - life of the 3/2@xmath238 state in @xmath53ne @xcite .
principal assumptions of the hae approach are ( i ) the adiabatic motion in the three - body system , and ( ii ) the cluster ansatz for the many - body wave function of the total system .
the latter assumption will be discussed in sect .
[ grig_theor ] in connection with another three - body cluster model , the so - called systematic @xmath239 model ( @xmath93 ) @xcite .
the adiabaticity assumption is better satisfied for certain collective states and for a spontaneous fission process .
however , this hypothesis becomes questionable if several adiabatic potentials are crossing as found in charged - particles / fragments emission @xcite .
[ [ the - three - body - cluster - model - the - systematic - ell2-model - grig_theor ] ] the three - body cluster model : the systematic @xmath239-model [ grig_theor ] + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the @xmath93 of grigorenko et al .
@xcite is a flexible tool to simulate various consequences of three - body dynamics in 2p emission decay .
this model is a variant of a three - cluster model with outgoing flux . like in hae approach ,
the @xmath93 uses hyperspherical harmonics to treat the decay of a three - body system .
however , unlike the hae method , the @xmath93 includes essentially only the direct three - body decay . in general ,
the three - cluster models are useful for understanding the structure of certain states in light nuclei , such as @xmath42he , @xmath42li , @xmath42be @xcite , @xmath240li , @xmath240b @xcite , and @xmath48li @xcite .
the basic assumption in @xmath93 is that the intrinsic constituent particle degrees of freedom can be implicitly treated through the judicious choice of both the active few - body degrees of freedom and the appropriate effective two - body interactions .
the cluster ansatz is physically more relevant for states close to the cluster - decay threshold @xcite .
hence , the three - cluster models are perhaps better suited for halo systems . on the other hand , couplings to the decay channels involving intrinsic states of a daughter nucleus with excitation energies up to @xmath241 mev @xcite
are strongly present for near - threshold states .
therefore , even in halo states the wave function is a superposition of several different cluster configurations . in heavier nuclei ,
the halo wave functions are too complex to be reduced into a simple three - cluster parameterization .
nevertheless , the @xmath93 provides in those nuclei an order of magnitude estimate of the relation between the three - body decay energy , the effective interactions in the proton - proton and proton - core subsystems , and the probabilities of the @xmath239 configuration of the 2p emitter .
such estimates are very helpful in phenomenological applications . moreover , by solving decay kinematics in hyperspherical variables , the @xmath93 allows to study angular correlations and energy / momentum correlations between decay products which presently can not be investigated by any other approach . in @xmath93 ,
one begins by calculating the internal wave function of the system @xmath242 at a discrete value of the energy @xmath243 .
this wave function is a solution of the schrdinger equation : @xmath244 with zero boundary condition at some finite value of the hyperradius @xmath245 .
the time - independent part of the wave function for decay particles @xmath246 with purely outgoing asymptotics and an energy @xmath243 is found by solving the inhomogeneous equation @xmath247 with the source term given by an internal wave function .
this way of modelling is based on the assumption that the decaying state is a narrow resonance with a negligible coupling to one- and two - particle continua of decay channels .
this assertion is questionable for couplings to one - particle decay channels , either closed or opened , which induce the dynamical effects in the 2p decay .
the 2p separation energy in @xmath93 is adjusted using a weak short - range three - body force .
it is assumed that this force does not modify the barrier penetration process .
an essential element of the @xmath93 is the proton - proton final - state interaction which is tuned to the proton - proton phase shift . in the proton - core channel ,
adjustable potentials of the woods - saxon form are employed .
the internal structure of the parent nucleus @xmath133 $ ] is tuned by changing parameters of the proton - core interaction potential for different @xmath237 values . in this way , one generates wave functions with different populations of @xmath239-configurations .
clearly , this treatment excludes genuine many - body effects , such as the configuration interaction and configuration mixing , or the pauli blocking which is approximated in @xmath93 by repulsive cores for occupied orbitals . both @xmath93 and hae approaches
are precisely tuned to describe the three - body structures and decays .
this gives them a certain conceptual advantage over sm - like approaches . on the other hand , an internal structure of the decaying state
is oversimplified , the treatment of pauli s exclusion principle is incomplete , and the information about the complexity of many - body resonance wave function is lost in adjusted interaction potentials .
certainly , some of deficiencies of three - body cluster models can be hidden by a judicious choice of phenomenological potentials . however , the calculated results can hardly be more reliable than the basic assumptions . due to an unrealistic structure of the decaying 2p pair in @xmath93
, the relative weight of various @xmath239-configurations obtained by fitting the experimental 2p half - life can not be used to extract meaningful information about proton - proton correlations inside a nucleus .
it is then astonishing that the @xmath93 can give an order of magnitude estimate of the experimental data .
results of configuration - interaction approaches ( smec and extended r - matrix approach ) can be compared to evaluate the role of nuclear dynamics and continuum couplings in the 2p decay .
unfortunately , no meaningful comparison is possible between sm - like approaches and three - body cluster models ( hae approach or @xmath93 ) . in this section
, we shall discuss results of smec and extended r - matrix approach for @xmath11fe , @xmath14ni and @xmath13zn assuming the diproton scenario of 2p radioactivity .
we shall also evaluate in smec an importance of the virtual sequential process in these decays . the experimental evidence for the 2p radioactivity in @xmath11fe , @xmath14ni and @xmath13zn has been discussed in sects . 6.2 , 6.4 and 6.3 , respectively ( see also table 1 ) . in smec studies of @xmath11fe , internal many - body states in @xmath127
are obtained by solving the standard sm with the iokin hamiltonian @xcite .
configurations with excitations from the @xmath248 shell to the @xmath209 shell are excluded .
the residual couplings between @xmath127 subspace and the embedding continuum is given by the wigner - bartlett interaction : @xmath249\delta(\bf{r_{2}}- \bf{r_{1}})$ ] , with @xmath250 ( @xmath251 ) and @xmath252 mev@xmath253@xmath254 .
@xmath255 is the spin - exchange operator .
the calculated diproton half - life is @xmath256ms @xcite .
the error bar is associated with an experimental uncertainty on the decay energy @xmath55 .
external mixing of sm wave functions due to the @xmath127-@xmath153 couplings gives a negligible correction .
more important are the @xmath127-@xmath128 couplings .
the external mixing generated by these couplings reduces the diproton half - lives by about 10% for @xmath94 in the interval @xmath257 .
this reduction depends weakly both on the magnitude of @xmath94 and its sign . the experimental partial half - life for 2p radioactivity is by factor @xmath258 shorter ( cf table 1 ) .
the extended r - matrix model yields the diproton half - life of @xmath259ms @xcite .
the nuclear structure enters in this model through @xmath19he spectroscopic factors which are calculated in @xmath209 shell using the gpfx1 hamiltonian @xcite .
the calculation using fpd6 interaction @xcite yields a similar value of the half - life .
this half - life is by a factor @xmath258 slower than the smec prediction .
the sequential 2p decay is always a part of the 2p decay , independently of the sign and magnituude of @xmath94 .
one should remind that the formal separation between the diproton decay and the virtual sequential emission can not be justified if these two modes yield comparable partial decay widths , i.e. @xmath260 .
let us consider the sequential emission of @xmath11fe through the @xmath261 and @xmath262 continua associated with the ground state ( @xmath263 ) and the first excited state ( @xmath264 ) of @xmath265 , respectively .
for @xmath266 , both @xmath267 and @xmath268 states are resonances which decay by 1p emission .
details of calculations can be found in ref .
@xcite .
the sequential emission half - life depends strongly both on the position of 1p emission threshold ( @xmath94 ) and on the total decay energy ( @xmath55 ) . assuming that @xmath11fe is stable with respect to 1p emission ( @xmath269mev ) , the virtual sequential 2p half - life is @xmath270 ms , i.e. about 10 times slower than the diproton decay .
this estimate includes an effect of external mixing of @xmath271 sm wave functions , which reduces the virtual sequential decay half - life by about 30% . for @xmath272mev , even though the standard sequential process through an intermediate resonance is possible , its contribution is totally screened by the coulomb barrier .
the dominant decay mechanism remains the virtual sequential process through the ghost of ground state of @xmath273mn , far away from the resonance region @xmath274 ( see fig .
[ seq_dep_eps_fig ] ) .
the half - life @xmath275 ms is about a factor of 8 slower than the diproton decay .
couplings to the decay channels @xmath276 associated with an excited state of @xmath273mn are relatively important and further reduce the half - life by about 30% for @xmath277 mev and about 38% for @xmath269 mev .
one can not exclude that also decay channels connected with higher lying states of @xmath273mn contribute significantly to the reduction of the 2p partial half - life @xcite .
an energy of the first emitted proton is limited by @xmath55 .
the most probable energy of first proton in the virtual sequential decay @xcite ( @xmath278 mev ) slightly decreases with increasing @xmath94 ( see fig .
[ seq_dep_eps_fig ] ) .
a small asymmetry seen in this distribution is a result of different barriers for proton emissions in @xmath11fe and in @xmath273mn .
the full width at half maximum of a ghost is @xmath279 mev at @xmath269 mev .
@xmath280 increases with @xmath94 in the range of @xmath94 values where the virtual 2p decay dominates .
experimental observation of symmetric energy distribution of individual protons @xcite indicates that either @xmath94 in @xmath11fe is less than @xmath281 mev , or the diproton decay mechanism dominates .
the latter scenario is favoured for typical effective sm interactions and @xmath282-values .
one should stress that these conclusions have been obtained assuming a coherent contribution of diproton and virtual sequential decay mechanisms . for @xmath283 mev
, the sequential 2p decay is predominantly related to the strength of a ghost in the energy interval @xmath284 and not to the strength of 1p resonance at @xmath94 . in this regime
, the sequential 2p decay width does not reduce to the product of the width for the first step ( @xmath285 ) and the branching ratio for the second step , as one would obtain semi - classically . above @xmath286 mev
, the sequential decay is dominated by the standard two - step process via the ground state of @xmath265 .
the @xmath94-dependence of the sequential 2p decay half - life is shown as full points in fig .
[ seq_dep_q1_fig ] .
squares in fig .
[ seq_dep_q1_fig ] represent the 1p decay half - life and the diproton half - life is depicted by the dashed - dotted line . the sequential 2p emission and the diproton emission , considered as independent , non - interfering decay paths , yield comparable half - lives for @xmath287 mev . in this interval ,
the three - body 2p decay can not be reduced to any sequence of two - body decays .
an interval of @xmath94 values for which the half - lifes of virtual sequential and diproton decays are expected to become similar lies close to the range of predicted @xmath94-values @xcite . hence , both processes could interfere strongly and , in spite of symmetric energy distribution of individual protons @xcite , the measurement of @xmath94 in @xmath11fe is mandatory before drawing definitive conclusions about the 2p emission mechanism in this nucleus .
the sequential 2p decay has two distinct dependences of the decay half - life on @xmath94 ( see fig .
[ seq_dep_q1_fig ] ) . for small positive values of @xmath94 ( @xmath288 mev )
as well as for closed 1p decay channels , the sequential 2p decay half - life changes linearly with @xmath94 .
this is the regime of virtual sequential 2p decay , where @xmath289 and the 2p decay goes via the ghost of the ground state . for @xmath290 mev ,
one enters in the regime @xmath291 , where the role of @xmath292 resonance in @xmath273mn is dominant . in this regime of standard sequential decay , the 2p partial half - life
: @xmath293 has a minimum at @xmath294 . in ref .
@xcite , several sm hamiltonians including iokin @xcite , kb3 @xcite or gxpf1 @xcite were used to evaluate the sensitivity of smec results to the choice of an effective interaction . for the iokin hamiltonian , the diproton half - life
is @xmath295ms .
the error bars are related with an experimental uncertainty of the decay energy @xmath55 .
external mixing of sm wave functions generated by the @xmath127-@xmath153 and @xmath127-@xmath128 couplings are unimportant , reducing the diproton half - life by @xmath296% in the interval @xmath257 .
somewhat stronger is the dependence of 2p partial half - lifes on the sm interaction . for gxfp1 and kb3 interactions ,
one finds @xmath297ms and @xmath298ms , respectively .
for all three sm interactions , the smec results are consistent with the experimental partial half - life .
the extended r - matrix model with gxfp1 interaction yields a diproton half - life @xmath299ms
@xcite which is about a factor of 2 slower than the smec result . in the description of virtual sequential emission ,
the 2p emission through the @xmath300 and @xmath301 continua associated with the ground state ( @xmath302 ) and the first excited state ( @xmath303 ) of @xmath304co were considered .
smec calculations were performed with the iokin hamiltonian .
principal uncertainties of the smec calculations are ( i ) an unknown position of 1p emission threshold in @xmath14ni , and ( ii ) an insufficient precision of the total decay energy . if @xmath269mev , then the virtual sequential 2p half - life is @xmath305ms , i.e. about 4 times slower than the diproton decay .
this estimate includes an external mixing of @xmath306 sm wave functions generated by the @xmath127-@xmath128 couplings .
these couplings reduce the virtual sequential decay half - life by about 15% . for @xmath272mev ,
the standard sequential decay through an intermediate resonance in @xmath304co is totally screened by the coulomb barrier .
hence , the dominant decay mechanism remains the virtual sequential process through the ghost of @xmath304co ground state .
the half - life in this case ( @xmath307ms ) is only @xmath308 slower than the diproton decay .
couplings to the decay channels @xmath309 associated with the excited state of @xmath304co are relatively unimportant , reducing the half - life by @xmath310 - @xmath311 .
the diproton half - life of @xmath13zn in smec with gxfp1 hamiltonian @xcite is @xmath312ms .
the error bar is due to the experimental uncertainty on the decay energy @xmath55 @xcite .
with kb3 hamiltonian , the diproton half - life is somewhat longer and equals @xmath313ms @xcite .
the experimental partial half - life for 2p radioactivity is about a factor of 4 faster .
as in @xmath11fe and @xmath14ni , a significant contribution from virtual sequential 2p decay can not be excluded but , again , the quantitative estimate of this contribution can not be made without knowing @xmath94 in @xmath13zn .
the extended r - matrix model with gxfp1 interaction yields the diproton half - life @xmath314ms
@xcite which is about 30% shorter than the smec prediction .
the coupling of a nuclear system to the environment of its decay channels may change its properties in a non - perturbative way @xcite .
the magnitude of the system - environment coupling depends explicitly on the location of various emission thresholds and on the structure of @xmath103-matrix polers .
the description of resonances and their decays are mainly sensitive to the two ingredients : the effective hamiltonian @xmath141 ( see sect .
[ smec_1p ] ) and the unitarity of the @xmath103-matrix which causes a non - trivial energy dependence of the coupling matrix elements between resonance states and continuum .
@xmath141 includes basic structural information contained in @xmath122 ( the closed quantum system hamiltonian ) and the coupling matrix elements between discrete and continuous states which depend on the structure of the decay channels .
smec studies of 2p decays in @xmath11fe , @xmath14ni and @xmath13zn have shown that the virtual sequential 2p emission can not be separated easily from the diproton decay , even though they describe two different limits .
the first one corresponds to two uncorrelated proton emissions separated in time , whereas the latter one represents an instantaneous emission of strongly correlated two protons .
the most probable energies of these protons are nearly the same in both scenarios .
if half - lives for these two pathways are strongly different , then neglecting their mutual interference and the omission of spatio - temporal correlations in the virtual sequential decay can be probably justified . otherwise , the 2p emission described by an effective hamiltonian ( [ h_eff_t2 ] ) has to be considered as a genuine three - body decay .
unfortunately , the 1p emission thresholds in @xmath11fe , @xmath14ni and @xmath13zn are unknown so that one can not rule out the possibility of a strong interference between ( virtual ) sequential and diproton decay modes .
smec calculations in @xmath11fe with @xmath94 values which are consistent with available theoretical estimates @xcite show that the virtual 2p sequential half - life is by a factor of about 10 longer than the diproton half - life .
this would favour an interpretation of the observed 2p emission in this nucleus as predominantly the diproton decay .
such a conclusion agrees also with the observation of a symmetric energy distribution of individual protons in the decay of @xmath11fe @xcite . for the sake of argument ,
let us suppose that the virtual sequential emission mechanism yields much longer half - lives in all three nuclei : @xmath11fe , @xmath14ni and @xmath13zn .
it is then interesting to notice that the diproton decay mechanism in smec reproduces quite well the experimental partial 2p decay half - life in a doubly magic nucleus ( @xmath14ni ) and fails by a factor of about 4 in both open - shell nuclei ( @xmath11fe and @xmath13zn ) .
if this tendency is not coincidental , or follows from a poor knowledge of @xmath282-values and an insufficient experimental statistics , then one might be tempted to associate it with pairing correlations .
as suggested in ref .
@xcite , the change of pairing correlations induced by the configuration mixing beyond the 1p0f shell may well be responsible for this enhancement . a similar effect has been seen a long time ago in ( p , t ) transfer reactions @xcite .
the absence of this component of pairing correlations is more important in open - shell nuclei than in a closed - shell nucleus @xmath14ni .
hence , one may expect that the fp shell smec calculations are better suited to describe the observed 2p partial half - life in @xmath14ni than in @xmath11fe and @xmath13zn .
one should mention that different effective interactions in the @xmath209 and @xmath315 shells give similar results to within @xmath316% .
external mixing , generated by the @xmath127-@xmath128 coupling , modifies the diproton decay width by less than 10% in @xmath11fe and @xmath296% in @xmath14ni .
again , the smec predictions for the closed - shell nucleus @xmath14ni seem to be more reliable since the direct @xmath94-dependence of the diproton half - life is weaker in this case .
the r - matrix model @xcite calculations show a gradual decrease of the 2p partial half - life going from @xmath11fe to @xmath13zn across the shell closure at z=28 , n=20 .
consequently , the diproton half - life in @xmath11fe is by a factor of @xmath317 longer than in @xmath13zn . on the contrary , both in the smec results and in the data , the 2p partial half - lives of @xmath11fe and @xmath13zn are very close . , @xmath318 , and @xmath319 , as calculated in the generalized r - matrix theory @xcite ( full triangles ) and in the smec @xcite(open circles ) .
the experimental values are shown with black points . for reasons of clarity ,
the uncertainty of theoretical results due to the uncertainty of the @xmath55 values and experimental error bars are not shown.,width=642,height=453 ] fig . [ systematic ]
compares the experimental partial half - lives for the diproton decay of @xmath320 , @xmath318 , and @xmath319 with the calculated ones in the smec and in the extended r - matrix model .
one may notice two characteristic features from this comparison .
firstly , smec and r - matrix models predict a different z - dependence of calculated 2p partial half - lives around the doubly - magic nucleus @xmath14ni .
secondly , smec yields an opposite tendency of @xmath11fe and @xmath13zn half - lives with respect to the half - life of @xmath14ni than seen experimentally . concerning the first point
, one may tentatively assume that this feature is related to the continuum coupling induced dynamical effects in the 2p decay which are missing in the r - matrix description .
certainly , one does not expect that this different systematic tendency of calculated half - lives in the smec approach and in the r - matrix model could be explained by the absence of configuration mixings beyond the 1p0f shell , since both in smec and in r - matrix calculations these configurations are not included .
concerning the second point , it may be related to the small number of channels in smec calculations of the @xmath127-@xmath128-coupling induced an external configuration mixing . in ref .
@xcite , the external mixing is a result of coupling between the ground state of nucleus @xmath133 $ ] and the two channels corresponding to the ground state and the first excited state of a nucleus @xmath166 $ ] .
this restriction on the number of channels is justified in closed - shell nuclei as in this case the ground - state to ground - state coupling of the @xmath133 $ ] and @xmath166 $ ] nuclei exhausts up to about 90% of the total coupling strength . for open - shell nuclei , the strength of continuum coupling
is largely spread over many channels , both open and closed .
indeed , the smec studies of continuum coupling effects in oxygen and fluorine isotopes @xcite have shown that a contribution of the ground - state to ground - state coupling to the total continuum coupling strength may vary between a few @xmath321 and @xmath322 in open - shell nuclei .
the size of a model space involved in the calculation of the 2p decay in @xmath320 , @xmath318 , and @xmath319 makes it impossible to verify this observation numerically , but if true it may lead to a strong enhancement of the contribution from the virtual sequential 2p decay path and , henceforth , to a decrease of the 2p partial half - life .
one should remind , that the interference between the virtual sequential and diproton emission paths can not be probably neglected if both processes become equally important . as said before , smec predictions are very sensitive to @xmath282 values
which for the nuclei studied are either not known experimentally with a sufficient precision ( @xmath55 values ) , or are unknown ( @xmath94 values ) .
the virtual sequential decay half - lives depend explicitly and strongly on the continuum coupling dynamics between the @xmath127 , @xmath128 and @xmath128 , @xmath153 subspaces which is governed by @xmath282 values .
a direct dependence of the diproton half - lives on @xmath127-@xmath128 virtual couplings is rather weak , i.e. dressing of the diproton decay mode by virtual 1p excitations does not change significantly its properties in @xmath11fe , @xmath14ni and @xmath13zn state of @xmath56ne @xcite . in this case ,
the diproton mode is dressed by virtual excitations to the _ opened _ 1p emission channels leading to the particle stable @xmath323 , @xmath324 final states in @xmath53f . in general , couplings to open 1p emission channels strongly modify properties of the diproton mode . ] . on the other hand , this idealized decay mode represents a sensible approximation of the 2p decay process described by ( [ h_eff_t ] ) only if the virtual sequential decay path is strongly suppressed .
this in turn depends directly on the dynamics of the @xmath127 , @xmath128 and @xmath128 , @xmath153 virtual couplings .
hence , further progress in describing spontaneous 2p radioactivity and understanding the role of both pairing correlations and virtual excitations to continuum states in this process depends on a precise experimental determination of the @xmath94 and @xmath55 values .
the study of 2p emission opens a new window to investigate nucleon - nucleon correlations and the structure of atomic nuclei .
although these studies , at least partially , could have been performed already with @xmath0-delayed 2p emitters , the field received a strong boost with the observation of ground - state 2p radioactivity .
this discovery triggered new experimental developments , such as time - projection chambers , and new theoretical activity related mainly to the unified description of structure and reactions with two - particle continua and the development of various methods to deal with three - body asymptotic states of charged particles / fragments .
both experiment and theory have reached a certain level of sophistication , but significant improvements are still needed to develop 2p radioactivity into a powerful tool for nuclear structure studies . on the theoretical side ,
the lack of three - body asymptotic channels hampers the calculation of various correlation functions in the smec framework .
description of direct 2p emission with three - body asymptotics in smec requires inclusion of finite - range interactions . at the moment
, the angular correlations between decay products can be studied solely in the three - body cluster models .
the simplified description of nuclear structure in cluster models does not allow to draw deeper conclusions about 2p in - medium correlations from these angular correlations . as one deals with three particles in the final state , the 2p emission can occur in different ways , i.e. the asymptotic channel can not be defined unambiguously .
consequently and in contrast to one - proton or @xmath2 radioactivity , the 2p radioactivity does not have a unique experimental signature and its interpretation is always depending on the theoretical interpretation .
a complete theoretical description of 2p emission has to take into account all different decay paths and include interferences between them .
it is not obvious whether different 2p decay paths , such as the diproton path and the virtual sequential 2p decay path which both yield similar proton energy distributions , can be distinguished by , e.g. , the proton - proton and/or proton - fragment correlations .
further studies along this line are necessary for a best possible characterization of 2p radioactivity . on the experimental side ,
more cases of 2p radioactivity have to be studied and the quality of the data has to be improved , as the possibilities to compare the experimental data existing today with modern theories of 2p radioactivity are hampered by large experimental error bars . to improve the experimental data , increased statistics is only one aspect . in the case of @xmath11fe ,
the major contribution to the uncertainty of the decay energy comes from the energy calibration .
therefore , new and improved calibration tools have to be developed as well .
a precise determination of @xmath94 and @xmath55 would help to check different features of smec calculations , such as the choice of effective interactions and valence space , or the absence of explicit three - body asymptotics .
all experiments performed up to now seem to indicate that the sequential decay prevails , as long as intermediate states for 1p emission are energetically accessible .
the only exceptions seem to be the decays of excited states in @xmath53ne and in @xmath17ag@xmath18 , where despite many intermediate states a strong proton - proton correlation was observed .
a halo - like structure and/or dynamical effects associated with the tunnelling through the anisotropic coulomb barrier could be at the origin of the observed correlation .
higher statistics data and advanced theoretical studies are needed to confirm or reject this interpretation .
another application of the 2p studies are direct resonant 2p
radiative capture processes @xcite .
it was demonstrated that the astrophysical 2p capture rates can be enhanced by as much as several orders of magnitude in certain ranges of excitation energies .
similar tendency was found for the size of a nonresonant e1 contribution to the 2p capture rates @xcite .
these predictions may have important consequences for the rp - process path changing earlier predictions concerning rp - process waiting points @xcite .
the advent of new radioactive beam machines like the ribf facility in riken , japan or the fair facility in gsi , germany should give access to new ground - state 2p emitters such as @xmath100ge , @xmath101se , or @xmath325kr and allow one to achieve much higher rates for the known ones .
these possibilities will most likely significantly increase the body of experimental data and may establish 2p radioactivity as a powerful tool of nuclear structure studies .
these new facilities may allow also to search for a similar phenomenon which is one- and two - neutron radioactivity .
from an experimental point of view , it seems to be rather difficult to find good candidates for these new decay modes . due to the absense of the coulomb barrier for neutrons ,
only the centrifugal barrier will prevent these neutrons from escaping `` immediately '' from the nucleus . to create a sufficiently high barrier ,
high - j orbitals are needed which form the nuclear ground state only in medium - mass nuclei .
however , these medium - mass neutron drip - line nuclei are rather far away from the valley of stability and extremely difficult to produce .
possibly , high - j isomers could be candidates for this kind of radioactivity . from a theoretical point of view
, two - neutron emitters would be ideal cases to study .
all theoretical formalisms can be applied to the two - neutron case without major changes .
actually , the two - neutron three - body problem can be solved exactly and not approximately as in the three - body case with charged particles . moreover
, angular correlations will be directly related to the internal two - neutron correlations without the uninteresting correlations due to a possible anisotropy of the coulomb barrier .
the experimental results on two - proton radioactivity were obtained in large collaborations including the cen bordeaux - gradignan , ganil , gsi , university of warsaw , and other laboratories .
b.b . would like to underline the decisive contributions of j. giovinazzo and m. pftzner to this work .
one of us ( m.p . ) wishes to thank j. okoowicz and j. rotureau for a fruitful collaboration which led to the development of the continuum shell - model for two - proton radioactivity . stimulating discussions with witek nazarewicz
are kindly acknowledged . | in the first part of this review , experimental results which lead to the discovery of two - proton radioactivity are examined . beyond two - proton emission from nuclear ground states ,
we also discuss experimental studies of two - proton emission from excited states populated either by nuclear @xmath0 decay or by inelastic reactions . in the second part
, we review the modern theory of two - proton radioactivity .
an outlook to future experimental studies and theoretical developments will conclude this review .
version : |
the purpose of this paper is to develop the geometric foundations for multisymplectic - momentum integrators for variational partial differential equations ( pdes ) .
these integrators are the pde generalizations of symplectic integrators that are popular for hamiltonian odes ( see , for example , the articles in marsden , patrick and shadwick [ 1996 ] , and especially the review article of mclachlan and scovel [ 1996 ] ) in that they are covariant spacetime integrators which preserve the geometric structures of the system . because of the covariance of our method which we shall describe below , the resulting integrators are spacetime localizable in the context of hyperbolic pdes , and generalize the notion of symplecticity and symmetry preservation in the context of elliptic problems . herein , we shall primarily focus on spacetime integrators ; however , we shall remark on the connection of our method with the finite element method for elliptic problems , as well as the gregory and lin [ 1991 ] method in optimal control .
historically , in the setting of odes , there have been many approaches devised for constructing symplectic integrators , beginning with the original derivations based on generating functions ( see de vogelaere [ 1956 ] ) and proceeding to symplectic runge - kutta algorithms , the shake algorithm , and many others .
in fact , in many areas of molecular dynamics , symplectic integrators such as the verlet algorithm and variants thereof are quite popular , as are symplectic integrators for the integration of the solar system . in these domains , integrators that are either symplectic or which are adaptations of symplectic integrators , are amongst the most
widely used .
a fundamentally new approach to symplectic integration is that of veselov [ 1988 ] , [ 1991 ] who developed a discrete mechanics based on a discretization of hamilton s principle .
their method leads in a natural way to symplectic - momentum integrators which include the shake and verlet integrators as special cases ( see wendlandt and marsden [ 1997 ] ) .
in addition , veselov integrators often have amazing properties with regard to preservation of integrable structures , as has been shown by moser and veselov [ 1991 ] .
this aspect has yet to be exploited numerically , but it seems to be quite important . the approach we take in this paper is to develop a veselov - type discretization for pde s in variational form .
the relevant geometry for this situation is multisymplectic geometry ( see gotay , isenberg , and marsden [ 1997 ] and marsden and shkoller [ 1997 ] ) and we develop it in a variational framework . as we have mentioned , this naturally leads to multisymplectic - momentum integrators .
it is well - known that such integrators can not in general preserve the hamiltonian _ exactly _ ( ge and marsden [ 1988 ] ) .
however , these integrators have , under appropriate circumstances , very good energy performance in the sense of the conservation of a nearby hamiltonian up to exponentially small errors , assuming small time steps , due to a result of neishtadt [ 1984 ] .
see also dragt and finn [ 1979 ] , and simo and gonzales [ 1993 ] .
this is related to backward error analysis ; see sanz - serna and calvo [ 1994 ] , calvo and hairer [ 1995 ] , and the recent work of hyman , newman and coworkers and references therein
. it would be quite interesting to develop the links with neishtadt s analysis more thoroughly .
an important part of our approach is to understand how the symplectic nature of the integrators is implied by the variational structure . in this way
we are able to identify the symplectic and momentum conserving properties after discretizing the variational principle itself .
inspired by a paper of wald [ 1993 ] , we obtain a formal method for locating the symplectic or multisymplectic structures directly from the action function and its derivatives .
we present the method in the context of ordinary lagrangian mechanics , and apply it to discrete lagrangian mechanics , and both continuous and discrete multisymplectic field theory .
while in these contexts our variational method merely uncovers the well - known differential - geometric structures , our method forms an excellent pedagogical approach to those theories .
[ [ outline - of - paper . ] ] outline of paper .
+ + + + + + + + + + + + + + + + + * in this section we sketch the three main aspects of our variational approach in the familiar context of particle mechanics .
we show that the usual symplectic @xmath0-form on the tangent bundle of the configuration manifold arises naturally as the boundary term in the first variational principle .
we then show that application of @xmath1 to the variational principle restricted to the space of solutions of the euler - lagrange equations produces the familiar concept of conservation of the symplectic form ; this statement is obtained variationally in a non - dynamic context ; that is , we do not require an evolutionary flow .
we then show that if the action function is left invariant by a symmetry group , then noether s theorem follows directly and simply from the variational principle as well .
* here we use our variational approach to construct discretization schemes for mechanics which preserve the discrete symplectic form and the associated discrete momentum mappings . *
this section defines the three aspects of our variational approach in the multisymplectic field - theoretic setting . unlike the traditional approach of defining the canonical multisymplectic form on the dual of the first jet bundle and then pulling back to the lagrangian side using the covariant legendre transform ,
we obtain the geometric structure by staying entirely on the lagrangian side .
we prove the covariant analogue of the fact that the flow of conservative systems consists of symplectic maps ; we call this result the _ multisymplectic form formula_. after variationally proving a covariant version of noether s theorem , we show that one can use the multisymplectic form formula to recover the usual notion of symplecticity of the flow in an infinite - dimensional space of fields by making a spacetime split .
we demonstrate this machinery using a nonlinear wave equation as an example .
* in this section we develop discrete field theories from which the covariant integrators follow .
we define discrete analogues of the first jet bundle of the configuration bundle whose sections are the fields of interest , and proceed to define the discrete action sum .
we then apply our variational algorithm to this discrete action function to produce the discrete euler - lagrange equations and the discrete multisymplectic forms . as a consequence of our methodology
, we show that the solutions of the discrete euler - lagrange equations satisfy the discrete version of the multisymplectic form formula as well as the discrete version of our generalized noether s theorem . using our nonlinear wave equation example , we develop various multisymplectic - momentum integrators for the sine - gordon equations , and compare our resulting numerical scheme with the energy - conserving methods of li and vu - quoc [ 1995 ] and guo , pascual , rodriguez , and vazquez [ 1986 ] .
results are presented for long - time simulations of kink - antikink solutions for over 5000 soliton collisions . *
this section contains some important remarks concerning the variational integrator methodology .
for example , we discuss integrators for reduced systems , the role of grid uniformity , and the interesting connections with the finite - element methods for elliptic problems .
we also make some comments on future work .
[ [ hamiltons - principle . ] ] hamilton s principle .
+ + + + + + + + + + + + + + + + + + + + + we begin by recalling a problem going back to euler , lagrange and hamilton in the period 17401830 .
consider an @xmath2-dimensional configuration manifold @xmath3 with its tangent bundle @xmath4 .
we denote coordinates on @xmath3 by @xmath5 and those on @xmath4 by @xmath6 .
consider a lagrangian @xmath7 . construct the corresponding action functional @xmath8 on @xmath9 curves @xmath10 in @xmath3 by integration of @xmath11 along the tangent to the curve . in coordinate notation , this reads @xmath12 the action functional depends on @xmath13 and @xmath14 , but this is not explicit in the notation . * _ _ * hamilton s principle seeks the curves @xmath15 for which the functional @xmath8 is stationary under variations of @xmath15 with fixed endpoints ; namely , we seek curves @xmath15 which satisfy @xmath16 for all @xmath17 , where @xmath18 is a smooth family of curves with @xmath19 and @xmath20 . using integration by parts ,
the calculation for this is simply @xmath21 the last term in ( [ 2 ] ) vanishes since @xmath22 , so that the requirement ( [ 1b ] ) for @xmath8 to be stationary yields the * _ _ * euler - lagrange equations @xmath23 recall that @xmath11 is called * _ _ * regular when the symmetric matrix @xmath24 $ ] is everywhere nonsingular . if @xmath11 is regular , the euler - lagrange equations are second order ordinary differential equations for the required curves .
[ [ the - standard - geometric - setting . ] ] the standard geometric setting .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the action ( [ 1 ] ) is independent of the choice of coordinates , and thus the euler - lagrange equations are coordinate independent as well .
consequently , it is natural that the euler - lagrange equations may be intrinsically expressed using the language of differential geometry .
this intrinsic development of mechanics is now standard , and can be seen , for example , in arnold [ 1978 ] , abraham and marsden [ 1978 ] , and marsden and ratiu [ 1994 ] .
the * _ _ * canonical @xmath25-form @xmath26 on the @xmath27-dimensional cotangent bundle of @xmath3 , @xmath28 is defined by @xmath29 where @xmath30 is the canonical projection .
the lagrangian @xmath11 intrinsically defines a fiber preserving bundle map @xmath31 , the * _ _ * legendre transformation , by vertical differentiation : @xmath32 we define the * _ _ * lagrange @xmath25-form on @xmath4 , the lagrangian side , by pull - back @xmath33 , and the * _ _ * lagrange @xmath0-form by @xmath34 .
we then seek a vector field @xmath35 ( called the * _ _ * lagrange vector field ) on @xmath4 such that @xmath36 , where the * _ _ * energy @xmath37 is defined by @xmath38 . if @xmath39 is a local diffeomorphism then @xmath35 exists and is unique , and its integral curves solve the euler - lagrange equations ( [ 3 ] ) .
in addition , the flow @xmath40 of @xmath35 preserves @xmath41 ; that is , @xmath42 .
such maps are * _ _ * symplectic , and the form @xmath41 is called a * _ _ * symplectic @xmath0-form .
this is an example of a * _ _ * symplectic manifold : a pair @xmath43 where @xmath44 is a manifold and @xmath45 is closed nondegenerate @xmath0-form . despite the compactness and precision of this differential - geometric approach , it is difficult to motivate and , furthermore , is not entirely contained on the lagrangian side .
the canonical @xmath25-form @xmath26 seems to appear from nowhere , as does the legendre transform @xmath39 . historically , after the lagrangian picture on @xmath4 was constructed , the canonical picture on @xmath28 emerged through the work of hamilton , but the modern approach described above treats the relation between the hamiltonian and lagrangian pictures of mechanics as a mathematical tautology , rather than what it is a discovery of the highest order .
[ [ the - variational - approach . ] ] the variational approach .
+ + + + + + + + + + + + + + + + + + + + + + + + + more and more , one is finding that there are advantages to staying on the `` lagrangian side '' .
many examples can be given , but the theory of lagrangian reduction ( the euler - poincar equations being an instance ) is one example ( see , for example , marsden and ratiu [ 1994 ] and holm , marsden and ratiu [ 1998a , b ] ) ; another , of many , is the direct variational approach to questions in black hole dynamics given by wald [ 1993 ] . in such studies ,
it is the variational principle that is the center of attention .
we next show that one can derive in a natural way the fundamental differential geometric structures , including momentum mappings , directly from the variational approach .
this development begins by removing the boundary condition @xmath46 from ( [ 2 ] ) .
equation ( [ 2 ] ) becomes @xmath47 where the left side now operates on more general @xmath48 ( this generalization will be described in detail in section ( 4 ) ) , while the last term on the right side does not vanish .
that last term of ( [ 5 ] ) is a linear pairing of the function @xmath49 , a function of @xmath5 and @xmath50 , with the tangent vector @xmath51 .
thus , one may consider it to be a @xmath25-form on @xmath4 ; namely the @xmath25-form @xmath52 .
this is exactly the lagrange @xmath25-form , and we can turn this into a formal theorem / definition : [ thm2.1 ] given a @xmath53 lagrangian @xmath11 , @xmath54 , there exists a unique @xmath55 mapping @xmath56 , defined on the second order submanifold @xmath57 of @xmath58 , and a unique @xmath59 @xmath25-form @xmath60 on @xmath4 , such that , for all @xmath9 variations @xmath61 , @xmath62 where @xmath63 the @xmath25-form so defined is called the * _ _ * lagrange @xmath25-form .
indeed , uniqueness and local existence follow from the calculation ( [ 2 ] ) , and the coordinate independence of the action , and then global existence is immediate . here then , is the first aspect of our method : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ using the variational principle , the lagrange @xmath25-form @xmath60 is the `` boundary part '' of the the functional derivative of the action when the boundary is varied .
the analogue of the symplectic form is the ( negative of ) the exterior derivative of @xmath60 . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for the mechanics example being discussed , we imagine a development wherein @xmath60 is so defined and we define @xmath64 . [
[ lagrangian - flows - are - symplectic . ] ] lagrangian flows are symplectic .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + one of lagrange s basic discoveries was that the solutions of the euler - lagrange equations give rise to a symplectic map .
it is a curious twist of history that he did this without the machinery of either differential forms , of the hamiltonian formalism or of hamilton s principle itself .
( see marsden and ratiu [ 1994 ] for an account of some of this history . ) assuming that @xmath11 is regular , the variational principle then gives coordinate independent second order ordinary differential equations , as we have noted .
we temporarily denote the vector field on @xmath4 so obtained by @xmath65 , and its flow by @xmath40 .
our further development relies on a change of viewpoint : we focus on the restriction of @xmath8 to the subspace @xmath66 of solutions of the variational principle .
the space @xmath66 may be identified with the initial conditions , elements of @xmath4 , for the flow : to @xmath67 , we associate the integral curve @xmath68 , @xmath69 $ ] .
the value of @xmath8 on that curve is denoted by @xmath70 , and again called the * _ _ * action .
thus , we define the map @xmath71 by @xmath72 where @xmath73 . the fundamental equation ( [ 7 ] ) becomes @xmath74 where @xmath75 is an arbitrary curve in @xmath4 such that @xmath76 and @xmath77 .
we have thus derived the equation @xmath78 taking the exterior derivative of ( [ 8 ] ) yields the fundamental fact that the flow of @xmath65 is symplectic : @xmath79 which is equivalent to @xmath80 this leads to the following : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ using the variational principle , the fact that the evolution is symplectic is a consequence of the equation @xmath81 , applied to the action restricted to the space of solutions of the variational principle . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in passing , we note that ( [ 8 ] ) also provides the differential - geometric equations for @xmath65 .
indeed , one time derivative of ( [ 8 ] ) and using ( [ actiont.equation ] ) gives @xmath82 , so that @xmath83 if we define @xmath84 . thus , we quite naturally find that @xmath85 .
of course , this set up also leads directly to hamilton - jacobi theory , which was one of the ways in which symplectic integrators were developed ( see mclachlan and scovel [ 1996 ] and references therein . )
however , we shall not pursue the hamilton - jacobi aspect of the theory here .
[ [ momentum - maps . ] ] momentum maps .
+ + + + + + + + + + + + + + suppose that a lie group @xmath86 , with lie algebra @xmath87 , acts on @xmath3 , and hence on curves in @xmath3 , in such a way that the action @xmath8 is invariant .
clearly , @xmath86 leaves the set of solutions of the variational principle invariant , so the action of @xmath86 restricts to @xmath66 , and the group action commutes with @xmath40 . denoting the infinitesimal generator of @xmath88 on @xmath4 by @xmath89 , we have by ( [ 8 ] ) , @xmath90 for @xmath88 , define @xmath91 by @xmath92 .
then ( [ 9 ] ) says that @xmath93 is an integral of the flow of @xmath35 .
we have arrived at a version of noether s theorem ( rather close to the original derivation of noether ) : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ using the variational principle , noether s theorem results from the infinitesimal invariance of the action restricted to space of solutions of the variational principle . the conserved momentum associated to a lie algebra element
@xmath94 is @xmath95 , where @xmath60 is the lagrange one - form .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [ [ reformulation - in - terms - of - first - variations . ] ] reformulation in terms of first variations .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we have just seen that symplecticity of the flow and noether s theorem result from restricting the action to the space of solutions .
one tacit assumption is that the space of solutions is a manifold in some appropriate sense .
this is a potential problem , since solution spaces for field theories are known to have singularities ( see , e.g. , arms , marsden and moncrief [ 1982 ] ) .
more seriously there is the problem of finding a multisymplectic analogue of the statement that the lagrangian flow map is symplectic , since for multisymplectic field theory one obtains an evolution picture only after splitting spacetime into space and time and adopting the `` function space '' point of view .
having the general formalism depend either on a spacetime split or an analysis of the associated cauchy problem would be contrary to the general thrust of this article .
we now give a formal argument , in the context of lagrangian mechanics , which shows how both these problems can be simultaneously avoided .
given a solution @xmath96 , a * _ _ * first variation at @xmath15 is a vector field @xmath97 on @xmath3 such that @xmath98 is also a solution curve ( i.e. a curve in @xmath66 ) .
we think of the solution space @xmath66 as being a ( possibly ) singular subset of the smooth space of all putative curves @xmath99 in @xmath4 , and the restriction of @xmath97 to @xmath15 as being the derivative of some curve in @xmath66 at @xmath15 .
when @xmath66 is a manifold , a first variation is a vector at @xmath15 tangent to @xmath66 .
temporarily define @xmath100 where by abuse of notation @xmath60 is the one form on @xmath99 defined by @xmath101 then @xmath66 is defined by @xmath102 and we have the equation @xmath103 so if @xmath97 and @xmath104 are first variations at @xmath15 , we obtain @xmath105 we have the identity @xmath106),\ ] ] which we will use to evaluate ( [ 54 ] ) at the curve @xmath15 .
let @xmath107 denote the flow of @xmath97 , define @xmath108 , and make similar definitions with @xmath104 replacing @xmath97 .
for the first term of ( [ 50 ] ) , we have @xmath109 which vanishes , since @xmath110 is zero along @xmath111 for every @xmath112 .
similarly the second term of ( [ 50 ] ) at @xmath15 also vanishes , while the third term vanishes since @xmath113 .
consequently , symplecticity of the lagrangian flow @xmath40 may be written @xmath114 for all first variations @xmath97 and @xmath104 .
this formulation is valid whether or not the solution space is a manifold , and it does not explicitly refer to any temporal notion . similarly , noether s theorem may be written in this way .
summarizing , _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ using the variational principle , the analogue of the evolution is symplectic is the equation @xmath115 restricted to first variations of the space of solutions of the variational principle .
the analogue of noether s theorem is infinitesimal invariance of @xmath116 restricted to first variations of the space of solutions of the variational principle . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the variational route to the differential - geometric formalism has obvious pedagogical advantages .
more than that , however , it systematizes searching for the corresponding formalism in other contexts .
we shall in the next sections show how this works in the context of discrete mechanics , classical field theory and multisymplectic geometry .
the discrete lagrangian formalism in veselov [ 1988 ] , [ 1991 ] fits nicely into our variational framework .
veselov uses @xmath117 for the discrete version of the tangent bundle of a configuration space @xmath3 ; heuristically , given some a priori choice of time interval @xmath118 , a point @xmath119 corresponds to the tangent vector @xmath120 . define a * _ _ * discrete lagrangian to be a smooth map @xmath121 , and the corresponding action to be @xmath122 the variational principle is to extremize @xmath8 for variations holding the endpoints @xmath123 and @xmath124 fixed .
this variational principle determines a `` discrete flow '' @xmath125 by @xmath126 , where @xmath127 is found from the * _ _ * discrete euler - lagrange equations ( del equations ) : @xmath128 in this section we work out the basic differential - geometric objects of this discrete mechanics directly from the variational point of view , consistent with our philosophy in the last section .
a mathematically significant aspect of this theory is how it relates to integrable systems , a point taken up by moser and veselov [ 1991 ] .
we will not explore this aspect in any detail in this paper , although later , we will briefly discuss the reduction process and we shall test an integrator for an integrable pde , the sine - gordon equation . [ [ the - lagrange-1-form . ] ] the lagrange @xmath25-form .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we begin by calculating @xmath116 for variations that do not fix the endpoints : @xmath129 it is the last two terms that arise from the boundary variations ( i.e. these are the ones that are zero if the boundary is fixed ) , and so these are the terms amongst which we expect to find the discrete analogue of the lagrange @xmath25-form
. actually , interpretation of the boundary terms gives the _ two _ @xmath25-forms on @xmath117 @xmath130 and @xmath131 and we regard _ the pair _
@xmath132 as being the analogue of the one form in this situation .
[ [ symplecticity - of - the - flow . ] ] symplecticity of the flow .
+ + + + + + + + + + + + + + + + + + + + + + + + + + we parameterize the solutions of the variational principle by the initial conditions @xmath133 , and restrict @xmath8 to that solution space
. then equation ( [ 12 ] ) becomes @xmath134 we should be able to obtain the symplecticity of @xmath135 by determining what the equation @xmath136 means for the right - hand - side of ( [ 15 ] ) . at first
, this does not appear to work , since @xmath136 gives @xmath137 which apparently says that @xmath135 pulls a certain @xmath0-form back to a different @xmath0-form .
the situation is aided by the observation that , from ( [ 13 ] ) and ( [ 14 ] ) , @xmath138 and consequently , @xmath139 thus , there are _ two _ generally distinct @xmath25-forms , but ( up to sign ) only _ one _ @xmath0-form . if we make the definition
@xmath140 then ( [ 16 ] ) becomes @xmath141 .
equation ( [ 13 ] ) , in coordinates , gives @xmath142 which agrees with the discrete symplectic form found in veselov [ 1988 ] , [ 1991 ] .
[ [ noethers - theorem . ] ] noether s theorem .
+ + + + + + + + + + + + + + + + + + suppose a lie group @xmath86 with lie algebra @xmath87 acts on @xmath3 , and hence diagonally on @xmath117 , and that @xmath11 is @xmath86-invariant .
clearly , @xmath8 is also @xmath86-invariant and @xmath86 sends critical points of @xmath8 to themselves .
thus , the action of @xmath86 restricts to the space of solutions , the map @xmath135 is @xmath86-equivariant , and from ( [ 15 ] ) , @xmath143 for @xmath144 , or equivalently , using the equivariance of @xmath135 , @xmath145 since @xmath11 is @xmath86-invariant , ( [ 20 ] ) gives @xmath146 , which in turn converts ( [ 21 ] ) to the conservation equation @xmath147 defining the discrete momentum to be @xmath148 we see that ( [ 24 ] ) becomes conservation of momentum .
a virtually identical derivation of this discrete noether theorem is found in marsden and wendlant [ 1997 ] .
[ [ reduction . ] ] reduction .
+ + + + + + + + + + as we mentioned above , this formalism lends itself to a discrete version of the theory of lagrangian reduction ( see marsden and scheurle [ 1993a , b ] , holm , marsden and ratiu [ 1998a ] and cendra , marsden and ratiu [ 1998 ] ) .
this theory is not the focus of this article , so we shall defer a brief discussion of it until the conclusions .
[ [ multisymplectic - geometry . ] ] multisymplectic geometry .
+ + + + + + + + + + + + + + + + + + + + + + + + + we now review some aspects of multisymplectic geometry , following gotay , isenberg and marsden [ 1997 ] and marsden and shkoller [ 1997 ] .
we let @xmath149 be a fiber bundle over an oriented manifold @xmath65 .
denote the first jet bundle over @xmath150 by @xmath151 or @xmath152 and identify it with the _ affine _ bundle over @xmath150 whose fiber over @xmath153 consists of aff@xmath154 , those linear mappings @xmath155 satisfying @xmath156 we let @xmath157 and the fiber dimension of @xmath150 be @xmath158 .
coordinates on @xmath65 are denoted @xmath159 , and fiber coordinates on @xmath150 are denoted by @xmath160 .
these induce coordinates @xmath161 on the fibers of @xmath151 .
if @xmath162 is a section of @xmath163 , its tangent map at @xmath164 , denoted @xmath165 , is an element of @xmath166 .
thus , the map @xmath167 defines a section of @xmath151 regarded as a bundle over @xmath65 .
this section is denoted @xmath168 or @xmath169 and is called the first jet of @xmath170 . in coordinates , @xmath168 is given by @xmath171 where @xmath172 .
higher order jet bundles of @xmath150 , @xmath173 , then follow as @xmath174 .
analogous to the tangent map of the projection @xmath175 , @xmath176 , we may define the jet map of this projection which takes @xmath177 onto @xmath151 let @xmath178 so that @xmath179
. then @xmath180 we define the subbundle @xmath181 of @xmath177 over @xmath65 which consists of second - order jets so that on each fiber @xmath182 in coordinates , if @xmath178 is given by @xmath183 , and @xmath184 is given by @xmath185 , then @xmath186 is a second - order jet if @xmath187 .
thus , the second jet of @xmath188 , @xmath189 , given in coordinates by the map @xmath190 , is an example of a second - order jet .
the * _ _ * dual jet bundle @xmath191 is the _ vector _ bundle over @xmath150 whose fiber at @xmath192 is the set of _ affine _ maps from @xmath193 to @xmath194 , the bundle of @xmath195-forms on @xmath65 .
a smooth section of @xmath191 is therefore an affine bundle map of @xmath151 to @xmath196 covering @xmath163 .
fiber coordinates on @xmath191 are @xmath197 , which correspond to the affine map given in coordinates by @xmath198 where @xmath199 .
analogous to the canonical one- and two - forms on a cotangent bundle , there exist canonical ( @xmath200)- and ( @xmath201)-forms on the dual jet bundle @xmath191 . in coordinates , with @xmath202 ,
these forms are given by @xmath203 a lagrangian density @xmath204 is a smooth bundle map over @xmath65 . in coordinates , we write @xmath205 the corresponding covariant legendre transformation for @xmath206 is a fiber preserving map over @xmath150 , @xmath207 , expressed intrinsically as the first order vertical taylor approximation to @xmath206 : @xmath208 where @xmath209 .
a straightforward calculation shows that the covariant legendre transformation is given in coordinates by @xmath210 we can then define the * _ _ * cartan form as the @xmath195-form @xmath211 on @xmath151 given by @xmath212 and the @xmath213-form @xmath214 by @xmath215 with local coordinate expressions @xmath216
{ \displaystyle \omega_\mathcal{l } = dy^a \wedge d\left ( \frac{\partial l } { \partial { v^a}_\mu } \right ) \wedge d^nx_\mu - d\left [ l - \frac{\partial l}{\partial { v^a}_\mu } { v^a}_\mu \right ] \wedge d^{n+1}x . }
\end{array } \label{d1a}\ ] ] this is the differential - geometric formulation of the multisymplectic structure .
subsequently , we shall show how we may obtain this structure directly from the variational principle , staying entirely on the lagrangian side @xmath151 .
[ [ the - multisymplectic - form - formula . ] ] the multisymplectic form formula .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in this subsection we prove a formula that is the multisymplectic counterpart to the fact that in finite - dimensional mechanics , the flow of a mechanical system consists of symplectic maps .
again , we do this by studying the action function .
let @xmath217 be a smooth manifold with ( piecewise ) smooth closed boundary .
we define the set of smooth maps @xmath218 for each @xmath219 , we set @xmath220 and @xmath221 so that @xmath222 is a diffeomorphism . @xmath223 .
we may then define the infinite - dimensional manifold @xmath99 to be the closure of @xmath224 in either a hilbert space or banach space topology .
for example , the manifold @xmath99 may be given the topology of a hilbert manifold of bundle mappings , @xmath225 , ( @xmath217 considered a bundle with fiber a point ) for any integer @xmath226 , so that the hilbert sections @xmath227 in @xmath150 are those whose distributional derivatives up to order @xmath186 are square - integrable in any chart . with our condition on @xmath186 ,
the sobolev embedding theorem makes such mappings well defined .
alternately , one may wish to consider the banach manifold @xmath99 as the closure of @xmath224 in the usual @xmath53-norm , or more generally , in a holder space @xmath228-norm .
see palais [ 1968 ] and ebin and marsden [ 1970 ] for a detailed account of manifolds of mappings .
the choice of topology for @xmath99 will not play a crucial role in this paper .
let @xmath229 be the lie group of @xmath163-bundle automorphisms @xmath230 covering diffeomorphisms @xmath231 , with lie algebra @xmath87 .
we define the * _ _ * action @xmath232 by @xmath233 furthermore , if @xmath234 , then @xmath235 .
the * _ _ * tangent space to the manifold @xmath99 at a point @xmath170 is the set @xmath236 defined by @xmath237 of course , when these objects are topologized as we have described , the definition of the tangent space becomes a theorem , but as we have mentioned , this functional analytic aspect plays a minor role in what follows .
given vectors @xmath238 we may extend them to vector fields @xmath239 on @xmath99 by fixing vector fields @xmath240 such that @xmath241 and @xmath242 , and letting @xmath243 and @xmath244 .
thus , the flow of @xmath245 on @xmath99 is given by @xmath246 where @xmath247 covering @xmath248 is the flow of @xmath249 .
the definition of the bracket of vector fields using their flows , then shows that @xmath250(\rho ) = [ v , w ] \circ ( \rho \circ \rho_x^{-1 } ) .\ ] ] whenever it is contextually clear , we shall , for convenience , write @xmath97 for @xmath251 .
the * _ _ * action function @xmath252 on @xmath99 is defined as follows : @xmath253 let @xmath254 be an arbitrary smooth path in @xmath229 such that @xmath255 , and let @xmath256 be given by @xmath257 we say that @xmath170 is a * _ _ * stationary point , critical point , or extremum of @xmath252 if @xmath258 then , @xmath259 where we have used the fact that @xmath260 for all holonomic sections @xmath261 of @xmath151 ( see corollary [ cor4.2 ] below ) , and that @xmath262 using the cartan formula , we obtain that @xmath263 \nonumber \\ & & \qquad + \int_{\partial u_x } j^1({\phi \circ \phi_x^{-1}})^ * [ { j^1(v ) } \intprod \theta_\mathcal{l } ] . \label{s6}\end{aligned}\ ] ] hence , a necessary condition for @xmath264 to be an extremum of @xmath252 is that the first term in ( [ s6 ] ) vanish .
one may readily verify that the integrand of the first term in ( [ s6 ] ) is equal to zero whenever @xmath265 is replaced by @xmath266 which is either @xmath175-vertical or tangent to @xmath267 ( see marsden and shkoller [ 1997 ] ) , so that using a standard argument from the calculus of variations , @xmath268 $ ] must vanish for all vectors @xmath104 on @xmath151 in order for @xmath170 to be an extremum of the action .
we shall call such elements @xmath264 covering @xmath269 , solutions of the euler - lagrange equations .
we let @xmath270=0 \ \ ; \mbox{for all } \ ; \
w\in tj^1(y)\}. \label{s8}\ ] ] in coordinates , @xmath271 is an element of @xmath272 if @xmath273 we are now ready to prove the multisymplectic form formula , a generalization of the symplectic flow theorem , but we first make the following remark
. if @xmath272 is a submanifold of @xmath99 , then for any @xmath274 , we may identify @xmath275 with the set @xmath276=0 \ \ ; \mbox{for all } \ ; \ w \in tj^1(y ) \}$ ] since such vectors arise by differentiating @xmath277 = 0 $ ] , where @xmath278 is a smooth curve of solutions of the euler - lagrange equations in @xmath272 ( when such solutions exist ) .
more generally , we do not require @xmath272 to be a submanifold in order to define the first variation solution of the euler - lagrange equations . for any @xmath279
, we define the set @xmath280=0 \ \ ; \mbox{for all } \ ; \ w \in tj^1(y)\}. \label{s10}\ ] ] elements of @xmath281 solve the first variation equations of the euler - lagrange equations .
@xmath223 [ thm1 ] if @xmath279 , then for all @xmath97 and @xmath104 in @xmath281 , @xmath282=0 .
\label{s11}\ ] ] [ [ proof . ] ] proof .
+ + + + + + we define the @xmath25-forms @xmath283 and @xmath284 on @xmath99 by @xmath285\ ] ] and @xmath286,\ ] ] so that by ( [ s6 ] ) , @xmath287 recall that for any @xmath25-form @xmath110 on @xmath99 and vector fields @xmath288 on @xmath99 , @xmath289 - w[\alpha(v ) ] - \alpha([v , w ] ) .
\label{dalpha}\ ] ] we let @xmath290 be a curve in @xmath99 through @xmath170 , where @xmath291 is a curve in @xmath229 through the identity such that @xmath292 and consider equation ( [ s13 ] ) restricted to all @xmath293 .
thus , @xmath294\\ & = & \int_{\partial u_x } { j^1}({\phi \circ \phi_x^{-1}})^ * { \mathfrak l}_{j^1(w)}[{j^1(v ) } \intprod \theta_\mathcal{l } ] \\ & = & -\int_{\partial u_x } { j^1}({\phi \circ \phi_x^{-1}})^ * [ { j^1(w ) } \intprod d({j^1(v ) } \intprod \theta_\mathcal{l } ) ] \\ & & \qquad\mbox+ \int_{\partial u_x } { j^1}({\phi \circ \phi_x^{-1}})^ * d [ { j^1(w ) } \intprod { j^1(v ) } \intprod \theta_\mathcal{l}],\end{aligned}\ ] ] where the last equality was obtained using cartan s formula . using stoke s theorem , noting that @xmath295 is empty , and applying cartan s formula once again , we obtain that @xmath296 \\ & & \qquad - \int_{\partial u_x } { j^1}({\phi \circ \phi_x^{-1}})^ * [ { j^1(w ) } \intprod { \mathfrak l}_{j^1(v)}\theta_\mathcal{l}],\end{aligned}\ ] ] and @xmath297 \\ & & \qquad - \int_{\partial u_x } { j^1}({\phi \circ \phi_x^{-1}})^ * [ { j^1(v ) } \intprod { \mathfrak l}_{j^1(w)}\theta_\mathcal{l}].\end{aligned}\ ] ] also , since @xmath298 = { j^1}([v , w])$ ] , we have @xmath299 ) = \int_{\partial u_x } { j^1}({\phi \circ \phi_x^{-1}})^*[{j^1(v)},{j^1(w ) } ] \intprod \theta_\mathcal{l}.\ ] ] now @xmath300\intprod \theta_\mathcal{l } = { \mathfrak l}_{j^1(v ) } ( { j^1(w)}\intprod \theta_\mathcal{l } ) - { j^1(w ) } \intprod { \mathfrak l}_{j^1(v ) } \theta_\mathcal{l},\ ] ] so that @xmath301 \\ & & + \int_{\partial u_x } { j^1}({\phi \circ \phi_x^{-1}})^ * [ { j^1(v)}\intprod { \mathfrak l}_{j^1(w ) } \theta_\mathcal{l } - { \mathfrak l}_{j^1(v)}({j^1(w)}\intprod \theta_\mathcal{l})].\end{aligned}\ ] ] but @xmath302 and @xmath303 hence , @xmath304 } \\ & = & \int_{\partial u_x } { j^1}({\phi \circ \phi_x^{-1}})^ * ( { j^1(v)}\intprod { j^1(w)}\intprod \omega_\mathcal{l } ) \\ & & - \int_{\partial u_x } { j^1}({\phi \circ \phi_x^{-1}})^ * d ( { j^1(v)}\intprod { j^1(w)}\intprod \theta_\mathcal{l}).\end{aligned}\ ] ] the last term once again vanishes by stokes theorem together with the fact that @xmath295 is empty , and we obtain that @xmath305 we now use ( [ dalpha ] ) on @xmath283 . a similar computation as above yields @xmath306 \nonumber\ ] ] which vanishes for all @xmath279 and @xmath307 .
similarly , @xmath308 for all @xmath279 and @xmath293 . finally , @xmath309 for all @xmath279 .
hence , since @xmath310 we obtain the formula ( [ s11 ] ) .
@xmath311 [ [ symplecticity - revisited . ] ] symplecticity revisited .
+ + + + + + + + + + + + + + + + + + + + + + + + let @xmath312 be a compact oriented connected boundaryless @xmath2-manifold which we think of as our reference cauchy surface , and consider the space of embeddings of @xmath312 into @xmath65 , emb@xmath313 ; again , although it is unnecessary for this paper , we may topologize emb@xmath314 by completing the space in the appropriate @xmath53 or @xmath315-norm .
let @xmath316 be an @xmath317-dimensional manifold .
for any fiber bundle @xmath318 , we shall , in addition to @xmath319 , use the corresponding script letter @xmath320 to denote the space of sections of @xmath321 .
the space of sections of a fiber bundle is an infinite - dimensional manifold ; in fact , it can be precisely defined and topologized as the manifold @xmath99 of the previous section , where the diffeomorphisms on the base manifold are taken to be the identity map , so that the tangent space to @xmath320 at @xmath322 is given simply by @xmath323 where @xmath324 denotes the vertical tangent bundle of @xmath325 .
we let @xmath326 be the vector bundle over @xmath325 whose fiber at @xmath327 , @xmath328 , is the set of linear mappings from @xmath329 to @xmath330 .
then the cotangent space to @xmath331 at @xmath322 is defined as @xmath332 integration provides the natural pairing of @xmath333 with @xmath334 : @xmath335 in practice , the manifold @xmath316 will either be @xmath65 or some ( @xmath200)-dimensional subset of @xmath65 , or the @xmath2-dimensional manifold @xmath336 , where for each @xmath337 , @xmath338 .
we shall use the notation @xmath339 for the bundle @xmath340 , and @xmath341 for sections of this bundle . for the remainder of this section
, we shall set the manifold @xmath99 introduced earlier to @xmath342 .
the infinite - dimensional manifold @xmath343 is called the * _ _ * @xmath344-configuration space , its tangent bundle is called the * _ _ * @xmath344-tangent space , and its cotangent bundle @xmath345 is called the * _ _ * @xmath344-phase space . just as we described in section [ mechanics ] , the cotangent bundle has a canonical @xmath25-form @xmath346 and a canonical @xmath0-form @xmath347 .
these differential forms are given by @xmath348 where @xmath349 , @xmath350 , and @xmath351 is the cotangent bundle projection map .
an infinitesimal slicing of the bundle @xmath163 consists of @xmath339 together with a vector field @xmath352 which is everywhere transverse to @xmath339 , and covers @xmath353 which is everywhere transverse to @xmath354 .
the existence of an infinitesimal slicing allows us to invariantly decompose the temporal from the spatial derivatives of the fields .
let @xmath355 , @xmath356 , and let @xmath357 be the inclusion map .
then we may define the map @xmath358 taking @xmath359 to @xmath360 over @xmath343 by @xmath361 in our notation , @xmath359 is the collection of restrictions of holonomic sections of @xmath151 to @xmath354 , while @xmath362 are the holonomic sections of @xmath363 .
it is easy to see that @xmath358 is an isomorphism ; it then follows that @xmath358 is an isomorphism of @xmath364 with @xmath365 , since @xmath366 is completely determined by @xmath367 .
this bundle map is called the jet decomposition map , and its inverse is called the jet reconstruction map . using this map , we can define the instantaneous lagrangian .
the instantaneous lagrangian @xmath368 is given by @xmath369 \label{t3}\ ] ] for all @xmath370 .
@xmath223 the instantaneous lagrangian @xmath371 has an instantaneous legendre transform @xmath372 which is defined in the usual way by vertical fiber differentiation of @xmath371 ( see , for example , abraham and marsden [ 1978 ] ) . using the instantaneous legendre transformation
, we can pull - back the canonical @xmath25- and @xmath0-forms on @xmath345 . denote , respectively , the instantaneous lagrange @xmath25- and @xmath0-forms on @xmath365 by @xmath373 alternatively , we may define @xmath374 using theorem [ thm2.1 ] , in which case no reference to the cotangent bundle is necessary .
we will show that our covariant multisymplectic form formula can be used to recover the fact that the flow of the euler - lagrange equations in the bundle @xmath375 is symplectic with respect to @xmath376 .
to do so , we must relate the multisymplectic cartan ( @xmath201)-form @xmath377 on @xmath151 with the symplectic @xmath0-form @xmath376 on @xmath365 .
[ thm3.2 ] let @xmath378 be the canonical @xmath25-form on @xmath364 given by @xmath379,\ ] ] where @xmath380 , @xmath381 .
( a ) if the @xmath0-form @xmath382 on @xmath359 is defined by @xmath383 , then for @xmath384 , @xmath385.\ ] ] ( b)let the diffeomorphism @xmath386 be a slicing of @xmath65 such that for @xmath387 , @xmath388 where @xmath389 is given by @xmath390 . for any @xmath279 ,
let @xmath391 so that for each @xmath392 , @xmath393 , and let @xmath394 . then @xmath395 [ [ proof.-1 ] ] proof .
+ + + + + + part ( a ) follows from the cartan formula together with stokes theorem using an argument like that in the proof of theorem [ thm1 ] . for part ( b ) , we recall that the multisymplectic form formula on @xmath396 states that for any subset @xmath397 with smooth closed boundary and vectors @xmath398 , @xmath399 , @xmath400=0.\ ] ] let @xmath401 } \sigma_\lambda.\ ] ] then @xmath402 , so that ( [ t8 ] ) can be written as @xmath403 \\ & & \qquad - \int_{\sigma_{\lambda_1 } } { j^1}(\phi \circ i_{\tau_{\lambda_1}})^ * [ { j^1v_{\tau_{\lambda_1 } } } \intprod { j^1w_{\tau_{\lambda_1 } } } \intprod \omega_\mathcal{l } ] \\ & = & \omega_{\tau_{\lambda_1}}^\mathcal{l}(j^1v_{\tau_{\lambda_1}},j^1w_{\tau_{\lambda_1}})- \omega_{\tau_{\lambda_2}}^\mathcal{l}(j^1v_{\tau_{\lambda_2}},j^1w_{\tau_{\lambda_2}}),\end{aligned}\ ] ] which proves ( [ t7 ] ) .
@xmath311 [ thm3.3 ] the identity @xmath404 holds .
[ [ proof.-2 ] ] proof .
+ + + + + + let @xmath405 , which we identify with @xmath406 , where @xmath407 is a @xmath408-vertical vector .
choose a coordinate chart which is adapted to the slicing so that @xmath409 .
with @xmath410 , we see that @xmath411 now , from ( [ t3 ] ) we get @xmath412\\ & = & \int_{\sigma_\tau } \frac{\partial l}{\partial { v^a}_0 } ( \phi^b , { \phi^b}_{,\mu})dy^a\otimes d^nx_0,\end{aligned}\ ] ] where we arrived at the last equality using the fact that @xmath413 in this adapted chart . since @xmath414 , we see that @xmath415 , and this completes the proof .
let the instantaneous energy @xmath416 associated with @xmath371 be given by @xmath417 and define the `` time''-dependent lagrangian vector field @xmath418 by @xmath419 since @xmath420 over emb@xmath313 is infinite - dimensional and @xmath421 is only weakly nondegenerate , the second - order vector field @xmath418 does not , in general , exist .
in the case that it does , we obtain the following result .
assume @xmath418 exists and let @xmath422 be its semiflow , defined on some subset @xmath423 of the bundle @xmath424 over emb@xmath313 .
fix @xmath425 so that @xmath426 where @xmath427 and @xmath428 . then @xmath429 .
[ [ proof.-3 ] ] proof .
+ + + + + + this follows immediately from theorem [ thm3.2](b ) and theorem [ thm3.3 ] and the fact that @xmath358 induces an isomorphism between @xmath359 and @xmath365 .
@xmath311 [ [ example - nonlinear - wave - equation .
] ] example : nonlinear wave equation .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + to illustrate the geometry that we have developed , let us consider the scalar nonlinear wave equation given by @xmath430 where @xmath431 is the laplace - beltrami operator and @xmath158 is a real - valued @xmath432 function of one variable . for concreteness ,
fix @xmath2@xmath433@xmath25 so that the spacetime manifold @xmath434 , the configuration bundle @xmath435 , and the first jet bundle @xmath436 . equation ( [ t11 ] ) is governed by the lagrangian density @xmath437 + n(\phi ) \right\ } dx^1 \wedge dx^0 .\ ] ] using coordinates @xmath438 for @xmath151 , we write the multisymplectic @xmath439-form for this nonlinear wave equation on @xmath440 in coordinates as @xmath441 a short computation verifies that solutions of ( [ t11 ] ) are elements of @xmath272 , or that @xmath442 = 0 $ ] for all @xmath443 ( see marsden and shkoller [ 1997 ] ) .
we will use this example to demonstrate that our multisymplectic form formula generalizes the notion of symplecticity given by bridges [ 1997 ] .
since the lagrangian ( [ lag ] ) does not explicitly depend on time , it is convenient to identify sections of @xmath150 as mappings from @xmath440 into @xmath444 , and similarly , sections of @xmath151 as mappings from @xmath440 into @xmath445 . thus , for @xmath446 , @xmath447 , and if we set @xmath448 , then ( [ t11 ] ) can be reformulated to @xmath449 \left[\begin{array}{cccccc } \phi \\ p^0 \\
p^1 \\ \end{array } \right ] _ { , 0 } + \left[\begin{array}{cccccc } 0 & 0 & -1\\ 0 & 0 & 0\\ 1 & 0 & 0\\ \end{array } \right ] \left[\begin{array}{cccccc } \phi \\
p^1 \\ \end{array } \right ] _ { , 1 } = \left[\begin{array}{cccccc } n'(\phi ) \\
-p^0 \\ p^1 \\ \end{array } \right ] .
\qquad \qquad \label{t12a}\end{aligned}\ ] ] to each degenerate matrix @xmath450 , we associate the contact form @xmath451 on @xmath445 given by @xmath452 , where @xmath453 and @xmath454 is the standard inner product on @xmath445 .
bridges obtains the following conservation of symplecticity : @xmath455 + \frac{\partial } { \partial x^1 } \left [ \omega^1 ( j^1(\phi_{,0}),j^1(\phi_{,1 } ) ) \right ] = 0.\ ] ] this result is interesting , but has somewhat limited scope in that the vector fields in ( [ t14 ] ) upon which the contact forms act are not general solutions to the first variation equations ; rather , they are the specific first variation solutions @xmath456 .
bridges obtains this result by crucially relying on the multi - hamiltonian structure of ( [ t11 ] ) ; in particular , the vector @xmath457 on the right - hand - side of ( [ t12a ] ) is the gradient of a smooth multi - hamiltonian function @xmath458 ( although the multi - hamiltonian formalism is not important for this article , we refer the reader to marsden and shkoller [ 1997 ] for the hamiltonian version of our covariant framework , and to bridges [ 1997 ] ) . using equation ( [ t12a ] ) ,
it is clear that @xmath459 so that ( [ t14 ] ) follows from the relation @xmath460 .
[ prop1 ] the multisymplectic form formula is an intrinsic generalization of the conservation law ( [ t14 ] ) ; namely , for any @xmath461 that are @xmath408-vertical , @xmath462 + \frac{\partial } { \partial x^1 } \left [ \omega^1 ( j^1(v),j^1(w ) ) \right ] = 0.\ ] ] [ [ proof.-4 ] ] proof .
+ + + + + + let @xmath265 and @xmath463 have the coordinate expressions @xmath464 and @xmath465 , respectively . using ( [ t12 ] ) , we compute @xmath466 so that with theorem [ thm1 ] and the definition of @xmath451 , we have , for @xmath397 , @xmath467 and hence by green s theorem , @xmath468 + \frac{\partial } { \partial x^1 } \left [ \omega^1 ( j^1(v),j^1(w ) ) \right ] \right\}dx^1 \wedge dx^0 = 0.\ ] ] since @xmath469 is arbitrary , we obtain the desired result .
@xmath311 in general , when @xmath97 is @xmath163-vertical , @xmath470 has the coordinate expression @xmath471 , but for the special case that @xmath472 , @xmath473 , and proposition [ prop1 ] gives @xmath474 - \frac{\partial } { \partial x^1 } \left [ \phi_{0 } \phi_{,1,1 } - \phi_{1 } \phi_{,0,1 } \right ] = 0,\ ] ] which simplifies to the trivial statement that @xmath475 [ [ the - variational - route - to - the - cartan - form . ] ] the variational route to the cartan form .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we may alternatively define the cartan form by beginning with equation ( [ s5 ] ) . using the infinitesimal generators defined in ( [ ig ] )
, we obtain that @xmath476 . \label{s18}\end{aligned}\ ] ] using the natural splitting of @xmath477 , any vector @xmath478 may decomposed as @xmath479 where we recall that @xmath480 .
[ lemma1 ] for any @xmath481 , @xmath482 , \label{s19}\ ] ] and @xmath483 [ [ proof.-5 ] ] proof .
+ + + + + + the equality ( [ s20 ] ) is obvious , since the second term in ( [ s18 ] ) clearly vanishes for all vertical vectors . for vectors @xmath484 , the first term in ( [ s18 ] ) vanishes ; indeed , using the chain rule , we need only compute that @xmath485 which is zero by ( [ s18a ] ) .
we then apply the cartan formula to the second term in ( [ s18 ] ) and note that @xmath486 is an ( @xmath201)-form on the ( @xmath200)-dimensional manifold @xmath469 so that we obtain ( [ s19 ] ) .
@xmath311 [ thm_cartan ] given a smooth lagrangian density @xmath487 , there exist a unique smooth section @xmath488 and a unique differential form @xmath489 such that for any @xmath481 , and any open subset @xmath469 such that @xmath490 , @xmath491 .
\label{s21}\ ] ] furthermore , @xmath492 \text { in } u_x .
\label{s21b}\ ] ] in coordinates , the action of the euler - lagrange derivative @xmath493 on @xmath181 is given by @xmath494 dy^a \wedge d^{n+1}x , \label{s21c}\end{aligned}\ ] ] while the form @xmath495 matches the definition of the cartan form given in ( [ d1a ] ) and has the coordinate expression @xmath496 [ [ proof.-6 ] ] proof .
+ + + + + + choose @xmath497 small enough so that it is contained in a coordinate chart , say @xmath498 .
in these coordinates , let @xmath499 so that along @xmath227 , our decomposition ( [ s18a ] ) may be written as @xmath500 and equation ( [ s20 ] ) gives @xmath501 d^{n+1}x , \label{s70}\ ] ] where we have used the fact that in coordinates along @xmath267 , @xmath502.\ ] ] integrating ( [ s70 ] ) by parts , we obtain @xmath503 \cdot v^a \right\ } d^{n+1}x \nonumber \\ & & \qquad + \int_{\partial u_x}\left\ { \frac{\partial l}{\partial { v^a}_\mu}(j^1({\phi \circ \phi_x^{-1 } } ) ) \cdot v^a d^nx_\mu \right .
\nonumber \\ & & \qquad \qquad + \left .
\frac{\partial l}{\partial { v^a}_\mu}(j^1({\phi \circ \phi_x^{-1 } } ) ) \frac{\partial ( { \phi \circ \phi_x^{-1}})^a}{\partial x^\nu } \cdot v^\nu d^{n}x_\mu \right\ } .
\label{s71}\end{aligned}\ ] ] let @xmath110 be the @xmath2-form integrand of the boundary integral in ( [ s71 ] ) ; then @xmath504 since @xmath110 is invariant under this lift .
additionally , from equation ( [ s19 ] ) , we obtain the horizontal contribution @xmath505 so combining equations ( [ s71 ] ) and ( [ s72 ] ) , a simply computation verifies that @xmath506 d^{n+1}x \otimes dy^a \right\ } \cdot v \nonumber \\ & & \qquad + \int_{\partial { j^1}(\phi\circ \phi_x^{-1})(u_x)}v \intprod \left\ { \frac{\partial l}{\partial { v^a}_\mu}(j^1({\phi \circ \phi_x^{-1 } } ) ) dy^a \wedge d^nx_\mu \right .
\nonumber \\ & & \qquad \qquad \qquad \left .
+ \left [ l - \frac{\partial l}{\partial { v^a}_\mu}(j^1({\phi \circ \phi_x^{-1 } } ) ) \frac{\partial ( { \phi \circ \phi_x^{-1}})^a}{\partial x^\mu } \right ] d^{n+1}x \right\}. \label{s73}\end{aligned}\ ] ] the vector @xmath97 in the second term of ( [ s73 ] ) may be replaced by @xmath265 since @xmath175-vertical vectors are clearly in the kernel of the form that @xmath97 is acting on .
this shows that ( [ s21c ] ) and ( [ s21d ] ) hold , and hence that the boundary integral in ( [ s73 ] ) may be written as @xmath507 .
\nonumber\ ] ] now , if we choose another coordinate chart @xmath508 , the coordinate expressions of @xmath509 and @xmath495 must agree on the overlap @xmath510 since the left - hand - side of ( [ s21 ] ) is intrinsically defined .
thus , we have uniquely defined @xmath493 and @xmath495 for any @xmath469 such that @xmath511 . finally , ( [ s21b ] ) holds , since @xmath512 is also intrinsically defined and both sides of the equation yield the same coordinate representation , the euler - lagrange equations in @xmath469 . @xmath311 [ [ remark ] ] remark + + + + + + to prove theorem [ thm_cartan ] for the case @xmath513 , we must modify the proof to take into account the boundary conditions which are prescribed on @xmath514 .
[ cor4.2 ] the ( @xmath200)-form @xmath495 defined by the variational principle satisfies the relationship @xmath515 for all holonomic sections @xmath516 .
[ [ proof.-7 ] ] proof .
+ + + + + + this follows immediately by substituting ( [ s21b ] ) into ( [ s21 ] ) and integrating by parts using cartan s formula .
@xmath311 [ [ remark-1 ] ] remark + + + + + + we have thus far focused on holonomic sections of @xmath151 , those that are the first jets of sections of @xmath150 , and correspondingly , we have restricted the general splitting of @xmath477 given by @xmath517 to @xmath518 , @xmath519 as we specified in ( [ s18a ] ) . for general sections @xmath520 ,
the horizontal bundle is given by image @xmath521 , and the frobenius theorem guarantees that @xmath521 is locally holonomic if the connection is flat , or equivalently if the curvature of the connection @xmath522 vanishes .
since this is a local statement , we may assume that @xmath523 , where @xmath524 is open , and that @xmath163 is simply the projection onto the first factor .
for @xmath525 , and @xmath526 , @xmath527 is a linear operator which is holonomic if @xmath528 , where @xmath529 is the differential of @xmath170 , and this is the case whenever the operator @xmath530 is symmetric .
equivalently , the operator @xmath531 is symmetric for all @xmath532 .
one may easily verify that the local curvature is given by @xmath533 and that @xmath534 locally for some @xmath525 , if and only if @xmath535 . [ [ the - variational - route - to - noethers - theorem . ] ] the variational route to noether s theorem .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + suppose the lie group @xmath229 acts on @xmath99 and leaves the action @xmath252 invariant so that @xmath536 this implies that for each @xmath537 , @xmath538 whenever @xmath279 .
we restrict the action of @xmath229 to @xmath272 , and let @xmath539 be the corresponding infinitesimal generator on @xmath99 restricted to points in @xmath272 ; then @xmath540 \\ & = & \int_{u_x } j^1({\phi \circ \phi_x^{-1}})^ * [ j^1(\xi ) \intprod \omega_\mathcal{l}],\end{aligned}\ ] ] since @xmath541 by ( [ s22a ] ) and corollary [ cor4.2 ] .
we denote the covariant momentum map on @xmath151 by @xmath542 which we define as @xmath543 using ( [ s22 ] ) , we find that @xmath544 = 0 $ ] , and since this must hold for all infinitesimal generators @xmath539 at @xmath545 , the integrand must also vanish so that @xmath546 = 0 , \label{s23}\ ] ] which is precisely a restatement of the covariant noether theorem .
we now generalize the veselov discretization given in section ( 3 ) to multisymplectic field theory , by discretizing the spacetime @xmath65 .
for simplicity we restrict to the discrete analogue of @xmath547 ; i.e. @xmath548 .
thus , we take @xmath549 and the fiber bundle @xmath150 to be @xmath550 for some smooth manifold @xmath551 . [ [ notation . ] ] notation .
+ + + + + + + + + the development in this section is aided by a small amount of notation and terminology .
elements of @xmath150 over the base point @xmath552 are written as @xmath553 and the projection @xmath163 acts on @xmath150 by @xmath554 .
the fiber over @xmath555 is denoted @xmath556 .
a * _ _ * triangle @xmath557 of @xmath65 is an ordered triple of the form @xmath558 the first component @xmath552 of @xmath557 is the * _ _ * first vertex of the triangle , denoted @xmath559 , and similarly for the * _ _ * second and * _ _ * third vertices .
the set of all triangles in @xmath65 is denoted @xmath560 . by abuse of notation the same symbol
is used for a triangle and the ( unordered ) set of its vertices .
a point @xmath555 is * _ _ * touched by a triangle if it is a vertex of that triangle .
if @xmath561 , then @xmath562 is an * _ _ * interior point of @xmath217 if @xmath217 contains all three triangles of @xmath65 that touch @xmath552 .
the * _ _ * interior @xmath563 of @xmath217 is the collection of the interior points of @xmath217 .
the * _ _ * closure @xmath564 of @xmath217 is the union of all triangles touching interior points of @xmath217 .
a * _ _ * boundary point of @xmath217 is a point in @xmath217 and @xmath564 which is not an interior point .
the * _ _ * boundary of @xmath217 is the set of boundary points of @xmath217 , so that @xmath565 generally , @xmath217 properly contains the union of its interior and boundary , and we call @xmath217 * _ _ * regular if it is exactly that union .
a * _ _ * section of @xmath150 is a map @xmath566 such that @xmath567 .
= 0.9 [ [ multisymplectic - phase - space . ] ] multisymplectic phase space .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + we define the * _ _ * first jet bundle , but may not lead to a good numerical scheme .
later , we shall also use four vertices together with averaging to define the partial derivatives of the fields .
] of @xmath150 to be @xmath568 heuristically ( see figure ( [ 25 ] ) ) , @xmath65 corresponds to some grid of elements @xmath569 in continuous spacetime , say @xmath570 , and @xmath571 corresponds to @xmath572 , where @xmath573 is `` inside '' the triangle bounded by @xmath574 , and @xmath170 is some smooth section of @xmath575 interpolating the field values @xmath576 .
the * _ _ * first jet extension of a section @xmath170 of @xmath150 is the map @xmath577 defined by @xmath578 given a vector field @xmath579 on @xmath150 , we denote its restriction to the fiber @xmath556 by @xmath580 , and similarly for vector fields on @xmath152 .
the * _ _ * first jet extension of a vector field @xmath579 on @xmath150 is the vector field @xmath581 on @xmath152 defined by @xmath582 for any triangle @xmath557 .
= 0.5 [ [ the - variational - principle . ] ] the variational principle .
+ + + + + + + + + + + + + + + + + + + + + + + + + + let us posit a * _ _ * discrete lagrangian @xmath583 . given a triangle @xmath557 , define the function @xmath584 by @xmath585 so that we may view the lagrangian @xmath11 as being a choice of a function @xmath586 for each triangle @xmath557 of @xmath65 .
the variables on the domain of @xmath586 will be labeled @xmath587 , irrespective of the particular @xmath557 .
let @xmath217 be regular and let @xmath588 be the set of sections of @xmath150 on @xmath217 , so @xmath588 is the manifold @xmath589 .
the * _ _ * action will assign real numbers to sections in @xmath588 by the rule @xmath590 given @xmath591 and a vector field @xmath97 , there is the 1-parameter family of sections @xmath592 where @xmath593 denotes the flow of @xmath594 on @xmath551 .
the * _ _ * variational principle is to seek those @xmath170 for which @xmath595 for all vector fields @xmath97 .
[ [ the - discrete - euler - lagrange - equations . ] ] the discrete euler - lagrange equations .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the variational principle gives certain field equations , the * _ _ * discrete euler - lagrange field equations ( delf equations ) , as follows .
focus upon some @xmath596 , and abuse notation by writing @xmath597 .
the action , written with its summands containing @xmath553 explicitly , is ( see figure ( [ 26 ] ) ) @xmath598 so by differentiating in @xmath553 , the delf equations are @xmath599 for all @xmath596 .
equivalently , these equations may be written @xmath600 for all @xmath596 .
[ [ the - discrete - cartan - form . ] ] the discrete cartan form .
+ + + + + + + + + + + + + + + + + + + + + + + + + now suppose we allow nonzero variations on the boundary @xmath601 , so we consider the effect on @xmath8 of a vector field @xmath97 which does not necessarily vanish on @xmath601 .
for each @xmath602 find the triangles in @xmath217 touching @xmath552 .
there is at least one such triangle since @xmath603 ; there are not three such triangles since @xmath604 . for each such triangle @xmath557
, @xmath552 occurs as the @xmath605 vertex , for one or two of @xmath606 , and those @xmath605 expressions from the list @xmath607 yielding one or two numbers .
the contribution to @xmath116 from the boundary is the sum of all such numbers . to bring this into a recognizable format ,
we take our cue from discrete lagrangian mechanics , which featured _ two _ @xmath25-forms . here the above list suggests the _ three _
@xmath25-forms on @xmath152 , the first of which we define to be @xmath608 @xmath609 and @xmath610 being defined analogously . with these notations
, the contribution to @xmath116 from the boundary can be written @xmath611 , where @xmath60 is the @xmath25-form on the space of sections @xmath588 defined by @xmath612(\delta)\right).\ ] ] in comparing ( [ 31 ] ) with ( [ s21 ] ) , the analogy with the multisymplectic formalism of section ( 4 ) is immediate .
[ [ the - discrete - multisymplectic - form - formula . ] ] the discrete multisymplectic form formula .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + given a triangle @xmath557 in @xmath65 , we define the projection @xmath613 by @xmath614 in this notation , it is easily verified that ( [ 31 ] ) takes the convenient form @xmath615 a * _ _ * first - variation at a solution @xmath170 of the delf equations is a vertical vector field @xmath97 such that the associated flow @xmath616 maps @xmath170 to other solutions of the delf equations . set @xmath617 . since @xmath618 one obtains @xmath619 so that only two of the three @xmath0-forms @xmath620 , @xmath606 are essentially distinct .
exactly as in section ( 2 ) , the equation @xmath621 , when specialized to two first - variations @xmath97 and @xmath104 now gives , by taking one exterior derivative of ( [ 43 ] ) , @xmath622 which in turn is equivalent to @xmath623(\delta)\right)=0.\ ] ] again , the analogy with the multisymplectic form formula for continuous spacetime ( [ s11 ] ) is immediate . [ [ the - discrete - noether - theorem . ] ] the discrete noether theorem .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + suppose that a lie group @xmath86 with lie algera @xmath624 acts on @xmath135 by vertical symmetries in such a way that the lagrangian @xmath11 is @xmath86-invariant .
then @xmath86 acts on @xmath150 and @xmath152 in the obvious ways . since there are three lagrange @xmath25-forms , there are three momentum maps @xmath625 , @xmath606 , each one a @xmath626-valued function on triangles in @xmath65 , and defined by @xmath627 for any @xmath88 .
invariance of @xmath11 and ( [ 35 ] ) imply that @xmath628 so , as in the case of the @xmath25-forms , only two of the three momenta are essentially distinct . for any @xmath94 ,
the infinitesimal generator @xmath629 is a first - variation , so invariance of @xmath8 , namely @xmath630 , becomes @xmath631 . by left insertion into ( [ 31 ] ) , this becomes the discrete version of noether s theorem : @xmath632 = 0.7 [ [ conservation - in - a - space - and - time - split . ] ] conservation in a space and time split .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + to understand the significance of ( [ 32 ] ) and ( [ 38 ] ) consider a discrete field theory with space a discrete version of the circle and time the real line , as depicted in figure ( [ 39 ] ) , where space is split into space and time , with `` constant time '' being constant @xmath633 and the `` space index '' @xmath634 being cyclic . applying ( [ 38 ] ) to the region @xmath635 shown in the figure , noether s theorem takes the conservation form @xmath636 similarly , the discrete multisymplectic form formula also takes a conservation form .
when there is spatial boundary , the discrete noether theorem and the discrete multisymplectic form formulas automatically account for it , and thus form nontrivial generalizations of these conservation results . furthermore , as in the continuous case , we can achieve `` evolution type '' symplectic systems ( i.e. discrete moser - veselov mechanical systems ) if we define @xmath3 as the space of fields at constant @xmath633 , so @xmath637 , and take as the discrete lagrangian @xmath638,[q_j^1])\equiv\sum_{i=1}^n l(q_i^0,q_i^1,q_{i+1}^1).\ ] ] then the moser - veselov del evolution - type equations ( [ 40 ] ) are equivalent to the delf equations ( [ 41 ] ) , the multisymplectic form formula implies symplecticity of the moser - veselov evolution map , and conservation of momentum gives identical results in both the `` field '' and `` evolution '' pictures .
[ [ example - nonlinear - wave - equation.-1 ] ] example : nonlinear wave equation .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + to illustrate the discretization method we have developed , let us consider the lagrangian ( [ lag ] ) of section ( 4 ) , which describes the nonlinear sine - gordon wave equation .
this is a completely integrable system with an extremely interesting hierarchy of soliton solutions , which we shall investigate by developing for it a variational multisymplectic - momentum integrator ; see the recent article by palais [ 1997 ] for a wonderful discussion on soliton theory .
to discretize the continuous lagrangian , we visualize each triangle @xmath557 as having base length @xmath639 and height @xmath640 , and we think of the discrete jet @xmath641 as corresponding to the continuous jet @xmath642 where @xmath643 is a the center of the triangle . for insertion into the nonlinear term instead of @xmath643 .
] this leads to the discrete lagrangian @xmath644 with corresponding delf equations @xmath645 when @xmath646 ( wave equation ) this gives the explicit method @xmath647 which is stable whenever the courant stability condition is satisfied . [
[ extensions - jets - from - rectangles - and - other - polygons . ] ] extensions : jets from rectangles and other polygons .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + our choice of discrete jet bundle is obviously not restricted to triangles , and can be extended to rectangles or more general polygons ( left of figure ( [ 60 ] ) ) .
a * _ _ * rectangle is a quadruple of the form , @xmath648 a point is an * _ _ * interior point of a subset @xmath217 of rectangles if @xmath217 contains all four rectangles touching that point , the discrete lagrangian depends on variables @xmath649 , and the delf equations become @xmath650 the extension to polygons with even higher numbers of sides is straightforward ; one example is illustrated on the right of figure ( [ 60 ] ) .
= 0.4 the motivation for consideration of these extensions is enhancing the stability of the triangle - based method in the nonlinear wave example just above .
[ [ example - nonlinear - wave - equation - rectangles . ] ] example : nonlinear wave equation , rectangles .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + think of each rectangle @xmath557 as having length @xmath639 and height @xmath640 , and each discrete jet @xmath651 being associated to the continuous jet @xmath652 where @xmath653 is a the center of the rectangle .
this leads to the discrete lagrangian @xmath654 if , for brevity , we set @xmath655 then one verifies that the delf equations become @xmath656 -\biggl[\frac { y_{{i\,j+1 } } -2\,y_{{ij}}+y_{{i\,j-1}}}{{h}^{2}}\biggr]\\ & & \qquad\qquad\mbox { } + \frac14\biggl[n^\prime(\bar y_{ij})+ n^\prime(\bar y_{i\,j-1})+ n^\prime(\bar y_{i-1\,j-1})+n^\prime(\bar y_{i-1\,j } ) \biggr]=0,\end{aligned}\ ] ] which , if we make the definitions @xmath657,&\end{aligned}\ ] ] is ( more compactly ) @xmath658 -\frac1{h^2}\partial^2_hy_{ij } + \bar n^\prime(\bar y_{ij})=0.\label{101}\ ] ] these are implicit equations which must be solved for @xmath659 , @xmath634 , given @xmath660 , @xmath661 , @xmath634 ; rearranging , an iterative form equivalent to ( [ 101 ] ) is @xmath662 in the case of the sine - gordon equation the values of the field ought to be considered as lying in @xmath663 , by virtue of the vertical symmetry @xmath664 .
soliton solutions for example will have a jump of @xmath665 and the method will fail unless field values at close - together spacetime points are differenced modulo @xmath665 . as a result it becomes important to calculate using integral multiples of small field - dependent quantities , so that it is clear when to discard multiples of @xmath665 , and for this the above iterative form is inconvenient .
but if we define @xmath666 then there is the following iterative form , again equivalent to ( [ 101 ] ) @xmath667 one can also modify ( [ 100 ] ) so as to treat space and time symmetrically , which leads to the discrete lagrangian @xmath668 and one verifies that the delf equations become @xmath669\nonumber\\ & & \qquad\mbox{}-\frac1{h^2}\biggl [ \frac14\partial^2_hy_{i+1\,j}+\frac12\partial^2_hy_{ij}+ \frac14\partial^2_hy_{i-1\,j}\biggr ] + \bar n^\prime(\bar y_{ij})=0,\label{102}\end{aligned}\ ] ] an equivalent iterative form of which is @xmath670 = 2.5 in = 2.5 in = 2.5 in = 2.5 in while the focus of this article is not the numerical implementation of the integrators which we have derived , we have , nevertheless , undertaken some preliminary numerical investigations of our multisymplectic methods in the context of the sine - gordon equation with periodic boundary conditions .
[ [ the - rectangle - based - multisymplectic - method . ] ] the rectangle - based multisymplectic method .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the top half of figure ( [ 200 ] ) shows a simulation of the collision of `` kink '' and `` antikink '' solitons for the sine - gordon equation , using the rectangle - based multisymplectic method ( [ 102 ] ) . in the bottom half of that figure
we show the result of running that simulation until the solitons have undergone about @xmath671 collisions ; shortly after this the simulation stops because the iteration ( [ 111 ] ) diverges .
the anomalous spatial variations in the waveform of the bottom left of figure ( [ 200 ] ) have period @xmath0 spatial grid divisions and are shown in finer scale on the bottom right of that figure .
these variations are reminiscent of those found in ablowitz , herbst and schober [ 1996 ] for the completely integrable discretization of hirota , where the variations are attributed to independent evolution of waveforms supported on even vs. odd grid points .
observation of ( [ 102 ] ) indicates what is wrong : the nonlinear term @xmath158 contributes to ( [ 102 ] ) in a way that will average out these variations , and consequently , once they have begun , ( [ 102 ] ) tends to continue such variations via the linear wave equation . in ablowitz et .
, the situation is rectified when the number of spatial grid points is not even , and this is the case for ( [ 102 ] ) as well .
this is indicated on the left of figure ( [ 202 ] ) , which shows the waveform after about @xmath672 soliton collisions when @xmath673 rather than @xmath674 .
figure ( [ 203 ] ) summarizes the evolution of energy error ] ] for that simulation .
[ [ initial - data . ] ] initial data .
+ + + + + + + + + + + + + for the two - soliton - collision simulations , we used the following initial data : @xmath675 ( except @xmath676 where noted ) , where @xmath677 and @xmath673 spatial grid points ( except figure ( [ 200 ] ) where @xmath674 ) .
the circle that is space should be visualized as having circumference @xmath678 .
let @xmath679 where @xmath680 , @xmath681 , @xmath682 and @xmath683 then @xmath684 is a kink solution if space has a circumference of @xmath685 .
this kink and an oppositely moving antikink ( but placed on the last quarter of space ) made up the initial field , so that @xmath686 , @xmath687 , where @xmath688 while @xmath689 where @xmath690 = 2.5 in = 2.5 in = 2.5 in = 2.5 in [ [ comparison - with - energy - conserving - methods . ] ] comparison with energy - conserving methods .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + as an example of how our method compares with an existing method , we considered the energy - conserving method of vu - quoc and li [ 1993 ] , page 354 : @xmath669 -\frac1{h^2}\partial^2_hy_{ij}\nonumber\\ & & \qquad\mbox{}+\frac12\left ( \frac{n(y_{i\,j+1})-n(y_{ij})}{y_{i\,j+1}-y_{ij } } + \frac{n(y_{ij})-n(y_{i\,j-1})}{y_{ij}-y_{i\,j-1 } } \right)=0.\label{155}\end{aligned}\ ] ] this has an iterative form similar to ( [ 111 ] ) and is quite comparable with ( [ 101 ] ) and ( [ 102 ] ) in terms of the computation required .
our method seems to preserve the soliton waveform better than ( [ 155 ] ) , as is indicated by comparison of the left and right figure ( [ 202 ] ) . in regards to the closely related papers vu - quoc and li [ 1993 ] and li and vu - quoc [ 1995 ]
, we could not verify in our simulations that their method conserves energy , nor could we verify their _ proof _ that their method conserves energy .
so , as a further check , we implemented the following energy - conserving method of guo , pascual , rodriguez , and vazquez [ 1986 ] : @xmath691 which conserves the discrete energy @xmath692 this method diverged after just 345 soliton collisions .
= 2.5 in = 2.5 in = 2.5 in = 2.5 in as can be seen from ( [ 500 ] ) , the nonlinear potential @xmath158 enters as a difference over two grid spacings , which suggests that halving the time step might result in a more fair comparison with the methods ( [ 102 ] ) or ( [ 155 ] ) . with this advantage ,
method ( [ 500 ] ) was able to simulate 5000 soliton collisions , with a waveform degradation similar to the energy - conserving method ( [ 155 ] ) , as shown at the bottom right of figure ( [ 705 ] ) .
the same figure also shows that , although the energy behavior of ( [ 500 ] ) is excellent for short time simulations , it drifts significantly over long times , and the final energy error has a peculiar appearance .
figure ( [ 706 ] ) shows the time evolution of the waveform through the soliton collision that occurs just before the simulation stops .
apparently , at the soliton collisions , significant high frequency oscillations are present , and these are causing the jumps in the energy error in the bottom left plot of figure ( [ 705 ] ) .
this error then accumulates due to the energy - conserving property of the method .
in these simulations , so as to guard against the possibility that this behavior of the energy was due to inadequately solving the implicit equation ( [ 500 ] ) , we imposed a minimum limit of 3 iterations in the corresponding iterative loop , whereas this loop would otherwise have converged after just 1 iteration .
= 5.5 in = 5.5 in = 5.5 in = 5.5 in [ [ comparison - with - the - triangle - based - multisymplectic - method . ] ] comparison with the triangle - based multisymplectic method .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the discrete second derivatives in the method ( [ 500 ] ) are the same as in the triangle - based multisymplectic method ( [ explicit ] ) ; these derivatives are simpler than either our rectangle - based multisymplectic method ( [ 102 ] ) or the energy - conserving method of vu - quoc and li ( [ 155 ] ) . to explore this we implemented the triangle - based multisymplectic method ( [ explicit ] ) .
even with the less complicated discrete second derivatives our triangle - based multisymplectic method simulated 5000 soliton collisions with comparable energy ] ] and waveform preservation properties as the rectangle - based multisymplectic method ( [ 102 ] ) , as shown in figure ( [ 703 ] ) .
figure ( [ 790 ] ) shows the time evolution of the waveform through the soliton collision just before the simulation stops , and may be compared to figure ( [ 706 ] ) .
as can be seen , the high frequency oscillations that are present during the soliton collisions are smaller and more smooth for the triangle - based multisymplectic method than for the energy - conserving method ( [ 500 ] ) .
a similar statement is true irrespective which of the two multisymplectic or two energy conserving methods we tested , and is true all along the waveform , irrespective of whether or not a soliton collision is occurring .
= 2.5 in = 2.5 in = 2.5 in = 2.5 in [ [ summary . ] ] summary .
+ + + + + + + + our multisymplectic methods are finite difference methods that are computationally competitive with existing finite difference methods .
our methods show promise for long - time simulations of conservative partial differential equations , in that , for long - time simulations of the sine - gordon equation , our method 1 ) had superior energy - conserving behavior , _ even when compared with energy - conserving methods _ ; 2 ) better preserved the waveform than energy - conserving methods ; and 3 ) exhibited superior stability , in that our methods excited smaller and more smooth high frequency oscillations than energy - conserving methods . however , further numerical investigation is certainly necessary to make any lasting conclusions about the long - time behavior of our integrator .
[ [ the - programs . ] ] the programs .
+ + + + + + + + + + + + + the programs that were used in the preceding simulations are `` c '' language implementations of the various methods .
a simple tridiagonal lud method was used to solve the linear equations ( e.g. the left side of ( [ 111 ] ) ) , as in vu - quoc and li [ 1993 ] , page 379 .
an @xmath693 order extrapolator was used to provide a seed for the implicit step .
all calculations were performed in double precision while the implicit step was terminated when the fields ceased to change to single precision ; the program s output was in single precision .
the extrapolation usually provided a seed accurate enough so that the methods became practically _
explicit _ , in that for many of the time - steps the first or second run through the iterative loops solving the implicit equations solved those equations to single precision .
however , in the absence of a regular spacetime grid the expenses of the extrapolation and solving the linear equation would grow .
our programs are freely available at url ` http://www.cds.caltech.edu/cds ` .
here we make a few miscellaneous comments and remark on some work planned for the future . as mentioned in the text , it is useful to have a discrete counterpart to the lagrangian reduction of marsden and scheurle [ 1993a , b ] , holm , marsden and ratiu [ 1998a ] and cendra , marsden and ratiu [ 1998 ] .
we sketch briefly how this theory might proceed .
this reduction can be done for both the case of `` particle mechanics '' and for field theory . for particle mechanics ,
the simplest case to start with is an invariant ( say left ) lagrangian on the tangent bundle of a lie group : @xmath694 .
the reduced lagrangian is @xmath695 and the corresponding euler
poincar equations have a variational principle of lagrange dalembert type in that there are constraints on the allowed variations .
this situation is described in marsden and ratiu [ 1994 ] .
the discrete analogue of this would be to replace a discrete lagrangian @xmath696 by a reduced discrete lagrangian @xmath697 related to @xmath698 by @xmath699 in this situation , the algorithm from @xmath700 to @xmath700 reduces to one from @xmath86 to @xmath86 and it is generated by @xmath701 in a way that is similar to that for @xmath702 .
in addition , the discrete variational principle for @xmath703 which states that one should find critical points of @xmath704 with respect to @xmath705 to implicitly define the map @xmath706 , reduces naturally to the following principle : find critical points of @xmath707 with respect to variations of @xmath708 and @xmath639 of the form @xmath709 and @xmath710 where @xmath711 and @xmath712 denote left and right translation and where @xmath713 . in other words ,
one sets to zero , the derivative of the sum @xmath714 with respect to @xmath112 at @xmath715 for a curve @xmath716 in @xmath86 that passes through the identity at @xmath717 .
this defines ( with caveats of regularity as before ) a map of @xmath86 to itself , which is the reduced algorithm .
this algorithm can then be used to advance points in @xmath718 itself , by advancing each component equally , reproducing the algorithm on @xmath700 .
in addition , this can be used with the adjoint or coadjoint action to advance points in @xmath87 or @xmath719 to approximate the euler poincar or lie poisson dynamics .
these equations for a discrete map , say @xmath720 generated by @xmath701 on @xmath86 are called the * _ _ * discrete euler poincar equations as they are the discrete analogue of the euler poincar equations on @xmath87 .
notice that , at least in theory , computation can be done for this map first and then the dynamics on @xmath700 is easily reconstructed by simply advancing each pair as follows : @xmath721 , where @xmath722 .
if one identifies the discrete lagrangians with generating functions ( as explained in wendlandt and marsden [ 1997 ] ) then the reduced lagrangian generates the reduced algorithm in the sense of ge and marsden [ 1988 ] , and this in turn is closely related to the lie
hamilton jacobi theory .
next , consider the more general case of @xmath4 with its discretization @xmath117 with a group action ( assumed to be free and proper ) by a lie group @xmath86 .
the reduction of @xmath4 by the action of @xmath86 is @xmath723 , which is a bundle over @xmath724 with fiber isomorphic to @xmath87 .
the discrete analogue of this is @xmath725 which is a bundle over @xmath726 with fiber isomorphic to @xmath86 itself .
the projection map @xmath727 is given by @xmath728 \mapsto ( [ q_1 ] , [ q_2])$ ] where @xmath729 $ ] denotes the relevant equivalence class .
notice that in the case in which @xmath730 this bundle is `` all fiber '' .
the reduced discrete euler - lagrange equations are similar to those in the continuous case , in which one has shape equations couples with a version of the discrete euler
poincar equations .
of course all of the machinery in the continuous case can be contemplated here too , such as stability theory , geometric phases , etc .
in addition , it would be useful to generalize this lagrangian reduction theory to the multisymplectic case .
all of these topics are planned for other papers .
consider an autonomous , continuous lagrangian @xmath731 where , for simplicity , @xmath3 is an open submanifold of euclidean space .
imagine some _ not necessarily uniform _ temporal grid ( @xmath732 ) of @xmath733 , so that @xmath734 .
in this situation , it is natural to consider the discrete action @xmath735 this action principle deviates from the action principle ( [ 850 ] ) of section 3 in that the discrete lagrangian density depends explicitly on @xmath640 .
of course nonautonomous continuous lagrangians also yield @xmath640-dependent discrete lagrangian densities , irrespective of uniformity of the grid .
thus , nonuniform temporal grids or nonautonomous lagrangians give rise to discrete lagrangian densities which are more general those those we have considered in section ( 3 ) . for field theories , the lagrangian in the action ( [ 30 ] ) depends on the spacetime variables already , through its explicit dependence on the triangle @xmath557 .
however , it is only in the context of a uniform grid that we have experimented numerically and only in that context that we have discussed the significance of the discrete multisymplectic form formula and the discrete noether theorem . using ( [ 800 ] ) as an example , will now indicate why the issue of grid uniformity may not be serious .
the del equations corresponding to the action ( [ 800 ] ) are @xmath736 and this gives evolution maps @xmath737 defined so that @xmath738 when ( [ 802 ] ) holds . for the canonical 1-forms corresponding to ( [ 13 ] ) and ( [ 14 ] ) we have the @xmath640-dependent one forms @xmath739 and @xmath740 and equations ( [ 16 ] ) and ( [ 2010 ] ) become @xmath741 respectively . together , these two equations give @xmath742 and if we set @xmath743 then ( [ 807 ] ) chain together to imply @xmath744 .
this appears less than adequate since it merely says that the pull back by the evolution of a certain 2-form is , in general , a different 2-form .
the significant point to note , however , is that _ this situation may be repaired at any @xmath640 simply by choosing @xmath745 . _
it is easily verified that the analogous statement is true with respect to momentum preservation via the discrete noether theorem .
specifically , imagine integrating a symmetric autonomous mechanical system in a timestep adaptive way with equations ( [ 802 ] ) . as the integration proceeds
, various timesteps are chosen , and if momentum is monitored it will show a dependence on those choices . _
a momentum - preserving symplectic simulation may be obtained by simply choosing the last timestep to be of equal duration to the first .
_ this is the highly desirable situation which gives us some confidence that grid uniformity is a nonissue .
there is one caveat : symplectic integration algorithms are evolutions which are high frequency perturbations of the actual system , the frequency being the inverse of the timestep , which is generally far smaller than the time scale of any process in the simulation .
however , timestep adaptation schemes will make choices on a much larger time scale than the timestep itself , and then drift in the energy will appear on this larger time scale .
a meaningful long - time simulation can not be expected in the unfortunate case that the timestep adaptation makes repeated choices in a way that resonates with some process of the system being simulated .
the sphere can not be generally uniformly subdivided into spherical triangles ; however , a good approximately uniform grid is obtained as follows : start from an inscribed icosahedron which produces a uniform subdivision into twenty spherical isosceles triangles ; these are further subdivided by halving their sides and joining the resulting points by short geodesics .
the variational approach we have developed allows us to examine the multisymplectic structure of elliptic boundary value problems as well .
for a given lagrangian , we form the associated action function , and by computing its first variation , we obtain the unique multisymplectic form of the elliptic operator .
the multisymplectic form formula contains information on how symplecticity interacts with spatial boundaries . in the case of two spatial dimensions , @xmath746 , @xmath747
, we see that equation ( [ t15 ] ) gives us the conservation law @xmath748 where the vector @xmath749 .
furthermore , using our generalized noether theory , we may define momentum - mappings of the elliptic operator associated with its symmetries .
it turns out that for important problems of spatial complexity arising in , for example , pattern formation systems , the covariant noether current intrinsically contains the constrained toral variational principles whose solutions are the complex patterns ( see marsden and shkoller [ 1997 ] ) .
there is an interesting connection between our variational construction of multisymplectic - momentum integrators and the finite element method ( fem ) for elliptic boundary value problems .
fem is also a variationally derived numerical scheme , fundamentally differing from our approach in the following way : whereas we form a discrete action sum and compute its first variation to obtain the discrete euler - lagrange equations , in fem , it is the original continuum action function which is used together with a projection of the fields and their variations onto appropriately chosen finite - dimensional spaces .
one varies the projected fields and integrates such variations over the spatial domain to recover the discrete equations . in general
, the two discretization schemes do not agree , but for certain classes of finite element bases with particular integral approximations , the resulting discrete equations match the discrete euler - lagrange equations obtained by our method , and are hence naturally multisymplectic .
to illustrate this concept , we consider the gregory and lin method of solving two - point boundary value problems in optimal control . in this scheme ,
the discrete equations are obtained using a finite element method with a basis of linear interpolants . over each one - dimensional element ,
let @xmath750 and @xmath751 be the two linear interpolating functions . as usual
, we define the action function by @xmath752 .
discretizing the interval @xmath753 $ ] into @xmath158@xmath754@xmath25 uniform elements , we may write the action with fields projected onto the linear basis as @xmath755 since the euler - lagrange equations are obtained by linearizing the action and hence the lagrangian , and as the functions @xmath756 are linear , one may easily check that by evaluating the integrals in the linearized equations using a trapezoidal rule , the discrete euler - lagrange equations given in ( [ 12 ] ) are obtained .
thus , the gregory and lin method is actually a multisymplectic - momentum algorithm .
fluid problems are not literally covered by the theory presented here because their symmetry groups ( particle relabeling symmetries ) are not vertical .
a generalization is needed to cover this case and we propose to work out such a generalization in a future paper , along with numerical implementation , especially for geophysical fluid problems in which conservation laws such as conservation of enstrophy and kelvin theorems more generally are quite important . it remains to link the approaches here with other types of integrators , such as volume preserving integrators ( see , eg , kang and shang [ 1995 ] , quispel [ 1995 ] ) and reversible integrators ( see , eg , stoffer [ 1995 ] ) . in particular
since volume manifolds may be regarded as multisymplectic manifolds , it seems reasonable that there is an interesting link .
one of the very nice things about the veselov construction is the way it handles constraints , both theoretically and numerically ( see wendlandt and marsden [ 1997 ] ) . for field theories
one would like to have a similar theory .
for example , it is interesting that for fluids , the incompressibility constraint can be expressed as a pointwise constraint on the first jet of the particle placement field , namely that its jacobian be unity .
when viewed this way , it appears as a holonomic constraint and it should be amenable to the present approach . under reduction by the particle relabeling group ,
such a constraint of course becomes the divergence free constraint and one would like to understand how these constraints behave under both reduction and discretization .
we would like to extend our gratitude to darryl holm , tudor ratiu and jeff wendlandt for their time , encouragement and invaluable input .
work of j. marsden was supported by the california institute of technology and nsf grant dms 9633161 .
work by g. patrick was partially supported by nserc grant ogp0105716 and that of s. shkoller was partially supported by the cecil and ida m. green foundation and doe .
we also thank the control and dynamical systems department at caltech for providing a valuable setting for part of this work .
arms , j.m .
marsden , and v. moncrief [ 1982 ] the structure of the space solutions of einstein s equations : ii several killings fields and the einstein - yang - mills equations .
_ ann . of phys . _
* 144 * , 81106 .
arms , j.m .
marsden , and v. moncrief [ 1982 ] the structure of the space solutions of einstein s equations : ii several killings fields and the einstein - yang - mills equations .
_ ann . of phys . _
* 144 * , 81106 .
holm , d. d. , j.e .
marsden and t. ratiu [ 1998b ] the euler - poincar equations in geophysical fluid dynamics , in _ proceedings of the isaac newton institute programme on the mathematics of atmospheric and ocean dynamics _ , cambridge university press ( to appear ) .
marsden , j.e . and j.m . wendlandt [ 1997 ] mechanical systems with symmetry , variational principles and integration algorithms .
_ current and future directions in applied mathematics _ , edited by m. alber , b. hu , and j. rosenthal , birkhuser , 219261 | this paper presents a geometric - variational approach to continuous and discrete mechanics and field theories . using multisymplectic geometry
, we show that the existence of the fundamental geometric structures as well as their preservation along solutions can be obtained directly from the variational principle .
in particular , we prove that a unique multisymplectic structure is obtained by taking the derivative of an action function , and use this structure to prove covariant generalizations of conservation of symplecticity and noether s theorem .
natural discretization schemes for pdes , which have these important preservation properties , then follow by choosing a discrete action functional . in the case of mechanics ,
we recover the variational symplectic integrators of veselov type , while for pdes we obtain covariant spacetime integrators which conserve the corresponding discrete multisymplectic form as well as the discrete momentum mappings corresponding to symmetries .
we show that the usual notion of symplecticity along an infinite - dimensional space of fields can be naturally obtained by making a spacetime split .
all of the aspects of our method are demonstrated with a nonlinear sine - gordon equation , including computational results and a comparison with other discretization schemes . |
let @xmath6 be the set of positive integers , @xmath7 , @xmath8 the ring of rational integers , @xmath9 the rational number field , @xmath10 the real number field , @xmath11 the complex number field , respectively . in our previous articles
@xcite , we defined the multi - variable version of witten zeta - functions , or `` zeta - functions of root systems '' , inspired by the original work of witten @xcite and of zagier @xcite .
we recall these results as follows .
let @xmath12 be a complex semisimple lie algebra with rank @xmath13 , @xmath14 be a cartan subalgebra of @xmath15 and @xmath16 be its dual .
let @xmath17 be the set of all roots of @xmath12 , @xmath18 the set of all positive roots of @xmath12 , @xmath19 the fundamental system of @xmath20 , and @xmath21 the coroot associated with @xmath22 ( @xmath23 ) .
let @xmath24 be the fundamental weights satisfying @xmath25 ( kronecker s delta ) . in the following
we denote the pairing @xmath26 of @xmath27 and @xmath28 by @xmath29 . in @xcite , we defined the multi - variable version of witten zeta - functions by @xmath30 for @xmath31 , where @xmath32 is the number of all positive roots . in the case
that @xmath12 is of type @xmath33 , we call ( [ 1 - 3 ] ) the zeta - function of the root system of type @xmath33 , and also denote it by @xmath34 , where @xmath35 .
note that the original witten zeta - function @xmath36 , studied by witten @xcite and zagier @xcite , coincides with @xmath37 where @xmath38 witten s motivation of introducing the above zeta - functions is to express the volumes of certain moduli spaces in terms of special values of @xmath36 .
this expression is called witten s volume formula , which implies that @xmath39 for any @xmath40 , where @xmath41 ( see ( * ? ? ?
* theorem , p.506 ) ) . in general , the explicit value of @xmath42 was not determined in their work .
in our previous work @xcite , we defined the bernoulli polynomials of root systems , and proved a formula which expresses @xmath42 in terms of those bernoulli polynomials .
consequently we were able to obtain a certain generalization of .
we further gave some functional relations for zeta - functions of root systems which include as special value - relations .
in fact , we studied explicit functional relations for zeta - functions of @xmath43 type in @xcite , and of @xmath44 and @xmath45 types in @xcite ( see also @xcite ) . in the present paper
we continue our research mentioned above .
the main aim of the present paper is to study the zeta - function of @xmath5-type . in section [ sec-2 ]
, we define the bernoulli polynomials of @xmath5-type and study the generating functions of them . by this consideration ,
we give for @xmath46 with explicit values of @xmath47 and more generalized results . in section [ sec-3 ] ,
we consider analytic properties of @xmath46 based on our previous paper @xcite .
actually we determine the possible singularities of @xmath46 . in section [ sec-4 ] ,
we quote several lemmas which were shown in our previous papers @xcite .
furthermore we prove a certain analogue of them .
these lemmas will play important roles in the next section .
finally , in section [ sec-5 ] , by using these lemmas , we give explicit functional relations for @xmath46 , which include at their special values .
recently , zhao expressed several values @xmath48 for @xmath49 in terms of double polylogarithms and gave their approximate values in @xcite .
we will be able to give some of these values exactly ( see example [ exam-2 - 1 ] and remark [ rem-4 - 7 ] ) .
a part of these results has also been announced in our previous paper @xcite .
finally we remark that it is theoretically possible to prove the same type of results for zeta - functions of other exceptional types @xmath50 , @xmath51 , @xmath52 and @xmath53 , by using our method .
however , it may be considerably hard to apply our method actually to those cases , while it is interesting to determine @xmath54 explicitly in those cases .
in our previous papers @xcite , we have already studied the general theory of zeta - functions of root systems .
we apply it to the case of @xmath5 as follows .
let @xmath55 be the root system of @xmath5-type and let @xmath56 be a real vector subspace .
let @xmath18 and @xmath57 be the set of all positive roots and negative roots respectively .
then we have a decomposition of the root system @xmath58 . we know that @xmath18 is given by @xmath59 where @xmath60 is the set of fundamental roots and @xmath61 ( see , for example , bourbaki @xcite ) .
g2 ( 73,48)@xmath62 ( 0,76)@xmath63 ( 82,76)@xmath64 ( 40,98)@xmath65 ( 26,76)@xmath66 ( 56,76)@xmath67 g2c ( 96,43)@xmath68 ( 20,62)@xmath69 ( 72,62)@xmath70 ( 48,74)@xmath71 ( 24,86)@xmath72 ( 70,86)@xmath73 let @xmath74 be the weyl group of @xmath5-type . for @xmath75 , we set @xmath76 from and , we see that the zeta - function of the root system of @xmath5-type can be given by @xmath77 where @xmath78 ( @xmath79 ) . furthermore , for @xmath80 and @xmath81 , we define @xmath82 note that we only use the case when @xmath83 in this paper . however , we study the case of general @xmath84 here for the convenience of our research in the future . with the above notation such as @xmath74 and @xmath85 ,
we let @xmath86 where @xmath87 with the identification @xmath88 and @xmath89 is the usual action of @xmath84 by @xmath90 .
this @xmath91 is a `` weyl group symmetric '' linear combination of zeta - functions of root systems , which plays a fundamental role in the study of value - relations and functional relations in @xcite . in order to evaluate @xmath91 at positive integers
, we consider bernoulli polynomials @xmath92 via their generating function @xmath93 .
this type of generalized bernoulli polynomials associated with any root system was first introduced in @xcite , and was further studied in @xcite . for a real number @xmath94 , let @xmath95 denote its fractional part
@xmath96 $ ] . applying theorem 4.1 in @xcite to the case of @xmath5-type , we obtain @xmath97 then @xmath98 is holomorphic at the origin and can be expanded as @xmath99 for @xmath100 . from our previous results (
* theorem 4.4,(4.19 ) and ( 4.20 ) ) , we obtain the following . [
t-2 - 1 ] for @xmath101 , @xmath102 [ exam-2 - 1 ] in the case @xmath103 for @xmath104 and @xmath105 , we see that @xmath106 on the other hand , from the definition of @xmath107 , we can calculate @xmath108 .
combining this fact with , we can obtain the explicit values of @xmath109 , for example , @xmath110 furthermore , in the case when @xmath111 @xmath112 , we have @xmath113 this is because the lengths of @xmath62 , @xmath66 and @xmath67 ( and of @xmath63 , @xmath64 and @xmath65 ) are the same , and the roots of the same length form a single weyl - group orbit .
hence we can obtain , for example , @xmath114 it is possible to compute the numerical values of the left - hand sides of the above formulas from the definition .
we have already checked that those numerical values agree with the above formulas .
in the preceding section we studied `` value - relations '' for @xmath46 , but we can further discuss `` functional relations '' for this function . for this purpose ,
we first consider analytic properties .
[ t-3 - 1 ] the function @xmath115 can be continued meromorphically to the whole space @xmath116 , and its possible singularities are located on the subsets of @xmath116 defined by one of the equations : @xmath117 the meromorphic continuation of @xmath46 , as well as zeta - functions of other exceptional algebras , can be deduced from earlier results given by essouabri @xcite , the second - named author ( * ? ? ?
* theorem 3 ) , and the first - named author @xcite .
however the following argument of determining the possible singularities also includes a proof of meromorphic continuation . at the end of (
* remark 6.4 ) , we discussed when the determination of possible singularities can be achieved just by shifting the path of integration .
the arrow from @xmath5 to @xmath118 in the diagram in ( * ? ? ?
* section 5 ) is horizontal , hence this is the case when the shifting of the path is sufficient .
therefore our argument here is not so complicated , similar to that in @xcite .
note that such simple shifting argument is not sufficient when one studies analytic properties of zeta - functions of other exceptional algebras . at first , assume @xmath119 @xmath120 are sufficiently large .
the mellin - barnes integral formula is @xmath121 where @xmath122 , @xmath123 , @xmath124 , @xmath125 with @xmath126 and the path @xmath127 of integration is the vertical line @xmath128 . by using , we first prove an integral expression of @xmath46 in terms of the zeta - function of @xmath118-type defined by @xmath129 ( see ( * ? ? ? * ( 6.1 ) ) and ( * ? ? ?
* example 7.3 ) ) .
writing @xmath130 and applying to the second factor of the right - hand side , we have @xmath131 with @xmath132 .
similarly @xmath133 with @xmath134 .
hence @xmath135 the singularities of @xmath136 are determined by ( * ? ? ?
* theorem 6.2 ) .
we find that the singularities of the zeta factor on the right - hand side of is @xmath137 let @xmath138 be a large positive integer , and define @xmath139 we can rewrite as @xmath140 where @xmath141 the singularities of the integrand on the right - hand side of are @xmath142 , @xmath143 @xmath144 , singularities of type , and of type with @xmath145 .
shifting the path to @xmath146 , where @xmath147 is a large positive integer and @xmath148 is a small positive number , and counting the residues at @xmath149 , we obtain @xmath150 where @xmath151 the above integral is holomorphic in the region @xmath152 and @xmath153 since @xmath147 is arbitrary , we now find that @xmath154 is continued meromorphically to the region
. we can also show that @xmath154 is of polynomial order with respect to the imaginary parts of variables .
the singularities of @xmath154 in the region are located only on @xmath155 which are coming from the zeta - factors in the sum - part on the right - hand side of .
now go back to the situation when @xmath119 @xmath156 are large , and consider .
let @xmath157 insert @xmath158 on the right - hand side of , and shift the path to @xmath159 to obtain @xmath160 this gives the continuation of @xmath46 to the region .
since @xmath138 is arbitrary , we obtain the meromorphic continuation of @xmath46 to the whole space .
its possible singularities are coming from @xmath158 and @xmath161 on the right - hand side of , which are exactly those stated in the theorem .
this completes the proof of theorem [ t-3 - 1 ] .
in this section , we quote several lemmas from our previous papers @xcite and further prove an analogue of them
. these will play important roles in the next section . from now on ,
the symbol @xmath162 implies ordinary curly parentheses , not the fractional part .
[ l-4 - 1 ] let @xmath163 , and @xmath164 be arbitrary functions .
then , for @xmath165 , @xmath166 } f(k-2\mu)\frac{(i\pi)^{2\mu}}{(2\mu ) ! } = \sum_{\xi=0}^{[a/2]}\zeta(2\xi)f(a-2\xi ) , \label{eq-4 - 1}\end{aligned}\ ] ] and @xmath167 } g(k-2\mu)\frac{(i\pi)^{2\mu}}{(2\mu+1 ) ! }
= -\frac{1}{2}g(a ) , \label{eq-4 - 2}\end{aligned}\ ] ] where @xmath168 for @xmath169 .
[ l-4 - 2 ] let @xmath170 be sequences such that @xmath171 for any @xmath172 .
then @xmath173 for any @xmath172 .
note that , in ( * ? ? ?
* lemma 4.4 ) , we proved only .
however , by just the same method , we can easily obtain .
[ l-4 - 3 ] let @xmath174 , and @xmath175 be sets of numbers indexed by integers . assume that the infinite series appearing in @xmath176 are absolutely convergent for @xmath177 $ ] , and that ( [ eq-4 - 3 ] ) is a constant function for @xmath177 $ ] .
then , for @xmath178 , @xmath179 holds for @xmath177 $ ] , where the infinite series appearing on the left - hand side of ( [ eq-4 - 4 ] ) are absolutely convergent for @xmath177 $ ] .
now we prepare the following lemma which is an analogue of lemma [ l-4 - 3 ] .
[ l-4 - 4 ] let @xmath174 , @xmath180 be sets of numbers indexed by integers , and @xmath181 assume that both of the right - hand sides of @xmath182 in are absolutely convergent for @xmath177 $ ] , and that both @xmath183 and @xmath184 are constant functions on @xmath185 $ ] . then
, for @xmath186 , @xmath187}\zeta(2k ) \sum_{\omega=0}^{c_\eta-2k}\binom{\omega+2d-1}{\omega}(-2)^{\omega } \sum_{m \in \mathbb{z } \atop m \not= 0}\frac{\beta(m;c_\eta-2k-\omega;\eta)}{m^{2d+\omega } } \\ & \ + 2\sum_{k=0}^{d}\zeta(2k)2^{-2k } \sum_{\eta=1}^{h } \sum_{\omega=0}^{c_\eta-1}\binom{\omega+2d-2k}{\omega}(-2)^{\omega } \\ & \hspace{1 in } \times \sum_{m \in \mathbb{z } \atop m \not=0}\frac{((-1)^m+1)\beta(m;c_\eta-1-\omega;\eta)}{m^{2d-2k+\omega+1 } } \\ & \ -2\sum_{k=0}^{d}\zeta(2k)\left(1 - 2^{-2k}\right ) \sum_{\eta=1}^{h } \sum_{\omega=0}^{c_\eta-1}\binom{\omega+2d-2k}{\omega}(-2)^{\omega } \\ & \hspace{1 in } \times \sum_{m \in \mathbb{z } \atop m \not=0}\frac{((-1)^m-1)\beta(m;c_\eta-1-\omega;\eta)}{m^{2d-2k+\omega+1}}=0 \end{split } \label{eq-4 - 11}\ ] ] for @xmath177 $ ] , where the infinite series appearing on the left - hand side of ( [ eq-4 - 11 ] ) are absolutely convergent for @xmath177 $ ] .
put @xmath188 from the assumption , we see that @xmath189 are constant functions for @xmath177 $ ] .
also , by using the relation @xmath190 we can check that @xmath191 repeating the indefinite integration , we can write @xmath192 for some @xmath193 . putting @xmath194 for @xmath186 and @xmath195 in , we obtain @xmath196 similarly , putting @xmath197 and @xmath195 in , we have @xmath198 by lemma [ l-4 - 2 ] , we have @xmath199 we will calculate each side of explicitly as follows .
note that @xmath200 by using , we have @xmath201 } \sum_{\omega=0}^{j-2\mu } \binom{2d-1+\omega}{\omega}(-2)^{\omega}\\ & \ \ \ \ \times \sum_{m \in \mathbb{z } \atop m \not= 0}\frac{\left((-1)^m + 1\right)\beta(m;j-2\mu-\omega;\eta)}{m^{2d+\omega}}\frac{(i\pi)^{2\mu}}{(2\mu)!}\bigg\}\\ & = ( -1)^d 2^{2d+1 } \bigg\ { \sum_{m \in \mathbb{z } \atop m \not= 0}\frac{\left((-1)^m+ 1\right)\alpha(m)}{m^{2d } } \\ & \ \
-2\sum_{\eta=1}^{h}\sum_{\xi=0}^{[c_\eta/2]}\zeta(2\xi ) \sum_{\omega=0}^{c_\eta-2\xi } \binom{2d-1+\omega}{\omega}(-2)^{\omega}\\ & \ \ \ \ \times \sum_{m \in \mathbb{z } \atop m \not= 0}\frac{\left((-1)^m + 1\right)\beta(m;c_\eta-2\xi - w;\eta)}{m^{2d+\omega}}\bigg\}. \end{split } \label{eq-4 - 18}\ ] ] similarly , by using , we have @xmath202 substituting and into , we have @xmath203}\zeta(2\xi ) \sum_{\omega=0}^{c_\eta-2\xi } \binom{2d-1+\omega}{\omega}(-2)^{\omega}\\ & \ \ \ \ \times \sum_{m \in \mathbb{z } \atop m \not= 0}\frac{\left((-1)^m + 1\right)\beta(m;c_\eta-2\xi - w;\eta)}{m^{2d+\omega}}\\ & \ \ + 4\sum_{\eta=1}^{h}\sum_{\xi=0}^{d}\sum_{\omega=0}^{c_\eta-1}\zeta(2\xi)2^{-2\xi}\binom{2d-2\xi+\omega}{\omega}(-2)^{\omega } \\ & \ \ \ \ \times \sum_{m \in \mathbb{z } \atop m \not= 0}\frac{\left((-1)^m + 1\right)\beta(m;c_\eta-1-w;\eta)}{m^{2d-2\xi+1+\omega}}=0 .
\end{split } \label{eq-4 - 20}\ ] ] similarly to , we obtain @xmath204 hence , as well as , we have @xmath205}\zeta(2\xi ) \sum_{\omega=0}^{c_\eta-2\xi } \binom{2d-1+\omega}{\omega}(-2)^{\omega}\\ & \ \ \ \ \times \sum_{m \in \mathbb{z } \atop m \not= 0}\frac{\left((-1)^m -1\right)\beta(m;c_\eta-2\xi - w;\eta)}{m^{2d+\omega}}\\ & \ \
+ 4\sum_{\eta=1}^{h}\sum_{\xi=0}^{d}\sum_{\omega=0}^{c_\eta-1}\zeta(2\xi)\left(1 - 2^{-2\xi}\right)\binom{2d-2\xi+\omega}{\omega}(-2)^{\omega } \\ & \ \ \ \ \times \sum_{m \in \mathbb{z } \atop m
\not= 0}\frac{\left((-1)^m -1\right)\beta(m;c_\eta-1-w;\eta)}{m^{2d-2\xi+1+\omega}}=0 . \end{split } \label{eq-4 - 22}\ ] ] combining and , we obtain .
now , using the results prepared in the previous section , we construct functional relations for @xmath206 and @xmath207 .
first we recall the relation for zeta - functions of @xmath118-type which was proved in @xcite , and will extend this relation to that of @xmath5-type .
the technique is essentially introduced in our previous papers ( see ( * ? ? ?
* remark 7.5 ) ) . from ( * ? ?
* ( 8.4 ) ) , we have @xmath208 for @xmath177 $ ] , @xmath209 , @xmath210 with @xmath211 and @xmath212 with @xmath213 . here
we use the same method as introduced in our previous papers @xcite by making use of polylogarithms as follows . replacing @xmath94 by @xmath214 and moving the terms corresponding to @xmath215 of the first member on the left - hand side of the above equation to the right - hand side
, we have @xmath216 if we fix @xmath209 and @xmath210 with @xmath211 , then we can apply lemma [ l-4 - 3 ] to the above equation with @xmath217 .
consequently we have @xmath218 where @xmath219 we can similarly write @xmath220 , @xmath221 and @xmath222 , but they are omitted for the purpose of saving space .
now we apply lemma [ l-4 - 4 ] to with @xmath227 for @xmath228 .
in fact , we can see that the left - hand side of is of the same form as .
furthermore , by the same method as in ( * ? ? ?
* section 7 ) , we can confirm that @xmath229 from these results and theorem [ t-3 - 1 ] , we obtain the following theorem .
[ fn - rel ] for @xmath230 , @xmath231 holds for all @xmath232 except for singularities of functions on the left - hand side , where @xmath233 are , by using the notation @xmath234 , defined as follows : @xmath235 @xmath236 @xmath237 @xmath238 @xmath239 @xmath240 @xmath241 @xmath242 [ rem-4 - 7 ] in @xcite , zhao expressed several values @xmath48 for @xmath49 in terms of double polylogarithms and gave approximate values of them , for example , @xmath247 by using the same method as stated above , we can explicitly obtain @xmath248 which agrees with zhao s numerical computation .
we can further give a functional relation between @xmath46 and @xmath207 including , which is an analogue of .
under more preparations , we will be able to give evaluation formulas for a certain class of values @xmath48 for @xmath49 in terms of @xmath249 for @xmath250 .
we will state the details in a forthcoming paper .
y. komori , k. matsumoto and h. tsumura , zeta - functions of root systems , in `` proceedings of the conference on @xmath138-functions '' ( fukuoka , 2006 ) , l. weng and m. kaneko ( eds . ) , world scientific , 2007 , pp .
115 - 140 . ,
zeta and @xmath138-functions and bernoulli polynomials of root systems , _ proc .
japan acad .
a , * 84 * ( 2008 ) , 57 - 62 . , on witten multiple zeta - functions associated with semisimple lie algebras ii , to appear in _ j. math .
japan_. , on witten multiple zeta - functions associated with semisimple lie algebras iii , preprint , arxiv:0907.0955 . , on multiple bernoulli polynomials and multiple @xmath138-functions of root systems , to appear in _ proc .
london math .
_ , on the analytic continuation of various multiple zeta - functions , in `` number theory for the millennium ii , proc
. millennial conference on number theory '' , m. a. bennett et al .
( eds . ) , a k peters , 2002 , pp . 417 - 440 . , on mordell - tornheim and other multiple zeta - functions , in `` proceedings of the session in analytic number theory and diophantine equations '' ( bonn , january - june 2002 ) , d. r. heath - brown and b. z. moroz ( eds . ) , bonner mathematische schriften nr .
360 , bonn 2003 , n.25 , 17pp . | in our previous work , we established the theory of multi - variable witten zeta - functions , which are called the zeta - functions of root systems .
we have already considered the cases of types @xmath0 , @xmath1 , @xmath2 , @xmath3 and @xmath4 . in this paper
, we consider the case of @xmath5-type .
we define certain analogues of bernoulli polynomials of @xmath5-type and study the generating functions of them to determine the coefficients of witten s volume formulas of @xmath5-type .
next we consider the meromorphic continuation of the zeta - function of @xmath5-type and determine its possible singularities . finally , by using our previous method
, we give explicit functional relations for them which include witten s volume formulas . |
a stably - stratified atmosphere , characterized by a positive vertical entropy gradient , can support gravity waves , or @xmath0-modes .
gravity waves are oscillations which arise from the buoyancy of parcels in the fluid .
such waves are readily excited by flow over thermal and surface topography , convective and shear instabilities , and flow adjustment processes .
they propagate through the atmosphere both horizontally and vertically .
gravity waves in the terrestrial atmosphere and ocean are much studied ( e.g. , * ? ? ?
* ; * ? ? ?
they have also been observed on other solar system bodies , such as jupiter @xcite and venus @xcite . in the terrestrial atmosphere
, a typical gravity wave has an energy flux of approximately 10@xmath3 to 10@xmath4 w m@xmath5 . despite being small ,
compared to the total amount of absorbed solar flux ( @xmath6237 w m@xmath5 ) , gravity waves are responsible for significantly modifying even dictating large - scale flow and temperature structures . several well known examples of this are the quasi - biennial oscillation , reversal of mean meridional temperature gradient in the upper middle atmosphere , and generation of turbulence ( e.g. , * ? ? ?
we expect similar effects to be present in the atmospheres ( and oceans ) of extrasolar planets
. moreover , due to the greater irradiation and scale heights on them , the acceleration and heating effects of gravity waves can be much stronger on hot extrasolar planets .
there has been much interest in modeling the atmospheric circulation of extrasolar planets ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
accurate simulation of atmospheric circulation is crucial for interpreting observations of extrasolar planets , as well as for improving theoretical understanding in general . for this
, the role of eddies and waves in transferring momentum and heat needs to be addressed @xcite .
this has long been recognized in solar system planet studies ( e.g. , * ? ? ?
* ; * ? ? ?
the plan of the paper is as follows . in [
sec : theory ] we derive the governing equation appropriate for linear monochromatic gravity waves on hot extrasolar planets .
we also discuss a simple parameterization of the key non - linear process , saturation .
in addition , we present solutions to the equation for simple isothermal atmospheres , with and without shear in the background mean flow . in [ sec : application ] we extend the calculation to a physically more realistic situation , by using background flow and temperature profiles derived from a three - dimensional ( 3-d ) hot
jupiter atmospheric circulation simulation .
this is the first such calculation to have been performed for extrasolar planets . through this ,
the significant effects of gravity waves on hot extrasolar planet atmospheric mean flows are demonstrated . in this section
, we also discuss a way in which gravity waves can transport momentum and heat horizontally
, from the dayside to nightside on tidally locked planets .
in [ sec : implications ] we discuss the implications of our work for current extrasolar planet atmospheric modeling work .
we conclude in [ sec : conclusion ] .
the dynamics of a linear gravity wave is described by the _ taylor goldstein equation _ ( tge ) .
this equation is derived from the full , 3-d hydrodynamics equations @xcite . in this work ,
we restrict the description to two dimensions and neglect rotation : [ hydroeqn ] @xmath7 where @xmath8 ; @xmath9 is the flow in the horizontal and vertical directions @xmath10 , respectively ; @xmath11 is the density ; @xmath12 ) ; @xmath13 is the pressure ; @xmath14 is the gravity ; @xmath15 is the potential temperature ; @xmath16 is the specific heat at constant pressure ; @xmath17 is the sensible temperature ; and @xmath18 is the net diabatic heating rate .
equations ( [ hydroeqn ] ) are supplemented with the ideal gas law : @xmath19 where @xmath20 is the specific gas constant .
note that @xmath21 where @xmath22 is some reference pressure ( here taken to be the pressure at the lower boundary of the model ) and @xmath23 ; @xmath15 is related to the entropy @xmath24 by @xmath25 .
the neglect of the third dimension and rotation requires that we restrict our analysis to waves with horizontal scale @xmath26 , where @xmath27 is the characteristic mean flow speed and @xmath28 is the planetary rotation rate .
this scale is adequate for all gravity waves , except for large - scale tides ( which has been recently considered by @xcite ) . as an example , @xmath29 m for hd209458b , based on @xmath27 in hot extrasolar planet simulations of @xcite .
the resulting value is approximately 1/10 of the planet s radius .
in addition , @xmath30 , @xmath20 , and @xmath16 are taken to be constant and @xmath18 is specified .
these restrictions do not mitigate the basic application and implications presented in this work .
however , for broader applications , relaxation of these and other restrictions will be considered in future work .
the variables in equations ( [ hydroeqn ] ) are all expanded as a small perturbation about a mean value , which is a function of height only : @xmath31 for the thermodynamic variables , we require that @xmath32 ; however , this is not required for the flow variables , @xmath33 and @xmath34 .
the mean state is assumed to be in hydrostatic balance , @xmath35 , and contains only horizontal flow so that @xmath36 .
we also assume the anelastic approximation ( e.g. , * ? ? ?
* ) , @xmath37 .
this implies the following : [ anelas ] @xmath38 where @xmath39^{1/2}$ ] is the brunt
visl frequency ; @xmath40 and @xmath41 are the density and pressure scale heights , respectively ; @xmath42 is the ratio of specific heats , with @xmath43 the specific heat at constant volume ; @xmath44 is the speed of sound ; and , @xmath45 is the vertical scale of the motion . with ( [ expan ] ) and ( [ anelas ] ) , we obtain [ linear ] @xmath46 where @xmath47 and @xmath48 .
since the coefficients in equation ( [ linear ] ) are independent of @xmath49 and @xmath50 , we assume perturbations of the form , @xmath51 where it is understood that the real part is to be taken .
this leads to the polarization equations : @xmath52 where @xmath53 is the forcing , @xmath54 is the constant ( possibly complex , see [ sec : saturation ] ) horizontal phase speed , and @xmath55 is the intrinsic phase speed .
now , transforming to a new variable so that the effect of decreasing density with height is compensated , [ eq : newvar ] @xmath56 where @xmath57 is @xmath58 at the bottom , we obtain the tge : @xmath59 here , @xmath60 is the index of refraction , and corresponds to the local vertical wavenumber .
it is given by @xmath61^{1/2}.\ ] ] in equation ( [ eq : indref ] ) we have used @xmath62 and the prime indicates differentiation with respect to @xmath58 .
the vertical structure of the perturbations , oscillating in @xmath58 , signify that we are dealing with _
internal _ waves . in equation ( [ eq : indref ] ) , the key terms contributing to @xmath63 are the first ( `` buoyancy '' ) term and the last ( `` non - hydrostatic '' ) term .
although the other three terms contribute , generally the buoyancy and non - hydrostatic terms control whether the wave propagates since the flow shear and curvature are small and the scale height is large in practice . for large wavelength waves , the waves are hydrostatic and the non - hydrostatic term is small ; then , the buoyancy term dominates .
in these cases , as long as the atmosphere is stratified ( i.e. , @xmath64 ) and @xmath65 , the wave will propagate vertically .
however , for shorter , non - hydrostatic waves , it is possible that @xmath66 ^ 2 $ ] . in these cases , @xmath67 is imaginary and the wave does not propagate .
this situation is discussed in more detail in [ subsec : wbi ] .
equation ( [ eq : tge ] ) can be thought of as a driven harmonic oscillator equation .
when @xmath67 is real and constant , its solution is a simple sinusoid .
since the transformation , equation ( [ eq : newvar ] ) , compensates for the exponential decay of density with height , @xmath34 grows ( decays ) rapidly with height ( depth ) .
here , we have dropped the tilde overscript .
from hereon we drop all tilde overscripts and `` 1 '' subscripts for notational clarity .
the growth can be clearly seen in figure [ fig : noshear ] , along with the corresponding constant stress ( vertical transport of horizontal momentum ) when there is no dissipation ( see [ sec : saturation ] ) .
if the temperature or the flow varies with height , @xmath67 is a function of height @xmath58 .
if @xmath67 varies slowly , the wkb approximation @xcite can then be used to obtain @xmath68 here , @xmath69^{1/2}$ ] . as in the constant @xmath67 case ,
the vertical perturbation velocity is wave - like with upwardly and downwardly propagating components ; the amplitude of the upward component grows with height and the downward component decays with depth .
however , when @xmath67 varies rapidly , the solution must be obtained numerically . at the boundaries we use the radiation condition , selecting the upwardly propagating solution at the top boundary @xmath70 and the downwardly propagating solution at the lower boundary @xmath71 .
this is achieved using the condition , @xmath72 here , @xmath73 , depending on the signs of the horizontal and intrinsic phase speeds and on whether the condition is at the upper or lower boundary .
note that condition ( [ eq : bc ] ) requires the wkb approximation to be valid at the boundaries .
for example , the boundaries can not be critical layers , regions where @xmath74 , since @xmath75 approaching such a layer and the wkb approximation ceases to be valid .
critical layers will be discussed further below .
as alluded to above , we use gaussian elimination @xcite in this work to numerically solve the tge . in the cases presented here , 3000
equally spaced levels are employed to solve for @xmath34 .
we have checked that this resolution is adequate by performing calculations with 10 , 000 levels and verifying convergence at the higher resolution .
extensive validations against known analytic solutions , where they exist , have also been performed .
the solution to the tge is a wave in the vertical velocity perturbation .
this can be related to the horizontal and temperature ( sensible and potential ) perturbations .
understanding these are essential for parameterizing the saturation process in the full non - linear situation and in general circulation models .
the perturbation quantities can be obtained from the polarization equations @xcite : [ eq : polar ] @xmath76 \\
\theta\ & = & \ -\frac{i}{k}\left[\frac{\theta_0 n^2 } { g \left(c - u_0\right)}\right ] w \\ t\ & = & \ -\frac{i}{k}\left[\frac{t_0 n^2 } { g \left(c - u_0\right)}\right ] w\ , , \end{aligned}\ ] ] where we have introduced the geopotential perturbation function @xmath77 ( @xmath78 ) .
equation ( [ eq : polar]a ) gives the relationship between the vertical and horizontal velocity perturbations .
note that the amplitude of @xmath33 is larger than the amplitude of @xmath34 , as it is scaled by @xmath79 .
note also that the geopotential ( pressure ) perturbation varies with the background flow via the dependence on the intrinsic phase speed , according to equation ( [ eq : polar]b ) .
equations ( [ eq : polar]c ) and ( [ eq : polar]d ) shows that the potential and sensible temperature perturbations , respectively , are both @xmath80 out of phase with the vertical velocity perturbation .
but , the phase between @xmath33 and @xmath34 varies locally through their dependence on the background .
gravity waves are an efficient means of transporting both momentum and energy .
the ( perturbation ) momentum and energy fluxes are simply obtained from equations ( [ eq : polar ] ) : [ eq : flux ] @xmath81 in equations ( [ eq : flux ] ) , @xmath82 is the vertical flux of horizontal momentum ( or , the wave stress ) ; @xmath83 and @xmath84 are the horizontal and vertical fluxes of energy , respectively ; and , the overline indicates an average over a wavelength , @xmath85 where @xmath86 and @xmath87 are arbitrary complex functions and `` * '' denotes the complex conjugate .
note that the energy fluxes depend on the background flow through their dependence on @xmath88 and @xmath77 .
however , as can be seen in figures [ fig : noshear ] and [ fig : excritlayer ] , the wave stress remains constant , away from the forcing and damping regions
saturation regions and critical layers .
this is in accordance with the second eliassen
palm theorem @xcite , which expresses non - interaction of the disturbance in the absence of dissipation and forcing . in figure
[ fig : excritlayer ] , it is important to note that , where the fluxes are changing , the wave is interacting with the background flow in those regions .
this should be contrasted with the behavior illustrated in figure [ fig : noshear ] , where the stress is not changing .
changes in the wave stress cause accelerations to the mean flow .
correspondingly , changes in the energy fluxes cause the temperature of the region to change .
the rates of these changes are given by [ eq : rate ] @xmath89 in this case , the wave causes the background flow to accelerate from 350 m s@xmath4 towards 500 m s@xmath4 . in the remainder of the paper , when speaking of vertically propagating waves , we consider only waves that propagate upwards from the region of excitation .
however , it must be remembered that downward propagating waves are also generated . on a giant planet without a solid surface , those waves may not be reflected or absorbed .
they can continue to penetrate downward until they encounter a critical level or a convective region . or ,
they are simply dissipated since the amplitudes of the downwardly propagating waves decrease exponentially , as already discussed ( and as also can be seen in figure [ fig : excritlayer ] ) .
the energy and momentum fluxes are linked by the first eliassen
palm theorem @xcite : @xmath90 which can be derived from equations ( [ eq : polar ] ) and ( [ eq : flux ] ) . for downwardly propagating waves
, we have @xmath91 ; and , therefore , via equations ( [ eq : ep1 ] ) and ( [ eq : rate]a ) , we see that for these waves the deposition of momentum also leads to acceleration of the flow toward the phase speed of the wave .
downwardly propagating planetary scale gravity waves ( i.e. , thermally excited tides ) are considered by @xcite . the present work concerns inviscid , linear , monochromatic waves .
such waves are infinite in extent and , in principle , can grow without limit when they propagate upward .
this is obviously not physical . in reality , such waves become unstable and saturate .
the saturation process can be treated by introducing a correction to the solution according to the physical criterion , @xmath92 ensuring that a wave does not become convectively unstable and remains at neutral stability .
when the saturation region is identified , @xmath15 is adjusted so that neutral stability is maintained .
this new value for @xmath15 is then used in equation ( [ eq : polar]c ) to obtain the corresponding @xmath34 in the saturated wave .
this value of @xmath34 is then used in equations ( [ eq : polar]a , b , c ) , ( [ eq : flux ] ) and ( [ eq : rate ] ) .
this saturation condition acquires the simple form , @xmath93 in situations where the wkb approximation is valid @xcite . as can be seen from the definitions of @xmath94 , @xmath95 , and @xmath15 , the brunt
visl frequency can be written as @xmath96^{1/2}.\ ] ] since @xmath0 and @xmath16 are essentially constants in the modeled height range , @xmath94 depends only on the temperature profile @xmath97 . for isothermal regions ,
@xmath94 is a constant . in general , the fractional change of @xmath17 with height is small compared to @xmath98 throughout the modeled region .
hence , @xmath94 is nearly constant in the entire domain , with a value that is roughly @xmath99 s@xmath4 .
the brunt
visl frequency for our atmospheric profile is shown in figure [ fig : flowtempprof ] .
further discussion of the effects of variation in @xmath94 are presented in
[ subsubsec : background ] .
if the background flow contains shear , it is possible for the wave to encounter a critical layer at some height . at the critical level ,
where @xmath100 , the tge becomes singular .
however , the equation can be solved using the method of frobenius @xcite , from which it is seen that the wave is drastically attenuated by the critical layer @xcite .
the amount of attenuation depends on the richardson number @xmath101 of the flow , @xmath102 figure [ fig : excritlayer ] illustrates a critical layer encountered by a wave . if the wave stress has magnitude @xmath82 below the critical layer , it then has magnitude , @xmath103^{1/2 } \right\}$ ] , after the encounter with the critical layer @xcite .
this is a substantial amount of attenuation .
for example , in our model atmosphere of figure [ fig : excritlayer ] , @xmath104 .
hence , the wave is essentially completely dissipated at the critical level , with the wave stress falling to practically zero and the momentum deposited into the mean flow .
note that , during its approach to the critical layer , the wave actually saturates and deposits momentum and energy over a finite layer ( cf . ,
figure [ fig : excritlayer ] ) .
it is important to note that while a wave saturates when approaching a critical layer , the presence of a critical layer is _ not _ required for saturation .
saturation is a general process referring to wave dissipation by many different mechanisms
radiation , conduction , breaking , and turbulence .
numerically , when a critical layer is present , we lift the phase speed from the real axis by adding a small imaginary component : @xmath105 , where @xmath106 .
this introduces a small amount of linear damping and ensures that the neglected nonlinear terms do not dominate in the regions where waves become steep and eventually break . of course , adding damping causes the wave - stress to decrease with height and the second eliassen - palm theorem to be no longer valid .
however , the effect is small .
this can be seen in figure [ fig : excritlayer ] , where the wave stress falls negligibly over the layer from @xmath107 to @xmath108 .
our aim in the present paper is to demonstrate several properties of gravity waves likely to be important for hot extrasolar planets . for this
, we choose hd209458b as a paradigm planet .
this planet has been the focus of many theoretical and observational studies , and it is expected to be generic with respect to the properties discussed here .
the physical parameters that characterize the planet s atmosphere are given in table [ tab : params ] .
llr + specific gas constant & @xmath20 & 3523 j kg@xmath4 k@xmath4 + specific heat at constant pressure & @xmath16 & 12300 j kg@xmath4 k@xmath4 + acceleration due to gravity & @xmath0 & 10 m s@xmath5 + rotation rate & @xmath28 & 2.08@xmath109 s@xmath4 + + [ tab : params ] in a stratified atmosphere , gravity waves are readily generated by many mechanisms thermal and mechanical such as the absorption of stellar radiation , convective release of latent heat , storms , flows over topography and coherent localized `` heat islands '' , and detonation by impacts . in this work ,
we consider small- and meso - scale thermally excited waves , rather than large - scale waves , as already discussed .
the horizontal wavelengths used in this work are 2500 km or less .
this is a reasonable range since it is well within the observed range of gravity waves on jupiter @xcite .
although not unimportant , we do not dwell on the precise nature of the source of the excitations .
the main interest here is in the propagation and deposition of momentum and energy .
the forcing in equation ( [ eq : tge ] ) is simply represented as a gaussian , modified so that it is zero beyond two half - widths above and below the center .
the center is located at @xmath110 above the bottom of the domain .
the half - width is 75 km , or @xmath111 .
the forcing location and width are chosen so that the vertical scale of the forcing is less than the vertical wavelength of the waves we present here .
we have extensively explored the effects of varying the location , width , and strength of forcing and present this case to illustrate several important points . not surprisingly , the dynamics do depend on the chosen parameter values , but the dependence is broadly predictable .
for example , if the forcing scale is much larger than the vertical wavelength , then only a very small amplitude wave is emitted from the forcing region , due to cancellations .
the heating rate , @xmath112 , is set to 10@xmath3 k s@xmath4 .
note that this is a modest value .
the forcing corresponds to roughly 300 k per rotation of the planet .
this is compared to @xmath6100 k per rotation at the chosen location in the circulation model of @xcite .
a forcing of @xmath61100 k per rotation , for a similar latitude - longitude location on the planet ( see [ subsubsec : background ] ) , is used in @xcite .
the latter value implies that , in the absence of motion , the location on the planet will cool completely in one rotation of the planet .
we stress that locally i.e .
, scales far below the grid scale of the current circulation models the forcing could actually be much stronger .
the actual value is presently uncertain and likely to be spatially and temporally variable over the planet . to provide a context , for the earth
the heating rate is @xmath113 k per day ( 1 day = 0.29 rotation of hd209458b ) over large areas ; but , locally , at the tops of low clouds on the earth the rate can be up to @xmath114 k per day @xcite .
figure [ fig : flowtempprof ] shows the mean flow and temperature profiles used to obtain much of the results presented in this section . the lower part of both profiles approximately the lower six scale heights is taken from global circulation simulations of the hot extrasolar giant planet hd209458b by @xcite , using the ncar community atmosphere model @xcite .
the profile is from a point near the equator , slightly away from the substellar point ( 70@xmath115e,10@xmath115n ) .
this point was chosen as it is within the equatorial jet and the coriolis parameter @xmath116 is not large .
the temperature profile generally increases with height over the lowest 4 scale heights and then becomes isothermal .
fortuitously , this provides an opportunity for a loose validation of our model : it is very similar to the temperature structure observed by the galileo probe in the same region of jupiter s atmosphere @xcite .
we extend the profile by keeping the temperature isothermal through the planet s stratosphere and having the thermosphere ( beginning of the temperature inversion near the top ) start between 12 and 13 scale heights at @xmath117 bar .
this profile has a richardson number of at least 3.4 , giving an attenuation of @xmath118 .
hence , any critical layer can be considered to fully dissipate the wave .
the chosen flow profile has two local flow velocity maxima .
the upper maximum is extended into a jet with a peak at @xmath119 .
this is similar to the structure of the jet in jupiter s stratosphere proposed by @xcite , which has a peak velocity between the 10@xmath5 and 10@xmath3 bar levels .
we then extend the profile further upwards without shear .
note , this is in keeping with the lower boundary of the model of @xcite .
the structure is somewhat different than those of @xcite , where there is just one jet with the peak approximately located at the 10@xmath4 bar level .
the peak flow speed is also much greater in those studies at 4 or 5 km s@xmath4 .
it is important to note , however , that these differences do not change qualitatively the basic points we are making in this paper .
the brunt
visl frequency profile @xmath120 is also shown in figure [ fig : flowtempprof ] .
as already discussed , the profile does not vary much over the whole domain : the maximum brunt
visl frequency is just 1.2 times the minimum value .
therefore , @xmath94 does not contribute much to the variation of the index of diffraction @xmath67 .
the main contributor to the variation of @xmath67 is the variation of the intrinsic phase speed , which is derived from the large variation of flow speeds .
this should be compared to the analogous terrestrial situation , where the range of flow speeds is much lower .
this allows @xmath94 to have a larger effect on the variation of @xmath67 on the earth .
figure [ fig : realcritlayer ] shows a gravity wave encountering a critical level in the upper jet in figure [ fig : flowtempprof ] .
the wave has a horizontal wavelength , @xmath121 , of 2500 km and @xmath122 m s@xmath4 .
here , since @xmath123 as the wave approaches the critical layer from below , the momentum deposited in the mean flow causes the mean flow to accelerate .
this acceleration peaks at over 250 m s@xmath4 per rotation .
this is large enough to double the flow speed at this layer in @xmath62 rotations a significant effect .
the effect is large enough to require its inclusion in any simulation of the atmospheric circulation @xcite .
the waves encountering critical layers are dissipated .
therefore , a flow with a range of flow speeds dissipates all gravity waves with phase speeds within this range .
that is , a spectrum of gravity waves is prevented from propagating high into the atmosphere .
there are other , secondary effects at the critical layer that may also affect the mean flow ; but , they are not modeled here .
they will form the basis of future work .
for example , the deposition of energy into the flow at the critical layer may well lead to the generation of new gravity waves , which then may propagate further , partly mitigating the filtering effect .
figure [ fig : realsaturate ] shows an example of a gravity wave saturating in the atmosphere of figure [ fig : flowtempprof ] . in general , it is possible that a wave may not encounter a critical level as it propagates upward .
however , such a wave then travels higher into the atmosphere , where it grows large and , if unabated , eventually breaks or suffers dissipation at higher altitudes .
although both momentum and energy are deposited in this case , we focus here on the effects on the temperature field .
the wave launched in figure [ fig : realsaturate ] has @xmath124 m s@xmath4 ( i.e. , westward ) .
the horizontal wavelength remains at 2500 km , as in the critical layer example of figure [ fig : realcritlayer ] .
the vertical velocity perturbation grows with height .
therefore , so do the zonal velocity perturbation @xmath33 and the potential temperature perturbation @xmath15 , as expected from equation ( [ eq : polar ] ) . in this case , the wave saturates near the top of the jet , where condition ( [ eq : satcrit1 ] ) is exactly satisfied . here , condition ( [ eq : satcrit2 ] ) is approximately satisfied since the zonal perturbation velocity and the intrinsic phase speed ( @xmath125 ) are both approximately 122 m s@xmath4 .
the saturation deposits energy that causes the atmosphere there to heat up .
the heating is significant , peaking at @xmath126 k per rotation a 5% change in one rotation . in the absence of other effects
, the ambient temperature can be doubled in about 20 planetary rotations ( or orbits , assuming 1:1 spin - orbit synchronization ) .
a wave that has a phase speed greater than the maximum flow speed will not encounter a critical layer .
those with phase speeds close to , but still above , the maximum flow speed will , in general , saturate in the regions just below the maximum flow , as the intrinsic flow speed will be small in that region .
similarly , waves with phase speeds just less than the minimum flow will saturate as well . in this way the filtering effect discussed above is extended beyond those waves with phase speeds equal to flow speeds .
the main effect of these filtered waves on the flow will be lower in the atmosphere . in the profile given in figure [ fig : flowtempprof ] , where the waves emanate from the @xmath110 level , this means that the upper layers of the lower jet will be slowed by gravity waves whereas the lower levels of the upper jet will be accelerated
. those waves that do not dissipate will be able to propagate into the upper atmosphere depositing their momentum and heat there . here
the changes to the flow can be very large .
for example , the wave shown in figure [ fig : realsaturate ] causes a deceleration of up to 6.8 km s@xmath4 rotation@xmath4 as it saturates .
this clearly dominates the flow at this level . moving the location of the wave origin down does not change the basic behavior in the qualitative sense . however , the amplitudes are much larger , compared with the case when the wave originates higher up in altitude .
thus , the possibility exists for stronger effects due to gravity wave interaction with the background .
this is illustrated in figures [ fig : flowtempprofdeep ] and [ fig : realsaturatedeep ] , which should be compared with figures [ fig : flowtempprof ] and [ fig : realsaturate ] .
here , we have extended the profiles downwards .
the wave is still launched at @xmath110 , but this is now deeper in the atmosphere .
the wave again has a phase speed of @xmath127 m s@xmath4 , and the horizontal wavelength remains at 2500 km .
this wave also saturates near the top of the upper jet , where the energy deposition into the mean flow causes the atmosphere there to heat up .
note that the region of heating is lower than when the wave originates higher up , as in figure [ fig : realsaturatedeep ] .
the heating is significant , peaking at @xmath128 k rotation@xmath4 .
the ambient temperature can be doubled in approximately half of a planetary rotation . in a more realistic scenario , dissipation which we have not included in our model
will likely reduce the heating rate .
so far we have been focusing on the vertical transport of momentum and energy by gravity waves . however , the waves can also transport momentum and energy _ horizontally _ ( cf . ,
[ subsubsec : ducting ] ) .
substituting @xmath129 into the the index of refraction , equation ( [ eq : indref ] ) , and rearranging gives the dispersion relation for gravity waves .
here , we consider the case where @xmath130 .
we will examine cases where @xmath131 in the sections following this one .
when there is no background mean flow , the dispersion relation simplifies to @xmath132^{1/2}}.\ ] ] we can then use the definitions , [ eq : grpvel ] @xmath133 to obtain the group velocities .
they are : [ eq : noflowgrpvel ] @xmath134}{\left[k^2 + m^2 + 1/(4h^2_{\rho})\right]^{3/2 } } \\
w_g & \ = \ & \pm\frac{n k m}{\left[k^2 + m^2 + 1/(4h^2_{\rho})\right]^{3/2}}.\end{aligned}\ ] ] thus , for propagating waves ( i.e. , waves for which @xmath67 is real ) , @xmath135 .
therefore , gravity waves always propagate obliquely and can not strictly propagate vertically when there is no background flow . from equations ( [ eq : noflowgrpvel ] )
we see that @xmath136 , the angle of propagation with respect to the horizontal , is given by @xmath137 since @xmath138 is nearly constant , with a value just under 500 km , @xmath139 varies with @xmath140 for @xmath141 m@xmath4 .
this gives rise to refraction .
as a wave propagates into a region of higher @xmath67 , it bends to a more horizontal path .
note that , as we are here considering a region with no flow , increasing @xmath67 is essentially equivalent to increasing @xmath94 .
as shown in figure [ fig : flowtempprof ] , @xmath94 increases in the thermosphere and so the paths of waves in this region will bend towards the horizontal even though the flow is small . from equation ( [ eq : wkbsoln ] ) we can see that in regions where @xmath67 is imaginary , the wave is evanescent : its amplitude decays towards zero and therefore it does not propagate vertically .
this can occur when @xmath142 ( i.e. , when the atmosphere is convectively unstable ) .
but , it can also occur for non - hydrostatic waves when the buoyancy term becomes dominated by the non - hydrostatic term .
in addition , when the index of refraction changes between layers , the wave is reflected at the boundary .
the amount of reflection is given by the magnitude of the coefficient of reflection @xmath143 , where @xmath144 in equation ( [ eq : coeff_reflect ] ) , @xmath145 and @xmath146 are the indices of refraction in two adjacent layers .
when @xmath146 is imaginary , total reflection occurs and the wave is evanescent in that region and its amplitude decays to zero .
however a region of propagation can exist between two regions of evanescence .
this readily occurs for jets , where the intrinsic phase speed varies enough to allow the hydrostatic term to dominate in some regions and not in others .
the region can also occur through variations of the brunt
visl frequency . in figure
[ fig : realtrapped ] we see a wave that is trapped in the relatively quiescent region between @xmath107 and @xmath147 . outside this region ,
the value of re@xmath148 is small , or zero .
trapped in this region the wave is able to interact with itself and , under appropriate conditions , resonate .
this is another mechanism via which waves may be filtered out by the flow ; and so the waves do not reach high altitudes at which saturation can occur . however , in this case , a trapped wave does not directly interact with a low level flow that changes its characteristics . indeed , between the two reflection layers the wave can propagate horizontally even in the absence of any flow , using the refractive mechanism described above .
as long as the layers do not allow much leakage , it is possible for a trapped wave to cover large horizontal distances transporting momentum and heat zonally ( east - west direction ) . as well as being trapped in relatively quiescent regions ,
it is possible for waves to be trapped in a jet . as alluded to above , it is possible for such a wave to travel within the region of trapping , which is known as a duct or a waveguide . in figure
[ fig : realducted ] , a non - hydrostatic wave with @xmath149 m s@xmath4 is trapped within the jet ( located at @xmath65 mbar level ) in our model atmosphere .
note the small values of re@xmath148 outside the jet . in this case , the flows are significant .
therefore , we use the full dispersion relation , + @xmath150 \\ \pm 2\left[h^2_{\rho}k^2\left(\left(1 + 4h^2_{\rho}\left(k^2+m^2\right)\right)n^2 + \left(u_0'+h_{\rho}u_0''\right)^2\right)\right]^{1/2}}{1 + 4h^2_{\rho}\left(k^2+m^2\right ) } \ , , \ ] ] to develop expressions for the group velocities . however , the expressions obtained are large and rather unilluminating
. they can be simplified by assuming that @xmath151 and @xmath152 are small .
this is realistic since the shear is of the order of @xmath153 s@xmath4 and @xmath152 of the order @xmath154
m@xmath4 s@xmath4 .
this is small compared with the other terms in the expressions .
this then gives @xmath155}{\left[k^2 + m^2 + 1/(4h^2_{\rho})\right]^{3/2 } } \\
w_g & \ = \ & \pm\frac{n k m}{\left[k^2 + m^2 + 1/(4h^2_{\rho})\right]^{3/2 } } \end{aligned}\ ] ] from these expressions , we can see that @xmath156 follows @xmath157 as this is the larger term on the right hand side of equation ( [ eq : grphorvelflow ] ) in our model atmosphere . in figure
[ fig : realducted ] , the values of @xmath156 and @xmath158 are shown .
note that in the center of the duct @xmath158 is very small while @xmath156 is large , so that energy is transported along the flow . at the the top and bottom of the duct the vertical group velocity increases , while the horizontal group velocity falls .
therefore , propagation here is nearly vertical . in figure
[ fig : realducted ] , we show @xmath158 as positive , however this is only for the upward propagation of energy , at the top of the duct the wave is reflected and the vertical component becomes negative .
this keeps the wave within the jet .
the ray path followed by the wave , before the reflection , is shown in figure [ fig : realductedpath ] .
the wave can travel large distances in this duct ; but , eventually , the wave will either escape the duct or be dissipated .
the range of speeds in the jet may change so that a critical level for the wave is created .
the wave will then be reabsorbed into the flow . alternatively , if the flow or brunt visl frequency outside the jet changes so that the buoyancy term is no longer small and is dominated by the non - hydrostatic term , then the reflection is no longer total and the wave will then leak out of the duct .
this can , for example , happen when propagating into a colder region , assuming the lapse rate remains constant . as can be seen from equation ( [ eq : bvf ] )
a fall in temperature with constant lapse rate will cause an increase in the brunt
visl frequency and thus an increase in the buoyancy term .
this may occur very far from the original region of wave excitation .
indeed , in the example given , it is possible to envisage jets ducting waves and so transporting energy from the dayside of a tidally locked planet to the colder nightside where the waves escape the jet and propagate away from the duct before dissipating .
the effects of gravity waves discussed in this work on the larger - scale circulation must be parameterized in global models because the spatial resolution both horizontal and vertical required to model them is currently prohibitive .
the waves important to the large - scale extrasolar planet atmospheric circulation have horizontal length scales ranging from approximately @xmath159 m to @xmath160 m and vertical wavelengths as small as @xmath161 m. waves with periods of few hours can carry significant momentum and energy flux vertically , but the sources of these waves include processes that are not included or resolvable in current circulation models .
the difficulty with representing gravity waves in gcms exists even for the gcms of the earth .
for example , the parameterization for convection is not aimed at producing realistic gravity waves @xcite .
however , not representing gravity waves can affect the accuracy of the gcms .
the lack of gravity wave drag can lead to the overestimation of wind speeds , resulting in faster and narrower jets than observed @xcite .
further , gravity waves introduce turbulence with subsequent mixing and thermal transport @xcite .
this leads to greater homogenization of the atmosphere with a reduction in , for example , temperature gradients .
gravity waves also interact with planetary waves , playing a role in important transient phenomena ( such as sudden stratospheric warming ) . in the absence of gravity waves ,
these phenomena are not accurately modeled @xcite .
there are many parametrization schemes currently incorporated or proposed for general circulation modeling @xcite . in all of the schemes , the basic components are 1 ) specification of the characteristic of the waves at the source level , 2 ) wave propagation and evolution as a function of altitude , and 3 ) effects on and by the atmosphere .
all of them are essentially linear and one - dimensional , in that waves only propagate vertically and that only vertical variation in the background influence the propagation . as seen in this work , linear theory still requires information such as the wave s phase speed @xmath162 and wavenumber @xmath163 , for example .
a more complete theory would need spatial and temporal spectral information .
intermittency is another crucial feature that would need to be taken into account .
the primary differences in various schemes pertain to the treatment of nonlinearity and specificity of wave dissipation mechanisms . currently , all global circulation models of hot jupiters suggest the presence of a low number of zonal jets . however , all the models do not have the resolution required to adequately resolve gravity waves and are subject to all of the limitations described above .
this issue has been previously raised by @xcite , in which they advocate caution against quantitative interpretation of current model results .
for example , without the inclusion of the wave effects discussed in this work , high jet speeds and precise eastward shift of putative `` hot spots '' can be questioned ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * )
gravity wave propagation and momentum and energy deposition are complicated by the environment in which the wave propagates .
for example , spatial variability of the background winds causes the wave to be refracted , reflected , focused , and ducted . additionally , temporal variability of the background winds cause the wave to alter its phase speed .
still further complications arise due to the wave s ability to generate turbulence , which can modify the source or serve as a secondary source , and the wave s interaction with the vortical ( rotational ) mode of the atmosphere .
many of these issues are as yet not well - understood and are currently areas of active research . in this work ,
we have emphasized only some of these issues .
we have shown that gravity waves propagate and transport momentum and heat in the atmospheres of hot extrasolar planets and that the waves play an important role in the atmosphere .
they modify the circulation through exerting accelerations on the mean flow whenever the wave encounters a critical level or saturates .
they also transport heat to the upper stratosphere and thermosphere , causing significant heating in these regions .
moreover , through ducting , they also provide a mechanism for transporting heat from the dayside of tidally synchronized planets . before relying on gcms for quantitative descriptions of hot extrasolar planet atmospheric circulations ,
further work needs to be performed to ensure that the effects of important sub - scale phenomena , such as the gravity waves discussed here , are accurately parameterized and included in the gcms .
the authors thank heidar thrastarson for generously sharing information from his simulation work on extrasolar planet atmospheric circulation .
we also thank orkan umurhan for helpful discussions and the anonymous referee for helpful suggestions .
is supported by the science and technology facilities council ( stfc ) .
j.y - k.c . is supported by nasa nng04gn82 g and stfc pp / e001858/1 grants . | we study the effects of gravity waves , or @xmath0-modes , on hot extrasolar planets .
these planets are expected to possess stably - stratified atmospheres , which support gravity waves . in this paper , we review the derivation of the equation that governs the linear dynamics of gravity waves and describe its application to a hot extrasolar planet , using hd209458b as a generic example .
we find that gravity waves can exhibit a wide range of behaviors , even for a single atmospheric profile .
the waves can significantly accelerate or decelerate the background mean flow , depending on the difference between the wave phase and mean flow speeds .
in addition , the waves can provide significant heating ( @xmath1 to @xmath2 k per planetary rotation ) , especially to the region of the atmosphere above about 10 scale heights from the excitation region .
furthermore , by propagating horizontally , gravity waves provide a mechanism for transporting momentum and heat from the dayside of a tidally locked planet to its nightside .
we discuss work that needs to be undertaken to incorporate these effects in current atmosphere models of extrasolar planets . |
let @xmath0 be an algebraic hypersurface in the algebraic torus @xmath1 , @xmath2 , i.e. @xmath3 where @xmath4 is the laurent polynomial .
recall that the _ newton polyhedron _ @xmath5 of @xmath6 is the convex hull in @xmath7 of @xmath8 .
let @xmath9 be the smooth toric variety associated to the fan @xmath10 , which is a refinement of the fan dual to the newton polyhedron @xmath11 .
we denote by @xmath12 the closure of @xmath0 in @xmath9 .
the polynomial @xmath6 is called _ non - singular for its newton polyhedron _ , if @xmath0 is smooth and for any face @xmath13 the truncation @xmath14 of @xmath6 to the face @xmath15 has non - vanishing jacobian at all @xmath16 : @xmath17 in accordance with singularity resolution theorem , cf .
@xcite , a generic polynomial @xmath6 is non - singular for its newton polyhedron and then @xmath18 is non - singular .
next we introduce the so - called _ logarithmic gauss map _ @xmath19 .
let @xmath20 denote the lie algebra of @xmath21 , which is identified with the tangent space of @xmath21 at the unit point @xmath22 .
for any point @xmath23 shift the tangent space @xmath24 by the torus multiplication ( with @xmath25 ) to a hyperplane @xmath26 , inducing a point in the projective space of the dual @xmath27 , which we define to be @xmath28 . in coordinates of @xmath21
the map @xmath29 is given by @xmath30 described in more geometric terms we have : let @xmath31 be a neighbourhood of a regular point @xmath32 on @xmath0 .
choose a branch of the logarithmic map ( restricted to @xmath33 ) @xmath34 , then the direction of the normal line at @xmath35 to transformed hypersurface @xmath36 has components @xmath37 .
this construction does not depend on the choice of the branch of @xmath38 . in @xcite , 3.2
, one can find the idea of a construction , how to extend @xmath29 in the non - singular case to a finite map .
@xmath39 having a finite map @xmath40 to a smooth variety , one can associate the _ ramification locus _ or the _ discriminant _ as image of the critical locus : @xmath41 , which is usually a hypersurface .
an analytic structure which is compatible with base chance was introduced by teissier , cf .
@xcite : the structure sheaf @xmath42 is defined to be the quotient by the @xmath43-th fitting ideal of @xmath44 . in local coordinates the defining equation is obtained as the ( classical ) discriminat of the polynomial , that generates the finite extension of the structure sheafs over an open affine subsets . from the well - known theorem of kouchnirenko , cf .
* th . 3 ) , mikhalkin obtains : [ lemma:1 ] if the polynomial @xmath6 is non - singular for its newton polyhedron , then the degree of @xmath45 is obtained as @xmath46 for later calculation we give a description of the logarithmic gauss map @xmath29 in local coordinates .
since @xmath0 is smooth , we may assume w.l.o.g .
@xmath47 locally , then a function @xmath48 , @xmath49 , exists , such that @xmath50 .
hence , @xmath51 and @xmath52 hold , and one obtains the formula @xmath53 then the fiber @xmath54 is given by the zeros of the local complete intersection ideal generate by @xmath6 and the 2-minors of @xmath55 i.e. ( in case of @xmath56 ) @xmath57 is defined by the complete intersection ideal @xmath58 where @xmath59 .
there are at most @xmath60 zeros in the torus by proposition 1 .
consider a rational function @xmath61 of @xmath62 complex variables and different laurent expansions @xmath63 of @xmath64 centered at @xmath65 .
the most natural way to describe these expansions uses the amoeba of polar hypersurface @xmath66 .
recall , that the _ amoeba _ @xmath67 of a toric hypersurface @xmath66 is the image of @xmath68 by the logarithmic map @xmath69 , @xmath70 the complement @xmath71 to the amoeba consists of a finite number of connected components @xmath72 , which are open and convex , cf .
* section 6.1 ) .
these components are characterized in the following proposition , which is a summary of propositions 2.5 , 2.6 in @xcite , theorem 10 and corollary 6 in ( * ? ? ?
* section i.5 ) .
[ thm:1 ] there exists an open subset @xmath73 in the set of polynomials with fixed newton polyhedron @xmath74 exists that satisfies the following property : if @xmath75 , then there is a bijection from the set of lattice points of @xmath76 to the set of components of @xmath77 , such that the normal cone @xmath78 to @xmath79 at the point @xmath80 is the recession cone of the component @xmath81 .
a recession cone is the maximal cone which can be put inside @xmath81 by a translation . if @xmath82 , the expected component @xmath81 may not exist for some non - vertice lattice points @xmath83 , because the associated laurent series below does not converge . given a component @xmath81 one obtains a laurent series of @xmath64 centered at @xmath65 using the term @xmath84 of @xmath6 as denominator in a corresponding geometric progression @xmath85 the set @xmath86 contains the domain of convergence for this laurent series .
the support of expansion ( [ eq:2 ] ) is the minimal cone @xmath87 , which after a translation by @xmath80 contains the face @xmath15 of @xmath79 , which has @xmath80 as interior point .
a non - zero vector @xmath88 defines a so - called _ diagonal subsequence _
@xmath89 of the set of coefficients of expansion ( [ eq:1 ] ) .
we will discuss its asymptotic in the next section .
the set of critical values of the map @xmath90 restricted to @xmath68 is called the _ contour _ @xmath91 of the amoeba @xmath67 ( see @xcite ) .
the contour is closely related to the logarithmic gauss map @xmath29 .
recall lemma 3 from @xcite .
[ lm:2 ] the preimage of the real points under the logarithmic gauss map is mapped by @xmath90 to the contour : @xmath92 let @xmath32 be a regular point on @xmath93 and @xmath33 its neighbourhood . since the map @xmath94 is a composition of @xmath95 and the projection @xmath96 , the point @xmath32 is critical for @xmath94 if the projection @xmath97 is not surjective at @xmath32 .
a fiber @xmath98 of the tangent bundle of the image by @xmath38 of the hypersurface @xmath68 is the hyperplane @xmath99 for real @xmath100 the projection @xmath101 is not surjective . if @xmath100 is not real one can consider @xmath102 as a system of linear equations with fixed real part @xmath103 and solve it with respect to @xmath104 hence , @xmath32 is not critical for @xmath105 therefore , the contour @xmath91 can be computed as the @xmath90-image of the zeros of the ideal @xmath106 where @xmath107 runs through all real points @xmath108 , ( here w.l.o.g @xmath109 ) .
consider the function @xmath110 introduce the _ phase function _
@xmath111 , later we show that it is indeed a phase function of some oscilllating integral .
denote by @xmath112 the set of critical ( or stationary ) points of function @xmath113 .
it coincides with the preimage of the logarithmic gauss map @xmath29 : [ thm:3 ] the relative critical locus of @xmath114 coincides with the graph of @xmath29 : @xmath115 assume @xmath116 , then we use local coordinates @xmath117 on @xmath68 , and consider the function @xmath48 such that @xmath50 .
we obtain @xmath118 up to a non - zero constant multiple the components of the gradient @xmath119 together with the defining polymonial @xmath4 of @xmath68 give us the defining ideal ( [ eq : cv ] ) of the fiber of @xmath120 by the logarithmetic gauss map @xmath40 .
the last statement shows us that the @xmath90-image of @xmath121 is contained in the contour @xmath91 of the amoeba @xmath67 , and the tangent hyperplane to @xmath91 at a point @xmath122 , @xmath123 , has normal vector @xmath124 .
another consequence of the above formula concerns the connection between the singularities in the fibers of the phase function and the fibers of the logarithmic gauss map : [ thm:4 ] let @xmath125 be a point of the graph of @xmath29 , then the jacobian matrix of @xmath29 at @xmath126 coincides with the hesse matrix of @xmath127 at @xmath126 up to multiplication with a regular constant diagonal matrix @xmath128 : @xmath129 as before we assume @xmath130 and use local coordinates @xmath117 . from ( [ eq : a ] ) we obtain the entries of the jacobian matrix @xmath131 of the map
@xmath29 @xmath132 where @xmath133 is the kronecker symbol . moreover , @xmath134 holds for the second derivatives of @xmath135 . since @xmath136 at a critical point @xmath126 of @xmath127
, we obtain the statement by putting the @xmath137-th entry of @xmath128 to be @xmath138 .
we obtain as corollary of the last proposition that for directions @xmath139 outside the ramification locus of the logarithmic gauss map @xmath29 the phase function @xmath113 has only morse critical points .
the logaritimic gauss map @xmath29 is umramified at @xmath140 iff the phase function @xmath113 has only morse critical points , e.g. non - degenerated singularities . the map @xmath29 is not ramified over @xmath120 iff its jacobian has full rank at all points of the fiber @xmath57 . from proposition 3
we get @xmath141 hence , the jacobian determinant does not vanish iff the hessian is not zero at corresponding points . next we want to discuss degenerated critical points of the phase function . by mather - yao type theorem the @xmath142-class ( right - equivalence ) of an analytic function @xmath143 at an isolated critical point @xmath144 is equivalent to the isomorphy type of the milnor algebra @xmath145 , but as @xmath146$]-algebra , the action of @xmath147 on @xmath148 induced by multiplication with @xmath149 , cf .
@xcite . the isomorphy type of the associated singularity @xmath150 , i.e. the @xmath151-class ( contact - equivalence ) of @xmath152 , is equivalent to the isomorphy class of the tjurina algebra @xmath153 itself , cf .
obviously , these equivalence classes coincide for quasi - homogeneous functions ( because @xmath154 , @xmath155 , @xmath156 ) .
the milnor algebra of the phase function at @xmath126 coincides with the local algebra of the fiber @xmath157 at @xmath126 .
if @xmath125 , denote by @xmath158 the milnor algebra of the function @xmath159 at @xmath126 , then we have @xmath160 by proposition 3 the germs coincide : @xmath161 .
the algebra of the critical locus is the milnor algebra of @xmath159 . by proposition 4
the jacobi determinant of @xmath29 at @xmath126 equals up to a constant multiple to the hessian of @xmath127 at @xmath126 , which generates the annulator of the maximal ideal in the local complete intersection algebra @xmath158 .
a function @xmath162 with isolated critical point is called _ almost quasihomogeneous _ , if @xmath163 .
this is equivalent to @xmath164 .
assume that the singularities in a fiber of a phase function are quasihomogeneous or almost quasihomogeneous , then in spite of mather - yao type theorems these singularities are determined by the fiber germs of the logarithmic gauss map , because @xmath165 or @xmath166 , respectively .
note , that all simple or unimodal critical points belong to these singularities .
the singularities of a phase function on their part determine the asymptotic of corresponding oscillating intergrals .
all degenerated critical points are lying over the singularities of the discriminant @xmath167 of @xmath29 .
many results could be found concerning the connection between singularites of discrimant and singularities in the fiber .
we try to discuss some consequnces with respect to our setting .
the finite map @xmath45 can be considered as family over @xmath168 of complete intersections ( of relative dimension zero ) .
let @xmath169 be a germ of an isolated complete intersection singularity , let @xmath170 its versal family with discriminant @xmath171 , the singularity of the discriminat @xmath172 determines the special fiber @xmath169 up to isomorphy by a result of wirthmuller , c.f @xcite . if @xmath173 the multiplicity of the discriminant fulfills @xmath174 , as a consequnce of @xcite , for instance .
this is globalized straight forward .
let @xmath175 be a finite morphism with discriminant @xmath171 and each @xmath176 is a complete intersection , then holds : @xmath177 moreover , equality holds at @xmath178 , if @xmath40 induces a versal deformation of @xmath176 .
the local branches of @xmath128 at @xmath179 are corresponding to the discriminants @xmath180 of the germs @xmath181 , hence the multiplicities of @xmath180 add up to the multiplicity of @xmath128 .
any family is locally induced from a versal one , hence the discriminant is induced by base chance from the discriminant of the versal family and its multiplicity can not become smaller .
note , versality is an open property and corresponds to some kind of _ stability _ in the sence of mather .
it is not clear for us , whether ( or under which additional assumtions ) the logarithmic gauss map @xmath29 for a generic function @xmath4 with fixed newton polyhedron @xmath182 has this stability property .
it holds in all computed examples .
but , an answer needs further investigation . inspecting the classification of hypersurface singularities
we get the types of possible critical points for small multiplicities of the discriminant , which are listed in the following corollary . given a laurent polynomial @xmath4 , non - singular with respect to its newton polyhedron , and let @xmath40 be the corresponding logarithmic gauss map with discriminant @xmath183 .
let @xmath184 , then the following configurations are met for the fiber @xmath185 , respectively for the collection of critical points of the phase function @xmath186 : * @xmath187 : @xmath188 has exactly one point @xmath189 of multiplicity 2 , @xmath113 has non - degenerated critical point and one @xmath190-singularity at @xmath189 .
* @xmath191 : @xmath188 either one point of multiplicity 3 or at most two points of multiplicity 2 , @xmath113 has at most one @xmath192 or two @xmath190-points .
* @xmath193 : besides @xmath194 can occur the following collections of critical points of @xmath113 : one @xmath195 or one @xmath196 or a combination @xmath197 with @xmath198 . *
@xmath199 : type of critical set of @xmath113 : only ( simple ) ade - critical points can occur @xmath200 such that @xmath201 and @xmath202 * @xmath203 : all critical points are quasihomogeneous ( and simple or unimodal ) . *
@xmath204 : all critical points are almost quasihomogeneous ( and simple or unimodal ) . the first critical point , which is not almost quasihomogeneous are the bimodal exceptional singularities with smallest milnor number @xmath205 of type @xmath206 or @xmath207 , cf.@xcite .
they can occur only at multiplicity @xmath208 .
in this section we return to laurent series ( [ eq:1 ] ) converging in @xmath209 .
we explane the residue asymptotics formula for its diagonal coefficient in the direction @xmath210 .
recall , that the laurent series coefficient can be represented in the form @xmath211 where @xmath212 and the cycle @xmath213 is @xmath62-dimensional real torus @xmath214 .
the direction @xmath107 induces series of diagonal coefficients @xmath215 we may assume that the point @xmath216 generates a line @xmath217 , which is transversal to the boundary @xmath218 and intersects it at a point @xmath219 and the normal vector at @xmath219 to @xmath218 coincides with the vector @xmath107 . in other words , @xmath219 is the @xmath220-image of points @xmath221 from the fiber @xmath222 of the logarithmic gauss mapping .
the torus @xmath223 intersects the hypersurface @xmath68 at most in @xmath224 points . consider a heighbourhood @xmath225 in @xmath226 of the point @xmath227 , then @xmath228 intersects the hypersurface @xmath68 in @xmath225 along an @xmath229-dimensional chain @xmath230 .
it can be shown , cf .
@xcite for the case @xmath231 , that integral ( [ eq:2a ] ) is asymptotically equivalent for @xmath232 to the sum @xmath233 where @xmath234 , @xmath235 is the residue form . in local coordinates @xmath236 of @xmath68 ( assuming @xmath116 ) we have @xmath237 .
therefore , the diagonal coefficient can be represented as the sum of oscillating integrals with the _ phase function _ @xmath238 . the critical points of this phase function give the main contribution to the asymptotic of such integrals . from proposition 3
follows , that the support of @xmath239 contains only one critical point of @xmath240 it is a point @xmath241 .
the asymptotics of an oscillating integral is most simple for morse critical points . in this case
it is given by stationary phase method ( also called saddle - point method , see @xcite ) .
the corollary 1 of proposition 4 states , that for directions @xmath139 outside the ramification locus of the logarithmic gauss map @xmath40 the phase function @xmath113 has only morse critical points .
the situation of a degenerated critical point is much more complicated .
first of all we are looking only for rational critical points ! by a result of varchenko
some information about asymptotics of oscillating integral can be read from the distance of the newton diagramm of the phase function at the corresponding point in case of a newton non - degenerated phase function ( and then it depends only of the @xmath151-equivanence class of the hypersurface singularity ) .
otherwise , the distance is only a lower bound .
so called _ adapted coordinates _ exist always in dimension 2 , such that the phase function is newton non - degenerated .
adapted coordinates can be computed algorithmically , for more details cf .
@xcite and @xcite .
consider the smooth hypersurface @xmath0 defined as a zero set of the polynomial @xmath242 which is non - degenerated for its newton polyhedron .
the cubic @xmath0 is a two - dimensional real torus with three removed points .
the solutions @xmath243 of @xmath244 for fixed parameter @xmath245 are zeroes of ideal ( [ eq : cv ] ) and for real parameter @xmath120 they are projected to the contour @xmath91 by @xmath90-map .
we are interested in the real ramification locus of @xmath29 .
we compute the resultant of @xmath246 with respect to the variable @xmath247 @xmath248 the multiplicity of an isolated zero @xmath249 of system ( [ eq : ex1 ] ) coincides with the multiplicity of the zero @xmath250 in @xmath251 .
the discriminant of the polynomial @xmath251 with respect to @xmath252 is the homogeneous polynomial @xmath253 in variables @xmath254 .
interested in roots of ( [ eq : ex1 ] ) in @xmath255 we can omit the factor @xmath256 in the last expression . substituting in @xmath257 an affine parameter @xmath258 we get the polynomial @xmath259
whose real zeroes @xmath260 give the points @xmath261 of the real ramification locus of @xmath29 .
choosing values of @xmath267 from the different intervals of @xmath262 we obtain different configurations of the contour @xmath91 and the amoeba @xmath67 . because the volume @xmath268 does not depend on @xmath267 , all these configurations have a following common property :
the number of preimages @xmath269 of a point @xmath270 with normal vector @xmath271 is equal to three , provided we count such preimages , which are solutions to ( [ eq : ex1 ] ) for corresponding @xmath272 , with their multiplicity in ( [ eq : ex1 ] ) .
hence , one can find for every @xmath273 three points on @xmath91 with the normal vector @xmath274 .
moreover , each one lies on its own colored or black part of the contour ( see .
fig . 1 ) . although , for @xmath275 the real ramification locus of @xmath29 is empty , we can distinguish two situations . if @xmath276 the hypersurface @xmath0 is a complexification of the so - called harnack curve and the complement of its amoeba has the maximal number of components . in this case
@xmath277 has only fold singularities , which coincide with @xmath278 . if @xmath279 the complement of the amoeba @xmath67 has no bounded component , and the mapping @xmath94 has three pleat singularities and other singularities are folds .
therefore , for @xmath275 @xmath29-fiber of any rational @xmath281 contains only the morse critical points of the phase function .
for example , set the parameter @xmath282 then the @xmath29-fiber of @xmath283 consists of the morse points @xmath284 , @xmath285 and @xmath286 .
for @xmath263 we can get degenerated rational points in a real ramification locus , e.g. there are six rational points @xmath287 in the real ramification locus of @xmath288 @xmath289 the @xmath29-fiber of such points has a simple point and an @xmath190-point of the phase function .
we consider the polynomial @xmath6 in @xmath290 variables and non - degenerated for its newton polyhedron , @xmath291 as in example 1 the real ramification locus of @xmath29 is determined by the following system @xmath292 with similar computations we obtain the discriminant @xmath257 of the logarithmic gauss map .
we do not consider zeroes of the first three factors in ( [ eq : ex22 ] ) , because they do not give us multiple roots of ( [ eq : ex2 ] ) in the torus .
the ramification locus of @xmath29 is given by zero set of @xmath294 .
let @xmath299 , @xmath300 be coordinates in affine part of @xmath301 , where @xmath302 .
2 shows the zero set of @xmath303 , which coincides with the affine part of the real ramification locus of @xmath29 .
the red points @xmath304 on fig .
2 are the degerated rational critical points of the discriminant with milnor number @xmath305 .
this example of the polynomial @xmath6 is a special one , because the existence of rational degenerated points in a real ramification locus is not a generic property .
but , we are interested in such points , because they lead to degenerated critical points of a phase function . in this example
the @xmath29-fiber of any @xmath190-point contains excately one @xmath192-critical point of the phase function ( see appendix for details ) .
+ * appendix : + computation with singular ( some experiences ) * + the computer algebra system singular , cf .
@xcite , was used for the computation of examples .
we tried several strategies for computing the discriminant of the log - gauss map with different success , i.e. to get a result for non - trivial examples without overflow and in reasonable time . here
we give a small introduction how proceed in singular , demonstrated with the equation of example 2 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ring r=0,(y1,y2,y3,z1,z2,z3),dp ; + poly f=1+z1+z2+z3 + 3*z1*z2 + 3*z1*z3 + 3*z2*z3 + 11*z1*z2*z3 ; + matrix a[2][3 ] = z1*diff(f , z1),z2*diff(f , z2),z3*diff(f , z3),y1,y2,y3 ; + ideal i = f , minor(a,2 ) ; _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ next we project the graph restricted to some affine chart @xmath308 into @xmath309 ( @xmath310 - a coordinate axes of @xmath311 ) .
the image is a hypersurface defined by a polynomial @xmath312 , which we could closure in @xmath313 by homogenizing in the @xmath314 . using the elimination of variable ,
the multiplicities of multiple factors may be lost , but it does not effect the result .
+ the choice of the projection direction was good , if @xmath315 .
the discriminant variety of @xmath29 is contained in the discriminant hypersurface of the projection @xmath316 , computed in the next step .
+ the plane curve @xmath317 has several components , it may have components with certain multiplicities , some of them induced from the closure @xmath0 or not belonging to the discriminant .
if our polynomial is generic , then we expect the discriminant of @xmath29 ( i.e. restricted to the torus ) being irreducible .
we should test which factor is correct .
some components of @xmath318 have empty fiber with respect to @xmath29 or no multiple points in its @xmath29-fibers .
we can reduce sometimes the number of factor as follows : compute for any coordinate @xmath319 ( as above for @xmath320 ) polynomials @xmath239 and @xmath321 and factorize only @xmath322 .
+ having found the equation of the discriminant polynomial @xmath323 , we can compute its ( discrete ) singular locus .
+ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ poly d0 = ld[1][2 ] ; + d0 = subst(d0,y3,1 ) ; + ring s = 0,(y1,y2,y3),dp ; + poly d0 = imap(r , d0 ) ; + ideal sl = slocus(d0 ) ; + list lsl = primdecgtz(sl ) ; _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ here , the singular locus has six rational double points @xmath324 , @xmath325 , @xmath326 , @xmath327 , @xmath328 , @xmath329 and more irrational singular points .
we choose @xmath330 and check , that it is an @xmath190-singularity of @xmath128 .
+ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ setring r ; + i = subst(i
, y1,1 ) ; i = subst(i , y2,3 ) ; + ring r0 = 0,(z1,z2,z3),dp ; + ideal i = imap(r , i ) ; + list lfib = primdecgtz(i);. + option(redsb ) ; + show(std(lfib[1][2 ] ) ) ; + `` mult = ' ' , vdim(std(lfib[1][1 ] ) ) ; _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ | the study of hypersurfaces in a torus leads to the beautiful zoo of amoebas and their contours , whose possible configurations are seen from combinatorical data .
there is a deep connection to the logarithmic gauss map and its critical points .
the theory has a lot of applications in many directions . in this report
we recall basic notions and results from the theory of amoebas , show some connection to algebraic singularity theory and discuss some consequences from the well known classification of singularities to this subject .
moreover , we have tried to compute some examples using the computer algebra system singular and discuss different possibilities and their effectivity to compute the critical points .
here we meet an essential obstacle : relevant examples need real or even rational solutions , which are found only by chance .
we have tried to unify different views to that subject . |
the characterization of correlations of a quantum state is object of an intense field of investigation , due to both its fundamental scientific interest and its importance towards the implementation of quantum technologies.@xcite entanglement has been traditionally considered as a fundamental resource to obtain quantum computational advantages , and has been used as the main indicator of the quantumness of correlation . indeed , as shown in ref .
, for pure - state computation , exponential speed - up only occurs if entanglement grows with the size of the system .
once mixed - state computation is considered , however , signatures of quantum speed - up can come out using factorized states as , for instance , in the so - called deterministic quantum computation with one qubit ( dqc1).@xcite decoherence effects due to dissipation are known to be detrimental for entanglement that is indeed disappearing after a transient time ( and not asymptotically).@xcite among many other attempts to quantify quantum correlations , a predominant role has been assumed by quantum discord.@xcite it has been introduced with the aim of capturing _ all _ quantum correlations , including entanglement
. however , the relationship between these two quantities is still unclear , since they seem to capture different properties of the states . in ref .
it is shown , for instance , that even if discord and entanglement are the same for pure states , @xmath0 states maximizing the discord in a given range of classical correlations are actually separable .
recently , the analytical expression of quantum discord has been obtained also for gaussian states @xcite opening the possibility to use it for continuous variables.@xcite from a different point of view , the quantumness of a system can be measured through other kinds of indicators , widely developed , for instance , in the field of quantum optics . a well - known example are
quantum correlations between two twin beams generated in optical parametric oscillators.@xcite the quantumness of the state of the emitted light is measured by the absence of fluctuations in their intensities difference .
this absence of noise is equivalent to the negativity of this variance for normal ordered operators and was first predicted by reynaud and collaborators@xcite and experimentally measured in ref . .
our aim in this work is to see if there is any connection between the latter correlations , which we will call `` twin oscillators correlations '' , with entanglement and discord , comparing their decaying and robustness . in this paper
, we will consider one of the most fundamental interacting systems , i.e. two coupled harmonic oscillators , in the presence of dissipation due to the interaction with a thermal environment .
two extreme scenarios we are going to investigate are represented by the so - called `` common bath '' , where the two oscillators are thought to be so `` close '' with each other that they interact with the same thermal modes , and the case of `` separate baths '' , where the dephasing channels are completely independent.@xcite it was recently shown@xcite that two identical oscillators in the presence of a common bath can exhibit asymptotic entanglement robust against decoherence , depending on the bath temperature and initial squeezing .
this is a very peculiar case and this behavior is generally lost if the two oscillators are not identical.@xcite still , slow decay of entanglement and robust quantum correlations appears in the presence of synchronization between detuned oscillators , as shown in ref . .
studying the dynamics of the system through the master equation approach , we want both to analyze the behavior of quantum correlations , considering entanglement and quantum discord , and classify the global quantumness of the state using the variance of the difference of the occupation numbers .
first , we will define the model and discuss its solution ; afterward , the relevant indicators will be defined ; as a final step , we will study the dynamics of such indicators .
let us consider two quantum harmonic oscillators allowing for diversity in their frequencies and direct coupling as well as dissipation in thermal environment.@xcite the model describing the system dipped into two identical but independent ( separate ) thermal baths is given by the total @xmath1 , where the system hamiltonian @xmath2 describes two oscillators with frequencies @xmath3 , unitary masses and coupled through their positions , @xmath4 is the free hamiltonian of two ( identical ) baths of harmonic oscillators ( labeled by @xmath5 ) , and @xmath6 encompasses the system - bath interaction .
the case of a common bath is obtained by considering only @xmath7 in @xmath8 and @xmath9 the effective dissipation takes place therefore only in the sum of positions @xmath10 , while in the case of separate baths both positions @xmath11 and @xmath12 are independently coupled with the thermal bath .
the thermal bath is assumed here to be ohmic , with a lorentz - drude cut - off parameter @xmath13 , and its spectral density@xcite is @xmath14 in the following we will consider the cut - off frequency always larger than any frequency involved in the oscillators free dynamics , that is @xmath15 and weak coupling @xmath16 .
the analysis of the dissipation of identical oscillators in common and separate baths was given in refs . .
in these works it is shown that in presence of a common bath entanglement can persist in the asymptotic state , in spite of dissipation .
the master equations describing the time evolution of the reduced density matrix of the system ( obtained by tracing out the degrees of freedom of the bath ) when including the frequency diversity of the oscillators have been reported in refs . in the weak coupling limit ( between system and environment ) and without relying on the rotating wave approximation .
once the master equation has been obtained , we can explicitly write down and solve the equations of motion for all the operators moments . if the initial state of the two oscillators is gaussian with vanishing average positions and momenta , the complete information is contained in the matrix of second moments ( the covariance matrix ) .
detailed dynamical equations , for the reduced density matrix and for the position and momenta second order moments , are given in ref . , also in the system eigenmodes basis@xcite and not reproduced here .
the limit case of high temperatures in presence of a non - markovian environment is studied in ref . .
in this section , we briefly review the quantifiers we will use to characterize the time evolution of our system . while , in general , measures of entanglement have only been developed for pure states , the case of gaussian density matrices , together with the case of qubits ,
is one of the exceptions , since a necessary and sufficient criterion of separability exists.@xcite for pure bipartite states @xmath17 , independently on their nature , entanglement can be calculated through the von neumann entropy ( entropy of entanglement ) of one of the two reduced density matrices : @xmath18 . in the case of mixed states
@xmath19 , however , the von neumann entropy can not be used since the mixedness of the reduced density matrices can not discriminate between entanglement and lack of purity of @xmath19 .
a sufficient criterion ( the so - called peres - horodecki criterion ) for detecting entanglement can be obtained by considering the positivity of the partial transpose @xmath20 ( or , equivalently , @xmath21 ) , i.e. of the matrix obtained by only transposing the degrees of freedom of one of the two sub parties.@xcite indeed , the presence of negative eigenvalues of @xmath20 witnesses that @xmath22 has not the form of a factorized density matrix .
as said before , in the case of gaussian states , the separability of the partial transpose is also necessary to detect entanglement,@xcite and the modulus of the sum of the negative eigenvalues of @xmath20 ( @xmath23 ) has been shown to be an entanglement monotone.@xcite since gausssian states are completely characterized by their first and second moments , and first moments can be set to zero with local operations that do no modify entanglement , the covariance matrix can be used to check the positivity of the partial transpose .
the logarithmic negativity , which represents an upper bound to the distillable entanglement , is defined as @xmath24 and is related to the smallest symplectic eigenvalue of the covariance matrix of @xmath20 ( @xmath25 ) : @xmath26.\ ] ] in contrast to other entanglement measures , logarithmic negativity does not reduce to entropy of entanglement on pure states .
this entanglement quantifier was also considered in the context of coupled dissipative harmonic oscillators in the mentioned works.@xcite in classical information theory , the mutual information of a bipartite system can be calculated through two equivalent formulae related by bayes rule : we have @xmath27 , with @xmath28 and @xmath29 , where @xmath30 is the shannon entropy and @xmath31 is the conditional shannon entropy of @xmath32 given @xmath33 .
the quantum counterparts of @xmath34 and @xmath35 , however , differ substantially.@xcite by replacing the shannon entropy with the von neumann entropy of a given bipartite state @xmath19 [ @xmath36 , we obtain the quantum mutual information @xmath37 where @xmath38 are the reduced states after tracing out party @xmath39 . due to the nature of measures in quantum mechanics , @xmath40 depends on the measurement realized on @xmath33 .
classical correlations are then defined as @xcite @xmath41,\ ] ] with the conditional entropy defined as @xmath42 , @xmath43 and where @xmath44 is the density matrix after a positive operator valued measure ( povm ) @xmath45 has been performed on @xmath33 .
quantum discord is defined as the difference between @xmath46 and @xmath47 : @xmath48.\ ] ] while the calculation of quantum discord , being based on a minimization procedure , is in general an unsolved problem , in the case of gaussian states , an analytical formula has been obtained.@xcite this allowed to consider quantum discord also in the context of continuous variable quantum information.@xcite the quantum discord measures in some sense how much disturbance is caused when trying to know about party a when measuring party b , and has been shown to be null only for a set of states with measure zero.@xcite it was shown to be a useful resource in the dqc1 algorithm,@xcite where the quantum speed up does not rely on entanglement and , given its inequivalence to entanglement ( except for pure states ) , it hints at more general definitions of what is quantum in a correlation .
the dynamics of quantum correlations , as quantified by the discord , and mutual information between quantum harmonic oscillators have been recently studied between different oscillators focusing on different parameters regimes and showing that the robustness ( slow decay ) of these correlations is related to the presence of a synchronous dynamics.@xcite in the context of quantum optics , the discrimination between the predictions of classical and quantum theories has been a wide field of investigation . a tool to investigate the violation of classical inequalities is given by the variance of the difference of the occupation numbers : @xmath49 where , as usual
, @xmath50 is the occupation number operator of each oscillator and @xmath51 indicates normal ordering .
the quantumness of correlations in the occupation numbers derives from the absence of noise when subtracting the oscillators intensity fluctuations .
this is equivalent ( in normal ordering ) to the negativity of the variance @xmath52 and is a consequence of the negativity of the glauber - sudarshan quasi - probability @xmath53 .
we note that for identical oscillators eq .
[ vard ] is identical to @xmath54 and this indicator characterize anti - bunching , being a cross correlation larger than the autocorrelation .
when the system or the state are not symmetrical in the two components , the negativity of the latter second order moment would be one of many possible quantum indicators , implying negativity of the correspondent @xmath53 distribution not associated to anti - bunching.@xcite the existence of these strong correlations in optics generally comes from the simultaneous generation of pairs of photons in nonlinear processes , and this generally characterizes twin beams.@xcite this is a rather robust phenomenon in complex spatiotemporal dynamics.@xcite in the following we will study the temporal dynamics of @xmath52 , that we will name twin oscillator correlations . given the gaussian character of the initial state we want to study , the fourth order moments can be obtained from the covariance matrix.@xcite the dynamics of this indicator was already considered in ref . in comparison with entanglement through few examples , showing a similar decay after a finite time transient .
however , the possibility to get asymptotically twin oscillators was not considered there and will be addressed in the next section where we will also fully analyze the role of the squeezing of the initial state and of the temperature . on the other hand , this indicator ( @xmath52 ) has been considered in ref . in a different regime , for high temperatures focusing on the short time decay in presence of non - markovian environment .
let us consider an initial two - mode squeezed state @xmath55 where @xmath56 and @xmath57 is the squeezing amplitude .
we know that for this state @xmath58 .
we want to study how the different indicators dynamically behave considering three different scenarios : ( i ) the case of different frequencies for a common bath ; ( ii ) the case of equal frequencies for separate baths ; ( iii ) the case of equal frequencies ( @xmath59 ) for a common bath . as shown in refs . and , for equal frequencies , in the case of a common environment
, the asymptotic state is expected to be entangled .
this is due to the fact that one of the degrees of freedom of the system [ the mode @xmath60 is actually frozen , since it does not interact with the bath , and represents a decoherence - free subspace .
moving away from the resonance condition , a full thermalization process takes place , and the final ( gibbs ) state can be entangled only in the very low temperature regime . motivated by these results , we then want to investigate whether the existence of this noiseless channel also protects other aspects of the quantumness of the state under evolution .
we start @xmath61 ( no direct coupling between the oscillators ) to avoid fast oscillations and @xmath62 as a scale unit of measure of energy.then , we choose for the system bath coupling @xmath63 ( weak coupling limit ) with cut - off @xmath64 , and squeezing @xmath65 in the initial state ( [ tms ] ) . [ figura2 ] in fig . 1
, we show the dynamics of logarithmic negativity , quantum discord and @xmath66 $ ] ( @xmath52 is scaled by a factor 4 for the sake of comparison ) .
we start considering resonant oscillators ( @xmath59 ) coupled to separate baths and detuned oscillators ( @xmath67 ) coupled to a common environment .
entanglement and @xmath66 $ ] vanish in a finite time,@xcite being more fragile than quantum discord , which is exponentially decaying.@xcite as a matter of fact , it is known that states with vanishing discord are rare,@xcite and in the presence of dissipation this indicator does not experience a sudden death process .
the case of equal frequencies and a common environment is that one showing entanglement in the asymptotic state.@xcite as we can observe from the right panel of fig . 2 , for temperature @xmath68 , both entanglement and quantum discord reach a stationary regime after a transient phase , while the negativity of @xmath52 disappears in a finite time .
by lowering @xmath69 up to @xmath70 , i.e. in a regime where quantum effects are stronger , we see that the asymptotic value of entanglement and discord is increased , but we also observe that the variance of the difference between the occupation numbers becomes negative for infinite time . in fig . 3 , we reproduce the left panel of fig . 2 by adding a finite direct coupling between the two oscillators @xmath71 .
while the presence of @xmath72 induces fast oscillations in the three observables , all of them are still present in the asymptotic regime .
finally , comparing these first figures , it seems then that the sub - poissonian character of fluctuations ( @xmath73 ) is rather fragile and that oscillators remain twin asymptotically only at very low temperature .
[ figura1 ] [ figura3 ]
in order to have a general view of the persistence of both entanglement and twin oscillators correlations for a common bath , we consider the respective asymptotic states . the asymptotic entanglement for identical
decoupled oscillators dissipating in a common bath was given in ref .
as a function of the temperature and the squeezing of the initial state .
we show the phase diagram for presence and absence of asymptotic entanglement in fig . 4 , in the weak coupling limit .
we then estimate @xmath52 in the asymptotic limit ( @xmath74 ) , assuming thermalization for the mode @xmath75 and free dynamics for the mode @xmath76 depending on the initial state . by explicitly writing down @xmath52 and by replacing all the entries of the two - oscillator covariance matrix by their asymptotic value ,
we conclude that the variance of the difference between the occupation numbers is negative when the following inequality is satisfied : @xmath77 since the observables of the non dissipating mode @xmath78 are oscillating in time , in order to find negative values for @xmath52 , we will take the minimum in a period .
it follows then , as for entanglement , it is equivalent to start from a two - mode squeezed state or two ( separable ) squeezed states
. finally we obtain the phase diagram represented in fig .
4 , where entanglement and negative @xmath52 ( or negative @xmath53 distribution ) are obtained below the corresponding lines .
we see that twin oscillators correlations are achieved only for very low temperatures , in contrast with entanglement for which the detrimental effect of temperature can be compensated by stronger initial squeezing . as a matter of fact , for @xmath79 there are no twin oscillators unless the temperature is very low ( @xmath80 ) .
this critical temperature corresponds to @xmath81 .
an important remark is that the threshold curve for @xmath73 is not continuous .
in fact , for @xmath82 we have @xmath83 for any @xmath69 .
on the other hand , for @xmath84 , the presence of twin correlations is determined by the sign of @xmath85 , and they only appear below a finite critical temperature .
[ figura4 ] from these results , we learn that the presence of a decoherence - free channel allows the preservation of the quantum character of the state , which manifests itself both through the presence of correlations ( entanglement and discord ) and the negativity of @xmath52 . this latter characteristic , signature of a negativity of the @xmath53 distribution , is more fragile than the other quantum markers considered here .
entanglement and twin correlations have been analyzed also in ref . in a different case , for a pair of optical field modes obtained from parametric down - conversion when the input light is in a thermal and separable state . even if the states obtained from ( different processes of ) thermalization of a squeezed state , as considered here , and the thermal state after unitary action of parametric down conversion , as in ref . , are different , we see that in both cases entanglement is more likely to be found that twin oscillators correlations .
we have studied the effects of dissipation on the quantum character of the state of two coupled harmonic oscillators .
we discussed the system quantumness considering the dynamics of different indicators , such as entanglement , quantum discord , and twin oscillators correlations , a signature of negative values of the @xmath53 distribution .
we find that in general all these quantumness indicators vanish in the asymptotic limit of large times when they reach the equilibrium ( apart from an exponentially small value of discord ) unless a decoherence - free channel exists . in our model
this channel is represented by the coupling of the two resonant ( identical ) oscillators to a common environment .
still , out of resonance between the two oscillators , it would possible , in line of principle , to restore the noiseless channel by unbalancing the coupling of the oscillators to the bath.@xcite we have also found that , whenever entanglement can not be asymptotically preserved , its death time becomes similar to that of the variance of the difference between the occupation numbers ( @xmath52 ) . after showing the effect of detuning between oscillators , ( common or separate ) environment , oscillators coupling and temperature through few examples
, we obtained the phase diagram for asymptotic entanglement and twin oscillators correlations . in the case of
identical decoupled oscillators , we have found that in order to find a negative value of @xmath52 at very long time , the temperature needs to be smaller than the one allowing for asymptotic conservation of entanglement .
moreover , the effect of temperature on entanglement can be compensated by stronger initial squeezing , being this not the case for twin oscillators correlations .
so , this quantifier ( @xmath73 ) is even more fragile under dissipation than the entanglement itself
. this observation could be more general as it is similar to what found in a rather different system in ref . .
this asymptotic phase diagram has been obtained for identical oscillators and a common thermal bath .
once the frozen degree of freedom disappears ( either by detuning the oscillators or by introducing separate baths ) , the system is expected to asymptotically thermalize in all degree of freedom .
its quantum character is then generally lost , unless other mechanisms are introduced , as for instance driving the system out of equilibrium.@xcite as a matter of fact , twin oscillators correlations are mostly related to the presence and the direction of squeezing of the damped eigenmode ( that is the sign of @xmath85 ) .
funding from fisicos ( fis2007 - 60327 ) , coqusys ( 200450e566 ) , `` accion especial '' caib ( aaee0113/09 ) projects are acknowledged .
glg is supported by the spanish ministry of science and innovation through the program `` juan de la cierva '' .
fg is supported by the csic through the program `` junta para la ampliacin de estudios '' e. knill and r. laflamme , phys .
* 81 * ( 1998 ) 5672 ; a. datta , a. shaji , and c. m. caves , phys
* 100 * ( 2008 ) 050502 ; b. p. lanyon , m. barbieri , m. p. almeida , and a. g. white , phys . rev
* 101 * ( 2008 ) 200501 .
i. p. degiovanni , m. genovese , v. schettini , m. bondani , a. andreoni , and m. g. a. paris phys .
a * 79 * ( 2009 ) 063836 ; g. brida , m. bondani , i. p. degiovanni , m. genovese , m. g.a .
paris , i. ruo berchera , and v. schettini , found phys * 41 * ( 2011 ) 305 . | the dynamics of a pair of coupled harmonic oscillators in separate or common thermal environments is studied , focusing on different indicators of quantumness , such as entanglement , twin oscillators correlations and quantum discord .
we compare their decay under the effect of dissipation and show , through a phase diagram , that entanglement is more likely to survive asymptotically than twin oscillators correlations . #
1#1 : |
massive multiple - input multiple - output ( mimo ) technique exploiting hundreds of antennas at the base station ( bs ) , is one of the key enabler for future 5 g wireless cellular systems . to achieve the theoretical performance gains in massive mimo systems ,
accurate channel state information at the transmitter ( csit ) is crucial @xcite . for csit acquisition , frequency - division duplexing ( fdd )
requires direct feedback of the csi from the users to the bs , but such process is unnecessary for time division duplexing ( tdd ) since the csit can be obtained from the uplink channel estimation by leveraging the channel reciprocity @xcite . while many of massive mimo works consider the tdd mode due to this reason , fdd has many benefits over tdd ( especially in delay - sensitive or traffic - symmetric applications @xcite ) and also dominates current cellular networks .
thus , it is of importance to come up with solutions to the csit acquisition problem for fdd massive mimo systems .
conventional csit acquisition for fdd mimo systems consists of two separate steps : channel estimation in the downlink and feedback of csi in the uplink .
first , the bs transmits orthogonal pilots in the downlink , and each user estimates its own channel using the pilot observation . commonly used channel estimation algorithms include least squares ( ls ) and minimum mean square error ( mmse ) .
then , the estimated channel is fed back to the bs via dedicated uplink channels .
since the number of pilots grows with the number of transmit antennas at the bs , overhead of downlink pilot signaling becomes overwhelming for massive mimo systems .
also , the overhead of csi feedback is a serious concern due to the same reason . in order to address these issues ,
various approaches have been proposed in recent years @xcite-@xcite . in @xcite and @xcite ,
authors propose to reduce the downlink training ovehead by carefully designing the training pilots . in @xcite , an approach to reduce the csi feedback overhead when
the bs antennas are highly correlated has been proposed . in @xcite ,
an approach based on compressive sensing ( cs ) has been proposed to reduce both the downlink training overhead and uplink csi feedback overhead .
while this approach is promising when the channel matrices of different users are sparse and partially share common support , such is not true when these assumptions are violated . in this letter ,
we propose a joint csit acquisition scheme based on low - rank matrix completion for fdd massive mimo systems .
specifically , the bs transmits pilots for downlink channel training and the scheduled users _ directly _ feed back the pilot observation to the bs without performing the individual channel estimation .
then , the joint recovery of the csi for all users is performed at the bs based on the low - rank matrix completion algorithm , whereby the low - rank property of the massive mimo channel matrix caused by correlation among users is exploited . in this way
, the overhead of downlink channel training as well as uplink channel feedback can be reduced , which will be verified by simulation results .
_ notation _ : lower - case and upper - case boldface letters denote vectors and matrices , respectively ; @xmath0 , @xmath1 and @xmath2 denote the transpose , conjugate transpose , and inverse of a matrix , respectively ; @xmath3 is the right moore - penrose pseudoinverse ; @xmath4 denotes the rank of @xmath5 ; @xmath6 and @xmath7 denote the vectorization and unvectorization , respectively ; @xmath8 denotes the kronecker product ; @xmath9 denotes the identity matrix of size @xmath10 ; @xmath11 is the @xmath12-@xmath13 ; @xmath14 is the nuclear norm denoting the sum of singular values of @xmath5 .
we consider the downlink of fdd massive mimo system with @xmath15 antennas at the bs and @xmath16 users with single receive antenna .
the bs transmits pilots @xmath17 at the @xmath18-th channel use ( @xmath19 ) . at the @xmath20-th user , the pilot observation @xmath21 during @xmath22 channel uses
can be expressed as @xmath23 where @xmath24 $ ] is an @xmath25 dimensional matrix constructed from the transmitted pilots during @xmath22 channel uses , @xmath26 represents the independent and identically distributed ( i.i.d . )
additive white gaussian noise ( awgn ) with elements having zero mean and the variance @xmath27 , the channel vector @xmath28 between the bs and the @xmath20-th user is given by @xcite @xmath29 where @xmath30 is the number of resolvable physical paths , @xmath31 is the propagation gain of the @xmath32-th path , @xmath33 is the angle - of - departure ( aod ) of the @xmath32-th path , and @xmath34 is the steering vector . in this work ,
we consider the typical uniform linear arrays model @xcite @xmath35 $ ] , where @xmath36 and @xmath37 denote the antenna spacing at the bs and carrier wavelength , respectively .
in conventional csit acquisition schemes , the channel vector @xmath38 of each user is estimated individually using classical algorithms such as ls or mmse , and then the estimated csi is fed back to the bs @xcite .
for example , ls algorithm generates the estimated channel vector @xmath39 . in our work , we propose a joint csit acquisition scheme , where each user directly feeds back its own pilot observation @xmath40 to the bs for the joint mimo channel recovery of all users .
the aggregate pilot observation @xmath41^t\in\mathcal{c}^{k\times t}$ ] for all scheduled @xmath16 users can be expressed as @xmath42 where @xmath43^t\in\mathcal{c}^{k\times m}$ ] is the mimo channel matrix to be recovered , and @xmath44^t$ ] is the downlink noise matrix . in the channel model ( [ eq2 ] ) ,
rich scattering is typically assumed at the user side , and most clusters around the bs are accessible for almost all users .
it has been shown that a cluster seen by different users , so called
joint clusters " , introduces correlation among users even when they are geographically separated - th cluster around the bs , which is accessible from all users except for the @xmath20-th user , we can model it by setting the corresponding propagation gains @xmath45 . ]
that is , the channel vectors associated with different users have the same steering vectors @xmath46 .
thus , we have @xmath47 where @xmath48 with the @xmath49-th entry being @xmath31 , and @xmath50^t\in\mathcal{c}^{p\times m}$ ] . as @xmath51
, we have @xmath52 . for massive mimo systems ,
@xmath15 and @xmath16 are usually large but the number of resolvable paths @xmath30 is relatively small due to the limited number of clusters around the bs @xcite , @xcite , so that @xmath53 . that is , the rank of @xmath5 of size @xmath54 is much smaller than its dimension . in the sequel
, we call this property as low - rank property " of the massive mimo channel matrix . the pilot observation @xmath55 at the bs can be expressed as @xmath56 where @xmath57 is the uplink rayleigh fading channel matrix whose entries follows @xmath58 @xcite , and @xmath59 is the uplink noise matrix whose entries follow @xmath60 . to recover the downlink channel matrix @xmath5 at the bs
, we firstly estimate the aggregate pilot observation @xmath61 by @xcite @xmath62 then , by exploiting the low - rank property of @xmath5 , the joint mimo channel recovery problem at the bs can be formulated as a rank minimization problem : @xmath63 note that this problem is non - convex and np - hard @xcite .
one possible solution to avoid the computional difficulty is to use the nuclear norm minimization problem @xmath64 note that this problem can be solved by semidefinite programming ( sdp ) @xcite , but the computational complexity of the solver ( e.g. , sedumi @xcite ) is still high especially when the problem dimension is large in massive mimo systems . to alleviate the computational complexity
, we need to reformulate the problem .
firstly , we vectorize ( [ eq4 ] ) as @xmath65 where @xmath66 , @xmath67 , @xmath68 and @xmath69 . then , the joint mimo channel recovery problem can be reformulated as a low - rank matrix completion problem : @xmath70 where @xmath71 . without the low - rank constraint @xmath53 , it is clear that the solution to the unconstrained optimization problem @xmath72 can be easily obtained by using the classical gradient descent algorithm or newton s algorithm @xcite . however , when the low - rank constraint is added , novel algorithm must be developed to solve the constrained optimization problem ( [ eq10 ] ) . the solution to ( [ eq10 ] ) can be obtained by using singular value projection ( svp ) based algorithms @xcite or riemannian pursuit ( rp ) algorithms @xcite . in this letter , we use the modified version of the svp - based algorithm . for traditional svp - based algorithms such as svp - based gradient decent algorithm
( svp - g ) and svp - based newton s algorithm ( svp - n ) , the solution satisfying the low - rank constraint can be achieved by svp at every iteration . in the @xmath73-th iteration , the current result @xmath74 of linear search
is projected onto a low - rank matrix @xmath75 , which is defined as @xmath76 , where @xmath77 is the @xmath78 most significant singular values of @xmath74 .
the resulting low - rank matrix @xmath75 will be the starting point of a linear search for the next iteration .
@xmath79 ; @xmath80 ; @xmath78 .
+ initialization : @xmath82 , @xmath83 , + @xmath84 , @xmath85 , @xmath86 .
@xmath87 , @xmath88 % svp - n @xmath89 , @xmath90 % svp - g @xmath91 @xmath92 @xmath93 @xmath94 @xmath95 @xmath96 .
however , as the cost function @xmath97 in ( [ eq10 ] ) is a quadratic convex function of @xmath98 , svp - n simply converges after one iteration ( see appendix a ) .
since the svp operation is performed only once ( i.e. , the low - rank constraint @xmath53 will be used only once ) , the performance of svp - n is generally not appealing . on the other hand ,
svp - g executes svp in every iteration and hence a better solution can be achieved at the cost of slow convergence . to combine the advantages of svp - n and svp - g
, we propose the svp - based hybrid low - rank marix completion algorithm ( svp - h ) as shown in * algorithm 1 * , where svp - n is used in the first iteration ( step 4 ) to realize fast convergence and svp - g is used for the rest iterations ( step 6 ) to achieve high accuracy . during the @xmath73-th iteration , the solution @xmath99 is obtained through a line search along the negative gradient or newton s direction ( step 8) .
after that , the unvectorized solution @xmath74 of @xmath99 is projected onto a low - rank matrix @xmath75 via svp ( step 9 ) . the vectorized solution @xmath100 of @xmath75
is used as the starting point of a linear search for the next iteration .
note that in the proposed svp - h algorithm , the search direction for svp - g is the gradient @xmath101 , while the search direction for svp - n is the newton s direction @xmath102 . the optimal step size @xmath103 is chosen to minimize @xmath104 .
that is , @xmath105 denoting @xmath106 , then the derivative of @xmath107 is @xmath108 combining this together with @xmath109 , we have @xmath110 , and thus the optimal step size @xmath103 is given by @xmath111 the existing algorithms to solve the sdp problem ( casted from ( [ eq9 ] ) ) have high complexity @xmath112 @xcite .
if the general - purpose sdp algorithm such as sedumi is employed , the complexity would be burdensome @xcite . for the svp - g algorithm , in each iteration
, the matrix multiplication to compute the search direction @xmath113 has the complexity @xmath114 , since @xmath115 .
the computation of the step size @xmath116 is complex but we can simply assume a constant step size by allowing marginal decrease in the convergence speed @xcite .
the svp operation @xmath117 requires the complexity @xmath118 .
thus , the complexity of svp - g algorithm is @xmath119 , where @xmath120 is the number of iterations . for the svp - n algorithm , since @xmath121 , the matrix multiplication to compute the solution
@xmath99 has the complexity @xmath122 ( see appendix a ) .
there is no need to compute the optimal step size for svp - n since the step size @xmath123 for svp - n is a constant ( @xmath124 ) as shown in appendix a. thus , the complexity of svp - n is @xmath125 . finally , we can conclude that the proposed svp - h algorithm has the complexity @xmath126 , which is much lower than that of existing sdp algorithms .
in this section , we investigate the performance of the proposed joint csit acquisition scheme as well as the svp - h algorithm .
the simulation parameters are set as : @xmath127 , @xmath128 , @xmath129 ; @xmath130 , @xmath131 @xcite ; @xmath132 , @xmath133 .
the overhead for downlink channel training as well as uplink channel feedback are @xmath22 channel uses . in fig .
[ svp_itr ] , we compare the normalized mean squared error ( nmse ) performance of the conventional csit acquisition scheme and the proposed joint csit acquisition scheme .
note that the nmse of the joint orthogonal matching pursuit ( j - omp ) algorithm based compressive csit estimation and feedback scheme proposed in @xcite is also presented for comparison .
the uplink channel is assumed to suffer from rayleigh fading @xcite , and both the downlink and uplink signal - to - noise ratio ( snr ) are set to 25 db . as a conventional channel estimation scheme at the user side , we use the widely used ls algorithm .
in addition , the proposed joint csit acquisition using conventional svp - n , svp - g algorithms and the proposed svp - h algorithm for joint mimo channel recovery at the bs side are also compared in fig .
[ svp_itr ] . due to the utilization of correlations among users and the resulting low - rank property of mimo channel matrix ,
it is clear that the proposed schemes using svp - g , svp - n , and svp - h outperform the conventional one using ls .
the nmse performance of j - omp is not good because the angular - domain channel matrix in our system model does not satisfy the sparse channel assumption , and the imperfect uplink channel degrades the performance of j - omp .
we also observe that both svp - h and svp - g achieve much smaller nmse than svp - n achieves because they repeatedly exploit the low - rank constraint as mentioned in section iii - b .
in addition , svp - h converges faster than svp - g due to the utilization of fast convergent svp - n in the first iteration .
that is to say , the proposed svp - h algorithm achieves accurate csit and fast convergence .
[ svp_n_c_t ] we next investigate the overhead reduction of the proposed joint csit acquisition scheme against the channel use @xmath22 . note that the nmse performance of sdp - based method and rp algorithm have also been shown in fig .
[ svp_n_c_t ] for comparison .
we can observe that both approaches perform similar to the proposed svp - h algorithm .
we can also observe that the channel use required for the proposed scheme using svp - h is much smaller than that required for the conventional scheme .
for example , to achieve the targeted nmse @xmath134 , the channel use required for the conventional scheme is @xmath135 , while that required for the proposed scheme is @xmath136 .
this clearly indicates that the proposed scheme can reduce the overhead of downlink channel training and uplink channel feedback .
in this letter , we investigate a novel csit acquisition scheme for fdd massive mimo systems by exploiting the property that the channel matrix of massive mimo system has low - rank structure . using this property
, we formulate the joint csit acquisition scheme as a low - rank matrix completion problem .
simulations have verified that the proposed svp - h algorithm can achieve accurate csit with fast convergence .
by substituting @xmath137 into ( [ eq14 ] ) , we can compute the step size of svp - n @xmath138 the solution in the @xmath73-th iteration is given by @xmath139 which is a constant vector and independent of the iteration index @xmath73 .
thus , svp - n algorithm obtains the solution after the first iteration .
f. zhu , f. gao , m. yao , and h. zou , joint information- and jamming- beamforming for physical layer security with full duplex base station , " _ ieee trans . signal process .
24 , pp . 6391 - 6401 , dec .
f. rusek , d. persson , b. lau , e. larsson , t. marzetta , o. edfors , and f. tufvesson , scaling up mimo : opportunities and challenges with very large arrays , " _ ieee signal process . mag .
_ , vol . 30 , no .
40 - 60 , jan . 2013 .
j. choi , d. j. love , and p. bidigare , downlink training techniques for fdd massive mimo systems : open - loop and closed - loop training with memory , " _ ieee j. sel . areas
signal process .
_ , vol . 8 , no .
802 - 814 , oct . 2014 .
n. song , m. d. zoltowski , and d. j. love , downlink training codebook design and hybrid precoding in fdd massive mimo systems , " in _ proc .
ieee global commun .
( ieee globecom14 ) _ , dec . 2014 , pp . 1631 - 1636 .
b. lee , j. choi , j. seol , d. j. love , and b. shim , antenna grouping based feedback reduction for fdd - based massive mimo systems , " in _ proc .
conf . on commun .
( ieee icc14 ) _ , jun .
2014 , pp .
4477 - 4482 .
y. shi , j. zhang , and k. b. letaief , low - rank matrix completion via riemannian pursuit for topological interference management , " in _ proc .
information theory ( ieee isit15 ) _ , jun .
2015 , pp .
1831 - 1835 .
t. d. bie and n. cristianini , fast sdp relaxations of graph cut clustering , transduction , and other combinatorial problems , " _ journal of machine learning research _ , vol . 7 , pp .
1409 - 1436 , dec . 2006 .
r. meka , p. jain , and i. s. dhillon , guaranteed rank minimization via singular value projection , " in _ proc .
neural information processing systems ( nips10 ) _ , 2010 . | channel state information at the transmitter ( csit ) is essential for frequency - division duplexing ( fdd ) massive mimo systems , but conventional solutions involve overwhelming overhead both for downlink channel training and uplink channel feedback . in this letter , we propose a joint csit acquisition scheme to reduce the overhead . particularly , unlike conventional schemes where each user individually estimates its own channel and
then feed it back to the base station ( bs ) , we propose that all scheduled users directly feed back the pilot observation to the bs , and then joint csit recovery can be realized at the bs .
we further formulate the joint csit recovery problem as a low - rank matrix completion problem by utilizing the low - rank property of the massive mimo channel matrix , which is caused by the correlation among users .
finally , we propose a hybrid low - rank matrix completion algorithm based on the singular value projection to solve this problem .
simulations demonstrate that the proposed scheme can provide accurate csit with lower overhead than conventional schemes .
massive mimo , fdd , csit , low - rank matrix completion . |
the history of quantum mechanics records that from the very beginning the founders of the quantum theory realized that there might exist formal connection between classical optics and quantum mechanics .
for example , schrdinger considered that classical dynamics of a point particle should be the geometrical optics approximation of a linear wave equation , in the same way as ray optics is a limiting approximation of wave optics ; schrdinger also searched for some quantum mechanical state which behaves like a classical ` particle ' , and this state was later recognized as the coherent state @xcite , which plays an essential role in quantum optics theory and laser physics ; as dirac wrote in his famous book @xmath1principles of quantum mechanics@xmath2@xcite : @xmath3 * for a quantum dynamic system that has a classical analogue , unitary transformation in the quantum theory is the analogue of contact transformation in the classical theory * " . according to dirac , there should exist a formal correspondence between quantum optics unitary - transform operators and classical optics transformations .
indeed , in the last century physicists also found some rigorous mathematical analogies between classical optics and quantum mechanics , i.e. the similarity between the optical helmholtz equation and the time - independent schrdinger equation ; since 1960s , the advent of a laser and the appearance of coherent state theory of radiation field @xcite , quantum optics has experienced rapid development and achieved great success in revealing and explaining the quantum mechanical features of optical field and non - classical behavior ( for instance , hanbury - brown - twiss effect , photon antibunching , squeezing , sub - poissonian photon statistics ) of photons in various photon - atom interactions @xcite . the relationship between classical and quantum coherence has been discussed in the book of mandel and wolf @xcite ; the hermite - gauss or the laguerre - gauss modes of a laser beam are described using the bosonic operator algebra by nienhuis and allen @xcite .
in addition , displaced light beams refracted by lenses according to the law of geometrical optics , were found to be the paraxial optics analog of a coherent state . besides , phase space correspondence between classical optics and quantum mechanics , say for example , the wigner function theory ,
is inspected in the literature @xcite . on the other hand ,
classical optics , which tackles vast majority of physical - optics experiments and is based on maxwell s equations , has never ceased its own evolving steps , physicists have endeavored to develop various optical transforms in light propagation through lens systems and various continuous media .
the two research fields , quantum optics and classical optics , have their own physical objects and conceptions . from the point of view of mathematics , classical optics is framed in the group transform and associated representations on appropriate function space , while quantum optics deals with operators and state vectors , and their overlap seems little at first glance .
it seems to us that if one wants to further relate them to each other , one needs some new theoretical method to `` bridge '' them .
for example , what is the quantum mechanical unitary operator corresponding to the fresnel transform in fourier optics ?
is there any so - called fresnel operator as the image of classical generalized fresnel transform ?
since generalized fresnel transforms are very popularly used in optical instrument design and optical propagation through lenses and various media , it is worth of studying these transforms in the context of quantum optics theory , especially based on coherent state , squeezed state @xcite and the newly invented entangled state theory @xcite .
fortunately , the recently developed technique of integration within an ordered product ( iwop ) of operators @xcite is of great aid to studying quantum optical version of classical optical transformations . using the iwop technique one may gain benefit for classical optics from quantum optics research , or vice - versa .
our present review is arranged as follows : in section 2 we briefly recall the classical diffraction theory @xcite , this is preparing for later sections in which we shall show that most frequently employed classical optical transforms have their counterparts in quantum optics theory . in section 3 we introduce the iwop technique and demonstrate that the completeness relation of fundamental quantum mechanical representations can be recast into normally ordered gaussian operator form . using the iwop technique we can directly perform the asymmetric ket - bra integration @xmath4 in the coordinate representation , which leads to the normally ordered single - mode squeezing operator , this seems to be a direct way to understanding the squeezing mechanism as a mapping from the classical scaling @xmath5 . in section 4 with the help of iwop technique and based on the concept of quantum entanglement of einstein - podolsky - rosen @xcite
we construct two mutually conjugate entangled states of continuum variables , @xmath6 versus @xmath7 and their _ deduced entangled states ( or named _ * correlated - amplitude number - difference entangled states * ) , they are all qualified to make up quantum mechanical representations .
it is remarkable that using the iwop technique to performing the asymmetric ket - bra integration @xmath8 leads to the two - mode normally ordered two - mode squeezing operator , this implies that the two - mode squeezed state is simultaneously an entangled state .
we point out that the entangled state @xmath9 also embodies entanglement in the aspect of correlative amplitude and the phase .
we are also encouraged that the overlap between two mutually conjugate _ deduced entangled states _ is just the bessel function the optical hankel transform kernel @xcite@xmath10 which again shows that the new representations in the context of physics theory match beautiful mathematical formalism exactly .
we then employ the _ deduced _ entangled states to derive * quantum optical version of classical circular harmonic correlation .
* section 5 is devoted to finding a quantum operator which corresponds to the optical fresnel transform , with use of the coherent state representation and by projecting the classical sympletic transform @xmath11 ( @xmath12 in phase space onto the quantum mechanical hilbert space , we are able to recognize which operator is the single - mode fresnel operator ( fo ) .
it turns out that the 1-dimensional optical fresnel transform is just the matrix element of the fresnel operator @xmath13 in the coordinate eigenstates .
besides , the coherent state projection operator representation of fo constitutes a loyal realization of symplectic group , which coincides with the fact that two successive optical fresnel transforms make up a new fresnel transform .
then in section 6 based on the coherent state projection representation of fo , we prove @xmath14 rule for optical propagation in the context of quantum optics . in section 7 the quadratic operator form of fo
is also presented and the four fundamental optical operators are derived by decomposing the fo . in section 8
we discuss how to apply the fresnel operator to quantum tomography theory , by introducing the fresnel quadrature phase @xmath15 we point out that wigner operator s radon transformation is just the pure state projection operator @xmath16 @xmath17 , where @xmath18 and @xmath19 is the position eigenstate , so the probability distribution for the fresnel quadrature phase is the radon transform of the wigner function .
moreover , the tomogram of quantum state @xmath20 is just the squared modulus of the wave function @xmath21 this new relation between quantum tomography and optical fresnel transform may provide experimentalists to figure out new approach for testing tomography .
in addition , we propose another new theorem for calculating tomogram , i.e. , the tomogram of a density operator @xmath22 is equal to the marginal integration of the classical weyl correspondence function of @xmath23 . in section 9 by virtue of the coherent state and iwop method we propose two - mode generalized fresnel operator ( gfo ) , in this case we employ the entangled state representation to relate the 2-mode gfo to classical transforms , since the 2-mode gfo is not simply the direct product of two 1-mode gfos .
the corresponding quantum optics @xmath14 rule for two - mode case is also proved .
the 2-mode gfo can also be expressed in quadratic operators form in entangled way .
the relation between optical ft and quantum tomography in two - mode case is also revealed . in section 10
we propose a kind of integration transformation , @xmath24 which is invertible and obeys parseval theorem .
remarkably , it can convert chirplet function to the kernel of fractional fourier transform ( frft ) .
this transformation can also serve for solving some operator ordering problems .
in section 11 we employ the entangled state representation to introduce the complex frft ( cfrft ) , which is not the direct product of two independent 1-dimensional frft transform .
the eigenmodes on cfrft is derived .
new eigen - modes for light propagation in graded - index medium and the fractional hankel transform are presented .
the wigner transform theory is extended to the complex form and its relation to cfrft is shown ; the integration transformation in section 10 is also extended to the entangled case . in section 12
we shall treat the adaption problem of collins diffraction formula to the cfrft with the use of two - mode ( 3 parameters ) squeezing operator and in the entangled state representation of continuous variables , in so doing the quantum mechanical version of associated theory of classical diffraction and classical cfrft is obtained , which connects classical optics and quantum optics in this aspect . in section 13
we introduce a convenient way for constructing the fractional radon transform .
the complex fractional randon transform is also proposed ; in sections 14 and 15 we discuss quantum optical version of classical wavelet transforms ( wts ) , including how to recast the condition of mother wavelet into the context of quantum optics ; how to introduce complex wavelet transform with use of the entangled state representations .
some properties , such as parseval theorem , inversion formula , and orthogonal property , the relation between wt and wigner - husimi distribution function are also discussed . in section 16 ,
we generalize the usual wavelet transform to symplectic wavelet transformation ( swt ) by using the coherent state representation and making transformation @xmath25 ( @xmath12 in phase space . the relation between swt and optical fresnel transformation
is revealed .
then the swt is extended to the entangled case by mapping the classical mixed transformation @xmath26 in 2-mode coherent state @xmath27 representation . at the end of this section ,
we introduce a new symplectic - dilation mixed wt by employing a new entangled - coherent state representation @xmath28 .
the corresponding classical optical transform is also presented . in the last section ,
we introduce the fresnel - hadamard combinatorial operator by virtue of the iwop technique and @xmath29 .
this unitary operator plays the role of both fresnel transformation for mode @xmath30 and hadamard transformation for mode @xmath31 respectively , and the two transformations are combinatorial .
all these sections are used to prove the existence of a one - to - one correspondence between quantum optical operators that transform state vectors in hilbert space and the classical optical transforms that change the distribution of optical field .
here we briefly review some typical optical transforms based on light diffraction theory . these transformations ,
as one can see in later sections , are just the correspondence of some representation transformations between certain quantum mechanical states of which some are newly constructed .
it was huygens who gave a first illustrative explanation to wave theory by proposing every point in the propagating space as a sub - excitation source of a new sub - wave .
an intuitive theory mathematically supporting huygens principle is the scalar diffraction approximation , so named because optical fields ( electromagnetic fields ) actually are vector fields , whereby the theory is valid approximately .
this theory is based on the superposition of the combined radiation field of multiple re - emission sources initiated by huygens .
light diffraction phenomena has played an important role in the development of the wave theory of light , and now underlies the fourier optics and information optics .
the formulation of a diffraction problem essentially considers an incident free - space wave whose propagation is interrupted by an obstacle or mask which changes the phase and/or amplitude of the wave locally by a well determined factor @xcite .
a more rigorous , but still in the scheme of scalar wave , derivation has been given by kirchhoff who reformulated the diffraction problem as a boundary - value problem , which essentially justifies the use of huygens principle .
the fresnel - kirchhoff ( or rayleigh - sommerfeld ) diffraction formula is practically reduced to the fresnel integral formula in paraxial and far - field approximation @xcite that reads : @xmath32 \right \ } dx_{1}dy_{1 } , \label{2.1}\ ] ] where @xmath33 is the optical distribution of a 2-dimensional light source and @xmath34 is its image on the observation plane , @xmath35 is the optical wavelength , @xmath36 is the wave number in the vacuum and @xmath37 is the propagation distance . when @xmath38 is satisfied , eq .
( [ 2.1 ] ) reduces to @xmath39 } { i\lambda z}\nonumber \\ & \times \int \int_{-\infty}^{\infty}u_{1}\left ( x_{1},y_{1}\right ) \exp \left [ -i\frac{2\pi}{\lambda z}\left ( x_{1}x_{2}+y_{1}y_{2}\right ) \right ] dx_{1}dy_{1 } , \label{2.3}\ ] ] which is named the fraunhofer diffraction formula .
the fresnel integral is closely related to the fractional fourier transform ( frft ) , actually , it has been proved that the fresnel transform can be interpreted as a scaled frft with a residual phase curvature @xcite .
the frft is a very useful tool in fourier optics and information optics .
this concept was firstly introduced in 1980 by namias @xcite but not brought enough attention until frft was defined physically , based on propagation in quadratic graded - index media ( grin media ) .
mendlovic and ozaktas @xcite defined the @xmath40th frft as follows : let the original function be input from one side of quadratic grin medium , at @xmath41 .
then , the light distribution observed at the plane @xmath42 corresponds to the @xmath40 equal to the ( @xmath43)th fractional fourier transform of the input fraction , where @xmath44 is a characteristic distance .
the frft can also be implemented by lenses .
another approach for introducing frft was made by lohmann who pointed out the algorithmic isomorphism among image rotation , rotation of the wigner distribution function @xcite , and fractional fourier transforming @xcite .
lohmann proposed the frft as the transform performed on a function that leads to a rotation with an angle of the associated wigner distribution function , in this sense , the frft bridges the gap between classical optics and optical wigner distribution theory .
recently , the frft has been paid more and more attention within different contexts of both mathematics and physics @xcite .
the frft is defined as @xmath45 \left ( x_{2},y_{2}\right ) = \frac{e^{i(1-\alpha)\frac{\pi}{2}}}{2\sin \left ( \frac{\pi}{2}\alpha \right ) } \exp \left [ -\frac{i\left ( x_{2}^{2}+y_{2}^{2}\right ) } { 2\tan \left ( \frac{\pi}{2}\alpha \right ) } \right ] \nonumber \\ & \times \int \int_{-\infty}^{\infty}\frac{dx_{1}dy_{1}}{\pi}\exp \left [ -\frac{i\left ( x_{1}^{2}+y_{1}^{2}\right ) } { 2\tan \left ( \frac{\pi}{2}\alpha \right ) } \right ] \exp \left [ \frac{i\left ( x_{2}x_{1}+y_{2}y_{1}\right ) } { \sin \left ( \frac{\pi}{2}\alpha \right ) } \right ] u_{1}\left ( x_{1},y_{1}\right ) .
\label{2.4}\ ] ] we can see that @xmath46 is the identity operator and @xmath47 is just the fourier transform .
the most important property of frft is that @xmath48 obeys the semigroup property , i.e. two successive frfts of order @xmath40 and @xmath49 makes up the frft of order @xmath50 a more general form describing the light propagation in an optical system characterized by the @xmath51 $ ] ray transfer matrix is the collins diffraction integral formula @xcite , @xmath52 \right \ } dx_{1}dy_{1 } , \label{2.5}\ ] ] where @xmath53 if the system is lossless .
one can easily find the similarity between collins formula and the frft by some scaling transform and relating the @xmath54 $ ] matrix to @xmath40 in the frft @xcite .
note that @xmath55{cc}a & b\\ c & d \end{array } \right ) $ ] is a ray transfer matrix describing optical systems belonging to the unimodular symplectic group .
when treating the light propagation in optical elements in near - axis approximation , matrices @xmath56 representing linear transformations are a convenient mathematical tool for calculating the fundamental properties of optical systems , which is the origin of the name of matrix optics . in cylindrical
coordinates the collins formula is expressed as @xcite @xmath57 \right \ } u_{1}\left ( r_{1},\theta \right ) r_{1}dr_{1}d\theta \label{2.6}\ ] ] where @xmath58 @xmath59 @xmath60 and @xmath61 when @xmath62 has rotational symmetry @xmath63 then ( [ 2.6 ] ) becomes @xmath64 { \int}_{0}^{\infty}\exp \left [ -\frac{i\pi } { \lambda b}\left [ ar_{1}^{2}+dr_{2}^{2}\right ] \right ] j_{m}\left ( \frac{2\pi r_{1}r_{2}}{\lambda b}\right ) u_{1}\left ( r_{1}\right ) r_{1}dr_{1 } , \label{2.8}\ ] ] where we have used the @xmath65-order bessel function @xmath66 d\theta . \label{2.9}\ ] ] when @xmath67 , ( [ 2.8 ] ) reduces to the standard hankel transform ( up to a phase factor)@xmath68 the compact form of one - dimensional collins formula is @xmath69 where the transform kernel is @xmath70 , \label{2.12}\ ] ] @xmath56 is the parameter matrix @xmath54 $ ] .
( [ 2.12 ] ) is called generalized fresnel transform @xcite . in the following sections we will show how we find the quantum optical counterpart for those transformations of classical optics . for this purpose in the next chapter
we introduce the iwop technique to demonstrate how dirac s symbolic method can be developed and be applied to quantum optics theory . also , we briefly review some properties of the entangled state @xcite and reveal the connection between the mutual transform generated by these entangled states and the hankel transform in classical optics .
the history of mathematics tells us that whenever there appears a new important mathematical symbol , there coexists certain operational rules for it , the quantum mechanical operators in ket - bra projective form ( the core of dirac s symbolic method ) also need their own operational rules .
the terminology symbolic method was first shown in the preface of dirac s book @xmath1the principle of quantum mechanics@xmath2 : _ _ the symbolic method , which deals directly in an abstract way with the quantities of fundamental importance__@xmath71 _
, however , seems to go more deeply into the nature of things .
it enables one to express the physical law in a neat and concise way , _ and will probably be increasingly used in the future as it becomes better understood and its own special mathematics gets developed**. * * @xcite then two questions naturally arise : how to better understand the _ symbolic method ? how to develop dirac s symbolic method , especially its mathematics ?
we noticed that _ * newton - leibniz integration rule only applies to commuting functions of continuum variables , while operators made of dirac s symbols ( ket versus bra , e.g. , * @xmath72 * * of continuous parameter * * @xmath73 * * ) in quantum mechanics are usually not commutative .
therefore integrations over the operators of type * * @xmath74 ( * where ket- and bra- state vectors need not to be hermitian - conjugate to each other ) can not be directly performed by the newton - leibniz rule .
thus we invented an innovative technique of integration within an ordered product ( iwop ) of operators that made the integration of non - commutative operators possible .
the core of iwop technique is to arrange non - commutable quantum operators within an ordered product ( say , normal ordering ) in a way that they become commutable , in this sense the gap between q - numbers and c - numbers is `` narrowed '' .
however , the nature of operators which which are within : : is not changed , they are still q - numbers , not c - numbers .
after the integration over c - numbers within ordered product is performed , we can get rid of the normal ordering symbol after putting the integration result in normal ordering . @xcite.the iwop technique thus bridges this mathematical gap between classical mechanics and quantum mechanics , and further reveals the beauty and elegance of dirac s symbolic method and transformation theory .
this technique develops symbolic method significantly , i. e. makes dirac s representation theory and the transformation theory more plentiful , and consequently to be better understood .
the beauty and elegance of dirac s symbolic method are further revealed .
various applications of the iwop technique , including constructing the entangled state , developing * the * nonlinear coherent state theory , wigner function theory , etc . are found ; many new unitary operators and operator - identities as well as new quantum mechanical representations can be derived too , which are partly * summarized in the review articles @xcite .
we begin with listing some properties of normal product of operators which means all the bosonic creation operators @xmath75 are standing on the left of annihilation operators @xmath76 in a monomial of @xmath75 and @xmath76 .
the order of bose operators @xmath76 and @xmath75 within a normally ordered product can be permuted .
that is to say , even though @xmath77 = 1 $ ] , we can have @xmath78 where @xmath79 @xmath79 denotes normal ordering . \2 .
@xmath80-numbers can be taken out of the symbol @xmath79 @xmath81 as one wishes .
the symbol @xmath79 : which is within another symbol @xmath79 @xmath81 can be deleted .
\4 . the vacuum projection operator @xmath82 has the normal product form @xmath83 \5 .
a normally ordered product can be integrated or differentiated with respect to a @xmath80-number provided the integration is convergent . as an application of iwop , ( in the following , unless particularly mentioned , we take @xmath84 for convenience . )
using the fock representation of the coordinate eigenvector @xmath85 ( @xmath86)@xmath87 we perform the integration below @xmath88 substituting ( [ 3.1 ] ) into ( [ 3.8a ] ) we see @xmath89 note that on the left of @xmath90 are all creation operators , while on its right are all annihilation operators , so the whole integral is in normal ordering , thus using property 1 we have @xmath91 as @xmath76 commutes with @xmath92 within @xmath79 @xmath79 , so @xmath93and @xmath76 can be considered as if they were parameters while the integration is performing .
therefore , by setting @xmath94 , sech@xmath95 tanh@xmath96 we are able to perform the integration and obtain@xmath97 which is just the single - mode squeezing operator in normal ordering appearing in many references .
it is worth mentioning that we have not used the su(1,1 ) lie algebra method in the derivation .
the integration automatically arranges the squeezing operator in normal ordering . using @xmath98\colon , \label{3.02}\ ] ]
( [ 3.01 ] ) becomes @xmath99 this shows the classical dilation @xmath100 maps into the normally ordered squeezing operator manifestly .
it also exhibits that the fundamental representation theory can be formulated in not so abstract way , as we can now directly perform the integral over ket - bra projection operators .
moreover , the iwop technique can be employed to perform many complicated integrations for ket - bra projection operators .
there is a deep ditch between quantum mechanical operators ( @xmath73-numbers ) theory and classical numbers ( @xmath80-numbers ) theory .
the iwop technique arranges non - commutable operators within an ordered product symbol in a way that they become commutable , in this sense the ` ditch ' between @xmath73-numbers and @xmath80-numbers is shoaled .
however , the nature of operators are not changed , they are still @xmath101numbers , not @xmath80-numbers .
after the integration over @xmath80-numbers within ordered product is performed , we can finally get rid of the normal ordering symbol by using ( [ 3.02 ] ) .
when @xmath102 eq .
( [ 3.03 ] ) becomes @xmath103 a real simple gaussian integration !
this immediately leads us to put the completeness relation of the momentum representation into the normally ordered gaussian form @xmath104 where @xmath105 and @xmath106 is the momentum eigenvector @xmath107 , @xmath108 \left \vert 0\right \rangle . \label{3.8}\ ] ] in addition , we should notice that @xmath109 and @xmath110 are related by the fourier transform ( ft ) , i.e. @xmath111 the integral kernel of the fraunhofer diffraction formula in 1-dimensional is such a ft , so ft in classical optics has its correspondence in quantum mechanical representations transform .
this enlightens us that in order to find more general analogy between unitary operators in quantum optics and transformations in classical optics we should construct new representations for quantum optics theory , and these are the bi - partite entangled state and many - particle entangled state .
these ideal states can be implemented by optical devices and optical network @xcite . in the following we focus on the bi - partite entangled state .
when we combine ( [ 3.04 ] ) and ( [ 3.05 ] ) we can obtain @xmath112 which is just the normally ordered wigner operator since its marginal integration gives @xmath113 and @xmath114 respectively , i.e. , @xmath115 thus the wigner function of quantum state @xmath22 can be calculated as @xmath116tr@xmath117 $ ] . on the other hand
, the wigner operator ( [ 3.06 ] ) can be recast into the coherent state representation , @xmath118 where @xmath119 is a coherent state .
in fact , using the iwop technique we can obtain @xmath120 which is the same as ( [ 3.06 ] ) .
the concept of quantum entanglement was first employed by einstein , rosen and poldosky ( epr ) to challenge that quantum mechanics is incomplete when they observed that two particles relative position @xmath121 and the total momentum @xmath122 are commutable .
hinted by epr , the bipartite entangled state @xmath6 is introduced as @xcite @xmath123 \left \vert 00\right \rangle . \label{3.11}\ ] ] @xmath124 is the common eigenstate of relative coordinate @xmath121 and the total momentum @xmath122 , @xmath125 where @xmath126 @xmath127 using the iwop technique , we can immediately prove that @xmath6 possesses the completeness relation @xmath128 \left [ \eta-(a_{1}-a_{2}^{\dagger})\right ] } \colon=1,\,d^{2}\eta = d\eta_{1}d\eta_{2 } , \label{3.13}\ ] ] and orthonormal relation@xmath129 the schmidt decomposition of @xmath6 is @xmath130 the @xmath6 state can also be schmidt - decomposed in momentum eigenvector space as @xmath131 the @xmath6 is physically appealing in quantum optics theory , because the two - mode squeezing operator has its natural representation on @xmath132 basis @xcite @xmath133 the proof of ( [ 3.23 ] ) is proceeded by virtue of the iwop technique@xmath134
so the necessity of introducing @xmath6 into quantum optics is clear .
@xmath135 squeezes @xmath6 in the manifest way @xmath136 and the two - mode squeezed state itself is an entangled state which entangles the idle mode and signal mode as an outcome of a parametric - down conversion process @xcite .
we can also introduce the conjugate state of @xmath6 @xcite , @xmath137 \left \vert 00\right \rangle , \text { } \xi=\xi_{1}+i\xi_{2 } , \label{3.17}\ ] ] which obeys the eigen - equations @xmath138 because @xmath139 = 2\mathtt{i},$ ] so we name the conjugacy between @xmath140 and @xmath141 the completeness and orthonormal relations of @xmath142 are@xmath143 \left [ \xi-(a_{1}+a_{2}^{\dagger})\right ] } \colon=1,\label{3.19a}\\ \left \langle \xi \right \vert \left .
\xi^{\prime}\right \rangle & = \pi \delta(\xi_{1}-\xi_{1}^{\prime})\delta(\xi_{2}-\xi_{2}^{\prime}),\text { } d^{2}\xi = d\xi_{1}d\xi_{2 } , \label{3.19}\ ] ] respectively . @xmath6 and @xmath144
can be related to each other by @xmath145 since @xmath146 is a pure imaginary number , eq .
( [ 3.20 ] ) is the fourier transform kernel in complex form ( or named entangled fourier transform , this concept should also be extended to multipartite entangled states . )
it will be shown in later sections that departing from entangled states @xmath6 and @xmath142 and the generalized fresnel operator a new entangled fresnel transforms in classical optics can be found . combining ( [ 3.13 ] ) and ( [ 3.19a ] ) we can construct the following operator @xmath147 \left [ \sigma-(a_{1}-a_{2}^{\dagger})\right ] -\left [ \gamma^{\ast}-(a_{1}^{\dagger}+a_{2})\right ] \left [ \gamma-(a_{1}+a_{2}^{\dagger})\right ] } \colon \nonumber \\ & = \delta \left ( \alpha,\alpha^{\ast}\right ) \otimes \delta \left ( \beta,\beta^{\ast}\right ) \equiv \delta \left ( \sigma,\gamma \right ) , \label{3.25}\ ] ] where@xmath148 eq .
( [ 3.25 ] ) is just equal to the direct product of two single - mode wigner operators .
it is convenient to express the wigner operator in the @xmath9 representation as @xcite @xmath149 for two - mode correlated system , it prefers to using @xmath150 to calculate quantum states wigner function .
for example , noticing @xmath151 the two - mode squeezed states wigner function is @xmath152 .
\label{3.28}\ ] ]
starting from the entangled state @xmath153 and introducing an integer @xmath65 , we can deduce new states @xcite , @xmath154 which is worth of paying attention because when we operate the number - difference operator , @xmath155 on @xmath156 using eq.([3.11 ] ) we see@xmath157 so the number - difference operator corresponds to a differential operation @xmath158 in the @xmath159representation , this is a remarkable property of @xmath160 .
it then follows@xmath161 on the other hand , by defining @xmath162 we see @xmath163 = 0,$ ] and @xmath164 is its eigenstate , @xmath165 where @xmath166 is named correlated - amplitude operator since @xmath167 thus we name @xmath164 correlated - amplitude number - difference entangled states .
it is not difficult to prove completeness and orthonormal property of @xmath164,@xmath168@xmath169 on the other hand , from @xmath142 we can derive another state @xmath170 which satisfies@xmath171 so @xmath172 in @xmath173 representation is equal to @xmath174 .
consequently , @xmath175 note @xmath176 = 0 $ ] and@xmath177 @xmath178 is qualified to be a new representation since @xmath179 since @xmath142 and @xmath6 are mutual conjugate , @xmath178 is the conjugate state of @xmath164 . from the definition of @xmath180 and @xmath178 and ( [ 3.20 ] )
we calculate the overlap @xcite@xmath181 d\theta d\varphi \nonumber \\ & = \frac{1}{8\pi^{2}}\int_{0}^{2\pi}\int_{0}^{2\pi}e^{is\varphi - im\theta}\sum_{l=-\infty}^{\infty}j_{l}\left ( rr^{\prime}\right ) e^{il\left ( \theta-\varphi \right ) } \nonumber \\ & = \frac{1}{2}\sum_{l=-\infty}^{\infty}\delta_{l , m}\delta_{l , s}j_{l}\left ( rr^{\prime}\right ) = \frac{1}{2}\delta_{s , m}j_{s}\left ( rr^{\prime}\right ) , \label{4.14}\ ] ] where we have identified the generating function of the s - order bessel function @xmath182@xmath183 and@xmath184 eq .
( [ 4.14 ] ) is remarkable , because @xmath185 is just the integral kernel of hankel transform .
in fact , if we define @xmath186 and use ( [ 4.7 ] ) as well as ( [ 4.14 ] ) , we obtain@xmath187 , \label{4.18}\ ] ] which is just the hankel transform of @xmath188 ( or it can be regarded as a simplified form of the collins formula in cylindrical coordinate , see ( [ 2.10 ] ) ) .
the inverse transform of ( [ 4.18 ] ) is@xmath189 .
\label{4.19}\ ] ] now we know that the quantum optical image of classical hankel transform just corresponds to the representation transformation between two mutually conjugate entangled states @xmath190 and @xmath191 this is like the case that the fourier transform kernel is just the matrix element between the coordinate state and the momentum state , a wonderful result unnoticed before . therefore the bipartite entangled state representations transforms , which can lead us to the hankel transform ,
was proposed first in classical optics , can find their way back in quantum optics . from eq.([4.1 ] ) we can see that its reciprocal relation is the circular harmonic expansion,@xmath192 or correlated - amplitude number - difference entangled state @xmath180 can be considered as circular harmonic decomposition of @xmath193 let @xmath194 , a general 2-dimensional function expressed in polar coordinates , be periodic in the variable @xmath195 it can be looked as the wavefunction of the state vector @xmath196 in the @xmath197 representation @xmath198 using ( [ 4.23 ] ) we have@xmath199 @xmath200 is the wavefunction of @xmath196 in @xmath201 representation . by using ( [ 4.7 ] ) and noticing that it then follows from ( [ 4.3])@xmath202 so @xmath203 behaves a rotation operator in @xmath6 representation , we see that the expectation value of @xmath204 in @xmath205 is@xmath206 which is just the cross - correlation between @xmath207 and an angularly rotated version of the same function , @xmath208 . on the other hand , using ( [ 4.7 ] ) we have@xmath209 substituting ( [ 4.28 ] ) into ( [ 4.27 ] ) and using the eigenvector equation ( [ 4.4 ] ) as well as ( [ 4.8 ] ) we obtain@xmath210 from which we see that each of the circular harmonic components of the crosscorrelation undergoes a different phase shift @xmath211 so @xmath212 is not rotation invariant .
however , when we consider only one harmonic component @xmath213 of a particular object is constructed , then if that some object is entered as an input to the system with any angular rotation , a correlation peak of strength proportional to @xmath214 will be produced , independent of rotation . hence an optical correlator can be constructed that will recognize that object independent of rotation @xcite .
so far we have studied the circular harmonic correlation in the context of quantum optics , we have endowed the crosscorrelation @xmath212 with a definite quantum mechanical meaning , i.e. the overlap between @xmath215 and the rotated state @xmath216 in the entangled state representation . note that fourier - based correlators is also very sensitive to magnification , however , the magnitude of mellin transform is independent of scale - size changes in the input @xcite .
now we examine when @xmath205 is both rotated and squeezed ( by a two - mode squeezing operator @xmath217)$ ] , then from ( [ 3.15 ] ) and ( [ 4.24 ] ) we have@xmath218 it follows the overlap between @xmath219 and the state @xmath220 @xmath221 which corresponds to the crosscorrelation arising from combination of squeezing and rotation ( joint transform correlator ) . on the other hand , from ( [ 4.1 ] ) and ( [ 3.15 ] ) we see @xmath222 and therefore@xmath223 from which one can see that to achieve simultaneous scale and rotation invariance , a two - dimensional object @xmath207 should be entered into the optical system in a distorted polar coordinate system , the distortation arising from the fact that the radial coordinate is stretched by a logarithmic transformation @xmath224 , which coincides with ref .
the quantum optical version is thus established which is a new tie connecting fourier optics and quantum optics @xcite . at the end of this section , using the two - variable hermite polynomials definition @xcite @xmath225 which is quite different from the product of two single - variable hermite polynomials , and its generating function formula is @xmath226 , \label{4.21}\ ] ] and noting @xmath227 we can directly perform the integral in ( [ 4.9 ] ) and derive the explicit form of @xmath228@xmath229 which is really an entangled state in two - mode fock space .
( [ 4.21 ] ) and ( [ 4.22 ] ) will be often used in the following discussions . in the following we concentrate on finding the generalized fresnel operators in both one- and two- mode cases with use of the iwop technique .
in this section we shall mainly introduce so - called generalized fresnel operators ( gfo ) ( in one - mode and two - mode cases both ) @xcite and some appropriate quantum optical representations ( e.g. coherent state representation and entangled state representation ) to manifestly link the formalisms in quantum optics to those in classical optics . in so doing , we find that the various transforms in classical optics are just the result of generalized fresnel operators inducing transforms on appropriate quantum state vectors , i.e. classical optical fresnel transforms have their counterpart in quantum optics . besides , we can study the important @xmath14 rule obeyed by gaussian beam propagation ( also the ray propagation in matrix optics ) @xcite in the domain of quantum optics . for the coherent state @xmath119 in quantum optics @xcite @xmath230 \begin{array } [ c]{c}z\\ z^{\ast}\end{array } \right ) \right \rangle , \label{3.2}\ ] ] which is the eigenstate of annihilation operator @xmath231 @xmath232 , using the iwop and ( [ 3.1 ] ) , we can put the over - completeness relation of @xmath119 into normal ordering@xmath233 the canonical form of coherent state @xmath119 is expressed as @xmath234 |0\rangle \equiv \left \vert \left ( \begin{array } [ c]{c}q\\ p \end{array } \right ) \right \rangle , \label{3.6}\ ] ] where @xmath235 .
it follows that @xmath236 @xmath237 this indicates that the states @xmath238 generate a canonical phase - space representation for a state @xmath239 @xmath240 thus the coherent state is a good candidate for providing with classical phase - space description of quantum systems . remembering that the fresnel transform s parameters @xmath241 are elements of a ray transfer matrix @xmath56 describing optical systems
, @xmath56 belongs to the unimodular symplectic group , and the coherent state @xmath242 is a good candidate for providing with classical phase - space description of quantum systems , we naturally think of that the symplectic transformation @xmath243{cc}a & b\\ c & d \end{array } \right ) \left ( \begin{array } [ c]{c}q\\ p \end{array } \right ) $ ] in classical phase space may mapping onto a generalized fresnel operator in hilbert space through the coherent state basis .
thus we construct the following ket - bra projection operator@xmath244{cc}a & b\\ c & d \end{array } \right ) \left ( \begin{array } [ c]{c}q\\ p \end{array } \right ) \right \rangle \left \langle \left ( \begin{array } [ c]{c}q\\ p \end{array } \right ) \right \vert \label{5.1}\ ] ] as the gfo .
in fact , using notation of @xmath119 ( coherent state)@xmath10 and introducing complex numbers @xmath245@xmath246 , \;r=-\frac{1}{2}\left [ a - d+i\left ( b+c\right ) \right ] , \;|s|^{2}-|r|^{2}=1 , \label{5.2}\ ] ] from ( [ 3.6 ] ) we know @xmath247{cc}a & b\\ c & d \end{array } \right ) \left ( \begin{array } [ c]{c}q\\ p \end{array } \right ) \right \rangle & = \left \vert \left ( \begin{array } [ c]{cc}s & -r\\ -r^{\ast } & s^{\ast}\end{array } \right ) \left ( \begin{array } [ c]{c}z\\ z^{\ast}\end{array } \right ) \right \rangle \equiv \left \vert sz - rz^{\ast}\right \rangle \nonumber \\ & = \exp \left [ -\frac{1}{2}|sz - rz^{\ast}|^{2}+(sz - rz^{\ast})a^{\dagger } \right ] |0\rangle , \label{5.3}\ ] ] @xmath243{cc}s & -r\\ -r^{\ast } & s^{\ast}\end{array } \right ) $ ] is still a symplectic group element , so ( [ 5.1 ] ) becomes @xcite@xmath248 where the factor @xmath249 is attached for anticipating the unitarity of the operator @xmath250 eq .
( [ 5.1 ] ) tells us that @xmath80-number transform @xmath251 in coherent state basis maps into @xmath252 .
now we prove @xmath253 is really the fo we want . using the iwop technique and eq .
( [ 5.3 ] ) and ( [ 3.1 ] ) we can perform the integral@xmath254 \colon \nonumber \\ & = \frac{1}{\sqrt{s^{\ast}}}\exp \left ( -\frac{r}{2s^{\ast}}a^{\dagger 2}\right ) \colon \exp \left \ { \left ( \frac{1}{s^{\ast}}-1\right ) a^{\dagger } a\right \ } \colon \exp \left ( \frac{r^{\ast}}{2s^{\ast}}a^{2}\right ) \nonumber \\ & = \exp \left ( -\frac{r}{2s^{\ast}}a^{\dagger2}\right ) \exp \left \ { \left ( a^{\dagger}a+\frac{1}{2}\right ) \ln \frac{1}{s^{\ast}}\right \ } \exp \left ( \frac{r^{\ast}}{2s^{\ast}}a^{2}\right ) , \label{5.5}\ ] ] where we have used the mathematical formula @xcite @xmath255 with the convergent condition @xmath256 it then follows @xmath257 .
\label{5.8}\ ] ] then using @xmath258 and the completeness relation of coherent state as well as ( [ 5.2 ] ) we obtain the matrix element of @xmath252 ( @xmath259 ) in coordinate representation @xmath260 , @xmath261 \equiv \mathcal{k}\left ( x_{2},x_{1}\right ) , \label{5.10}\ ] ] which is just the kernel of generalized fresnel transform @xmath262 in ( [ 2.12 ] ) .
the above discussions demonstrate how to transit classical fresnel transform to gfo through the coherent state and the iwop technique . now if we define @xmath263 , @xmath264 and using eq .
( [ 3.10 ] ) , we can rewrite eq .
( [ 2.11 ] ) as@xmath265 which is just the quantum mechanical version of gfo . therefore
, the 1-dimensional gft in classical optics corresponds to the 1-mode gfo @xmath266 operating on state vector @xmath267 in hilbert space , i.e. @xmath268 . one merit of gfo is : using coordinate - momentum representation transform we can immediately obtain gft in frequency " domain , i.e. @xmath269 \nonumber \\ & = \frac{1}{\sqrt{2\pi i\left ( -c\right ) } } \exp \left [ \frac{i}{2\left ( -c\right ) } \left ( dp_{1}^{2}-2p_{2}p_{1}+ap_{2}^{2}\right ) \right ] .
\label{5.12}\ ] ] obviously , @xmath266 induces the following transform @xmath270{c}q\\ p \end{array } \right ) f_{1}\left ( a , b , c\right ) = \left ( \begin{array } [ c]{cc}a & b\\ c & d \end{array } \right ) \left ( \begin{array } [ c]{c}q\\ p \end{array } \right ) .
\label{5.13}\ ] ] because two successive optical fresnel transforms is still a fresnel transform , we wonder if the product of two gfo is still a gfo . on the other hand
, we have known that the gfo is the image of the symplectic transform @xmath271 we expect that the product of two symplectic transforms maps into the gfo which is just the product of two gfos . if this is so , then correspondence between gft and gfo is perfect . using ( [ 5.4 ] ) , @xmath272 $ ] and the iwop technique we can directly perform the following integrals@xmath273 \colon \exp \left \
{ \left ( \frac{1}{s^{\prime \prime \ast}}-1\right ) a^{\dagger}a\right \ } \colon \exp \left [ \frac{r^{\prime \prime \ast}}{2s^{\prime \prime \ast}}a^{2}\right ] \nonumber \\ & = \sqrt{s^{\prime \prime}}\int \frac{d^{2}z}{\pi}\left \vert s^{\prime \prime \ast}z - r^{\prime \prime}z^{\ast}\right \rangle \left \langle z\right \vert = f_{1}\left ( s^{\prime \prime},r^{\prime \prime}\right ) , \label{7.3}\ ] ] where we have set@xmath274 or @xmath275{cc}s^{\prime \prime } & -r^{\prime \prime}\\ -r^{\ast \prime \prime } & s^{\ast \prime \prime}\end{array } \right ) = \left ( \begin{array } [ c]{cc}s & -r\\ -r^{\ast } & s^{\ast}\end{array } \right ) \left ( \begin{array } [ c]{cc}s^{\prime } & -r^{\prime}\\ -r^{\prime \ast } & s^{\prime \ast}\end{array } \right ) = mm^{\prime},\ \left \vert s^{\prime \prime}\right \vert ^{2}-\left \vert r^{\prime \prime}\right \vert ^{2}=1 , \label{7.4}\ ] ] from which we see that it is just the mapping of the above ( @xmath276 matrices multiplication .
hence @xmath277 is the loyal representation of the product of two symplectic group elements shown in ( [ 7.4 ] ) .
the above discussion actually reveals an important property of coherent states , though two coherent state vectors are not orthogonal , but the equation @xmath278 seems as if their overlap @xmath279 was a @xmath280-function .
the coherent state representation for gfos product may be visualized very easily , but it achieves striking importance , because it does not change its form when treating symplectic transform according to @xmath281 . as a result of this group multiplication rule of gfo , we immediately obtain @xmath282 provided that the parameter matrices @xmath283 satisfy ( [ 7.2 ] ) .
thus by virtue of the group multiplication property of gfo we immediately find the successive transform property of gfts .
in classical optics , ray - transfer matrices , @xmath284{cc}a & b\\ c & d \end{array } \right ) , $ ] @xmath53 , have been used to describe the geometrical formation of images by a centered lens system . for an optical ray ( a centered spherical wavefront ) passing through optical instruments
there is a famous law , named abcd law , governing the relation between input ray @xmath285 and output ray @xmath286
i.e. @xmath287{c}r_{2}\\ \alpha_{2}\end{array } \right ) = n\left ( \begin{array } [ c]{c}r_{1}\\ \alpha_{1}\end{array } \right ) , \label{8.1}\ ] ] where @xmath288 is the ray height from the optical axis , and @xmath289 is named the optical direction - cosine , @xmath290 specifies the ray s wavefront shape .
( [ 8.1 ] ) implies @xmath291 this law is the core of matrix optics , since it tells us how the curvature of a centered spherical wavefront changes from one reference plane to the next .
besides , the multiplication rule of matrix optics implies that if the ray - transfer matrices of the @xmath292 optical components are @xmath293 , respectively , then the whole system is determined by a matrix @xmath294 one of the remarkable things of modern optics is the case with which geometrical ray - transfer methods , constituting the matrix optics , can be adapted to describe the generation and propagation of laser beams . in 1965
kogelnik @xcite pointed out that propagation of gaussian beam also obeys abcd law via optical diffraction integration , i.e. the input light field @xmath295 and output light field @xmath296 are related to each other by so - called fresnel integration @xcite
@xmath297 where@xmath298 .\ ] ] the abcd law for gaussian beam passing through an optical system is @xcite@xmath299 where @xmath300 @xmath301 represents the complex curvature of the input ( output ) gaussian beam , eq . ( [ 8.3 ] ) has the similar form as eq .
( [ 8.2 ] ) .
an interesting and important question naturally arises @xcite : does abcd law also exhibit in quantum optics ? since classical fresnel
transform should have its quantum optical counterpart ? to see the abcd law more explicitly , using eq.([5.2 ] ) we can re - express eq.([5.5 ] ) as @xmath302 } a^{\dagger2}\right .
\nonumber \\ & \left .
+ \left [ \frac{2}{a+d+i\left ( b - c\right ) } -1\right ] a^{\dagger } a-\frac{a - d - i\left ( b+c\right ) } { 2\left [ a+d+i\left ( b - c\right ) \right ] } a^{2}\right \ } \colon , \label{8.4}\ ] ] and the multiplication rule for @xmath303 is @xmath304 where@xmath287{cc}a^{\prime \prime } & b^{\prime \prime}\\ c^{\prime \prime } & d^{\prime \prime}\end{array } \right ) = \left ( \begin{array } [ c]{cc}a^{\prime } & b^{\prime}\\ c^{\prime } & d^{\prime}\end{array } \right ) \left ( \begin{array } [ c]{cc}a & b\\ c & d \end{array } \right ) .
\label{8.5}\ ] ] next we directly use the gfo to derive abcd law in quantum optics . from eq.([8.5 ] ) we see that the gfo generates @xmath305 } a^{\dagger2}\right \ } \left \vert 0\right \rangle , \label{8.6}\ ] ] if we identify @xmath306 then@xmath307 \left \vert 0\right \rangle , \label{8.8}\ ] ] the solution of eq.([8.7 ] ) is @xmath308 let @xmath309 expressed by ( [ 8.8 ] ) be an input state for an optical system which is characteristic by parameters @xmath310 then the _ quantum optical abcd law
_ states that the output state is @xmath311 \left \vert 0\right \rangle , \label{8.10}\ ] ] which has the similar form as eq.([8.8])@xmath10 where @xmath312 is determined by eq.([8.5])@xmath10 and @xmath313 which resembles eq.([8.3 ] ) .
proof : according to the multiplication rule of two gfos and eqs.([8.4])-([8.5 ] ) we have@xmath314 } a^{\dagger2}\right \ } \left \vert 0\right \rangle \nonumber \\ & = \sqrt{\frac{2}{a^{\prime}\left ( a+ib\right ) + b^{\prime}\left ( c+id\right ) -ic^{\prime}\left ( a+ib\right ) -id^{\prime}\left ( c+id\right ) } } \nonumber \\ & \times \exp \left \ { \frac{a^{\prime}\left ( a+ib\right ) + b^{\prime}\left ( c+id\right ) + ic^{\prime}\left ( a+ib\right ) + id^{\prime}\left ( c+id\right ) } { 2\left [ a^{\prime}\left ( a+ib\right ) + b^{\prime}\left ( c+id\right ) -ic^{\prime}\left ( a+ib\right ) -id^{\prime}\left ( c+id\right ) \right ] } a^{\dagger2}\right \ } \left \vert 0\right \rangle \nonumber \\ & = \sqrt{\frac{-2/\left ( c+id\right ) } { a^{\prime}q_{1}-b^{\prime}-i\left ( c^{\prime}q_{1}-d^{\prime}\right ) } } \exp \left \ { \frac{a^{\prime}q_{1}-b^{\prime}+i\left ( c^{\prime}q_{1}-d^{\prime}\right ) } { 2\left [ a^{\prime}q_{1}-b^{\prime}-i\left ( c^{\prime}q_{1}-d^{\prime}\right ) \right ] } a^{\dagger2}\right \ } \left \vert 0\right \rangle .
\label{8.12}\ ] ] using eq.([8.9 ] ) we see @xmath315 together using eq.([8.11 ] ) we can reach eq.([8.10 ] ) , thus the law is proved . using eq .
( [ 8.8 ] ) we can re - express eq.([8.11 ] ) as @xmath316 which is in consistent to eq.([8.9 ] ) .
( [ 8.8])-([8.13 ] ) are therefore self - consistent .
as an application of quantum optical abcd law , we apply it to tackle the time - evolution of a time - dependent harmonic oscillator whose hamiltonian is @xmath317 where we have set the initial mass @xmath318 @xmath319 denotes damping . using @xmath320to perform the transformation@xmath321 then @xmath322leads to @xmath323 @xmath324 @xmath325 where @xmath326 @xmath327 does not contain @xmath328 explicitly .
the dynamic evolution of a mass - varying harmonic oscillator from the fock state @xmath329 at initial time to a squeezed state at time @xmath328 is @xmath330 if we let @xmath331 and @xmath332 then @xmath333 @xmath334 , according to eq.([8.10 ] ) we directly obtain@xmath335 \left \vert 0\right \rangle , \label{8.18}\ ] ] so the time evolution of the damping oscillator embodies the quantum optical abcd law .
fresnel diffraction is the core of fourier optics @xcite , fresnel transform is frequently used in optical imaging , optical propagation and optical instrument design .
the gft represents a class of optical transforms which are of great importance for their applications to describe various optical systems .
it is easily seen that when we let the transform kernel @xmath336 , the gft changes into the well - known fourier transform , which is adapted to express mathematically the fraunhofer diffraction . and
if @xmath337 $ ] , the gft then describes a fresnel diffraction . in studying various optical transformations
one also proposed so - called optical operator method @xcite which used quantum mechanical operators ordered product to express the mechanism of optical systems , such that the ray transfer through optical instruments and the diffraction can be discussed by virtue of the commutative relations of operators and the matrix algebra .
two important questions thus naturally arises : how to directly map the classical optical transformations to the optical operator method ? how to combine the usual optical transformation operators , such as the square phase operators , scaling operator , fourier transform operator and the propagation operator in free space , into a concise and unified form ? in this section
we shall solve these two problems and develop the optical operator method onto a new stage .
the gfo @xmath266 can also be expressed in the form of quadratic combination of canonical operators @xmath338 and @xmath339 @xcite , i.e. , @xmath340 where we have set @xmath341 @xmath342 to confirm this , we first calculate matrix element @xmath343 where we have used the squeezing property @xmath344 \left \vert p\right \rangle = \frac{1}{\sqrt{a}}\left \vert p / a\right \rangle . \label{9.3}\ ] ]
it then follows from ( [ 9.2 ] ) and @xmath53 , we have @xmath345 thus @xmath266 in ( [ 9.1 ] ) is really the expected gfo .
next we directly use ( [ 5.1 ] ) and the canonical operator @xmath346 representation ( [ 9.1 ] ) to develop the optical operator method . by noticing the matrix decompositions @xcite@xmath287{cc}a & b\\ c & d \end{array }
\right ) = \left ( \begin{array } [ c]{cc}1 & 0\\ c / a & 1 \end{array } \right ) \left ( \begin{array } [ c]{cc}a & 0\\ 0 & a^{-1}\end{array } \right ) \left ( \begin{array } [ c]{cc}1 & b / a\\ 0 & 1 \end{array } \right ) , \label{9.5}\ ] ] and comparing ( [ 5.1 ] ) and ( [ 9.1 ] ) as well as using ( [ 8.5 ] ) we know@xmath347 where@xmath348{cc}1 & 0\\ c / a & 1 \end{array } \right ) \left ( \begin{array } [ c]{c}x\\ p \end{array } \right ) \right \rangle \left \langle \left ( \begin{array } [ c]{c}x\\ p \end{array } \right ) \right \vert \nonumber \\ & = \exp \left ( \frac{ic}{2a}q^{2}\right ) , \label{9.7}\ ] ] which is named quadrature phase operator ; and @xmath349{cc}1 & b / a\\ 0 & 1 \end{array } \right ) \left ( \begin{array } [ c]{c}x\\ p \end{array } \right ) \right \rangle \left \langle \left ( \begin{array } [ c]{c}x\\ p \end{array } \right ) \right \vert \nonumber \\ & = \exp \left ( -\frac{ib}{2a}p^{2}\right ) , \label{9.8}\ ] ] which is named fresnel propagator in free space ; as well as@xmath350{cc}a & 0\\ 0 & a^{-1}\end{array } \right ) \left ( \begin{array } [ c]{c}x\\ p \end{array } \right ) \right \rangle \left \langle \left ( \begin{array } [ c]{c}x\\ p \end{array } \right ) \right \vert \nonumber \\ & = \exp \left [ -\frac{i}{2}\left ( qp+pq\right ) \ln a\right ] , \label{9.9}\ ] ] which is named scaling operator ( squeezed operator @xcite ) . when @xmath351 from ( [ 8.4 ] ) we see @xmath352{cc}0 & 1\\ -1 & 0
\end{array } \right ) \left ( \begin{array } [ c]{c}x\\ p \end{array } \right ) \right \rangle \left \langle \left ( \begin{array } [ c]{c}x\\ p \end{array } \right ) \right \vert \nonumber \\ & = \exp \left [ -\left ( a^{\dagger}a+\frac{1}{2}\right ) \ln i\right ] \nonumber \\ & = \exp \left [ -i\frac{\pi}{2}\left ( a^{\dagger}a+\frac{1}{2}\right ) \right ] , \label{9.10}\ ] ] which is named the fourier operator , since it quantum mechanically transforms @xcite @xmath353 q\exp \left [ -i\frac{\pi}{2}\left ( a^{\dagger}a+\frac{1}{2}\right ) \right ] & = p,\nonumber \\
\left [ i\frac{\pi}{2}\left ( a^{\dagger}a+\frac{1}{2}\right ) \right ] p\exp \left [ -i\frac{\pi}{2}\left ( a^{\dagger}a+\frac{1}{2}\right ) \right ] & = -q . \label{9.11}\ ] ] note that when @xmath67 , the decomposition ( [ 9.1 ] ) is not available , instead , from@xmath287{cc}a & b\\ c & d \end{array } \right ) ^{-1}=\left ( \begin{array } [ c]{cc}d & -b\\ -c & a \end{array } \right ) , \label{9.12}\ ] ] and ( [ 8.4 ] ) , ( [ 8.5 ] ) and ( [ 9.1 ] ) we have@xmath354 \exp \left ( \frac{ib}{2d}p^{2}\right ) , \label{9.13}\ ] ] it then follows @xmath355 \exp \left ( \frac{ic}{2d}q^{2}\right ) , \text { } d\neq0 . \label{9.14}\ ] ] besides , when we notice @xmath287{cc}a & b\\ c & d \end{array } \right ) = \left ( \begin{array } [ c]{cc}1 & 0\\ d / b & 1 \end{array } \right ) \left ( \begin{array } [ c]{cc}b & 0\\ 0 & 1/b \end{array } \right ) \left ( \begin{array } [ c]{cc}0 & 1\\ -1 & 0 \end{array } \right ) \left ( \begin{array } [ c]{cc}1 & 0\\ a / b & 1 \end{array } \right ) , \label{9.15}\ ] ] and @xmath287{cc}a & b\\ c & d \end{array } \right ) = \left ( \begin{array } [ c]{cc}1 & a / c\\ 0 & 1 \end{array } \right ) \left ( \begin{array } [ c]{cc}-1/c & 0\\ 0 & -c \end{array } \right ) \left ( \begin{array } [ c]{cc}0 & 1\\ -1 & 0 \end{array } \right ) \left ( \begin{array } [ c]{cc}1 & d / c\\ 0 & 1 \end{array } \right ) , \label{9.16}\ ] ] we have another decomposition for @xmath356 @xmath357 \exp \left ( \frac{ia}{2b}q^{2}\right ) , \text { } \label{9.17}\ ] ] and for @xmath358 @xmath359 \nonumber \\ & \times \exp \left [ -\frac{i\pi}{2}\left ( a^{\dagger}a+\frac{1}{2}\right ) \right ] \exp \left ( -\frac{id}{2c}p^{2}\right ) .
\label{9.18}\ ] ] for a special optical systems with the parameter @xmath360 @xmath361@xmath287{cc}0 & b\\ -b^{-1 } & d \end{array } \right ) = \left ( \begin{array } [ c]{cc}1 & 0\\ d / b & 1 \end{array } \right ) \left ( \begin{array } [ c]{cc}b & 0\\ 0 & b^{-1}\end{array } \right ) \left ( \begin{array } [ c]{cc}0 & 1\\ -1 & 0 \end{array } \right ) , \label{9.20}\ ] ] we have@xmath362 .
\label{9.21}\ ] ] in particular , when @xmath363 @xmath361 from@xmath287{cc}0 & b\\ -\frac{1}{b } & 0 \end{array } \right ) = \left ( \begin{array } [ c]{cc}b & 0\\ 0 & \frac{1}{b}\end{array } \right ) \left ( \begin{array } [ c]{cc}0 & 1\\ -1 & 0 \end{array } \right ) , \label{9.22}\ ] ] we have@xmath364 \exp \left [ \left ( a^{\dagger}a+\frac{1}{2}\right ) \ln \left ( \frac { -2bi}{b^{2}+1}\right ) \right ] \exp \left [ -\frac{b^{2}-1}{2\left ( b^{2}+1\right ) } a^{2}\right ] \nonumber \\ & = \exp \left [ -\frac{i}{2}\left ( qp+pq\right ) \ln b\right ] \exp \left [ -i\frac{\pi}{2}\left ( a^{\dagger}a+\frac{1}{2}\right ) \right ] .
\label{9.23}\ ] ] using the following relations @xmath287{cc}a & b\\ c & d \end{array } \right ) = \left ( \begin{array } [ c]{cc}1 & \left ( a-1\right ) /c\\ 0 & 1 \end{array }
\right ) \left ( \begin{array } [ c]{cc}1 & 0\\ c & 1 \end{array } \right ) \left ( \begin{array } [ c]{cc}1 & \left ( d-1\right ) /c\\ 0 & 1 \end{array } \right ) , \label{9.24}\ ] ] it then follows that @xmath365 while from@xmath287{cc}a & b\\ c & d \end{array } \right ) = \left ( \begin{array } [ c]{cc}1 & 0\\ \left ( d-1\right ) /b & 1 \end{array } \right ) \left ( \begin{array } [ c]{cc}1 & b\\ 0 & 1 \end{array } \right ) \left ( \begin{array } [ c]{cc}1 & 0\\ \left ( a-1\right ) /b & 1 \end{array } \right ) \label{9.26}\ ] ] we obtain@xmath366 so we have @xmath367 in this section , based on a one - to - one correspondence between classical fresnel transform in phase space and quantum unitary transform in state - vector space and the iwop technique as well as the coherent state representation we have found a way to directly map the classical optical transformations to the optical operator method .
we have combined the usual optical transformation operators , such as the square phase operators , scaling operator , fourier transform operator and the propagation operator in free space , into a concise and unified form .
the various decompositions of fresnel operator into the exponential canonical operators are also obtained .
in quantum optics theory all possible linear combinations of quadratures @xmath338 and @xmath339 of the oscillator field mode @xmath76 and @xmath75 can be measured by the homodyne measurement just by varying the phase of the local oscillator .
the average of the random outcomes of the measurement , at a given local oscillator phase , is connected with the marginal distribution of wigner function ( wf ) , thus the homodyne measurement of light field permits the reconstruction of the wf of a quantum system by varying the phase shift between two oscillators . in ref .
@xcite vogel and risken pointed out that the probability distribution for the rotated quadrature phase @xmath368/\sqrt{2},$ ] @xmath369 = 1,$]which depends on only one @xmath370 angle , can be expressed in terms of wf , and that the reverse is also true ( named as vogel - risken relation ) , i.e. , one can obtain the wigner distribution by tomographic inversion of a set of measured probability distributions , @xmath371 of the quadrature amplitude .
once the distribution @xmath372 are obtained , one can use the inverse radon transformation familiar in tomographic imaging to obtain the wigner distribution and density matrix .
the radon transform of the wf is closely related to the expectation values or densities formed with the eigenstates to the rotated canonical observables .
the field of problems of the reconstruction of the density operator from such data is called quantum tomography .
( optical tomographic imaging techniques derive two - dimensional data from a three - dimensional object to obtain a slice image of the internal structure and thus have the ability to peer inside the object noninvasively , the slice image is equated with tomogram . ) the theoretical development in quantum tomography in the last decade has progressed in the direction of determining more physical relevant parameters of the density from tomographic data@xcite . in @xcite
the radon transform of wf which depends on two continuous parameters is introduced , this has the advantage in conveniently associating quantum tomography theory with squeezed coherent state theory . in this subsection
we want to derive relations between the fresnel transform and the radon transform of wf in quantum optics in tomography theory . by extending the rotated quadrature phase @xmath373 to the fresnel quadrature phase @xmath374 where @xmath375 and @xmath376 are related to abcd through([5.2 ] ) , @xmath246 , \;r=-\frac{1}{2}\left [ a - d+i\left ( b+c\right ) \right ] , \;|s|^{2}-|r|^{2}=1,\ ] ] we shall prove that the @xmath377 related radon transform of wigner operator @xmath378 is just the pure state density operator @xmath379 ( named as the tomographic density operator ) formed with the eigenstates belonging to the quadrature @xmath380 , ( @xmath381 @xmath338 is the coordinate operator),@xmath382 \delta \left ( q^{\prime},p^{\prime}\right ) , \label{10.2}\]]@xmath383 since @xmath13 corresponds to classical fresnel transform in optical diffraction theory , so eq . (
[ 10.2 ] ) indicates that the probability distribution for the fresnel quadrature phase is the radon transform of wf @xcite .
proof : firstly , from ( [ 5.5 ] ) we see @xmath384 so from @xmath385 indeed we have@xmath386 secondly , we can derive the explicit form of @xmath387 starting from @xmath388 @xmath389 we set up the eigenvector equation@xmath390 it follows@xmath391 in the coordinate and momentum representations we have@xmath392 the normalizable solutions to ( [ 10.9 ] ) and ( [ 10.10 ] ) are@xmath393 , \label{10.11}\\ \left \langle p\right \vert \left . q\right \rangle _ { s , r } & = d\left ( q\right ) \exp \left [ \frac{ip\left ( -bp-2q\right ) } { 2d}\right ]
. \label{10.12}\ ] ] using the fock representation of @xmath109 and @xmath106 in eqs.([3.7 ] ) and ( [ 3.8 ] ) , we obtain@xmath394 \left \vert 0\right \rangle , \label{10.13}\ ] ] and@xmath395 \left \vert 0\right \rangle . \label{10.14}\ ] ] comparing eq.([10.13 ] ) with ( [ 10.14 ] ) we see@xmath396
. \label{10.15}\ ] ] on the other hand , according to the orthogonalization of @xmath397 , @xmath398 we have @xmath399 thus combining eq.([10.15 ] ) and ( [ 10.16 ] ) we deduce @xmath400 , \text { } d\left ( q\right ) = \frac{1}{\sqrt{2\pi d}}\exp \left [ \frac{icq^{2}}{2d}\right ] , \label{10.17}\ ] ] and @xmath401 or@xmath402 it is easily seen that that @xmath403 make up a complete set ( so @xmath403 can be named as the tomography representation),@xmath404 then according to the weyl quantization scheme @xcite@xmath405 where @xmath406 is the weyl correspondence of @xmath407 @xmath408 , \label{10.22}\ ] ] @xmath378 is the wigner operator @xcite,@xmath409 and using ( [ 10.22 ] ) , ( [ 10.23 ] ) and ( [ 10.13 ] ) we know that the classical weyl correspondence ( weyl image ) of the projection operator @xmath379 is @xmath410 \nonumber \\ & = \left . _
{ s , r}\left
\langle q\right \vert \right .
\int_{-\infty}^{\infty } due^{ip^{\prime}u}\left \vert q^{\prime}+\frac{u}{2}\right \rangle \left \langle q^{\prime}-\frac{u}{2}\right \vert \left .
q\right \rangle _ { s , r}\nonumber \\ & = \frac{1}{2\pi b}\int_{-\infty}^{\infty}du\exp \left [ ip^{\prime}u+\frac { i}{b}u\left ( q - dq^{\prime}\right ) \right ] \nonumber \\ & = \delta \left [ q-\left ( dq^{\prime}-bp^{\prime}\right ) \right ] , \label{10.24}\ ] ] which means@xmath411 \delta \left ( q^{\prime},p^{\prime}\right ) . \label{10.25}\ ] ] combining eqs .
( [ 10.4])-([10.7 ] ) together we complete the proof .
therefore , the probability distribution for the fresnel quadrature phase is the radon transform of wf@xmath412 \left \langle \psi \right \vert \delta \left ( q^{\prime},p^{\prime}\right ) \left \vert \psi \right \rangle , \label{10.26}\ ] ] so we name @xmath379 * the tomographic density*.
moreover , the tomogram of quantum state @xmath413 is just the squared modulus of the wave function @xmath414 this new relation between quantum tomography and optical fresnel transform may provide experimentalists to figure out new approach for generating tomography .
the introduction of @xmath403 also bring convenience to obtain the inverse of radon transformation , using ( [ 10.20 ] ) we have@xmath415 considering its right hand - side as a fourier transformation , its reciprocal transform is@xmath416 , \end{aligned}\ ] ] where @xmath417 @xmath418 @xmath419so once the distribution @xmath420 related radon transform of @xmath378 is just the pure state density operator @xmath421 formed with the eigenstates belonging to the conjugate quadrature of @xmath422@xmath423 \delta \left ( q^{\prime},p^{\prime}\right ) , \nonumber \\ a & = \frac{1}{2}\left ( s^{\ast}-r^{\ast}+s - r\right ) , \text { \ } c=\frac { 1}{2i}\left ( s - r - s^{\ast}+r^{\ast}\right ) .
\label{10.27}\ ] ] similarly , we find that for the momentum density , @xmath424
\delta \left ( q^{\prime},p^{\prime}\right ) , \label{10.28}\ ] ] where @xmath425 as an application of the relation ( [ 10.2 ] ) , recalling that the @xmath426 makes up a faithful representation of the symplectic group @xcite , it then follows from ( [ 10.2 ] ) that @xmath427 \delta \left ( q^{\prime } , p^{\prime}\right ) , \label{10.30}\ ] ] in this way a complicated radon transform of tomography can be viewed as the sequential operation of two fresnel transforms .
this confirms that the continuous radon transformation corresponds to the symplectic group transformation @xcite , this is an advantage of introducing the fresnel operator .
the group property of fresnel operators help us to analyze complicated radon transforms in terms of some sequential fresnel transformations
. the new relation may provide experimentalists to figure out new approach for realizing tomography . in this subsection
, we introduce a new theorem , i.e. , the tomogram of a density operator @xmath22 is equal to the marginal integration of the classical weyl correspondence function of @xmath428 where @xmath13 is the fresnel operator . multiplying both sides of eq .
( [ 10.25 ] ) by a density matrix @xmath22 and then performing the trace , noting the wigner function @xmath429 , $ ] one can see@xmath430 \delta \left ( q\acute{},p\acute{}\right ) \rho \right ] \nonumber \\ & = tr\left ( \left \vert q\right \rangle _ { s , rs , r}\left \langle q\right \vert \rho \right ) = _ { s , r}\left \langle q\right \vert \rho \left \vert q\right \rangle _ { s , r}=\left
\langle q\right \vert f^{\dagger}\rho f\left \vert q\right \rangle \nonumber \\ & = { \displaystyle \iint \nolimits_{-\infty}^{\infty } } dq\acute{}dp\acute{}\delta \left [ q-\left ( dq\acute{}-bp\acute{}\right ) \right ] w(p , q ) .
\label{j1}\ ] ] the right hand side of eq .
( [ j1 ] ) is commonly defined as the tomogram of quantum states in @xmath431 direction , so in our view the calculation of tomogram in @xmath431 direction is ascribed to calculating @xmath432 this is a concise and neat formula .
similarly , the tomogram in @xmath433 direction is ascribed to @xmath434 . according to the weyl correspondence rule @xmath435 and the weyl ordering form of @xmath436 @xmath437 where the symbol@xmath438 denotes weyl ordering , the classical correspondence of a weyl ordered operator @xmath439
is obtained just by replacing @xmath440 in @xmath441 i.e. , @xmath442 let the classical weyl correspondence of @xmath23 be @xmath443@xmath444 then using ( [ j14 ] ) and ( [ 10.23 ] ) we have @xmath445 thus we reach a theorem : the tomogram of a density operator @xmath22 is equal to the marginal integration of the classical weyl correspondence @xmath443 of @xmath428 where @xmath13 is the fresnel operator , expressed by@xmath446 = \int_{-\infty}^{\infty}\frac{dp}{2\pi}h(p , q ) , \label{19}\ ] ] or@xmath447 = \int_{-\infty}^{\infty}\frac{dx}{2\pi}h(p , q ) .
\label{20}\ ] ] in this way the relationship between tomogram of a density operator @xmath22 and the fresnel transformed @xmath448s classical weyl function is established .
for two - dimensional optical fresnel transforms ( see ( [ 2.5 ] ) ) in the @xmath449 plane one may naturally think that the 2-mode gfo is just the direct product of two independent 1-mode gfos but with the same @xmath450 matrix .
however , here we present another 2-mode fresnel operator which can not only lead to the usual 2-dimensional optical fresnel transforms in some appropriate quantum mechanical representations , but also provide us with some new classical transformations ( we name them entangled fresnel transformations ) .
similar in spirit to the single - mode case , we introduce the two - mode gfo @xmath451 through the following 2-mode coherent state representation @xcite @xmath452 which indicates that @xmath451 is a mapping of classical sympletic transform @xmath453 in phase space . concretely , the ket in ( [ 11.1 ] ) is @xmath454 @xmath375 and @xmath376 are complex and satisfy the unimodularity condition . using the iwop technique we perform the integral in ( [ 11.1 ] ) and obtain@xmath455\colon \nonumber \\ & = \frac{1}{s^{\ast}}\exp \left ( \frac{r}{s^{\ast}}a_{1}^{\dagger}a_{2}^{\dagger}\right ) \colon \exp \left [ \left ( \frac{1}{s^{\ast}}-1\right ) \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) \right ] \colon \exp \left ( -\frac{r^{\ast}}{s^{\ast}}a_{1}a_{2}\right ) \nonumber \\ & = \exp \left ( \frac{r}{s^{\ast}}a_{1}^{\dagger}a_{2}^{\dagger}\right ) \exp[\left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}+1\right ) \ln \left ( s^{\ast}\right ) ^{-1}]\exp \left ( -\frac{r^{\ast}}{s^{\ast}}a_{1}a_{2}\right ) .
\label{11.3}\ ] ] thus @xmath451 induces the transform @xmath456 and @xmath457 is actually a general 2-mode squeezing operator . recall that ( [ 3.15 ] ) implies the intrinsic relation between the epr entangled state and the two - mode squeezed state , which has physical implementation , i.e. in the output of a parametric down - conversion the idler - mode and the signal - mode constitute a two - mode squeezed state , meanwhile are entangled with each other in frequency domain , we naturally select the entangled state representation to relate @xmath451 to two - dimensional gft . letting @xmath458 and then projecting @xmath413 onto the entangled state @xmath459 defined by ( [ 3.11 ] ) and using the completeness relation ( [ 3.13 ] ) of @xmath6
, we obtain @xmath460 then using the overcompleteness relation of the coherent state , and@xmath461 we can calculate the integral kernel@xmath462 \nonumber \\ & \times \exp \left [ -\frac{r^{\ast}}{s^{\ast}}z_{1}z_{2}+z_{1}^{\ast}z_{2}^{\ast}+\eta z_{1}^{\ast}+\frac{1}{s^{\ast}}\left ( z_{1}^{\prime \ast } z_{1}+z_{2}^{\prime \ast}z_{2}\right ) + \frac{r}{s^{\ast}}z_{1}^{\prime \ast } z_{2}^{\prime \ast}+z_{1}^{\prime}z_{2}^{\prime}+\eta^{\prime \ast}z_{1}^{\prime}-\eta^{\prime}z_{2}^{\prime}-\eta^{\ast}z_{2}^{\ast}\right ] \nonumber \\ & = \frac{1}{\left ( -r - s+r^{\ast}+s^{\ast}\right ) \pi}\exp \left [ \frac{\left ( -s+r^{\ast}\right ) \left \vert \eta \right \vert ^{2}-\left ( r+s\right ) \left \vert \eta^{\prime}\right \vert ^{2}+\eta \eta^{\prime \ast } + \eta^{\ast}\eta^{\prime}}{-r - s+r^{\ast}+s^{\ast}}-\frac{1}{2}\left ( \left \vert \eta^{\prime}\right \vert ^{2}+\left \vert \eta
\right \vert ^{2}\right ) \right ] .
\label{11.6}\ ] ] using the relation between @xmath463 and @xmath241 in eq.([5.2 ] ) eq .
( [ 11.6 ] ) becomes@xmath464 \equiv \mathcal{k}_{2}^{m}\left ( \eta^{\prime},\eta \right ) , \label{11.7}\ ] ] where the superscript @xmath56 only means the parameters of @xmath465 are @xmath51 $ ] , and the subscript @xmath466 means the two - dimensional kernel .
( [ 11.7 ] ) has the similar form as ( [ 2.12 ] ) except for its complex form . taking @xmath467 @xmath468 and @xmath469 , @xmath470 , we have@xmath471 this shows that @xmath451 is really the counterpart of the 2-dimensional gft .
if taking the matrix element of @xmath451 in the @xmath140 representation which is conjugate to @xmath9 , we obtain the 2-dimensional gft in its ` frequency domain ' , i.e. , @xmath472 \equiv \mathcal{k}_{2}^{n}\left ( \xi^{\prime},\xi \right ) , \label{11.9}\ ] ] where the superscript @xmath473 means that this transform kernel corresponds to the parameter matrix @xmath474 $ ] .
the two - mode gfo also abides by the group multiplication rule . using the iwop technique and ( [ 11.1 ] )
we obtain@xmath475\nonumber \\ & + \left ( sz_{1}+rz_{2}^{\ast}\right ) a_{1}^{\dagger}+\left ( rz_{1}^{\ast } + sz_{2}\right ) a_{2}^{\dagger}+z_{1}^{\prime \ast}a_{1}+z_{2}^{\prime \ast } a_{2}\nonumber \\ & + z_{1}^{\ast}\left ( s^{\prime}z_{1}^{\prime}+r^{\prime}z_{2}^{\prime \ast } \right ) + z_{2}^{\ast}\left ( r^{\prime}z_{1}^{\prime \ast}+s^{\prime}z_{2}^{\prime}\right ) -a_{1}^{\dagger}a_{1}-a_{2}^{\dagger}a_{2}\ } \colon \nonumber \\ & = \frac{1}{s^{\prime \prime \ast}}\exp \left ( \frac{r^{\prime \prime}}{2s^{\prime \prime \ast}}a_{1}^{\dagger}a_{2}^{\dagger}\right ) \colon \exp \left \ { \left ( \frac{1}{s^{\prime \prime \ast}}-1\right ) \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) \right \ } \colon \exp \left ( -\frac{r^{\prime \prime \ast}}{2s^{\prime \prime \ast}}a_{1}a_{2}\right ) \nonumber \\ & = f_{2}\left ( r^{\prime \prime},s^{\prime \prime}\right ) , \label{11.10}\ ] ] where @xmath476 are given by eq.([7.2 ] ) or ( [ 7.4 ] ) .
therefore , ( [ 11.10 ] ) is a loyal representation of the multiplication rule for ray transfer matrices in the sense of _ matrix optics_.@xmath477 next we extended quantum optical abcd law to two - mode case . operating with @xmath478 on two - mode number state @xmath479 and using the overlap between coherent state and number state , i.e. @xmath480 , \label{11.11}\ ] ] and the integral formula
@xcite@xmath481 we can calculate@xmath482 \nonumber \\ & \times \exp \left [ -sr^{\ast}z_{1}z_{2}-rs^{\ast}z_{1}^{\ast}z_{2}^{\ast } + \left ( sz_{1}+rz_{2}^{\ast}\right ) a_{1}^{\dagger}+\left ( rz_{1}^{\ast } + sz_{2}\right ) a_{2}^{\dagger}\right ] \left \vert 00\right \rangle \nonumber \\ & = \frac{s}{\left \vert s\right \vert ^{2m+2}\sqrt{m!n!}}\int \frac{d^{2}z_{2}}{\pi}z_{2}^{\ast n}\left ( sa_{1}^{\dagger}-sr^{\ast}z_{2}\right ) ^{m}\exp \left ( -\left \vert z_{2}\right \vert ^{2}+\frac{1}{s^{\ast}}a_{2}^{\dagger}z_{2}+\frac{ra_{1}^{\dagger}a_{2}^{\dagger}}{s^{\ast}}\right ) \left
\vert 00\right \rangle \nonumber \\ & = \frac{r^{\ast n}}{s^{\ast n+1}\sqrt{m!n!}}h_{m , n}\left [ \frac { a_{1}^{\dagger}}{s^{\ast}},\frac{a_{2}^{\dagger}}{r^{\ast}}\right ] \exp \left ( \frac{ra_{1}^{\dagger}a_{2}^{\dagger}}{s^{\ast}}\right ) \left \vert 00\right \rangle , \label{11.13}\ ] ] where @xmath483 is the two variables hermite polynomial @xcite , shown in ( [ 4.20 ] ) and ( [ 4.21 ] ) . using eqs.([5.2 ] ) and ( [ 8.9 ] ) , we recast eq.([11.13 ] ) into@xmath484 \exp \left ( -\frac{q_{1}-i}{q_{1}+i}a_{1}^{\dagger}a_{2}^{\dagger } \right ) \left \vert 00\right \rangle . \label{11.14}\ ] ] noticing the multiplication rule of @xmath451 in eq.([11.10 ] ) , which is equivalent to @xmath485 where @xmath486 and @xmath487 are related to each other by eq.([8.5 ] ) . next
we directly use the gfo to derive abcd rule in quantum optics for gaussian beam in two - mode case . according to eq.([11.13 ] ) and eq .
( [ 11.15 ] ) we obtain@xmath488 \exp \left [ \frac { r^{\prime \prime}a_{1}^{\dagger}a_{2}^{\dagger}}{s^{\prime \prime \ast}}\right ] \left \vert 00\right \rangle , \label{11.16}\ ] ] similar to the way of deriving eq .
( [ 11.14 ] ) , we can simplify eq .
( [ 11.16 ] ) as@xmath489 \exp \left [ -\frac{q_{2}-i}{q_{2}+i}a_{1}^{\dagger}a_{2}^{\dagger}\right ]
\left \vert 00\right \rangle , \label{11.17}\ ] ] where the relation between @xmath490 and @xmath300 are determined by eq.([8.11 ] ) which resembles eq.([8.3 ] ) , this is just the new abcd law for two - mode case in quantum optics . in order to obtain the quadratic combinations of canonical operators , let first derive an operator identity .
note @xmath491 @xmath492 and eq.([3.12]),([3.13 ] ) we can prove the operator identity@xmath493 } & = \int \frac{d^{2}\eta}{\pi}e^{\frac{\lambda}{2}\left [ \left ( q_{1}-q_{2}\right ) ^{2}+\left ( p_{1}+p_{2}\right ) ^{2}\right ] } \left \vert \eta \right \rangle \left \langle \eta \right \vert \nonumber \\ & = \frac{1}{1-\lambda}\colon \exp \left [ \frac{2\lambda}{1-\lambda}k_{+}\right ] \colon , \label{11.18}\ ] ] where we have set @xmath494 . \label{k1}\ ] ] when @xmath495 , @xmath496 @xmath497 @xmath498 and using eq.([11.3 ] ) we see that @xmath499 which is corresponding to the square phase operator in single - mode case . in a similar way ,
using ( [ 3.18 ] ) and ( [ 3.19 ] ) we can derive another operator identity @xmath500 } & = \int \frac{d^{2}\xi}{\pi}e^{\frac{\lambda}{2}\left [ \left ( q_{1}+q_{2}\right ) ^{2}+\left ( p_{1}-p_{2}\right ) ^{2}\right ] } \left \vert \xi \right \rangle \left \langle \xi \right \vert \nonumber \\ & = \frac{1}{1-\lambda}\colon \exp \left [ \frac{2\lambda}{1-\lambda}k_{-}\right ] \colon , \label{11.20}\ ] ] where @xmath501 .
\label{k2}\ ] ] it then follows from eqs.([11.3 ] ) and ( [ 11.20 ] ) @xmath502 which is corresponding to fresnel propagator in free space ( single - mode case ) .
in particular , when @xmath503 and @xmath504 eq . (
[ 11.3 ] ) becomes @xmath505 \colon , \label{11.22}\ ] ] where@xmath506 @xmath507 eq .
( [ 11.22 ] ) is just the two - mode squeezing operator , @xmath508 \equiv \exp \left [ -2k_{0}\ln a\right ] , \label{11.22b}\\ k_{0 } & \equiv-\frac{\mathtt{i}}{2}\left ( q_{1}p_{2}+q_{2}p_{1}\right ) , \label{k0}\ ] ] which actually squeezes the entangled state @xmath142 ( its conjugate state is @xmath6),@xmath509 using the decomposition ( [ 9.5 ] ) of the matrix and combining equations ( [ 11.19 ] ) , ( [ 11.21 ] ) and ( [ 11.22b ] ) together , we see that @xmath510 this is the two - mode quadratic canonical operator representation of @xmath511 . to prove eq.([11.24 ] ) , using ( [ 11.23 ] ) and ( [ 3.20 ] ) we see@xmath512 .
\label{11.25}\ ] ] it then follows@xmath513 \nonumber \\ & \equiv \mathcal{k}_{2}^{m}\left ( \eta^{\prime},\eta \right ) , \label{11.26}\ ] ] which is just the transform kernel of a 2-dimensional gft and the definition given in ( [ 11.24 ] ) is true .
note that the quadratic combinations in eqs.([k1 ] ) , ( [ k2 ] ) and ( [ k0 ] ) of the four canonical operators @xmath514 obey the commutative relations @xmath515
= 2k_{0},$ ] @xmath516 = \pm k_{\pm},$ ] so @xmath517 involves a su(2 ) lie algebra structure ( this structure is also compiled by @xmath518 , @xmath519 and @xmath520 that have been used in constructing @xmath266 ) .
when @xmath351 from eq.([11.3 ] ) we see @xmath521 \nonumber \\ & = \exp \left [ -\mathtt{i}\frac{\pi}{2}\left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}+1\right ) \right ] \equiv \mathcal{f } , \label{11.27}\ ] ] which can also be named the fourier operator , since it induces the quantum mechanically transforms @xcite@xmath522 it then follows that @xmath523 on the other hand , in order to obtain the decomposition of @xmath517 for @xmath360 similar to deriving eq.([9.14 ] ) , we have @xmath524 \exp \left [ 2k_{0}\ln d\right ] \exp \left [ \frac{ic}{d}k_{+}\right ] , \text { for } d\neq0 .
\label{11.30}\ ] ] while for @xmath525 or @xmath358 , using eqs.([9.15 ] ) and ( [ 9.16 ] ) we have another decomposition of @xmath511 , i.e.,@xmath526 \exp \left [ -2k_{0}\ln b\right ] \mathcal{f}\exp \left [ \frac{ia}{b}k_{+}\right ] , \text { } b\neq0 , \label{11.31}\ ] ] and @xmath527 \exp \left [ -2k_{0}\ln \frac{-1}{c}\right ] \mathcal{f}\exp \left [ -\frac{id}{c}k_{-}\right ] , c\neq0 . \label{11.32}\ ] ] in addition , noticing eqs.([9.26 ] ) and ( [ 9.24 ] ) , we can rewrite eqs.([11.31 ] ) and ( [ 11.32 ] ) as follows @xmath528 \exp \left [ -ibk_{-}\right ] \exp \left [ \frac{i}{b}\left ( a-1\right ) k_{+}\right ] , b\neq0 , \label{11.33}\ ] ] and @xmath529 \exp \left [ ick_{+}\right ] \exp \left [ \frac{-i}{c}\left ( d-1\right ) k_{-}\right ] , c\neq0 , \label{11.34}\ ] ] respectively .
next , according to some optical systems used frequently in physical optics , we derive some new entangled optical operator identities . for a special optical system with the parameter @xmath360 @xmath361 ( [ 9.20 ] ) which corresponds to the fourier transform system , we have@xmath530 \exp \left [ 2k_{0}\ln d\right ] \exp \left [ -\frac{i}{bd}k_{+}\right ] = \exp \left [ \frac{id}{b}k_{+}\right ] \exp \left [ -2k_{0}\ln b\right ] \mathcal{f}. \label{11.35}\ ] ] in particular , when @xmath363 @xmath361 eq.([9.22 ] ) corresponding to the ideal spectrum analyzer , we have @xmath531 \exp \left [ -\frac{i}{b}k_{+}\right ] \exp \left [ -ibk_{-}\right ] = \exp \left [ -2k_{0}\ln b\right ] \mathcal{f}. \label{11.36}\ ] ] when @xmath532 @xmath504@xmath287{cc}a & 0\\ c & a^{-1}\end{array } \right ) = \left ( \begin{array } [ c]{cc}1 & 0\\ c / a & 1 \end{array } \right ) \left ( \begin{array } [ c]{cc}a & 0\\ 0 & a^{-1}\end{array } \right ) , \ ] ] which corresponds to the form of image system , another operator identity is given by@xmath533 \exp \left [ iack_{+}\right ] = \exp \left [ \frac{ic}{a}k_{+}\right ] \exp \left [ -2k_{0}\ln a\right ] .
\label{11.37}\ ] ] when @xmath534 @xmath535@xmath287{cc}d^{-1 } & b\\ 0 & d \end{array } \right ) = \left ( \begin{array } [ c]{cc}1 & b / d\\ 0 & 1 \end{array } \right ) \left ( \begin{array } [ c]{cc}d^{-1 } & 0\\ 0 & d \end{array } \right ) , \ ] ] which corresponds to the far foci system , @xmath536 \exp \left [ -2k_{0}\ln b\right ] \mathcal{f}\exp \left [ \frac{i}{bd}k_{+}\right ] = \exp \left [ -\frac{ib}{d}k_{-}\right ] \exp \left [ 2k_{0}\ln d\right ] . \label{11.38}\ ] ] when @xmath537 @xmath361 corresponding to the fresnel transform system , @xmath287{cc}a & b\\ -b^{-1 } & 0 \end{array } \right ) = \left ( \begin{array } [ c]{cc}b & 0\\ 0 & b^{-1}\end{array } \right ) \left ( \begin{array } [ c]{cc}0 & 1\\ -1 & 0 \end{array } \right ) \left ( \begin{array } [ c]{cc}1 & 0\\ a / b & 1 \end{array } \right ) , \ ] ] we have @xmath538 \exp \left [ -2k_{0}\ln a\right ] \exp \left [ -\frac{ib}{a}k_{-}\right ] = \exp \left [ -2k_{0}\ln b\right ] \mathcal{f}\exp \left [ \frac{ia}{b}k_{+}\right ]
. \label{11.39}\ ] ] the gfo can unify those optical operators in two - mode case .
various decompositions of the gfo into the exponential canonical operators , corresponding to the decomposition of ray transfer matrix @xmath539 , $ ] are also derived . in our derivation ,
the entangled state representation is of useness in our research .
in section 8 we have found that under the fresnel transformation the pure position density @xmath540 becomes the tomographic density @xmath379 , which is just the radon transform of the wigner operator @xmath541 in this section we want to generalize the above conclusion to two - mode entangled case .
here we shall prove @xmath542 i.e. , we show that @xmath543 is just the radon transform of the entangled wigner operator @xmath544 similar in spirit to the single - mode case , operating @xmath545 on entangled state representation @xmath6 we see @xmath546 \left \vert z_{1},z_{2}\right \rangle \left \langle z_{1},z_{2}\right \vert \left .
\eta \right \rangle \nonumber \\ & = \frac{1}{s^{\ast}}\int \frac{d^{2}z_{1}d^{2}z_{2}}{\pi^{2}}\exp \left [ -\left \vert z_{1}\right \vert ^{2}+\frac{1}{s^{\ast}}\left ( a_{1}^{\dagger } -r^{\ast}z_{2}\right ) z_{1}+\left ( \eta+z_{2}^{\ast}\right ) z_{1}^{\ast } \right ] \nonumber \\ & \times \exp \left [ -\frac{1}{2}\left \vert
\eta \right \vert ^{2}-\left \vert z_{2}\right \vert ^{2}+\frac{1}{s^{\ast}}z_{2}a_{2}^{\dagger}-\eta^{\ast}z_{2}^{\ast}+\frac{r}{s^{\ast}}a_{1}^{\dagger}a_{2}^{\dagger}\right ] \left \vert 00\right \rangle \nonumber \\ & = \frac{1}{s^{\ast}}\int \frac{d^{2}z_{2}}{\pi}\exp \left [ -\frac{s^{\ast } + r^{\ast}}{s^{\ast}}\left \vert z_{2}\right \vert ^{2}+\frac{1}{s^{\ast}}\left ( a_{2}^{\dagger}-\eta r^{\ast}\right ) z_{2}+\frac{1}{s^{\ast}}\left ( a_{1}^{\dagger}-s^{\ast}\eta^{\ast}\right ) z_{2}^{\ast}\right ] \nonumber \\ & \times \exp \left [ + \frac{\eta}{s^{\ast}}a_{1}^{\dagger}+\frac{r}{s^{\ast}}a_{1}^{\dagger}a_{2}^{\dagger}-\frac{1}{2}\left \vert \eta \right \vert ^{2}\right ] \left \vert 00\right \rangle \nonumber \\ & = \frac{1}{s^{\ast}+r^{\ast}}\exp \left \ { -\allowbreak \frac{s^{\ast}-r^{\ast}}{2\left ( s^{\ast}+r^{\ast}\right ) } \left \vert \eta \right \vert ^{2}+\allowbreak \frac{\eta a_{1}^{\dagger}}{s^{\ast}+r^{\ast}}\allowbreak -\allowbreak \frac{\eta^{\ast}a_{2}^{\dagger}}{s^{\ast}+r^{\ast}}+\frac { s+r}{s^{\ast}+r^{\ast}}\allowbreak a_{1}^{\dagger}a_{2}^{\dagger}\right \ }
\left \vert 00\right \rangle \equiv \left \vert \eta \right \rangle _ { s , r } , \label{22}\ ] ] or @xmath547 where we have used the integration formula@xmath548 noticing the completeness relation and the orthogonality of @xmath9 we immediately derive @xmath549 a generalized entangled state representation @xmath550 with the completeness relation ( [ 25 ] ) . from ( [ 23 ] ) we can see that@xmath551 so we have the eigen - equations for @xmath552 as follows@xmath553 \left \vert \eta \right \rangle _ { s , r } & = \sqrt{2}\eta_{1}\left \vert \eta \right \rangle _ { s , r},\text { } \label{28}\\ \left [ b\left ( q_{1}+q_{2}\right ) + d\left ( p_{1}+p_{2}\right ) \right ] \left
\vert \eta \right \rangle _ { s , r } & = \sqrt{2}\eta_{2}\left \vert \eta \right \rangle _ { s , r } , \label{29}\ ] ] we can also check eqs.([26])-([29 ] ) by another way .
for two - mode correlated system , we have introduced the wigner operator in ( [ 3.27 ] ) . according to the wely correspondence rule @xcite@xmath555 where @xmath556 is the weyl correspondence of @xmath557 and @xmath558 , \label{32}\ ] ] the classical weyl correspondence of the projection operator @xmath559 can be calculated as @xmath560 \nonumber \\ & = 4\pi^{2}\int \frac{d^{2}\eta^{\prime}}{\pi^{3}}\left . _ { r , s}\left \langle \eta \right \vert \left . \sigma-\eta^{\prime}\right \rangle \left \langle \sigma+\eta^{\prime}\right \vert \left . \eta \right \rangle _ { r , s}\right .
\exp(\eta^{\prime}\gamma^{\ast}-\eta^{\prime \ast}\gamma)\nonumber \\ & = 4\pi^{2}\int \frac{d^{2}\eta^{\prime}}{\pi^{3}}\left \langle \eta \right \vert f_{2}^{\dagger}\left \vert \sigma-\eta^{\prime}\right \rangle \left \langle \sigma+\eta^{\prime}\right \vert f_{2}\left \vert \eta \right \rangle \exp ( \eta^{\prime}\gamma^{\ast}-\eta^{\prime \ast}\gamma ) .
\label{33}\ ] ] then using eq.([11.7 ] ) we have@xmath561 = \pi \delta \left ( \eta_{2}-d\sigma_{2}+b\gamma_{1}\right ) \delta \left ( \eta_{1}-d\sigma _ { 1}-b\gamma_{2}\right ) , \label{34}\ ] ] which means the following weyl correspondence @xmath562 so the projector operator @xmath543 is just the radon transformation of @xmath150 , @xmath172 and @xmath563 are the radon transformation parameter . combining eqs .
( [ 22])-([35 ] ) together we complete the proof ( [ 9 ] ) .
therefore , the quantum tomography in two - mode entangled case is expressed as @xmath564 where @xmath565 is the wigner function .
so the probability distribution for the fresnel quadrature phase is the tomography ( radon transform of the two - mode wigner function ) .
this new relation between quantum tomography and optical fresnel transform may provide experimentalists to figure out new approach for generating tomography .
next we turn to the frequency domain , that is to say , we shall prove that the @xmath433 related radon transform of entangled wigner operator @xmath566 is just the pure state density operator @xmath567 i.e. , @xmath568 where @xmath142 is the conjugated entangled state to @xmath6 . by analogy with the above procedure
, we obtain the 2-dimensional fresnel transformation in its ` frequency domain ' , i.e. , @xmath569 , \label{39}\ ] ] where the superscript @xmath473 means that this transform kernel corresponds to the parameter matrix @xmath474 $ ] . thus the 2d fresnel transformation in its ` frequency domain '
is given by@xmath570 operating @xmath451 on @xmath142 we also have@xmath571 or @xmath572 noticing that the entangled wigner operator in @xmath573 representation is expressed as @xmath574 and using the classical correspondence of @xmath575 which is calculated by @xmath576 \nonumber \\ \ & = 4\int \frac{d^{2}\xi}{\pi}\left \langle \gamma-\xi \right \vert f_{2}\left \vert \xi \right \rangle \left \langle \xi \right \vert f_{2}^{\dag } |\gamma+\xi \rangle \exp(\xi^{\ast}\sigma-\sigma^{\ast}\xi)\nonumber \\ & = \pi \delta \left ( \xi_{1}-a\sigma_{1}-c\gamma_{2}\right ) \delta \left ( \xi_{2}-a\sigma_{2}+c\gamma_{1}\right ) , \label{44}\ ] ] we obtain @xmath577 so the projector operator @xmath578 is another radon transformation of the two - mode wigner operator , with @xmath579 and @xmath580 being the radon transformation parameter ( ` frequency ' domain ) .
therefore , the quantum tomography in @xmath581 representation is expressed as the radon transformation of the wigner function@xmath582 and @xmath583 now we consider the inverse radon transformation .
for instance , using eq.([35 ] ) we see the fourier transformation of @xmath584 is@xmath585 , \label{47}\ ] ] the right - hand side of ( [ 47 ] ) can be regarded as a special fourier transformation of @xmath566 , so by making its inverse fourier transformation , we get@xmath586 where @xmath587 and @xmath588 \nonumber \\ & \times \exp \left [ -ir_{2}\left ( \frac{\eta_{2}}{\sqrt{b^{2}+d^{2}}}-\sigma_{2}\cos \theta_{2}+\gamma_{1}\sin \theta_{2}\right ) \right ] .
\label{49}\ ] ] eq.([48 ] ) is just the inverse radon transformation of entangled wigner operator in the entangled state representation .
this is different from the two independent radon transformations direct product of the two independent single - mode wigner operators , because in ( [ 23 ] ) the @xmath589 is an entangled state .
therefore the wigner function of quantum state @xmath413 can be reconstructed from the tomographic inversion of a set of measured probability distributions @xmath590 , i.e.,@xmath591 thus , based on the previous section , we have further extended the relation connecting optical fresnel transformation with quantum tomography to the entangled case .
the tomography representation @xmath592 is set up , based on which the tomogram of quantum state @xmath413 is just the squared modulus of the wave function @xmath593 i.e. the probability distribution for the fresnel quadrature phase is the tomogram ( radon transform of the wigner function ) .
the fractional fourier transform ( frft ) has been shown to be a very useful tool in fourier optics and information optics .
the concept of frft was firstly introduced mathematically in 1980 by namias @xcite as a mathematical tool for solving theoretical physical problems @xcite , but did not brought enough attention until mendlovic and ozaktas @xcite defined the @xmath40-th frft physically , based on propagation in quadratic graded - index media ( grin media with medium parameters @xmath594 ) . since
then a lot of works have been done on its properties , optical implementations and applications @xcite .
the frft of @xmath370-order is defined in a manner , i.e. , @xmath595 = \sqrt{\frac { e^{i\left ( \frac{\pi}{2}-\theta \right ) } } { 2\pi \sin \theta}}\int_{-\infty } ^{\infty}\exp \left \ { -i\frac{x^{2}+y^{2}}{2\tan \theta}+\frac{ixx^{\prime}}{\sin \theta}\right \ } f\left ( x\right ) dx , \label{12.1}\ ] ] where the exponential function is an integral kernel . in order to find the quantum correspondence of frft , multiplying the function @xmath596 by the ket @xmath597 and bra @xmath598 from left and right , respectively , where @xmath599 and @xmath600 are coordinate eigenvectors , @xmath601 , and then using ( [ 3.7 ] ) and the iwop technique to perform the integration , we obtain@xmath602 where we have used the operator identity in the last step of eq.([12.2 ] ) @xmath603 from the orthogonal relation @xmath604 we know that eq.([12.2 ] ) indicates @xmath605 which implies that the integral kernel in eq.([12.1 ] ) is just the matrix element of operator @xmath606 in coordinate state ( @xmath606 called as fractional fourier operator @xcite .
therefore , if we consider @xmath607 as @xmath608 , the wave function of quantum state @xmath609 in the coordinate representation , from eqs .
( [ 12.1 ] ) and ( [ 12.4 ] ) it then follows @xmath595 = \int_{-\infty } ^{\infty}dx\left \langle y\right \vert e^{i\theta a^{\dagger}a}\left \vert x\right \rangle f\left ( x\right ) = \left \langle y\right \vert e^{i\theta a^{\dagger}a}\left \vert f\right \rangle \equiv g\left ( y\right ) , \label{12.5}\ ] ] which suggests @xmath610 from eqs.([12.5 ] ) and ( [ 12.1 ] ) one can see that the frft in eq.([12.1 ] ) corresponds actually to the rotating operator @xmath611 transform in eq.([12.5 ] ) between two quantum states , which is just the quantum version of frft .
in fact , using quantum version of frft , one can directly derive various properties of the frfts .
an important feature of the frft is that they are composed according to @xmath612 ( the additivity property ) . without losing generality ,
we examine@xmath613 \equiv \int_{-\infty}^{\infty}\frac{d^{2}\eta}{\pi}\left \langle y\right \vert e^{i\left ( \theta+\theta^{\prime}\right ) a^{\dagger}a}\left \vert x\right \rangle f\left ( x\right ) .
\label{12.7}\ ] ] according to the completeness relation of coordinate eigenvector , @xmath614 eq.([12.7 ] ) yields@xmath615 & = \int_{-\infty}^{\infty}dx\left \langle y\right \vert e^{i\theta a^{\dagger}a}e^{i\theta^{\prime}a^{\dagger}a}\left \vert x\right \rangle f\left ( x\right ) \nonumber \\ & = \int_{-\infty}^{\infty}dx^{\prime}\left \langle y\right \vert e^{i\theta a^{\dagger}a}\left \vert x^{\prime}\right \rangle \int_{-\infty}^{\infty } dx\left \langle x^{\prime}\right \vert e^{i\theta^{\prime}a^{\dagger}a}\left \vert x\right \rangle f\left ( x\right ) \nonumber \\ & = \int_{-\infty}^{\infty}dx^{\prime}\left \langle y\right \vert e^{i\theta a^{\dagger}a}\left \vert x^{\prime}\right \rangle \mathcal{f}_{\theta}\left [ f\left ( x\right ) \right ] = \mathcal{f}_{\theta}\mathcal{f}_{\theta^{\prime}}\left [ f\left ( x\right ) \right ] , \label{12.8}\ ] ] which is just the additivity of frft . in particular , when @xmath609 is the number state , @xmath616 ,
its wavefunction in coordinate representation is @xmath617 the frft of @xmath618 is @xmath619 = \left \langle y\right \vert e^{i\theta a^{\dagger}a}\left \vert n\right \rangle = e^{in\theta}\left \langle y\right .
\left \vert n\right \rangle , \label{12.10}\ ] ] or @xmath620 = e^{in\theta } e^{-y^{2}/2}h_{n}(y ) , \label{12.11}\ ] ] which indicates that the eigenfunction is hermite - gaussian function with the corresponding eigenvalue being @xmath621 . in studying various optical transformations
the optical operator method is proposed @xcite as mapping of ray - transfer abcd matrix , such that the ray transfer through optical instruments and the diffraction can be discussed by virtue of the commutative relations of operators and the matrix algebra .
the square phase operators , scaling operator , fourier transform operator and the propagation operator in free space have been proposed in the literature , two important questions thus naturally arise : 1 . what is the scaled frft ( sfrft ) operator which corresponds to the sfrft s integration kernel @xcite@xmath622 where @xmath623 is standard focal length ( or a scaled parameter ) ; 2 .
if this operator is found , can it be further decomposed into simpler operators and what are their physical meaning ?
since sfrft has wide application in optical information detection and can be implemented even by using a thick lens @xcite , so our questions are worth of paying attention @xcite .
[ fig2 ] let start with a thick lens ( shown in fig.1 ) which represents a transfer matrix @xcite @xmath287{cc}\mathcal{a } & \mathcal{b}\\ \mathcal{c } & \mathcal{d}\end{array } \right ) = \left ( \begin{array } [ c]{cc}1-\frac{\left ( 1 - 1/n\right ) l}{r_{1 } } & \frac{l}{n}\\ -[\left ( n-1\right ) \frac{r_{1}+r_{2}}{r_{1}r_{2}}-\frac{l\left ( n-1\right ) ^{2}}{nr_{1}r_{2 } } ] & 1-\frac{\left ( 1 - 1/n\right ) l}{r_{2}}\end{array } \right ) , \label{12.13}\ ] ] where @xmath292 is the reflective index ; @xmath624 is the thickness of thick lens ; @xmath625 and @xmath626 denotes the curvature radius of the two surfaces of the lens , respectively .
when we choose @xmath627 then eq.([12.13 ] ) reduces to@xmath287{cc}\mathcal{a } & \mathcal{b}\\ \mathcal{c } & \mathcal{d}\end{array } \right ) = \left ( \begin{array } [ c]{cc}1-\frac{\left ( 1 - 1/n\right ) l}{r } & \frac{l}{n}\\ -[\left ( n-1\right ) \frac{2}{r}-\frac{l\left ( n-1\right ) ^{2}}{nr^{2 } } ] & 1-\frac{\left ( 1 - 1/n\right ) l}{r}\end{array } \right ) .
\label{12.14}\ ] ] by defining @xmath628 @xmath629 and @xmath630 @xmath631 we can recast ( [ 12.14 ] ) into the simple form@xmath287{cc}\mathcal{a } & \mathcal{b}\\ \mathcal{c } & \mathcal{d}\end{array } \right ) = \left ( \begin{array } [ c]{cc}\cos \phi & f_{e}\sin \phi \\
-\sin \phi / f_{e } & \cos \phi \end{array } \right ) , \text { } \det \left ( \begin{array } [ c]{cc}\mathcal{a } & \mathcal{b}\\ \mathcal{c } & \mathcal{d}\end{array } \right ) = 1 .
\label{12.15}\ ] ] according to ( [ 8.4 ] ) we immediately know that the operator of sfrft is@xmath632 .\nonumber\end{aligned}\ ] ] noting that the matrix @xmath243{cc}\mathcal{a } & \mathcal{b}\\ \mathcal{c } & \mathcal{d}\end{array } \right ) $ ] can be decomposed into@xmath287{cc}\mathcal{a } & \mathcal{b}\\ \mathcal{c } & \mathcal{d}\end{array } \right ) = \left ( \begin{array } [ c]{cc}1 & 0\\ -\frac{1}{f_{e}}\tan \phi & 1 \end{array } \right ) \left ( \begin{array } [ c]{cc}\cos \phi & 0\\ 0 & \cos \phi \end{array } \right ) \left ( \begin{array } [ c]{cc}1 & f_{e}\tan \phi \\ 0 & 1 \end{array } \right ) , \label{12.17}\ ] ] according to the previous section we have@xmath633 where @xmath634 @xmath635 and @xmath636 @xmath637 and @xmath638 are the quadrature phase operator , the squeezing operator and the free propagation operator , respectively . on the other hand , from @xmath243{cc}\mathcal{a } & \mathcal{b}\\ \mathcal{c } & \mathcal{d}\end{array } \right ) ^{-1}=\left ( \begin{array } [ c]{cc}\mathcal{d } & \mathcal{-b}\\ \mathcal{-c } & \mathcal{a}\end{array } \right ) , $ ] and ( [ 12.18 ] ) we see@xmath639 it then follows @xmath640 using the canonical operator form ( [ 12.18 ] ) or ( [ 12.20 ] ) of @xmath641 we can deduce its matrix element in the coordinate states @xmath600 ( its conjugate state is @xmath110 ) @xmath642 which is just the kernel of sfrft , thus we name @xmath641 sfrft operator .
noticing that the @xmath643 @xmath519 and @xmath644 construct a close su(2 ) lie algebra , we can put eq.([12.20 ] ) into a more compact form , i.e.,@xmath645 eqs .
( [ 12.18 ] ) , ( [ 12.20 ] ) and ( [ 12.22 ] ) are different forms of the same operator of sfrft .
especially , when @xmath646 @xmath647 which is the usual frft operator .
using ( [ 12.21 ] ) the sfrft of @xmath648 denoted as @xmath649 , $ ] can be expressed as an matrix element in quantum optics context , @xmath650 = \int dx\left \langle x^{\prime}\right \vert f_{1}\left ( \mathcal{a},\mathcal{b},\mathcal{c}\right ) \left \vert x\right \rangle \left \langle x\right \vert \left .
f\right \rangle = \left \langle x^{\prime}\right \vert f_{1}\left ( \mathcal{a},\mathcal{b},\mathcal{c}\right ) \left \vert f\right \rangle . \label{12.23}\ ] ] the above discussions are useful since any unimodular matrix @xmath243{cc}a & b\\ c & d \end{array } \right ) $ ] can be decomposed into @xcite@xmath287{cc}a & b\\ c & d \end{array }
\right ) = \left ( \begin{array } [ c]{cc}1 & 0\\ -\mathcal{p } & 1 \end{array } \right ) \left ( \begin{array } [ c]{cc}m & 0\\ 0 & m^{-1}\end{array } \right ) \left ( \begin{array } [ c]{cc}\cos \phi & f_{e}\sin \phi \\
-\sin \phi / f_{e } & \cos \phi \end{array } \right ) , \label{12.24}\ ] ] where the parameters @xmath651 are all real , @xmath652 correspondingly , the operator of frft is given by@xmath653 where @xmath654 $ ] is the quadratic phase operator .
thus the general fresnel transform can always be expressed by sfrft as follows @xmath655 \left ( \frac{x^{\prime}}{m}\right ) .
\label{12.27}\ ] ] i.e. , the output @xmath656 is the sfrft of the input @xmath657 plus a quadratic phase term @xmath658 in the history of developing optics we have known that each optical setup corresponds to an optical transformation , for example , thick lens as a fractional fourier transformer . in turn , once a new integration transform is found , its experimental implementation is expected . in this subsection
we report a new integration transformation which can convert chirplet function to frft kernel @xcite , as this new transformation is invertible and obeys parseval theorem , we expect it be realized by experimentalists .
the new transform we propose here is @xmath659 which differs from the usual two - fold fourier transformation @xmath660 in particular , when @xmath661 eq .
( [ 12.28 ] ) reduces to @xmath662 so @xmath663 can be considered a basis funtion in @xmath664 phase space , or eq .
( [ 12.28 ] ) can be looked as an expansion of @xmath665 with the expansion coefficient being @xmath666 we can prove that the reciprocal transformation of ( [ 12.28 ] ) is@xmath667 in fact , substituting ( [ 12.28 ] ) into the left - hand side of ( [ 12.30 ] ) yields@xmath668 } \nonumber \\ & = \iint_{-\infty}^{\infty}dp^{\prime}dq^{\prime}h(p^{\prime},q^{\prime } ) e^{2i\left ( p^{\prime}q^{\prime}-pq\right ) } \nonumber \\ & \times \delta \left ( p - p^{\prime}\right ) \delta \left ( q - q^{\prime}\right ) \left .
= h(p , q)\right . .
\label{12.31}\ ] ] this transformation s parseval - like theorem is @xmath669 } \nonumber \\ & = \iint \frac{dxdy}{\pi}|f\left ( x , y\right ) |^{2}\iint dx^{\prime } dy^{\prime}e^{2i\left ( x^{\prime}y^{\prime}-xy\right ) } \nonumber \\ & \times \delta \left ( x - x^{\prime}\right ) \delta \left ( p - p^{\prime}\right ) \nonumber \\ & = \iint \frac{dxdy}{\pi}|f\left ( x , y\right ) |^{2}. \label{12.32}\ ] ] now we apply eq .
( [ 12.28 ] ) to phase space transformation in quantum optics .
recall that a signal @xmath670 s wigner transform @xcite is @xmath671 using dirac s symbol @xcite to write @xmath672 @xmath109 is the eigenvector of coordinate @xmath338 , the wigner operator emerges from ( [ 12.33]),@xmath673 if @xmath406 is quantized as the operator @xmath674 through the weyl - wigner correspondence @xcite@xmath675 then@xmath676 this in the literature is named the weyl transform , @xmath406 is the weyl classical correspondence of the operator @xmath677 . substituting ( [ 12.36 ] ) into ( [ 12.28 ] )
we have @xmath678 using @xmath679 and @xmath680 where @xmath681 is the momentum eigenvector , and@xmath682 then eq . ( [ 12.37 ] ) becomes@xmath683 thus through the new integration transformation a new relationship between a phase space function @xmath443 and its weyl - wigner correspondence operator @xmath684 is revealed .
the inverse of ( [ 12.39 ] ) , according to ( [ 12.30 ] ) , is@xmath685 for example , when @xmath686 its classical correspondence is@xmath687 substituting ( [ 12.41 ] ) into ( [ 12.39 ] ) we have@xmath688 using the gaussian integration formula @xmath689 in particular , when @xmath690 with @xmath691 @xmath692 eq .
( [ 12.43 ] ) becomes@xmath693@xmath694 where @xmath695 represents an infinite long chirplet function . comparing ( [ 12.44 ] ) with ( [ 12.42 ] )
we see @xmath696 @xmath697 it then follows@xmath698 where the right - hand side of ( [ 12.45 ] ) is just the frft kernel .
therefore the new integration transformation ( [ 12.28 ] ) can convert spherical wave to frft kernel .
we expect this transformation could be implemented by experimentalists .
moreover , this transformation can also serve for solving some operator ordering problems .
we notice @xmath699\nonumber \\ & = \int_{-\infty}^{\infty}\frac{dv}{2\pi}\delta \left ( q - y-\frac{v}{2}\right ) \exp \left \ { i\left ( p - x\right ) v\right \ } , \label{12.46}\ ] ] so the transformation ( [ 12.28 ] ) is equivalent to@xmath700 for example , using ( [ 12.34 ] ) and ( [ 12.46 ] ) we have @xmath701 so@xmath702 thus this new transformation can convert the wigner operator to @xmath703 similarly , we have @xmath704 then for the wigner function of a density operator @xmath22 , @xmath705 , $ ] we have@xmath706 e^{2\mathtt{i}\left ( p - p^{\prime}\right ) \left ( q - q^{\prime}\right ) } \nonumber \\ & = \mathtt{tr}\left [ \rho \delta \left ( q - q\right ) \delta \left ( p - p\right ) \right ] \nonumber \\ & = \int \frac{dudv}{4\pi^{2}}\mathtt{tr}\left [ \rho e^{i\left ( q - q\right ) u}e^{i\left ( p - p\right ) v}\right ] , \label{12.50}\ ] ] we may define @xmath707 $ ] as the @xmath708 characteristic function .
similarly , @xmath706 e^{-2\mathtt{i}\left ( p - p^{\prime}\right ) \left ( q - q^{\prime}\right ) } \nonumber \\ & = \mathtt{tr}\left [ \rho \delta \left ( p - p\right ) \delta \left ( q - q\right ) \right ] \nonumber \\ & = \int \frac{dudv}{4\pi^{2}}\mathtt{tr}\left [ \rho e^{i\left ( p - p\right ) v}e^{i\left ( q - q\right ) u}\right ] \label{12.51}\ ] ] we name @xmath709 $ ] as the @xmath710 characteristic function .
in this section , we extend 1-d frft to the complex fractional fourier transformation ( cfrft ) . according to ref .
@xcite , based on the entangled state @xmath9 in two - mode fock space and its orthonormal property , we can take the matrix element of @xmath711 $ ] in the entangled state @xmath156 @xmath712 \left \vert \eta \right \rangle , \label{13.1}\ ] ] as the integral transform kernel of cfrft , @xmath713 \left ( \eta \right ) = \int \frac { d^{2}\eta}{\pi}\mathcal{k}^{f}\left ( \eta^{\prime},\eta \right ) f\left ( \eta \right ) .
\label{13.2}\ ] ] using the normally ordered expansion of @xmath714 \colon$ ] and the completeness relation of the coherent state representation , @xmath715 we calculate that @xmath716 is @xmath717 \colon \nonumber \\ & \times \int \frac{d^{2}z_{1}d^{2}z_{2}}{\pi^{2}}\left \vert z_{1},z_{2}\right \rangle \left \langle z_{1},z_{2}\right \vert \left .
\eta \right \rangle \nonumber \\ & = \frac{e^{i(\alpha-\frac{\pi}{2})}}{2\sin \alpha}\exp \left [ \frac { i(\left \vert \eta^{\prime}\right \vert ^{2}+\left \vert \eta \right \vert ^{2})}{2\tan \alpha}-\frac{i\left ( \eta^{\prime \ast}\eta+\eta^{\ast}\eta^{\prime } \right ) } { 2\sin \alpha}\right ] , \label{13.3}\ ] ] which is just the integral kernel of the cfrft in @xcite .
thus we see that the matrix element of @xmath718 $ ] between two entangled state representations @xmath6 and @xmath719 corresponds to cfrft@xmath720 this is a new route from quantum optical transform to classical cfrft transform .
let @xmath721 @xmath722 ( [ 13.2 ] ) becomes@xmath723 \left ( x_{2},y_{2}\right ) & = \frac{e^{i(\alpha-\frac{\pi}{2})}}{2\sin \alpha}\exp \left [ \frac{i\left ( x_{2}^{2}+y_{2}^{2}\right ) } { 2\tan \alpha}\right ] \nonumber \\ & \times \int \frac{dx_{1}dy_{1}}{\pi}\exp \left [ \frac{i\left ( x_{1}^{2}+y_{1}^{2}\right ) } { 2\tan \alpha}-i\frac{\left ( x_{1}x_{2}+y_{1}y_{2}\right ) } { \sin \alpha}\right ] f\left ( x_{1},y_{1}\right ) .
\label{13.4}\ ] ] in fact , letting @xmath724 then using eqs.([3.13 ] ) and ( [ 13.3 ] ) we have @xmath725 \left \vert f\right \rangle \nonumber \\ & = \int \frac{d^{2}\eta}{\pi}\left \langle \eta^{\prime}\right \vert \exp \left [ -i\alpha \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) \right ] \left \vert \eta \right \rangle \left \langle \eta \right \vert \left .
f\right \rangle = \int \frac{d^{2}\eta}{\pi}\mathcal{k}^{f}\left ( \eta^{\prime } , \eta \right ) f\left ( \eta \right ) \nonumber \\ & = \frac{e^{i(\alpha-\frac{\pi}{2})}}{2\sin \alpha}\int \frac{d^{2}\eta}{\pi } \exp \left [ \frac{i(\left \vert \eta^{\prime}\right \vert ^{2}+\left \vert \eta \right \vert ^{2})}{2\tan \alpha}-\frac{i\left ( \eta^{\prime \ast}\eta + \eta^{\ast}\eta^{\prime}\right ) } { 2\sin \alpha}\right ] f\left ( \eta \right ) .
\label{13.5}\ ] ] thus the quantum mechanical version of cfrft is given by@xmath713
\left ( \eta^{\prime}\right ) \equiv
\left \langle \eta^{\prime}\right \vert \exp \left [ -i\alpha \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) \right ] \left \vert f\right \rangle . \label{13.6}\ ] ] the standard complex fourier transform is @xmath726 @xmath727 is the identity operator .
we will show later that this cfrft can help us to reveal some new property which has been overlooked in the formulation of the direct product of two real frfts @xcite .
the definition ( [ 13.16 ] ) is of course required to satisfy the basic postulate that @xmath728 \left ( \eta^{\prime}\right ) = \mathcal{f}_{\alpha+\beta } \left [ f\left ( \eta \right ) \right ] $ ] ( the additivity property ) . for this purpose , using eq.([13.16 ] ) and eq.([3.13 ] ) we see@xmath729 & \equiv \left \langle \eta^{\prime}\right \vert e^{-i(\alpha+\beta)\left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) } \left \vert f\right \rangle \nonumber \\ &
= \int_{-\infty}^{\infty}\frac{d^{2}\eta^{\prime \prime}}{\pi}\left \langle \eta^{\prime}\right \vert e^{-i\alpha \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) } \left \vert \eta^{\prime \prime}\right \rangle \nonumber \\ & \times \int_{-\infty}^{\infty}\frac{d^{2}\eta}{\pi}\left \langle \eta ^{\prime \prime}\right \vert e^{-i\beta \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) } \left \vert \eta \right \rangle f\left ( \eta \right ) \nonumber \\ & = \int_{-\infty}^{\infty}\frac{d^{2}\eta^{\prime \prime}}{\pi}\left \langle \eta^{\prime}\right \vert e^{-i\alpha \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) } \left \vert \eta^{\prime \prime}\right \rangle \mathcal{f}_{\beta}\left [ f\left ( \eta \right ) \right ] \nonumber \\ & = \mathcal{f}_{\alpha}\mathcal{f}_{\beta}\left [ f\left ( \eta \right ) \right ] .
\label{13.7}\ ] ] this derivation is clear and concise by employing the @xmath9 representation and quantum mechanical version of cfrft . on the other hand ,
the formula ( [ 13.6 ] ) can help us to derive cfrft of some wave functions easily .
for example , when @xmath609 is a two - mode number state @xmath730 , then the cfrft of the wave function @xmath731 is @xmath732 & = \left \langle \eta^{\prime}\right \vert e^{i(\alpha+\beta)\left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) } \left \vert m , n\right \rangle \nonumber \\ & = e^{i(\alpha+\beta)\left ( m+n\right ) } \left \langle \eta^{\prime } \right \vert \left .
m , n\right \rangle .
\label{13.8}\ ] ] to calculate @xmath733 , let us recall the definition of two - variable hermite polynomial @xmath734 ( [ 4.20,4.21 ] ) , we can expand @xmath459 as@xmath735 thus@xmath736 as a result of ( [ 13.10 ] ) we see that equation ( [ 13.8 ] ) becomes@xmath737 = e^{i(\alpha+\beta)\left ( m+n\right ) } h_{m , n}\left ( -i\eta^{\prime \ast},i\eta^{\prime}\right ) e^{-\left \vert \eta^{\prime}\right \vert ^{2}/2}. \label{13.11}\ ] ] if we consider the operation @xmath738 as an operator , one can say that the eigenfunction of @xmath738 ( the eigenmodes of cfrft ) is the two - variable hermite polynomials @xmath739 with the eigenvalue being @xmath740 .
this is a new property of cfrft . since the function space spanned by @xmath741 is complete,@xmath742 ^{\ast } = \sqrt{m!n!m^{\prime}!n^{\prime}!}\delta_{m ,
m^{\prime}}\delta_{n , n^{\prime } } , \label{13.12}\ ] ] and @xmath743 ^{\ast}e^{-\left \vert \eta \right \vert ^{2}}=\pi \delta \left ( \eta-\eta ^{\prime}\right ) \delta \left ( \eta^{\ast}-\eta^{\prime \ast}\right ) , \label{13.13}\ ] ] one can confirms that the eigenmodes of cfrft form an orthogonal and complete basis set @xcite .
note that the two variable hermite polynomial @xmath741 is not the direct product of two independent ordinary hermite polynomials , so cfrft differs from the direct product of two frfts . in this subsection , by developing eq .
( [ 12.28 ] ) to more general case which can be further related to the transformation between two mutually conjugate entangled state representations @xmath142 and @xmath9 , we shall propose a new integration transformation in @xmath744 phase space ( see eq .
( [ 13.14 ] ) below ) and its inverse transformation .
we find that eq .
( [ 13.14 ] ) also possesses some well - behaved transformation properties and can be used to obtain the cfrft kernel from a chirplet @xcite . corresponding to the structure of phase space spanned by @xmath140 and @xmath6 and enlightened by eq .
( [ 12.28 ] ) , we propose a new complex integration transformation in @xmath744 phase space @xmath745 when @xmath746 ( [ 13.14 ] ) becomes@xmath747 so @xmath748 can be considered a basis function in @xmath744 phase space , or eq .
( [ 13.14 ] ) can be looked as an expansion of @xmath749 in terms of @xmath750 with the expansion coefficient being @xmath751 we can prove that the inverse transform of ( [ 13.14 ] ) is @xmath752 in fact , substituting ( [ 13.14 ] ) into the left - hand side of ( [ 13.16 ] ) yields@xmath753 this parseval - like theorem for this transformation can also be demonstrated , @xmath754 \nonumber \\ & \times \int \frac{d^{2}\xi d^{2}\eta}{\pi^{2}}\exp \left [ \left ( \mu ^{\prime \ast}-\mu^{\ast}\right ) \eta+\left ( \mu-\mu^{\prime}\right ) \eta^{\ast}\right ] \nonumber \\ & \times \exp \left [ ( \nu^{\ast}-\nu^{\prime \ast})\xi+(\nu^{\prime}-\nu ) \xi^{\ast}\right ] \nonumber \\ & = \int \frac{d^{2}\mu d^{2}\nu}{\pi^{2}}|d\left ( \nu,\mu \right ) |^{2}\iint d^{2}\mu^{\prime}d^{2}\nu^{\prime}\nonumber \\ & \times \exp \left [ \left ( \mu^{\ast}\nu-\nu^{\ast}\mu \right ) + \left ( \mu^{\prime}\nu^{\prime \ast}-\nu^{\prime}\mu^{\prime \ast}\right ) \right ] \nonumber \\ & \times \delta^{\left ( 2\right ) } \left ( \mu-\mu^{\prime}\right ) \delta^{\left ( 2\right ) } \delta(\nu^{\prime}-\nu)\nonumber \\ & = \int \frac{d^{2}\mu d^{2}\nu}{\pi^{2}}|d\left ( \nu,\mu \right ) |^{2}. \label{13.18}\ ] ] in ref .
@xcite for correlated two - body systems , we have successfully established the so - called entangled wigner operator , expressed in the entangled state @xmath132 representation as ( [ 3.27 ] ) , @xmath755 the advantage of introducing @xmath756 can be seen in ref .
the corresponding wigner function for a density matrix @xmath22 is@xmath757 if @xmath758 is quantized as the operator @xmath759 through the weyl - wigner correspondence@xmath760 then using ( [ 13.20 ] ) we see@xmath761 \nonumber \\ & = 4\int \frac{d^{2}\sigma}{\pi}e^{\sigma \xi^{\ast}-\sigma^{\ast}\xi } \left \langle \eta+\sigma \right \vert f\left ( q_{1},q_{2},p_{1},p_{2}\right ) \left \vert \eta-\sigma \right \rangle , \label{13.22}\ ] ] which is named as the complex weyl transform , and @xmath762 is the weyl classical correspondence of @xmath763 . substituting ( [ 13.22 ] ) into ( [ 13.14 ] ) we get@xmath764 using ( [ 3.12 ] ) , we have @xmath765 \right \ } , \label{13.24}\\ \nu & = \nu_{1}+i\nu_{2}.\nonumber\end{aligned}\ ] ] as a result of ( [ 13.24 ] ) and @xmath766 we see @xmath767 \ } \nonumber \\ & = 2\pi \left \langle \xi_{=\mu}\right \vert \exp \ { \frac{i}{\sqrt{2}}\left [ \nu_{1}\left ( p_{1}-p_{2}\right ) -\nu_{2}\left ( q_{1}+q_{2}\right ) \right ] \ } \nonumber \\ & = 2\pi \left \langle \xi_{=\mu}\right \vert e^{i\left ( \mu_{2}\nu_{1}-\mu _ { 1}\nu_{2}\right ) } . \label{13.25}\ ] ] using ( [ 13.25 ] ) , we convert eq .
( [ 13.23 ] ) as@xmath768 the inverse of ( [ 13.26 ] ) , according to ( [ 13.16 ] ) , is@xmath769 thus through the new integration transformation , a new relationship between a phase space function @xmath762 and its weyl - wigner correspondence operator @xmath759 is revealed .
for example , from the following weyl - wigner correspondence@xmath770 \rightarrow \exp \{f[k_{+}+k_{-}-1]\ } , \label{13.28}\ ] ] ( @xmath771 and @xmath772 are defined in eqs.([k1 ] ) and ( [ k2 ] ) ) and ( [ 13.27 ] ) we have@xmath773 \nonumber \\ & = 2\pi \left \langle \xi_{=\mu}\right \vert f\left ( q_{1},q_{2},p_{1},p_{2}\right ) \left \vert \nu \right \rangle e^{i\left ( \nu_{1}\mu_{2}-\nu _ { 2}\mu_{1}\right ) } . \label{13.29}\ ] ] using the gaussian integration formula@xmath774 , \label{13.30}\ ] ] in particular , when @xmath690 with @xmath691 @xmath775eq .
( [ 13.30 ] ) becomes@xmath693@xmath776 \nonumber \\ & = \frac{e^{i\alpha}}{i\sin \alpha}\exp \left [ \frac{i(\left \vert
\mu \right \vert ^{2}+\left \vert \nu \right \vert ^{2})}{2\tan \alpha}-\frac{\mu \nu^{\ast}-\mu^{\ast}\nu}{2\sin \alpha}+i\mu_{2}\nu_{1}-i\mu_{1}\nu_{2}\right ] , \label{13.31}\ ] ] where @xmath777 $ ] represents an infinite long chirplet function . by taking @xmath778 in ( [ 13.29 ] ) , such that @xmath696 and comparing with ( [ 13.31 ] ) we obtain@xmath779 , \label{13.32}\ ] ] where the right - hand side of ( [ 13.32 ] ) is just the cfrft kernel whose properties can be seen in ref .
( one may compare the forms ( [ 13.3 ] ) and ( [ 13.32 ] ) to see their slight difference .
for the relation between them we refer to ref.@xcite ) .
dragoman has shown that the kernel of the cfrft can be classically produced with rotated astigmatic optical systems that mimic the quantum entanglement .
therefore the new integration transformation ( [ 13.14 ] ) can convert spherical wave to cfrft kernel .
we expect this transformation could be implemented by experimentalists . in some practical applications
it is necessary to introduce input and output scale parameters @xcite into frft , i.e. , scaled frft .
the reason lies in that two facts : ( 1 ) the scaled frft may be more useful and convenient for optical information processing due to the scale parameters ( free parameters ) introduced into frft ; ( 2 ) it can be reduced to the conventional frft under a given condition . in this subsection , by establishing the relation between the optical scaled frft and quantum mechanical squeezing - rotating operator transform in one - mode case , we employ the iwop technique and the bipartite entangled state representation of two - mode squeezing operator to extend the scaled frft to more general cases , such as scaled complex frft and entangled scaled frft .
the properties of scaled frfts can be seen more clearly from the viewpoint of representation transform in quantum mechanics .
the scaled frft @xcite of @xmath40-order is defined in a manner such that the usual frft is its special case , i.e. , @xmath780 = \sqrt{\frac { e^{i\left ( \frac{\pi}{2}-\alpha \right ) } } { 2\pi \mu \nu \sin \alpha}}\int_{-\infty}^{\infty}\exp \left \ { -i\frac{x^{2}/\mu^{2}+y^{2}/\nu^{2}}{2\tan \alpha}+\frac{ixx^{\prime}}{\mu \nu \sin \alpha}\right \ } f\left ( x\right ) dx , \label{13.33}\ ] ] where the exponential function is an integral kernel . in a similar way to deriving the quantum correspondence of frft in ( [ 12.4 ] ) , and using the natural repression of single - mode squeezing operator @xmath781 in coordinate representation @xcite , @xmath782 we have @xmath783 which implies that the integral kernel in eq.([13.33 ] ) is just the matrix element of operator @xmath784 in coordinate states . from eq.([13.35 ] )
it then follows that @xmath785 & = \int_{-\infty } ^{\infty}dx\left \langle y\right \vert s_{1}^{\dagger}\left
( \nu \right ) e^{i\alpha a^{\dagger}a}s_{1}\left
( \mu \right ) \left \vert x\right \rangle f\left ( x\right ) \nonumber \\ & = \left \langle y\right \vert s_{1}^{\dagger}\left ( \nu \right ) e^{i\alpha a^{\dagger}a}s_{1}\left ( \mu \right ) \left \vert f\right \rangle \equiv g\left ( y\right ) , \label{13.36}\ ] ] which suggests @xmath786 from eqs.([13.36 ] ) and ( [ 13.33 ] ) one can see that the scaled frft in eq.([13.36 ] ) corresponds actually to the squeezing - rotating operator @xmath787 transform in eq.([13.36 ] ) between two quantum states . on the basis of quantum mechanical version of one - mode
scaled frft , we generalize it to two - mode case , i.e. , we can introduce the integral @xmath788 & \equiv \left \langle \eta^{\prime}\right \vert s_{2}^{\dagger}\left ( \nu \right ) e^{i\alpha \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) } s_{2}\left ( \mu \right ) \left \vert
f\right \rangle \nonumber \\ & = \int_{-\infty}^{\infty}\frac{d^{2}\eta}{\pi}\left \langle \eta^{\prime } \right \vert s_{2}^{\dagger}\left ( \nu \right ) e^{i\alpha \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) } s_{2}\left ( \mu \right ) \left \vert \eta \right \rangle f\left ( \eta \right ) , \label{13.38}\ ] ] where @xmath789 @xmath790 . using the natural expression of the two - mode squeezing operator @xmath135 ( [ 3.16 ] ) , and noticing that @xmath791 is just the integral kernel of cfrft ( [ 13.3 ] ) ,
we can reform ( [ 13.38 ] ) as @xmath788 & = \frac{e^{i\left ( \frac{\pi}{2}-\alpha \right ) } } { 2\mu \nu \sin \alpha}\int \frac{d^{2}\eta}{\pi}f\left ( \eta \right ) \nonumber \\ & \times \exp \left \ { -\frac{i(\left \vert \eta^{\prime}\right \vert ^{2}/\nu^{2}+\allowbreak \left \vert \eta \right \vert ^{2}/\mu^{2})}{2\tan \alpha } + \frac{i\left ( \eta^{\prime}{}^{\ast}\allowbreak \eta+\eta^{\ast}\allowbreak \eta^{\prime}\right ) } { 2\mu \nu \sin \alpha}\right \ } .
\label{13.39}\ ] ] it is obvious that eq.([13.39 ] ) is just a generalized cfrft with squeezing parameters , we name it the scaled cfrft .
thus we link a two - mode squeezing - rotating operator transform to the scaled cfrft of complex functions . on the other hand , recall that the entangled state @xmath792 can be schmidt - decomposed as @xcite @xmath793
we see that @xmath794 using eq.([3.13 ] ) we have @xmath795 where @xmath796 on substituting eqs .
( [ 13.40 ] ) and ( [ 13.41 ] ) into eq.([13.42 ] ) , we can derive@xmath797 \right \ } \nonumber \\ & \times \left \ { \sqrt{\frac{e^{i\left ( \frac{\pi}{2}-\alpha \right ) } } { 2\pi \sin \alpha}}\exp \left [ -i\frac{\kappa_{\nu}^{\prime2}+\kappa_{\mu}^{2}}{2\tan \alpha}+\frac{i\kappa_{\mu}\kappa_{\nu}^{\prime}}{\sin \alpha}\right ] \right \ } , \label{13.43}\ ] ] where @xmath798 @xmath799 from eq.([13.43 ] ) one can see that a new 2-dimensional ( 2d ) scaled frft can be composed of one 1d scaled fft in its space domain and the other in its `` frequency '' domain , while the transform variables being the combination of two coordinates as shown in eq.([13.44 ] ) , so eq.([13.43 ] ) is quite different from the direct product two 1d scaled frfts that are both in space domain are indicated in eq.([13.36 ] ) .
note that the new 2d scaled fft is still characterized by only 3-parameter .
therefore , for any function @xmath800 we can define an entangled scaled frft , i.e. , @xmath801 & = \int_{-\infty}^{\infty}k\left ( x_{1}^{\prime},x_{2}^{\prime},x_{1},x_{2}\right ) f\left ( x_{1},x_{2}\right ) dx_{1}dx_{2}\nonumber \\ & = \left \langle x_{1}^{\prime},x_{2}^{\prime}\right \vert s_{2}^{\dagger } \left ( \nu \right ) e^{i\alpha \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger } a_{2}\right ) } s_{2}\left ( \mu \right ) \left \vert f\right \rangle . \label{13.45}\ ] ] next we examine the properties of these scaled frfts in the quantum optics context . without losing generality , for the additivity property
, we consider the scaled cfrft , @xmath802 \equiv \int \frac{d^{2}\eta}{\pi}\left \langle \eta^{\prime}\right \vert s_{2}^{\dagger}\left ( \nu \right ) e^{i(\alpha+\beta)\left ( a_{1}^{\dagger } a_{1}+a_{2}^{\dagger}a_{2}\right ) } s_{2}\left ( \mu \right ) \left \vert \eta \right \rangle f\left ( \eta \right ) .
\label{13.46}\ ] ] inserting the completeness relation of @xmath6 into eq.([13.46 ] ) yields@xmath803 & = \int \frac{d^{2}\eta}{\pi}\left \langle \eta^{\prime}\right \vert s_{2}^{\dagger}\left ( \nu \right ) e^{i\alpha \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) } s_{2}\left ( \tau \right ) s_{2}^{\dagger } \left ( \tau \right ) e^{i\beta \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger } a_{2}\right ) } s_{2}\left ( \mu \right ) \left \vert \eta \right \rangle f\left ( \eta \right ) \nonumber \\ & = \int \frac{d^{2}\eta^{\prime \prime}}{\pi}\left \langle \eta^{\prime } \right \vert s_{2}^{\dagger}\left ( \nu \right ) e^{i\alpha \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) } s_{2}\left ( \mu_{=\tau } ^{\prime}\right ) \left \vert \eta^{\prime \prime}\right \rangle \nonumber \\ & \times \int \frac{d^{2}\eta}{\pi}\left \langle \eta^{\prime \prime}\right \vert s_{2}^{\dagger}\left ( \nu_{=\tau}^{\prime}\right ) e^{i\beta \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) } s_{2}\left ( \mu \right ) \left \vert \eta \right \rangle f\left ( \eta \right ) \nonumber \\ & = \int \frac{d^{2}\eta^{\prime \prime}}{\pi}\left \langle \eta^{\prime } \right \vert s_{2}^{\dagger}\left ( \nu \right ) e^{i\alpha \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) } s_{2}\left ( \tau \right ) \left \vert \eta_{=\tau}^{\prime \prime}\right \rangle \mathcal{f}_{\beta}\left [ f\left ( \eta \right ) \right ] = \mathcal{f}_{\alpha}\mathcal{f}_{\beta}\left [ f\left ( \eta \right ) \right ] , \label{13.47}\ ] ] which is just the additivity property .
it should be pointed out that the condition of additive operator for the scaled frfts is that the parameter @xmath804 of the prior cascade should be equal to the parameter @xmath805 of the next one , i.e. , @xmath806 for other scaled frfts , the properties can also be discussed in the similar way ( according to their quantum versions ) . to this end , we should emphasize that different scaled frfts correspond to different quantum mechanical squeezing operators or representations .
that is to say , it is possible that some other scaled frft can be presented by using different quantum mechanical squeezing operators or representations .
the connection between the fresnel diffraction in free space and the frft had been bridged by pellat - finet @xcite who found that frfts are adapted to the mathematical expression of fresnel diffraction , just as the standard fourier transform is adapted to fraunhofer diffraction . in previous sections ,
a new formulation of the cfrft and the collins diffraction formula are respectively derived in the context of representation transform of quantum optics . in this section
we inquire if the adaption problem of collins diffraction formula to the cfrft can also be tackled in the context of quantum optics .
we shall treat this topic with the use of two - mode ( 3 parameters ) squeezing operator and in the entangled state representation of continuous variables , in so doing the quantum mechanical version of associated theory of classical diffraction and classical cfrft is obtained , which connects classical optics and quantum optics in this aspect . [ fig1 ] for gaussian beam , the @xmath14 rule is equally derived via optical diffraction integral theory the collins integral formula .
as shown in fig.2 , if @xmath807 represents the input field amplitude at point @xmath808 on @xmath781 , and @xmath809 denotes the diffraction field amplitude at point @xmath810 on @xmath811 then collins formula in complex form takes the form ( [ 11.5,11.7 ] ) .
next we shall examine adaption of the collins formula to the cfrft by virtue of the entangled state representation in quantum optics @xcite . using the completeness relation of @xmath156
we can further put eq.([11.5 ] ) into@xmath812 and taking @xmath813 @xmath814 as well as writing @xmath815 where @xmath166 and @xmath816 are two constants to be determined later , then according to eqs .
( [ 11.5 ] ) and ( [ 14.1 ] ) we have@xmath817 \int \frac{d^{2}\eta}{\pi}\exp \left \ { \frac{i\left \vert \eta \right \vert ^{2}}{2l^{2}}-\frac{i\left ( \sigma^{\ast}\eta+\sigma \eta^{\ast}\right ) } { 2lk\sqrt{ad}}\right \ } f\left ( \eta \right ) .
\label{14.3}\ ] ] comparing eq.([14.3 ] ) with eq.([11.7 ] ) leads us to choose@xmath818 then eq.([14.3 ] ) becomes@xmath819 \nonumber \\ & \times \int \frac{d^{2}\eta}{\pi}\exp \left \ { \frac{i\left ( \left \vert \eta \right \vert ^{2}+\left \vert \sigma \right \vert ^{2}\right ) } { 2\tan \alpha } -\frac{i\left ( \sigma^{\ast}\eta+\sigma \eta^{\ast}\right ) } { 2\sin \alpha } \right \ } f\left ( \eta \right ) \nonumber \\ & = \frac{\cos \alpha}{a}e^{-i\alpha}\exp \left [ i\frac{ad-\cos^{2}\alpha}{\sin2\alpha}\left \vert \sigma \right \vert ^{2}\right ] \mathcal{f}_{\alpha } \left [ f\right ] \left ( \sigma \right ) , \label{14.5}\ ] ] so eq .
( [ 14.5 ] ) is a standard cfrft up to a quadratic phase term @xmath820 $ ] , according to eq.([14.4 ] ) and @xmath821 , it can also be written as@xmath822 = \exp \left [ \frac{i}{r}\left
\vert \eta^{\prime}\right \vert ^{2}\right ] , \label{14.6}\ ] ] which represents a quadratic approximation to a sphere wave diverging from a luminous point at distance@xmath823 from @xmath824 let @xmath825 be the sphere tangent to @xmath135 with radius @xmath826 ( see fig.2 ) .
a point on @xmath825 is located by its projection on @xmath135 , this means that coordinates on @xmath135 can also be used as coordinates on @xmath825 .
therefore , the quadratic phase term can be compensated if the output field is observed on @xmath825 but @xmath135 .
then , after considering the phase compensation , the field transforms from @xmath781 to @xmath825 is @xmath827 \left ( \sigma \right ) , \label{14.8}\ ] ] in this way , the field amplitude on @xmath825 is the perfect @xmath828 fft - c of the field amplitude on @xmath781 .
the most important property of frft is that @xmath738 obeys the additivity rule , i.e. , two successive frft of order @xmath40 and @xmath49 makes up the fft of order @xmath829 . for the cfrft ,
its additivity property is proven in eq.([13.7 ] ) . for collins diffraction from @xmath781 to @xmath830
( see fig.1 ) , the additivity means that the diffraction pattern observed on @xmath830(@xmath831 ) ( the sphere tangent to @xmath832 with radius @xmath833 ) and associated with @xmath834 should be the result of a first diffraction phenomenon ( associated with @xmath835 on @xmath825 ( with @xmath810 ) , followed by a second diffraction phenomenon ( associated with @xmath836 from @xmath825 to @xmath830 .
this is a necessary consequence of the huygens principle .
next we prove that such is indeed the case .
firstly , let us consider the field transform from @xmath781 ( with @xmath808 ) to @xmath830 ( see fig.2 ) described by the ray transfer matrix [ @xmath837 .
similar to deriving eq.([14.7 ] ) , after the squeezing transform and the phase compensation,@xmath838 thus we can obtain the expression of cfrft for collins diffraction from @xmath781 to @xmath830 ( not @xmath832)@xmath10@xmath839 \left ( \sigma^{\prime}\right ) , \label{14.10}\ ] ] where @xmath840and @xmath841 @xmath842 eq.([14.10 ] ) is the same in form as eq.([14.8 ] ) but with primed variables . using eqs.([14.9 ] ) and ( [ 11.5 ] ) one can prove that the transform from @xmath781 to @xmath830 is ( see eqs .
( [ 11.5 ] ) , ( [ 14.5])-([14.7]))@xmath477@xmath843 in eq.([14.12 ] ) we have taken the phase compensation term ( [ 14.9 ] ) into account . secondly , let us consider the second diffraction from @xmath825 to @xmath832 determined by the ray transfer matrix [ @xmath844 . for this purpose , using the group multiplication rule of @xmath451 , we can decompose the diffraction from @xmath781 to @xmath830 into two parts : one is described as the matrix @xmath845 $ ] from plane @xmath781 ( with @xmath808 ) to @xmath846 ( with @xmath810 ) , the other is @xmath847 $ ] from plane @xmath135 to @xmath848 ( with @xmath831 ) , then the total matrix from @xmath781 to @xmath832 is@xmath287{cc}a^{\prime } & b^{\prime}\\ c^{\prime } & d^{\prime}\end{array } \right ) = \left ( \begin{array } [ c]{cc}a^{\prime \prime } & b^{\prime \prime}\\ c^{\prime \prime } & d^{\prime \prime}\end{array } \right ) \left ( \begin{array } [ c]{cc}a & b\\ c & d \end{array } \right ) . \label{14.13}\ ] ] using eq.([3.13 ] ) and the group multiplication rule of @xmath849 , we can further put eq.([14.12 ] ) into another form @xmath850 \left \langle \bar{\eta}\right \vert u_{2}\left ( r^{\prime \prime},s^{\prime \prime}\right ) u_{2}\left ( r , s\right ) \left \vert f\right \rangle \nonumber \\ &
= \exp \left [ -\frac{i}{r^{\prime}}\left \vert \bar{\eta}\right \vert ^{2}\right ] \left \langle \bar{\eta}\right \vert u_{2}\left ( r^{\prime \prime } , s^{\prime \prime}\right ) \mu_{2}^{\prime2}\int \frac{d^{2}\sigma}{\pi } \left \vert \mu_{2}^{\prime}\sigma \right \rangle \left \langle \mu_{2}^{\prime } \sigma \right \vert u_{2}\left ( r , s\right ) \mu_{1}^{\prime2}\int \frac { d^{2}\eta}{\pi}\left \vert \mu_{1}^{\prime}\eta \right \rangle \left \langle \mu_{1}^{\prime}\eta \right . \left \vert f\right \rangle \nonumber \\ & = \mu_{2}^{\prime2}\int \frac{d^{2}\sigma}{\pi}\exp \left ( -\frac{i\left \vert \bar{\eta}\right \vert ^{2}}{r^{\prime}}\right ) \left \langle \bar{\eta } \right \vert u_{2}\left ( r^{\prime \prime},s^{\prime \prime}\right ) \left \vert
\mu_{2}^{\prime}\sigma \right \rangle \left [ \mu_{1}^{\prime2}\int \frac { d^{2}\eta}{\pi}\left \langle \mu_{2}^{\prime}\sigma \right \vert u_{2}\left ( r , s\right ) \left \vert \mu_{1}^{\prime}\eta \right \rangle f\left ( \eta \right ) \right ] \nonumber \\ & = \frac{b}{dk^{2}}\int \frac{d^{2}\sigma}{\pi}\exp \left ( -\frac{i\left \vert \bar{\eta}\right \vert ^{2}}{r^{\prime}}\right ) \left \langle \bar{\eta } \right \vert u_{2}\left ( r^{\prime \prime},s^{\prime \prime}\right ) \left \vert \mu_{2}^{\prime}\sigma \right \rangle g\left ( \sigma \right ) , \label{14.14}\ ] ] where @xmath851 and we have made a reasonable assumption that @xmath852(so @xmath853 ) , which means that there are same scaled variants for the input field amplitudes on @xmath781 of the diffractions from @xmath781 to @xmath825 and from @xmath781 to @xmath830 . in order to examine the second diffraction domain from @xmath825 to @xmath830 ( not @xmath832 ) , we need to translate the output field amplitude @xmath854 observed on plane @xmath135 to the field amplitude observed on sphere plane @xmath855 i.e. , putting @xmath856 into @xmath857 ( see eq.([14.8 ] ) ) by taking the phase compensation ( see eq.([14.3 ] ) ) into account .
thus the field transform from @xmath825 to @xmath830 is ( @xmath858)@xmath859 comparing eq.([14.15 ] ) with eq.([11.7 ] ) leads us to choose@xmath860 and noticing that @xmath852 yields @xmath861 thus we have @xmath862 combining eqs.([14.1 ] ) and ( [ 14.17 ] ) it then follows ( letting @xmath863)@xmath864 and @xmath865 where we have used eqs.([14.7 ] ) , ( [ 14.9 ] ) and ( [ 14.13 ] ) .
substitution of eqs.([14.8 ] ) , ( [ 14.16 ] ) , ( [ 14.18 ] ) and ( [ 14.19 ] ) into eq.([14.15 ] ) yields@xmath866 \left
( \sigma \right ) \nonumber \\ & = \frac{\cos \alpha^{\prime}}{a^{\prime}}e^{-i\left ( \alpha+\beta \right ) } \mathcal{f}_{\beta}\mathcal{f}_{\alpha}\left [ f\right ] \left ( \sigma^{\prime}\right ) .
\label{14.20}\ ] ] the first equation of eq.([14.20 ] ) indicates that it is just a cfrft of @xmath867 from @xmath825 to @xmath868 comparing eq.([14.20 ] ) with eq.([14.10 ] ) , we see @xmath869 \left ( \sigma^{\prime}\right ) = \mathcal{f}_{\alpha+\beta}\left [ f\right ] \left ( \sigma^{\prime}\right ) .
\label{14.21}\ ] ] thus we complete the study of adaption of cfrft to the mathematical representation of collins diffraction formula in quantum optics context .
optical tomographic imaging techniques derive two - dimensional data from a three - dimensional object to obtain a slice image of the internal structure and thus have the ability to peer inside the object noninvasively . the mathematical method which complete this task is the radon transformation .
similarly , one can use the inverse radon transformation to obtain the wigner distribution by tomographic inversion of a set of measured probability distributions of the quadrature amplitude @xcite . based on the radon
transform @xcite and the frft we can introduce the conception of fractional radon transformation ( frt ) which combines both of them in a reasonable way . we notice the well - known fact that the usual radon transform of a function @xmath870 can be proceeded in two successive steps , the first step is an @xmath871dimensional ordinary fourier transform , i.e. performing a usual ft of @xmath870 in @xmath292-dimensional @xmath872 space , @xmath873 where @xmath874 @xmath875 is a unit vector , @xmath328 is a real number .
its inverse is @xmath876 letting @xmath877 and rewriting ( [ 18.1 ] ) as @xmath878 one can see that the integration over @xmath879 has been defined as a radon transform of @xmath870 , denoted as @xmath880 so @xmath881 can be considered as a @xmath882dimensional fourier transform of @xmath883 @xmath884 its inverse transform is@xmath885 this ordinary @xmath882dimensional fourier transform is considered as the second step . combining result of ( [ 18.1 ] ) and ( [ 18.6 ] ) we have@xmath886 i. e. two usual fts
make up a radon transform , the inverse of ( [ 18.7 ] ) is @xmath887 by analogy with these procedures we can make two successively frfts to realize the new fractional radon transformation @xcite .
the @xmath292-dimensional frft of @xmath870 is defined as @xmath888 = \left ( c_{\alpha}\right ) ^{n}\int \exp \left ( \frac{i\left ( \vec{r}^{2}+\vec{k}^{2}\right ) } { 2\tan \alpha}-\frac{i\vec{k}\cdot \vec{r}}{\sin \alpha}\right ) f\left ( \vec { r}\right ) d\vec{r}\equiv f_{\alpha}\left ( t\hat{e}\right ) , \text { \ } \vec{k}=t\hat{e}. \label{18.9}\ ] ] where @xmath40 is named as the order of frft , @xmath889 ^{1/2}.$ ] firstly , we perform an @xmath882dimensional inverse fractional fourier transform for @xmath890 in @xmath328-space,@xmath891 ^{1-n}\mathfrak{f}_{-\alpha , t}\left [ f_{\alpha}\left ( t\hat{e}\right ) \right ] \nonumber \\ & = \left [ c_{\alpha}\right ] ^{1-n}c_{-\alpha}\int_{-\infty}^{\infty}\exp \left ( -\frac{i\left ( \lambda^{2}+t^{2}\right ) } { 2\tan \alpha}+\frac{i\lambda t}{\sin \alpha}\right ) f_{\alpha}\left ( t\hat{e}\right ) dt , \label{18.11}\ ] ] @xmath892 ^{1-n}$ ] was introduced for later s convenience . then substituting ( [ 18.9 ] ) into ( [ 18.11 ] ) we have@xmath891 ^{1-n}c_{-\alpha}\left ( c_{\alpha}\right ) ^{n}\int \exp \left ( -\frac { i\left ( t^{2}+\lambda^{2}\right ) } { 2\tan \alpha}+\frac{i\lambda t}{\sin \alpha}+\frac{i\left ( \vec{r}^{2}+t^{2}\right ) } { 2\tan \alpha}-\frac{it\hat{e}\cdot \vec{r}}{\sin \alpha}\right ) f\left ( \vec{r}\right ) d\vec{r}dt\nonumber \\ & = \int \exp \left ( \frac{i\left ( \vec{r}^{2}-\lambda^{2}\right ) } { 2\tan \alpha}\right ) \delta \left ( \lambda-\hat{e}\cdot \vec{r}\right ) f\left ( \vec{r}\right ) d\vec{r } , \label{18.12}\ ] ] which completes the @xmath292-dimensional fractional radon transformation .
especially , when @xmath893 ( [ 18.12 ] ) reduces to the usual radon transform ( [ 18.4 ] ) .
now we examine if the additive property of frft is consistent with ( [ 18.12 ] ) . according to the additive property of frft @xmath894 and ( [ 18.9 ] ) we see@xmath895 \nonumber
\\ & = \left ( c_{\beta}\right ) ^{n}\left ( c_{\alpha}\right ) ^{n}\int \int \exp \left ( \frac{i\left ( \vec{r}^{2}+\vec{\xi}^{2}\right ) } { 2\tan \alpha } + \frac{i\left ( \vec{\xi}^{2}+\vec{k}^{2}\right ) } { 2\tan \beta}\right ) \nonumber \\ & \times \exp \left ( -\frac{i\vec{\xi}\cdot \vec{r}}{\sin \alpha}-\frac{i\vec { k}\cdot \vec{\xi}}{\sin \beta}\right ) f\left ( \vec{r}\right ) d\vec{r}d\vec{\xi}\nonumber \\ & = \mathfrak{f}_{\beta+\alpha,\vec{k}}\left [ f\right ] .
\label{18.13}\ ] ] the corresponding one - dimensional inverse frft should be@xmath896 ^{\left ( 1-n\right ) /2}c_{-\beta}c_{-\alpha } \int_{-\infty}^{\infty}\exp \left ( -\frac{i\left ( \lambda^{2}+\mu ^{2}\right ) } { 2\tan \beta}-\frac{i\left ( t^{2}+\mu^{2}\right ) } { 2\tan \alpha } \right ) \nonumber \\ & \times \exp \left ( + \frac{i\lambda \mu}{\sin \beta}+\frac{i\mu t}{\sin \alpha } \right ) f_{\alpha+\beta}\left ( t\hat{e}\right ) dtd\mu \nonumber \\ & = \left [ \frac{e^{i\left ( \alpha+\beta \right ) } } { 2\pi i\sin \left ( \alpha+\beta \right ) } \right ] ^{\left ( 1-n\right ) /2}c_{-\left ( \alpha+\beta \right ) } \int_{-\infty}^{\infty}f_{\alpha+\beta}\left ( t\hat { e}\right ) \exp \left ( -\frac{i\left ( t^{2}+\lambda^{2}\right ) } { 2\tan \left ( \alpha+\beta \right ) } + \frac{i\lambda t}{\sin \left ( \alpha+\beta \right ) } \right ) dt\nonumber \\ & = \int \exp \left ( \frac{i\left ( \vec{r}^{2}-\lambda^{2}\right ) } { 2\tan \left ( \alpha+\beta \right ) } \right ) \delta \left ( \lambda-\hat { e}\cdot \vec{r}\right ) f\left ( \vec{r}\right ) d\vec{r}=f_{r,\alpha+\beta } \left ( \lambda,\hat{e}\right ) , \label{18.14}\ ] ] which coincides with ( [ 18.12 ] ) . from ( [ 18.12 ] ) and ( [ 18.14 ] ) , we can confirm that the transform kernel of @xmath40th frft is @xmath897 for example , one can calculate the fractional radon transform of the @xmath871mode wigner operator to obtain some new quantum mechanical representations .
finally we give the inversion of the fractional radon transformation , from ( [ 18.12 ] ) we have @xmath898 which is an extension of ( [ 18.8 ] ) . in summary ,
based on the radon transform and fractional fourier transform we have naturally introduced the @xmath292-dimensional frft , in ref .
zalevsky and mendlovic @xcite also defined 2-dimensional frft , but in different approach .
we have identified the transform kernel for frft .
the generalization to complex fractional radon transformation is also possible @xcite .
in recent years wavelet transforms @xcite have been developed which can overcome some shortcomings of the classical fourier analysis and therefore has been widely used in fourier optics and information science since 1980s . here
we present a quantum optical version of classical wavelet transform ( wt ) by virtue of the iwop technique .
a wavelet has its energy concentrated in time to give a tool for the analysis of transient , nonstationary , or time - varying phenomena .
( it is a wavelet because it is localized and it resembles a wave because it oscillates . )
mathematically , wavelets are defined by starting with a function @xmath899 of the real variable @xmath900 , named a mother wavelet which is required to decrease rapidly to zero as @xmath901 tends to infinity , @xmath902 a more general requirement for a mother wavelet is to demanded @xmath903 to have vanishing moments @xmath904 @xmath905 ( a greater degree of smoothness than continuity also leads to vanishing moments for the mother wavelet ) .
the theory of wavelets is concerned with the representation of a function in terms of a two - parameter family of dilates and translates of a fixed function .
the mother wavelet @xmath899 generates the other wavelets of the family @xmath906 , ( @xmath907 is scaling parameter , @xmath375 is a translation parameter , @xmath908 the dilated - translated function is defined as @xmath909 while the wavelet integral transform of a signal function @xmath910 by @xmath899 is defined by @xmath911 we can express ( [ 15.3 ] ) as @xmath912 where @xmath913 is the state vector corresponding to the given mother wavelet , @xmath609 is the state to be transformed , and @xmath914 is the squeezing - translating operator @xcite , @xmath17 is the eigenvector of coordinate operator . in order to combine the wavelet transform with quantum states transform more tightly , using the iwop technique we can directly perform the integral in ( [ 15.5 ] ) ( @xmath915)@xmath916 \colon \nonumber \\ & = \sqrt{\frac{2\mu}{1+\mu^{2}}}\colon \exp \left [ \frac{1}{2\left ( 1+\mu ^{2}\right ) } \left ( \frac{s}{\mu}+\sqrt{2}a^{\dagger}+\sqrt{2}\mu a\right ) ^{2}-\sqrt{2}\frac{s}{\mu}a^{\dagger}-\frac{s^{2}}{2\mu^{2}}-q^{2}\right ] \colon . \label{15.6}\ ] ] this is the explicitly normal product form .
let @xmath94 , @xmath917 @xmath918 using the operator identity @xmath919 \colon,$ ] eq .
( [ 15.6 ] ) becomes@xmath920 \nonumber \\ & \times \exp \left [ \left ( a^{\dagger}a+\frac{1}{2}\right ) \ln \operatorname{sech}\lambda \right ] \nonumber \\ & \times \exp \left [ \frac{a^{2}}{2}\tanh \lambda+\frac{sa}{\sqrt{2}}\operatorname{sech}\lambda \right ] .
\label{15.7}\ ] ] in particular , when @xmath921 , it reduces to the well - known squeezing operator , @xmath922 for a review of the squeezed state theory we refer to @xcite .
now we analyze the condition ( [ 15.1 ] ) for mother wavelet from the point of view of quantum optics . due to @xmath923 where @xmath110 is the momentum eigenstate , we can recast the condition into quantum mechanics as @xmath924 which indicates that the probability of a measurement of @xmath925 by the projection operator @xmath926 with value @xmath927 is zero . without loss of generality , we suppose @xmath928 where @xmath929 are such chosen as to letting @xmath413 obeying the condition ( [ 15.1 ] ) . using the coherent states overcompleteness relation we have@xmath930 eq.([15.12 ] ) provides a general formalism to find the qualified wavelets .
for example , assuming @xmath931 for @xmath932 , so the coefficients of the survived terms should satisfy @xmath933 and @xmath413 becomes @xmath934 projecting it onto the coordinate representation , we get the qualified wavelets@xmath935 , \label{15.66}\ ] ] where we have used @xmath936 , and @xmath937 is the hermite polynomials .
now we take some examples .
case 1 : in ( [ 15.12 ] ) by taking @xmath938 @xmath939 @xmath931 ( otherwise)@xmath10 we have @xmath940 it then follows @xmath941 which is just the maxican hat wavelet , satisfying the condition @xmath942hence @xmath943 is the state vector corresponding to the maxican hat mother wavelet ( see fig .
once the state vector @xmath913 corresponding to mother wavelet is known , for any state @xmath609 the matrix element @xmath944 is just the wavelet transform of @xmath945 with respect to @xmath946 [ fig3 ] [ fig4 ] when @xmath947 , @xmath948 , @xmath949and @xmath950.,title="fig:",width=226 ] [ fig5 ] , @xmath951 , @xmath949and @xmath950.,title="fig:",width=226 ] case 2 : when @xmath947 , @xmath948 , @xmath949 and @xmath950 , from ( [ 15.66 ] ) we obtain ( see fig . 4)@xmath952 which also satisfies @xmath953 .
note that when @xmath954 , @xmath951 , @xmath949 and @xmath950 , we obtain a slightly different wavelet ( see fig .
therefore , as long as the parameters @xmath955 conforms to condition ( [ 15.64 ] ) , we can adjust their values to control the shape of the wavelets .
case 3 : when @xmath956 , @xmath957 , @xmath958 and @xmath959 , we get ( see fig . 6)@xmath960 and @xmath961 . from these figures
we observe that the number of the nodes of the curves at the @xmath900-axis is equal to the highest power of the wavelet functions .
[ fig6 ] when @xmath956 , @xmath957 , @xmath958and @xmath959.,title="fig:",width=226 ] to further reveal the properties of the newly found wavelets , we compare the wavelet transform computed with the well - known mexican hat wavelet @xmath962 and that with our new wavelet @xmath963 .
concretely , we map a simple cosine signal @xmath964 by performing the wavelet transforms with @xmath965 $ ] , @xmath966 , into a two - dimensional space @xmath967 , where @xmath968 denotes the location of a wavelet and @xmath76 its size . the resulting wavelet transforms by @xmath969 ( = @xmath970 ) and @xmath971 are@xmath972 \cos \pi x,\label{15.69}\\ \omega_{2}\left ( x , t\right ) & = \frac{1}{\sqrt{30}}\int_{-\infty}^{\infty } dx\psi_{2}\left [ t\left ( x - x\right ) \right ] \cos \pi x , \label{15.70}\ ] ] where @xmath973 and @xmath974 are the normalization factors for @xmath975 and @xmath976 respectively , the wavelet integral @xmath977 are also called wavelet coefficient which measures the variation of cos@xmath978 in a neighborhood of @xmath979 whose size is proportional to @xmath980 .
the contour line representation of @xmath981 and @xmath982 are depicted in fig . 7 and fig .
8 , respectively , where the transverse axis is @xmath968-axis ( time axis ) , while the longitudinal axis ( @xmath983-axis ) is the frequency axis .
[ fig7 ] , title="fig:",width=302 ] [ fig8 ] , title="fig:",width=302 ] it is remarkable that although two overall shapes of the two contour lines look similar , there exist two notable differences between these two figures : 1 ) along @xmath983-axis @xmath984 has one maximum , while @xmath982 has one main maximum and one subsidiary maximum ( two islands `` ) , so when @xmath976 scales its size people have one more chance to identify the frequency information of the cosine wave than using @xmath985 .
interesting enough , the two islands '' of @xmath982 in fig .
8 can be imagined as if they were produced while the figure of @xmath984 deforms into two sub - structures along @xmath76-axis .
2 ) near the maximum of @xmath982 the density of the contour lines along @xmath76-axis is higher than that of @xmath984 , which indicates that the new wavelet @xmath976 is more sensitive in detecting frequency information of the signal at this point . therefore , @xmath963 may be superior to @xmath986 in analyzing some signals . finally , we mention that there exist some remarkable qualitative similarities between the mother wavelets presented in figs . 3 through 6 and some of the amplitude envelopes of higher order laser spatial modes and spatial supermodes of phase locked diode laser arrays @xcite , which are due to spatial coherence . in this subsection , we shall prove that the parseval theorem of 1d wt @xcite:@xmath987 where @xmath988 is a mother wavelet whose fourier transform is @xmath989 @xmath990 . in the context of quantum mechanics , according to eq.([15.4 ] ) we see that the quantum mechanical version of parseval theorem should be @xmath991 and since @xmath992 , so @xmath993 involved in @xmath994 is @xmath995 @xmath996 is the momentum eigenvector@xmath997 eq.([15.17 ] ) indicates that once the state vector @xmath998 corresponding to mother wavelet is known , for any two states @xmath999 and @xmath1000 , their overlap up to the factor @xmath994 ( determined by eq.([15.31 ] ) ) is just their corresponding overlap of wts in the ( @xmath1001 parametric space .
_ proof of equation _ * ( [ 15.17 ] ) * : in order to show eq.(*[15.17 ] * ) , we calculate @xmath1002 which leads to @xmath1003 where we have used the formula@xmath1004 inserting the completeness relation @xmath1005 into the left side of eq.([15.16 ] ) and then using eq.([15.20 ] ) we have@xmath1006 where @xmath1007 \left \langle f_{2}\right .
\left \vert p\right \rangle \left \langle p\right \vert \left .
f_{1}\right \rangle , \label{15.23}\ ] ] and @xmath1008 \left \langle f_{2}\right .
\left \vert p\right \rangle \left \langle p\right \vert \left .
f_{1}\right \rangle . \label{15.24}\ ] ] further , we can put eqs.([15.23 ] ) and ( [ 15.24 ] ) into the following forms,@xmath1009 \left \langle f_{2}\right .
\left \vert p\right \rangle \left \langle p\right \vert \left .
f_{1}\right \rangle \nonumber \\ & = c_{\psi}\int_{0}^{\infty}dp\left \langle f_{2}\right .
\left \vert p\right \rangle \left \langle p\right \vert \left .
f_{1}\right \rangle + 2\pi \int_{-\infty}^{0}dp\left [ \int_{0}^{-\infty}\left \vert \psi \left ( p^{\prime}\right ) \right \vert ^{2}\frac{dp^{\prime}}{p^{\prime}}\right ] \left \langle f_{2}\right .
\left \vert p\right \rangle \left \langle p\right \vert \left .
f_{1}\right \rangle \nonumber \\ & = c_{\psi}\int_{0}^{\infty}dp\left \langle f_{2}\right .
\left \vert p\right \rangle \left \langle p\right \vert \left .
f_{1}\right \rangle + c_{\psi } ^{\prime}\int_{-\infty}^{0}dp\left \langle f_{2}\right .
\left \vert p\right \rangle \left \langle p\right \vert \left .
f_{1}\right \rangle \label{15.25}\ ] ] and@xmath1010 where @xmath1011 thus when the definite integration satisfies the admissible condition , i.e. , @xmath1012 which leads to @xmath1013 eq .
( [ 15.22 ] ) can be transformed to@xmath1014 where@xmath1015 thus the theorem is proved .
especially , when @xmath1016 @xmath1017 eq.([15.17 ] ) becomes@xmath1018 which is named isometry of energy .
now we can directly derive the inversion formula of wt , i.e. we take @xmath1019 in eq.([15.17 ] ) , then using eq.([15.4 ] ) we see that eq.([15.17 ] ) reduces to @xmath1020 due to eq.([15.5 ] ) we have@xmath1021 it then follows@xmath1022 which means@xmath1023 this is the inversion formula of wt . form the parserval theorem ( [ 15.16 ] ) of wt in quantum mechanics we can derive some new property of mother wavelet @xcite . taking @xmath1024 , @xmath1025 in ( [ 15.16 ] ) one can see that@xmath1026 which is a new orthogonal property of mother wavelet in parameter space spanned by @xmath1027 . in a similar way , we take @xmath1028 a number state , since @xmath1029 then we have @xmath1030 or take @xmath1031 @xmath1032 is the coherent state , then @xmath1033 this indicates that @xmath994 is @xmath1034-independent , which coincides with the expression in ( [ 15.31 ] ) .
next , we consider a special example .
when the mother wavelet is the mexican hat ( [ 15.14 ] ) , we have@xmath1035 where@xmath1036 here @xmath1037 is the single - variable hermit polynomial @xcite . substituting eq.([15.40 ] ) into eq.([15.31 ] ) we have@xmath1038 thus , for the mexican hat wavelet ( [ 15.14 ] ) , we see @xmath1039 eq.([15.43 ] ) can be checked as follows . using eq.([15.14 ] ) and noticing that @xmath1040
, we can put the left hand side of eq.([15.43 ] ) into@xmath1041{cc}0 , & x\neq x^{\prime}\\ \frac{3}{2}\int_{0}^{\infty}du\rightarrow \infty , & x = x^{\prime}\end{array } \right .
= \mathtt{r.h.s.of}\text { } ( \ref{15.43 } ) .
\label{15.44}\ ] ] where we have used the integration formula@xmath693s
@xmath1042 and@xmath1043 e^{-s^{2}/2-\left ( b - s\right ) ^{2}/2}ds\nonumber \\ & = \frac{\sqrt{\pi}}{16}e^{-\frac{b^{2}}{4}}\left [ 12+b^{2}\left ( b^{2}-12\right ) \right ] .
\label{15.46}\ ] ] next , we examine if the morlet wavelet obey the formalism ( [ 15.37 ] ) . the the morlet wavelet is defined as @xcite @xmath1044 substituting ( [ 15.47 ] ) into the left hand side of ( [ 15.37 ] ) yields@xmath1045{cc}0 , & x\neq x^{\prime}\\ 2\left ( 1+e^{-f^{2}}-2e^{-3f^{2}/4}\right ) \int_{0}^{\infty}\frac{d\mu}{\mu^{2}}\rightarrow \infty , & x = x^{\prime}\end{array } \right . .
\label{15.48}\ ] ] thus the morlet wavelet satisfies eq.([15.37 ] ) .
phase space technique has been proved very useful in various branches of physics .
distribution functions in phase space have been a major topic in studying quantum mechanics and quantum statistics . among various phase space distributions
the wigner function @xmath1046 @xcite is the most popularly used , since its two marginal distributions lead to measuring probability density in coordinate space and momentum space , respectively .
but the wigner distribution function itself is not a probability distribution due to being both positive and negative . in spite of its some attractive formal properties
, it needs to be improved . to overcome this inconvenience , the husimi distribution function @xmath1047 is introduced @xcite , which is defined in a manner that guarantees it to be nonnegative .
its definition is smoothing out the wigner function by averaging over a coarse graining " function , @xmath1048 , \label{15.49}\ ] ] where @xmath1049 is the gaussian spatial width parameter , which is free to be chosen and which determines the relative resolution in @xmath1050-space versus @xmath73-space . in the following
, we shall employ the optical wavelet transformation to study the husimi distribution function , this is to say , we shall show that the husimi distribution function of a quantum state @xmath1051 can be obtained by making a wt of the gaussian function @xmath1052 i.e. , @xmath1053 where @xmath1054 and @xmath1055 is the husimi distribution function as well as @xmath1056 is the husimi operator , @xmath1057 here @xmath1058 denotes normal ordering ; @xmath338 and @xmath339 are the coordinate and the momentum operator .
* proof of eq.([15.50 ] ) .
* according to eqs.([15.3 ] ) and ( [ 15.4 ] ) , when @xmath609 is the vacuum state @xmath329 , @xmath1059 , we see that @xmath1060 from eq.([15.7 ] ) it then follows that @xmath1061 \left \vert 0\right \rangle . \label{15.54}\ ] ] substituting eq.([15.51 ] ) and @xmath1062 @xmath1063 into eq.([15.54 ] ) yields@xmath1064 then the wt of eq.([15.53 ] ) can be further expressed as@xmath1065 using normally ordered form of the vacuum state projector @xmath1066and the iwop method as well as eq.([15.55 ] ) we have@xmath1067 \colon=\delta_{h}\left ( q , p,\kappa \right ) .
\label{15.57}\ ] ] now we explain why @xmath1068 is the husimi operator . using the formula for converting an operator @xmath579 into its weyl ordering form @xcite@xmath1069 where the symbol @xmath1070 denotes the weyl ordering , @xmath1071 is the usual coherent state , substituting eq.([15.57 ] ) into eq.([15.58 ] ) and performing the integration by virtue of the technique of integration within a weyl ordered product of operators , we obtain@xmath1072 \genfrac{}{}{0pt}{}{:}{:}. \label{15.59}\ ] ] this is the weyl ordering form of @xmath1073 then according to weyl quantization scheme @xcite we know the classical corresponding function of a weyl ordered operator is obtained by just replacing @xmath1074 @xmath1075 \genfrac{}{}{0pt}{}{:}{:}\rightarrow \exp \left [ -\kappa \left ( q - q^{\prime}\right ) ^{2}-\frac{\left ( p - p^{\prime}\right ) ^{2}}{\kappa}\right ] , \label{15.60}\ ] ] and in this case the weyl rule is expressed as @xmath1076 \nonumber \\ & = 2\int dq^{\prime}dp^{\prime}\delta_{w}\left ( q^{\prime},p^{\prime } \right ) \exp \left [ -\kappa \left ( q^{\prime}-q\right ) ^{2}-\frac{\left ( p^{\prime}-p\right ) ^{2}}{\kappa}\right ] , \label{15.61}\ ] ] where at the last step we used the weyl ordering form of the wigner operator @xmath1077 @xcite @xmath1078 in reference to eq.([15.49 ] ) in which the relation between the husimi function and the wf is shown , we know that the right - hand side of eq .
( [ 15.61 ] ) should be just the husimi operator , i.e. @xmath1079 \nonumber \\ & = \delta_{h}\left ( q , p,\kappa \right ) , \label{15.63}\ ] ] thus eq .
( [ 15.50 ] ) is proved by combining eqs.([15.63 ] ) and ( [ 15.56 ] ) .
thus the optical wt can be used to study the husimi distribution function in quantum optics phase space theory @xcite .
we now turn to 2-dimensional complex wavelet transform ( cwt ) @xcite . since wavelet family involves squeezing transform , we recall that the two - mode squeezing operator has a natural representation in the entangled state representation ( esr ) , @xmath1080=\frac{1}{\mu}\int_{-\infty}^{\infty}\left \vert \frac{\eta}{\mu}\right \rangle \left \langle
\eta \right \vert dx,$ ] @xmath1081 , thus we are naturally led to studying 2-dimensional cwt in esr . using esr we can derive some new results more conveniently than using the direct - product of two single - particle coordinate eigenstates . to be concrete ,
we impose the condition on qualified mother wavelets also in @xmath9 representation , @xmath1082 where @xmath1083 @xmath1084 thus we see @xmath1085 and the condition ( [ 16.1 ] ) becomes @xmath1086 without loss of generality , assuming @xmath1087 then using the two - mode coherent @xmath1088 state we can write ( [ 16.3 ] ) as @xmath1089 \nonumber \\ \ & = \sum_{n , m=0}^{\infty}k_{n , m}\int \frac{d^{2}z_{2}}{\pi}\exp \left [ -|z_{2}|^{2}\right ] z_{2}^{n}z_{2}^{\ast m}\left ( -1\right ) ^{n}\nonumber \\ & = \sum_{n=0}^{\infty}n!k_{n , n}\left ( -1\right ) ^{n}=0 , \label{16.5}\ ] ] this is the constraint on the coefficient @xmath1090 in ( [ 16.5 ] ) , i.e. , the admissibility condition for @xmath413 .
( [ 16.4 ] ) is in the form:@xmath1091 to derive the qualified mother wavelet @xmath1083 @xmath1092 from @xmath925 , noticing eq.([13.9 ] ) and ( [ 16.6 ] ) we have@xmath1093 where @xmath1094 is the laguerre polynomial . in this case
, we may name the wavelet in eq .
( [ 16.7 ] ) as the laguerre
gaussian mother wavelets , analogous to the name of laguerre
gaussian modes in optical propagation .
for example : \(1 ) when taking @xmath1095 @xmath1096 @xmath1097 for @xmath1098 so we see @xmath1099 which differs from the direct - product state @xmath1100 .
it then follows from eq .
( [ 16.7 ] ) that @xmath1101 which differs from @xmath1102 , the direct product of two 1d mexican hat wavelets ( see also the difference between figs . 9 and 10 ) .
\(2 ) when @xmath1103 , @xmath1104 , @xmath1105 , @xmath1097 for @xmath1106 we have ( see fig .
11)@xmath1107 \(3 ) when @xmath1103 , @xmath1108 , @xmath1109 , @xmath1110 , @xmath1097 for @xmath1111 the mother wavelet @xmath1112 ( see fig .
12 ) reads @xmath1113 from the figures we can see that as long as the coefficients @xmath1090 satisfy condition ( [ 16.7 ] ) , we can construct arbitrary complex mother wavelet by adding or reducing the number of coefficients , or by adjusting the value of them . and
since only @xmath1114 ( @xmath1115 ) survive in all the coefficients , the mother wavelets obtained are all circularly symmetric on the complex plane .
moreover , the cwt of a signal function @xmath1116 by @xmath1117 is defined by @xmath1118 using the @xmath132 representation we can treat it from the quantum mechanically , @xmath1119 where @xmath1120 is the two - mode squeezing - displacing operator . using the iwop technique we can calculate its normally ordered form,@xmath1121 when @xmath1122 it reduces to the usual normally ordered two - mode squeezing operator . once the state vector @xmath1123 corresponding to mother wavelet is known , for any state @xmath1124
the matrix element @xmath1125 is just the wavelet transform of @xmath1126 with respect to @xmath1127 therefore , various quantum optical field states can then be analyzed by their wavelet transforms . in order to complete the cwt theory ,
we must ask if the corresponding parseval theorem exists @xcite . this is important since the inversion formula of cwt may appear as a lemma of this theorem . noting that cwt involves two - mode squeezing transform ,
so the corresponding parseval theorem differs from that of the direct - product of two 1d wavelet transforms , too .
next let us prove the parseval theorem for cwt,@xmath1128 where @xmath1129 and @xmath1130 @xmath1131 is the complex fourier transform of @xmath1132 . according to ( [ 16.13 ] ) and ( [ 16.14 ] )
the quantum mechanical version of parseval theorem should be@xmath1133 eq.([16.18 ] ) indicates that once the state vector @xmath998 corresponding to mother wavelet is known , for any two states @xmath1134 and @xmath1135 , their overlap up to the factor @xmath994 ( determined by ( [ 16.17 ] ) ) is just their corresponding overlap of cwts in the ( @xmath1136 ) parametric space . next we prove eq.(*[16.16 ] ) or ( [ 16.18 ] ) .
* in the same procedure as * * the proof of eq.([15.17])*. * we start with calculating @xmath1137 using ( [ 3.20 ] ) and ( [ 16.14 ] ) , we have@xmath1138 it then follows@xmath1139 } \left \vert \frac{\xi^{\prime}}{\mu } \right \rangle \left \langle \frac{\xi}{\mu}\right \vert \nonumber \\ & = 4\pi \left \vert \frac{\xi}{\mu}\right \rangle \left \langle \frac{\xi}{\mu } \right \vert \delta \left ( \xi_{1}^{\prime}-\xi_{1}\right ) \delta \left ( \xi_{2}-\xi_{2}^{\prime}\right ) .
\label{16.20}\ ] ] using the completeness of @xmath142 and ( [ 16.20 ] ) the left - hand side ( lhs ) of ( [ 16.18 ] ) can be reformed as@xmath1140 where the integration value in @xmath1141 is actually @xmath1142independent .
noting that the mother wavelet @xmath1143 in eq.([16.7 ] ) is just the function of @xmath1144 so @xmath1131 is also the function of @xmath1145 in fact , using eqs.([16.7 ] ) and ( [ 3.20 ] ) , we have @xmath1146 where we have used the integral formula@xmath1147 so we can rewrite ( [ 16.21 ] ) as @xmath1148 where@xmath1149 then we have completed the proof of the parseval theorem for cwt in ( [ 16.18 ] ) . here , we should emphasize that ( [ 16.18 ] ) is not only different from the product of two 1d wts , but also different from the usual wt in 2d . when @xmath1150 by using ( [ 16.14 ] ) we see @xmath1151 then substituting it into ( [ 16.18 ] ) yields@xmath1152 which is just the inverse transform of the cwt .
especially , when @xmath1153 @xmath1154 eq .
( [ 16.18 ] ) reduces to@xmath1155 which is named isometry of energy . on the other hand ,
when @xmath1156 @xmath1157 , eq.([16.18 ] ) becomes@xmath1158 which is a new orthogonal property of mother wavelet in parameter space spanned by @xmath1159 . in a similar way
, we take @xmath1160 a two - mode number state , since @xmath1161 then we have @xmath1162 or take @xmath1163 @xmath1164 is the coherent state , then @xmath1165 next we examine a special example . when the mother wavelet is @xmath1166 in ( [ 16.9 ] ) , using ( [ 3.20 ] ) we have @xmath1167 which leads to @xmath1168 thus for @xmath1166 , we see@xmath1169 eq . ( [ 16.31 ] ) can be checked as follows . using ( [ 16.9 ] ) and the integral formula@xmath1170 we can put the left - hand side ( lhs ) of ( [ 16.31 ] ) into @xmath1171 where @xmath1172 @xmath1173 when @xmath1174 @xmath1175@xmath1176 on the other hand , when @xmath1177 and noticing that @xmath1178 which leads to @xmath1179 , where @xmath1180{cc}\kappa_{1}-\eta_{1}^{\prime } & \kappa_{2}-\eta_{2}^{\prime}\\ \kappa_{1}-\eta_{1 } & \kappa_{2}-\eta_{2}\end{array } \right \vert $ ] . as a result of ( [ 16.35 ] ) , ( [ 16.33 ] ) reduces to @xmath1181 where we have noticed that @xmath1182 is the funtion of @xmath1183 thus we have@xmath1184{cc}\infty , & \eta=\eta^{\prime},\\ 0 , & \eta \neq \eta^{\prime}. \end{array } \right .
= \text{rhs of ( \ref{16.31 } ) . } \label{16.37}\ ] ] recalling that in ref.@xcite , the so - called entangled husimi operator @xmath1185 has been introduced , which is endowed with definite physical meaning , and it is found that the two - mode squeezed coherent state @xmath1186 representation of @xmath1187 @xmath1188 @xmath1189 . the entangled husimi operator @xmath1190 and the entangled husimi distribution @xmath1191 of quantum state @xmath413 are given by@xmath1192 and @xmath1193 respectively , where @xmath1194 is two - mode wigner function , with @xmath1195 being the two - mode wigner operator .
thus we are naturally led to studying the entangled husimi distribution function from the viewpoint of wavelet transformation . in this subsection , we shall extend the relation between wavelet transformation and wigner - husimi distribution function to the entangled case , that is to say , we employ the cwt to investigate the entangled husimi distribution function ( ehdf ) by bridging the relation between cwt and ehdf .
this is a convenient approach for calculating various entangled husimi distribution functions of miscellaneous two - mode quantum states . in ref.@xcite ,
the cwt has been proposed , i.e. , the cwt of a complex signal function @xmath1196 by @xmath899 is defined by @xmath1197 whose admissibility condition for mother wavelets , @xmath1198 is examined in the entangled state representations @xmath132 and a family of new mother wavelets ( named the laguerre
gaussian wavelets ) are found to match the cwt @xcite .
in fact , by introducing the bipartite entangled state representation @xmath1199we can treat ( [ e5 ] ) quantum mechanically , @xmath1200 where @xmath1201 @xmath1202 @xmath1203and @xmath1204 are the wavefunction of state vector @xmath196 and the mother wavelet state vector @xmath413 in @xmath160 representation , respectively , and @xmath1205 is the two - mode squeezing - displacing operator . noticing that the two - mode squeezing operator has its natural expression in @xmath1206 representation ( [ 3.24 ] ) , which is different from the direct product of two single - mode squeezing ( dilation ) operators , and the two - mode squeezed state is simultaneously an entangled state , thus we can put eq.([e9 ] ) into the following form,@xmath1207 where @xmath1208 is a two - mode displacement operator , @xmath1209 and @xmath1210 \nonumber \\ & = d_{1}\left ( -z/2\right ) d_{2}\left ( z^{\ast}/2\right ) .
\label{e12}\ ] ] it the follows the quantum mechanical version of cwt is@xmath1211 eq.([e13 ] ) indicates that the cwt can be put into a matrix element in the @xmath132 representation of the two - mode displacing and the two - mode squeezing operators in eq.([e10 ] ) between the mother wavelet state vector @xmath413 and the state vector @xmath205 to be transformed .
once the state vector @xmath998 as mother wavelet is chosen , for any state @xmath205 the matrix element @xmath1212 is just the wavelet transform of @xmath1213 with respect to @xmath946 therefore , various quantum optical field states can then be analyzed by their wavelet transforms . in the following
we shall show that the ehdf of a quantum state @xmath925 can be obtained by making a complex wavelet transform of the gaussian function @xmath1214 i.e. , @xmath1215 where @xmath1216 @xmath1217 and @xmath1218 , \label{e15}\\ z_{2 } & = \frac{i\cosh \lambda}{1+\kappa}\left [ \gamma+\gamma^{\ast}+\kappa \left ( \sigma-\sigma^{\ast}\right ) \right ] , \label{e16}\ ] ] @xmath1185 is named the entangled husimi operator by us , @xmath1219 @xmath1220 is the husimi distribution function .
* proof of eq.([e14 ] ) . * when the state to be transformed is @xmath1221 ( the two - mode vacuum state ) , by noticing that @xmath1222 we can express eq.([e8 ] ) as@xmath1223 to combine the cwts with transforms of quantum states more tightly and clearly , using the iwop technique we can directly perform the integral in eq.([e9 ] ) @xcite@xmath1224 \nonumber \\ & \times \exp \left [ \left ( a_{1}^{\dagger}a_{1}+a_{2}^{\dagger}a_{2}\right ) \ln
\operatorname{sech}\lambda \right ] \exp \left ( \frac{z^{\ast}a_{1}-za_{2}}{1+\mu^{2}}-a_{1}a_{2}\tanh \lambda \right ) .
\label{e19}\ ] ] where we have set @xmath94 , @xmath1225 , @xmath1226 , and we have used the operator identity @xmath1227 \colon$ ] . in particular , when @xmath1228 @xmath1229 becomes to the usual normally ordered two - mode squeezing operator @xmath1230 . from eq.([e19 ] ) it then follows that @xmath1231 \operatorname{sech}\lambda \right \ }
\left \vert 00\right \rangle .
\label{e20}\ ] ] substituting eqs.([e15 ] ) , ( [ e16 ] ) and @xmath1232 @xmath1063 into eq.([e20 ] ) yields@xmath1233 then the cwt of eq.([e18 ] ) can be further expressed as@xmath1234 using normally ordered form of the vacuum state projector @xmath1235 and the iwop method as well as eq.([e21 ] ) we have@xmath1236 \colon \nonumber \\ & = \frac{4\kappa}{\left ( 1+\kappa \right ) ^{2}}\colon \exp \left \
{ -\frac{\left ( a_{1}+a_{2}^{\dag}-\gamma \right ) \left ( a_{1}^{\dag}+a_{2}-\gamma^{\ast}\right ) } { 1+\kappa}\right .
\nonumber \\ & -\left .
( a_{1}-a_{2}^{\dag}-\sigma \right ) \left ( a_{1}^{\dag}-a_{2}-\sigma^{\ast}\right ) } { 1+\kappa}\right \ } \colon \left .
= \right .
\delta_{h}\left ( \sigma,\gamma,\kappa \right ) . \label{e23}\ ] ] now we explain why @xmath1185 is the entangled husimi operator . using the formula for converting an operator @xmath579 into its weyl ordering form @xcite@xmath1237 where the symbol @xmath1070 denotes the weyl ordering , @xmath1071 is the usual coherent state , substituting eq.([e23 ] ) into eq.([e24 ] ) and performing the integration by virtue of the technique of integration within a weyl ordered product of operators , we obtain@xmath1238 where we have used the integral formula@xmath1239 eq.([e25 ] ) is the weyl ordering form of @xmath1240 then according to weyl quantization scheme we know the weyl ordering form of two - mode wigner operator is given by@xmath1241 thus the classical corresponding function of a weyl ordered operator is obtained by just replacing @xmath1242 i.e.,@xmath1243 and in this case the weyl rule is expressed as@xmath1244 in reference to eq.([e5 ] ) in which the relation between the entangled husimi function and the two - mode wigner function is shown , we know that the right - hand side of eq .
( [ e29 ] ) should be just the entangled husimi operator , i.e. @xmath1245 thus eq .
( [ e14 ] ) is proved by combining eqs.([e30 ] ) and ( [ e22 ] ) .
thus we have further extended the relation between wavelet transformation and wigner - husimi distribution function to the entangled case . that is to say , we prove that the entangled husimi distribution function of a two - mode quantum state @xmath413 is just the modulus square of the complex wavelet transform of @xmath1246 with @xmath1143 being the mother wavelet up to a gaussian function , i.e. , @xmath1247 . thus is a convenient approach for calculating various entangled husimi distribution functions of miscellaneous quantum states .
in this section we shall generalize the usual wavelet transform to symplectic wavelet transformation ( swt ) by using the coherent state representation @xcite .
first we are motivated to generalize the usual wavelet transform , which concerns about dilation , to optical fresnel transform ( we will explain this in detail in section below ) , i.e. we shall use the symplectic - transformed translated versions of the mother wavelet @xmath1248 \label{17.1}\ ] ] as a weighting function to synthesize the original complex signal @xmath1249 , @xmath1250 this is named the symplectic - transformed translated wavelet transform .
one can see that the mother wavelet @xmath899 generates the other wavelets of the family @xmath1251 $ ] through a translating transform followed by a symplectic transform , ( @xmath1252 are the symplectic transform parameter , @xmath1253 @xmath1254 is a translation parameter , @xmath375 , @xmath376 and @xmath1255 ) this can be seen more clearly by writing the second transform in matrix form @xmath287{c}z-\kappa \\ z^{\ast}-\kappa^{\ast}\end{array } \right ) \rightarrow m\left ( \begin{array } [ c]{c}z-\kappa \\ z^{\ast}-\kappa^{\ast}\end{array } \right ) , \text { } m\equiv \left ( \begin{array } [ c]{cc}s & -r\\ -r^{\ast } & s^{\ast}\end{array } \right ) , \label{17.3}\ ] ] where @xmath56 is a symplectic matrix satisfies @xmath1256 , @xmath1257{cc}0 & i\\ -i & 0 \end{array } \right ) $ ] .
symplectic matrices in hamiltonian dynamics correspond to canonical transformations and keep the poisson bracket invariant , while in matrix optics they represent ray transfer matrices of optical instruments , such as lenses and fibers .
it is straightforward to evaluate this transform and its reciprocal transform when @xmath1258 is the complex fourier exponentials , @xmath1259 ( note that @xmath1260 is pure imaginary):@xmath1261 \nonumber \\ & = \sqrt{s}\int \frac{d^{2}z}{\pi}\exp \left [ \left ( z+\kappa \right ) \beta^{\ast}-\left ( z^{\ast}+\kappa^{\ast}\right ) \beta \right ] \psi^{\ast } \left ( sz - rz^{\ast}\right ) \nonumber \\ & = \sqrt{s}\int \frac{d^{2}z^{\prime}}{\pi}\exp \left [ \left ( s^{\ast } z^{\prime}+rz^{\prime \ast}+\kappa \right ) \beta^{\ast}-\left ( sz^{\prime \ast}+r^{\ast}z^{\prime}+\kappa^{\ast}\right ) \beta \right ] \psi^{\ast } \left ( z^{\prime}\right ) \nonumber \\ & = \exp \left [ \kappa \beta^{\ast}-\kappa^{\ast}\beta \right ] \sqrt{s}\int \frac{d^{2}z^{\prime}}{\pi}\exp \left [ z^{\prime}\left ( s^{\ast}\beta^{\ast}-r^{\ast}\beta \right ) -z^{\prime \ast}\left ( s\beta - r\beta^{\ast } \right ) \right ] \psi^{\ast}\left ( z^{\prime}\right ) \nonumber \\ & = \sqrt{s}\exp \left [ \kappa \beta^{\ast}-\kappa^{\ast}\beta \right ] \phi \left ( s^{\ast}\beta^{\ast}-r^{\ast}\beta \right ) , \label{17.4}\ ] ] where @xmath1262 is the complex fourier transform of @xmath1263 then we form the adjoint operation@xmath1264 \nonumber \\ & = \left \vert s\right \vert \phi \left ( s^{\ast}\beta^{\ast}-r^{\ast}\beta \right ) \int \frac{d^{2}\kappa}{\pi}\exp \left [ \left ( \kappa+z\right ) \beta^{\ast}-\left ( \kappa^{\ast}+z^{\ast}\right ) \beta \right ] \psi \left [ -s\kappa+r\kappa^{\ast}\right ] \nonumber \\ & = \left \vert s\right \vert \exp \left ( z\beta^{\ast}-z^{\ast}\beta \right ) \phi \left ( s^{\ast}\beta^{\ast}-r^{\ast}\beta \right ) \nonumber \\ & \times \int \frac{d^{2}\kappa^{\prime}}{\pi}\exp \left [ -\kappa^{\prime } \left ( s^{\ast}\beta^{\ast}-r^{\ast}\beta \right ) + \kappa^{\prime \ast } \left ( s\beta - r\beta^{\ast}\right ) \right ] \psi \left ( \kappa^{\prime } \right ) \nonumber \\ & = \left \vert s\right \vert \exp \left ( z\beta^{\ast}-z^{\ast}\beta \right ) \left
\vert \phi \left ( s^{\ast}\beta^{\ast}-r^{\ast}\beta \right ) \right \vert ^{2 } , \label{17.5}\ ] ] from which we have@xmath1265 so we get the inversion formula@xmath1266 eq.([17.7 ] ) leads us to impose the normalization @xmath1267 in order to get the wavelet representation@xmath1268 then we can have a form of parseval s theorem for this new wavelet transform : * preposition : * for any @xmath1269 and @xmath1270 we have@xmath1271 proof : let us assume @xmath1272 and @xmath1273 are the complex fourier transform of @xmath1258 and @xmath1274 respectively , @xmath1275 recall the convolution theorem defined on complex fourier transform , @xmath1276 so from ( [ 17.12 ] ) and ( [ 17.1 ] ) , ( [ 17.2 ] ) we see that @xmath1277 can be considered as a convolution in the form @xmath1278 \nonumber \\ & = \int d^{2}\beta f(\beta)\phi^{\ast}\left ( s\beta - r\beta^{\ast}\right ) \exp \left ( \kappa \beta^{\ast}-\kappa^{\ast}\beta \right ) \label{17.13}\ ] ] it then follows from ( [ 17.12 ] ) that @xmath1279 therefore , using ( [ 17.8 ] ) we see that the further integration yields@xmath1280 which completes the proof .
* theorem * : from the proposition ( [ 17.10 ] ) we have @xmath1281 that is , there exists an inversion formula for arbitrary function @xmath1249 .
in fact , in eq .
( [ 17.2 ] ) when we take @xmath1282 then @xmath1283 substituting ( [ 17.17 ] ) into ( [ 17.15 ] ) we obtain ( [ 17.16 ] ) .
now we explain why the idea of @xmath1284 is originated from the optical fresnel transform .
we can visualize the symplectic - transformed translated wavelet transform in the context of quantum mechanics , letting @xmath1285 , @xmath1286 is the coherent state , @xmath1287 \equiv \left \vert \left ( \begin{array } [ c]{c}z\\ z^{\ast}\end{array } \right ) \right \rangle $ ] , @xmath1288 is the vacuum state in fock space , then eq . (
[ 17.1 ] ) can be expressed as@xmath1289
f\left ( z\right ) \nonumber \\ & = \sqrt{s}\int \frac{d^{2}z}{\pi}\left \langle \psi \right \vert \left .
\left ( \begin{array } [ c]{cc}s & -r\\ -r^{\ast } & s^{\ast}\end{array } \right ) \left ( \begin{array } [ c]{c}z-\kappa \\ z^{\ast}-\kappa^{\ast}\end{array } \right ) \right \rangle \left \langle z\right \vert \left .
f\right \rangle \nonumber \\ & = \left \langle \psi \right \vert f_{1}\left ( r , s,\kappa \right ) \left \vert f\right \rangle , \label{17.18}\ ] ] where @xmath1290 is defined as @xmath1291 and @xmath1292{cc}s & -r\\ -r^{\ast } & s^{\ast}\end{array } \right ) \left ( \begin{array } [ c]{c}z\\ z^{\ast}\end{array } \right ) \right \rangle .$]to know the explicit form of @xmath1293 , we employ the normal ordering of the vacuum projector @xmath1294 and the iwop technique to perform the integration in ( [ 17.19 ] ) , which leads to@xmath1295 \exp \left ( \frac{r^{\ast}}{2s^{\ast}}a^{2}-\frac{1}{2s^{\ast}}\left ( s^{\ast}\kappa^{\ast}+r^{\ast}\kappa \right ) a\right ) .
\label{17.20}\ ] ] the transformation matrix element of @xmath1296 in the coordinate representation @xmath600 is just the kernel of optical diffraction integration ( [ 5.10 ] ) ( fresnel transform ) , this explains our motivation to introduce @xmath1284 .
in particular , when @xmath1297 , @xmath1298 @xmath1296 reduces to it the well - known single - mode squeezing operator @xmath1299 $ ] which corresponds to dilation in the usual wt . in the above subsection
, the mother wavelet is gained through a translating transform followed by a symplectic transform .
this motivation arises from the consideration that symplectic transforms are more general than the dilated transform , and are useful in fresnel transform of fourier optics , e.g. ray transfer matrices of optical instruments , such as lenses and fibers in matrix optics , while in quantum optics symplectic transforms correspond to single - mode fresnel operator ( or generalized su(1,1 ) squeezing operator ) . recalling that in section 9 we have introduced the 2-mode entangled fresnel operator which is a mapping of classical mixed transformation @xmath1300 in 2-mode coherent state @xmath1301 representation onto quantum operator @xmath1302 ,
thus we are naturally led to develop the swt in ( [ 17.2 ] ) to the so - called entangled swt ( eswt ) @xcite for signals @xmath1303 defined in two complex planes , @xmath1304 here @xmath1305 , \label{17.22}\ ] ] is used as a weighting function to synthesize the signal @xmath1303 regarding to two complex planes .
one can see that the mother wavelet @xmath1306 generates the family @xmath1307 $ ] through a translating transform followed by an entangled symplectic transform .
we emphasize that this transform mixes the two complex planes , which is different from the tensor product of two independent transforms @xmath1308 $ ] given by ( [ 17.18 ] ) .
the new symplectic transform can be seen more clearly by writing it in matrix form:@xmath287{c}z - k\\ z^{\ast}-k^{\ast}\\ z^{\prime}-k^{\prime}\\ z^{\prime \ast}-k^{^{\prime}\ast}\end{array } \right ) \longrightarrow \mathcal{m}\left ( \begin{array } [ c]{c}z - k\\ z^{\ast}-k^{\ast}\\ z^{\prime}-k^{\prime}\\ z^{\prime \ast}-k^{^{\prime}\ast}\end{array } \right ) \text { , } \mathcal{m}=\left [ \begin{array } [ c]{cccc}s & 0 & 0 & r\\ 0 & s^{\ast } & r^{\ast } & 0\\ 0 & r & s & 0\\ r^{\ast } & 0 & 0 & s^{\ast}\end{array } \right ] \label{17.23}\ ] ] where @xmath1309 is symplectic satisfying @xmath1310 , @xmath1311{cc}0 & \mathcal{i}\\ -\mathcal{i } & 0 \end{array } \right ] $ ] , @xmath1312 is the @xmath1313 unit matrix . for eq .
( [ 17.21 ] ) being qualified as a new wavelet transform we must prove that it possesses fundamental properties of the usual wavelet transforms , such as the _ admissibility condition , _ the parseval s theorem and the inversion formula . when @xmath1314 is the complex fourier exponential , @xmath1315
according to ( [ 17.21])-([17.22 ] ) we evaluate its eswt @xmath1316 \nonumber \\ & \times \exp \left [ \left ( z+k\right ) \beta^{\ast}-\left ( z^{\ast}+k^{\ast}\right ) \beta+\left ( z^{\prime}+k^{\prime}\right ) \gamma^{\ast } -\left ( z^{\prime \ast}+k^{\prime \ast}\right ) \gamma \right ] .
\label{17.25}\ ] ] making the integration variables transform @xmath1317 @xmath1318 eq .
( [ 17.25 ] ) becomes@xmath1319 , \label{17.26}\ ] ] the last integration is just the complex fourier transform ( cft ) of @xmath1320 denoting it as @xmath1321 we have @xmath1322 then we form the adjoint operation of ( [ 17.27 ] ) , @xmath1323 \nonumber \\ & = \left \vert s\right \vert ^{2}\phi^{\ast}\left ( s^{\ast}\beta^{\ast } + r^{\ast}\gamma,\text { } s^{\ast}\gamma^{\ast}+r^{\ast}\beta \right ) { \displaystyle \iint } \frac{d^{2}kd^{2}k^{\prime}}{\pi^{2}}\exp \left ( k\beta^{\ast}-k^{\ast}\beta+k^{\prime}\gamma^{\ast}-k^{\prime \ast}\gamma \right ) \nonumber \\ & \times \phi \left [ s\left ( z - k\right ) + r\left ( z^{\prime \ast}-k^{\prime \ast}\right ) , s\left ( z^{\prime}-k^{\prime}\right ) + r\left ( z^{\ast}-k^{\ast}\right ) \right ] \nonumber \\ & = \left \vert s\right \vert ^{2}\phi^{\ast}\left ( s^{\ast}\beta^{\ast } + r^{\ast}\gamma,\text { } s^{\ast}\gamma^{\ast}+r^{\ast}\beta \right ) { \displaystyle \iint } \frac{d^{2}kd^{2}k^{\prime}}{\pi^{2}}\phi \left [ -sk - rk^{\prime \ast } , -sk^{\prime}-rk^{\ast}\right ] \nonumber \\ & \times \exp \left [ \left ( k+z\right ) \beta^{\ast}-\left ( k^{\ast}+z^{\ast}\right ) \beta+\left ( k^{\prime}+z^{\prime}\right ) \gamma^{\ast } -\left ( k^{\prime \ast}+z^{\prime \ast}\right ) \gamma \right ] \nonumber \\ & = \left \vert s\right \vert ^{2}\phi^{\ast}\left ( s^{\ast}\beta^{\ast } + r^{\ast}\gamma,\text { } s^{\ast}\gamma^{\ast}+r^{\ast}\beta \right ) \exp \left ( z\beta^{\ast}-z^{\ast}\beta+z^{\prime}\gamma^{\ast}-z^{\prime \ast } \gamma \right ) { \displaystyle \iint } \frac{d^{2}vd^{2}v^{\prime}}{\pi^{2}}\phi \left ( v , v^{\prime}\right ) \nonumber \\ & \times \exp \left [ -v\left ( s^{\ast}\beta^{\ast}+r^{\ast}\gamma \right ) + v^{\ast}\left ( s\beta+r\gamma^{\ast}\right ) -v^{\prime}\left ( s^{\ast } \gamma^{\ast}+r^{\ast}\beta \right ) + v^{\prime \ast}\left ( s\gamma + r\beta^{\ast}\right ) \right ] , \label{17.28}\ ] ] where the integration in the last line is just the cft of @xmath1306 ( comparing with ( [ 17.26 ] ) ) , thus ( [ 17.28 ] ) leads to @xmath1324 from eq .
( [ 17.29 ] ) we have@xmath1325 which together with ( [ 17.24 ] ) lead to @xmath1326 eq .
( [ 17.31 ] ) implies that we should impose the normalization @xmath1327 such that the reproducing process _ _ _ _ exists @xmath1328 ( [ 17.32 ] ) may be named the generalized admissibility condition .
now we can have the corresponding _ parseval theorem _ : for any @xmath1329 and @xmath1330 we have@xmath1331 _ _ proof : _ _ * * * * assuming @xmath1332 and @xmath1333 be cft of @xmath1334 and @xmath1335 , respectively,@xmath1336 recalling the corresponding convolution theorem @xmath1337 so from eqs .
( [ 17.21 ] ) and ( [ 17.35])-([17.36 ] ) we see that @xmath1338 can be considered as a convolution in the form ( noting that the cft of @xmath1339 is @xmath1340 see ( [ 17.26])-([17.27]))@xmath1341 \nonumber \\ & = { \displaystyle \iint } d^{2}\beta d^{2}\gamma f\left ( \beta,\gamma \right ) \phi^{\ast}\left ( s^{\ast}\beta^{\ast}+r^{\ast}\gamma,\text { } s^{\ast}\gamma^{\ast}+r^{\ast } \beta \right ) \exp \left ( k\beta^{\ast}-k^{\ast}\beta+k^{\prime}\gamma^{\ast } -k^{\prime \ast}\gamma \right ) .
\label{17.37}\ ] ] using eq .
( [ 17.37 ] ) we calculate@xmath1342 as a consequence of ( [ 17.32 ] ) and ( [ 17.38 ] ) the further integration yields@xmath1343 which completes the proof . _
inversion formula _ : _ _
_ _ from eq .
( [ 17.34 ] ) we have @xmath1344 that is , there exists an inversion formula for @xmath1345 which represents the original signal @xmath1314 as a superposition of wavelet functions @xmath1346 with the value of entangled wavelet transform @xmath1347 serving as coefficients .
in fact , in eq .
( [ 17.21 ] ) when we take @xmath1348 then @xmath1349 substituting ( [ 17.41])-([17.42 ] ) into ( [ 17.39 ] ) , we obtain ( [ 17.40 ] ) .
we can visualize the eswt in the context of quantum mechanics , letting @xmath1350and using eqs .
( [ 17.22])-([17.23 ] ) , eq .
( [ 17.21 ] ) is expressed as@xmath1351 g\left ( z , z^{\prime}\right ) \nonumber \\ & = s{\displaystyle \iint } \frac{d^{2}zd^{2}z^{\prime}}{\pi^{2}}\left \langle \phi \right . \left \vert \mathcal{m}\left ( \begin{array } [ c]{c}z - k\\ z^{\ast}-k^{\ast}\\ z^{\prime}-k^{\prime}\\ z^{\prime \ast}-k^{\prime \ast}\end{array } \right ) \right \rangle \left \langle z , z^{\prime}\right \vert \left .
g\right \rangle = \left \langle \phi \right \vert f_{2}\left ( r , s;k , k^{\prime } \right ) \left \vert g\right \rangle \label{17.43}\ ] ] where @xmath1352 is defined as@xmath1353{c}f_{2}\left ( r , s;k , k^{\prime}\right ) = s{\displaystyle \iint } \frac{d^{2}zd^{2}z^{\prime}}{\pi^{2}}\left \vert sz+rz^{\prime \ast},sz^{\prime } + rz^{\ast}\right \rangle \left \langle z+k , z^{\prime}+k^{\prime}\right \vert ,
\\ \left \vert sz+rz^{\prime \ast},sz^{\prime}+rz^{\ast}\right \rangle = \left \vert sz+rz^{\prime \ast}\right \rangle _ { 1}\otimes
\left \vert sz^{\prime}+rz^{\ast } \right \rangle _ { 2}. \end{array } \label{17.44}\ ] ] when @xmath1354 and @xmath1355 , @xmath1356 is just the 2-mode fresnel operator .
thus , we have extended the swt of signals in one complex plane to eswt of signals defined in two complex planes , the latter is not the tensor product of two independent swts , this generalization is inevitable , since it resembles the extending from the single - mode squeezing transform ( or fresnel operator ) to the two - mode squeezing transform ( or entangled fresnel operator ) in quantum optics .
next we shall introduce a new kind of wt , i.e. , symplectic - dilation mixed wt @xcite . recalling that in ref .
@xcite we have constructed a new entangled - coherent state representation ( ecsr ) @xmath1357,@xmath1358
\left \vert 00\right \rangle , \label{17.45}\ ] ] which is the common eigenvector of the operator @xmath1359 and @xmath1360 i.e. , @xmath1361 and @xmath1362 where @xmath1363 is the coordinate operator , ( @xmath1364 .
@xmath1365 is complete,@xmath1366 and exhibits partly non - orthogonal property ( for @xmath1367 and orthonormal property ( for @xmath1368 @xmath1369 \delta \left ( x^{\prime}-x\right ) , \label{17.47}\ ] ] so @xmath29 possess behavior of both the coherent state and the entangled state .
an interesting question is : can we introduce a new kind of continuous wt for which the @xmath29 representation underlies ?
the answer is affirmative .
our motivation of this issue comes from the mixed lens - fresnel transform in classical optics @xcite ( see ( [ 44 ] ) below ) . by synthesizing ( [ 15.3 ] ) and ( [ 17.2 ] ) and in reference to ( [ 17.46 ] ) we propose the mixed wt for @xmath1370 ( @xmath1371 ) : @xmath1372 where @xmath1373 the family of mother wavelet @xmath899 involves both the the symplectic transform of @xmath40 and the dilation - transform of @xmath900,@xmath1374 .
\label{17.49}\ ] ] letting @xmath1375 @xmath1376 then ( [ 17.48 ] ) can be expressed as quantum mechanical version@xmath1377 where @xmath1378 is defined as@xmath1379 @xmath1380 is just the generalized squeezing operator , which causes a lens - fresnel mixed transform . for eq .
( [ 17.48 ] ) being qualified as a new wt we must prove that it possesses fundamental properties of the usual wts , such as the admissibility condition , the parseval theorem and the inversion formula .
it is straightforward to evaluate the transform ( [ 17.48 ] ) and its reciprocal transform when @xmath1370 is the exponential @xmath1381@xmath1382 making the integration variables transform @xmath1383 leading to @xmath1384 and @xmath1385 , ( [ 17.52 ] ) becomes@xmath1386 where @xmath1387 is just the fourier transform of @xmath1388@xmath1389 then we perform the adjoint wt of ( [ 17.48 ] ) , using ( [ 17.49 ] ) and ( [ 17.53 ] ) we see@xmath1390 \nonumber \\ & = \left \vert s\right \vert \left \vert \mathrm{a}\right \vert g_{1}\left ( \alpha , x\right ) \phi^{\ast}\left ( s^{\ast}\beta^{\ast}-r^{\ast}\beta,\text { } \mathrm{a}p\right ) \int_{-\infty}^{\infty}\frac{\mathtt{d}b^{\prime}}{\sqrt{\pi}}\nonumber \\ & \times \int \frac{\mathtt{d}^{2}\kappa^{\prime}}{2\pi}e^{\kappa^{\prime}\beta^{\ast}-\kappa^{\prime \ast}\beta+\mathtt{i}\mathrm{a}pb^{\prime}}\psi \left ( s\kappa^{\prime}-r\kappa^{\prime \ast},b^{\prime}\right ) \nonumber \\ & = \left \vert s\right \vert \left \vert \mathrm{a}\right \vert g_{1}\left ( \alpha , x\right ) \left \vert \phi \left ( s^{\ast}\beta^{\ast}-r^{\ast}\beta,\text { } \mathrm{a}p\right ) \right \vert ^{2}. \label{17.55}\ ] ] from eq .
( [ 17.55 ] ) we obtain@xmath1391 which leads to @xmath1392 eq .
( [ 17.57 ] ) implies that we should impose the normalization @xmath1393 such that the reproducing process _ _ _ _ exists @xmath1394 ( [ 17.58 ] ) may be named the generalized _
admissibility condition_. now we can have the corresponding _ parseval theorem _ : for any @xmath1329 and @xmath1330 we have@xmath1395 _ _
proof : _ _ * * * * assuming @xmath1396 and @xmath1397 be the fourier transforms of @xmath1398 and @xmath1399 , respectively,@xmath1400 in order to prove ( [ 17.60 ] ) , we first calculate @xmath1401 . in similar to deriving eq.([19 ] ) , using ( [ 14 ] ) , ( [ 15 ] ) and the inversion formula of ( [ 17.61 ] ) we have@xmath1402 \nonumber \\ & = \sqrt{s\left \vert \mathrm{a}\right \vert } \int_{-\infty}^{\infty}\frac{\mathtt{d}p}{\sqrt{2\pi}}\int \frac{\mathtt{d}^{2}\beta}{\pi}f\left ( \beta , p\right ) \nonumber \\ & \times \phi^{\ast}\left ( s^{\ast}\beta^{\ast}-r^{\ast}\beta,\text { } \mathrm{a}p\right ) e^{\kappa^{\ast}\beta-\kappa \beta^{\ast}-\mathtt{i}pb}. \label{17.62}\ ] ] it then follows @xmath1403 substituting ( [ 17.63 ] ) into the left - hand side ( lhs ) of ( [ 17.60 ] ) and using ( [ 17.58 ] ) we see@xmath1404 thus we complete the proof of eq.([17.60 ] ) .
_ inversion formula _ : _ _ _ _ from eq .
( [ 17.60 ] ) we have @xmath1405 that is the inversion formula for the original signal @xmath1406 expressed by a superposition of wavelet functions @xmath1407 with the value of continuous wt @xmath1408 serving as coefficients .
in fact , in eq .
( [ 17.48 ] ) when we take @xmath1409 then@xmath1410 substituting ( [ 17.66 ] ) into ( [ 17.64 ] ) yields ( [ 17.65 ] ) .
we can visualize the new wt @xmath1411 in the context of quantum optics .
noticing that the generalized squeezing operator @xmath1380 in ( [ 17.51 ] ) is an image of the combined mapping of the classical real dilation transform @xmath1412 @xmath1413 ( @xmath1414 ) and the classical complex symplectic transform @xmath1415 in @xmath29 representation , one can use the technique of integration within normal product of operators to perform the integration in ( [ 17.51 ] ) to derive its explicit form ( see eq .
( 15 ) in ref .
the transform matrix element of @xmath1416 in the entangled state representation @xmath9 is@xmath1417 in fock space @xmath124 is two - mode epr entangled state in ( [ 3.11 ] ) . then using ( [ 17.45 ] ) and ( [ 3.11 ] )
, we obtain @xmath1418 .
\label{17.68}\ ] ] substituting ( [ 17.68 ] ) into ( [ 17.67 ] ) and using ( [ 5.2])@xmath10 we obtain@xmath1419 .
\label{17.69}\ ] ] which is just the kernel of a mixed lens@xmath1420fresnel transform , i.e. , the variable @xmath1421 of the object experiences a generalized fresnel transform , while @xmath1422 undergoes a lens transformation .
thus , based on @xmath1365 we have introduced sdwt which involves both the real variable dilation - transform and complex variable symplectic transform , corresponding to the lens - fresnel mixed transform in classical optics .
in the theoretical study of quantum computer , of great importance is the hadamard transform .
this operation is @xmath292 hadamard gates acting in parallel on @xmath292 qubits .
the hadamard transform produces an equal superposition of all computational basis states . from the point of view of deutsch - jozsa quantum algorithm ,
the hadamard transform is an example of the @xmath1423 quantum fourier transform , which can be expressed as @xcite@xmath1424 now the continuous hadamard transform , used to go from the coordinate basis @xmath600 to the momentum basis@xmath10 is defined as@xcite @xmath1425 where @xmath1426 is the scale length .
@xmath1427 is named hadamard operator . using the completeness of @xmath1428 we have@xmath1429 the above two transforms ( fresnel transform and hadmard transform ) are independent of each other
, an interesting question thus naturally arises : can we combine the two transforms together ? to put it in another way , can we construct a combinatorial operator which play the role of both fresnel transform and hadmard transform for two independent optical modes ?
the answer is affirmative , in this section we try to construct so - called fresnel - hadmard combinatorial transform .
based on the coherent - entangled representation @xmath1357 , and enlightened by eq .
( [ 5.4 ] ) and ( [ j18.2 ] ) we now construct the following ket - bra integration @xcite@xmath1430 we name @xmath1431 the hadamard - fresnel combinatorial operator . substituting eq.([j18.2 ] ) into eq.([j18.4 ] ) , and using the two - mode vacuum projector s normally ordered form @xmath1432 \colon$ ] as well as the iwop technique we get@xmath1433 where@xmath1434@xmath1435 , \ ] ] and @xmath1436 , \ ] ] they are all within the normal ordering symbol @xmath1437 @xmath720 now performing the integration over @xmath1438 within @xmath1437 and remembering that all creation operators are commute with all annihilation operators ( the essence of the iwop technique ) so that they can be considered c - number during the integration , we can finally obtain@xmath1439 which is the normally ordered form of hadamard - fresnel combinatorial operator .
note @xmath1440 = 0 , \label{j18.6}\ ] ] and @xmath1441 = 1,\text { \ } \left [ \frac{a_{1}+a_{2}}{\sqrt{2}},\frac { a_{1}^{\dag}+a_{2}^{\dag}}{\sqrt{2}}\right ] = 1 , \label{j18.7}\ ] ] @xmath30 can be considered a mode independent of another mode @xmath1442 thus we have the operator identity @xmath1443 = \colon \exp[\frac{1}{2}\left ( e^{2f}-1\right ) \left ( a_{1}^{\dag}\pm a_{2}^{\dag}\right ) \left ( a_{1}\pm a_{2}\right ) ] \colon .
\label{j18.8}\ ] ] using ( [ j18.8 ] ) we can rewrite eq.([j18.7 ] ) as@xmath1444 where@xmath1445 \nonumber \\ & \exp \left [ \frac{a_{1}^{\dag}+a_{2}^{\dag}}{\sqrt{2}}\frac{a_{1}+a_{2}}{\sqrt{2}}\ln \frac{4i\sigma^{2}}{\left ( \sigma^{4}+4\right ) } \right ] \exp \left [ \frac{\sigma^{4}-4}{2\left ( \sigma^{4}+4\right ) } \left ( \frac{a_{1}+a_{2}}{\sqrt{2}}\right ) ^{2}\right ] \label{j18.10}\ ] ] and@xmath1446 \exp \left [ \left ( \frac { a_{1}^{\dag}-a_{2}^{\dag}}{\sqrt{2}}\frac{a_{1}-a_{2}}{\sqrt{2}}+\frac{1}{2}\right ) \ln \frac{1}{s^{\ast}}\right ] \nonumber \\ & \exp \left [ \frac{r^{\ast}}{2s^{\ast}}\left ( \frac{a_{1}-a_{2}}{\sqrt{2}}\right ) ^{2}\right ] , \label{j18.11}\ ] ] while @xmath1447 is the fresnel operator for mode @xmath1448 @xmath1449 is named the hadamard operator for mode @xmath1450 it then follows @xmath1451 from which we see the hadamard - fresnel combinatorial operator can play the role of fresnel transformation for @xmath1452 physically , @xmath30 and @xmath1453 can be two output fields of a beamsplitter . in a similar way , we have @xmath1454 , \nonumber \\
u\frac{a_{1}^{\dag}+a_{2}^{\dag}}{\sqrt{2}}u^{-1 } & = u_{1}\frac{a_{1}^{\dag } + a_{2}^{\dag}}{\sqrt{2}}u_{1}^{-1}=\frac{1}{4i\sigma^{2}}\left [ -\left ( \sigma^{4}+4\right ) \frac{a_{1}^{\dag}+a_{2}^{\dag}}{\sqrt{2}}+\left ( \sigma^{4}-4\right ) \frac{a_{1}+a_{2}}{\sqrt{2}}\right ] \label{j18.13}\ ] ] which for the quadrature @xmath1455 @xmath1456 @xmath1457 leads to @xmath1458 from which we see that the hadamard - fresnel combinatorial operator also plays the role of exchanging the total momentum average position followed by a squeezing transform , with the squeezing parameter being @xmath1459 the mutual transform in ( [ j18.14 ] ) can be realized by @xmath1460 while the two - mode squeezing operator is @xmath1461 , $ ] therefore@xmath1462 from eq.([j18.9 ] ) and eq.([j18.17 ] ) , we see that the hadamard - fresnel combinatorial operator can be decomposed as @xmath1463 it can be also seen that @xmath1431 is unitary , @xmath1464 . in this section , we have introduced the fresnel - hadamard combinatorial operator by virtue of the iwop technique . this unitary
operator plays the role of both fresnel transformation for mode @xmath30 and hadamard transformation for mode @xmath31 respectively , and the two transformations are combinatorial .
we have shown that the two transformations are concisely expressed in the coherent - entangled state representation as a projective operator in integration form .
we also found that the fresnel - hadamard operator can be decomposed as @xmath1465 a fresnel operator @xmath1447 , a two - mode squeezing operator @xmath1466 and the total momentum - average position exchanging operator .
physically , @xmath1467 and @xmath1453 can be two output fields of a beamsplitter .
if an optical device can be designed for fresnel - hadamard combinatorial transform , then it can be directly applied to these two output fields of the beamsplitter . in summary , although quantum optics and classical optics are so different , no matter in the mathematical tools they employed or in a conceptual view ( quantum optics concerning the wave - particle duality of optical field with an emphasis on its nonclassical properties , whereas classical optics works on the distribution aqnd propagation of the light waves ) , that it is a new exploration to link them systematically .
however , in this review , via the route of developing dirac s symbolic method we have revealed some links between them by mapping classical symplectic transformation in the coherent state representation onto quantum unitary operators ( gfo ) , throughout our discussion the iwop technique is indispensable for the derivation .
we have resorted to the quantum optical interpretation of various classical optical transformations by adopting quantum optics concepts such as the coherent states , squeezed states , and entangled states , etc .
remarkably , we have endowed complex fractional fourier transform , hankel transform with quantum optical representation - transform interpretation .
our formalism , starting from quantum optics theory , not only provides quantum mechanical account of various classical optical transformations , but also have found their way back to some new classical transformations , e.g. entangled fresnel transform , fresnel - wavelet transform , etc , which may have realistic optical interpretation in the future .
as dirac predicted , functions that have been applied in classical optical problems may be translated in an operator language in quantum mechanics , and vice - versa .
we expect that the content of this work may play some role in quantum states engineering , i.e. , optical field states preparation and design .
once the correspondence in this respect between the two distinct fields is established , the power of dirac s symbolic method can be fully displayed to solve some new problems in classical optics , e.g. , to find new eigen - modes of some optical transforms ; to extend the research region of classical optics theoretically by introducing new transforms ( for example , the entangled fresnel transforms ) , which may bring attention of experimentalists , who may get new ideas to implement these new classical optical transformations .
* acknowledgement : * this work supported by the national natural science foundation of china , grant no .
10775097 and 10874174 , and a grant from the key programs foundation of ministry of education of china ( no .
210115 ) , and the research foundation of the education department of jiangxi province of china ( grant no .
gjj10097 ) .
j. r. klauder , annals phys .
( new york ) * 11 * ( 1960 ) 123 ; j. r. klauder and b .- s .
skagerstam , coherent states , world scientific , singapore ( 1985 ) ; j. r. klauder , e. c. g. sudarshan , _ fundamentals of quantum optic _ w. a. benjamin , new york , 1968 l. mandel and e. wolf , optical coherence and quantum optics , cambridge , 1996 ; m. o. scully and m. s. zubairy , quantum optics , cambridge university press , 1997 ; w. p. schleich , quantum optics in phase space , wiley - vch , 2001 ; d. f. walls and g. j. milburn , quantum optics , springer - verlag , 1994 hong - yi fan , h. r. zaidi and j. r. klauder , phys . rev .
d 35 ( 1987 ) 1831 ; fan hongyi and h. r. zaidi , phys .
a 37 ( 1988 ) 2985 ; hongyi fan and j. r. klauder , j. phys . a 21 ( 1988 ) l715 ; hong - yi fan and hui zou , phys .
lett . a 252 ( 1999 ) 281 ; hong - yi fan , _ representation and transformation theory in quantum mechanics _
( in chinese ) , shanghai scientific and technical publishers , shanghai , 1997 ; hong - yi fan , entangled state representations in quantum mechanics and their applications , shanghai jiao tong university press , 2001 p.a . belanger . beam propagation and the abcd ray matrix .
( 1991 ) 196 ; v. magni , g. cerullo , s.d .
abcd matrix analysis of propagation of gaussian beams through kerr media .
96 ( 1993 ) 348 h. y. fan , _ representation and transformation theory in quantum mechanics _
( in chinese ) , shanghai scientific and technical publishers , shanghai 1997 ; entangled state representations in quantum mechanics and their applications , shanghai jiao tong university press , 2001 h.y .
fan , h.r .
zaidi , and j.r .
klauder , _ phys .
* 35 * ( 1987 ) 1831 ; h.y .
fan and y. fan , _ phys .
a _ * 54 * ( 1996 ) 958 ; h.y .
fan , _ commun .
* 31 * ( 1999 ) 285 ; h.y .
fan and y. fan , c__ommun .
_ _ 33 ( 2000 ) 701 ; h.y .
fan and l.s .
li , _ commun .
* 29 * ( 1998 ) 477 .
m.a.nielsen and i.l.chuang , _ the quantum computation and quantum information _ , cambridge university press , cambridge ( 2000 ) ; see e.g. , j.preskill , _ quantum information and computation _ , california institute of technology ( 1998 ) . | by virtue of the newly developed technique of integration within an ordered product ( iwop ) of operators , we explore quantum optical version of classical optical transformations such as optical fresnel transform , hankel transform , fractional fourier transform , wigner transform , wavelet transform and fresnel - hadmard combinatorial transform etc . in this way one may gain benefit for developing classical optics theory from the research in quantum optics , or vice - versa .
we can not only find some new quantum mechanical unitary operators which correspond to the known optical transformations , deriving a new theorem for calculating quantum tomogram of density operators , but also can reveal some new classical optical transformations . for examples , we find the generalized fresnel operator ( gfo ) to correspond to the generalized fresnel transform ( gft ) in classical optics .
we derive gfo s normal product form and its canonical coherent state representation and find that gfo is the loyal representation of symplectic group multiplication rule .
we show that gft is just the transformation matrix element of gfo in the coordinate representation such that two successive gfts is still a gft .
the abcd rule of the gaussian beam propagation is directly demonstrated in the context of quantum optics .
especially , the introduction of quantum mechanical entangled state representations opens up a new area to finding new classical optical transformations .
the complex wavelet transform and the condition of mother wavelet are studied in the context of quantum optics too . throughout our discussions , the coherent state , the entangled state representation of the two - mode squeezing operators and the technique of integration within an ordered product ( iwop ) of operators
are fully used .
all these confirms dirac s assertion : @xmath0for a quantum dynamic system that has a classical analogue , unitary transformation in the quantum theory is the analogue of contact transformation in the classical theory " .
keywords : dirac s symbolic method ; iwop technique ; entangled state of continuum variables ; entangled fresnel transform ; collins formula ; generalized fresnel operator ; complex wavelet transform ; complex wigner transform ; complex fractional fourier transform ; symplectic wavelet transform ; entangled symplectic wavelet transform ; symplectic - dilation mixed wavelet transform ; fractional radon transform ; new eigenmodes of fractional fourier transform |
cavity qed ( cqed ) investigates the interaction of atoms with confined electromagnetic field modes .
when atoms are coupled to a high - finesse optical resonator , the usual free - space dipole force they experience is strongly enhanced and , at the same time , the back - action of atoms on the confined light field can not be ignored any longer . as a consequence both atoms and light must be treated on the same footing and the dynamics for the atomic motion and the cavity field becomes strongly nonlinear .
these hybrid systems open the way to new physical situations , in particular when cold atoms , bosonic or fermionic , are loaded inside optical cavities @xcite .
furthermore , it is possible to probe the properties of these systems in a non - destructive way by using the field leaking outside the cavity @xcite . trapping bose - einstein condensates ( bec ) in laser - driven high - finesse optical cavities @xcite has been realized experimentally recently @xcite .
the strong atom - cavity coupling enhances nonlinear effects and bistable behavior @xcite and even chaos @xcite set in . in the dispersive regime
where the pump and cavity fields are both far - detuned from the atomic transition , the light fields impart forces on the atoms which thus move . in turn , the light fields pick up phase shifts induced by the refractive index of this moving atomic dielectric medium .
this alters the light forces , thus the atomic motion , thus the accumulated dispersive phase shifts , and this combined atom - field process loops self - consistently . as a result , when the pump field strength is larger than some critical value , the atomic cloud scatters constructively the pump photons into the cavity modes .
this causes an abrupt increase of the number of photons inside the cavity , a phenomenon referred to as superradiance , and the atoms achieve self - organization into the effective optical lattice created by the coherent superposition of the pump and cavity fields @xcite .
self - organization breaks the initial translation symmetry .
in other words , photon scattering couples the initial zero - momentum state of the atomic cloud to a superposition of higher recoil momentum states @xcite . this effect has been used to simulate the dicke superradiance quantum phase transition @xcite in a bec - cavity system where two collective motional modes of the condensate play the role of the two hyperfine spin states of the original dicke model @xcite .
the non - equilibrium dynamics of such systems have been studied in @xcite .
recent theoretical proposals have generalized the model even further by introducing cavity - assisted raman coupling in order to reach phases such as self - organized magnetic lattices of bosons @xcite or topologically non - trivial phases of fermions @xcite . in this article
we consider a two - dimensional system similar to the experiment in @xcite but with two crossing cavities in order to examine the possible formation of self - organized triangular and honeycomb lattices of cold bosons as a result of superradiance . lately
, the honeycomb lattice has attracted a lot of attention in the cold atom community because of its unique band structure mimicking weyl - dirac quasi - particles at the so - called dirac points @xcite . by numerically simulating the real - time dynamics of the cavity - atoms system
we show that , depending on the relative phase @xmath0 of the cavity fields with respect to the pump field , three distinct atomic lattice structures are possible when non - interacting atoms and identical cavities are considered . the first one is a triangular lattice , the cavity fields being in phase with the pump laser ( @xmath1 ) .
the second one is a honeycomb lattice with density - balanced sites , the cavities and pump fields being in quadrature ( @xmath2 ) .
the last one is a honeycomb lattice with density - imbalanced sites , the relative phase @xmath0 being in between zero and @xmath3 .
we address the long - time stability of these atomic lattices as well as the possibility of driving the atoms from one structure to another .
the rest of this paper is organized as follows . in sec .
[ model ] we introduce the experimental setup that we consider and give the relevant hamiltonian in the dispersive regime ( sec .
[ hamiltonian ] ) .
the dynamical equations of the coupled atoms - cavities system are derived in sec .
[ equations ] .
they give rise to an effective lattice potential for the center of mass motion of the atoms which is studied in details in sec .
[ potential ] , [ symmetry ] and [ idcav ] . using the symmetry properties of the effective lattice potential
, we derive the atomic equations of motion in reciprocal space in sec .
[ dma ] . the condition for the normal to superradiance phase transition
is obtained by studying the linear response of the system in sec .
[ instability ] . finally , in sec .
[ numerics ] , we present our numerical results about the atoms and cavity fields dynamics , the superradiance phase transition ( sec .
[ ntos ] ) , the long - time stability of the different atomic lattices which are obtained ( sec .
[ longtime ] ) as well as the possibility of switching the system between these different lattice structures ( sec . [ switch ] ) .
we summarize the work in sec .
about axis @xmath4 .
the atoms are then pumped by a retro - reflected classical laser field ( @xmath4 direction ) and release photons into the modes of the cavities . in return , the coherent superposition of the pump and cavity fields creates an effective periodic potential and atoms self - organize themselves into the corresponding potential wells .
a triangular or honeycomb lattice structure of atoms thus emerges as a result of superradiance processes .
[ fig : system],scaledwidth=30.0% ] to study the self - organization of bosonic atoms into a triangular or a honeycomb lattice as a result of superradiance , we consider two high - finesse optical cavities with frequencies @xmath5 ( @xmath6 ) located in the @xmath7-plane and oriented symmetrically about the @xmath8-axis by an angle @xmath9 ( see fig . [ fig : system ] ) .
a two - dimensional ( 2d ) dilute cloud of non - interacting two - level ultracold bosonic atoms ( resonance frequency @xmath10 , excited - state lifetime @xmath11 ) is then loaded inside these cavities , which we assume do not contain any photons initially .
a monochromatic standing - wave at frequency @xmath12 ( obtained by retro - reflecting a classical laser field propagating along unit vector @xmath13 ) is used to pump the atoms . in such a system ,
the atoms scatter the pump field photons into the cavity modes and self - organize in the potential wells of the effective potential created by the standing - wave and cavity fields . in the following ,
we assume that spontaneous emission processes are fully negligible so that the atomic dynamics is hamiltonian .
this is achieved when both the pump and cavity fields are far detuned from the atomic resonance frequency , @xmath14 and @xmath15 .
we further assume that the rotating - wave approximation is valid such that fast variables can be eliminated .
this is the case when @xmath16 and @xmath17 .
finally , we assume that one can adiabatically eliminate the atomic excited state amplitude so that atoms , initially prepared in their ground state , mostly evolve in their ground state .
this is the case when the rabi oscillation has a small amplitude , which happens when the pump detuning @xmath18 is much larger in magnitude than the kinetic energy of the atoms in their excited state , the pump and cavity rabi frequencies , and the pump - cavity detuning @xmath19 . under all these assumptions , and since the pump is described by a classical field , the hamiltonian describing the coupled dynamics between the cavity fields and the 2d motion of atoms , in their internal ground state and at center - of - mass position @xmath20 , reads : @xmath21 where @xmath22 is the pump rabi frequency , @xmath23 the atom - cavity coupling strength , and @xmath24 and @xmath25 the atomic and cavity bosonic annihilation operators .
the atomic operator is normalized to the total number of atoms @xmath26 , @xmath27 . with a suitable choice of the origin of coordinates ,
the pump and cavity mode functions can be written as @xmath28 where @xmath0 is a controllable phase that can be changed , for example , by moving the pump mirror along @xmath4 .
the pump and cavity wave vectors are @xmath29 and @xmath30 respectively , with the unit - length vectors @xmath31 the pump and cavity wavelengths are @xmath32 and @xmath33 , @xmath34 being the speed of light . ,
eq . , and its bravais vectors @xmath35 and @xmath36 , eq . .
right panel : the corresponding triangular reciprocal lattice spanned by vectors @xmath37 and @xmath38 , eq . .
the initial zero - momentum state of the atomic cloud ( black disk at the center ) is directly coupled to eighteen points of the reciprocal lattice ( other colorful disks ) by schrdinger s equation eq . .
disks with the same color refer to a given coupling term in schrdinger s equation eq . .
pump - cavity coupling : @xmath39 ( purple ) and @xmath40 ( blue ) terms .
inter - cavity coupling : @xmath41 term ( red ) .
intra - cavity coupling : @xmath42 ( green ) and @xmath43 ( brown ) terms .
pump coupling : @xmath44 term ( orange ) .
[ fig : points],title="fig:",scaledwidth=22.0% ] , eq . , and its bravais vectors @xmath35 and @xmath36 , eq . .
right panel : the corresponding triangular reciprocal lattice spanned by vectors @xmath37 and @xmath38 , eq . .
the initial zero - momentum state of the atomic cloud ( black disk at the center ) is directly coupled to eighteen points of the reciprocal lattice ( other colorful disks ) by schrdinger s equation eq . .
disks with the same color refer to a given coupling term in schrdinger s equation eq . .
pump - cavity coupling : @xmath39 ( purple ) and @xmath40 ( blue ) terms .
inter - cavity coupling : @xmath41 term ( red ) .
intra - cavity coupling : @xmath42 ( green ) and @xmath43 ( brown ) terms .
pump coupling : @xmath44 term ( orange ) .
[ fig : points],title="fig:",scaledwidth=20.0% ] in the rest of this paper , we consider the mean - field regime where the field operators are replaced by their mean values , @xmath45 and @xmath46 , and quantum fluctuations are discarded .
we next define the pump recoil energy @xmath47 .
using @xmath48 , @xmath49 and @xmath50 as space , time and energy units , the dimensionless schrdinger equation for the atomic wavefunction reads : @xmath51\varphi(\vec{\rho},\tau)\end{aligned}\ ] ] where @xmath52 , @xmath53 , @xmath54 , and @xmath55 .
note that the normalization condition for the reduced atomic wavefunction is unchanged , @xmath56 .
the various dimensionless coupling constants appearing in eq .
are : @xmath57 introducing the dimensionless cavity decay constants @xmath58 , the equations of motion for the cavity field amplitudes read : @xmath59 where @xmath60 is the shifted cavity resonance frequency and where : @xmath61 note that , because of the normalization of the atomic wave function , the @xmath62-integrals do not depend on the actual number of atoms @xmath26 . from eq .
( [ eq : psidot ] ) , one immediately sees that the pump and cavity fields create an effective potential for the atom center - of - mass dynamics : @xmath63 this effective potential can be recast under the simpler and suggestive form @xmath64 featuring the complex rabi field amplitude @xmath65 associated , in the mean - field regime , to the coherent superposition of the pump and cavity classical standing - wave electrical field amplitudes .
as one can see , the pump and cavity fields are balanced when @xmath66 ( @xmath67 .
we will restrict our subsequent analysis to a red - detuned pump field , @xmath68 , for which the minima of the effective potential , eqs .
- , are found at the maxima of the norm of the rabi field , eq . , and will thus be the deepest .
let us define the two vectors @xmath69 ( @xmath6 ) with @xmath70 and @xmath71 .
as readily checked , they satisfy the property @xmath72 . as a consequence , and by construction , the cavity mode functions @xmath73 are left invariant when @xmath74 is translated by any integer combination of the @xmath75 .
the pump mode function @xmath76 is also left invariant provided @xmath77 is a multiple integer of @xmath78 .
the simplest choice , which also fulfills the conditions leading to eq .
, is @xmath79 .
if this condition is not strictly obeyed , @xmath76 will still be approximately invariant , to a very good accuracy , if @xmath80 , where @xmath81 is the size of the atomic cloud . in the rest of the paper
we will assume this condition to hold such that the approximation @xmath82 is perfectly justified .
in turn , it means that the cavity and pump fields have almost the same frequencies , @xmath83 , the cavity fields are also red - detuned from the atomic transition . ] . within these assumptions
, it follows that the bravais lattice associated to the effective potential @xmath84 is @xmath85 , see fig .
[ fig : points ] .
it is triangular and its unit cell is @xmath86 with dimensionless area @xmath87 .
the corresponding reciprocal lattice @xmath88 is also triangular and its first brillouin zone is @xmath89 with dimensionless area @xmath90 . in the following
, we will examine the simplest case where the cavities have exactly the same characteristics , @xmath91 , @xmath92 and @xmath93 . then a reflection symmetry about axis @xmath4
( @xmath94 ) amounts to exchange the cavity fields , @xmath95 .
if we further assume that the cavities are initially empty and the initial atomic cloud is also symmetric with respect to @xmath4 , then the subsequent time evolution will always enforce @xmath96 at all times where @xmath97 and @xmath98 are not perfectly equal . ] . in this case the rabi field , eq .
, writes : @xmath99 with @xmath100 .
it is immediately seen that , in this case , the effective potential is always reflection - symmetric about planes located at @xmath101 ( @xmath102 ) and in particular about axis @xmath4 ( @xmath103 ) .
it is also easy to prove that the potential for @xmath104 ( @xmath105 ) is obtained by a mere translation of the potential for @xmath0 .
furthermore , the potential for @xmath106 is simply obtained by reflecting the potential for @xmath0 about axis @xmath107 ( @xmath108 ) .
this means that , for all practical purposes , the range of cavity phases can be restricted to @xmath109 . for @xmath110 , the bravais unit cell hosts two energy - balanced minima which can be labeled @xmath111 and @xmath112 , so that the full lattice of minima is now a honeycomb lattice made of two shifted @xmath111 and @xmath112 triangular lattices , see fig .
[ fig : veff ] . interestingly enough ,
this honeycomb structure does not depend on the relative weight between the pump and cavity rabi amplitudes .
indeed , when @xmath113 , the potential is symmetric under @xmath114 .
this immediately ensures the existence of an even number of energy - balanced minima in the bravais unit cell , here two .
when @xmath0 is continuously decreased from @xmath3 , the @xmath111 and @xmath112 minima become energy - imbalanced , still retaining a nice - looking honeycomb structure when @xmath0 is not too small .
this situation is interesting for producing bands with non - vanishing chern numbers .
when @xmath0 gets closer to zero , the honeycomb structure is lost for all practical purposes and the lattice of deepest minima is triangular .
irrespective of the value of @xmath0 , and for a priori different mean - field values
@xmath115 , it is worth mentioning that the effective potential gets shifted by @xmath116 when the sign of @xmath117 is flipped , by @xmath118 when the sign of @xmath119 is flipped and by @xmath120 ( which corresponds to a shift of @xmath121 along axis @xmath4 ) when both signs are flipped .
it is straightforward to check that eqs and indeed remain invariant when the sign of @xmath119 and/or @xmath117 is flipped and the corresponding translation by @xmath122 is implemented on the mode functions and atomic wave function @xmath123
. therefore , for each set of coupling parameters , there are four possible solutions associated to given numbers of cavity photons ( @xmath124 ) , all related by translations along @xmath116 and/or @xmath118 .
depending on the initial conditions , the system may select any one of these solutions .
+ +
we now rewrite schrdinger s equation , eq . , by expanding the atomic wave function @xmath123 in reciprocal space :
@xmath125 where @xmath126 is the number of atoms in the unit bravais cell @xmath127 .
the normalization conditions now read : @xmath128 it is straightforward to see that @xmath129 is actually conserved during the evolution .
indeed , the effective potential @xmath84 being periodic under bravais translations , it can not scatter and change @xmath129 . substituting eq . into eq . and using eq .
, we obtain the following dynamical equations for the coefficients @xmath130 : @xmath131 where the @xmath129-dependence of the @xmath132-coefficients is the same on both sides of the equation . the dynamical equations for the cavity fields @xmath115 , eq . ,
remain the same but with the integrals in eq . now reading : @xmath133 generally , in superradiance and self - organization problems ,
the atomic system is initially homogeneous ( zero - momentum state ) , meaning that only the @xmath132-coefficient with @xmath134 and @xmath135 is nonzero initially ( and equal to 1 because of the normalization condition ) . in the course of time , as a result of superradiance , higher - momentum states will be occupied . for our system , the dynamical equations ( [ eq : cdot ] ) couple @xmath136 to the eighteen points of the reciprocal lattice @xmath137 shown in fig .
[ fig : points ] .
each of these points will be in turn coupled to other points in the reciprocal lattice with even higher momenta . in the following sections , we restrict our calculations to momenta transfers less than @xmath138 , i.e. to reciprocal lattice points @xmath139 with @xmath140 and @xmath141 .
these thirty - seven points in total include the initial zero - momentum state , the eighteen states to which this state is directly coupled , and the eighteen other states to which the states with momentum @xmath142 ( the points closest to the initial zero - momentum state ) are directly coupled .
in this section , we study the conditions for superradiance to take place and focus on the pump strength needed to destabilize the initial state for given cavity parameters . for this purpose
we consider the linear response of the system to perturbations added to the initial cavity fields and atomic wave function .
we thus substitute @xmath143 and @xmath144 into the dynamical equations eqs and and in the integrals given in eq . .
our initial state is @xmath145 ( empty cavities ) and @xmath146 ( homogeneous cloud ) . for this initial state ,
@xmath147 and @xmath148 . since @xmath129 can not change under evolution , as explained in sec .
[ dma ] , its value remains zero and will thus be omitted .
the linearized perturbed equations then read : @xmath149 with @xmath150 and @xmath151 eqs and only involve reciprocal lattice vectors of the form @xmath152 with a transfer lattice vector @xmath153 satisfying @xmath154 and @xmath155 .
then it can be seen that all phase factors can be gauged away : @xmath156 where @xmath157 , with @xmath105 , @xmath158 and @xmath159 the ( positive or negative ) remainder with smallest absolute value .
in other words , the stability of our initial state is independent from the cavity phase which can thus be set to @xmath160 in the stability analysis .
let us define the column vector @xmath161 , where the superscript @xmath162 denotes transposition .
the linearized equations eqs and can then be rewritten in compact form @xmath163 where the stability matrix @xmath164 controls the dynamics of the ( doubled ) perturbation column vector @xmath165 around our initial condition .
as one can see , the perturbation will grow , and the initial system is unstable , as soon as the eigenvalues @xmath166 of @xmath164 have a positive imaginary part .
the matrix @xmath164 has the following 4-block structure @xmath167 where the star stands for complex conjugation and where the sub - matrix @xmath168 is symmetric @xmath169 .
it is easy to show that @xmath170 , where the usual entries of the pauli matrix @xmath171 have been replaced by the null and identity matrices . as a consequence , @xmath166 and @xmath172
are both eigenvalues of @xmath164 , with eigenstates @xmath173 and @xmath174 , and the real parts must come in opposite pairs .
[ fig : omega ] shows the real and imaginary parts of the eigenvalues of the stability matrix @xmath164 as functions of the pump coupling parameter @xmath44 and obtained for fixed typical cavity experimental parameters . note that we have not reproduced in these plots eigenvalues with large imaginary parts ( equal to the cavity damping rates @xmath58 ) .
as expected from the symmetry of @xmath164 , one can see that the real parts come indeed in opposite pairs .
however , contrary to what a quick and misleading glance at the plots may let think , imaginary parts should not come in opposite pairs .
indeed , the sub - matrix @xmath175 becomes hermitian , @xmath176 , only when the cavities are lossless which implies the additional symmetry @xmath177 . in this particular case ,
both @xmath166 and @xmath178 are eigenvalues and the imaginary parts would also come in opposite pairs .
this is not the case here since the cavities are lossy ( @xmath179 ) . and
indeed , a careful check shows that seemingly opposite values of the imaginary parts are in fact slightly different .
we have checked that this difference approaches zero when the cavity damping rates @xmath58 go to zero and the cavities become lossless . another important point to mention is that all the imaginary parts ( positive or negative ) are either two- or four - fold degenerate .
the two - fold degeneracy is always present and comes from the fact that @xmath166 and @xmath172 are both eigenvalues .
the occasional four - fold degeneracy comes from an additional spatial symmetry of the system : when cavities are identical , the system is reflection - symmetric about axis @xmath4 regardless of the value of @xmath0 , see discussion after eq . .
therefore , in some cases , there are two additional eigenstates of @xmath164 co - existing with @xmath173 and @xmath174 and sharing the same imaginary part .
they are linear combination of the images of these eigenstates obtained by reflecting them about axis @xmath4 .
writing @xmath139 , these images are obtained by exchanging @xmath180 and @xmath181 .
we have confirmed this fact by verifying that the 4-dimensional eigenspace associated to four - fold degeneracies is indeed stable under reflection about axis @xmath4 . for the cavity parameters used in the plots , the top - left panel of fig .
[ fig : omega ] suggests that our initial state is stable and superradiance can not take place when @xmath182 .
a few other stable regions appear when @xmath183 increases further . for larger values @xmath184 , a continuous instability develops ,
see top - right panel of fig .
[ fig : omega ] where eigenvalues with growing positive imaginary parts are visible
. the threshold for unstable behavior at weak pump fields does not change when the number of reciprocal lattice vectors used to compute and diagonalize @xmath164 is increased .
however , by expanding the momentum transfer limits and including more reciprocal lattice vectors , additional eigenvalues with positive imaginary parts appear in the range @xmath185 .
a word of caution about the conclusions drawn from the linear stability analysis is necessary .
indeed , the picture may change when subsequent nonlinear terms are included to refine the analysis . as a matter of fact , as we have numerically checked , our initial state is still dynamically unstable when @xmath183 is chosen in between the linear instability regions visible in the top - left panel of fig .
[ fig : omega ] .
however , when @xmath183 is chosen within the linear instability regions , a better atomic lattice structure is obtained numerically .
we present now the results of our numerical simulations of the dynamical equations and with the integrals given in eq . .
as mentioned in sec . [ dma ] , we only consider momenta transfer to atoms up to @xmath138 , which amounts to include thirty - seven points of the reciprocal lattice in our simulations .
starting from empty cavities and a homogeneous atomic cloud , the system undergoes , for appropriate laser fields strengths , a superradiant phase transition and subsequently self - organizes into either a triangular or a honeycomb lattice of atoms depending on the value of the cavity phase @xmath0 . to check that the system does enter a superradiant phase , we monitor the dynamics of the system for a pump field strength larger than the lower bound found in the linear stability analysis , see sec . [ instability ] .
our numerical simulations confirm the superradiant state for @xmath186 .
however the effective potential is not deep enough to support a sharp and stable lattice of atoms .
this is because the pump field is too weak and can not feed a sufficient number of photons inside the cavities .
this problem is overcome by decreasing further the pump field strength down to @xmath187 see fig .
[ fig:0 ] .
as shown in the top - left panel and the insets , the number of photons inside the cavities increases steadily and reaches @xmath188 after a short time @xmath189 .
the atomic density has been computed at three different times . at small times , and before superradiance takes place ( @xmath190 ) , the atomic cloud self - organizes according to the minima of the pump standing - wave potential alone .
as superradiance takes place and the number of photons inside the cavities increases ( @xmath191 and @xmath192 ) , the superposition of the pump and cavity fields creates a sufficiently deep potential and the atoms self - organize into the triangular lattice of potential minima , see left panel of fig .
[ fig : veff ] .
it should be noted that , within the time span shown in fig .
[ fig:0 ] , the number of cavity photons fluctuates a lot , even if it remains sizable , and the effective potential fluctuates too .
this is because the system has not yet reached the steady state .
however these fluctuations do not alter the triangular nature of the lattice of potential minima and , in turn , the triangular nature of the self - organized atomic lattice . as suggested by the results of the linear stability analysis developed in sec .
[ instability ] , superradiance takes place irrespective of the value of the cavity phase @xmath0 .
however the lattice structure of the potential minima is triangular only when @xmath194 , otherwise it is rather hexagonal ( see sec . [ idcav ] ) . to check this point ,
we have considered the same initial state and we have numerically solved eqs . , and with @xmath195 . in this case
, we expect superradiance to organize atoms according to the energy - imbalanced wells of the honeycomb potential , see middle panel of fig .
[ fig : veff ] for an example of such a potential , with more atoms trapped in the deeper wells .
this is indeed the behavior observed in fig .
[ fig : pi4th ] where the obtained atomic density distributions are shown at times @xmath191 and @xmath196 and display the form of a density - imbalanced honeycomb lattice . by the same token , when @xmath110 , the wells of the honeycomb potential have all the same depth and we expect superradiance to drive the atoms into a density - balanced honeycomb lattice , see right panel of fig .
[ fig : veff ] .
this is indeed what our numerical results show , see fig .
[ fig : pihalf ] where the atomic distributions obtained at times @xmath191 and @xmath197 are given .
one may have noticed that , for the same coupling parameters , the atomic honeycomb lattices have lower contrasts than the triangular one .
as one can see in fig .
[ fig : veff ] , the honeycomb minima are already shallower than the triangular ones when the number of cavity photons is the same . here
the dynamics develops less cavity photons for the honeycomb lattice than for the triangular one .
moreover , for the honeycomb lattice , the atoms are distributed over twice many sites than for the triangular lattice .
all these reasons conspire to produce less contrasted atomic lattices for the honeycomb structure than for the triangular one . for the parameters chosen here ,
we infer from eq . that the pump and cavity fields have same order of magnitude when @xmath198 .
as seen from figs .
[ fig:0]-[fig : pihalf ] , this is the case for @xmath199 where @xmath200 fluctuates roughly around @xmath201 . for @xmath113 and @xmath195 , @xmath200 fluctuates roughly around @xmath202 , or a bit less , and the pump field is larger than the cavity fields . as seen in figs .
[ fig:0]-[fig : pihalf ] , the cavity photon numbers @xmath200 suffer strong temporal fluctuations .
this means that the effective potential fluctuates too and thus , in turn , so does the atomic distribution .
in fact , the system needs more time to reach the steady state .
[ fig:0long ] shows the long - time evolution of the system when @xmath160 ( triangular lattice ) .
the steady state is reached around time @xmath203 with a number of cavity photons @xmath204 . in this case
the cavity fields largely dominate over the pump field and the wells of the effective potential organize in a rectangular array rendering the ( exact ) triangular symmetry less apparent as is shown in fig .
[ fig:0long ] . however , the atoms accumulate in the deepest minima of this effective potential which still form a nice regular triangular lattice . the contour plots in fig .
[ fig:0long ] show an example of the atomic density distribution and the effective potential at time @xmath205 .
this distribution does not change in time once the system has reached the steady state . when @xmath206 , as we have seen , atomic honeycomb lattices with density - imbalanced consecutive sites are synthesized .
however these structures are not stable in the long - time limit and atoms re - organize themselves into the deepest potential wells which form a triangular lattice . as an example , fig .
[ fig : pi4thlong ] shows the atomic density distribution and the effective potential obtained at time @xmath207 for @xmath195 .
as clearly seen the density - imbalanced atomic honeycomb lattice found at early times , see fig .
[ fig : pi4th ] , has disappeared in favor of the triangular atomic sub - lattice of deepest wells .
these density - imbalanced atomic honeycomb lattices destabilize in favor of the triangular lattice after a certain latency time @xmath162 . after the time @xmath162
, there is no visible occupation of the shallower lattice sites . approximately . ]
the atomic lattice is purely triangular and becomes the stable structure in the long - time regime .
the latency time @xmath162 depends on the potential energy difference between consecutive sites . since this
potential mismatch decreases when @xmath0 increases , @xmath162 gets longer when @xmath208 , see fig . [
fig : t ] .
note however that the cavity photon numbers take a time longer than @xmath162 to reach their steady - state . against the cavity phase @xmath0 .
after @xmath162 , atoms have re - organized into the deepest wells of the energy - imbalanced effective potential obtained for @xmath2090,\pi/2[$ ] and exhibit a triangular lattice structure .
this time @xmath162 increases when the energy mismatch between the principal and secondary potential minima decreases , i.e. when @xmath0 increases .
it diverges when @xmath208 .
the atomic honeycomb lattice obtained at @xmath110 is a stable structure .
[ fig : t],scaledwidth=40.0% ] when @xmath110 , the effective potential is hexagonal with perfectly energy - balanced minima .
we thus expect a stable density - balanced atomic honeycomb lattice to emerge from the self - organization process . as one can see from fig .
[ fig : pihalflong ] , this is indeed the case , but the cavity photon numbers now reach their steady - state after a much longer time . they even fluctuate a lot during their temporal evolution and induce in turn fluctuations of the atomic lattice .
these atomic density fluctuations get strongly reduced after the time @xmath210 but , even at time @xmath211 , the system has not yet fully reached its steady - state . as one may have noticed , our numerical results produce different cavity photon numbers @xmath212 even if their average behavior is the same and their asymptotic values are equal .
this asymmetry is due to the sensitivity of this system to numerical errors .
however , the difference in @xmath97 and @xmath98 approaches zero as the system reaches its steady state .
so far , we have considered the dynamical behavior of the cavity - atom system for a fixed pump strength , chosen above the threshold for superradiance , and we have studied the self - organization of atoms into triangular or hexagonal lattice structures for different cavity phases .
we now investigate the possibility of driving atoms from one lattice structure to another by changing in time both the pump strength and the cavity phase .
the protocol we explore is the following .
the pump strength increases linearly in time from @xmath213 to some final value @xmath214 over a certain time window @xmath215 .
meanwhile , the cavity phase is set at @xmath216 initially .
after superradiance takes place and enough number of photons are produced inside the cavities , the phase is then changed linearly in a time interval of @xmath217 to its final value @xmath218 .
finally , the cavity phase keeps this target value for the rest of the time sequence .
we present below the time evolution of the cavity photon numbers and the atomic distribution obtained numerically for this experimental protocol for 3 possible cavity phase values : @xmath160 , @xmath110 and some intermediate value that we choose to be @xmath193 . when the system starts with @xmath220
, the atoms organize in a triangular lattice . by switching the cavity phase to @xmath221
, a new optical potential is produced featuring additional shallower wells .
however we know from sec .
[ longtime ] that when @xmath206 , the stable structure at long times is the triangular lattice formed by atoms occupying the deepest potential wells .
since atoms have already self - organized in a triangular lattice , changing the cavity phase from @xmath213 to @xmath219 will not affect the atomic distribution which thus remains triangular .
on the other hand , when the system starts instead with @xmath222 and the cavity phase is changed to @xmath223 , the transition to the triangular lattice is inevitable and is illustrated in fig .
[ fig : pi4h-0 ] .
as the cavity phase decreases , the secondary minima of the effective potential obtained for @xmath222 become shallower and shallower and are occupied by less and less atoms .
the triangular lattice becomes the prominent structure around @xmath224 , well before @xmath0 reaches the target value @xmath213 , which triggers a considerable rise of the cavity photon numbers . in this case ,
both the triangular and density - balanced honeycomb lattices are stable structures of the static situation .
[ fig:0-pihalf ] shows what happens when we start from the triangular lattice ( @xmath220 ) and gradually increase the cavity phase to @xmath225 .
as one expects from the previous case , the atomic lattice structure remains triangular as long as the cavity phase has not reached its target value , @xmath226 .
interestingly enough , once the cavity phase has reached @xmath225 , the atomic density distribution starts to oscillate between the wells of the two triangular sub - lattices of the energy - balanced effective potential , taking the form of density - imbalanced atomic honeycomb lattices in between .
these two triangular sub - lattices are simply shifted by @xmath227 . as a net result
, the atomic density looks like `` blinking '' between two shifted triangular lattices , a sign of a bi - stable behavior .
this blinking does not stop if the pump strength is kept fixed , rather than linearly increasing , after the cavity phase has reached @xmath225 .
it even speeds up if the pump strength is ramped up faster .
the bottom panels in fig .
[ fig:0-pihalf ] display the atomic densities at times @xmath228 and @xmath229 , showing that , for the experimental parameters chosen , the oscillation period between the two triangular lattices is comparable to the pump recoil time @xmath230 .
it is worth recalling that , once the cavity phase is set to @xmath3 , the minima of the effective potential have same depth and display a honeycomb structure , irrespective of the cavity photon numbers ( as long as they are not zero ) .
this case should not thus be confused with the previous case ( @xmath206 ) where the lattice potential has minima with different depths .
our present results then show that , for time - driven parameters , the density - balanced honeycomb structure is a transient state between two stable triangular structures .
[ fig : pihalf-0 ] shows what trivially happens for the reverse process , i. e. when we start with @xmath231 and decrease gradually the cavity phase to @xmath223 .
once the cavity phase departs from @xmath3 , the effective potential has minima with different depths .
as long as the potential mismatch is weak enough , the initial density - balanced honeycomb structure resists but , with increasing potential mismatch , finally gives in and a density - imbalanced honeycomb lattice is formed around @xmath232 .
then , as the cavity phase decreases further , the depth of the secondary minima , as well as the number of atoms sitting on them , decreases rapidly . before the cavity phase finally reaches @xmath160 at @xmath233 , the atoms choose the deeper triangular sub - lattice which becomes the stable structure as expected .
we address now the last case , namely the transition between the two possible , density - balanced or density - imbalanced , atomic honeycomb lattices .
we first consider switching the cavity phase from @xmath222 to @xmath225 . in this case , some of the atoms are sitting in the shallower potential wells of the effective honeycomb potential while the majority of them are hosted in the deeper ones . by increasing the cavity phase ,
the shallower sites become deeper .
however , the atoms in the deeper wells tend to stay where they are ( see sec .
[ tri - hon1 ] ) until the cavity phase becomes exactly @xmath3 and all potential wells have same depth . at this point ,
the atomic lattice is hexagonal with almost density - balanced sites .
then , for the rest of the time sequence , atoms start oscillating in time between the two stable triangular sub - lattices , featuring the `` lattice blinking '' already observed in the preceding sec .
[ tri - hon2 ] . when the system starts with @xmath231 ,
the initial atomic lattice is a density - balanced honeycomb lattice .
by gradually decreasing the cavity phase , the potential wells of one of the two triangular sub - lattices become shallower .
however , the initial density - balanced atomic honeycomb lattice resists the potential mismatch even after the cavity phase has reached its final value @xmath195 .
it gets finally destabilized around @xmath234 and takes the form of a density - imbalanced honeycomb lattice , see fig .
[ fig : pihalf - pi4th ] . as explained in sec .
[ longtime ] , this structure is not stable in the long - time limit and destabilizes subsequently into a triangular lattice as the atoms in the shallower sites move to the deeper sites .
in this work we have proposed to load a two - dimensional cloud of non - interacting cold bosonic atoms inside two identical initially - empty optical cavities with an angle of @xmath235 between their axes .
the atoms are driven by an external laser field .
we have given the hamiltonian of this hybrid system in the dispersive regime and we have derived the corresponding dynamical equations in the mean - field regime .
the coherent superposition of the cavity and pump fields creates a dynamical effective lattice potential with a triangular bravais structure in which the atoms move . as a result of superradiance , the atoms
self - organize into a triangular or a honeycomb lattice inside the cavities , the nature of the lattice depending on the relative phase between the pump and the cavity fields . using the symmetry properties of the effective potential ,
we have derived the dynamical equations in reciprocal space and we have investigated the condition required for superradiance to take place .
linear response theory shows that superradiance takes place irrespective of the relative phase between the cavity and pump fields .
this is confirmed by our numerical results : atoms self - organize into a triangular lattice when pump and cavity fields are in phase , into a density - balanced honeycomb lattice when the fields are in quadrature and into a density - imbalanced honeycomb lattice in between .
the stable atomic structures in the long - time limit are the triangular and density - balanced honeycomb lattices .
the density - imbalanced honeycomb lattice only survives for a limited amount of time which becomes longer and longer as the cavity and pump fields approach quadrature . in the end ,
atoms redistribute equally into the triangular sub - lattice made of the deepest potential wells .
we have also studied the transition between these different lattice structures when the relative phase between the fields is dynamically changed and confirmed the stability of the triangular and density - balanced honeycomb structures .
a natural extension of this work would include the study of the survival time of the density - imbalanced honeycomb lattices , the ( possibly chaotic ) dynamics leading to superradiance when the fields are in quadrature and , importantly , the effect of atomic interactions on the self - organization process .
safaei would like to gratefully acknowledge inspiring discussions with zgur esat mstecapliolu .
the centre for quantum technologies is a research centre of excellence funded by the ministry of education and national research foundation of singapore .
h. ritsch , p. domokos , f. brennecke , and t. esslinger , rev .
phys . * 85 * , 553 ( 2013 ) . i. b. mekhov and h. ritsch , j. phys .
b : at . mol .
* 45 * , 102001 ( 2012 ) .
i. b. mekhov , c. maschler , and h. ritsch , nature physics * 3 * , 319 ( 2007 ) .
p. horak , s. m. barnett , and h. ritsch , * 61 * , 033609 ( 2000 ) .
f. brennecke , t. donner , s. ritter , t. bourdel , m. khl , and t. esslinger , nature * 450 * , 268 ( 2007 ) .
y. colombe , t. steinmetz , g. dubois , f. linke , d. hunger , and j. reichel , nature * 450 * , 272 ( 2007 ) .
s. slama , s. bux , g. krenz , c. zimmermann , and ph . w. courteille , * 98 * , 053603 ( 2007 ) .
m. wolke , j. klinner , h. keler , and a. hemmerich , science * 337 * , 75 ( 2012 ) .
elssser , b. nagorny , and a. hemmerich , * 69 * , 033403 ( 2004 ) . s. gupta , k. l. moore , k. w. murch , and d. m. stamper - kurn , * 99 * , 213601 ( 2007 ) .
s. ritter , f. brennecke , k. baumann , t. donner , c. guerlin , and t. esslinger , app .
b * 95 * , 213 ( 2009 ) .
s. safaei , . e. mstecapliolu , and b. tanatar , laser phys .
* 23 * , 035501 ( 2013 ) .
m. diver , g. r. m. robb , and g .- l .
oppo , * 89 * , 033602 ( 2014 ) .
p. domokos and h. ritsch , * 89 * , 253003 ( 2002 ) .
a. t. black , h. w. chan , and v. vuleti , * 91 * , 203001 ( 2003 ) .
d. nagy , g. szirmai , and p. domokos , eur .
j. d * 48 * , 127 ( 2008 ) .
g. knya , g. szirmai , and p. domokos , eur .
j. d * 65 * , 33 ( 2011 ) .
s. bux , c. gnahm , r. a. w. maier , c. zimmermann , and ph .
w. courteille , * 106 * , 203601 ( 2011 ) .
m. j. bhaseen , j. mayoh , b. d. simons , and j. keeling , * 85 * , 013817 ( 2012 ) .
r. h. dicke , phys .
93 * , 99 ( 1954 ) .
k. hepp and e. h. lieb , ann .
phys . * 76 * , 360 ( 1973 ) .
y. k. wang and f. t. hioe , * 7 * , 831 ( 1973 ) .
d. nagy , g. knya , g. szirmai , and p. domokos , * 104 * , 130401 ( 2010 ) .
k. baumann , c. guerlin , f. brennecke , and t. esslinger , nature * 464 * , 1301 ( 2010 ) .
k. baumann , r. mottl , f. brennecke , and t. esslinger , * 107 * , 140402 ( 2011 ) .
e. g. dalla torre , s. diehl , m. d. lukin , s. sachdev , and p. strack , * 87 * , 023831 ( 2013 ) .
m. reza bakhtiari , a. hemmerich , h. ritsch , and m. thorwart , * 114 * , 123601 ( 2015 ) .
g. knya , d. nagy , g. szirmai , and p. domokos , * 86 * 013641 ( 2012 ) .
b. ztop , m. bordyuh , . e. mstecapliolu , and h. e. treci , new j. phys .
* 14 * , 085011 ( 2012 ) .
d. nagy , g. szirmai , and p. domokos , * 84 * 043637 ( 2011 ) .
s. safaei , . e. mstecapliolu , and b. tanatar , new j. phys . *
15 * 083037 ( 2013 ) .
jian - song pan , xiong - jun liu , wei zhang , wei yi , and guang - can guo , arxiv:1410.8431 .
corinna kollath , and ferdinand brennecke arxiv:1502.01817 .
k. l. lee , b. grmaud , r. han , b .-
englert , and c. miniatura , * 80 * 043411 ( 2009 ) .
a. h. castro neto , f. guinea , n. m. r. peres , k. s. novoselov , and a. k. geim , rev .
81 * , 109 ( 2009 ) .
l. tarruell , d. greif , t. uehlinger , g. jotzu , and t. esslinger , nature * 483 * , 302 ( 2012 ) .
n . fuchs , and g. montambaux , * 108 * , 175303 ( 2012 ) . | we consider a two - dimensional homogeneous ensemble of cold bosonic atoms loaded inside two optical cavities and pumped by a far - detuned external laser field . we examine the conditions for these atoms
to self - organize into triangular and honeycomb lattices as a result of superradiance . by collectively scattering the pump photons ,
the atoms feed the initially empty cavity modes . as a result ,
the superposition of the pump and cavity fields creates a space - periodic light - shift external potential and atoms self - organize into the potential wells of this optical lattice . depending on the phase of the cavity fields with respect to the pump laser , these minima
can either form a triangular or a hexagonal lattice . by numerically solving the dynamical equations of the coupled atom - cavity system
, we have shown that the two stable atomic structures at long times are the triangular lattice and the honeycomb lattice with equally - populated sites .
we have also studied how to drive atoms from one lattice structure to another by dynamically changing the phase of the cavity fields with respect to the pump laser . |
the goal of quantum state tomography , or quantum state estimation , is to arrive at a best guess of the unknown state @xmath0 of a quantum system , based on data collected from measuring a number of identical copies of the state .
an accurate guess is needed in all aspects of quantum information or quantum computation , ranging from the characterization of an unknown quantum communication channel , to a check of a quantum gate implementation , or to the verification of a state preparation procedure in the lab . whether the guess from a particular tomography recipe can be considered the best , or most accurate , depends on one s figure - of - merit , which should be chosen according to the quantum information processing task at hand . in many situations ,
one is interested not in the state of the system itself , but in a quantity computed from it , _
e.g. _ , the amount of entanglement in the state . in such cases , rather than reporting a best guess for @xmath0 , one expects to get a more accurate answer by directly estimating the quantity of interest from the data , as done in a related procedure carrying the name of `` parameter estimation . '' in other situations , one is interested in a range of quantities related to @xmath0 , and reporting a best guess for the state itself [ _ e.g. _ , by maximizing the likelihood for the data over all physical states , as is done for the maximum - likelihood estimator ( mle ) ] can be a convenient and self - consistent way of summarizing and interpreting the data . in the latter case , one might assess the accuracy of the estimate obtained from a particular tomography scheme by examining some measure of closeness between the estimate @xmath1 , and the true state @xmath0 , for a variety of known true states ( _ e.g. _ , from sources that have previously been fully characterized ) .
a poorer , but possibly useful , gauge is to look at the accuracy of the prediction of one of the quantities of interest computed from @xmath1 . reference compares the performance between an estimator @xmath2 from a procedure the authors refer to as linear inversion ( lin ) , and two other estimators , @xmath3 ( the standard mle @xcite ) , and @xmath4 ( from a procedure known as `` free least squares '' ( fls ) @xcite ) .
the article assesses the three estimators by looking at the accuracy in the prediction for target fidelity , a relevant quantity when the source is supposed to produce a certain target state .
the fidelity @xmath5 takes value between @xmath6 and @xmath7 for physical @xmath8 and @xmath9 @xcite . in ref .
, the quality of an estimator is measured by the target fidelity @xmath10 , the fidelity between the target state @xmath11 and the estimator @xmath1 , which is compared with the actual true value @xmath12 computed for the true state @xmath0 . here , the `` true '' state yields the probabilities that are used for the generation of the simulated data from which the various estimators are derived .
the authors of ref . perform this comparison for different @xmath11s and @xmath0s , for many repeated simulations of the measurement data ( and hence , many @xmath1s , one for each data ) .
they draw the conclusion that @xmath2 is always the best , because it is an unbiased estimator , _
i.e. _ , fluctuations from different runs of the same experiment lead to fluctuations in the predicted @xmath13 value , but all centered about the true value @xmath14 ; @xmath3 and @xmath4 , on the other hand , give predictions that are biased , _
i.e. _ , have a systematic shift away from @xmath14 ( see fig . 1 of ref . and fig .
[ fig : noisyghz ] below ) .
the authors go on to point out that any estimation procedure that always produces a physical ( _ i.e. _ , nonnegative ) state will unavoidably be biased ; their @xmath2 , coming from an unbiased estimation procedure , is not guaranteed to be a physical density operator , and in fact , generically has negative eigenvalues . here , the qualifiers _ biased _ and
_ unbiased _ have the technical meaning that is discussed below in the context of eq .
( [ twopieces ] ) .
contrary to their connotation in common parlance , they are not synonyms of `` bad '' and `` good . ''
one must not fall into the trap of regarding a biased estimator as automatically inferior to an unbiased one .
indeed , it is well known in classical statistics that unbiased estimators are not always the best choice . instead , minimizing the mean squared error ( mse , a popular measure of estimation accuracy ) is key , and this is often not accomplished by minimizing the bias .
in fact , we will show that the lin approach yields mses that are comparable to ( and sometimes worse than ) what one obtains from the mle procedure ; yet the lin technique forces us to give up physicality , which leads to many severe problems and highly restricts the usefulness of the estimator @xmath2 .
the mle itself also does not and was never designed to minimize the mse , but it does a comparably good job as @xmath2 while enforcing physicality .
we hence see little utility at all in employing the lin strategy .
below , we remind the reader why , in the quantum context , it is usually not a good idea to treat relative frequencies obtained from the data as probabilities , as is prescribed by the lin procedure of ref . .
then , we explain why focusing on reducing bias only , and not the overall mse , constitutes a conceptual misunderstanding .
lastly , we compare the mses obtained from the mle and the lin approaches and observe that one can easily find examples in which the biased mles have smaller mses than the unbiased lin estimators .
before we begin , a brief note on notation is in order .
the tomography measurement is described by a positive - operator - valued measure ( povm ) , or , if we use a more descriptive name , a probability - operator measurement ( pom ) : it comprises a set of outcomes @xmath15 , one for each detector , with @xmath16 for all @xmath17 and @xmath18 .
the probability of getting a click in detector @xmath17 , corresponding to outcome @xmath15 , is given by the born rule , @xmath19 .
in the tomography experiment , @xmath20 identically prepared copies of the ( unknown ) state @xmath0 are measured using the pom . the data @xmath21 consist of a particular sequence of detector clicks , summarized by the set of relative frequencies @xmath22 , where @xmath23 is the number of clicks in detector @xmath17 . from @xmath21 ,
one estimates the probabilities @xmath24 using a chosen procedure like mle or lin , and from these ( if one can , _
e.g. _ , in the case of tomographically complete poms ) , one constructs the estimator @xmath1 .
lin , as proposed in ref .
, sets the estimated probabilities equal to the relative frequencies of the observed data , @xmath25 and then obtains @xmath2 by `` linear inversion '' of the born rule @xmath26 . while relative frequencies will be close to probabilities when there is a lot of data , they are most certainly not the same thing : relative frequencies satisfy only one constraint , that of unit sum : @xmath27 ; probabilities @xmath28 ( for pom @xmath29 ) that arise from a physical state @xmath0 through the born rule satisfy further constraints imposed by the positivity of @xmath0 .
the latter constraints can be easily stated for the case of measuring a qubit state with the symmetric informationally complete pom ( sic pom ) , the tetrahedron measurement @xcite , where @xmath30 requires the four tetrahedron probabilities to satisfy @xmath31 , in addition to the unit - sum constraint . since the relative frequencies do not themselves satisfy these physicality constraints , @xmath2 is hence not necessarily a physical state , as is also emphasized in ref .
( and many other existing references in the literature ) . that @xmath2 is not necessarily nonnegative , is not a minor nuisance : many quantities associated with a physical state @xmath0 are ill - defined for @xmath32 and can no longer be computed ,
_ e.g. _ , entropy , negativity , and the fidelity with another state .
other quantities , such as the purity @xmath33 or the expectation value @xmath34 of an observable @xmath35 , are computable for @xmath32 , but the numbers so obtained do not mean purity , expectation value , etc .
hence , @xmath2 may not just lack a reasonable physical interpretation , but may also not be useful at all . in the case of uncomputable quantities , the proposal of ref . is to be content with the bounds that can be computed from linear approximations .
these bounds , however , also lack a physical meaning if they are evaluated for @xmath32 . while one might choose not to be too concerned if @xmath2 is only slightly unphysical ( however one may want to quantify that statement ) , or if an unphysical @xmath2 occurs only rarely , getting an unphysical @xmath2 can be generic in certain situations .
for example , imagine a qubit state measured with the tetrahedron measurement , and suppose that the true state is orthogonal to one of the tetrahedron outcomes ( say the one labeled by @xmath36 ) .
then , the only relative frequencies that can give a physical @xmath2 are @xmath37 and @xmath38 , _ i.e. _ , the detector counts for all outcomes , other than the tetrahedron leg orthogonal to the true state , must be exactly equal .
this is not even possible if the total number of counts is different from a multiple of 3 .
mse-fig1.pdf lest the reader complains that the above is a pathological case , another situation where one sees a stark contrast between frequencies and probabilities can be found in the commonly used `` bb84-like '' measurements , _
i.e. _ , measure the pauli @xmath39 and @xmath40 on a single qubit with equal probability . in an optical implementation ,
where the qubit is the photon polarization , the usual way this measurement is implemented is by having a 50 - 50 beam - splitter direct the incoming photons into two possible paths , one carrying out the @xmath39 measurement , the other the @xmath40 measurement ; see fig .
[ fig : bb84 ] .
now , the probabilities for such a measurement , by the very nature of the measurement structure , satisfy @xmath41 , alongside the positivity constraint @xmath42 .
the relative frequencies , however , obey no such constraints : despite the 50 - 50 nature of the beam - splitter , one hardly ever encounters the situation where _ exactly _ half the photons travel down the @xmath39 path , and half down the other .
the procedure of finding @xmath2 from such relative frequencies will then typically be internally inconsistent , and yields no solution .
one common fix used to circumvent the above problem requires one to ignore the counts in one of the detectors , _
e.g. _ , the one measuring the @xmath43 eigenstate for @xmath39 ( outcome @xmath44 in fig .
[ fig : bb84 ] ) @xcite . to ensure that the relative frequencies comply with the constraint of @xmath45 , in imitation of the probabilities , one replaces the actual count @xmath46 obtained by @xmath47 , and modifies the total number @xmath20 to be @xmath48 .
however , this ad - hockery , which involves discarding data , does not guarantee that the relative frequencies satisfy the remaining positivity constraint on the probabilities . the simulations in ref .
mimic the tomography of four - qubit states using product pauli poms , _
i.e. _ , measure @xmath49\equiv o_{1}\otimes o_{2}\otimes o_{3}\otimes o_{4}}$ ] , with each of the four @xmath50s equal to one of the three pauli operators @xmath39 , @xmath51 , and @xmath40 . rather than just having two settings of @xmath39 and @xmath40 as in the single - qubit case of fig .
[ fig : bb84 ] , there are now @xmath52 settings , all to be measured with equal probability
. a practical way of carrying out this tomography measurement , having ( usually ) no access to a 1-to-81 beam - splitter , would be to divide the total number of counts by @xmath53 , and measure @xmath54 copies of @xmath0 with each setting ( this is another way of overcoming , by hand , the difficulty discussed in the previous two paragraphs ) .
this automatically ensures that the relative frequencies for each setting @xmath55 $ ] sum to @xmath56 , as do the corresponding probabilities .
however , these constraints are not the only ones needed to ensure internal consistency .
there is , for example , the problem that frequencies obtained when measuring , say , @xmath57 $ ] and @xmath58 $ ] give two values for the expectation values of the two - qubit observable @xmath59=x\otimes1\otimes1\otimes y$ ] and also for the single - qubit observables @xmath60 $ ] and @xmath61 $ ] . in ref .
, these conflicts are resolved by a fix that may appear plausible , but is ad - hockery nevertheless and ultimately difficult to justify .
furthermore , there is still the issue of positivity constraints , and our own simulations show that almost all the estimators constructed by lin violate positivity and , therefore , are unphysical .
none of these ad - hoc fixes or discard of data are necessary if we do not insist on setting frequencies equal to probabilities . the usual approach of finding a _ physical _ density operator that best fits the data , _
e.g. _ , by maximizing the likelihood for the data over all physical states in the case of the mle , automatically handles all these constraints .
there is no need to include constraints explicitly by hand , such a need becoming more severe and difficult to carry out as the dimensionality of the system and the complexity of the pom increase . of course
, finding the best - fit physical estimator for the data is also not easy , but the estimator is assuredly physical , as is the true state , and one interprets the data honestly without additional tweaks .
the accuracy of an estimator is often quantified by the mean - squared error ( mse ) , _ i.e. _ , the mean , over all possible data , of the squared deviation of the estimated parameter from the true value , but this is only one of many measures of inaccuracy . to be concrete , we take as example the estimate @xmath62 for the target fidelity ( with respect to some target state @xmath11 ) as discussed in ref . .
the mse in this case is @xmath63=[\mathbb{e}_{{\rho_{\textsc{t}}^{\ } } } ( \hat{f})-f_0]^2 + \mathbb{v}_{{\rho_{\textsc{t}}^{\ } } } ( \hat{f})\,,\ ] ] where we adopt the notation of ref . :
@xmath64 denotes the expected value and @xmath65 the variance , for the true state @xmath11 .
the mse in eq .
( [ twopieces ] ) is the sum of two contributions : the first is the so - called _ bias _ of the estimator ( as compared with the true value @xmath14 ) ; the second is the variance of @xmath62 .
as it is written , the formula for the mse makes no statement about the relative importance of the two pieces , and a general estimation strategy can have different relative sizes for , and compromises between , the two terms .
other measures of inaccuracy may not have a break - up analogous to eq .
( [ twopieces ] ) , and the bias may not be a relevant notion .
it is easy to understand intuitively why both pieces , not just the bias , must be small in order for us to have reasonable confidence of obtaining a good estimate from a single set of data .
if the bias is small , and one has many runs of the same experiment , yielding very many estimates , one might imagine getting a good answer by looking for the centre of all these estimates .
however , one usually has access to only one set of data , and hence a single estimate @xcite ; other estimates may then be generated by bootstrapping the data , but this just gives a distribution of values centered at the data - based estimate .
if the estimation strategy yields no bias , but a large variance , one will often end up with an estimate that is quite far away from the true value .
hence , one needs both the bias , as well as the variance , summarized as the mse , to be small for an accurate guess .
most statistics textbooks ( see , for example , @xcite ) do focus on the class of unbiased estimators for a given problem .
however , this is always discussed together with the search for one with the minimum variance , _
i.e. _ , to minimize the mse _ within the class of unbiased estimators_. the resulting theory of point estimation for unbiased estimators has rather nice and concise results , with wide applicability in many areas of statistics . yet , at the same point where unbiased estimators are discussed , textbooks usually remind readers that unbiased estimators are not the whole story .
the constraint of unbiasedness is often too strong : an estimator with a slight bias but a significantly smaller variance can very well be the estimator that minimizes the mse . certainly , considering only unbiasedness without also trying to minimize the variance , as suggested in ref . , is altogether insufficient to give a good estimator .
furthermore , as most textbooks will also explain , reasonable unbiased estimators do not exist in many problems .
the nonexistence of an unbiased estimator that satisfies the positivity constraint for quantum states , as is emphasized in ref . , is but another example of the difficulty in requiring unbiasedness in constrained estimation problems . as a result of constraints , pure states and other rank - deficient states
are extremal and the valid ( _ i.e. _ , satisfy the constraints ) estimators can not approach them from all sides : the estimators are unavoidably biased .
this is clearly visible in the examples discussed in the next section .
the trustworthy estimators , such as the mles , are _ consistent _ , _ i.e. _ , they converge to the true state when more and more data are taken into account .
this consistency is an important properties that is worth insisting upon , while unbiasedness is not .
the various virtues of , and problems with , unbiased estimators have been repeatedly expounded upon in the statistics literature . as an example , we point the reader to the excellent discussion found in jaynes s book on probability theory , ref .
@xcite ( see chapter 17 ) . within jaynes
s book is an example where the requirement of unbiased estimators leads one to waste half the data as compared with a biased estimator that attains the same mse @xcite .
to illustrate the issues mentioned in the previous sections , we re - examine the example of target fidelity discussed in ref . .
the unknown true state is a four - qubit state , and the parameter of interest is its fidelity with some target state @xmath11 .
the measurement comprises the @xmath53 product pauli operators described above , with @xmath66 copies measured per setting , and @xmath67 possible outcomes per setting . to explore the statistics ,
for each set of target and true states , we simulate @xmath68 runs of the experiment , obtain the estimators @xmath3 and @xmath2 for each set of data , and from these , estimate the target fidelity .
as observed above , there is a problem with evaluating ( not to mention , interpreting the meaning of ) the target fidelity @xmath69 when @xmath2 is not nonnegative .
reference also admits this problem , and suggests to fix it by using the formula @xmath70 instead , which is correct _ if _ @xmath2 is nonnegative and @xmath11 is a pure state .
although @xmath71 can be computed also for an unphysical @xmath2 , it does not have the significance of the target fidelity then .
the other issue , _ viz . _ that @xmath11 must be pure , is of lesser concern if all target states of interest are indeed pure ; the smolin state considered in ref . is not pure , and this is why the respective line for @xmath2 is missing in their fig . 2 .
clearly , one already sees the problems with using a nonphysical estimator .
fidelity histograms of simulated quantum - state tomography experiments for the noisy ghz state from ref . .
the true state is a ghz state with added white noise so that its fidelity is @xmath72 with respect to the original target ghz state . ] nevertheless , we can still compare the estimated target fidelities obtained from the mle and lin procedures for pure target states only , using their suggested formula of @xmath73 for @xmath74 , whether @xmath1 is physical or not .
as the first target state , we consider the four - qubit greenberger - horne - zeilinger ( ghz ) state with the ket @xmath75 .
this is the main example discussed in ref . ; we simulate the identical situation and the results are shown in fig .
[ fig : noisyghz ] , which are very similar to those in fig . 1 of ref . .
fidelity histograms of simulated quantum - state tomography experiments for various states ( indicated in plots ) .
the ghz and w4 states of plots ( a ) and ( b ) have true fidelity of @xmath7 ; plot ( c ) is for a true state that has fidelity @xmath76 with a randomly chosen pure target state . in plots ( b ) and ( c ) ,
the histograms for the lin values extend into the unphysical range of fidelity values exceeding unity .
] instead of mixing the ghz state with white noise , we can also take the ghz state itself as the true state .
then the true fidelity with respect to the target state takes the maximal value of @xmath7 .
in this situation , we find that all the fidelity values estimated using lin give the true value of @xmath7 , while the mle values slightly underestimate the fidelity ( see fig .
[ fig : plots](a ) ; @xmath77 , the mean and the standard deviation obtained from 500 data sets ) . that lin always gives the correct fidelity value , regardless of the statistical fluctuations in the data , is a special feature of the ghz target state in conjunction with the product pauli pom @xcite .
such a singular situation ( as can be seen from more generic examples below ) does not provide a fair benchmark from which to draw reliable conclusions about the efficacy of the estimation procedures . as the second example , we use a different target state , also taken from ref . , namely the four - qubit w - state , @xmath78 .
to illustrate the problems of a nonphysical @xmath2 , we again consider the extreme scenario where the true state is exactly the target state so that the true fidelity is @xmath7 .
lin gives fidelity values ( @xmath79 ) that are distributed around the true value but have a wide spread , with a large fraction of the values exceeding @xmath7 , a situation without physical meaning .
the results obtained via mle ( @xmath80 ) again slightly underestimate the fidelity , but not by much , and the spread of values is much narrower than for lin ; see fig .
[ fig : plots](b ) . that the mle underestimates
the true fidelity value is unavoidable because @xmath81 for physical states and this demonstrates our earlier point that one can only approach an extremal value from one side .
the ghz and w states considered thus far are quite special states , and particularly so with respect to the product pauli pom .
one wonders about the behaviour for a generic target state . to this end
, we randomly pick a pure target state @xmath82 and a random true state @xmath0 with a target fidelity close to @xmath7 .
the selected true state has target fidelity of @xmath76 ; the fidelities calculated via lin give @xmath83 and those gotten via mle give @xmath84 ; see fig .
[ fig : plots](c ) . again , a significant fraction of fidelity values obtained from lin are greater than @xmath7 ; no such problems appear in the mle case , by construction , and the mle range of values has a smaller spread .
fidelity histograms of simulated quantum - state tomography experiments for the same noisy random pure state as in fig .
[ fig : plots](c ) , but with the true fidelity being @xmath85 .
the histograms for the lin values extend into the unphysical range of negative fidelity values . ]
the same problems of unphysical fidelity values appear at the other end of the allowed fidelity range , _
i.e. _ , for @xmath13 close to @xmath6 .
we again use the same randomly generated pure state @xmath82 as the target state , but now choose a true state with a target fidelity of @xmath85 .
this is the situation in which one s intent is to prepare a state orthogonal to the target state . in this setting
, we find that the fidelities calculated via lin ( @xmath86 ) still center around the true value , but with a large fraction being less than @xmath6 , another situation that is physically meaningless . on the other hand ,
the values obtained via mle ( @xmath87 ) in this case slightly overestimate the fidelity , but again the spread of mle values is smaller than the spread of lin values ; see fig . [
fig : rand_ort ] .
c@c@r@@xmath88l@r@@xmath88l@r@@xmath88l ' '' '' & & & & + ' '' '' target & @xmath14 & lin & mle & lin & mle & lin & mle + ' '' '' @xmath89 & 0.8 & @xmath90 & @xmath91 & @xmath92 & @xmath93 & @xmath94 & @xmath95 + ' '' '' @xmath89 & 1.0 & @xmath96 & @xmath97 & @xmath96 & @xmath98 & @xmath96 & @xmath99 + ' '' '' @xmath100 & 1.0 & @xmath101 & @xmath102 & @xmath103 & @xmath104 & @xmath105 & @xmath106 + ' '' '' @xmath82 & 0.981 & @xmath107 & @xmath108 & @xmath109 & @xmath110 & @xmath111 & @xmath112 + ' '' ''
@xmath82 & 0.016 & @xmath113 & @xmath114 & @xmath115 & @xmath116 & @xmath117 & @xmath118 + ' '' '' @xmath82 & 0.8 & @xmath119 & @xmath120 & @xmath121 & @xmath122 & @xmath123 & @xmath124 + ' '' '' @xmath125 & 0.8 & @xmath126 & @xmath127 & @xmath126 & @xmath128 & @xmath129 & @xmath130 + ' '' '' @xmath131 & 0.8 & @xmath132 & @xmath133 & @xmath132 & @xmath134 & @xmath135 & @xmath136 + ' '' '' @xmath137 & 0.8 & @xmath138 & @xmath139 & @xmath138 & @xmath140 & @xmath141 & @xmath142 + in all the examples above , lin suffers from the problem of having unphysical fidelity values .
note that any attempt to fix these larger - than-1 or smaller - than-0 @xmath13 values from lin , _
e.g. _ , set all values @xmath143 to be equal to @xmath7 , will bias the originally unbiased lin estimator . in all cases ,
lin indeed gives a _
fidelity closer to the true value ( in fact it should be exactly equal , if not for the fluctuations from only @xmath68 runs ) than mle but , apart from the singular case of the ghz state , has a significantly larger spread than the mle values , suggesting a possibly larger , or at least comparable , mse . as mentioned in the previous section , for a single run of the experiment , what matters is not the mean value over many repeated runs , but the mse ; a large variance means that one has a high chance of winding up rather far away from the true value .
thus a meaningful comparison between mle and lin requires us to look at the mse for a variety of states .
table [ tbl : mse ] presents the results of such a comparison .
the first six rows of the table give the variance , bias and mse values for the examples described above ; the remaining rows give the values for a further three randomly chosen target states , with the true state in each case being the target state with added white noise to yield a true target fidelity of @xmath72 .
in all cases , lin has zero bias ( by construction ) , with the small deviations indicated in the table attributable to finite statistics ; however , it has a variance comparable to that from the mle in every single case .
for the randomly chosen states , the mse values from lin are slightly smaller than those from mle , but again , mle is not designed to reduce the mse ( neither bias nor variance ) , and sacrificing physicality in lin is a very high price to pay for the occasional small reduction of the mse .
the method of linear inversion may appear simple and intuitive
after all , relative frequencies are asymptotically the correct frequentist s interpretation of probabilities .
however , many difficulties arise , both in the construction of @xmath2 ( having to put in ad - hoc fixes for internal consistency ) , as well as in the interpretation of the estimated quantities ( if one can compute them at all from @xmath2 , and having to deal with parameter values of no physical meaning even when computation is possible ) .
both these problems disappear when insisting on physical estimators .
this requirement is natural in all cases , since the true state must be physical , not just asymptotically so .
even if one is willing to deal with the problems of having a nonphysical estimator , reducing the mse , not just the bias , is the key , and the reduction had better be substantial enough for one to want to cope with the nonphysicality issues . to end the discussion ,
we remind the reader of a point briefly mentioned in the opening paragraphs : if one is interested in estimating only a single parameter , a direct estimation from the data without first going through an estimator for the state generally works better .
thus , our assessment here of the lin versus mle is also lacking in that we should compare their performance for a variety of parameters computed from @xmath2 and @xmath3 , for one would not report these estimators unless the interest is in a few different parameters computed from the state . yet , given that lin already does poorly in terms of target fidelity , it is hardly necessary to provide more evidence against the use of lin by exploring other parameters .
furthermore , a more comprehensive summary of the data is provided not by the _ point _
estimators @xmath2 and @xmath3 , but by regions of estimators , which can circumvent some of the problems highlighted here ; see @xcite for confidence regions , and @xcite for credible regions .
we thank yong siah teo and david nott for stimulating discussions .
the centre for quantum technologies is a research centre of excellence funded by the ministry of education and the national research foundation of singapore .
this @xmath13 is actually the square of the fidelity , if a close analog of the fidelity of two probability distributions is wanted .
we here adopt the convention of ref . and use the term `` fidelity '' for @xmath13 thus defined .
what is particularly ironic is jaynes s point that , for his example , the estimator that minimizes the mse actually requires one to `` double the bias '' by dividing the variance by @xmath144 ( for sample size @xmath20 ) instead of the usual unbiased correction of @xmath145 , as in ref .
[ 37 ] of ref . .
if the ghz state is the target , then the target fidelity involves the expectation values of @xmath146 $ ] , @xmath147 $ ] , @xmath148 $ ] as well as @xmath149 $ ] and @xmath150 $ ] and the observables obtained by permuting the four qubits .
the ghz state is an eigenstate of every one of these @xmath151 observables .
hence , there are no statistical fluctuations when their expectation values are estimated from the data obtained from the product pauli pom . | because of the constraint that the estimators be _ bona fide _ physical states , any quantum state tomography scheme including the widely used maximum likelihood estimation yields estimators that may have a bias , although they are consistent estimators .
schwemmer _
et al . _
( arxiv:1310.8465 [ quant - ph ] ) illustrate this by observing a systematic underestimation of the fidelity and an overestimation of entanglement in estimators obtained from simulated data .
further , these authors argue that the simple method of linear inversion overcomes this ( perceived ) problem of bias , and there is the suggestion to abandon time - tested estimation procedures in favor of linear inversion . here , we discuss the pros and cons of using biased and unbiased estimators for quantum state tomography .
we conclude that the little occasional benefit from the unbiased linear - inversion estimation does not justify the high price of using unphysical estimators , which are typically the case in that scheme . |
geometrically frustrated antiferromagnets have been a topic of significant interest due to their abundant and exotic physical properties.@xcite the rare - earth titanates @xmath1ti@xmath2o@xmath3 ( @xmath7 rare earth ) are one of the families with three - dimensional spin frustration.@xcite these materials have pyrochlore crystal structure with the space group @xmath8 , in which the magnetic rare - earth ions form a network of corner - sharing tetrahedra and are prone to a high degree of geometric frustration . for this reason ,
antiferromagnetically coupled classical heisenberg spins on the pyrochlore lattice can not form a long - range order at finite temperature .
however , the real materials are greatly sensitive to weak perturbations ( e.g. single - ion anisotropy , dipolar interaction or quantum fluctuations ) beyond the nearest - neighboring exchange , which results in rich unconventional low - temperature magnetic and thermodynamic properties .
example phenomena include the disordered classical spin ice in ho@xmath2ti@xmath2o@xmath3 and dy@xmath2ti@xmath2o@xmath3 ( with effective ferromagnetic exchange and ising anisotropy),@xcite the quantum spin liquid in tb@xmath2ti@xmath2o@xmath3,@xcite and the quantum spin ice in yb@xmath2ti@xmath2o@xmath3.@xcite furthermore , the long - range antiferromagnetic ( af ) order can be formed in gd@xmath2ti@xmath2o@xmath3 and er@xmath2ti@xmath2o@xmath3.@xcite recently , the low-@xmath5 heat transport properties of @xmath1ti@xmath2o@xmath3 have started to attract research interests for the purpose of detecting the transport of magnetic excitations in these materials .
the spin - ice compounds exhibit considerably strong interactions between the magnetic excitations and phonons and hence show drastic changes of thermal conductivity at the field - induced magnetic transitions.@xcite in the spin - liquid material tb@xmath2ti@xmath2o@xmath3 , the coupling between spin fluctuations and phonons is so strong that the phonon heat transport is extremely weak.@xcite in this work , we study the low-@xmath5 thermal conductivity of two long - range ordered systems @xmath1ti@xmath2o@xmath3 ( @xmath7 gd and er ) to probe the roles of magnon excitations and magnetic fluctuations . in gd@xmath2ti@xmath2o@xmath3
, the magnetic gd@xmath9 ion has a spin momentum @xmath10 7/2 and an orbital momentum @xmath11 0 , and therefore has a negligible single - ion anisotropy.@xcite in zero field , gd@xmath2ti@xmath2o@xmath3 shows the magnetic - order transition induced by thermal fluctuation at about 1 k to the so - called @xmath12 state.@xcite lowering temperature to about 0.7 k , the @xmath12 state changes to a so - called @xmath13 state .
it was found that both @xmath12 and @xmath13 states have a specific 4-@xmath14 magnetic structure with * k * = ( 1/2 , 1/2 , 1/2 ) , in which all the gd@xmath9 spins are perpendicular to the local [ 111 ] axes.@xcite however , in the @xmath13 state each pyrochlore unit cell consists of three fully ordered spin tetrahedra and one weakly ordered spin tetrahedron , while in the @xmath12 state that weakly ordered spin tetrahedron becomes fully disordered.@xcite with applying magnetic field , the gd@xmath9 spin structures present further changes.@xcite at @xmath15 3 t and @xmath16 1 k , a collinear spin state is formed , in which three spins on each tetrahedron are aligned with the field and the rest one points to the opposite direction.@xcite in high magnetic fields , the gd@xmath9 spins enter a polarized paramagnetic state .
the mechanisms of these field - driven transitions are likely that magnetic field can break the degenerate ground states and force the system to select a particular ordered structure .
er@xmath2ti@xmath2o@xmath3 was proposed to be a realization of the xy antiferromagnet on pyrochlore lattice .
er@xmath9 ion has a large orbital momentum @xmath17 = 6 and a spin momentum @xmath18 = 3/2 .
however , the crystal - field ground state of er@xmath9 ion is a kramers doublet , with large anisotropy respective to the local [ 111 ] axis , and can be described as an effective spin of 1/2.@xcite this implies a significant quantum effect in this material .
er@xmath2ti@xmath2o@xmath3 undergoes a second - order transition to a long - range ordered state at about 1.2 k with an unusual * k * = 0 non - coplanar af structure , the so - called @xmath19 state.@xcite the mechanism of this long - range ordered phenomenon had been discussed to be the quantum order by disorder effect,@xcite for which the quantum fluctuations lift the degeneracy of the ground state . applying magnetic field above a critical value of 1.52 t ( slightly different for @xmath20 [ 100 ] , [ 110 ] , and [ 111 ] ) , the system enters a high - field quantum paramagnetic state through a second - order transition driven by the quantum fluctuations . in the high - field state ,
the er@xmath9 spins are fully polarized in the xy plane and maximize their projections along the field direction.@xcite both gd@xmath2ti@xmath2o@xmath3 and er@xmath2ti@xmath2o@xmath3 exhibit the ground states of a coexistence of the long - range order and short - range order or spin fluctuations,@xcite which are obviously different from the conventional magnetic materials .
one can naturally expect that their heat transport properties would show some peculiar phenomena . in this paper
, we perform detailed studies of the low-@xmath5 thermal conductivity of @xmath1ti@xmath2o@xmath3 ( @xmath4 = gd and er ) single crystals and find that they have rather complicated temperature and magnetic - field dependencies . to probe the mechanism of the heat transport properties , the low-@xmath5 specific heat are also studied .
the data demonstrate that the magnetic excitations in these materials can effectively scatter phonons instead of transport heat . due to the spin - phonon coupling ,
the field - induced phase transitions cause drastic changes of the phonon thermal conductivity .
high - quality @xmath1ti@xmath2o@xmath3 ( @xmath4 = gd and er ) single crystals were grown using a floating - zone technique.@xcite in particular , gd@xmath2ti@xmath2o@xmath3 and er@xmath2ti@xmath2o@xmath3 were grown in 0.4 mpa pure oxygen with a growth rate of 2.5 mm / h and in 0.1 mpa pure oxygen with a rate of 4 mm / h , respectively .
the crystals were oriented by the x - ray laue photograph and cut precisely along those crystallographic axes like [ 111 ] , [ 110 ] ( or [ 1@xmath210 ] ) , and [ 11@xmath22 into long - bar and thin - plate shapes for the thermal conductivity and specific heat measurements , respectively .
the thermal conductivity was measured at low temperatures down to 0.3 k and in magnetic fields up to 14 t by using a conventional steady - state technique.@xcite the specific heat was measured at low temperatures down to 0.4 k by using the relaxation method in a commercial physical property measurement system ( ppms , quantum design ) .
ti@xmath2o@xmath3 single crystals for @xmath23 [ 110 ] and @xmath20 [ 111 ] .
( b , d ) zoom - in of the low - temperature plots .
the sample sizes are [email protected]@xmath240.17 mm@xmath25 for @xmath20 [ 110 ] and [email protected]@xmath240.09 mm@xmath25 for @xmath20 [ 111 ] , respectively .
the magnetic fields were applied along the shortest dimensions of the samples.,width=321 ] figure 1 shows the low-@xmath5 specific heat data of gd@xmath2ti@xmath2o@xmath3 single crystals for @xmath20 [ 110 ] and @xmath20 [ 111 ] . in zero magnetic field , there are two peaks at 1.0 and 0.71 k , which represent two consecutive transitions from paramagnetic state to the @xmath12 state and then to the @xmath13 state , respectively . in applied fields along the [ 110 ] direction ,
the two peaks merge with each other and a broad peak can be seen at 1 t. increasing field to 2 t , the peak becomes much larger and a bit sharper and is locating at 0.96 k. however , this peak splits into a peak at 0.95 k and a lower shoulder at 0.77 k when magnetic field is increased to 3 t. increasing field further to 4 t , the peak shifts to 0.92 k while the shoulder - like feature disappears . for @xmath26 4 t ,
the remained peak gradually moves to lower temperatures and can not be seen at temperature down to 0.4 k when the field is larger than 6 t. finally , a hump - like feature appears and gradually shifts to high temperatures for @xmath26 6 t. it resembles a schottky anomaly and is apparently related to the zeeman effect of the gd@xmath9 moments in the high - field paramagnetic state .
all these phenomena are essentially consistent with the results reported previously,@xcite which were discussed to originate from the partial lift of degenerate states.@xcite the specific - heat data in @xmath20 [ 111 ] show rather similar behaviors to those in @xmath20 [ 110 ] , but with some obvious differences in the positions and magnitudes of those peaks , which indicates a non - negligible magnetic anisotropy of gd@xmath2ti@xmath2o@xmath3.@xcite figures 2(a)-2(c ) show the thermal conductivity as a function of temperature in zero and 14 t fields for different directions of magnetic field and heat current ( @xmath27 ) .
apparently , the heat transport of gd@xmath2ti@xmath2o@xmath3 is almost isotropic in both zero and 14 t fields and is independent of the field direction .
the zero - field @xmath28 curves exhibit a typical phonon transport behavior at high temperatures and large phonon peaks at about 10 k.@xcite it is notable that at the phase transitions from the high-@xmath5 paramagnetic state to the low-@xmath5 magnetically ordered state at 1.0 and 0.7 k , indicated by the specific - heat data , there is no obvious change of @xmath28 .
it seems that the magnetic excitations of gd@xmath2ti@xmath2o@xmath3 have no obvious contribution to transporting heat .
however , at low temperatures the @xmath0 is increased in 14 t , indicating that there is magnetic scattering on phonons in zero field that can be removed by applying high field.@xcite the detailed magnetic - field dependencies of the low-@xmath5 thermal conductivities are shown in figs .
2(d)-2(f ) .
for three different configurations , @xmath6 isotherms show very similar behaviors . at very low temperatures
, the @xmath0 shows some reduction at low fields , a step - like increase at about 6 t , and a high - field plateau . upon increasing temperature ,
the low - field reduction is weakened and the transition at 6 t is broadened , accompanied with a decrease of the high - field plateau . at 1.95
k , the @xmath0 is almost independent of field . at lowest temperature ,
the high - field plateau is started from 78 t. from the specific - heat data , it is known that above 8 t the samples enter the paramagnetic or the high - field polarized state .
therefore , the step - like increase and high - field plateau of @xmath0 are apparently originated from the quench of the phonon scattering by magnetic excitations ( magnons from the ordered state ) , when entering the spin - polarized state.@xcite this means that the low-@xmath5 thermal conductivity in high fields would be a phonon transport free from magnetic scattering .
however , one may note that at sub - kelvin temperatures the 14-t @xmath28 curves have some obvious curvatures in the log - log plot instead of showing a power law close to @xmath29 . it is deviated from the expectation of a phonon heat transport in the boundary scattering limit.@xcite this unusual temperature dependence points to some peculiarity of the high - field magnetic state of gd@xmath2ti@xmath2o@xmath3 .
nevertheless , it is rather clear that magnons in this material mainly play a role of scattering phonons and there is no signature that they can contribute substantially to transporting heat .
it can also be seen that the low - field behaviors of @xmath6 are a bit different for three different configurations of heat current vs field direction . for @xmath20 [ 111 ] , as shown in fig .
2(d ) , the 0.36-k @xmath6 curve has two pronounced dips " at 1 and 3.25 t. with increasing temperature , they become more shallow and disappear above 1 k. for @xmath20 [ 110 ] ( or @xmath30 $ ] ) , as shown in figs .
2(e ) and 2(f ) , the 0.36-k @xmath6 curves show two weak dips " , which are slightly different between @xmath31 [ 111 ] and @xmath31 [ 110 ] .
in the former case , they locate at about 1 and 3.5 t , while in the latter case they locate at about 1 and 3 t ( the 1 t dip in fig .
2(f ) is very weak ) .
note that in these cases , the relative directions between the heat current and magnetic field are different , which yields some difference in the demagnetization effect . considering this factor , the positions of two dips depend on the direction of magnetic field rather than that of heat current .
these two dips disappear quickly with increasing temperature .
it should be noted that these low-@xmath5 @xmath6 isotherms do not display obvious hysteresis , which indicates that the possible magnetic domains@xcite are not coupled with the phonon transport . in this regard ,
the @xmath6 results are not able to distinguish the order of phase transition at 6 t , which either a weak first - order transition or a second - order one were suggested by some earlier experimental results.@xcite ti@xmath2o@xmath3 single crystals obtained from the specific - heat data for @xmath20 [ 110 ] and @xmath20 [ 111 ] .
the characteristic fields of the @xmath6 curves are also plotted for the comparison .
@xmath32 is the high - temperature paramagnetic state .
the @xmath12 , @xmath13 , and collinear states are three different phases mentioned in the main text.,width=226 ] to probe the origin of the @xmath6 behaviors , we compare the characteristic fields , including the dips and step - like increase , with those transitions from the specific - heat data . in fig .
3 , the @xmath33 phase diagrams of gd@xmath2ti@xmath2o@xmath3 for either @xmath20 [ 111 ] or @xmath20 [ 110 ] are determined by the transitions from the specific heat .
they are nearly the same as those in some earlier literature.@xcite it is known that there are at least four different magnetic phases in the phase diagram , that is , the high-@xmath5 and high - field paramagnetic ( pm ) phase , the low-@xmath5 and low - field @xmath13 state , a lower - field @xmath12 state , and a higher - field collinear state ( see the introduction section ) .
it is found that the characteristic fields from the @xmath6 data have a good correspondence with the phase boundaries among these four phases .
the step - like increase of @xmath0 at about 6 t is caused by the transition from the low - field ordered state to the paramagnetic state.@xcite it can be explained as the elimination of the magnon scattering on phonons in a spin - polarized state .
the dip of @xmath6 at about 3 t is caused by the transition from the @xmath13 state to the collinear state .
it seems that this transition is a rearrangement of the spin structure , which is usually associated with the collapse of the spin anisotropy gap and the increased numbers of low - energy magnons .
the phonon scattering is enhanced by a large number of magnons at the transition fields.@xcite another dip of @xmath0 at about 1 t , however , has no direct relationship to the specific - heat data . in the phase diagrams determined by the specific heat , this transition is locating deeply in the @xmath13 state . in this regard , it may be related to a very recent experimental finding of a new phase boundary at 1 t for @xmath20 [ 100].@xcite this boundary separates some unknown phases ( named as @xmath34 and @xmath35 in ref . ) from the @xmath13 state . since the phase diagrams of gd@xmath2ti@xmath2o@xmath3 are rather similar for different field directions,@xcite it is reasonable to believe that these newly found phases and phase boundaries exist also in the cases of @xmath20 [ 111 ] and @xmath20 [ 110 ] .
finally , it should be pointed out that the small differences in the critical fields between the @xmath6 data and the specific - heat data are probably due to either the uncertainty in defining the transition fields in two different physical properties or the difference in the demagnetization effect , or both .
ti@xmath2o@xmath3 single crystal for magnetic field along the [ 111 ] axis .
inset : zoom - in of the low - temperature plot .
the sample size is [email protected]@xmath240.08 mm@xmath25 , with the field direction along the shortest dimension .
( b ) the fits of the high - field data ( @xmath36 2 t ) to eq .
( [ sh ] ) as described in the main text .
the solid lines represents the fitting results and the dashed line shows the phonon contribution .
inset displays the fitting parameter @xmath37.,width=226 ] figure 4(a ) shows the low-@xmath5 specific heat of er@xmath2ti@xmath2o@xmath3 single crystal . in zero field
, a sharp @xmath38 peak appears at about 1.2 k , which reveals a second - order phase transition from the paramagnetic state to the long - range ordered state . applying magnetic field along the [ 111 ] direction ,
the peak is gradually suppressed and shifts to lower temperatures .
it is not observable at temperatures down to 0.4 k with @xmath36 2 t. at the same time , a much broader peak emerges at about 1.3 k when @xmath39 1.5 t. it shifts to higher temperatures and becomes even broader with increasing field , which behaves like a schottky anomaly.@xcite all these are in good agreement with the previous results.@xcite in literatures , the specific heat of er@xmath2ti@xmath2o@xmath3 were studied for three different directions of magnetic field , that is , @xmath23 [ 100 ] , [ 110 ] , and [ 111 ] , and were found to show some anisotropy of the transition fields between the ground state and the high - field quantum paramagnetic state . furthermore , the high - field specific heat were dependent on the field direction .
in particular , the high - field humps or broad peaks in the [ 100 ] and [ 110 ] fields could be described by a simple two - level schottky anomaly , while those in the [ 111 ] fields could not.@xcite our present data confirmed this .
as shown in fig .
4(b ) , the high - field data ( @xmath36 2 t ) are fitted using the following formula @xmath40 where the first term represents the two - level schottky anomaly and the second one is the phonon contribution . here
, @xmath4 is the universal gas constant , @xmath41 is a numerical coefficient , @xmath37 is the gap value and is dependent on the zeeman effect , and @xmath42 is the coefficient for low-@xmath5 phonon specific heat .
it is found that this formula could qualitatively describe the data but the fittings are not perfect , which is very similar to the earlier study by sosin _
et al._@xcite note that the parameters in our fittings , @xmath41 = 1.65 and @xmath42 = 2.8 @xmath43 j / k@xmath44mol , are also almost the same as those in ref . .
all these indicate that for @xmath23 [ 111 ] , higher - field schottky - like peaks are not a simple behavior of paramagnetic moments . some soft modes related to the spin fluctuations
are probably involved.@xcite in a theoretical scenario , er@xmath2ti@xmath2o@xmath3 was found to have a ground state of quantum order by disorder and exhibit a quantum critical point at 1.52 t , where the long - range order is suppressed.@xcite however , the zero - field long - range order actually coexists with spin fluctuations , as revealed by the magnetic resonance spectroscopy and inelastic neutron scattering.@xcite these soft spin excitations would not only contribute to the specific heat but also be easily coupled with low - energy phonons and significantly damp the phonon transport.@xcite across the quantum critical field ( 1.52 t ) , the magnetic excitations of ground state change drastically , which in principle could be detected by the thermal conductivity .
figure 5 shows the temperature and magnetic - field dependencies of the thermal conductivity for different directions of heat current and magnetic field .
both the @xmath28 curves and the low-@xmath5 @xmath6 isotherms show no obvious anisotropy .
as far as the zero - field @xmath28 data are concerned , the phonon peaks locate at about @xmath45 k ; below about 1.2 k , as shown in the inset to fig .
5(c ) , some slight changes in the slopes of these @xmath28 can be seen , which is related to the phase transition from the paramagnetic state to the long - range ordered state . with applying field of 1.5 or 2 t
, the @xmath0 become much smaller than the zero - field data and the changes of the @xmath28 slopes at @xmath46 1.2 k are more significant . in a high field of 14 t ,
the @xmath0 values are larger than the zero - field values at low temperatures , which indicates that in the spin - polarized state , the phonon scattering by magnetic excitations is quenched and the phonon transport is recovered . however ,
at relatively higher temperatures , the @xmath0 in 14 t field is much smaller than the zero - field value , suggesting a field - induced magnetic scattering effect .
this is different from the high-@xmath5 behavior of gd@xmath2ti@xmath2o@xmath3 .
the @xmath6 isotherms show a dip - like behavior at low fields and an increase at high fields , particularly at very low temperatures . at the first glance , it is similar to that caused by paramagnetic scattering on phonons.@xcite but one may note a remarkable difference between the er@xmath2ti@xmath2o@xmath3 results and the paramagnetic scattering phenomenon . in the latter case ,
the minimum or dip of the @xmath6 isotherm is due to a phonon resonant scattering by some magnetic excitations between the energy levels that are affected by the zeeman effect;@xcite therefore , the position of the dip should shift almost linearly with the magnetic field . for er@xmath2ti@xmath2o@xmath3 , at higher temperatures ( @xmath47 1 k ) , the dip of @xmath0 becomes broader and shifts to higher fields with increasing temperature , which is indeed the same as the paramagnetic scattering effect .
however , as can be seen in figs .
5(d)-5(f ) , the dip fields are temperature independent for @xmath16 1 k. apparently , this sub - kelvin behavior must have some other origin .
since the low-@xmath5 dip fields ( 1.5 or 2 t ) are close to the characteristic field in the specific heat , above which the @xmath38 peak was suppressed , it is likely that they correspond to the phase transition from the low - field ordered state to the high - field quantum paramagnetic state . in passing ,
it should be pointed out that the dip field is essentially isotropic for different field directions , as shown by figs .
5(d ) and 5(f ) , while the higher dip fields in fig .
5(e ) could be due to the demagnetization effect .
in addition , the @xmath6 curves at 0.36 and 0.5 k in fig .
5(f ) display shallow minimums at high field ( @xmath47 10 t ) .
it is unclear whether they are related to some kind of transition since the feature is rather weak and broad .
ti@xmath2o@xmath3 single crystal for @xmath20 [ 111 ] from the specific - heat data .
the low-@xmath5 @xmath38 peak gives the antiferromagnetic transition temperature @xmath48 .
the field dependence of @xmath48 displays the phase boundary of the long - range ordered ( lro ) ground state , as shown by the dashed line .
the higher - field schottky - like peaks may separate the
high-@xmath5 , low - field " thermally paramagnetic state and the high - field , low-@xmath5 " quantum paramagnetic state , as suggested by ref . .
the characteristic dip fields of the @xmath6 isotherms are also shown for three configurations of the heat current ( @xmath27 ) and magnetic field.,width=264 ] figure 6 shows the comparison of dip fields of the @xmath6 with the characteristic fields of the specific heat .
it is clear that the dips of @xmath6 have no good correspondence with the phase boundary of the af state , and at different temperature regions they have very different origins . at very low temperatures ( @xmath16 1 k )
, the dips are coincided with the quantum critical field that suppresses the long - range - ordered ground state , which can be ascribed to be a simple explanation of the strong phonon scattering by the critical magnetic fluctuations .
this is supported by a complete mode softening of spin excitations at @xmath49 as revealed by the inelastic neutron scattering.@xcite the higher - field enhancement of @xmath0 is due to the removal of low - lying magnetic excitations at @xmath50.@xcite on the other hand , the higher-@xmath5 ( @xmath51 1 k ) dips have a good correspondence with the higher - field schottky - like anomalies of the specific heat .
although this anomaly was discussed to separate the
high-@xmath5 , low - field " thermally paramagnetic state and the high - field , low-@xmath5 " quantum paramagnetic state,@xcite the paramagnetic excitations are the main contribution and are likely playing a role of scattering phonon .
therefore , it is understandable that the @xmath6 behaviors at @xmath51 1 k match well with the phonon resonant scattering effect by the paramagnetic ions .
both gd@xmath2ti@xmath2o@xmath3 and er@xmath2ti@xmath2o@xmath3 display the magnetic excitations scattering on phonons .
it seems to be a common phenomenon in pyrochlore @xmath1ti@xmath2o@xmath3 . in spin - ice members
dy@xmath2ti@xmath2o@xmath3 and ho@xmath2ti@xmath2o@xmath3 , there is negligible scattering on phonons by the magnetic excitations in zero field.@xcite this is due to the fact that the peculiar magnetic excitations , magnetic monopoles , in the spin - ice compounds have sizeable energy gaps that prevent the thermal activations at very low temperatures.@xcite however , in magnetic fields the magnetic monopoles can be well populated and were found to scatter phonons rather strongly , particularly at the field - induced magnetic transitions between the spin - ice " , the kagom - ice " , and the three - in - one - out " states , although they might be able to transport heat also.@xcite in the quantum - spin - liquid material tb@xmath2ti@xmath2o@xmath3 , the coupling between spin fluctuations and phonons is so strong that the phonon heat transport exhibits a glassy - like behavior.@xcite in comparison , the coupling or scattering effect between magnetic excitations and phonons in the long - range ordered gd@xmath2ti@xmath2o@xmath3 and er@xmath2ti@xmath2o@xmath3 are not very strong but detectable at zero field .
one reason that the magnetic excitations act mainly as phonon scatterers rather than heat carriers in @xmath1ti@xmath2o@xmath3 is apparently related to the rather weak exchange interactions in these materials . as a result ,
the magnetic excitations have weak dispersion or small velocity .
an example is that the magnetic monopoles in the spin - ice compounds were estimated to have a small velocity of @xmath46 20 m / s.@xcite the magnon velocity of er@xmath2ti@xmath2o@xmath3 , calculated from the dispersion curve , is about 80 m / s only.@xcite gd@xmath2ti@xmath2o@xmath3 is expected to have a similar situation , since the exchange energies of these materials are essentially comparable.@xcite in contrast , the magnon velocity is about 24 orders of magnitude larger in many materials that show substantial magnon heat transport.@xcite although the heat transport of gd@xmath2ti@xmath2o@xmath3 and er@xmath2ti@xmath2o@xmath3 have some similarities in the weak anisotropy and magnetic scattering effect , there are several notable differences between them .
first , gd@xmath2ti@xmath2o@xmath3 shows much larger phonon peaks in the zero - field @xmath28 curves , even though these two compounds have the same crystal structures . for nonmagnetic insulators ,
larger phonon peak may indicate less structural imperfections in crystals .
however , for magnetic materials , the magnetic scattering can play a role .
this possibility seems to be supported by the details of the temperature dependence of @xmath0 . as can be seen in figs . 2 and 5 , near the long - range order transition in zero field
, the @xmath28 of gd@xmath2ti@xmath2o@xmath3 do not show any observable changes , while those of er@xmath2ti@xmath2o@xmath3 show weak changes of the @xmath28 slopes .
second , the magnetic scatterings actually have quite different origins in these two materials . in gd@xmath2ti@xmath2o@xmath3 , the magnetic excitations of the long - range ordered state , _
i.e. _ , magnons are effectively phonon scatterers .
for this reason , the magnetic - field - induced changes of @xmath0 are only presented at low temperatures ; above 2 k , the @xmath0 is nearly unaffected by applying field . in er@xmath2ti@xmath2o@xmath3 , however , the role of magnons in the long - range ordered state is not so obvious . nevertheless , in zero field and at very low temperatures , magnons are likely to scatter phonons , since the high - field @xmath0 are about 50% larger than the zero - field values . on the other hand , the low - energy spin excitations associated with the quantum critical point
induce a strong scattering on phonons in a rather broad temperature range .
another contribution of the magnetic excitations is the high-@xmath5 paramagnetic scattering effect . in this regard ,
it is rather difficult to understand why there is no any signature of paramagnetic scattering effect in gd@xmath2ti@xmath2o@xmath3 .
third , gd@xmath2ti@xmath2o@xmath3 shows much more complicated @xmath6 behaviors than er@xmath2ti@xmath2o@xmath3 .
this is clearly due to a more complicated low-@xmath5 phase diagram of gd@xmath2ti@xmath2o@xmath3 . in present work
, the low-@xmath5 heat transport of these materials again shows a close relationship to the magnetism and magnetic transitions , as also found in many other magnetic materials.@xcite it therefore provides a useful tool to probe some unknown magnetic transitions , as shown in fig .
3 . there are also some details remained to be further studied . as shown in fig .
2 , the temperature dependencies of @xmath0 of gd@xmath2ti@xmath2o@xmath3 show some clear deviations from the power low at sub - kelvin temperatures even in 14 t field . this may suggest that the magnetic scattering in the high - field spin - polarized state is still active at such low temperatures .
this unusual temperature dependence points to some peculiarity of the high - field magnetic state of gd@xmath2ti@xmath2o@xmath3 .
one possible reason is that the ground state of gd@xmath2ti@xmath2o@xmath3 is a coexisting of long - range and short - ranges orders .
the short - range order or spin fluctuations could survive in high field even when the long - range af order was suppressed .
a supportive experimental result is that the magnetic resonance spectroscopy has detected some weakly dispersive soft modes in the spin - saturated phase , which could be explained based on the spin - wave calculations.@xcite
thermal conductivities of pyrochlore @xmath1ti@xmath2o@xmath3 ( @xmath4 = gd and er ) are studied at low temperatures down to 0.3 k and in high fields up to 14 t. it is found that the magnetic excitations play a role of scattering phonons rather than transporting heat , although these two materials have long - range magnetic orders at low temperatures . in gd@xmath2ti@xmath2o@xmath3
, the field - induced magnetic transitions can cause drastic changes of @xmath0 at the phase boundaries and lead to rather complicated field dependencies of @xmath0 .
the main phenomena include the dip - like features in @xmath6 isotherms at the low - field phase transitions and a step - like increase of @xmath0 at the spin - polarization transition .
er@xmath2ti@xmath2o@xmath3 shows much simpler field dependencies of @xmath0 , with a dip - like feature caused by the quantum phase transition at about 1.52 t or the paramagnetic phonon scattering at higher temperatures .
these data demonstrate that heat transport is an effective means to probe the magnetic properties of pyrochlore rare - earth titanates .
this work was supported by the national natural science foundation of china , the national basic research program of china ( grants no .
2009cb929502 and no .
2011cba00111 ) , and the fundamental research funds for the central universities ( program no .
wk2340000035 ) .
j. s. gardner , s. r. dunsiger , b. d. gaulin , m. j. p. gingras , j. e. greedan , r. f. kiefl , m. d. lumsden , w. a. macfarlane , n. p. raju , j. e. sonier , i. swainson , and z. tun , phys .
* 82 * , 1012 ( 1999 ) .
j. d. m. champion , m. j. harris , p. c. w. holdsworth , a. s. wills , g. balakrishnan , s. t. bramwell , e.cizmr , t. fennell , j. s. gardner , j. lago , d. f. mcmorrow , m. orendc , a. orendcov , d. mck .
paul , r. i. smith , m. t. f. telling , and a. wildes , phys .
b * 68 * , 020401(r ) ( 2003 ) . | we report a systematic study on the low - temperature thermal conductivity ( @xmath0 ) of @xmath1ti@xmath2o@xmath3 ( @xmath4 = gd and er ) single crystals with different directions of magnetic field and heat current .
it is found that the magnetic excitations mainly act as phonon scatterers rather than heat carriers , although these two materials have long - range magnetic orders at low temperatures .
the low-@xmath5 @xmath6 isotherms of both compounds show rather complicated behaviors and have good correspondences with the magnetic transitions , where the @xmath6 curves show drastic dip- or step - like changes . in comparison ,
the field dependencies of @xmath0 are more complicated in gd@xmath2ti@xmath2o@xmath3 , due to the complexity of its low-@xmath5 phase diagram and field - induced magnetic transitions .
these results demonstrate the significant coupling between spins and phonons in these materials and the ability of heat - transport properties probing the magnetic transitions . |
we consider solving the laplace - beltrami problem on a smooth two dimensional surface embedded into a three dimensional space partitioned into a mesh consisting of shape regular tetrahedra .
the mesh does not respect the surface and thus the surface cuts through the elements .
following olshanskii , reusken , and grande @xcite we construct a galerkin method by using the restrictions of continuous piecewise linears defined on the tetrahedra to the surface . the resulting discrete method may be severely ill - conditioned and the main purpose of this paper is to suggest a remedy for this problem based on adding a consistent stabilization term to the original bilinear form .
the stabilization term we consider here controls jumps in the normal gradient on the faces of the tetrahedra and provides a certain control of the derivative of the discrete functions in the direction normal to the surface .
similar terms have recently been used for stabilization of cut finite element methods for fictitious domain methods @xcite , @xcite , @xcite , and @xcite . note
that none of these references involve any partial differential equations on surfaces only regular boundary and interface conditions . in principle , it is possible , in this situation , to deal with the ill conditioning problem in the linear algebra using a scaling , see @xcite .
starting from a stable method has clear advantages in more complex applications that may need stabilization anyway , such as problems with hyperbolic character or coupled bulk - surface problems .
it is also not clear that matrix based algebraic scaling procedures is possible in all situations and thus alternative approaches must be investigated . using the additional stability we first prove an optimal estimate for the condition number , independent of the location of the surface , in terms of the mesh size of the underlying tetrahedra .
the key step in the proof is certain discrete poincar estimates that are also of general interest .
then we prove a priori error estimates in the energy and @xmath0 norms . in a companion paper
, we will consider the more challenging problems of the surface helmholtz equation and show error estimates for a stabilized method under a suitable condition on the product of the mesh size and the wave number .
finally , we refer to @xcite , @xcite , @xcite , and @xcite for general background on finite element methods for partial differential equations on surfaces .
the outline of the reminder of this paper is as follows : in section 2 we formulate the model problem and the finite element method , in section 3 we summarize some preliminary results involving lifting of functions from the discrete surface to the continuous surface , in section 4 we prove an optimal bound on the condition number of the stabilized method , in section 5 we prove a priori error estimates in the energy and @xmath0 norms , and finally in section 6 we present numerical investigations confirming our theoretical results .
let @xmath1 be a smooth @xmath2-dimensional closed surface embedded in @xmath3 , @xmath4 or @xmath5 , with signed distance function @xmath6 such that the exterior surface unit normal is given by @xmath7 .
let @xmath8 be the nearest point projection mapping onto @xmath1 , i.e. , @xmath8 is the point on @xmath1 that minimizes the euclidian distance to @xmath9 . for @xmath10 let @xmath11 be the tubular neighborhood @xmath12 of @xmath1 .
then @xmath13 and there is a @xmath14 such that for each @xmath15 there is a unique @xmath16 .
using @xmath17 we may extend any function @xmath18 defined on @xmath1 to @xmath19 by defining @xmath20 we consider the following problem : find @xmath21 such that @xmath22 where @xmath23 is a given function such that @xmath24 . here
@xmath25 is the laplace - beltrami operator defined by @xmath26 where @xmath27 is the tangent gradient @xmath28 with @xmath29 the projection of @xmath30 onto the tangent plane of @xmath1 at @xmath31 , defined by @xmath32 where @xmath33 is the identity matrix , and @xmath34 the @xmath35 gradient .
let @xmath36 and @xmath37 be the @xmath38 inner product and norm on the set @xmath39 equipped with the appropriate measure .
let @xmath40 be the sobolev spaces on @xmath1 with norm @xmath41 where @xmath42 and the @xmath0 norm for a matrix is based on the pointwise frobenius norm .
we then have the weak problem : find @xmath43 such that @xmath44 where @xmath45 it follows from the lax - milgram lemma that the weak problem has a unique solution for @xmath46 such that @xmath24 . for smooth surfaces we also have the elliptic regularity estimate @xmath47 here and below @xmath48 denotes less or equal up to a positive constant .
let @xmath49 be a quasi uniform partition into shape regular tetrahedra for @xmath50 and triangles for @xmath4 with mesh parameter @xmath51 of a polygonal domain @xmath52 in @xmath53 completely containing @xmath19 .
let @xmath54 be the space of continuous piecewise linear polynomials defined on @xmath49 .
let @xmath55 be an approximation of the distance function @xmath6 and let @xmath56 be the zero levelset @xmath57 then @xmath56 is piecewise linear and we define the exterior normal @xmath58 to be the exact exterior unit normal to @xmath56 .
we consider a family of such surfaces @xmath59 such that ( a ) @xmath60 , ( b ) the closest point mapping @xmath61 , is a bijection , and ( c ) the following estimates hold @xmath62 for @xmath63 .
these properties are , for instance , satisfied if @xmath64 is the lagrange interpolant of @xmath6 and @xmath65 is small enough .
let @xmath66 and @xmath67 be the continuous piecewise linear functions defined on @xmath68 with average zero .
the finite element method on @xmath56 takes the form : find @xmath69 such that @xmath70 here the bilinear form @xmath71 is defined by @xmath72 with @xmath73 and @xmath74,[{\boldsymbol n}_f \cdot \nabla w])_f\ ] ] where @xmath75 denotes the set of internal interfaces in @xmath68 , @xmath76 = ( { \boldsymbol n}_f \cdot \nabla v)^+ - ( { \boldsymbol n}_f \cdot \nabla v)^-$ ] with @xmath77 , is the jump in the normal gradient across the face @xmath78 , @xmath79 denotes a fixed unit normal to the face @xmath80 , and @xmath81 is a constant of @xmath82 .
the tangent gradients are defined using the normal to the discrete surface @xmath83 and the right hand side is given by @xmath84 introducing the mesh dependent norm @xmath85 where @xmath86 we note that @xmath87 is indeed a norm on @xmath88 ( for fixed @xmath56 ) since if @xmath89 we have @xmath90 . to see this
we first note that if @xmath91 then @xmath18 must be a linear polynomial @xmath92 with @xmath93 , @xmath94 , and @xmath95 the center of gravity of @xmath56 , here @xmath96 since @xmath97 , secondly if @xmath98 and thus @xmath99 since @xmath56 is a closed surface and thus @xmath100 can not be normal to @xmath56 everywhere .
thus it follows from the lax - milgram lemma that there exists a unique solution to ( [ eq : fem ] ) .
in this section we collect some essentially standard results , see @xcite , @xcite , and @xcite , related to lifting of functions from the discrete surface to the exact surface . for each @xmath101
we let @xmath102 be the distance function to the hyperplane , with normal @xmath103 , that contains @xmath104 and @xmath105 be the associated nearest point projection @xmath106 onto the hyperplane .
then we define the mapping @xmath107 where @xmath108 and we defined @xmath109 for @xmath15 .
we note that @xmath110 is an invertible mapping , @xmath111 , and @xmath112 is a partition of @xmath1 .
the derivative @xmath113 of @xmath110 is given by @xmath114 where we used the identity @xmath115 . here
@xmath116 and we have the identity @xmath117 where @xmath118 are the principal curvatures of @xmath1 with corresponding principal curvature vectors @xmath119 , see @xcite lemma 14.7 .
thus we note that @xmath120 is tangential to @xmath1 and that for @xmath121 small enough we have the estimate @xmath122 .
in particular , on @xmath56 we have @xmath123 and thus we obtain the simplified expression @xmath124 where we used the fact that @xmath120 is a tangential tensor to @xmath1 , i.e. @xmath125 . the mapping @xmath113 maps the tangent and normal spaces of @xmath56 at @xmath126 onto the tangent and normal spaces of @xmath1 at @xmath8 .
we define the lift @xmath127 of a function @xmath128 to @xmath129 by @xmath130 the tangent gradient of @xmath127 is given by @xmath131 using the fact that @xmath132 and @xmath133 preserves the normal and tangent directions we finally obtain the identity @xmath134 where we introduced the notation @xmath135 here @xmath132 is a symmetric matrix with eigenvalues @xmath136 , which are all strictly greater than zero on @xmath19 for @xmath121 small enough , and @xmath137 is nonsymmetric with singular values @xmath138 , which are also strictly greater than zero .
we will need the following estimates for @xmath139 : @xmath140 the first estimate follows directly from the definition of @xmath139 , the second can be proved using the the fact that eigenvalues and singular values discussed above are all strictly greater than zero . to prove the third we proceed as follows @xmath141 where we collected the terms involving the distance function @xmath6 in the last term and used the assumption ( [ eq : rhobounds ] ) that @xmath142 .
next for the remaining term we write @xmath143 and then we note that the identity @xmath144 holds . using the bound @xmath145
the estimate follows .
the surface measure @xmath146 on @xmath1 is related to the surface measure @xmath147 on @xmath56 by the identity @xmath148 where @xmath149 is the determinant of @xmath139 which is given by @xmath150 using this identity we obtain the estimate @xmath151 where we finally used the bound @xmath152 . in summary
we have the following estimates for the determinant @xmath153
in this section we derive several discrete poincar estimates .
we begin with the standard poincar inequality on @xmath56 for functions in @xmath154 with average zero and a constant uniform in @xmath51 for @xmath51 small enough .
then we show estimates that essentially quantifies the improved control of the solution and its gradient provided by the gradient jump stabilization term . in order to prepare for the proof of our main estimates lemma [ lemmacondb ] and lemma [ lemmacondbb ]
we first prove a poincar inequality for piecewise constant functions defined on @xmath68 in lemma [ lemmaconda ] and then in lemma [ lemmacondaa ] we quantify the improved control of the total gradient provided by the stabilization term .
the proof of lemma [ lemmaconda ] builds on the idea of using a covering of @xmath155 in terms of sets consisting of a uniformly bounded number of elements . on these sets a local poincar estimate holds for functions with local average zero .
the local averages can then be approximated by a smooth function for which a standard poincar estimate on the exact surface can finally be applied .
this approach is used in @xcite to prove korn s inequality in a tubular neighborhood of a smooth surface .
in contrast to @xcite our proof handles discrete functions and the fact that @xmath155 is a polygon that changes with the mesh size @xmath51 .
[ lemmapoincaresigmah ] let @xmath156 be the average @xmath157
. then the following estimate holds @xmath158 for @xmath159 with @xmath65 small enough .
let @xmath160 be the average of @xmath161 .
using the fact that @xmath162 is the constant that minimizes @xmath163 and then changing coordinates to @xmath1 followed by the standard poincar estimate on @xmath1 we obtain @xmath164 where we mapped back to @xmath56 in the last step .
this concludes the proof .
[ lemmaconda ] let @xmath18 be a piecewise constant function on @xmath68 and let @xmath165 be the average on @xmath155
. then the following estimate holds @xmath166\|^2_f\ ] ] for @xmath159 with @xmath65 small enough .
let @xmath167 and let @xmath168 for @xmath169 .
next we let @xmath170 we note that there is a uniform bound @xmath171 on the number of elements in @xmath172 since the mesh is quasiuniform . furthermore
, if @xmath173 is a set of points on @xmath1 such that @xmath174 is a covering of @xmath1 then @xmath175 is a covering of @xmath155 .
next let @xmath176 be smooth , nonnegative , have compact support , and be constant equal to @xmath177 in a neighborhood of @xmath178 .
define @xmath179 and @xmath180 then @xmath181 and we have the estimates @xmath182 on all of the sets @xmath183 we have the local poincar estimate @xmath184\|_f^2\ ] ] where @xmath185 is the average of @xmath18 over @xmath183 and @xmath186 is the set of interior faces in @xmath172 .
we finally let @xmath187 be defined by @xmath188 with average @xmath189 with these definitions we have the estimates @xmath190\|_f^2 + \sum_{{\boldsymbol x}\in \mathcal{x}_h } h\| a_{{\boldsymbol x } } - a \|^2_{d_{h,{\boldsymbol x}}}\end{aligned}\ ] ] where we used the fact that @xmath162 is the constant that minimizes @xmath191 in ( [ eq : conda : aaa ] ) and the local poincar estimate ( [ eq : localpoincare ] ) to estimate the first term and the estimate @xmath192 together with the fact that @xmath193 is a constant to estimate the second term in ( [ eq : conda : bbb ] ) .
next we split the second term in ( [ eq : conda : ccc ] ) by adding and subtracting the constant @xmath194 and the function @xmath195 in each term in the sum as follows @xmath196 we proceed with estimates of terms @xmath197 .
[ [ term - boldsymbol - i ] ] term @xmath33 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + using the fact that @xmath198 is constant on @xmath183 , the bound @xmath199 , the definition of @xmath195 , and the identity @xmath200 , we obtain @xmath201\|_f^2 \end{aligned}\ ] ] where we used cauchy - schwarz in ( [ lemma : a : i : a ] ) , the bounds ( [ eq : varphibounds ] ) for @xmath202 in ( [ lemma : a : i : b ] ) , an element wise inverse inequality in ( [ lemma : a : i : c ] ) , and finally the local poincar estimate ( [ eq : localpoincare ] ) in ( [ lemma : a : i : d ] ) .
thus we have the estimate @xmath203\|_f^2\ ] ] [ [ term - boldsymbol - iboldsymbol - i ] ] term @xmath204 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + using the estimate @xmath205 followed by the fundamental theorem of calculus we obtain @xmath206 for each @xmath207 we have @xmath208 and thus we find that @xmath209\|_f^2\end{aligned}\ ] ] where we used the local poincar estimate ( [ eq : localpoincare ] ) . taking the supremum over @xmath210 we obtain @xmath211\|_f^2\ ] ] where @xmath212 is the set of interior faces contained in the set @xmath213 .
thus we conclude that @xmath214\|_f^2\ ] ] [ [ term - boldsymbol - iboldsymbol - iboldsymbol - i ] ] term @xmath215 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + using the standard poincar estimate on @xmath1 we obtain @xmath216\|_f^2\end{aligned}\ ] ] where we used ( [ eq : maxatilde ] ) .
starting from the bound ( [ eq : conda : ccc ] ) and using the bounds of terms @xmath217 and @xmath218 for the second term we arrive at @xmath219\|_f^2 \lesssim \sum_{f \in { \mathcal{f}}_{h } } h^{-1}\|[v]\|_f^2\ ] ] where , in the last inequality , we used the fact that there is a covering such that we have a uniform bound on the number of sets @xmath220 that each edge belongs to . to construct such a covering
we let @xmath221 and @xmath222 then @xmath223 and we note that @xmath224 .
thus @xmath225 is a covering of @xmath1 .
next we note that there is a constant @xmath226 such that @xmath227 since the mesh is quasiuniform . the covering number of @xmath228 is uniformly bounded by the number of points in the set @xmath229 .
[ lemmacondaa ] the following estimate holds @xmath230 for @xmath159 with @xmath65 small enough .
we have @xmath231 for all @xmath232 .
choosing @xmath100 such that @xmath233 @xmath234 , it follows from lemma [ lemmaconda ] that @xmath235 next , to estimate @xmath236 we first map to @xmath1 and then use the fact that @xmath1 is a closed surface followed by finite dimensionality of the constant functions to conclude that @xmath237 where we mapped back to the discrete surface @xmath56 and used the estimate @xmath238 .
finally , for @xmath159 with @xmath65 sufficiently small , a kick back argument leads to the estimate @xmath239 next , writing @xmath240 , we have @xmath241 combining the estimates ( [ eq : condaaa ] ) , ( [ eq : condaab ] ) , ( [ eq : condaac ] ) , and ( [ eq : condaad ] ) , we obtain the desired result .
we are now ready to state and prove our poincar inequality .
[ lemmacondb ] the following estimate holds @xmath242 for @xmath159 with @xmath65 small enough . using the same notation as in lemma [ lemmaconda ]
we first show that there is a constant @xmath226 such that for each @xmath169 and @xmath51 , @xmath159 , there exists an element @xmath243 such that @xmath244 we say that such an element has a large intersection with @xmath56 .
assume that there is no such element .
then there is a sequence @xmath245 such that @xmath246 since there is a uniform bound on the number of elements in @xmath172 we obtain the estimate @xmath247 which is a contradiction since @xmath248 since @xmath249 .
furthermore , the following estimate holds @xmath250 where we introduced the notation @xmath251\|_f^2\ ] ] to prove ( [ eq : condba ] ) , let @xmath252 and @xmath253 be two elements sharing a common face @xmath78 .
then we have the following identity @xmath254 { \boldsymbol n}_f \cdot ( { \boldsymbol x}-{\boldsymbol x}_f)\ ] ] where @xmath255 is the center of gravity of the face @xmath78 .
thus @xmath256 { \boldsymbol n}_f \cdot ( { \boldsymbol x}-{\boldsymbol x}_f)\|^2_{k_2 } \\
\label{condterm2 } & \lesssim \| v_1 \|^2_{k_1 } + h^{3 } \| [ { \boldsymbol n}_f \cdot \nabla v ] \|^2_f\end{aligned}\ ] ] iterating this bound and summing over all elements in @xmath172 we obtain @xmath257 here @xmath258 may be estimated using the inverse bound @xmath259 which holds due to quasiuniformity and the fact that the intersection with @xmath56 of @xmath260 satisfies @xmath261 . combining ( [ eq : condb : b ] ) and ( [ eq : condb : c ] ) we obtain ( [ eq : condba ] ) . using a covering @xmath262 of @xmath155 with
a uniformly bounded covering number as constructed in lemma [ lemmaconda ] , we obtain @xmath263 where we used the poincar inequality , see lemma [ lemmapoincaresigmah ] .
this concludes the proof .
we conclude this section with a version of the previous lemma which involves the @xmath0 norm instead of the energy norm .
[ lemmacondbb ] the following estimate holds @xmath264 for @xmath159 with @xmath65 small enough . starting from ( [ eq
: conda : b ] ) we have @xmath265 and thus we need to estimate @xmath266 .
using the same notation as above we obtain @xmath267 where we used the fact that @xmath260 has a large intersection with @xmath56 so that an inverse estimate holds for the tangential derivative and we also added and subtracted @xmath268 where @xmath269 is the tangential gradient to @xmath270 .
we proceed with estimates of @xmath271 and @xmath272 .
[ [ term - i_boldsymbol - x ] ] term @xmath271 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + using the definition of the tangential derivative we obtain the estimate @xmath273 where @xmath103 is the ( constant ) normal associated with element @xmath274 .
now @xmath275 since @xmath276 for any @xmath277 . using ( [ eq : rhobounds ] )
we conclude that the first two terms are @xmath278 and using the fundamental theorem of calculus we have the estimate @xmath279 for the third term .
thus we have the estimate @xmath280 which gives @xmath281 where we used an inverse estimate at last . [
[ term - ii_boldsymbol - x ] ] term @xmath272 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we let @xmath269 act on identity ( [ eq : condb : aa ] ) , which gives @xmath282 { \boldsymbol p}_{\sigma_{h,{\boldsymbol x } } } { \boldsymbol n}_f\ ] ] and thus we have the estimate @xmath283\|^2_f\ ] ] iterating this bound and summing over all the elements in @xmath172 , again using the fact that the number of elements in @xmath172 is uniformly bounded , we arrive at @xmath284 using ( [ eq : condbb : aa ] ) we obtain @xmath285 where we used the estimate @xmath286 which holds since element @xmath260 has a large intersection with @xmath56 .
collecting the estimates ( [ eq : condbb : c ] ) , ( [ eq : condbb : i ] ) , and ( [ eq : condbb : ii ] ) we have @xmath287 combining ( [ eq : condbb : a ] ) and ( [ eq : condbb : d ] ) yields @xmath288 and the lemma follows for @xmath63 , with @xmath65 small enough , using a kick back argument . here
we derive the inverse inequality needed in the proof of the condition number estimate .
[ lemmacondc ] the following estimate holds @xmath289 for @xmath159 with @xmath65 small enough . using the fact that @xmath290 is
constant we note that @xmath291 , which leads to the estimate @xmath292 where we used an element wise inverse inequality at last .
furthermore , using standard inverse inequalities , we obtain the following estimate for the jump term @xmath293 to derive an estimate of the condition number of the stiffness matrix we use the poincar inequality in lemma [ lemmacondb ] and the inverse estimate in lemma [ lemmacondc ] together with the approach in @xcite .
let @xmath294 be the standard piecewise linear basis functions associated with the nodes in @xmath68 and let @xmath295 be the stiffness matrix with elements @xmath296 .
we recall that the condition number is defined by @xmath297 where @xmath298 for @xmath299 and @xmath300 for @xmath301 .
the expansion @xmath302 defines an isomorphism that maps @xmath303 to @xmath304 and satisfies the following well known estimates @xmath305 the following estimate of the condition number of the stiffness matrix holds @xmath306 for @xmath159 with @xmath65 small enough .
we need to estimate @xmath307 and @xmath308 .
starting with @xmath307 we have @xmath309 where we used the estimate @xmath310 together with ( [ inverse ] ) and ( [ rneqv ] ) .
thus latexmath:[\[\label{aest } next we turn to the estimate of @xmath308 . using ( [ rneqv ] ) and
( [ poincare ] ) , we get @xmath312 and thus we conclude that @xmath313 . setting @xmath314 we obtain @xmath315 ) and ( [ ainvest ] ) of @xmath316 and @xmath308 the theorem follows .
in this section we derive a priori error estimates in the energy and @xmath0-norms .
the main technical difficulty is to handle the fact that the surface is approximated by a discrete surface .
our approach essentially follows @xcite , @xcite , and @xcite . in order to define an interpolation operator we note that the extension @xmath317 of @xmath318 satisfies the stability estimate @xmath319 with constant only dependent on the curvature of the surface @xmath1 .
we let @xmath320 denote the standard scott - zhang interpolation operator and recall the interpolation error estimate @xmath321 where @xmath322 is the union of the neighboring elements of @xmath274 .
we also define an interpolation operator @xmath323 as follows @xmath324 introducing the energy norm @xmath325 associated with the exact surface @xmath326 we have the following approximation property .
[ lem : approx ] the following estimate holds @xmath327 for @xmath159 with @xmath65 small enough .
we first recall the element wise trace inequality @xmath328 which holds with a uniform constant independent of the intersection , see lemma 4.2 in @xcite . to estimate the first term we change domain of integration from @xmath1 to @xmath56 and then use the trace inequality ( [ eq : trace ] ) as follows @xmath329 where we used the interpolation estimate ( [ interpolstandard ] ) followed by the stability estimate ( [ eq : ext_stab ] ) for the extension operator . observing that @xmath330 with @xmath331 we arrive at @xmath332
the second term can be directly estimated using the elementwise trace inequality followed by ( [ interpolstandard ] ) .
[ thmenergy ] the following a priori error estimate holds @xmath333 for @xmath159 with @xmath65 small enough . adding and subtracting an interpolant @xmath334 , defined by ( [ pihl ] ) , and @xmath335 , and using the triangle inequality we have @xmath336 here the first two terms can be immediately estimated using the interpolation error estimate ( [ interpol ] ) . for the third and fourth we have the following identity @xmath337 estimating the right hand side we obtain @xmath338 [ [ term - boldsymbol - i-1 ] ] term @xmath33 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the first term can be directly estimated using the interpolation inequality ( [ interpol ] ) .
[ [ term - boldsymbol - iboldsymbol - i-1 ] ] term @xmath204 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + changing domain of integration we obtain the estimate @xmath339 where we used the estimates ( [ detbbounds ] ) , the poincar inequality in lemma [ lemmapoincaresigmah ] on @xmath56 , and finally we mapped from @xmath56 to @xmath1 . [ [ term - boldsymbol - iboldsymbol - iboldsymbol - i-1 ] ] term @xmath215 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + changing domain of integration we obtain @xmath340 where at last we used the stability estimate @xmath341 for the method .
furthermore , we used ( [ bestimates ] ) and ( [ detbbounds ] ) to show the following estimate @xmath342 finally , collecting the estimates of terms @xmath197 the proof follows . the following a priori error estimate holds @xmath343 for @xmath159 with @xmath65 small enough .
recall that @xmath344 satisfies @xmath345 and define @xmath346 such that @xmath347 then @xmath348 and we have the estimate @xmath349 [ [ term - boldsymbol - i-2 ] ] term @xmath33 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath350 be the solution to the dual problem @xmath351 .
then it follows from the lax - milgram lemma that there exists a unique solution in @xmath352 and we also have the elliptic regularity estimate @xmath353 . in order to estimate @xmath354
we multiply the dual problem by @xmath355 , integrate using green s formula , and add and subtract suitable terms @xmath356 these terms may now be estimated using cauchy - schwarz , the energy norm estimate ( [ eqenergy ] ) , together with the estimates of terms @xmath357 and @xmath218 in the proof of theorem [ thmenergy ] .
we note in particular that @xmath358 here the first term is estimated by observing that @xmath359 and the second term @xmath360 using lemma [ lem : approx ] .
[ [ term - boldsymbol - iboldsymbol - i-2 ] ] term @xmath204 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + using the fact that @xmath345 we obtain @xmath361 combining the estimates of @xmath362 and @xmath357 we obtain the desired estimate .
in order to assess the effect of our stabilization method on the condition number , we discretize a circle , solve the eigenvalue problem of the laplace - beltrami operator and sort these in ascending order .
the problem is posed so that the integral of the solution is set to zero by use of a lagrange multiplier ( for well - posedness ) .
the first eigenvalue ( corresponding to the multiplier ) is then negative . in the unstabilized method
the next eigenvalue is zero with eigenfunction equal to the discrete piecewise linear distance function used to define the circle .
this is due to our choice to define the circle using a level set function on the same mesh used for computations ; this particular problem can be avoided using a finer mesh for the level set function .
we illustrate the effect by showing the zero isoline of the level set function ( thick line ) together with the isolines of the eigenfunction corresponding to the zero eigenvalue in figure [ zeroeigen ] . to avoid this effect , we base the condition number on the quotient between the largest eigenvalue and the first positive eigenvalue . for the unstabilized method ,
the first nonzero eigenvalue is thus the third , whereas for the stabilized method it is the second . in the stabilized method we used @xmath363 for all computations . in fig .
[ mesh2d ] we show the initial position of the circle in a 2d mesh .
the circle is then moved to the left , to end up a distance @xmath364 to the left of its original position . for each increment @xmath365 ,
we plot the condition numbers of the two methods , shown in fig .
[ cond2d ] .
note the large variation in condition number of the unstabilized method . for our convergence / conditioning comparison , we discretize a sphere of radius @xmath366 with center at @xmath367 and with a load @xmath368 corresponding to the exact solution @xmath369 compute @xmath23 on @xmath56 to solve for @xmath344 , and define an approximate @xmath0error as @xmath370 a plot of the approximate ( unstabilized ) solution on a coarse mesh
is given in figure [ solution ] , shown on the planes intersected by the level set function .
we compare the error for the stabilized ( using different values for @xmath81 ) and unstabilized methods in figure [ errors ] , where ndof stands for the total number of degrees of freedom on the active tetrahedra , so that @xmath371 .
note that the error constant is slightly worse for the stabilized methods but that all choices converge at the optimal rate of @xmath372 .
the numbers underlying figure [ errors ] are given in table [ table : conv ] , where @xmath373 stands for the number of unknowns , the errors @xmath374 are listed for different @xmath81 , and @xmath375 is the rate of convergence . in figure [ conditioning ]
we show the condition number computed for the same problem ( with the same approach as in section [ condest2d ] ) with different choices for @xmath81 and also for the preconditioning by diagonal scaling suggested in @xcite .
no stabilization results in a condition number that grows faster than the standard rate of @xmath376 .
the condition number is most improved by diagonal scaling .
note , however , that diagonal scaling does not remedy the zero eigenvalue induced by the level set .
the numbers underlying figure [ conditioning ] are given in table [ table : cond ] .
.errors and convergence for different @xmath81[table : conv ] [ cols=">,>,>,>,>,>,>,>,>",options="header " , ]
this research was supported in part by epsrc , uk , grant no .
ep / j002313/1 , the swedish foundation for strategic research grant no . am13 - 0029 , and the swedish research council grants nos .
2011 - 4992 and 2013 - 4708 .
10 g. dziuk .
finite elements for the beltrami operator on arbitrary surfaces . in _ partial differential equations and calculus of variations _ , volume 1357 of _ lecture notes in math .
_ , pages 142155 .
springer , berlin , 1988 . | we consider solving the laplace - beltrami problem on a smooth two dimensional surface embedded into a three dimensional space meshed with tetrahedra .
the mesh does not respect the surface and thus the surface cuts through the elements .
we consider a galerkin method based on using the restrictions of continuous piecewise linears defined on the tetrahedra to the surface as trial and test functions . the resulting discrete method may be severely ill - conditioned , and the main purpose of this paper is to suggest a remedy for this problem based on adding a consistent stabilization term to the original bilinear form .
we show optimal estimates for the condition number of the stabilized method independent of the location of the surface .
we also prove optimal a priori error estimates for the stabilized method . laplace beltrami , embedded surface , tangential calculus . |
the technique of ice mapping was developed to determine the spatial distribution of solid phase molecular species in star - forming regions @xcite .
one question remaining to be answered is whether obtaining information on both the abundances and spatial distribution of solid molecular material simply provides an insight into the global chemical networks in molecular clouds , or whether , additionally , it probes the prevailing physical conditions in the local environment when the ices first formed and can reveal chemical diversity across a star - forming region , thereby illuminating the very small scale feedback between solid and gas phase chemistries . in general , gas phase molecules
are observed in emission , making it comparatively simple to use different chemical species or isotopes to detect molecules and use them to trace astrophysical properties such as density , temperature , and velocity across vast regions of a molecular cloud ( e.g. * ? ? ?
* ) , or in great detail over very small astronomical scales i.e. tens to hundreds of au ( e.g. * ? ? ?
* ) . since ices can only be directly observed as absorption features in the spectrum of a bright background source , mapping their spatial distribution is much more challenging , necessarily relying upon the serendipitous positioning of stars directly behind the cloud to probe individual lines of sight ( e.g. * ? ? ?
one observational approach used to better constrain solid phase chemistry has been to infer that ice must be present in order to explain observed gas phase emission in those cases when molecules with only solid phase formation mechanisms are probed e.g. the detection of gas phase water , generated by non - thermal desorption of water ices deep in a prestellar core @xcite .
such analysis is reliant on gas - grain models , with many data inputs which can be independently changed to influence the output gas abundances for comparison with observations .
similarly , in the study of protoplanetary disks , the gas phase observation of `` complex '' organic species presumed to originate exclusively from the solid phase is used to infer the presence of ice hidden within the disk .
the conjecture is that a `` jump '' in the abundance of such molecules in the gas phase is directly relatable to their solid phase abundance through the onset of photo- or thermal - desorption mechanisms ( e.g. * ? ? ?
very recently the first observation of a water ice snowline in a protoplanetary disk was made using alma , by correlating an increase in continuum dust opacity in the inner disk to the onset of water ice evaporation from cm - sized icy particles migrating from outside the snowline @xcite .
the detection of the first branched organic molecule in the ism , iso - propylcyanide , was also recently reported , and its presence attributed ( through astrochemical modelling ) to radical addition reactions within the solid phase followed by ice desorption @xcite .
[ table1 ] [ cols="<,^,^,^,^,^,^,^,^,^,^ " , ] from a solid phase perspective , there is still much work to do to provide the fundamental data necessary to constrain the origins of these molecules observed in the gas phase , e.g. by determining precise values of ice abundances and compositions , studying ice morphologies , and understanding how solid phase processes are affected by these parameters .
solid phase com production in astrochemical models is reliant on methanol formation @xcite , which itself depends on co ice hydrogenation @xcite , and by inference , how much co ice originally formed .
this process is known to occur from critical freezeout of co gas in dense interstellar regions @xcite , and co ice abundances have been shown to vary more signifiantly than any other ice component from star - forming cloud to cloud @xcite .
gas phase mapping can now be conducted on scales of a few au , probing structure on small astronomical scales within dense molecular cores and even into a disk .
the first `` image '' of a co snowline in a protoplanetary disk was recently constructed from alma detections of n@xmath0h@xmath2 emission ( an ion that astrochemical models show increases in abundance as co freezes out ) , locating its origin to around 30 au from the central star @xcite .
although such approaches are ideal for locating the onset of co ice freezeout in the disk , and remain our only spatially resolved probe of ice distributions within protoplanetary disks at the current time , one can not be absolutely certain that ice abundances are consistent ( predictable ) or even generic over spatial scales of a few 10 s to 1000 s of au in star - forming regions .
the spatially resolved distributions of co , h@xmath0o , and co@xmath0 ices in dense cores on scales of around a few 1000 au are reported here for the first time . with these data
we aim to address the assumptions inherent in the current observational approach of deriving ice abundances indirectly from gas phase mapping .
we generate spatially resolved ice abundance maps and investigate how they compare to gas and dust abundances in the same regions , as well as to the globally - adopted , generic set of interstellar ice abundances @xcite .
the ice absorption spectroscopy method adopted in this study is observationally more challenging than inferring solid phase abundances from gas phase observations .
the major advantage of such an approach is that the solid phase material is probed directly ; the obvious disadvantages are the limitations on the number of sources and the spatial scale .
the traditional approach to ice observations has been to observe single lines of sight in single pointings , then to utilise one - dimensional correlation plots to describe the general trends in ice chemistry over a number of clouds or regions @xcite , or histograms to show the range of ice abundances , mixtures , and environments across a statistically significant sample @xcite .
the utilisation of a two - dimensional ice mapping approach enables us to test the spread of values in such correlations , probing whether spatial ( ice abundance ) differences in a core are correlated with localised effects , such as dust evolution , heating , density , turbulence , and radiation field , that are already know to vary considerably over short distances in star - forming regions .
realistically , most of these factors even vary on scales smaller than we are able to probe with current ice maps . the first ice maps were reported by @xcite , who made partial observations of the 3 @xmath3 m h@xmath0o stretching absorption band towards 50 stars behind the taurus molecular cloud complex , where the optical depth at 3.1 @xmath3 m was found to correlate well with a gas phase c@xmath1o map , suggesting that h@xmath0o ice abundance increases with gas density tracers . the subsequent , more detailed , studies of @xcite highlighted the power of the technique in elucidating ice chemistry and abundance changes on relatively small scales .
a map of h@xmath0o and ch@xmath4oh ice abundances showed that , while h@xmath0o ice abundances remain constant over a range of densities , ch@xmath4oh ice abundances increase by at least a factor of ten towards the class 0 protostar smm 4 in the serpens star forming cloud @xcite .
the ch@xmath4oh is likely formed from co frozen out on dust grains at high densities .
this work was followed by h@xmath0o , co@xmath0 , and co observations on lines of sight towards five ysos in the ophiuchus - f star forming core which , when mapped against distance from the core , implied that co@xmath0 ice formation occurs in two distinct periods : concurrently with h@xmath0o at low extinctions , observed in a h@xmath0o - rich ice formed on the grain surface ; and during co freezeout , observed in a co - rich overlayer @xcite .
finally , a study of the column densities of solid co@xmath0 and h@xmath0o towards the cepheus a east star - forming region , was performed using _ spitzer _ @xcite .
uniquely , these ice absorption features were observed , not towards specific ir sources , but against continuum background emission .
the results suggested that co@xmath0 and h@xmath0o were formed concurrently on dust grains , across the region sampled ( much in line with the findings of laboratory experiments on concurrent h@xmath0o and co@xmath0 ice formation @xcite ) , while peak abundances were found in regions closest to the embedded protostar .
this type of study provides evidence that processing in one part of a star - forming region facilitates a different chemistry in that local area , but not necessarily throughout the surrounding areas .
the ice mapping method has even been extended to extragalactic sources ; @xcite showed that h@xmath0o ice is widely distributed in the nearby star - burst galaxy m82 , whilst co@xmath0 ice is concentrated near the galactic centre .
whilst again a very different astronomical spatial scale , they also attribute variations in co@xmath0:h@xmath0o ice abundance ratios to changes in the `` local '' interstellar environments within the galaxy .
likewise , the same group showed that ice distributions of h@xmath0o , co@xmath0 , and ocn@xmath5 varied significantly across the edge - on galaxy ngc 253 , and differed significantly from that of the h@xmath0 gas and pah emission @xcite . previously , we reported the detection of ices towards a sample of 22 background stars and eight young stellar objects ( ysos ) @xcite .
the aims of the current work are : using these data , to move from individual spectra to two - dimensional maps of ice column densities for co , co@xmath0 , and h@xmath0o ices with spatial resolutions on the scale of a few 1000 au ; to test the ice column densities against the spatial distribution of dust and gas phase co in the region ; and to test the reliability of multi - object , slitless spectroscopy as a technique for ice mapping .
we anticipate that the data acquired with akari will set the precedent for future spectroscopic observations of this type with jwst and the e - elt . , ordered by decreasing irac 1 flux .
overplotted on each spectrum is the fitted baseline ( dotted line ) and the photometric points irac 1 and irac 2 ( diamonds ) . using our extraction technique
, we can probe spectra with a flux lower than 10 mjy , as seen in objects 714.,scaledwidth=50.0% ]
we observed the dense cores of four molecular clouds
b 35a , dc 274.2 - 00.4 , bhr 59 , and dc 300.7 - 01.0 using the nir spectroscopy mode of the akari satellite ( aot irc04 ; ng disperser , 2.5 5 @xmath3 m , 1 @xmath6 1 slit ) .
data reduction was fully detailed in a previous publication @xcite .
the aim of the current study was to construct spatially resolved ice maps of h@xmath0o , co@xmath0 , and co towards these four molecular cores , chosen from the european users open time observing programme imape programme as they were the only observations from which the spectra of three or more objects could be concurrently extracted from a single akari pointing .
figure [ fig1spectra ] shows the extracted spectra , ordered by descending flux at 3.6 @xmath3 m ( object numbers 1 14 ) . for clarity ,
the corresponding object numbers used in @xcite are given in table [ table1 ] along with target co - ordinates , and the associated ice abundances ; targets are identified as either background stars or low - mass young stellar objects ( ysos ) according to their attribution in the c2d _ spitzer _ legacy programme catalogues @xcite .
it is interesting to note that ice absorptions are clearly visible by eye in all of these spectra , even at fluxes as low as a few millijansky ( e.g. objects 12 and 13 ) .
water ice is detected in all 14 spectra .
co@xmath0 ice is also detected in all objects , but those for which the proximity of other objects in the observed field of view could not be fully corrected for have their extracted column densities given as upper limits in table 1 ( for a more detailed discussion , see @xcite ) .
co ice was detected in all but two objects and again around half of these detections include some minor confusion with extraneous light , and so the values are reported as upper limits .
a comprehensive process of baseline subtraction combined with fitting of laboratory spectra was used to extract the ice column densities ( again , see @xcite for full details ) . in a few cases , e.g. objects 2 , 5 , 13 and 14
, the baseline fitted to the full spectrum does not perform as well beyond around 4.5 @xmath3 m , so a small additional linear baseline correction was made across the co absorption feature .
the total h@xmath0o , co@xmath0 , and co column densities are given in table [ table1 ] .
in fact , the total column densities of co and co@xmath0 ice are the sum of multiple components which reflect the chemical environment surrounding the molecules . the column densities of individual co and co@xmath0 components , their associated uncertainties , and the component analysis method are reported in @xcite .
briefly , in the objects observed in this work , co@xmath0 is present in co - rich and h@xmath0o - rich ices ( fitted using cde - corrected laboratory spectra ) , while co is observed in pure ( `` co@xmath7 '' ) and h@xmath0o - rich ( `` co@xmath8 '' ) ices ( fitted following the @xcite component analysis approach ) .
additionally , column densities of ocn@xmath5 , co gas - grain ( co@xmath9 * ? ? ?
* ) , and @xmath10co@xmath0 were calculated where these spectral features were detected .
figure [ figav ] presents the ice column densities of h@xmath0o , co@xmath0 , and co plotted against visual extinction ( a@xmath11 ) .
the data in figure [ figav ] is presented two ways : firstly in terms of ice species ( squares for h@xmath0o , triangles for co@xmath0 , and circles for co ) , and secondly in terms of cores ( b 35a in cyan , dc 274.2 - 00.4 in red , bhr 59 in blue , and dc 300.7 - 01.0 in purple ) .
it is known that , generally , as the dust density increases so does the total ice column density , and therefore some correlation is expected between ice column density and a@xmath11 ( e.g. * ? ? ?
* ; * ? ? ?
we can corroborate this conclusion with figure [ figav ] , showing that for our data , broadly speaking , the higher the a@xmath11 , the more ice is present . considering the individual ice species , the column densities of co@xmath0 and h@xmath0o ice globally
increase with a@xmath11 , whereas the co column density remains broadly constant .
these results are consistent with co@xmath0 and h@xmath0o ices forming on the dust surfaces ( and therefore increasing in abundance if more dust ( _ i.e. _ a greater surface area ) is available ) , but co freezing out from the gas phase in a process closely linked to the gas density @xcite .
the dense cores observed here have extinctions between 10 and 50 a@xmath11 , so we are not probing the regions of ice formation onset but rather regions with established icy dust grain mantles . if we consider the data core by core a more complex picture emerges . in b
35a ( cyan ) , a@xmath11 values could only be obtained for two of four sources ( target data was extracted from the c2d _ spitzer _ legacy programme catalogues @xcite ) , but for the two datapoints available the ice column density increases with dust density .
likewise , in bhr 59 ( blue ) , the ice column densities of h@xmath0o and co@xmath0 increase with dust density , and co ice column densities are approximately constant , as one would expect .
however in dc 274.2 - 00.4 ( red ) and more particularly in dc 300.7 - 01.0 ( purple ) the trends are different .
only the h@xmath0o ice column density in dc 274.2 - 00.4 follows the presumed increase with dust density .
the co@xmath0 and co ice column densities are lower at higher a@xmath11 . in dc
300.7 - 01.0 , the co , co@xmath0 , and h@xmath0o ice column densities decrease with increasing a@xmath11 , with the trend being most prominent for co@xmath0 .
this suggests that ice constituents may not reflect the prestellar chemistry across an entire core , but be more heavily influenced by very local conditions . to explain these ice column densities , some ice destruction mechanisms ( such as non - thermal desorption )
must be dominating , producing effectively lower freezeout rates as a@xmath11 increases .
o , co@xmath0 , and co plotted against visual extinction ( a@xmath11 ) .
ice species are represented by shapes ( squares ( @xmath12 ) for h@xmath0o , triangles ( @xmath13 ) for co@xmath0 , and circles ( @xmath14 ) for co ) and cores are identified by different colours ( b 35a in cyan , dc 274.2 - 00.4 in red , bhr 59 in blue , and dc 300.7 - 01.0 in purple).,scaledwidth=50.0% ] we can further probe the adage `` more is more '' , _
e.g. _ as the water ice column density increases , so does the co@xmath0 ice column density , by comparing the derived column densities of one ice species to another .
this relationship is not entirely true for co ice , where gas density has been shown to be more critical to the ice freezeout rate @xcite , but holds for h@xmath0o and co@xmath0 . in figure
[ figcorrelationplot ] , we plot all published column densities of h@xmath0o , co@xmath0 , and co ices ( of which we are aware , per single source , excluding upper - limits ) in two correlation plots : h@xmath0o versus co and h@xmath0o versus co@xmath0 @xcite .
the data exploited in this work , which originate from background sources and low - mass ysos , are shown in colour on these plots and entirely overlap with existing literature values .
we will revisit the specific core - by - core distribution of our data below .
it is clear that , although the derived column densities of each species span between two and three orders of magnitude , a general correlation is apparent in each plot .
the limitation of traditional column density correlation plots is that they give us no information about the astronomical source of each data point , or the extent to which ice column densities change on scales of up to a few thousand au _
i.e. _ which data point is associated with which particular line of sight , nor which objects lie within , or probe , the same cloud or core .
clearly , the correlations contain a significant spread of values ( in the anti - correlation direction ) , even when plotted on log - log scales , which leads to the question : how can we account for this diversity ?
one might suspect ice column density of varying consistently with source type .
the plots in figure [ figcorrelationplot ] include both extragalactic data ( squares ) and galactic data , the latter divided into low - mass ysos ( stars ) , high - mass ysos ( triangles ) and background sources ( circles ) .
extragalactic values are typically closer to those found in high - mass ysos in our own galaxy @xcite .
however data from galactic low - mass ysos , background stars and high - mass ysos are well spread across figure [ figcorrelationplot ] with no evident source - dependent trends .
this spread of values was rationalised in a detailed statistical analysis by @xcite of ice observations encompassing 63 low - mass ysos , eight high - mass ysos and a handful of background ( galactic ) sources ( without access to the more recent field star surveys of @xcite ) .
@xcite derived the median generic ice abundances of ( amongst other solid phase species ) h@xmath0o , co , and co@xmath0 , giving typical values of 100:29:29 ( in low - mass ysos ) , 100:31:38 ( in cloud cores _
i.e. _ towards background stars ) , and 100:13:13 ( in high - mass ysos ) .
these generic ratios are superimposed on figure [ figcorrelationplot ] as black lines : solid for background sources and dashed for low - mass ysos , which for these three ice species have remarkably similar ratios .
however , they also noted that ( in their dataset ) the actual abundance values towards low - mass ysos could vary for co and co@xmath0 by up to a factor of two ( although towards background stars the variation was only a few percent ) .
figure [ figcorrelationplot ] shows that the published ice column density data from h@xmath0o , co , and co@xmath0 ices are mostly encompassed by the limits calculated by @xcite ( shown as grey solid lines for background sources and grey dashed lines for ysos ) .
however , we include more data points than @xcite and observe that the spread of values is even broader ( in this case , for h@xmath0o : co@xmath0 we plot 250 objects ( comprising 62 background stars , 80 low - mass stars , 31 high - mass stars and 76 extragalactic sources ) and for h@xmath0o : co we plot 127 objects ( comprising 33 background stars , 47 low - mass stars , 30 high - mass stars and 17 extragalactic sources ) ) . this large data spread is also observed within the small subset of our own data points ( plotted in colour ) .
o , co@xmath0 , and co plotted as correlation plots .
object types are plotted as : field stars ( @xmath14 ) , low - mass ysos ( @xmath15 ) , high - mass galactic ysos ( @xmath16 ) , extragalactic sources ( @xmath17 ) .
literature values are plotted in grey , and the four cores of this work are identified by different colours ( b 35a in cyan , dc 274.2 - 00.4 in red , bhr 59 in blue , and dc 300.7 - 01.0 in purple ) .
the straight lines represent average ice abundances calculated by @xcite and their derivation is explained in the text.,scaledwidth=50.0% ] clearly then , source type alone can not account for the diversity of ice column densities observed .
an alternative explanation is that from one cloud to another one occupies a slightly different region of parameter space on the correlation plot , which somehow reflects the differences on astronomically `` regional '' scales . on a cloud - by - cloud basis
, variations in ice formation and processing in prestellar phases would be carried forward to the protostellar phases , meaning that different cores within the same cloud would demonstrate similar ice abundances .
however , it is also possible that a more local core - to - core variation is seen within one cloud . in figure
[ figcorrelationplot ] , the ice column densitites from each of our cores are highlighted in cyan ( b 35a ) , red ( dc 274.2 - 00.4 ) , blue ( bhr 59 ) , or purple ( dc 300.7 - 01.0 ) . since h@xmath0o and co@xmath0 ices were detected in all 14 objects in this data set , the column density correlation plot for co@xmath0 versus h@xmath0o contains a full compliment of data , whereas the co versus h@xmath0o correlations exclude two objects ( where no co ice was detected ) .
although the statistics are relatively poor ( because the field of view is only 1 arcminute square and therefore contains a maximum of four non - confused objects in each case ) , of the four cores presented here , two
b 35a and bhr 59 adhere globally to the `` more is more '' paradigm while the other two
dc 300.7 - 01.0 and dc 274.2 - 00.4
manifestly do not .
it is clear from figure [ figcorrelationplot ] , that the correlation between the co@xmath0 and h@xmath0o column densities in the four ysos in the b 35a field of view ( highlighted in cyan ) lie almost entirely on a straight line .
however , the lines of sight towards background stars in bhr 59 ( blue ) already deviate slightly from a perfect correlation , while the background stars in dc 300.7 - 01.0 ( purple ) and the combination of background stars and ysos in dc 274.2 - 00.4 ( red ) clearly show no correlation , and potentially even an anti - correlation , between the h@xmath0o and co@xmath0 column densities .
likewise , even though the correlation between h@xmath0o and co column densities is generally less robust than that between co@xmath0 and h@xmath0o , in b 35a and bhr 59 the `` more is more '' adage still seems to hold , whereas again in dc 300.7 - 01.0 and dc 274.2 - 00.4 it does not .
realistically there is no clear trend in ice abundances , even within those cores that contain only background sources , so it appears that ice column densities must be reliant on much more local influences than cloud - to - cloud or even core - to - core variations .
this conclusion is entirely consistent with the statistical kruskal - wallis analysis of @xcite and an earlier principal component analysis by @xcite .
individual variations in ice abundances must therefore be attributable to very localised conditions during both the prestellar formation phase of the ice and its subsequent processing during protostellar evolution , even within individual cores , as suggested by the monte carlo simulations of @xcite . for the remainder of this paper
we consider three distinct phases of ice formation and processing ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and discuss our ice data in this context . of specific relevance to the three ice species discussed here :
* firstly , ices form such that the co@xmath0:h@xmath0o column density ratio is determined in the `` early '' stages of ice formation , in prestellar cores , when the co@xmath0 ice forms in competition with h@xmath0o ice , driven by hydrogenation reactions , thereby justifying why these ices generically follow the `` more is more '' mantra .
such mechanisms are corroborated by laboratory experiments which quantify the relative rates of co@xmath0 and h@xmath0o formation ( e.g. * ? ? ?
* ; * ? ? ?
* ) and by models which exploit such empirical data ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
devations from `` more is more '' must point towards specific localised conditions which accellerate or decellerate ice formation with increasing a@xmath11 . * secondly , ices evolve such that the co : h@xmath0o column density ratios are constrained mostly during the `` late '' prestellar phases , where co ice freezes out and ice composition is driven by variations in co freezeout rate , combined with a co - based chemistry .
this has been shown in the first multi - molecular 2d ice maps of the ophiuchus - f core @xcite and the conclusions of the statistical analysis of @xcite . *
finally , the co : h@xmath0o column density ratio varies again in protostellar environments , where ice chemistry is subject to thermal and non - thermal processing , leading ( of particular relevance here ) to co ice desorption , ice mixing and/or segregation between ice constituents , and processing to form more complex chemical species .
this highlights the need to be able to map in at least two dimensions the positions and ice abundances in objects concurrently , and to be able to compare such maps with other observational constraints on the prevailing physical conditions in the ism . for the remainder of this paper we focus on this challenge
maps of the calculated ice column densities are presented in figure [ fig2maps ] .
the akari imaging fields of view at 3.2 @xmath3 m are shown in figure [ fig2maps]a . adopting the c2d definitions @xcite , the bhr 59 and dc 300.7 - 01.0 fields of view contain only background star observations ( marked @xmath14 ) , and so are likely to trace prestellar regions , while the map of b 35a contains only data from embedded objects ( marked @xmath15 ) and dc 274.2 - 00.4 contains both .
the two closest objects which could be extracted from the akari field of view using our reduction pipeline are objects 4 & 14 . as dc
300.7 - 01.0 lies at a distance of 175 parsecs @xcite , the resolution between these objects represents @xmath18 2200 au , and thus the map probes chemistry over small spatial distributions ( but still a significantly larger scale than gas phase observations ) . figure [ fig2maps]b shows the a@xmath11 towards each object plotted as bubble plots , the values being extracted from the c2d catalogues .
objects 1 & 11 have no calculated a@xmath11 values , but based on gas phase observations of this cloud , it is estimated that the values will be roughly equivalent to the other two objects in the field of view @xcite .
visual extinction is correlated to the dust abundance in the cloud , and as pointings were centred on the most dense region of each core , we are likely probing the highest visual extinctions possible
. h@xmath0o ice maps ( figure [ fig2maps]c ) show column densities of h@xmath0o which are closely related to the a@xmath11 values in the panels above .
the ice maps are testing the spread in a 1d correlation plot of h@xmath0o and a@xmath11 .
our data agree well with previous studies ( e.g. * ? ? ?
* ) that a@xmath11 and n(h@xmath0o ) are closely correlated .
although the relationship is consistent within the clouds , it is not possible to predict the column density of h@xmath0o given the a@xmath11 , nor vice - versa .
for example , in dc 300.7 - 01.0 , the a@xmath11 is approximately constant towards all lines of sight , but the column density of h@xmath0o varies by a factor of two over a small scale .
there must , therefore , be other physical and astronomical influences to consider in the formation and spatial distribution of ice in clouds , and ice mapping provides a powerful tool for investigating them .
the information in an ice map illustrates the link between chemical reagents and products on a spatial scale , and co@xmath0 and co are a good test of this . in figure
[ fig2maps]d , the maps of total co@xmath0 ice column density show a general agreement with the h@xmath0o ice maps .
this is unsurprising , as previous studies have already established the existence of a correlation between the abundances of h@xmath0o and co@xmath0 in interstellar ices ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . a notable exception to this pattern is object 3 , where the column density of co@xmath0 appears to be high in relation to that of h@xmath0o .
thus , in clouds such as b 35a , co@xmath0 appears to be a tracer of h@xmath0o , while in dc 300.7 - 01.0 this does not seem to be the case . however , given that the values calculated in the former are upper limits only , it is not possible to draw any firm conclusions based upon this single aberration .
co ice maps are shown in figure [ fig2maps]e and are in general agreement with both the a@xmath11 and h@xmath0o maps .
all of the lines of sight are above the critical extinction threshold for co freezeout , and so the column densities of co are expected to be approximately 20 30 % h@xmath0o @xcite .
however , we would expect to see discrepancies in the co distribution as it begins to freeze out . towards objects 5 and 9
there is no co ice , but this is a reasonable result when the spectra in figure [ fig1spectra ] are considered .
there is little or no co absorption band visible in these spectra , suggesting that , within the detector limits , there is no co ice towards these lines of sight .
a potential explanation for this is that heating in these isolated regions has caused co to desorb into the gas phase . in general
, observations to date have shown that ice abundances toward background stars and protostars are largely similar , except for a lack of features associated with ice heating ( for example pure co@xmath0 ice ) toward background stars .
in particular , there is no evidence for a different range of ch@xmath4oh and co@xmath0 ice abundances .
additionally , the co@xmath0:h@xmath0o ratio , and ch@xmath19 and nh@xmath4 abundances vary little with respect to h@xmath0o , suggesting co - formation of these ices . in contrast , the co : h@xmath0o and co@xmath0:co ratios , as well as the abundances of ocn@xmath20 and ch@xmath4oh vary by factors of two to three with respect to h@xmath0o , indicative of a separate formation pathway from h@xmath0o ice .
pure co and co@xmath0 ice abundances vary even more , consistent with their sensitivity to protostellar heating @xcite .
interestingly , across all ice species mapped in figure [ fig2maps]c , d & e , there is no real trend in the column densities derived for either background stars or for ysos . in these data , the highest column density of h@xmath0o
is observed towards a background star ( object 13 ) .
however , it must be noted that this value is an upper limit , and the next three highest h@xmath0o column densities are seen towards ysos in b 35a . for co
, the two highest column densities are seen towards background stars ( objects 8 and 14 ) , only one of which ( 14 ) is an upper limit .
for co@xmath0 , the three highest column densities are again seen towards ysos in b 35a , which fits better with the standard scenario of ice formation .
when considering the component analysis of co@xmath0 and co ( as shown in figure [ fig2maps]d , e by the overplotting of orange circles denoting water - rich environments ) , it is interesting to note that in the mapped data there is no obvious pattern in any of the fields of view .
for example , towards b 35a , the co@xmath0 towards objects 6 and 11 is entirely in a co - rich ice ( the h@xmath0o - rich component , denoted in these maps by orange circles , being competely absent ) , and towards objects 1 and 12 , the majority is in a co - rich ice . for co ,
objects 6 and 12 are exclusively co in a pure co ice , but object 11 has predominantly co in a h@xmath0o - rich ice , and object 1 is more equally mixed .
thus , for object 11 , ( almost ) all co is in a h@xmath0o - rich environment , rather than pure co , but the co@xmath0 is exclusively in a co - rich ice .
is it possible that co@xmath0 is trapped in co pockets in a h@xmath0o ice ? in object 6 , however , all the co and co@xmath0 is in a co - rich environment . towards this line of sight
, did co@xmath0 form exclusively from a pure co ice ? or is the ice porous , and diffusion of co and/or co@xmath0 has occurred on a large scale ?
is it possible that the co freezeout occurred extremely rapidly in this line of sight ?
furthermore , co and co@xmath0 ice components ( e.g. in water ) can vary by factors of at least an order of magnitude from the `` median '' values derived .
these results somewhat challenge the standard model of ice formation , which is based upon the freeze out of a co layer on top of a mixed co@xmath0:h@xmath0o layer . considering the other fields of view
, it seems that there are also some other objects with non - standard column densities . towards object 3 in dc 274.2 - 00.4 ,
all co@xmath0 is in a co - rich ice , while all co is in a h@xmath0o - rich ice ( which is similar to object 11 in b 35a ) . towards object 14 in dc 300.7 - 01.0 ,
there is an anomalously high co column density , while the little co@xmath0 which is present is again exclusively in a co - rich ice .
these data suggest that even within a single cloud , the chemistry can be different towards different lines of sight and that making generalisations might not be a reasonable approach to take .
the relationships established by the component analysis comparisons in 1d correlation plots become more complex still when spatial distribution is considered .
if it is assumed that some co@xmath0 formation occurs at early times , concurrent with h@xmath0o , it would be reasonable to assume that some co@xmath0 in a h@xmath0o - rich component would be visible towards all lines of sight .
the component analysis in figure [ fig2maps]d shows that no co@xmath0 in h@xmath0o is observed towards objects 6 , 8 , 10 , 11 & 14 .
this is most likely due to a limitation in the detection method , with very small abundances of co@xmath0 in fact being present on these lines of sight , but requires observations at higher sensitivity and resolution for clarification . as was discussed earlier , most of the co@xmath0 in a h@xmath0o - rich environment is likely formed at later times , so correlation should be seen between co@xmath8 and co@xmath0 in h@xmath0o .
this is generally the case in b 35a and bhr 59 , but there are anomalies . for example , object 9 , which has no co ice , has almost all its co@xmath0 ice in a h@xmath0o - rich phase .
this could suggest that all of the co has desorbed or has reacted to form co@xmath0 in the h@xmath0o - rich ice , as seen in the ice models of @xcite , and could be evidence of heating , uv processing , or the age of object 9 .
considering minor components which were observed towards some lines of sight , but are not mapped here , objects 3 , 4 , 5 , 6 & 9 were all found to have ocn@xmath5 , which is a tracer of uv or temperature processing .
this corresponds well to the assertions made about the non - detection of co towards objects 5 & 9 .
object 6 has the highest ( available ) a@xmath11 of any of the mapped objects , and thus is the most highly embedded .
it may be that the observations are more sensitive to ocn@xmath5 at higher extinctions .
object 5 has low co , co@xmath0 column densities , and an ocn@xmath5 detection , which could suggest that this region of dc 274.2 - 00.4 is heated .
there is some evidence of the presence of co@xmath9 ( gas - grain ) towards all clouds , but not towards all lines of sight .
as co@xmath9 is believed to be a component that corresponds to co interacting directly with bare dust grain surfaces , it could trace a co component that can form co@xmath0 at very early times in a cloud , before there is accumulation of an icy mantle on grains . in b 35a , objects 1 , 6 & 11 all have co@xmath9 components , and it is also present towards objects 2 , 7 , 13 & 14 . as
this is a mixture of background stars and ysos , with varying extinctions , a correlating factor is not immediately evident .
further observations would provide more insight into this issue .
these data highlight the potential importance of the ice mapping technique in extracting detailed chemistry from prestellar regions , and represent an important first step in our understanding .
they strongly suggests that chemistry in interstellar regions could be very local , changing even from object to object within one cloud .
ultimately , higher resolution spectra over a wider spectral range ( e.g. combined nirspec / miri observations using jwst ) are required to fully investigate the environments in which ice and gas phase molecular species are present towards these lines of sight .
gas - grain coupled astrochemical models can be used to predict ice abundances , either at the `` end '' of a prestellar phase @xcite , or as a function of time during protostellar evolution @xcite , and an ideal observational constraint for comparison would be predictions of ice abundances for all species , even if observations only existed for one or two of them .
one of the original aims of this work was to compare our ice maps to gas phase maps of the same clouds , and investigate the relationships between the solid phase and gas density , temperature , chemistry , and astrophysics .
an example of such solid - gas mapping for the core b 35a is presented in figure [ figscuba3 ] , combining our ice data , gas phase c@xmath1o maps @xcite and archival scuba @xcite and _
herschel_/spire @xcite data .
a contour map of the core constructed from scuba maps at 450 and 850 @xmath3 m is shown in panel 1 .
the 450/850 @xmath3 m ratio is related to both the dust temperature and its opacity @xcite .
as the temperature increases through grain heating processes , or the opacity index @xmath21 increases due to grain growth @xcite , the 450/850 @xmath3 m ratio increases . in b 35a , the highest values are obtained just south of objects 1 and 12 .
h@xmath0o abundances towards all four objects , calculated from our h@xmath0o column densities and h@xmath0 gas phase maps of the region @xcite , are overplotted on panel 1 .
if the temperature and density are constant across the core , an increase in the 450/850 @xmath3 m ratio is likely related to coagulating grains or icy mantles @xcite , and therefore at the highest ratios we d expect to observe the highest ice abundances .
this is true for objects 12 , 1 and 11 , but not for object 6 .
this indicates that , were we to do quantitative ice mapping , we might be able to see whether ice abundances can establish if changes with dust opacity are directly related to the ice via processes such as coagulation . in figure [ figscuba3 ] , panel 2 we present the first combined gas - ice map of a single molecular species , comparing the total co ice abundance to the c@xmath1o@xmath22 integrated intensity .
c@xmath1o probes the densest regions of the core , and by comparison with panel 1 , we see that it is approximately colocated with the dust emission .
the relationship between solid and gas phase co is complex , as it relies on the freeze out of co under certain conditions of temperature and density .
we would expect that , if co freezeout is simply a function of density , there would be higher co ice abundances where co gas is the most abundant , but in this case there is no obvious trend in the column density of solid phase co compared to gas phase co. as was the case for h@xmath0o ( panel 1 ) the `` more is more '' pattern is observed for three objects , but object 6 is once again aberrant . in panel 3 , we consider the component analysis of the co ice band . here
we plot c@xmath1o@xmath23 integrated intensity contours , which traces more highly excited co gas but is observed in essentially the same region as the c@xmath1o@xmath24 in panel 2 . in panel 3
we also plot the solid co observed in a water - rich ice environment , co@xmath8 .
we only observe co@xmath8 towards the embedded yso object 1 and object 11 , with only pure co observed towards objects 6 and 12 .
this would seem to confirm that the co@xmath8 is a tracer of warmer ice , as co migrates into the water ice layer in warmer , denser regions @xcite .
it is interesting to note that objects 1 and 6 have almost identical total co abundances ( panel 2 ) but very different co@xmath8 abundances , as has previously been observed in studies of co ice @xcite .
as shown in figure [ fig2maps ] above , co@xmath0 ice was observed upon all four lines of sight in this core , with a variability rougly similar to the total co abundance variability .
however , due to the complexities of extracting column densities for the co@xmath0 data @xcite we choose instead to map @xmath10co@xmath0 ice against the 850 @xmath3 m scuba map in figure [ figscuba3 ] , panel 4 .
@xmath10co@xmath0 ice bands were observed towards three objects studied in @xcite , and the derived column density upper limits are presented in table [ table1 ] . for this small sample of objects ,
we obtain @xmath25c/@xmath10c ratios of @xmath18 10 15 , significantly lower than previous studies ( e.g. @xmath18 50 110 , @xcite ) .
although some of the difference can be attributed to the choices of laboratory ice mixtures used to analyse the two datasets , there is most likely an error introduced by the systematic undersampling of @xmath25co@xmath0 in our data due to the aforementioned column density extraction issues ( detailed in @xcite ) . in this case
, the interest is to determine whether the continuum dust emission can be related to isotopic fractionation in the ice , as would be the case if fractionation had a large temperature dependence .
however , it seems that there is no direct relationship between the cooler 850 @xmath3 m emission map and the distribution of @xmath10co@xmath0 in the core .
the final solid - gas map is the co@xmath9 component mapped against the 250 @xmath3 m scuba data mapping the peak dust emissivity ( figure [ figscuba3 ] , panel 5 ) .
as co@xmath9 probes the co directly adsorbed to bare dust grain surfaces , we would expect to see its abundance increase at higher density as there would be a greater surface area of dust upon which co could condense .
thus , even if there is more co ice due to greater density - induced freezeout , it may not fully cover the bare grains and thus the co@xmath9 component would increase .
however , in this case we see a roughly constant abundance of co@xmath9 across three objects and no co@xmath9 towards object 12 . for all of the gas - solid maps presented here ,
full analysis requires a lot more knowledge about the astrophysical environment on a local scale in this small core .
it is clear from these maps that the dust in the core of b 35a is centred around objects 1 and 11 , explaining the high column densities of h@xmath0o , co@xmath0 , and co ice present towards these objects .
however , the column densities of all three species are lower towards object 12 than object 6 , despite the fact that the dust is more abundant towards object 12 . thus
, dust density is not a direct indicator of ice column density .
the interesting thing with these comparisons is that where dust emission , ice absorption , and gas emission originate from may not necessarily be tracing the same spatial scales . in absorption ,
the matter is probed along the whole line of sight , whereas emission spectra may probe different depths .
this is beyond the scope of the current study , but work is ongoing to answer these questions and in subsequent ice maps , this will be an important factor to incorporate .
the ice mapping results presented in this work indicate that the simple adage `` more is more '' can not be applied to astrophysical ices on a local scale .
although globally this trend is observed when ice abundances are averaged across many lines of sight , mapping probes the large spread of data points in 1d correlation plots .
ice mapping , in this study probing resolutions of @xmath18 2200 au , gives a striking and immediate visual picture of the `` small scale '' variations which can occur within a cloud or even a core , highlighting the fact that there is no one interstellar ice composition .
combined solid - gas maps go one step further , providing vital information on the complex interplay between gas , ice , and dust in star forming regions .
overall , the maps presented in this work reiterate that the presence of more dust in a line of sight does not always indicate that there is more ice , and that both dust density and gas density must be considered to ascertain ice growth mechanisms .
there is no simple relationship linking dust ( or gas ) to ice .
they also indicate that features which are often identified as tracers of thermal history or ice processing ca nt necessarily be immediately linked to the local environment .
even the simple approach presented in this paper , of a direct comparison of ice abundances with available dust and gas data without supporting astrophysical modelling , indicates that ice chemistry is being affected by very local astrophysical conditions . with this technique
, we are tracing the chemistry of a region beyond that which can be ascertained from the gas phase alone .
in particular , ice mapping could be a powerful method to probe the prestellar to protostellar ice transition .
if we can identify specific ice constituents present just before the collapse phase we can set the initial chemical conditions for star formation . with additional modelling and observations , this will lead to a better understanding of how the thermal and processing history of a star forming region leads to ice evolution .
ice mapping is a powerful tool , but it is not as trivial , observationally , as gas phase mapping .
ideally , linking both mapping techniques , using telescopes such as jwst or the e - elt to produce combined solid - gas maps with less confusion and higher sensitivity , will yield more information on the complex interplay between gas , ice , and dust in star forming regions .
j.a.n . acknowledges the financial support of the royal commission for the exhibition of 1851 , the university of strathclyde , the scottish universities physics alliance , and the japan society for the promotion of science .
the research leading to these results has received funding from the european community s seventh framework programme fp7/2007 - 2013 under grant agreement no . 238258 ( lassie ) . h.j.f . and j.a.n .
are grateful to eu funded cost action cm0805 `` the chemical cosmos : understanding chemistry in astronomical environments '' for a funding contribution towards the final production of this work .
belloche , a. , garrod , r. t. , mller , h. s. p. , & menten , k. m. 2014 , science , 345 , 1584 bergin , e. a. , melnick , g. j. , gerakines , p. a. , neufeld , d. a. , & whittet , d. c. b. 2005 , , 627 , l33 boogert , a. c. a. , ehrenfreund , p. , gerakines , p. a. , et al .
2000 , , 353 , 349 boogert , a. c. a. , & ehrenfreund , p. 2004
, astrophysics of dust , 309 , 547 boogert , a. c. a. , pontoppidan , k. m. , knez , c. , et al .
2008 , , 678 , 985 - 1004 boogert , a. c. a. , huard , t. l. , cook , a .. m. , et al .
2011 , , 729 , 92 boogert , a. c. a. , chiar , j. e. , knez , c. , et al .
2013 , , 777 , 73 boogert , a. c. a. , gerakines , p. a. , & whittet , d. c. b. 2015 , , 53 , 541 van broekhuizen , f. a. , keane , j. v. , & schutte , w. a. 2004 , , 415 , 425 buckle , j. v. , hills , r. e. , smith , h. , et al .
2009 , , 399 , 1026 burke , d. j. , & brown , w. a. 2010 , phys .
, 12 , 5947 caselli , p. , keto , e. , bergin , e. a. , et al .
2012 , , 759 , l37 chiar , j. e. , adamson , a. j. , kerr , t. h. , & whittet , d. c. b. 1995 , , 455 , 234 chiar , j. e. , pendleton , y. j. , allamandola , l. j. , et al .
2011 , , 731 , 9 cieza , l. a. , casassus , s. , tobin , j. , et al .
2016 , , 535 , 258 craigon , a. 2015 , phd thesis , university of strathclyde collings , m. p. , dever , j. w. , fraser , h. j. , mccoustra , m. r. s. , & williams , d. a. 2003 , , 583 , 1058 di francesco , j. , johnstone , d. , kirk , h. , mackenzie , t. , & ledwosinska , e. 2008 , , 175 , 277 drozdovskaya , m. n. , walsh , c. , van dishoeck , e. f. , et al .
2016 , , 462 , 977 esplugues , g. b. , cazaux , s. , meijerink , r. , spaans , m. , & caselli , p. 2016
, , 591 , a52 evans , n. j. , ii , et al .
2003 , , 115 , 965 evans , n. j. , et al . 2009 , vizier online data catalog , 218 , 10321 fraser , h. j. , bisschop , s. e. , pontoppidan , k. m. , tielens , a. g. g. m. , & van dishoeck , e. f. 2005 , , 356 , 1283 friesen , r. k. , di francesco , j. , bourke , t. l. , et al .
2014 , , 797 , 27 garrod , r. t. 2008 , , 491 , 239 gerakines , p. a. , whittet , d. c. b. , ehrenfreund , p. , et al . 1999 , , 522 , 357 gibb , e. l. , whittet , d. c. b. , boogert , a. c. a. , & tielens , a. g. g. m. 2004 , , 151 , 35 griffin , m. j. , abergel , a. , abreu , a. , et al .
2010 , , 518 , l3 hatchell , j. , wilson , t. , drabek , e. , et al .
2013 , , 429 , l10 jrgensen , j. k. , schier , f. l. , & van dishoeck , e. f. 2005 , , 437 , 501 karssemeijer , l. j. , ioppolo , s. , van hemert , m. c. , et al .
2014 , , 781 , 16 knez , c. , boogert , a. c. a. , pontoppidan , k. m. , et al .
2005 , , 635 , l145 lauck , t. , karssemeijer , l. , shulenberger , k. , et al .
2015 , , 801 , 118 lee , c. w. , & myers , p. c. 1999 , , 123 , 233 li , b. , pei , c .- c . , &
ma , h .- j .
2007 , , 31 , 372 makrymallis , a. , & viti , s. 2014 , , 794 , 45 martin , p. g. , roy , a. , bontemps , s. , et al .
2012 , , 751 , 28 murakawa , k. , tamura , m. , & nagata , t. 2000 , , 128 , 603 murdin , p. , & penston , m. v. 1977 , , 181 , 657 noble , j. a. , dulieu , f. , congiu , e. , & fraser , h. j. 2011 , , 735 , 121 noble , j. a. , fraser , h. j. , aikawa , y. , pontoppidan , k. m. , & sakon , i. 2013 , , 775 , 85 nummelin , a. , whittet , d. c. b. , gibb , e. l. , gerakines , p. a. , & chiar , j. e. 2001 , , 558 , 185 oba , y. , watanabe , n. , kouchi , a. , hama , t. , & pirronello , v. 2010 , , 712 , l174 berg , k. i. , bottinelli , s. , & van dishoeck , e. f. 2009 , , 494 , l13 berg , k. i. , boogert , a. c. a. , pontoppidan , k. m. , et al .
2011 , , 740 , 109 oliveira , j. m. , van loon , j. t. , chen , c .- h .
r. , et al .
2009 , , 707 , 1269 oliveira , j. m. , van loon , j. t. , sloan , g. c. , et al .
2013 , , 428 , 3001 ossenkopf , v. , & henning , t. 1994 , , 291 , 943 pontoppidan , k. m. , et al . 2003 , , 408 , 981 pontoppidan , k. m. , van dishoeck , e. f. , & dartois , e. 2004 , , 426 , 925 pontoppidan , k. m. 2006 , , 453 , l47 pontoppidan , k. m. , et al .
2008 , , 678 , 1005 poteet , c. a. , pontoppidan , k. m. , megeath , s. t. , et al .
2013 , , 766 , 117 qi , c. , berg , k. i. , wilner , d. j. , et al .
2013 , science , 341 , 630 racca , g. a. , vilas - boas , j. w. s. , & de la reza , r. 2009 , , 703 , 1444 reboussin , l. , wakelam , v. , guilloteau , s. , & hersant , f. 2014 , , 440 , 3557 shimonishi , t. , onaka , t. , kato , d. , et al .
2010 , , 514 , a12 sonnentrucker , p. , neufeld , d. a. , gerakines , p. a. , bergin , e. a. , melnick , g. j. , forrest , w. j. , pipher , j. l. , & whittet , d. c. b. 2008 , , 672 , 361 viti , s. , collings , m. p. , dever , j. w. , mccoustra , m. r. s. , & williams , d. a. 2004 , , 354 , 1141 walsh , c. , millar , t. j. , nomura , h. , et al .
2014 , , 563 , a33 ward - thompson , d. , di francesco , j. , hatchell , j. , et al .
2007 , , 119 , 855 watanabe , n. , & kouchi , a. 2002 , , 571 , l173 whittet , d. c. b. , bode , m. f. , longmore , a. j. , adamson , a. j. , mcfadzean , a. d. , aitken , d. k. , & roche , p. f. 1988 , , 233 , 321 whittet , d. c. b. , shenoy , s. s. , bergin , e. a. , et al .
2007 , , 655 , 332 whittet , d. c. b. , cook , a. m. , chiar , j. e. , et al . 2009 , , 695 , 94 whittet , d. c. b. , poteet , c. a. , chiar , j. e. , et al .
2013 , , 774 , 102 yamagishi , m. , kaneda , h. , ishihara , d. , et al .
2011 , , 731 , l20 yamagishi , m. , kaneda , h. , ishihara , d. , et al . 2013 , , 773 , l37 yamagishi , m. , kaneda , h. , ishihara , d. , et al . 2015 , , 807 , 29 zasowski , g. , kemper , f. , watson , d. m. , et al .
2009 , , 694 , 459 | we present maps of the column densities of h@xmath0o , co@xmath0 , and co ices towards the molecular cores b 35a , dc 274.2 - 00.4 , bhr 59 , and dc 300.7 - 01.0 .
these ice maps , probing spatial distances in molecular cores as low as 2200 au , challenge the traditional hypothesis that the denser the region observed , the more ice is present , providing evidence that the relationships between solid molecular species are more varied than the generic picture we often adopt to model gas - grain chemical processes and explain feedback between solid phase processes and gas phase abundances .
we present the first combined solid - gas maps of a single molecular species , based upon observations of both co ice and gas phase c@xmath1o towards b 35a , a star - forming dense core in orion .
we conclude that molecular species in the solid phase are powerful tracers of `` small scale '' chemical diversity , prior to the onset of star formation . with a component analysis approach
, we can probe the solid phase chemistry of a region at a level of detail greater than that provided by statistical analyses or generic conclusions drawn from single pointing line - of - sight observations alone .
[ firstpage ] astrochemistry star - formation
ism . |
weak gravitational lensing produces coherent distortions in the images of background galaxies .
this effect provides a unique method to measure directly the distribution of mass in the universe ( for reviews see schneider et al .
1992 ; narayan & bartelmann 1996 ; mellier 1999 ) .
this technique is now routinely used to map the mass of clusters of galaxies ( for a review , see fort & mellier 1994 ) .
a search for weak lensing by large - scale structure is the subject of much recent and on - going theoretical and observational effort ( eg . villumsen 1995 ; stebbins , et al .
1995 ; kaiser 1996 ; schneider et al . 1997
; van waerbeke et al . 1998 ; refregier et al .
1998 ; see refregier 1999 for a bibliography ) .
the main difficulty lies in the fact that the lensing distortions are small ( @xmath0 for clusters and @xmath1 for large - scale structure ) , thus requiring high precision measurements and tight control of systematic effects .
kaiser , squires and broadhurst ( 1995 , ksb ) have developed a method to correct for the major systematic effects , namely the anisotropy of the point - spread function ( psf ) and camera distortions , and to calibrate the relation between galaxy ellipticities and lensing shear ( for other methods , see also bonnet & mellier 1995 ; schneider & seitz 1995 ) .
further elements of their method were presented in luppino & kaiser ( 1997 , lk ) and in hoekstra et al .
( 1998 , hfks ) .
recently , kaiser ( 1999 ) pointed out that the ksb method had several shortcomings , all stemming from the fact that most psfs encountered in practice are not sufficiently compact .
kaiser then proposed another method based on the explicit construction of the post - convolution shear operator .
another alternative method was recently proposed by kuijken ( 1999 ) . in this different approach
, a sheared and convolved isotropic model is fitted to the galaxy image , so as to derive an estimator for the shear .
these two methods are promising , but both require complete knowledge of the 2-dimensional psf function , while psf measurements are sparse due to the finite number of stellar images . here
, we revisit the ksb method , which has the advantage both of being linear and of relying only on the first multipole moments of the psf and galaxy images .
we focus on the demanding search for weak lensing by large - scale structure .
we show that both the psf and the camera distortions can be corrected for using source moments , as opposed to ellipticities .
we clarify the applicability of some of the approximations made in this method , and show how the weight function for stars can be chosen to be different from that for galaxies .
we derive expressions for the corrections in term of the moments only . by decomposing the moments into spinors
, we derive an explicit relation between the shear and the average ellipticity .
we discuss the shortcomings of the method discussed in kaiser ( 1999 ) and kuijken ( 1999 ) , and test its validity using numerical simulations . as an application
, we consider weak lensing measurements with the hubble space telescope ( hst ) .
the small psf and absence of atmospheric seeing makes hst an ideal instrument for weak lensing measurements ( eg .
kneib et al .
1996 ; hfks ) .
we reproduce the analysis of hfks who studied , in detail , the psf and camera distortion of the wfpc2 camera onboard hst .
we apply these calibrations to the galaxies in the hst survey strip known as the `` groth strip '' ( groth et al .
1995 , rhodes 1999 ) .
in particular , we show how the psf anisotropy and camera distortions depend on the size of the galaxies .
we also discuss the prospects of weak lensing measurements with hst .
a description of our search for weak lensing by large - scale structure with the survey strip will be presented in rhodes , refregier , & groth 1999 ( see also rhodes 1999 ) . because the method is somewhat complex ,
we provide a practical summary in [ overview ]
in which we point to the results and equations which are of direct practical interest . in [ shapes ] , we describe measures of object shapes .
next , in [ distortion ] and [ convolution ] , we show how these measures are affected by the two main classes of deformations , namely convolutions and distortions , and derive explicit expressions to correct for them . in [ comb ] , we study the effect of shear combined with a convolution on the observed ellipticity . in [ validity ] , we discuss the shortcomings of the method , and test its validity using numerical simulations . in [ hst ]
, we apply our method to hst observations .
in particular , we consider measurements of the camera distortion , globular cluster observations , and finally apply our results to the survey strip .
our conclusions are summarized in [ conclusions ] .
the purpose of this method is to provide a measure of the shapes of galaxies , to correct for instrumental effects , and to derive an estimate for the weak lensing shear .
the images of galaxies are assumed to be altered by two kinds of operations : _ distortions _ ( sometimes simply called shear ) , which are intensity - conserving mappings between the source plane and the image plane ( see eq.[[eq : fprime ] ] ) , and _ convolutions _ ( or smear ; see eq.[[eq : iprime_conv ] ] ) .
specifically , we follow ksb and hfks and assume that a galaxy image is altered by a distortion due to gravitational lensing , followed by a convolution by a generally anisotropic psf , and then by a distortion due to the camera optics .
schematically , we can write @xmath2 where @xmath3 and @xmath4 symbolize distortions and convolutions , respectively . here
, we revisit the ksb method .
we show how to correct for the two latter effects and how to calibrate the sought - after effect of lensing .
the synopsis of our method is shown on figure [ fig : synopsis ] .
since the method is somewhat complex , we provide here a summary of the method , and point to the equations which are of direct practical interest .
references to these equations can also be found on this figure .
the basis of the method is to characterize the shapes of galaxies and stars by measuring their multipole moments @xmath5 , @xmath6 , etc .
( eq . [ [ eq : jij_def ] ] ) . to enforce convergence in the presence of image noise
, these moments are measured with a weight function @xmath7 which we choose to be a gaussian of width @xmath8 ( eq . [ [ eq : w_gaussian ] ] ) .
this choice allows us to write all the expressions for the corrections in terms of the moments . unlike ksb ,
we perform all the corrections using moments , and postpone the use of ellipticities ( eq . [ [ eq : epsilon_def ] ] ) until the last step .
this has the advantage of keeping track of the size of the psf and of the galaxies , and thus of reducing the noise resulting from the corrections , if the psf size varies across the field .
the first step consists of deriving the distortion matrix @xmath9 ( eq . [ [ eq : phi_def ] ] ) for the camera distortion .
hfks showed that this could be achieved by considering astrometric shift solutions , such as that of holtzmann et al .
( 1995 ) for the wfpc2 camera on hst . the resulting shear pattern for wfpc2 is shown on figure [ fig : wf_shear ] .
( see discussion in
[ hfks_camera ] about the difference between this figure and the results of hfks ) . in the second step ,
we derive the psf moments from stellar images .
for this purpose , the stellar multipole moments @xmath10 and @xmath11 are measured using an optimally chosen weight - function width @xmath12 .
these moments are then corrected for the weight function to provide ( an approximation to ) the unweighted moments of the psf ( eqs .
[ [ eq : pij_correction],[eq : pijkl_correction ] ] ) .
this allows us to use a different weight - function width @xmath8 for the galaxies , and thus to improve the sensitivity .
the psf moments are then corrected for the camera distortion ( eqs [ [ eq : jij_dcorrection],[eq : jijkl_dcorrection ] ] with @xmath13 ) .
this provides the corrected unweighted psf moments @xmath14 and @xmath15 ( eqs .
[ [ eq : pij_def],[eq : pijkl_def ] ] ) , that can be then be interpolated across the field using low - order polynomial fits for each component . figures [ fig : m4ellipt ] and [ fig : ngc6572 ] show the resulting ellipticities of the wfpc2 psf derived from two globular clusters , while figure [ fig : m4sm ] shows the interpolated psf ellipticities derived from a combination of the two after the moments have been corrected for weighting and camera distortion . in the third step
, we measure the galaxy moments and correct them for instrumental effects . the galaxy moments @xmath5 and @xmath6
are measured using an optimal weight - function width @xmath8 .
they are then corrected for the camera distortion using @xmath16 ( eqs .
[ [ eq : jij_dcorrection],[eq : jijkl_dcorrection ] ] ) .
the psf can be decomposed into an anisotropic and an isotropic part ( eq . [ [ eq : p_decomp ] ] ) .
we correct the galaxy moments @xmath5 for the psf anisotropy , by expanding in powers of the psf ellipticity ( eq . [ [ eq : jij_acorrection ] ] ) .
since , by construction , the isotropic part of the psf is a gaussian ( eq . [ [ eq : pi_gaussian ] ] ) , we can perform the isotropic correction exactly for @xmath5 ( eq . [
[ eq : jij_icorrection ] ] ) .
the fourth order moments @xmath6 are then approximately corrected for the psf ( eq . [ [ eq : jijkl_correction ] ] ) .
this provides us with the corrected galaxy moments @xmath17 and @xmath18 which are now effectively weighted by the new weight - function width @xmath19 ( eq . [ [ eq : omega_g ] ] ) .
the final step consists of measuring the weak lensing shear by averaging over an ensemble of galaxies in a region of the sky . for this purpose ,
we compute the ellipticity @xmath20 of each galaxy from its corrected moments @xmath17 ( eq . [ [ eq : epsilon_def ] ] ) .
the weak lensing shear @xmath21 is then computed from the average ellipticity @xmath22 ( eq . [
[ eq : gamma_epsilon ] ] ) . in this last step
, one should remember to use @xmath19 rather than @xmath8 as the weight - function width .
the relationship between @xmath21 and @xmath23 is greatly simplified by considering the rotation properties of the multipole moments ( see [ rotations ] ) .
note that our method avoids the complications of the pre - smear and post - smear shear susceptibilities discussed in lk and hfks . in
[ hst ] , we discuss the application of our method to the hst survey strip .
in this section , we show how object shapes can be characterised using multipole moments and related quantities .
we also study the rotational properties of the moments and decompose them into spinor representations .
let us consider a source with intensity @xmath24 .
as a first step , we find the centroid @xmath25 of the source by solving @xmath26 where @xmath7 is a weight function introduced to ensure convergence in the presence of noise . in this paper , we will consider a normalized gaussian weight function , @xmath27 which has convenient analytical properties . in practice ,
equation ( [ eq : centroid ] ) can be solved iteratively , by fixing the weight - function width @xmath8 to an initial estimate of the source size ( see [ w_choice ] for a description of our choice of @xmath8 for the hst survey strip ) . to characterize the source shape , we consider the weighted multipole moments of the source intensity , @xmath28 where the origin of the coordinate system was chosen to coincide with @xmath25 .
( by definition , the dipole moment @xmath29 vanishes in this coordinate system ) .
it is usually more convenient to consider the normalized moments @xmath30 the normalized quadrupole moments @xmath5 can be diagonalised as @xmath31 where @xmath32 and @xmath33 are the ( weighted ) major and minor radii , @xmath34 is the position angle measured counter - clockwise from the positive @xmath35-axis , and @xmath36 stands for the transpose operation .
the rotation matrix @xmath37 is defined as @xmath38 inverting this relation yields @xmath39 \\ \tan 2 \alpha & = & \frac{2 j_{12}}{j_{11}-j_{22}}.\end{aligned}\ ] ] to study the rotational properties of the tensors defined above , let us consider a new coordinate system which is rotated counter - clockwise by an angle @xmath40 from the original positive @xmath35-axis . in this new coordinate system ,
the components @xmath41 of the normalized quadrupole moments are related to the unrotated components @xmath5 by @xmath42 , and similarly for tensors of higher order , where @xmath37 is the rotation matrix defined in equation ( [ eq : r_matrix ] ) . in general , any tensor in 2 dimensions can be decomposed into 2-component ( or 1-component for scalars ) @xmath43-spinors @xmath44 which rotate as @xmath45 under this change of coordinates .
( these spinors form irreducible representations of @xmath46 , the rotation group in 2 dimensions ) .
for instance , the quadrupole moment @xmath5 , which , being symmetric , consists of 3 independent components , can be decomposed into a scalar @xmath47 which is the mean square radius of the source , and a spin-2 tensor @xmath48 which is the ellipticity and has been normalized to follow the weak - lensing nomenclature .
the component @xmath49 ( @xmath50 ) corresponds to stretches and compressions parallel to ( at @xmath51 from ) the coordinate axes . by considering infinitesimal rotations
, we can also decompose the fourth - order moment @xmath6 , which has 5 independent components .
we find that it can be decomposed into a scalar , @xmath52 a spin-2 tensor , @xmath53 and a spin-4 tensor , @xmath54 for future convenience , these spinors have been normalized with the weight - function width @xmath8 ( eq . [ [ eq : w_gaussian ] ] ) and the scalar @xmath55 .
these decompositions are useful for simplifying tensors which are averaged over an ensemble of randomly - oriented galaxies ( see [ ellip_shear ] ) .
in this section , we study the effect of distortions on the source moments . in practice , distortions arise from the instrument optics and from weak lensing .
we show how the former can be corrected , and how the latter can be measured by averaging the ellipticities of an ensemble of galaxies .
a distortion is an intensity - conserving mapping between the true position @xmath56 and the observed position @xmath57 of the form @xmath58 the observed intensity @xmath59 is thus related to the true intensity @xmath60 by @xmath61 the local properties of the distortion are quantified by the distortion matrix @xmath62 which is defined as @xmath63 where @xmath64 is the convergence , @xmath21 is the shear , and @xmath65 the rotation parameter .
the convergence @xmath64 describes overall rescalings , and the shear components @xmath66 and @xmath67 describe contractions and dilations parallel to , and at @xmath51 from , the coordinate axes . in the case of weak lensing , @xmath68 is directly related to the second - order derivative of the gravitational potential projected along the line of sight .
while the rotation parameter @xmath65 is expected to be negligible for weak lensing , it is generally not so for instrumental distortions .
let us consider a true position @xmath69 , which is close to a reference position @xmath70 ( eg . the centroid of a source ) . for small distortions (
i.e. if the offset @xmath71 is small compared to the scale on which the distortion mapping varies ) , the corresponding distorted position is given by @xmath72 , where @xmath73 and @xmath74 denotes higher - order derivatives of the distortion mapping .
similarly , the true position corresponding to a distorted position @xmath75 , is given by @xmath76 . by inserting this expression in equation ( [ eq : i ] ) and by letting @xmath77 and @xmath78
, we obtain @xmath79 the quadrupole moment @xmath80 ( eq .
[ [ eq : iij_def ] ] ) for the distorted image @xmath81 ( eq . [ [ eq : fprime ] ] ) is given by @xmath82 after taylor expanding and integrating by parts we get , for a gaussian weight function ( eq .
[ [ eq : w_gaussian ] ] ) , @xmath83k } + i_{ij } \phi_{kk } - \omega^{-2 } i_{ijkl } \phi_{kl } + o(\phi^{2})$ ] , where the unprimed moments correspond to the undistorted moments .
similarly , the distorted monopole moment is related to the undistorted moments by @xmath84 . as a result
, the distorted normalized moments ( eq . [ [ eq : jij_def ] ] ) are given by @xmath85 where the distortion susceptibility tensor @xmath86 is given by @xmath87l } + \omega^{-2 } \left ( j_{ij } j_{kl } - j_{ijkl } \right).\ ] ] the brackets denote the symmetrizer , which , for an arbitrary tensor @xmath88 of rank @xmath89 , is defined by @xmath90 } \equiv \frac{1}{n ! }
\left[a_{i_{1}i_{2 } \ldots i_{n}}+ \mbox{\rm all $ n!$ permutations of } \{i_{1},i_{2 } , \ldots , i_{n}\ } \right].\ ] ] inverting equation ( [ eq : jij_distorted ] ) yields @xmath91 where @xmath92 .
this expression can be used to correct the normalized moments for a known distortion . in principle , the correction for the fourth - order moments @xmath6 can be derived in a similar way .
however , when the weight function @xmath7 is taken into account , the resulting expressions contain sixth - order moments and are very cumbersome .
as we will see in [ hst ] , it is sufficient , in practice , to consider the unweighted ( @xmath13 ) corrections to @xmath6 . in this approximation ,
the corrected fourth - order moments @xmath6 are related to the distorted moments @xmath93 by @xmath94 m } + o(\phi^{2}),\ ] ] where , as before , the brackets denote the symmetrizer ( eq . [ [ eq : symmetrizer ] ] ) .
this expression can be used to correct @xmath93 for instrumental distortions .
we now show how the gravitational shear can be measured by averaging over galaxy ellipticities . for this purpose
, we consider galaxy moments that have been corrected for all instrumental effects , i.e. from instrumental distortion and psf convolution , using the prescriptions presented in the other sections of this paper .
the effect of a weak general distortion @xmath95 ( eq . [ [ eq : phi_def ] ] ) on the ( corrected ) ellipticity of a galaxy can be derived by substituting the distorted moments @xmath41 ( eq . [ [ eq : jij_distorted ] ] ) into the definition of the ellipticity @xmath20 ( eq . [
[ eq : epsilon_def ] ] ) .
this results in a relation between the distorted ellipticity @xmath20 and the distortion parameters @xmath96 of the form @xmath97 , with @xmath98 where @xmath99 and @xmath100 are the spinors defined in [ rotations ] .
note that @xmath65 is kept here for completeness , but is expected to vanish for weak lensing distortions . to measure the shear @xmath21
, we then average over an ensemble of galaxies which are assumed to be randomly oriented , intrinsically .
thus , all we need is the rotational average of the above relation , which we write as @xmath101 . after discarding all terms which are not rotationally invariant
, we obtain @xmath102 where @xmath103 and @xmath104 , @xmath105 , and @xmath106 .
note that , to this order , the convergence @xmath64 ( and the rotation parameter @xmath65 ) do not affect the mean ellipticity @xmath107 .
since we do not expect the galaxy population to have a preferred handedness , we can also discard terms which are not parity invariant .
it is easy to check that @xmath108 is such a term since it changes sign when it is transformed to a left - handed coordinate system ( eg .
@xmath109 ) .
we are therefore left with a remarkably simple relation between the mean ellipticity @xmath23 and the shear @xmath21 given by @xmath110 where @xmath111 is the shear susceptibility .
this expression agrees with equation ( b13 ) in ksb , as corrected by hfks .
the inverse relation , @xmath112 where @xmath113 can be used to measure the shear from the averaged ellipticity .
* unweighted moments : in this case , the effect of a general distortion @xmath95 ( eq . [ [ eq : phi_def ] ] ) on the ellipticity is more tractable .
we find that the distorted ellipticity @xmath114 is given by @xmath115 where @xmath20 is the undistorted ellipticity , and the levi - civit symbol @xmath116 is defined by @xmath117 .
when averaged over an ensemble of randomly - distributed sources , this reduces to @xmath118 in agreement with equation ( [ eq : epsilon_gamma ] ) in the @xmath119 limit .
* circular source , unweighted moments : if we make the further simplifying assumption that the undistorted source is circular ( @xmath120 ) , the distorted ellipticity ( eq . [ [ eq : epsilon_dist ] ] ) becomes @xmath121 it is useful to plot the distortion ellipticity @xmath122 ( rather than @xmath21 ) as a function of position on the chip , as a measure of the effect of the camera distortion ( see figure [ fig : wf_shear ] ) .
* radial displacement : let us consider the case where the displacement field ( eq . [ [ eq : deltax ] ] ) is radial , i.e. where @xmath123 , where @xmath124 , and @xmath125 is an arbitrary function .
it is easy to show that , in this case , the distortion tensor ( eq . [ [ eq : phi_def ] ] ) is @xmath126 the corresponding distortion ellipticity ( eq . [
[ eq : epsilon_phi ] ] ) is @xmath127 where @xmath128 is the unit radial ellipticity field . from this expression , it is easy to see that @xmath129 will be radial ( tangential ) if @xmath130 is positive ( negative ) .
in this section , we study the effect of convolution by a weakly - anisotropic psf on the source moments .
the psf can be decomposed into an isotropic and an anisotropic part .
we show how each part can be corrected for , and how the psf moments can be derived from stellar images .
let us consider the case where the true galaxy image @xmath131 is convolved by a kernel @xmath132 .
we take @xmath132 to be normalized so that @xmath133 , and centered so that @xmath134 .
the observed image is given by @xmath135 so that the observed moments are @xmath136 there are three angular scales in this equation , namely @xmath8 , @xmath137 , and @xmath32 , corresponding to the size of the weight function @xmath138 , the psf kernel @xmath139 , and the ( unconvolved ) source @xmath140 , respectively .
to simplify this expression , we need to make an expansion with respect to the ratio between two of these angular scales .
we choose to expand with respect to @xmath141 , thereby assuming that the window function width is much larger than that of the psf . in practice ,
the weight function scale is always chosen to be at least as large as the source size , i.e. @xmath142 . to be conservative , we thus take @xmath143 to collect the terms in the expansion .
ksb instead , effectively expanded in powers of @xmath144 .
it is interesting to note that the expression we derive below for @xmath145 is nevertheless identical to theirs , to second order in @xmath141 . after a change of variable ( @xmath146 ) , and a taylor expansion of @xmath147 about @xmath148
, the previous equation becomes @xmath149k } - \frac{1}{2\omega^{2 } } i_{ij } p_{kk } + \frac{1}{2 \omega^{4 } } i_{ijkl}p_{kl } + o\!\left(\frac{g^4}{\omega^2}\right)$ ] , where the brackets stand for the symmetrizer ( eq . [ [ eq : symmetrizer ] ] ) , and @xmath150 and @xmath151 are the ( undistorted ) weighted moments ( eq . [ [ eq : iij_def ] ] ) . as before ,
we take the weight function @xmath7 to be a normalized gaussian ( eq . [ eq : w_gaussian ] ) .
the moments @xmath14 are the ( unweighted ) moments of the convolution kernel , i.e. @xmath152 as stated above , this agrees , to this order , with ksb , as corrected for a factor of @xmath153 by hfks . in [ stars ] , we will show how the unweighted psf moments @xmath14 can be derived from weighted stellar moments . in a similar fashion , it is easy to show that the convolved monopole moment is @xmath154 . as a result , the convolved normalized moments ( eq . [ [ eq : jij_def ] ] ) are given by @xmath155 where the convolution susceptibility tensor @xmath156 is given by @xmath157l } + \frac{1}{2 \omega^4 } \left [ j_{ijkl } - j_{ij } j_{kl } \right].\ ] ] inverting these equations yields @xmath158 where @xmath159 .
the last two equations can be used to correct the observed moments @xmath41 for the psf , and only require knowledge of the observed quadrupole and fourth - order moments , @xmath5 and @xmath6 , of the galaxy , and of the quadrupole moments @xmath14 of the psf .
we now show how this approximation can be used to correct for a weakly anisotropic psf . in practice
, it is often required to include galaxies with size @xmath32 ( and therefore weight - function width @xmath8 ) only marginally larger than the psf size @xmath137 .
the psf correction given by equation ( [ eq : jij_correction ] ) , is thus not directly applicable , as the expansion series do not converge sufficiently fast .
we can nevertheless apply the above correction scheme when the psf is weakly anisotropic and sufficiently compact . in this case , we write the unweighted psf moments as @xmath160 where @xmath137 is the psf radius , and @xmath161 is the psf ellipticity , which is assumed to be small .
we can then decompose the kernel @xmath162 into the convolution of an isotropic part @xmath163 with an anisotropic part @xmath164 , as @xmath165 it is easy to show that this implies that @xmath166 , where @xmath167 and @xmath168 are the unweighted moments of @xmath169 and @xmath170 , respectively . without loss of generality , we further require that @xmath171 be a normalized circular gaussian with standard deviation @xmath137 .
this implies that @xmath172 we now show how to correct each of these components in turn .
if we consider a convolution with the anisotropic kernel @xmath173 alone , the corrected moments ( eq . [ [ eq : jij_correction ] ] ) become @xmath174 where , as before , @xmath159 is defined in equation ( [ eq : c_tensor ] ) .
the residual terms are now supressed by the factor @xmath175 , and are thus negligible in practice .
as explained earlier , the above approximation can not be applied to correct for the isotropic part of the psf . however , since , by construction , the isotropic part @xmath169 of the kernel is a gaussian , we can perform the isotropic correction exactly . indeed , by inserting the form of @xmath169 ( eq . [ [ eq : pi_gaussian ] ] ) into equation ( [ eq : iij_conv ] ) and by integrating twice , we find the convolved moments to be @xmath176 where @xmath177 and the superscript @xmath178 and @xmath179 in the moments @xmath150 , @xmath180 , and @xmath145 indicate the standard deviation of their respective weight functions .
the convolved monopole moment is simply @xmath181 .
consequently , the convolved normalized moments are @xmath182 this convenient relation allows us to relate the convolved moments @xmath183 with the unconvolved moments @xmath184 , but this time weighted with the wider weight standard deviation @xmath19 . for later reference , we explicitly write the inverse relation @xmath185 which can be used to correct for the isotropic part of the psf . as in the case of distortions ( [ jijkl_distortion ] ) , the general expression for the correction of the fourth - order moments @xmath6 for convolutions contains sixth - order moments and is very cumbersome . here
again , it is usually sufficient to consider the unweighted ( @xmath13 ) corrections to @xmath6 . with this approximation ,
the corrected moments @xmath6 are related to the convolved moments @xmath41 and @xmath93 by @xmath186 } ' + 6 p_{[ij } p_{kl]},\ ] ] where @xmath187 is the ( unweighted ) fourth - order psf moment , and the brackets denote the symmetrizer ( eq . [ [ eq : symmetrizer ] ] ) .
this expression can be used to correct @xmath93 for convolutions . in the next section
, we will show how @xmath15 can be estimated from stellar moments .
stars are point - like and therefore have an intensity profile given by @xmath188 , where @xmath189 is the flux . the intrinsic moments ( eq .
[ [ eq : iij_def ] ] ) of a star are thus @xmath190 and @xmath191 .
we allow for the possiblity that the weight function , @xmath192 , for the stars be different from that for galaxies , @xmath193 .
after convolution ( see eq . [ [ eq : jij_correction ] ] and above ) , the moments become @xmath194 and @xmath195 , where @xmath15 was defined in equation ( [ eq : pijkl_def ] ) . as a result ,
the normalized moments become @xmath196 + o\!\left ( \frac{g^{6}}{\omega_{*}^{4 } } \right ) .\ ] ] it is also easy to show that the observed fourth - order moments are given by @xmath197 , while their normalized version is @xmath198 we can invert these equations to obtain @xmath199 + o\!\left ( \frac{g^{6}}{\omega_{*}^{4 } } \right),\ ] ] and @xmath200 with these expressions , the unweighted psf moments @xmath14 and @xmath15 can be derived from the observed stellar moments @xmath201 and @xmath202 .
these can then be corrected for the camera distortion using equations ( [ eq : jij_dcorrection],[eq : jijkl_dcorrection ] ) with @xmath13 , and then used in equations ( [ eq : jij_acorrection],[eq : jij_icorrection],[eq : jijkl_correction ] ) to correct the galaxy moments . for this purpose , @xmath14 and @xmath15
need to be interpolated across the chip . in practice , this can be done by fitting a low order polynomial to each component separately .
figure [ fig : m4sm ] shows the psf ellipticities for the wfpc2 camera derived from globular cluster observations ( figures [ fig : m4ellipt ] and [ fig : ngc6572 ] ) , after correction and interpolation .
if the star and galaxy weight functions are equal ( @xmath203 ) , then one only needs to keep the first term in equation ( [ eq : pij_correction ] ) .
however , more accurate measurements of the stellar shapes can be achieved by taking a narrower weight function . in addition
, it is desirable to avoid recomputing the stellar moments for each value of @xmath8 , which is often taken to be a function of the galaxy size .
one would then choose @xmath204 , making equation ( [ eq : pij_correction ] ) converge more slowly than equation ( [ eq : jij_correction ] ) . in this case , one would thus need to keep the second term in equation ( [ eq : pij_correction ] ) and to ensure that the residual error ( @xmath205 ) is acceptable compared to that for equation ( [ eq : jij_correction ] ) ( @xmath206 ) .
* unweighted moments : in this case , the convolved moments become ( eq . [ [ eq : jij_convolved ] ] with @xmath13 ) @xmath207 .
thus , the observed square radius is @xmath208 and the observed ellipticity is @xmath209 , where @xmath210 and @xmath161 are defined by equation ( [ eq : pij_decomp ] ) .
it is sometimes useful to consider the moments @xmath211 , which have been deconvolved from the isotropic part of the psf only .
the associated radius is @xmath212 , while the associated ellipticity is @xmath213 . * unweighted moments , circular source : in this case , @xmath214 and
thus the observed ellipticity becomes @xmath215 while the isotropically deconvolved ellipticity becomes @xmath216 we have introduced the pre- and post - deconvolution reduction factors @xmath217 and @xmath218 .
these expressions are useful to estimate the effect of the psf anisotropy on a galaxy .
in practice , an image is deformed by a series of distortions and convolutions .
the full treatment of the combined effect of a distortion and convolution is impractical given the complexity of the @xmath86 and @xmath156 tensors ( eqs .
[ [ eq : c_tensor],[eq : d_tensor ] ] ) .
however , the general behavior of the ellipticity is captured by considering the simplified case of a circular unweighted source which is subject to a weak distortion .
this provides a good model of ellipticities observed in practice ( see [ strip ] below ) . in the unweighted case ( @xmath219 )
, the normalized quadrupole moments @xmath41 of an image deformed by a convolution followed by a distortion are ( see eqs .
[ [ eq : jij_correction],[eq : jij_dcorrection ] ] with @xmath119 ) @xmath220 where @xmath14 are the psf moments ( eq . [
[ eq : pij_def ] ] ) , @xmath95 is the distortion matrix ( eq . [
[ eq : phi_def ] ] ) , and @xmath5 are the undeformed moments .
we will assume that the psf is weakly anisotropic , i.e. that @xmath161 defined in equation ( [ eq : pij_decomp ] ) is small .
then , for an intrinsically circular source ( @xmath221 ) , the ellipticity @xmath114 of the deformed source is @xmath222 where the deformation matrix @xmath95 was parametrized as in equation ( [ eq : phi_def ] ) , and @xmath137 is the psf radius defined in equation ( [ eq : pij_decomp ] ) .
notice that the convergence @xmath64 and rotation parameter @xmath65 do not appear in this first order expression .
the observed square radius @xmath223 is given by @xmath224 for stars ( @xmath225 ) , the observed ellipticity @xmath226 becomes @xmath227 we can thus rewrite @xmath114 in terms of observables as @xmath228 2 \gamma_{i } + \left ( \frac{g}{d ' } \right)^{2 } \epsilon_{i}^ { * } + o((\phi,\epsilon^{p})^2 ) .
\label{eq : comb}\ ] ] this expression is useful for computing the expected deformation of objects as a function of their observed size .
this simplified model agrees well with the observed deformations in the survey strip ( see [ strip ] ) .
as was recently pointed out ( kaiser 1999 ; kuijken 1999 ) , the ksb method has several shortcomings which limit its validity and accuracy .
first , the method requires a decomposition of the psf into the convolution of an isotropic part and a compact anisotropic part . for most
psfs encountered in practice , this decomposition is formally ill - defined and the anisotropic part is not necessarily sufficiently compact .
another problem results from the fact that most psfs do not fall off fast enough for the second moments ( and higher moments ) to converge .
consequently , the weighted second moments of the psf depend strongly on the size of the window function .
these problems are particularly severe for the psf of hst which has broad wings extending beyond a central core , and thus cast doubts on the validity of the numerous weak - lensing analyses based on hst observations .
since our method is based on the same principles as that of the ksb method , it also suffers from the same shortcomings .
in addition , we have made the further assumption that , in the decomposition of the psf , the isotropic part can be taken to be a gaussian ( see [ psf_decomposition ] ) .
this has the advantage of greatly simplifying the deconvolution , but can arguably produce further inaccuracies .
these problems are certainly worrisome , and might eventually be solved by considering different methods such as those proposed by kaiser ( 1999 ) and kuijken ( 1999 ) . however , the ksb approach has the advantage of being relatively simple and of requiring only a small amount of information about the psf and galaxy shapes , namely the first few multipole moments .
moreover , these shortcomings are mainly formal in nature . in practice ,
one indeed always measures moments with a weight function width that is close to the width of the central core of the psf , and which therefore do not diverge .
for instance , for the hst psf , the secondary sidelobes which compose the extended wings are smaller in amplitude than the central core by about one order of magnitude .
since the central core dominates in the convolution , one therefore expects that the corrections will be approximately correct . in the following paragraphs
, we describe numerical simulations designed to test this assertion quantitatively . to test the weak link in our method , namely the psf correction
, we performed a series of numerical simulations .
we first generated a wfpc2 psf using the tiny tim software ( krist & hook 1997 ) .
we chose the psf to be particularly non - circular by placing it at the lower - right - hand corner of chip 2 , at pixel coordinate ( 100,100 ) ( see [ hst ] for a description of the wfpc2 camera and of its psf ) .
we oversampled the psf by using pixels which are 10 times smaller than the wfpc2 pixels , and did not add any noise to the images .
this isolates the systematic deconvolution errors from the random errors produced by pixelization and noise . to maintain uniformity
, we will nevertheless quote angular sizes in wfpc2 pixels .
we then measured the weighted moments @xmath14 of the psf as described in
[ stars ] , using a range of weight function widths @xmath12 . for @xmath229 pixels
, we found the psf size and ellipticity components ( eq . [ [ eq : pij_decomp ] ] ) to be @xmath230 pixels , and @xmath231 .
as expected , we found both @xmath137 and @xmath20 to diverge as @xmath12 increases .
fo the rest of the simulations , we set @xmath229 pixels , which is the stellar weight width that we used in our analysis of the survey strip ( see [ w_choice ] ) .
we then convolved the psf with elliptical gaussian `` galaxies '' of various intrinsic sizes @xmath55 and ellipticities @xmath20 . since the source of the problem lies with the psf shape , this simplified galaxy model is sufficient for our purposes .
we measured the moments of the convolved gaussian using a weight function width of @xmath232 pixels , just as we did for the analysis of the survey strip ( [ w_choice ] ) .
we then applied our psf correction method ( [ convolution ] ) to obtain the corrected ellipticity @xmath233 , weighted with an effective width @xmath19 ( eq .
[ [ eq : omega_g ] ] ) . for an elliptical gaussian source , it is easy to show that the true weighted ellipticity is @xmath234^{-1}$ ] .
the error @xmath235 induced by the psf correction on the galaxy ellipticity is thus @xmath236 we computed this residual error for a range of intrinsic galaxy sizes and ellipticities .
the results of these simulations are presented in the next section . for the range of galaxy shapes which we considered ( @xmath237 pixels and @xmath238 ) , we found that , for a given galaxy size @xmath55 , the ellipticity error @xmath235 is very well approximated by @xmath239 where @xmath240 is a constant , @xmath241 is the unit radial ellipticity vector , and @xmath242 is a function describing the radial behavior . the constant term @xmath240 is shown in figure [ fig : errors]a for several relevant galaxy sizes , and has a modulus less than .004 for @xmath243 pixels .
figure [ fig : errors]b shows the radial term @xmath242 as a function of the unweighted galaxy ellipticity modulus @xmath244 , for the same galaxy sizes .
the radial term depends strongly on the galaxy size @xmath55 and ellipticity @xmath244 , and has an amplitude of several percent for moderate ellipticities ( @xmath245 ) and sizes ( @xmath243 pixels ) .
this amplitude is the size of the systematic error made in measuring the ellipticity of a single galaxy .
however , for weak lensing measurements , one is not eventually interested in the ellipticity of a single galaxy , but in measuring a global change in ellipticity averaged over a galaxy ensemble . let us consider a change in ellipticity @xmath246 produced by lensing produced by weak lensing has a weak dependence on @xmath20 ( see eq.[[eq : epsilon_dist ] ] ) . while this effect is only important for large ellipticities , the ensemble average is dominated by small ellipticities . for the purpose of estimating the size of the average correction error , we can thus ignore this small effect and consider a constant ellipticity shift . ] .
the error in the measurement of @xmath247 is @xmath248 where the brackets refer to the average over the galaxy ensemble .
for the functional form of equation ( [ eq : error_model ] ) , this becomes @xmath249 where @xmath250 $ ] .
the first term was plotted on figure [ fig : errors]a .
the second term is shown as a function of @xmath244 and @xmath55 in figure [ fig : errors]c , for an ellipticity change of @xmath251 . since , for our sample @xmath252 ( eq . [ [ eq : epsilon_gamma ] ] ) , this value of @xmath247 corresponds to a shear of @xmath253 and is therefore representative of shear signals expected from large - scale structure . for moderate ellipticities ( @xmath245 ) and sizes ( @xmath243 pixels )
, the second term has an amplitude less than 0.004 .
interestingly , both terms in equation ( [ eq : delta_delta_e ] ) tend to average out , when the distribution of galaxy sizes is considered . indeed , in our treatment of the survey strip
( see [ budget ] ) , we selected a sample of `` large '' galaxies with @xmath243 pixels ( and @xmath254 , which is not always satisfied because of noise ) . the average size for this sample is @xmath255 pixels , which is close to the value of @xmath55 for which both terms change sign ( see figure [ fig : errors ] ) .
the ellipticity dispersion for this sample is @xmath256 .
taking @xmath257 and @xmath258 to be representative for this sample , we find the average residual errors to be @xmath259 .
this cancellation may , however , depend on the galaxy profile , while we have only considered gaussians in these simulations . to be conservative
, we therefore take the residual error for our sample to be about 0.004 , which is the error before averaging over @xmath55 values .
this is close to the residual error estimated from the anisotropic correction of the stars in our analysis of globular cluster fields ( [ budget ] ) .
we also studied how the choice of the weight function widths @xmath8 and @xmath12 affect these results .
we found that for moderate values ( @xmath260 and @xmath261 ) , the magnitude of the residual errors did not change significantly from the above .
however , the errors were much larger for @xmath8 and @xmath12 outside of these ranges .
this is , of course , a consequence of the divergences present in the psf moments .
we conclude that , the ellipticity error produced by the psf correction can be several percent for an individual galaxy observed with wfpc2 . however , it is only about 0.004 when averaged over a galaxy ensemble with @xmath243 pixels , provided moderate values of the weight function widths are used .
this does not alleviate the serious formal shortcomings of the method , and the need for a search for more robust methods ( see kaiser ( 1999 ) ; kuijken ( 1999 ) ) .
as an application of our method , we consider weak - lensing measurements with the wfpc2 camera onboard hst .
the camera consists of 3 @xmath262 pixel chips , with a pixel size of 0.1 arcsec . with the f814w filter
, the psf has a fwhm of about 0.09 arcsec .
the instrumental effects for wfpc2 have been studied in detail by hfks .
they showed that the two main systematic effects for this instrument are the camera distortion and the psf .
we repeat their analysis by considering each of these effects in turn .
we then apply these results to the galaxies found in the survey strip .
the camera distortion is caused mainly by the flat fielder , a refractive element in the wfpc2 .
this effect has been studied by holtzman et al .
( 1995 ) , who quantified this effect using astrometric measurements of globular cluster observations .
they modeled the distortion by a cubic polynomial in chip position for each chip .
the polynomial coefficients can be used to derive a transformation matrix @xmath263 between the wfpc2 chip coordinates ( which are the distorted coordinates @xmath57 in eq . [ [ eq : deltax ] ] ) and the undistorted coordinates ( @xmath56 in the same equation ) .
the distortion field is shown in figure [ fig : wf_shear ] .
the quantity plotted is @xmath264 , the ellipticity induced by the distortion field @xmath95 on an intrinsically circular source ( see eq .
[ [ eq : epsilon_phi ] ] ) .
the pattern is radial and increases in magnitude with distance from the chip center .
this differs from the results of hfks who derived a tangential rather than radial camera distortion field .
the radial nature of the pattern is confirmed by the fact that , using the coefficients of holtzman et al .
( 1995 ) in equation ( [ eq : ephi_radial ] ) , we find @xmath265 .
( this can also be confirmed by drawing a box connecting the heads and tails of four adjacent arrows in figure 15 in holtzman et al . )
[ hfks_camera ] to study the profile of the ellipticity pattern , it is useful to define the rotated ellipticity by @xmath266 where @xmath20 is the ellipticity of an object in the chip frame , @xmath267 is the rotation matrix defined in equation [ eq : r_matrix ] , and @xmath40 is the polar angle of the source position about the center of the chip , measured counter - clockwise from the x - axis .
a positive ( negative ) value of @xmath268 corresponds to a radial ( tangential ) ellipticity pattern .
a positive ( negative ) value of @xmath269 corresponds to an anti - clockwise ( clockwise ) swirl pattern .
the meaning of @xmath270 is illustrated in figure [ fig : e_rotate ] .
the profile of the rotated distortion ellipticity is shown as the dot - dashed line on figure [ fig : inst ] .
the mean ellipticity averaged over the 3 chips is listed in table [ tab : stats ] .
the mean rotated ellipticities are found to be @xmath271 and @xmath272 , while the mean absolute ellipticity @xmath23 is less than 0.001 .
the effect of the camera distortion is thus small , but nevertheless comparable to the lensing signal expected from large - scale structure .
the psf is affected by diffraction by the telescope and scattering of light from different parts of the telescope and the optics .
the psf depends both on wavelength and time .
the time dependence has two components ( krist and hook , 1997 ) .
the first one is due to orbit - to - orbit `` breathing '' of the telescope .
as the telescope orbits the earth every 90 minutes , it passes into and out of sunlight .
the heating and cooling of the telescope changes its size , and thus its focus and psf .
the second time - dependent factor arises from a change in the focus of the telescope over longer periods of time .
this change is produced by the outgassing of the graphite epoxy truss which supports the primary mirror .
approximately every six months , the secondary mirror is moved to bring the telescope back into optimal focus .
the time - dependence of the focus position is summarized in figure [ fig : focus ] . as a first step
, we used the tiny tim software to model the psf of the hst ( krist & hook 1997 ) .
tiny tim takes into account the diffraction of light by the telescope to create a psf , given the instrument ( wfpc2 in our case ) , chip number , chip position , filter , spectrum of the object being modeled , and the focus of the telescope .
tiny tim does not include the geometric distortion discussed above .
we used tiny tim to model the psf across the three wfpc2 chips .
we chose the f814w filter and the spectrum of the objects to be similar to that of a type g star ( b@xmath273v=0.619 ) , which is typical of galaxies in the survey strip ( see [ strip ] ) .
we measured the moments ( eq . [ [ eq : jij_def ] ] ) and weighted them with a circular gaussian ( eq . [ [ eq : w_gaussian ] ] ) .
for all stellar measurements , we adopted a weight - function width of @xmath229 pixels , which resulted in an optimal sensitivity . to study the variation of the psf across the chip , we used tiny tim to create a grid of psfs .
these psfs were generated for a telescope with a mirror at a focus position of @xmath274 microns , the value of the focus when the survey strip was taken ( see [ strip ] ) .
the resulting ellipticity pattern is shown in figure [ fig : tt ] .
the corresponding ellipticity profile averaged over all chips is shown as the central dashed line in figure [ fig : inst ] . in these figures , we have added the effects of the detector shear shown in figure [ fig : wf_shear ] , so that a direct comparison can be made between these simulations and the globular cluster measurements which are presented below .
the predicted psf anisotropy is quite large ( @xmath275 at the edge ) and varies significantly across each chip and from chip to chip .
to test the effect of focus changes , we created tiny tim psfs at focus values of + 2.0 microns and -5.0 microns , the full range of telescope focus ( see figure [ fig : focus ] ) .
the ellipticity profiles are shown are the upper and lower dashed curves in figure [ fig : inst ] .
the chip - averaged ellipticities for each focus value are listed in table [ tab : stats ] .
extreme focus changes from the central value of @xmath274 microns produce variations of about 0.002 for @xmath276 and about 0.006 for @xmath277 and @xmath278 , when averaged over all chips . for a more direct measurement of the psf
, we obtained archival images of the globular cluster m4 observed by richer et al.(1995 ) .
we also obtained archival images of the globular cluster ngc6572 .
this was the method of calibration used by hfks .
it is fortuitous that the m4 , ngc6572 , and the survey strip data all have focus values of approximately @xmath274 microns ( see figure [ fig : focus ] ) .
the m4 data consisted of 3 different fields in each cluster .
each field had 8 different exposures : four pairs of dithered pointings separated by very nearly 2 pixels .
other dithers were separated by non - integer numbers of pixels .
we shifted the dithers which are separated by integer numbers of pixels and compared them pairwise to detect cosmic rays .
a pixel was flagged as a cosmic ray if its flux was significantly above that of the corresponding pixel in its pair image .
we selected stars from the m4 exposures by excluding saturated stars and those that are faint and dominated by noise .
we also excluded stars confused by close neighbors , and stars containing a cosmic ray flagged pixel within a 5 pixel radius .
our resulting sample consisted of about 160 stars per chip .
we then averaged the three moments @xmath279 measured in each of the eight exposures .
the analysis of the ngc6572 data was simplified by the fact that we had only one field and the images were not dithered .
the ellipticity fields for m4 and ngc6572 are shown in figures [ fig : m4ellipt ] and [ fig : ngc6572 ] , respectively . the ellipticity profile for each globular cluster
is shown as the upper and lower solid line in figure [ fig : inst ] , respectively .
the chip - averaged ellipticities for each globular cluster data set are listed in table [ tab : stats ] .
the m4 and ngc6572 ellipticity patterns are qualitatively similar , and both show tangential ellipticities of the order of 0.1 at the edge of all chips .
the m4 and ngc6572 average ellipticities @xmath276 are @xmath280 and @xmath281 , respectively , and are both larger than that predicted by tiny tim ( about @xmath282 ) .
the fact that the mean tangential ellipticity for ngc6572 is larger than that for m4 by about 0.013 , shows that the psf varies substantially over periods as short as a few days .
this change is larger than that predicted by tiny tim for extreme focus changes ( see [ tiny tim ] ) .
moreover , the globular cluster ellipticities are qualitatively different from that predicted by tiny tim .
for example , figure [ fig : m4sm ] shows a `` teardrop '' shape to the m4 ellipticity pattern of chip 2 , with a cusp at the top of the chip .
in contrast , the tiny tim ellipticity pattern of the same chip , shown in figure [ fig : tt ] has a cusp on the right - hand side .
small changes in this cusp result by changing the focus of the telescope in tiny tim , but it does not change the qualitative pattern of the ellipticities . because of these shortcomings , we used the observed globular cluster moments for the corrections , rather than the tiny tim predictions . to reduce patchiness , we combined the m4 and ngc6572 stars , and corrected the moments for the weight function ( eqs .
[ [ eq : pij_correction],[eq : pijkl_correction ] ] ) and for the camera distortion ( eqs [ [ eq : jij_dcorrection],[eq : jijkl_dcorrection ] ] with @xmath13 ) to obtain the psf moments @xmath14 and @xmath15 .
we then fitted a fifth order polynomial in chip position for each moment component and for each chip .
because the star counts in the two globular cluster are close to one another ( i.e. about 160 per chip ) , this amounts to giving equal weight to each of them .
the ellipticity profile for the combined globular clusters is shown as the central solid line in figure [ fig : inst ] .
figure [ fig : m4sm ] shows the resulting fitted ellipticity pattern of the psf . the mean psf radius for the combined m4/ngc6572 was found to be @xmath283 pixels ( see eq . [
[ eq : pij_decomp ] ] ) .
this is much smaller than the formal psf radius that one would measure using a tiny tim psf with @xmath284 . as explained in [ shortcomings ] , the quadrupole moments indeed formally diverge because of the extended wings of the hst psf .
our value is however commensurate with the psf fwhm ( 0.9 pixels ) , and should thus be considered as an effective psf radius .
[ psf_radius ] the survey strip is a set of 28 contiguous pointings with the hst in two colors , v ( f606w ) and i ( f814w ) ( groth et al.1995 ; rhodes 1999 ) .
the images were taken in march and april 1994 with the wfpc2 .
the strip has about 10,000 galaxies down to @xmath285 and covers an area of about 108 square arcmin over a @xmath286 region .
the strip has already proved a useful data set in exploring the distribution of luminous matter through number counts and the 2-point angular correlation function ( rhodes , groth and refregier 1998 ; rhodes 1999 ) .
we are in the process of exploring the large scale distribution of matter through weak lensing of the strip galaxies ( rhodes 1999 ; rhodes , refregier , & groth 1999 ) . here
, we use the strip galaxies to test our shear measurement method .
for this purpose , we used the i images only .
the catalog of galaxies was created using the faint object classification and analysis system ( focas ) within iraf ( jarvis & tyson 1981 ) .
an object was considered as detected if two contiguous pixels were more than @xmath287 above the sky background .
however , an object was included in the final catalog only if both the i and v images had a detection within an error box of approximately 5 pixels ( 0.5 arcsec ) .
this resulted in a sample of 9448 galaxies with @xmath288 .
the moments of the galaxies were calculated as described above ( eq . [ [ eq : jij_def ] ] ) .
the variable weight function was chosen to be @xmath289 , where @xmath290 is the detection area calculated by focas .
the lower bound of @xmath8 was chosen to match the width @xmath12 used for stars .
[ w_choice ] the importance of the camera distortion and of the psf depends on the radius @xmath55 of the galaxy ( eq . [ [ eq : d_def ] ] ) .
we thus considered subsamples of small and large galaxies , with @xmath291 pixels and @xmath292 pixels respectively . here
@xmath293 is the root mean squared observed radius ( eq . [ [ eq : d_obs ] ] ) and @xmath283 pixels is the psf radius ( see [ psf_radius ] ) .
because of the noise , a fraction of the galaxies have unphysical ellipticity values ( @xmath294 ) . because these outliers would dominate ellipticity statistics , we only retained galaxies with @xmath254 .
the mean observed radius and reduction factor ( eq . [ [ eq : fd ] ] )
are @xmath295 pixels and @xmath296 , and @xmath297 pixels and @xmath298 , for the small and large galaxies , respectively .
each subsample respectively comprises 41% and 77% of the total number of @xmath288 galaxies .
the ellipticity profile for each subsample is shown as a solid line in figures [ fig : er_small ] and [ fig : er_large ] .
also shown are the asymptotic cases , namely the combined m4/ngc6572 stars ( @xmath225 ) , and the camera distortion ellipticity ( @xmath299 ) .
the dashed line shows the prediction of the simplified model described in [ comb ] ( see eq .
[ [ eq : comb ] ] ) .
the small galaxy sample shows a tangential ellipticity which increases with radius from the chip center , in agreement with the model prediction .
the large galaxies , on the other hand , do not display any significant tangential ellipticity , again in agreement with the model . in both cases , @xmath300 is consistent with zero . using the camera distortion matrix derived in [ hst_distortion ] and the psf moments derived in [ hst_psf ] , we then corrected the strip galaxies .
the ellipticity profiles at different stages of the correction for the small and large galaxy subsamples are shown in figures [ fig : correction_small ] and [ fig : correction_large ] . since the camera distortion is radial , its correction increases the tangential ellipticity of the galaxies .
on the other hand , the correction for the psf , which is mainly tangential , reduces the tangential component of the ellipticity . for both subsamples ,
the resulting corrected profile is consistent with 0 .
as expected , the psf correction has a smaller effect on the large galaxies than on the small galaxies .
the effect of the corrections for the chip - averaged ellipticities are summarized in table [ tab : stats ] . to ensure statistical independence
, we considered in this table a subsample of large galaxies with @xmath301 pixels ( and @xmath254 ) .
this subsample contains 51% of the total number of galaxies with @xmath288 , corresponding to a surface density of @xmath302 arcmin@xmath303 .
their mean observed radius is @xmath304 pixels , corresponding to a post - deconvolution reduction factor of @xmath305 ( eq . [ [ eq : fd_deconv ] ] ) . as a test of the correction algorithm
, we also corrected the globular cluster stars , exactly as we corrected the galaxies .
( we have not corrected either the stars or the small galaxies for the isotropic psf , as this would produce diverging ellipticities ) .
the results for the corrected stars are also shown in the table .
the self - corrected m4/ngc6572 stars have a residual ellipticity of @xmath306 and @xmath307 .
this is a measure of the errors resulting from the approximations in our correction method and from the fit to the stellar moments .
another estimation of the errors from the method was presented in [ validity ] . by performing numerical simulations
, we showed that the residual ellipticity errors in the psf correction was 0.004 , after averaging over a galaxy ensemble with @xmath243 pixels .
given the agreement of these two estimates , we take the error in the correction method to be about 0.004 . as we noted in
[ hst_psf ] , the m4 and ngc6572 ellipticities differ by about 0.01 for both rotated and absolute ellipticities . after the anisotropic psf correction ,
the small galaxies have a residual ellipticity of @xmath308 , @xmath309 and @xmath310 .
since the small galaxies have a reduction factor of @xmath296 , their residual @xmath311 ellipticity indicates that the psf variation is about 0.02 , which is larger than that given by the comparison of m4/ngc6572 .
we therefore take the variations of the chip - average psf ellipticity to be about 0.02 .
how does this uncertainty affect the ellipticities of the large galaxies ?
as noted above , the post - deconvolution reduction factor for the large galaxies ( with @xmath301 pixels ) is @xmath312 .
consequently , our uncertainty of 0.02 in the psf ellipticity produces an uncertainty in the ellipticities of the large galaxies of only 0.003 .
this is of course a commendable consequence of the small size of the wfpc2 psf .
note also that , for the large galaxies , the camera distortion correction is about the same size as that for the psf , so that the two corrections almost cancel each other .
the changes in ellipticities produced by each of these corrections is less than 0.005 for the large galaxies .
the total systematic uncertainty is a combination of the correction uncertainty and of the psf variability , and is thus about @xmath313 .
the statistical uncertainties are determined by the ellipticity variance @xmath314 . both noise in the image and
the intrinsic shapes of the galaxies contribute to this dispersion .
the _ rms _ ellipticity @xmath258 is listed in the last column of table [ tab : stats ] , at various stages of the correction .
for both the large and small galaxies , @xmath258 does not increase when we correct for the camera distortion and for the anisotropic psf .
this shows that our correction method does not introduce any appreciable noise in the ellipticity measurements .
not surprisingly , @xmath258 increases moderately when we correct for the isotropic psf .
this is expected since this deconvolution reduces the galaxy size , and thus reduces the denominator in the definition of the ellipticity .
for the large galaxies , the _ rms _ ellipticity after all corrections is about @xmath315 .
since the mean number of such galaxies per chip is about @xmath316 , the @xmath317 sensitivity to detect the shear in a chip is @xmath318 .
this is close to the expected _ rms _ shear expected from weak lensing by large - scale structure in 1.3 arcmin cells , for cluster normalized cdm models and for a source redshift @xmath319 ( jain & seljak 1997 ) .
this shows that the expected signal - to - noise ratio of a single wfpc2 chip is about 1 .
we have revisited the ksb method to measure the weak lensing shear from the shapes of galaxies . in our method , the corrections for the camera distortion and psf convolution are performed using moments rather than ellipticities . using a gaussian weight function , we derived explicit expressions for the corrections , which involve only second and fourth order moments .
we clarified the convergence of some of the approximations made by ksb , and showed how the weight function for stars can be chosen to be different from that for galaxies .
we also showed how the isotropic part of the psf can be assumed to be a gaussian , to the required level of precision , and can thus be corrected for exactly .
we derived the explicit relation between the shear and the ellipticities by decomposing moments into tensors with definite rotational properties .
we addressed the recently exposed shortcomings of the ksb method ( kaiser 1999 ; kuijken 1999 ) .
our method , as well as the ksb method , has formal problems arising from the fact that psfs encountered in practice are not sufficiently compact .
we used numerical simulations to assess the importance of these problems in the analysis of wfpc2 images .
we found that the ellipticity error produced by the psf correction can be of several percent for an individual galaxy .
however , it is only about 0.004 when averaged over a galaxy ensemble with @xmath243 pixels , provided moderate weight function widths are used .
we studied systematic effects arising in wfpc2 images . from globular cluster observations ,
we confirm the results of hfks , who found @xmath32010% psf ellipticities at the edge of each chip .
we find however that the camera distortion is radial , rather than tangential .
it produces ellipticities of the order of 0.7% .
we further find that the psf ellipticity varies by as much as 2% over time .
we applied our correction method to the hst survey strip .
we showed that the different stages of our correction do not introduce any appreciable noise .
we studied the dependence of galaxy ellipticities on the galaxy size .
small galaxies are more sensitive to the psf and also indicate that the psf varies with time . for large galaxies ( observed radii @xmath301 pixels ) , the total systematic uncertainty is about 0.4% , and results from a nearly equal contribution from the correction uncertainty and from the psf variability .
the statistical @xmath317 uncertainty in measuring the shear in a single wfpc2 @xmath321 square arcmin chip is about 4% , for this subsample of galaxies .
this provides good prospects for detecting a cosmic shear signal with the strip and other deep hst surveys . in rhodes
et al . ( 1999 ) and rhodes ( 1999 ) , we will describe our search for such a signal .
we thank richard ellis , peter schneider , stella seitz , henk hoekstra , christophe alard , david bacon and meghan gray for useful discussions .
we are grateful to nick kaiser , the referee , for insightful comments and criticisms .
ar was supported by the nasa map / midex program and by the nasa atp grant nag5 - 7154 .
eg and jr were supported by nasa grant nag5 - 6279 , and would like to thank the wfpc1 idt for cooperation on this project .
baggett , s. , 1998 , private communication .
biretta , j. , editor , 1996 , the wfpc2 instrument handbook , ( stsci ) .
bonnet , h. , & mellier , y. 1995 , , 303 , 331 fort , b. , & mellier , y. , 1994 , a&ar , 5 , 239 groth et al .
, 1998 , baas .
hoekstra , h. , franx , m. , kuijken , k. , & squires , g. , 1998 , ( hfks ) , , 504 , 636 holtzman , j.a . , et al .
1995 , , 107 , 156 jarvis , j.f . , & tyson , j.a .
, 1981 , , 86 , 476 kaiser , n. , & squires , g. , & broadhurst , t. 1995 , ( ksb ) , , 449 , 460 kaiser , n. 1996 , preprint astro - ph/9610120 kaiser , n. 1999 , preprint astro - ph/9904003 kneib , j .-
p . , ellis , r.s . ,
smail , i. , couch , w.j . , &
sharples , r.m . , 1996 , , 471 ,
643 kuijken , k. 1999 , submitted to , preprint astro - ph/9904418 luppino , g.a . , & kaiser , n. 1997 , ( lk ) , , 475 , 20 mellier , y. 1999 , ara&a , in press narayan , r. , & bartelmann , m. , 1996 , preprint astro - ph/9606001 rhodes , j. , groth , e. , & wfpc1 idt 1997 , baas , 29 , 1207 rhodes , j. 1999 , phd thesis , department of physics , princeton university rhodes , j. , refregier , a. , & groth , e.j .
1999 , in preparation refregier , a. , 1999 , weak lensing by large - scale structure links and bibliography , http://www.astro.princeton.edu/~refreg refregier , a. brown , s.t .
, kamionkowski , m. , helfand , d.j . ,
cress , c.m .
, babul , a. , becker , r. , & white , r.l . , 1998 , to appear in proc . of the xivth iap meeting `` wide - field surveys in cosmology '' , held in paris in may 1998 , eds .
mellier , y. , & colombi , s. , ( paris : frontiere ) , preprint astro - ph/9810025 richer , h.b .
, et al . 1997 , 484 , 741 schneider , p. , ehlers , j. , & falco , e.e .
, 1992 , gravitational lenses , ( new york : springer - verlag ) schneider , p. , & seitz , c. 1995 , , 294 , 411 schneider , p. , van waerbeke , l. , mellier , y. , jain , y. , seitz , s. , & fort , b. 1997 , preprint astro - ph/9705122 stebbins , a. , mckay , t. , & frieman , j.a . , 1995 , proc . of iau
symposium 173 , preprint astro - ph/9510012 villumsen , j. 1995 , preprint astro - ph/9507007 van waerbeke , l. , bernardeau , f. , mellier , y. , 1998 , submitted to a&a , preprint astro - ph/9807007 lcrrrrr camera distortion & & @xmath322 & @xmath323 & @xmath324 & @xmath323 & + m4+ngc6572 & & @xmath325 & @xmath326 & @xmath327 & @xmath328 & + & d , a & @xmath329 & @xmath324 & @xmath330 & @xmath331 & + m4 & & @xmath332 & @xmath333 & @xmath334 & @xmath335 & + & d , a & @xmath336 & @xmath324 & @xmath337 & @xmath338 & + ngc6572 & & @xmath339 & @xmath340 & @xmath341 & @xmath342 & + & d , a & @xmath343 & @xmath323 & @xmath344 & @xmath345 & + tiny tim ( @xmath346 ) & & @xmath347 & @xmath348 & @xmath349 & @xmath350 & + ( @xmath351 ) & & @xmath352 & @xmath353 & @xmath354 & @xmath355 & + ( @xmath356 ) & & @xmath357 & @xmath348 & @xmath358 & @xmath359 & + small galaxies & & @xmath360 & @xmath361 & @xmath362 & @xmath363 & 0.24 + & d & @xmath364 & @xmath365 & @xmath366 & @xmath367 & 0.24 + & d , a & @xmath366 & @xmath368 & @xmath369 & @xmath370 & 0.24 + large galaxies & & @xmath371 & @xmath348 & @xmath372 & @xmath373 & 0.24 + & d & @xmath374 & @xmath348 & @xmath375 & @xmath376 & 0.24 + & d , a & @xmath377 & @xmath378 & @xmath379 & @xmath380 & 0.24 + & d , a , i & @xmath381 & @xmath382 & @xmath383 & @xmath365 & 0.31 + | the weak distortions produced by gravitational lensing in the images of background galaxies provide a unique method to measure directly the distribution of mass in the universe . however , because the induced distortions are only of a few percent , this technique requires high precision measurements of the lensing shear and cautious corrections for systematic effects .
kaiser , squires , & broadhurst ( 1995 ) proposed a method to calibrate the ellipticity shear relation in the presence of point spread function ( psf ) anisotropies and camera distortions . here , we revisit the ksb method in the context of the demanding search for weak lensing by large - scale structure .
we show that both the psf and the camera distortions can be corrected for using source moments , as opposed to ellipticities .
we clarify the applicability of some of the approximations made in this method .
we derive expressions for the corrections which only involve the galaxy moments . by decomposing the moments into spinors
, we derive an explicit relation between the shear and the average ellipticity .
we discuss the shortcomings of the method , and test its validity using numerical simulations . as an application of the method , we repeat the analysis of the hst wfpc2 camera performed by hoekstra et al .
( 1998 ) .
we confirm the presence of sizable ( @xmath0 ) psf ellipticities at the edge of the wfpc2 chips .
however , we find that the camera distortion is radial , rather than tangential .
we also show that the psf ellipticity varies by as much as 2% over time .
we use these measurements to correct the shape of galaxies in the hst survey strip ( `` groth '' strip ) . by considering the dependence of the ellipticities on object size ,
we show that , after corrections , the residual systematic uncertainty for galaxies with radii greater than 0.15 arcsec , is about 0.4% , when averaged over each chip .
we discuss how these results provide good prospects for measuring weak lensing by large - scale structure with deep hst surveys . |
the asymmetric simple exclusion process ( asep ) is a minimal model for many traffic and queueing processes @xcite .
it describes a non - equilibrium system of many driven particles , interacting via hard core repulsion .
this markov process is defined on a lattice , each site of which can be empty or occupied by one particle .
particles jump independently after an exponentially distributed random time with mean @xmath2 to a nearest neighbor site on the right or on the left , provided that the target site is empty ( hard core exclusion rule ) . the probability to choose the right neighbour is @xmath3 while the probability of choosing the left neighbour is @xmath4 . in spite of its simple formulation , and very simple product stationary state ,
the time - dependent characteristics are very nontrivial to obtain , even in the simplest periodic one - dimensional case ( closed ring of @xmath5 sites ) .
this is due to the fact that e.g. the time - dependent operators of type @xmath6 are complicated objects because the generator of the stochastic dynamics @xmath7 is equivalent to an interacting many - body quantum hamiltonian .
recently , it has been found that the asep , under the restriction to produce an extreme flux or extreme activity is governed by effective long range interactions @xcite . in this
setting one considers realizations of the process for a duration @xmath8 which for a long interval between large times @xmath9 and @xmath10 ( where @xmath10 is itself also large ) have carried an atypically large flux . this extreme event quite surprisingly makes the conditioned process intrinsically related to a much simpler system than the original one , namely to a system of non - interacting fermions .
this fact suggests to use field theoretic free fermion techniques to compute time - dependent correlations in this particular large deviation limit of the classical stochastic dynamics of the conditioned asep . in the present contribution we shall employ these techniques to calculate two particular time - dependent correlation functions : the conditional probabilities @xmath11 for the transition from a microscopic @xmath12-particle initial configuration @xmath13 to a final configuration @xmath14 at time @xmath9 and the dynamical structure factor which is the fourier transform of the time - dependent density - density correlation function @xmath15 in the stationary asep in the two above mentioned limits : under the restriction to produce a very large flux and under the restriction to produce a very large activity .
we note that for the totally asymmetric version of the asep ( tasep ) on an infinite lattice and on a ring the conditional probabilities were found in @xcite and @xcite respectively using the bethe ansatz .
the dynamical structure for the infinite system was obtained by prhofer and spohn using random matrix techniques @xcite .
we remark that that the time - dependent correlations for atypical stochastic trajectories ( in the large deviation limit ) are practically unaccessible by numerical methods , even though some promising approaches have been developed recently @xcite .
the paper is organized as follows . in sec .
[ condition ] we introduce the asep conditioned to carry a certain nontypical flux or to exhibit a nontypical activity . in the first two subsections we review for self - containedness the our previous work @xcite which among other things clarifies how conditioning on large flux and activity is related to free fermions . in the third subsection of sec .
[ condition ] we present a new result , viz .
the form of temporal correlations in the large current or large activity regime . in the following main sections we focus on large flux and derive the conditional transition probabilities ( sec . [ probabilities ] ) and dynamical structure factor ( sec . [ structure ] ) .
we conclude with a brief summary and some remarks on the large limit activity in the asep and symmetric exclusion .
the flux ( or synonymously the current ) in the asep is the mean net number of jumps in some time interval .
conditioning the asep to carry an typical flux is achieved by first ascribing to the asep space - time trajectories weighting factors @xmath16 where @xmath17 is a generalized chemical potential , conjugated to the total current @xmath18 , registered along the trajectory up to time @xmath9 ( see e.g. @xcite for a elaborate treatment ) . fixing @xmath17 , one picks up an ensemble of space - time trajectories characterized by any desired average current @xmath19 .
in particular , fixing the desired current to be very large ( in the sense described in the introduction ) , one obtains a particular set of trajectories , where several tendencies are clearly seen @xcite when taking the limit @xmath20 : ( a ) the backward hoppings are totally suppressed ( b ) formation of clusters of particles , potentially leading to the flux reduction , is strongly suppressed ( c ) long range interactions between the particles are generated . in more details , it was noted that the asep under the restriction to produce extreme flux after proper rescaling of time is described by a master equation @xmath21 with effective stochastic hamiltonian@xmath22 where @xmath23 is a diagonal matrix with positive entries given further below , and @xmath24 is the lowest eigenvalue of the hamiltonian @xmath7 @xmath25 with the spin 1/2 lowering and raising operators of the lie algebra @xmath26 .
we point out that in the limit of very large flux the same stochastic process describes the partially asymmetric simple exclusion process ( asep ) or even the symmetric case , since in this limit any backward hopping does not contribute .
similar considerations can be carried out for the activity where space - time trajectories are weighted by a factor @xmath27 where @xmath28 is the total number of jumps along the trajectory , irrespective of the direction of the jump . in the case of asep
restricted to have an extremely large activity ( counting hoppings , irrespective of the direction ) , the effective hamiltonian has the form ( [ heff ] ) where @xmath7 is substituted by @xmath29 and @xmath24 is substituted by the lowest eigenvalue of @xmath30 and @xmath31 is the hopping rate to the right ( clockwise ) while @xmath32 is the hopping rate to the left ( anti - clockwise ) of the asep . both hamiltonians ( [ unidirxx0]),([asepactivity ] ) are free fermion hamiltonians , and in particular for the case of symmetric hopping @xmath33 , the so called symmetric simple exclusion process , or ssep , the dynamics
is governed by the effective hamiltonian ( [ heff ] ) with @xmath7 substituted by the physical @xmath34 hamiltonian describing quantum fermions on a ring of @xmath5 sites , @xmath35 and @xmath24 substituted by the lowest eigenvalue of @xmath36 .
all the above mentioned hamiltonians commute with each other and share the same ground state @xmath37 .
the components of @xmath37 in the natural basis constitute the respective entries of the diagonal matrix @xmath23 from ( [ heff ] ) @xmath38 . by the perron - frobenius theorem the ground state is nondegenerate , all its components are positive @xmath39 , so the matrix @xmath40 exists .
it is straightforward to verify that the stationary state ( the state with zero eigenvalue ) of the stochastic process ( [ masterequation ] ) is given by @xmath41 .
the stationary state is thus common for all hamiltonians mentioned above , for details see @xcite,@xcite .
it may be worthwhile noting that the effective process is an interesting object of study in its own right , i.e. , without reference to atypically large currents in the usual asep .
this process is a totally asymmetric exclusion process with long - range interaction .
it describes nearest neighbour hopping , but the hopping rate depends on the positions of the other particles .
specifically , the move of the @xmath42-th particle , located at position @xmath43 in a configuration @xmath14 , to the consecutive position @xmath44 in the new configuration @xmath45 has the rate @xcite @xmath46 where the product is over @xmath47 and @xmath42 from 1 through @xmath12 . if the target site @xmath44 is occupied , the hopping rate is zero , in agreement with the exclusion rule . for the case of extreme activity one obtains in the same fashion a partially asymmetric or symmetric exclusion process with long range interaction .
the symmetric case has been studied some time ago by spohn @xcite who derived a hydrodynamic fluctuation theory for the large scale dynamics of the process .
noticing that fluctuations are driven by a gaussian field he presented a general form of the dynamical structure function which represents a universality different from the edwards wilkinson equation .
he also noted the link of the hopping rates ( [ hoppingrates ] ) to dyson s brownian motion , which was pointed out independently in @xcite for the totally asymmetric case .
thus our discrete model has a natural interpretation as a discrete , biased random walk version of dyson s brownian motion driven by an external field .
the complete set of eigenstates of @xmath48 and @xmath36 in a sector with @xmath12 particles is characterized by a combination of @xmath12 plane waves where each plane has a quasimomentum @xmath49 satisfying @xmath50 because of the periodic boundary conditions @xcite .
therefore each @xmath49 takes one of the quantized values @xmath51 and thus each eigenstate is defined by an @xmath12-tuple of integers @xmath52 where each @xmath53 . with these definitions we can write the eigenvectors as @xmath54 where the vector @xmath55 denotes a state with particles at respective positions @xmath56 ( not necessarily ordered ) and the sum is over all possible particle positions @xmath57
the free fermion wave function in coordinate representation has the form @xmath58 where @xmath59 is a permutation of indexes @xmath60 and @xmath61 denotes the sign of the permutation .
the factor @xmath62 is zero if in the set @xmath63 some @xmath56 coincide , and otherwise is equal to @xmath64 or @xmath65 , depending on whether the set @xmath63 was obtained from the ordered set by an even or odd number of pair permutations .
any exchange of particle positions or quasimomenta results in the eigenfunction @xmath66 changing sign , which reflects the pauli principle .
the states @xmath66 with all possible sets @xmath67 are normalized , orthogonal and form an orthonormal basis in the respective sector of hilbert space with @xmath12 number of particles .
the eigenvalues @xmath68 corresponding to the @xmath66 are sums of those for each quasiparticle @xmath69 where the quasiparticle energies are @xmath70 for the hamiltonians ( [ unidirxx0]),([asepactivity ] ) and ( [ xx0 ] ) , respectively .
the ground state @xmath71 which minimizes the energies @xmath72 is characterized by the choice of quasimomenta @xmath73 where @xmath74 , i.e. for the set @xmath75 for this choice the ground state energies @xmath76 for all three cases are real . in the following by default
we treat the hamiltonian ( [ unidirxx0 ] ) with quasiparticle `` energies '' ( [ defaultenergy ] ) and drop the superscript @xmath77 .
consider a general time - dependent correlation function @xmath78 in the stationary distribution of the effective stochastic process defined by ( [ heff ] ) . here
@xmath79 and @xmath80 are functions of the occupation numbers and hence represented in the quantum hamiltonian formalism by diagonal operators and @xmath81 .
the basis of our present work is the following simple , but fundamental property of conditioned processes of the form ( [ heff ] ) .
* theorem*. let the operators @xmath82 be diagonal in the natural basis defined by @xmath83 .
then @xmath84 where @xmath85 and @xmath7 is the respective free fermion hamiltonian ( [ unidirxx0]),([asepactivity ] ) or ( [ xx0 ] ) .
the proof is simple , but for readers not familiar with the machinery we provide here the details . by definition , @xmath86 where
the summation vector @xmath87 and stationary distribution vector @xmath88 are left and right lowest eigenvectors of @xmath83 with eigenvalue 0 .
we recall that @xmath89 @xcite and that the summation vector @xmath90 is a vector with all unit components .
the property @xmath91 is simply a stochasticity condition ( conservation of probability ) for the hamiltonian @xmath83 .
using the latter condition and the fact that @xmath92=0 $ ] when both @xmath23 and @xmath80 diagonal operators we obtain @xmath93 now we note that @xmath94 , which is verified by expanding the exponent on both sides in taylor series and comparing the series term by term , using ( [ heff ] ) . inserting this equality into ( [ th3 ] ) , and using @xmath95=0 $ ] , we obtain @xmath96 finally , noting @xmath97 and @xmath98
, we obtain the right hand side of ( [ th ] ) .
@xmath99 some remarks are in order .
theorem ( [ th ] ) is straighforwardly generalized to multipoint space - time correlation functions and one obtains @xmath100 , provided that all operators @xmath101 are diagonal in the natural basis , i.e. , represent observables of the classical stochastic process generated by @xmath83 .
we also point out that the theorem is phrased for our specific case of the current or activity in the asep
. however , a similar results applies for any conditioned stochastic dynamics in which case @xmath102 is the weighted generator of the unconditioned process .
we pick two arbitrary configurations @xmath13 and @xmath14 of our system containing @xmath12 particles at the positions @xmath103 and @xmath104 respectively , i.e. , @xmath105 , @xmath106 .
we assume both sets to be ordered , @xmath107 and @xmath108 .
the probability to find the system in configuration @xmath14 at time @xmath9 , provided it has been in configuration @xmath13 at time @xmath109 is given by @xmath110 where the subscript denotes that the average is computed with respect to the stationary state defined by the effective dynamics ( [ masterequation ] ) .
the operators @xmath111 and @xmath112 are diagonal operators , which allows to use the theorem ( [ th ] ) .
the denominator in ( [ condprob ] ) can be obtained by using ( [ th ] ) with @xmath113 and @xmath114 . using the results for the stationary probabilities of ref .
@xcite we find @xmath115 the right hand side of ( [ eta0 ] ) are ( modulo squares ) the components of a slater determinant with quasimomenta ( [ gsmomenta ] ) chosen so as to fill the fermi sea . for details , see e.g. @xcite . for the numerator we have , using ( [ th]),@xmath116 since the eigenvectors ( [ mu_q ] ) form an orthonormal basis the sum of projectors @xmath117 over all @xmath12-tuples @xmath67 is a unit operator .
the factor @xmath118 appears because in the sum @xmath119 each different set of @xmath49-s occurs @xmath120 times .
inserting it in ( [ nom ] ) , we obtain@xmath121 where @xmath122 is the ground state energy corresponding to the specific set of quasimomenta @xmath123 given by ( [ gsmomenta ] ) , and @xmath124 are found using ( [ mu_q ] ) .
substituting @xmath125 into ( [ nom1 ] ) , one obtains @xmath126 using the explicit form of @xmath127 ( [ ksi ] ) , the sum over @xmath67 in ( [ nom2 ] ) can be rewritten as @xmath128 in the sum over the permutations @xmath129 there are @xmath130 terms ; however , under the summation over @xmath131 and reshuffling of @xmath42-s only @xmath120 terms are independent , namely , @xmath132 .
substituting this into ( [ nom3 ] ) , we get @xmath133 so far the discussion has been general and applicable to all three hamiltonians ( [ unidirxx0]),([asepactivity ] ) and ( [ xx0 ] ) .
focussing now on the default case ( [ unidirxx0 ] ) we use ( [ e ] ) and ( [ defaultenergy ] ) for further simplification .
we expand the exponents @xmath134 in ( [ nom4 ] ) , and collect the terms with the same @xmath73 . under the summation over @xmath67 ,
only the terms with @xmath135 contribute to the sum , where @xmath136 if @xmath137 and @xmath138 otherwise .
since each @xmath139 satisfies @xmath140 , for odd number of particles @xmath12 each contributing term in ( [ nom4 ] ) is equal to @xmath141 for even @xmath12 , the signs of the terms with different @xmath142 will alternate .
now we introduce the function @xmath143 ^{n+1}\frac{t^{d_{l}+\kappa l}}{(d_{l}+\kappa l)!},\label{g(d)}\ ] ] where @xmath144 is an integer ranging from @xmath145 to @xmath146 and @xmath147 for @xmath148 and @xmath149 for @xmath150 .
the function @xmath151 for @xmath152 and @xmath153 for @xmath150 . with this function
the expression ( [ nom4 ] ) can be rewritten in compact determinantal form as @xmath154\label{nom5}\ ] ] where we used leibnitz formula for the determinant . finally , using ( [ eta0 ] ) , ( [ nom2 ] ) and
( [ nom5 ] ) , the conditional probabilities ( [ condprob ] ) can be brought after some algebra into the final form @xmath155\label{condprobres}\ ] ] the @xmath156 are stationary probabilities of the initial and the final state respectively , given by ( [ eta0 ] ) . at the last step of the calculation we have used the fact that all the components @xmath157 of the ground state eigenvector can be made positive ,
see the discussion after eq .
( [ xx0 ] ) . consequently @xmath158 .
( [ condprobres ] ) is the main result of this section .
the determinantal structure of the conditional probabilities was first noticed by spohn @xcite for the symmetric case .
in contrast to the symmetric case , here the matrix elements of the determinant are the propagators of a totally asymmetric random walk . extending the link of the symmetric model to dyson s brownian motion we may regard our model as a totally asymmetric dyson random walk . let us discuss some limiting cases of ( [ condprobres ] )
. for @xmath109 one has @xmath159=\delta_{\eta\eta_{0}}$ ] , yielding the correct normalization @xmath160 . in the simplest case of one particle @xmath161 , @xmath162 @xmath163 , and we obtain @xmath164 where @xmath165 for @xmath166 and @xmath167 otherwise , and @xmath168 are initial and final particle positions . each term @xmath169 in the sum is a contribution of a poisson process @xmath170 where an event ( hopping of a particle to the right with rate @xmath64 ) has happened @xmath171 times during time @xmath9 . along this trajectory
a particle starts from site @xmath172 , and arrives at site @xmath173 after making @xmath142 complete circles on the ring of size @xmath5 .
thus the parameter @xmath142 in ( [ g(d ) ] ) is the winding number . the formula ( [ condprobres ] ) has been obtained for the continuous time markov process ( [ masterequation ] ) . for the discrete time update @xcite we expect the poisson distribution terms in ( [ g(d ) ] ) to be substituted with the binomial distribution
terms@xmath174 ^{n+1}\binom{t}{d_{l}+\kappa l}p^{d_{l}+\kappa l } ( 1-p)^{t - d_{l}-\kappa l } , \label{g(d ) discrete}\ ] ] where @xmath31 is the probability of a tasep particle to hop . unlike the function ( [ g(d ) ] ) , the sum ( [ g(d ) discrete ] ) is truncated for any finite @xmath9 .
the dynamic structure factor in the large current regime of a periodic chain with @xmath5 sites is defined as the fourier transform of the stationary correlation function @xmath175 , where @xmath176 are particle number operators and @xmath177 is the stationary particle density . without losing generality , we assume @xmath178 .
because of the translational symmetry and time independence of the hamiltonian @xmath83 , the correlation function @xmath179 depends only on the differences @xmath180 @xmath181 , i.e. @xmath182 . thus we can write the real - space representation of the dynamic structure factor as @xmath183 . moreover , by particle - hole symmetry we have that @xmath184 and trivially @xmath185 .
hence we can limit our discussion to the range @xmath186 . in order to simplify notation
we drop the dependence on @xmath187 in the structure function and the correlation function and simply write @xmath188 and @xmath189 .
the number operator is diagonal and therefore the formula ( [ th ] ) is applicable .
@xmath189 can then be calculated in similar manner as the conditional probabilities in the previous section .
we shall present only the final result ( the derivation proceeds analogously to the respective quantum mechanical calculation of @xmath190 in @xcite),@xmath191 where the @xmath73 are of the form ( [ quasimomenta ] ) with the ground state choice @xmath74 , and @xmath192 is given by one of expressions ( [ defaultenergy]-[e_xx0 ] ) .
notice that even though the expression ( [ h(n , t ) ] ) formally looks like the corresponding quantum formula in imaginary time , its analytical properties and limits are crucially different .
all the sums in ( [ h(n , t ) ] ) are strictly real which can be verified straighforwardly , using symmetricity of the ground state set ( [ gsmomenta ] ) of pseudomomenta @xmath73 .
we remark that for our default choice @xmath193 , the summation from 1 through @xmath5 in eq.([h(n , t ) ] ) attains a simple form by expanding the exponent .
we obtain then @xmath194 where @xmath195 is given by ( [ g(d ) ] ) . from the two - point correlation function ( [ h(n , t ) ] ) we compute the dynamic structure factor @xmath196 with the integer momentum variable @xmath197 .
obviously @xmath198 for any integer @xmath172 and @xmath199 which allows us to restrict the subsequent study of the dynamic structure factor to the range @xmath200 . to evaluate ( [ dsfdef ] ) in this range
we first observe that the fourier transformation turns the exponentials of the summation variables @xmath201 into the kronecker - delta @xmath202 .
then we write the second sum in ( [ h(n , t ) ] ) ( which runs up to @xmath12 ) as a sum from 1 to @xmath5 and subtract the part from @xmath203 to @xmath5 .
this yields as an intermediate expression @xmath204 \nonumber\\ & + \frac{1}{l } \sum_{k=1}^{n } \sum_{l = n+1}^{l } \mathrm{e}^{-(\varepsilon ( \alpha_{k})-\varepsilon(\alpha_{l}))t } \delta_{p , k - l}\\ & : = \hat{s}_{l}^{(1)}(p , t ) + \hat{s}_{l}^{(2)}(p , t)\end{aligned}\ ] ] for which we analyse next the double sum @xmath205 .
we focus on the default case ( [ defaultenergy ] ) . for the default case the difference of relaxation times in
the exponential takes the simple form @xmath206 and for notational convenience we introduce @xmath207 bearing in mind the range of definition of the momentum variable @xmath200 , the kronecker - delta in conjunction with the summation limits of the double sum gives rise to three distinct regimes for @xmath31 .
careful analysis yields @xmath208{ll}\displaystyle \frac{1}{l } \sum_{k=1}^{p } \mathrm{e}^{t_{p}\mathrm{e}^{-i\alpha_{k } } } & p = 1,\dots , n-1\\ \displaystyle \frac{1}{l } \sum_{k=1}^{n } \mathrm{e}^{t_{p}\mathrm{e}^{-i\alpha_{k } } } & p = n,\dots , l - n\\ \displaystyle \frac{1}{l } \sum_{k = n+1-l+p}^{n } \mathrm{e}^{t_{p}\mathrm{e}^{-i\alpha_{k } } } & p = l - n+1,\dots , l-1 . \end{array } \right.\ ] ] in the thermodynamic limit @xmath209 the sums over @xmath42 and @xmath210 in ( [ h(n , t ) ] ) turn into integrals , @xmath211 .
notice that here @xmath31 is real - valued and we define it to be in the interval @xmath212 $ ] .
this yields @xmath213 where @xmath214 corresponds to the limit of the function @xmath215 , where only the first term in ( [ g(d ) ] ) appears ( no winding condition ) . with a view in large , but still finite systems we point out that this observation imposes obvious validity limitations on the integral expression ( [ h(n , t)thermodynamiclimit ] ) as an approximation for large system size : the limiting behaviour can not be used as an approximation for @xmath216 and @xmath217 .
the integrals @xmath218 , apart from obvious special cases @xmath214 , and @xmath219 are not expressed in elementary functions and must be evaluated numerically . in fig . [ fig_evenodd ] we show the function @xmath220 for even and odd @xmath172 for the particular case of half - filling @xmath221 at different times @xmath9 . for @xmath109
the difference in the expression @xmath222 for energies the quasiparticles become irrelevant and , using @xmath223 , we obtain from ( [ h(n , t)thermodynamiclimit ] ) the static density - density correlation function @xmath224 first derived in @xcite . in order to explore the large - scale behaviour of the dynamic structure factor we study the behaviour for small momentum @xmath31 and large times @xmath9 . to this end
we return to the definition ( [ dsfdef ] ) and first observe that in the thermodynamic limit @xmath225\ ] ] where now @xmath226 .
using @xmath227 allows us to rewrite this expression as a difference of integrals over the same function @xmath228 with integration intervals @xmath229 $ ] ( positive term ) and @xmath230 $ ] ( negative term ) respectively . on the other hand @xmath231{ll}\displaystyle \frac{1}{2\pi } \int_{-\rho\pi}^{-\rho\pi+p } dx\ , \mathrm{e}^{t_{p}\mathrm{e}^{-ix } } & p \in[0,2\rho\pi]\\ \displaystyle \frac{1}{2\pi } \int_{-\rho\pi}^{\rho\pi } dx\ , \mathrm{e}^{t_{p}\mathrm{e}^{-ix } } & p \in[-\pi,\dots,-2\rho\pi ] \cup[2\rho\pi,\dots , \pi]\\ \displaystyle \frac{1}{2\pi } \int_{\rho\pi+p}^{\rho\pi } dx\ , \mathrm{e}^{t_{p}\mathrm{e}^{-ix } } & p \in[-2\rho\pi,0 ] \end{array } \right.\ ] ] putting everything together we finally obtain the dynamic structure factor for @xmath232 @xmath233{ll}\displaystyle \frac{1}{2\pi } \int_{\rho\pi}^{\rho\pi+p } dx\ , \mathrm{e}^{t_{p}\mathrm{e}^{-ix } } & p \in[0,2\rho\pi]\\ \displaystyle \frac{1}{2\pi } \int_{-\rho\pi+p}^{\rho\pi+p } dx\ , \mathrm{e}^{t_{p}\mathrm{e}^{-ix } } & p \in[-\pi,\dots,-2\rho\pi ] \cup[2\rho\pi,\dots , \pi]\\ \displaystyle \frac{1}{2\pi } \int_{-\rho\pi+p}^{-\rho\pi } dx\ , \mathrm{e}^{t_{p}\mathrm{e}^{-ix } } & p \in[-2\rho\pi,0 ] \end{array } \right.\ ] ] this , along with the symmetry relation @xmath234 provides an exact integral presentation valid for all densities @xmath187 , momenta @xmath31 and times @xmath9 .
the static structure factor takes the simple form ( @xmath232 ) @xmath235{ll}\displaystyle \frac{|p|}{2\pi } & p \in[-2\rho\pi,2\rho\pi]\\ \displaystyle \rho & p \in[-\pi,\dots,-2\rho\pi ] \cup[2\rho\pi,\dots,\pi ] , \end{array } \right.\ ] ] cf . the real - space result ( [ staticcorrelation ] ) .
we are particularly interested in the large scale behaviour as expressed in the scaling limit of small @xmath31 and large @xmath9 of the form @xmath236 where @xmath237 is the dynamical exponent and @xmath238 is the scaling variable . in the limit @xmath239 only the first and third expression in ( [ dsfexact ] ) are relevant . from the occurrence of the factor @xmath240
we conclude that there is non - trivial scaling behaviour for @xmath0 , i.e. for @xmath241 . in this scaling
we have @xmath242 which yields the desired result @xmath243 which is valid for all @xmath244 $ ] and in agreement with the universal form the dynamic structure factor derived in @xcite for the symmetric case .
the presence of the particle drift does not change the universality class as it does for the usual unconditioned exclusion process where the undriven model is in the universality class of the edwards - wilkinson equation with dynamical exponent @xmath245 while the driven model is in the kpz universality class with @xmath1 .
this is in agreement with an earlier observation that for stochastic dynamics which have an underlying free - fermion structure an external drift can be absorbed into a galilei transformation @xcite .
we obtained analytically conditional probabilities and the two point time - dependent density correlation functions for the asep conditioned to carry a very large average current .
the conditional probabilities have determinantal form and can be expressed through elementary functions .
the density correlation functions are obtained both for a finite system and in the thermodynamic limit . by fourier transformation
we have computed the exact dynamical structure factor and derived its large - scale behaviour .
the natural scaling variable turns out to be @xmath241 which proves that the dynamical exponent of the conditioned asep is @xmath0 . from the explicit scaling form we read off the collective velocity @xmath246 of density fluctuations which is in contrast to @xmath247 of the usual asep in the regime of typical currents .
the relaxation part is symmetric in @xmath238 and very different from the corresponding quantity in the usual asep @xcite which has dynamical exponent @xmath1 for the driven case and @xmath245 for the symmetric case . in our model
the presence of a drift does not change the universality class .
our results have a natural generalization to the study of large activity , i.e. to choosing the hamiltonians ( [ asepactivity ] ) and ( [ xx0 ] ) .
one has to replace the energies @xmath192 in ( [ h(n , t ) ] ) by the respective expressions @xmath248 and @xmath249
. it will be interesting to study not only the hydrodynamic limit , but also the microscopic structure of shocks for the general case .
it would also be interesting to extend the analysis of the effective dynamics of driven systems under large deviation constraints to the non - diagonal case , e.g. to compute current - current time - dependent averages .
we thank m. salerno and h. spohn for stimulating discussions and d. simon for valuable comments on a preliminary version of the manuscript . this work was supported by deutsche forschungsgemeinschaft . | we consider the asymmetric simple exclusion processes ( asep ) on a ring constrained to produce an atypically large flux , or an extreme activity . using quantum free fermion techniques we find the time - dependent conditional transition probabilities and the exact dynamical structure factor under such conditioned dynamics . in the thermodynamic limit
we obtain the explicit scaling form .
this gives a direct proof that the dynamical exponent in the extreme current regime is @xmath0 rather than the kpz exponent @xmath1 which characterizes the asep in the regime of typical currents .
some of our results extend to the activity in the partially asymmetric simple exclusion process , including the symmetric case . |
for both mathematical and physical reasons one would like to be able to understand anisotropic and inhomogeneous cosmological models in general . as a first step spatially homogeneous spacetimes can be studied . for the case where the matter model is a perfect fluid a lot of results
have been already obtained which are summarized in @xcite , see also @xcite for a critical discussion .
usually in observational cosmology it is assumed that there exists an era which is `` matter - dominated '' where the matter model is a perfect fluid with zero pressure , i.e. the dust model . a kinetic description via collisionless matter enables to study the stability of this model in the following sense .
suppose we have an expanding universe where the particles have certain velocity dispersion .
one might think that due to the expansion the velocity dispersion will decay . that this is true has been shown in @xcite , @xcite and @xcite for locally rotationally symmetric ( lrs ) models in the cases of bianchi i , ii and iii .
these results have been generalized ( in a different direction ) to lrs bianchi ix and in a context which also goes beyond collisionless matter in @xcite and @xcite .
+ the lrs models can be diagonalized in a suitable frame where they then stay diagonal if one makes some extra assumptions on the distribution function .
a natural question is , what happens if the model is not diagonal ? in this paper we will treat the late time dynamics of the _ non - diagonal _ bianchi i case assuming small data .
that this makes sense was established in @xcite where it was shown that geodesic completeness holds for the general bianchi i - vlasov case . in a sense which will be specified later , we assume that the universe is close to isotropic and that the velocity dispersion of the particles is bounded .
then we conclude that the universe will isotropize and have a dust - like behaviour asymptotically .
in fact these two properties are intimately linked in the proof , something which is not expected to happen in other bianchi types , except in the case of a positive cosmological constant where it has been shown @xcite that isotropization and asymptotic dust - like behaviour occurs in all bianchi models except bianchi ix and without the lrs assumption .
+ we hope to be able to extend this result in the near future a ) to other bianchi types , where the spacetime does not ( necessarily ) isotropize , but some other solution still may act as an attractor b ) to remove the small data assumption(s ) .
a _ bianchi spacetime _ is defined to be a spatially homogeneous spacetime whose isometry group possesses a three - dimensional subgroup @xmath4 that acts simply transitively on the spacelike orbits .
they can be classified by the structure constants of the lie algebra associated to the lie group .
we will only consider the simplest case where the structure constants vanish , i.e. the case of bianchi i with the abelian group of translations in @xmath5 as the lie group , where the metric has the following form using gauss coordinates : @xmath6 we will use the 3 + 1 decomposition of the einstein equations as made in @xcite .
we use gauss coordinates , which implies that the lapse function is the identity and the shift vector vanishes , so comparing our metric with ( 2.28 ) of @xcite we have that @xmath7 and @xmath8 .
the only non - trivial christoffel symbols are the following ( see ( 2.44)-(2.49 ) of @xcite ) : @xmath9 in terms of coordinate expressions we have then that ( @xmath10 ) @xmath11 where @xmath12 , @xmath13 and @xmath14 are the energy density , matter current and energy momentum tensor respectively . in the 3 + 1 formulation the second fundamental form @xmath15 is used and rewriting its definition as in ( 2.29 ) of @xcite we have : @xmath16 using the einstein equations as in ( 2.34 ) of @xcite and the fact that @xmath17 does not depend on the spatial variables , we have : @xmath18 where we have used the notations @xmath19 , @xmath20 .
it has been assumed that the cosmological constant vanishes . from the constraint equations ( 2.26)-(2.27 ) of @xcite : @xmath21 and we see that no matter current is present since the `` spatial '' christoffel symbols vanish : @xmath22 from ( [ c ] ) we see that @xmath23 never vanishes except for the minkowski spacetime .
we exclude this case and assume now without loss of generality that @xmath24 for all time .
if this is not true for a given solution it may be arranged by doing the transformation @xmath25 .
+ for the matter model we will take the point of view of kinetic theory .
this means that we have a collection of particles ( in a cosmological context the particles are galaxies or clusters of galaxies ) which are described statistically by a non - negative distribution function @xmath26 which is the density of particles at a given spacetime point with given four - momentum .
we will assume that all the particles have equal mass ( one can relax this condition if necessary , see @xcite ) .
we want that our matter model is compatible with our symmetry assumption , so we will also assume that @xmath27 does not depend on @xmath28 .
in addition to that we will assume that there are no collisions between the particles .
in this case the distribution function satisfies the vlasov equation ( see ( 3.38 ) of @xcite ) : @xmath29 where f is defined on the set determined by the equation @xmath30 called the mass shell .
the energy momentum tensor is ( compare with ( 3.37 ) of @xcite ) : @xmath31 here @xmath32 and @xmath33 . for this kind of matter all the energy conditions hold . in particular @xmath34 .
our system of equations consists of the equations ( [ a])-([cu ] ) . for a given bianchi i geometry the vlasov equation
can be solved explicitly with the result that if @xmath27 is expressed in terms of the covariant components @xmath35 then it is independent of time .
this has the consequence that if @xmath36 is some fixed time and @xmath37 then we can express the non - trivial components of the energy momentum tensor as follows : @xmath38 a useful relation concerns the determinant of the metric ( ( 2.30 ) of @xcite ) : @xmath39=-2 h \end{aligned}\ ] ] taking the trace of the mixed version of the second fundamental form ( 2.36 of @xcite ) : @xmath40 with ( [ c ] ) one can eliminate the energy density and ( [ i m ] ) reads : @xmath41 we can decompose the second fundamental form introducing @xmath42 as the trace - free part : @xmath43 then @xmath44 and ( [ in ] ) takes the following form : @xmath45 it is obvious that @xmath46 from the constraint equation ( [ c ] ) and ( [ yes ] ) it also follows that @xmath47 from ( [ ie ] ) and ( [ up ] ) we conclude that @xmath48 let us use the following notation : @xmath49 this quantity is related to the so called _ shear parameter
_ , which is bounded by the cosmic microwave background radiation and is a dimensionless measure of the anisotropy of the universe ( see chapter 5.2.2 of @xcite ) . using this notation and with the help of ( [ a ] ) , ( [ b ] ) and ( [ yes ] ) we have @xmath50\end{aligned}\ ] ] from the constraint equation ( [ c ] ) @xmath51 in general
we see that @xmath52 .
+ we see that in the vacuum case @xmath53 and @xmath54 . actually in this case one can write explicitly the solution known as _ kasner solution _ : @xmath55 where the constants @xmath35 are called the _ kasner exponents _ which satisfy the two _ kasner relations _
given by : @xmath56 the mean curvature in this case is @xmath57 let @xmath58 be the eigenvalues of @xmath59 with respect to @xmath60 , i.e. , the solutions of @xmath61 we define @xmath62 as the _ generalized kasner exponents_. they satisfy the first but not in general not the second kasner relation . a special case of bianchi
i is the friedmann model with flat spatial geometry .
the metric in this case is : @xmath63 where @xmath64 is the scale factor . in this case @xmath65 which means that @xmath66 and also that @xmath54 .
+ moreover in the dust case @xmath67 ( the einstein - dust system can be thought of as singular case of the einstein - vlasov system , see chapter 3.4 of @xcite for more information ) and we can solve ( [ yes ] ) obtaining : @xmath68 we can also write down the explicit solution of the metric in this case : @xmath69 this is also called the einstein - de sitter model .
let us look at the dust case not necessarily isotropic .
the general solution is known and one can see from the solution ( 11 - 1.12 ) of @xcite that the spacetime will isotropize .
nevertheless we will analyze this case with care , since we will show that the general case behaves asymptotically like it assuming small data . in the dust case : @xmath70 we are interested in the asymptotic behaviour at late times and will assume that @xmath71 is small , i.e. : @xmath72 .
then it follows from ( [ yes ] ) : @xmath73 integration leads to @xmath74 using this inequality in ( [ dd ] ) and integrating : @xmath75 we can put the equality ( [ hd ] ) in the following form : @xmath76 where @xmath77 is @xmath78 now we will use the fact that we can choose freely the time origin setting @xmath79 . note that for the general bianchi i symmetric einstein - vlasov - system we know that @xmath23 takes all values in the range @xmath80 ( lemma 2.1 of @xcite )
we then obtain @xmath81 using ( [ yoy ] ) it is clear that @xmath82 .
note that with our time origin choice @xmath83 in ( [ if ] ) , so @xmath84 our result for @xmath23 is the following : @xmath85\end{aligned}\ ] ] this equation in ( [ dd ] ) leads after integration to : @xmath86 from ( [ hi ] ) and ( [ 2 ] ) we can conclude that : @xmath87 now from ( [ tf ] ) we see that the eigenvalues ( [ ev ] ) of the second fundamental form with respect to the induced metric are also the solutions of @xmath88\delta^i_j)\end{aligned}\ ] ] let us define the eigenvalues of @xmath89 with respect to @xmath60 by @xmath90 , we have that : @xmath91 note that @xmath92 . from ( [ hi ] ) and ( [ 3 ] )
we can see that the spacetime isotropizes at late times , in the sense that @xmath93 where @xmath35 are the generalized kasner exponents .
+ now using ( [ det ] ) and ( [ hi ] ) we have : @xmath94 so the hope would be to show that @xmath95 can be bounded by a constant .
let us define @xmath96 we have then with ( [ a ] ) that : @xmath97 making similar computations as in @xcite we arrive at : @xmath98 \vert \bar{g}_{ab}(s ) \vert ds\end{aligned}\ ] ] and with gronwall s inequality we obtain : @xmath99ds\}\leq c\end{aligned}\ ] ] therefore @xmath100 is bounded for all @xmath101 .
the same holds for @xmath102 by similar computations .
thus : @xmath103 from this we can conclude that : @xmath104 or @xmath105 looking again at the derivative of @xmath100 and putting the facts which have been obtained together , we see that : @xmath106 this is enough to conclude that : @xmath107\\ g^{ab}=t^{-\frac{4}{3}}[\mathcal{g}^{ab}+o(t^{-2})]\end{aligned}\ ] ] where @xmath108 is the limit of @xmath100 as @xmath3 goes to infinity .
for the general case the basic equations are ( [ yes ] ) , a modified version of ( [ b ] ) where @xmath109 is a small and positive quantity which is introduced for technical reasons and ( [ no ] ) @xmath110 + 2t^{-\gamma+\frac{4}{3}}\label{bb } \sigma^{ab } \\ & & \dot{f}=h[f(1-\frac{3}{2}f-\frac{8\pi } { h^2 } \operatorname{tr}s)-\frac{16\pi}{h^3 } s_{ab}\sigma^{ab}]\label{cc}\end{aligned}\ ] ] we have a number ( different from zero ) of particles at possibly different momenta and we will define @xmath111 as the supremum of the absolute value of these momenta at a given time @xmath3 : @xmath112 consider any @xmath113 solution of the einstein - vlasov system with bianchi i - symmetry and with @xmath113 initial data .
assume that @xmath114 and @xmath115 are sufficiently small .
then at late times one can make the following estimates : @xmath116 * remark * ( [ pp ] ) implies that asymptotically there is a dust - like behaviour ( see ( [ pp ] ) ) . * proof * we will use a bootstrap argument ( see chapter 10.3 of @xcite for more information ) .
let us look at the interval @xmath117 .
our bootstrap assumptions are the following : @xmath118 where a and b are positive constants which we can choose as small as we want .
+ + _ 1 .
estimate of @xmath23 _
+ + integrating ( [ aa ] ) we obtain : @xmath119 with @xmath120ds\end{aligned}\ ] ] where the inequality comes from ( [ a1 ] ) and ( [ c ] ) .
consider now an orthonormal frame and denote the components of the spatial part of the energy - momentum tensor in this frame by @xmath121 .
the components can be bounded by @xmath122 so we have that @xmath123 from which follows that : @xmath124 now returning back to ( [ h ] ) and setting as in the dust case @xmath125 we have besides @xmath126 that @xmath127 + _ 2 . estimate of @xmath111 _ + + now we will use the second equation ( [ bb ] ) to obtain an estimate for the metric .
one can show that in the sense of quadratic forms the following is true : @xmath128 then with the definitions of @xmath102 and @xmath71 and the estimates ( [ a1 ] ) and ( [ ho ] ) we have : : @xmath129 now from the last inequality of the first step ( [ hh ] ) equation ( [ bb ] ) leads to : @xmath130t^{-1}=-\eta t^{-\gamma}\bar{g}^{ab}t^{-1}\end{aligned}\ ] ] @xmath131 and @xmath132 have to be chosen in such a way that @xmath133 is positive , then @xmath134 from which follows that @xmath135 or @xmath136 now from ( [ neg ] ) we have : @xmath137 using the fact that @xmath138 is constant along the geodesics we can conclude that : @xmath139 since we can choose @xmath115 and @xmath132 independently as small as we want . in order to improve ( [ a2 ] )
@xmath109 has to be smaller then @xmath140 . using the notation
@xmath141 the last inequality can be expressed as @xmath142 where @xmath143 .
+ + _ 3 .
estimate of @xmath71 _
+ + until now we have an estimate for @xmath23 and for @xmath111 in the interval @xmath117 .
now we have to improve the estimate for @xmath71 coming from the bootstrap assumption .
the desired estimate is @xmath144 .
if this is the case ( case i ) the bootstrap argument will work and there is nothing more to do .
_ case i _
let us suppose now the opposite , that @xmath145 . then define @xmath146 as the smallest number not smaller than @xmath36 with the property that @xmath147 . in this case
, we have to distinguish between the case that @xmath148 ( case iia ) and @xmath149 ( case iib ) .
let us look now at ( [ cc ] ) , in particular at the terms in square brackets . by using ( [ a1 ] ) @xmath150 where @xmath151 is a positive small constant . using ( [ c ] ) , ( [ a2 ] ) and ( [ ss ] ) : @xmath152 with the cauchy - schwarz inequality , ( [ c ] ) , ( [ pp ] ) and supposing that @xmath153 in the interval @xmath154 $ ] : @xmath155 note that although @xmath156 may be a big quantity , since @xmath131 and @xmath132 are independent we can make @xmath132 smaller to `` correct '' this . using ( [ hh ] ) and
the last three inequalities in ( [ cc ] ) leads to : @xmath157 where @xmath158 . now setting @xmath159 we end up with : @xmath160 which means that @xmath161 _ case iia .
_ in this case @xmath162 , so ( [ b2 ] ) means @xmath163 since we can choose @xmath114 as small as we want .
so it follows that in this case : @xmath164 situation which is schematically depicted in the following figure .
_ _ case iib .
_ in the case iib we can use the fact that by continuity @xmath165 holds and then @xmath166 the @xmath2 here is also a quantity which we can choose as small as we want and then it follows that in this case @xmath167 we can choose @xmath2 to be smaller then @xmath168 , so @xmath169 _ case iib
results of the bootstrap argument _
+ we have arrived at the statement that at least for a small interval @xmath117 and assuming ( [ a1])-([a2 ] ) we obtain the estimates : @xmath170 thus both assumptions ( [ a1 ] ) and ( [ a2 ] ) have been improved and using a bootstrap argument we know that the estimates obtained are valid for @xmath171 assuming small data .
( [ hh ] ) can be expressed as : @xmath172 and is then also valid for the whole interval .
+ + _ 5 . improving the estimate of @xmath71 _ + we want to improve ( [ yy ] ) , but before that we need an inequality in the other direction . from ( [ cc ] )
we have : @xmath173 implementing now ( [ yy])-([xx ] ) in this inequality , in particular having in mind the b-dependence of the estimate of @xmath111 ( [ b ] ) : @xmath174 using now ( [ ho ] ) : @xmath175 from which follows @xmath176 and @xmath177 @xmath114 is strictly positive .
choosing now @xmath132 small enough we have the following estimate : @xmath178 putting the last term of ( [ cc ] ) with the help of ( [ xx ] ) and ( [ u ] ) in the following manner : @xmath179 we can improve ( [ yy ] ) implementing ( [ yy])-([zz ] ) in ( [ cc ] ) with the result : @xmath180 + with these results we obtain the following theorem . [ t2 ] consider the same assumptions as in the previous theorem .
then @xmath181 and @xmath182\\\label{ll } g^{ab}=t^{-\frac{4}{3}}[\mathcal{g}^{ab}+o(t^{-2 } ) ] \end{aligned}\ ] ] where @xmath108 and @xmath183 are independent of @xmath3 . *
proof * the conclusions made in the dust case after ( [ 2 ] ) only depend on the estimate of @xmath71 and @xmath23 and apply directly to the general case .
the result concerning the asymptotics of the metric implies `` asymptotic freezing '' in the expanding direction in the following sense .
consider the metric ( 4.12 ) of @xcite , then by comparison with ( [ lp ] ) one sees that the off - diagonal degrees of freedom @xmath184,@xmath185 and @xmath186 tend to constants and thus are not important for the dynamics .
see @xcite for the importance of asymptotic freezing in the `` other '' direction which has been studied , namely the initial singularity , and @xcite for consequences of that in a quantum version .
+ as already mentioned in the introduction there exist several results concerning the bianchi i - symmetric einstein - vlasov system ( without cosmological constant ) .
almost all of them assume additional symmetries namely the lrs and the reflection symmetry condition ( see @xcite for a precise definition of these symmetries ) .
however there exist some results where only the reflection symmetry is assumed .
concerning the expanding direction there is theorem 5.4 of @xcite .
our theorems can be seen as a generalization of that theorem since we obtain the same the result , but a ) we also obtain how fast the expressions converge b ) we obtain an asymptotic expression for the spatial metric c ) we do not assume any of the additional symmetries mentioned .
however we used a different kind of restriction namely the small data assumptions .
+ in any case we think it interesting to study the non - diagonal case because although this time there was not an essential difference in the result with respect to the diagonal case , we do not expect that this will always be the case , especially when analyzing the initial singularity . in @xcite
the possible dynamical behaviour towards the past has been determined assuming only the reflection symmetry and already there surprising new features like the existence of heteroclinic networks arose .
+ the situation of several ( tilted ) fluids leads naturally to the non - diagonal case .
the bianchi i - symmetry implies the absence of a matter current such that a single _ tilted _ fluid is not compatible with this assumption .
however there can be several tilted fluids such that the total current vanishes . in @xcite and @xcite
this has been considered in the case of two fluids .
what was found is that isotropization occurs if at least one of the two fluids has a speed of sound which is less or equal @xmath0 the speed of light .
it is shown in particular for the case of two pressure free fluids in @xcite , which can be seen as a singular solution of the einstein - vlasov system .
+ of course there are many other ways of generalizing results which have been obtained for ( single ) perfect fluids .
see for instance @xcite and references therein for the inclusion of a maxwell field in the bianchi i case . in presence of a cosmological constant
the results of @xcite have been generalized even to the einstein - vlasov - maxwell case @xcite .
a natural generalization of the vlasov equation is the case where the collision term is not zero , i.e. the boltzmann equation . for this case
dust - like asymptotics have already been obtained in @xcite for an isotropic spacetime with a cosmological constant and @xcite provides a basis for a possible extension to the asymptotics in the case of bianchi i with lrs symmetry .
+ finally we would like to mention that non - diagonal bianchi i spacetimes are not only of interest in the context of cosmology , see for instance a recent work on the so called ultra - local limit @xcite .
* acknowledgements * + the author would like to thank alan d. rendall for helping along all the steps of this project , from proposing the problem to actually helping to solve it .
i am also thankful for many discussions and a lot of concrete advice
. this work has been funded by the deutsche forschungsgemeinschaft via the sfb 647-project b7 .
s. calogero and j.m .
heinzle 2009 bianchi cosmologies with anisotropic matter : locally rotationally symmetric models ( preprint arxiv : 0911.0667v1 ) s. calogero and j.m .
heinzle 2010 oscillations toward the singularity of lrs bianchi type ix cosmological models with vlasov matter ( to appear in siam journal for appl .
dyn . syst . ) b. cropp and m. visser 2010 any spacetime has a bianchi type i spacetime as a limit ( preprint arxiv : 1008.4639v1 ) t. damour , m. henneaux and h. nicolai 2003 cosmological billiards . class .
quantum grav .
20 r145-r200 j.m .
heinzle and c. uggla 2006 dynamics of the spatially homogeneous bianchi type i einstein - vlasov equations .
23 , 3463 - 3490 j.m .
heinzle and c. uggla 2009 mixmaster : fact and belief . class .
quantum grav .
26 , 075016 a. kleinschmidt , m. koehn and h. nicolai 2009 supersymmetric quantum cosmological billiards .
phys . rev .
d 80:061701 v.g .
leblanc 1997 asymptotic states of magnetic bianchi i cosmologies . class .
quantum grav .
14 , 2281 - 2301 h. lee 2004 asymptotic behaviour of the einstein - vlasov system with a positive cosmological constant .
cambridge phil .
137 , 495 - 509 n. noutchegueme and d. dongo 2006 global existence of solutions for the einstein - boltzmann system in a bianchi type i spacetime for arbitrarily large initial data . class .
quantum grav .
23 , 2979 - 3004 n. noutchegueme and e.m .
tetsadjio 2009 global dynamics for a collisionless charged plasma in bianchi spacetimes . class .
quantum grav .
26 , 195001 o. heckmann and e. schcking 1962 gravitation : an introduction to current research , ed .
l. witten .
academic press , new york . chap .
rendall 1994 cosmic censorship for some spatially homogeneous cosmological models .
233 , 82 - 96 .
rendall 1996 the initial singularity in solutions of the einstein - vlasov system of bianchi type i. j. math phys .
37 , 438 - 451 .
a.d . rendall and k.p .
tod 1998 dynamics of spatially homogeneous solutions of the einstein - vlasov equations which are locally rotationally symmetric . class .
quantum grav .
16 , 1705 - 1726 a.d .
rendall and c. uggla 2000 dynamics of spatially homogeneous locally rotationally symmetric solutions of the einstein - vlasov equations . class .
quantum grav .
17 , 4697 - 4713 a.d .
rendall 2002 cosmological models and centre manifold theory .
rel . grav .
34 , 1277 - 1294 a.d .
rendall 2008 partial differential equations in general relativity .
oxford university press , oxford .
p. sandin and c. uggla 2008 bianchi type i models with two tilted fluids . class .
quantum grav .
25 , 225013 p. sandin 2009 tilted two - fluid bianchi type i models .
41 , 2707 - 2724 e. takou 2009 global properties of the solutions of the einstein - boltzmann system with cosmological constant in the robertson - walker space - time . commun .
sci . 7 , 399 - 410 j. wainwright and g.f.r .
ellis 1997 dynamical systems in cosmology .
cambridge university press , cambridge . | assuming that the space - time is close to isotropic in the sense that the shear parameter is small and that the maximal velocity of the particles is bounded , we have been able to show that for non - diagonal bianchi i - symmetric spacetimes with collisionless matter the asymptotic behaviour at late times is close to the special case of dust .
we also have been able to show that all the kasner exponents converge to @xmath0 and an asymptotic expression for the induced metric has been obtained .
the key was a bootstrap argument .
the sign conventions of @xcite are used . in particular , we use metric signature
+ + + and geometrized units , i.e. the gravitational constant g and the speed of light c are set equal to one .
also the einstein summation convention that repeated indices are to be summed over is used .
latin indices run from one to three .
@xmath1 will be an arbitrary constant and @xmath2 will denote a small and strictly positive constant .
they both may appear several times in different equations or inequalities without being the same constant .
a dot above a letter will denote a derivative with respect to the cosmological ( gaussian ) time @xmath3 . |
the qcd sum rules ( sr ) approach suggested in @xcite for description of the static properties of mesons and later employed for nucleons @xcite is based on the dispersion relation for the function describing the propagation of the system which carries the quantum numbers of the hadron . in the case of the proton this function , usually referred to as the `` polarization operator '' ,
can be written as @xmath2 with @xmath3 the four - momentum of the system , @xmath4 , @xmath5 is the unit matrix .
the dispersion relations @xmath6 are considered at @xmath7 .
this enables to expand the left hand sides ( lhs ) of eq .
( [ 2 ] ) in powers of @xmath8 .
the coefficients of the expansion are the qcd condensates.this is known as the operator product expansion ( ope ) @xcite .
this provides the perturbative expansion of the short distance effects , while the nonperturbative physics is contained in the condensates .
the higher terms of the ope contain the condensates of the higher dimension . the left hand side of eq .
( [ 2 ] ) can be written as @xmath9 where the lower index @xmath10 is the dimension of the corresponding qcd condensate .
the lowest order ope contribution to the structure @xmath11 is the quark loops , and does not contain condensates .
we denote it as @xmath12 . the lowest order ope contribution to the structure
@xmath13 is proportional to the scalar quark condensate @xmath14 , which has dimension @xmath0 .
this contribution contains a quark loop .
next comes the term with the gluon condensate @xmath15 with dimension @xmath1 , also containing the quark loops .
the ope terms of the higher dimension do not contain loops .
the loops provide the logarithmic terms @xmath16 on the left hand side , while the contributions with @xmath17 provide the power corrections @xmath18 .
the physical spectrum of polarization operator contains the single - particle states @xmath19 , etc . and the many - particle states , e.g. protons and pions .
the former manifest themselves as the poles of the integrand on the right hand side ( rhs ) of eq .
( [ 2 ] ) while the latter correspond to its cuts .
both the logarithmic terms and the power corrections on the lhs contribute to the terms originated by the poles and the cuts on the rhs .
the rhs of eq .
( [ 2 ] ) is usually approximated by the `` pole+continuum '' model @xcite in which the lowest lying pole is written down exactly , while the higher states are described by continuum .
the spectral function on the cut corresponding to continuum is approximated by discontinuity of the logarithmic terms on the lhs . under the latest assumption eqs .
( [ 2 ] ) take the form @xmath20 here @xmath21 , @xmath22 .
the upper index ope means that several lowest ope terms are included .
one usually applies the borel transform [ 1 - 3 ] to both sides of eq .
( [ 2 ] ) , which converts the functions of @xmath23 to the functions of the borel mass @xmath24 .
an important assumption is that there is an interval of the values of @xmath24 where the two sides of the sr have a good overlap , approximating also the true functions .
the standard interval ( `` duality interval '' ) is @xmath25 gev@xmath26 gev@xmath27 .
the proton mass @xmath28 , the residue at the nucleon pole @xmath29 and the continuum threshold @xmath30 are the unknowns of the qcd sr equations .
the set of parameters @xmath31 which minimize the function @xmath32 with @xmath33 and @xmath34 the borel transforms of the rhs and lhs of eq.([3 ] ) correspondingly is referred to as the solution of the qcd sum rules equations . here
@xmath35 is a set of @xmath36 points in the duality interval .
we put @xmath37 . the `` pole + continuum '' model for the spectrum makes sense only if the contribution of the proton pole ( treated exactly ) to the rhs of eq .
( [ 3 ] ) exceeds that of the continuum ( treated approximately ) .
we treat a solution which satisfy this condition as the physical one .
otherwise we call it an unphysical solution @xcite . in order to obtain the value of @xmath28 close to the physical value @xmath38 mev
one should include the ope terms up to those with @xmath39 .
if only the loop terms @xmath12 , @xmath40 and @xmath41 are included on the lhs of eq .
( [ 3 ] ) ( `` loop approximation '' ) , we immediately find a trivial solution @xmath42 , @xmath43 , in which the logarithmic terms on the lhs correspond to the second term on the rhs of eq.([3 ] ) .
this solution exists for any value of the nucleon mass @xmath28 .
there was a general believe that this is the only solution in the loop approximation .
this was viewed as one of the week points of the `` pole+continuum '' model and of the qcd sr approach as a whole @xcite . here
we demonstrate that limiting ourselves to the terms with @xmath44 and even with @xmath45 we have also another solution , which can be treated as a good lowest order approximation to the solution with @xmath46 .
keeping only the terms which contain the logarithmic loops we find @xmath47 and @xmath48 .
the explicit form of the contributions depends on the form of the local operator @xmath49 with the proton quantum numbers , often referred to as current " .
the form of @xmath49 is not unique , and we assume @xmath50 , suggested in @xcite .
one can find @xcite , @xcite @xmath51 with @xmath52 the ultraviolet cutoff , @xmath53 , while @xmath54 and @xmath55 are the scalar quark condensate and the gluon condensates multiplied by certain numerical factors @xmath56 the factors @xmath57 include the radiative corrections .
following @xcite we consider several cases .
the radiative corrections are neglected , in this case @xmath58 .
ii ) . leading logarithmic approximation ( lla ) , in which only the terms @xmath59 are included , but they are summed to all orders of @xmath10 .
the lla terms and the nonlogarithmic corrections of the order @xmath60 are taken into account .
all radiative corrections are included in the lowest order of @xmath60 .
the lla terms are expressed in terms of the function @xmath61 in eq .
( [ 4a ] ) @xmath62 is the qcd scale , while @xmath63 is the normalization point , the standard choice is @xmath64gev . in the case
ii we have @xmath65 , while @xmath66 . in the case
iii @xcite , @xcite @xmath67 with @xmath68gev@xmath69 . in the case
iv the value of @xmath70 changes to @xmath71 while @xmath72 remains the same as in the case ii .
since the contribution of the term @xmath41 is numerically small , we do not include the nonlogarithmic corrections here , keeping @xmath73 for the cases ii and iii .
actually the standard procedure is to consider the sr for the operators @xmath74 .
the factor @xmath75 is introduced in order to deal with the values of the order of unity ( in gev units ) . finally the sr equations in the loop approximation , except the case iv can be written as @xmath76 and @xmath77 with @xmath78 , while @xmath79 if all radiative corrections are neglected ( i ) , @xmath58 . in the lla ( ii )
@xmath80 , while @xmath66 . in these cases eq .
( [ 22 ] ) and eq.([22a ] ) are just the lowest terms of the standard nucleon sum rules [ 2,3 ] . adding the nonlogarithmic corrections of the order @xmath60 to the lla terms we obtain ( iii ) @xmath81 for the case when all corrections of the order @xmath60 are treated perturbatively ( iv ) , the first term on the lhs of eq . ( [ 22 ] ) takes a more complicated form .
it should be modified as @xcite @xmath82-\frac{\alpha_s}{\pi } \big[m^4w^2(1+\frac{3}{4}\gamma)e^{-\gamma}+m^6{\cal e}(-\gamma)\big ] , \label{25}\ ] ] with @xmath83 .
now we try to find the values of @xmath28 , @xmath29 and @xmath30 which minimize the difference between the lhs and rhs of eqs .
( [ 22 ] ) and ( [ 22a ] ) .
we employ the value @xmath84 mev .
thus in the radiative corrections @xmath85gev@xmath86 .
we vary the value of the scalar condensate around the point @xmath87mev@xmath88 , corresponding to @xmath89 gev@xmath90 we present also the results for @xmath91 gev@xmath90 corresponding to @xmath92mev@xmath88 and for @xmath93 gev@xmath90 corresponding to @xmath94mev@xmath88 .
they are shown in table i for the case when only the terms @xmath12 and @xmath40 are included , and in table ii , where the contribution @xmath41 is added .
here we employ the numerical value @xmath95 gev@xmath96 .
since the contribution of @xmath41 is numerically small , we do not vary the value of @xmath55 .
we see that the nucleon mass becomes closer to the physical value when we include the condensate with dimension @xmath1 .
it also becomes closer to the physical value while we include the radiative corrections .
the accuracy of the solution for the case when radiative corrections are included perturbatively is illustrated by fig . 1 . in .
2@xmath54-@xmath97 we show dependence of the parameters @xmath28 , @xmath29 and @xmath30 on the actual value of the scalar condensate for the same case with the largest dimension of the condensates involved @xmath98 and @xmath99 @xcite .
the qcd sum rules provide a solution with observed value of the nucleon mass if the terms with the dimension up to @xmath39 are included @xcite @xmath100 now we study the influence of the terms with @xmath17 by `` smooth switching on '' their contribution to the sum rules . to do this we construct the operators @xmath101 tracing dependence on the solutions on @xmath102 .
the results are shown in fig . 3
one can see that our solution converts smoothly to a physical one at @xmath103 .
the solution with @xmath104 , @xmath43 turns to the unphysical one .
the latter was studied in details in @xcite .
taking into account only the condensates of the lowest dimension @xmath0 or @xmath105 one immediately finds a trivial exact solution @xmath42 , @xmath43 for the nucleon qcd sum rules in framework of the `` pole+continuum '' model for the spectrum of the polarization operator . for a long time
there was a general believe that this is the only solution if the condensates with @xmath106 are included .
we find a nontrivial solution for the qcd sum rules equations in this case . assuming that @xmath107 we find a set of values @xmath108 and @xmath30 , which minimize the function @xmath109 defined by eq.([3x ] ) . it can be treated as the lowest order approximation to the ope solution .
this is important since the condensates of the lowest dimension @xmath14 and @xmath15 can be either calculated or related in terms of observables , while the condensates with larger dimensions are known with bigger uncertainties .
the convergence of the ope series is illustrated by fig .
2 . in fig .
3 we show how the trivial solution turns to an unphysical one during the smooth switching on of the condensates of higher dimension . in the same figure
we show how our solution turns to the physical one .
the authors acknowledge support by the grant rfbr 12 - 02 - 00158 .
shifman , a.i .
vainshtein and v.i .
zakharov , nucl .
* b147 * , 385 ; 448 ; 519 ( 1979 ) . b.l .
ioffe , nucl . phys . *
b188 * , 317 ( 1981 ) ; * b191 * , 591(e ) ( 1981 ) . b.l .
ioffe , l.n .
lipatov and v.s .
fadin , _ quantum chromodynamics _ ( cambridge univ . press , 2010 ) .
e. g. drukarev , m. g. ryskin , and v. a. sadovnikova , phys .
rev . d * 80 * , 014008 ( 2009 ) .
d. v. leinweber , ann . phys . * 254 * , 328 ( 1997 ) .
ioffe , z. phys .
c * 18 * , 67 ( 1983 ) .
v. a. sadovnikova , e. g. drukarev and m. g. ryskin , phys .
d * 72 * , 114015 ( 2005 ) .
a. a. ovchinnikov , a. a. pivovarov and l. r. surguladze , int . j. mod .
phys . a * 6 * , 2025 ( 1991 ) .
@xmath54,gev@xmath90&&@xmath28 , mev&@xmath29 , gev@xmath110&@xmath30 , gev@xmath27 & @xmath111 + & i&1.31&3.46&3.14&6.2(-2 ) + 0.50 & ii&1.44&5.15&4.56&9.2(-2 ) + & iii&1.23&3.47&2.81&1.6(-2 ) + & iv&1.18&2.90&2.41&2.7(-2 ) + & i&1.36&4.35&3.47&8.1(-2 ) + 0.55&ii&1.49&6.49&5.27&1.4(-1 ) + & iii&1.27&4.33&3.08&2.3(-2 ) + & iv&1.23&3.65&2.64&3.4(-2 ) + & i&1.40&5.37&3.84&1.0(-1 ) + 0.60&ii&1.53&8.01&6.19&2.0(-1 ) + & iii&1.31&5.32&3.37&2.3(-2 ) + & iv&1.27&4.50&2.89&4.3(-2 ) + @xmath54,gev@xmath90&&@xmath28 , mev&@xmath29 , gev@xmath110&@xmath30 , gev@xmath27 & @xmath111 + & i&1.20&2.61&2.65&1.2(-2 ) + 0.50 & ii&1.35&4.12&3.94&2.2(-2 ) + & iii&1.16&2.98&2.57&2.9(-2 ) + & iv&1.10&2.31&2.10&6.4(-3 ) + & i&1.25&3.34&2.96&1.9(-2 ) + 0.55&ii&1.40&5.26&4.51&4.1(-2 ) + & iii&1.21&3.75&2.83&6.1(-3 ) + & iv&1.15&2.96&2.33&9.9(-3 ) + & i&1.30&4.18&3.29&2.9(-2 ) + 0.60&ii&1.45&6.56&5.21&7.0(-2 ) + & iii&1.25&4.64&3.11&1.1(-2 ) + & iv&1.19&3.71&2.57&1.1(-2 ) +
1 . consistency of the lhs and rhs of the sum rules for @xmath112 gev@xmath90 .
the terms @xmath12 and @xmath41 are included on the lhs of eqs .
( [ 22 ] ) for @xmath113 , and the term @xmath40 is included on the rhs of ( [ 22a ] ) for @xmath114 .
radiative corrections of the order @xmath60 are included perturbatively ( iv ) .
solid lines and dashed lines show the lhs - to - rhs ratio for eqs .
( [ 22 ] ) and ( [ 22a ] ) correspondingly .
+ fig . 2 .
dependence of the values of nucleon parameters on the value of quark scalar condensate for various values of the largest dimension @xmath115 of the condensates .
the dotted , dashed and solid lines are for @xmath116 , @xmath106 and @xmath99 correspondingly .
radiative corrections of the order @xmath60 are included perturbatively ( iv ) + fig .
evaluation of solutions during the `` smooth switching on '' of the condensates of higher dimension . dashed line evaluation of the trivial solution to the unphysical one .
dashed - dotted line
evaluation of our solution to the physical one .
the radiative corrections of the order @xmath60 are included perturbatively ( iv ) , @xmath112 gev@xmath90 . | there was a general believe that the nucleon qcd sum rules which include only the quark loops and thus contain only the condensates of dimension @xmath0 and @xmath1 have only a trivial solution .
we demonstrate that there is also a nontrivial solution .
we show that it can be treated as the lowest order approximation to the solution which includes the higher terms of the operator product expansion .
inclusion of the radiative corrections improves the convergence of the series . |
diluted magnetic semiconductors ( dms ) are attracting much attention lately due to their potential for device applications in the growing field of spintronics .
in particular , a large number of dms studies have focused on iii - v compounds where mn doping in inas and gaas has been achieved using molecular beam epitaxy ( mbe ) techniques .
the main result of recent experimental efforts is the discovery@xcite of ferromagnetism at a curie temperature @xmath1@xmath2110 k in ga@xmath3mn@xmath4as , with mn concentrations @xmath5 up to 10% .
it is widely believed that this ferromagnetism is `` carried induced '' , with holes donated by mn ions mediating a ferromagnetic interaction between the randomly localized mn@xmath6 spins . in practice ,
anti - site defects reduce the number of holes @xmath7 from its ideal value @xmath7=@xmath5 , leading to a ratio @xmath8=@xmath9 substantially smaller than 1 . until recently
, theoretical descriptions of dms materials could be roughly classified in two categories .
on one hand , the multi - band nature of the problem is emphasized as a crucial aspect to quantitatively understand these materials.@xcite in this context the lattice does not play a key role and a continuum formulation is sufficient .
the influence of disorder is considered on average . on the other hand , formulations based on the possible strong localization of carriers at the mn - spin sites
have also been proposed.@xcite in this context a single impurity - band description is considered sufficient for these materials . still within the single - band framework , but with carrier hopping not restricted to the mn locations , recent approaches to the problem have used dynamical mean - field@xcite or reduced - basis@xcite approximations .
an effective hamiltonian for ga@xmath3mn@xmath4as was derived in the dilute limit and studied in ref . .
all these calculations are important in our collective effort to understand dms materials . however , it is desirable to obtain a more general view of the problem of ferromagnetism induced by a diluted set of spins and holes . to reach this goal
it would be better to use techniques that do not rely on mean - field approximations and , in addition , select a model that has both the continuum and impurity - band formulations as limiting cases .
such an approach would provide information on potential procedures to further enhance @xmath1 and clarify the role of the many parameters in the problem .
in addition , these general considerations will be useful beyond the specific details of ga@xmath3mn@xmath4as , allowing us to reach conclusions for other dms .
an effort in this direction was recently initiated by mayr and two of the authors.@xcite a detailed monte carlo study of a simple model for dms already allowed us to argue that @xmath1 could be further enhanced perhaps even to room temperature by further increasing @xmath5 and @xmath8 from current values in dms samples .
these predictions seem in agreement with recent experimental developments since very recently ga@xmath3mn@xmath4as samples with @xmath1 as high as @xmath10 k were prepared,@xcite a result believed to be caused by an enhanced free - hole density . also samples with @xmath11
k and @xmath12 were reported in ref . , and , very recently , a @xmath1 of 140k was achieved on high quality gamnas films grown with arsenic dimers.@xcite there seems to be plenty of room to further increase the critical temperatures according to these theoretical calculations .
the model used in ref . was a single - band lattice hamiltonian , also studied by other groups , which does not have the multi - band characteristics and spin - orbit couplings of the real problem .
however , it contains spin and holes in interaction and it is expected to capture the main qualitative aspects of carrier - induced ferromagnetism in dms materials .
the choice of a single - band model allows us to focus on the essential aspects of the problem , leaving aside the numerical complexity of the multi - band hamiltonian , which unfortunately can not be studied with reliable numerical techniques at present without introducing further approximations.@xcite note that a quite similar philosophy has been followed in the related area of manganites for many years .
in fact , it is well - known that the orbital order present in those compounds can only be studied with a two - band model . however , single - band approaches are sufficient to investigate the competition between ferromagnetic and anti - ferromagnetic states , an effect recently argued to be at the heart of the colossal magnetoresistance phenomenon.@xcite as a consequence , a qualitatively reliable study of the single - band dms model with itinerant holes and localized classical spins
is expected to be important for progress in the dms context as well . in our previous publication@xcite the phase diagram of the single - band model for dms
was already sketched.@xcite this phase diagram is reproduced in fig.[fig : summary ] .
the only parameter in the hamiltonian is the ratio @xmath0 , where @xmath13 is the local anti - ferromagnetic coupling between the spins of carriers and localized mn@xmath6 ions , and @xmath14 is the carrier - hopping amplitude ( for details , see the full form of the hamiltonian in the next section ) . due to its large value 5/2
, the mn spin is assumed to be classical for numerical simplicity . at small @xmath0 ,
individual carriers are only weakly bounded to the mn spins .
even for small realistic values of @xmath5 , the carrier wave functions strongly overlap and a large - bandwidth itinerant band dominates the physics . in this regime
, @xmath1 grows with @xmath0 .
this portion of the phase diagram is referred to as ` weak coupling ' in fig.[fig : summary ] , and in our opinion it qualitatively corresponds to the continuum - limit approach pursued in refs . .
in the other extreme of large @xmath0 , the previous effort@xcite found a @xmath1 that is much suppressed , converging to zero as @xmath0 increases ( at least at small values of @xmath5 ) .
the reason for this behavior is the strong localization of carriers at mn spin locations , to take advantage of the strong @xmath13 coupling .
the localization suppresses the mobility and the carrier - induced mechanism is no longer operative .
however , when several mn spins are close to one another small regions can be magnetized efficiently . as a consequence , a picture emerges where small islands of ferromagnetism are produced at a fairly large temperature scale @xmath15 ( @xmath16@xmath1 ) that grows with @xmath13 , but a global ferromagnetic state is only achieved upon further reducing the temperature such that the overlaps of wave functions induce a percolation - like process that aligns the individual preformed clusters the proper procedure to distinguish between these two regimes is through an analysis of short- and long - distance spin correlations . in our opinion , the effort of refs
. belongs to this class , and it corresponds to the ` strong coupling ' region of fig.[fig : summary ] . between the weak and strong coupling domains , an _ intermediate region _ provides a natural interpolation between those two extreme cases .
the curie temperature is optimal ( i.e. , the largest ) in this intermediate zone , at fixed values of @xmath8 and @xmath5 . in this regime , there is sufficient interaction between the clusters to become globally ferromagnetic , and at the same time the @xmath13 is sufficiently strong to induce a robust @xmath1 . if chemical control over @xmath0 were possible , this intermediate region would be the most promising to increase the curie temperature ( again , at fixed @xmath5 and @xmath8 ) .
however , it may even occur that the dms materials are already in the intermediate optimal range of @xmath0 couplings . an indirect way to test this hypothesis relies on calculations of dynamical and transport properties , which are presented in this paper , and its comparison with experiments .
in particular , the experimentally measured d.c .
resistivity @xmath17 of dms compounds has a nontrivial shape with a ( poor ) metallic ( d@xmath17/d@xmath18@xmath160 ) behavior below @xmath1 , which turns to insulating ( d@xmath17/d@xmath18@xmath190 ) at higher temperatures .
this nontrivial profile resembles results reported in the area of manganites , which also have a metal - insulator transition as a function of temperature,@xcite although in those materials the changes in resistivity with temperature are far more dramatic than in dms .
the formal similarities dms - manganites have already been remarked in previous literature,@xcite and suggest a common origin of the @xmath17 vs. temperature curves . in mn - oxides
it is believed that above the curie temperature , preformed ferromagnetic clusters with random orientations contribute to the insulating behavior of those materials.@xcite for dms , a similar rationalization can be envisioned if the state of relevance above @xmath1 has preformed magnetic clusters .
from the previous discussion@xcite it is known that the intermediate @xmath0 region can have clusters without collapsing @xmath1 to a very small value . at the curie temperature
, the alignment of these preformed moments leads to a metallic state .
an explicit calculation of the resistivity using techniques borrowed from the mesoscopic physics context is provided in this paper , and the results support the conjecture that clustered states can explain the transport properties of dms . these unveiled analogies with manganites are not just accidental . clearly , approaches to dms that rely on mobile carriers and localized spins in interaction have close similarities with the standard single - band double - exchange model where only one @xmath20-orbital is considered .
the key difference is the presence of strong dilution in dms as compared with manganites .
it is interesting to remark that in agreement with the ` clustered ' state described here , recently timm and von oppen have shown that _ correlated _ defects in dms are needed to describe experimental data.@xcite their simulations with coulombic effects incorporated lead to cluster formation , with sizes well beyond those obtained from a random distribution of mn sites as considered here . in this respect , the results of ref .
provide an even more dramatic clustered state than ours .
if the mn spins were not distributed randomly in our simulations but in a correlated manner , the state above @xmath1 would be even more insulating than reported below and the physics would resemble much closer that observed in manganites . note also that kaminski and das sarma have also independently arrived to a polaronic state@xcite that qualitatively resembles the clustered state discussed here .
it is also interesting that in recent ion - implanted ( ga , mn)p : c experiments that reported a high curie temperature , the presence of ferromagnetic clusters were observed and they were attributed to disorder and the proximity to a metal - insulator transition.@xcite even directly in ( ga , mn)as , the inverse magnetic susceptibility deviates from the curie law at temperatures above @xmath1 ( fig.2(b ) of ref . ) , which may be an indication of an anomalous behavior . in manganites , these same anomalies in the susceptibility
are indeed taken as evidence of the formation of clusters.@xcite the paper is organized as follows : in section [ sec : method ] the details of the method of simulation , definitions , and conventions are presented .
section [ sec : dos ] describes the results for the density - of - states ( dos ) of this model , the appearance of an `` impurity band '' at large @xmath0 , and the relation of the dos with the optimal @xmath1 .
the temperature and carrier dependence of the optical conductivity and drude weight are presented in section [ sec : sigma ] , with results compared with experiments . in section
[ sec : conductance ] , the resistance of a small cluster is calculated and also compared with experimental data .
the intermediate coupling regime is the only one found to qualitatively reproduce the available resistivity vs. temperature curves . from this perspective , both the weak and strong coupling limits are not realistic to address dms materials . in section [ sec : magnet ] , the influence of a magnetic field is studied by calculating the magnetoresistance .
large effects are observed .
section [ sec : cluster ] describes the characteristics of the clustered - state regime above @xmath1 at strong coupling .
finally , in section [ sec : conclusion ] the ` transport ' phase diagram is presented , and it is discussed in what region it appears to be the most relevant to qualitatively reproduce results for ga@xmath3mn@xmath4as mbe - grown films .
there are two degrees of freedom in the dms problem described here : ( i ) the local magnetic moments corresponding to the 5 electrons in the @xmath21 shell of each mn impurity , with a total spin 5/2 , and ( ii ) the itinerant carriers produced by the mn impurities .
since these mn impurities substitute ga atoms in the zinc - blende structure of gaas , there is , in principle , one hole per mn .
however , it has been found that the system is heavily compensated and , as a consequence , the actual concentration of carriers is lower . in the present study ,
both the density of mn atoms , @xmath5 , and the density of carriers , @xmath7 , are treated as independent input parameters .
moreover , the ratio @xmath22 is defined , which is a measure of the compensation of the system , e.g. for @xmath23 the system is totally compensated and for @xmath24 there is no compensation .
+ the hamiltonian of the system in the one - band approximation can be written as : @xmath25 where @xmath26 creates a carrier at site @xmath27 with spin @xmath28 .
the carrier - spin operator interacting antiferromagnetically with the localized mn - spin @xmath29 is @xmath30 . through nearest - neighbor hopping
, the carriers can hop to @xmath31 site of the square or cubic lattice .
the interaction term is restricted to a randomly selected but fixed set of sites , denoted by @xmath32 .
note that in addition to the use of a single band , there are other approximation in the model described here : ( i ) the hubbard @xmath33 is not included .
this is justified based on the low - carrier concentration of the problem , since in this case the probability of double occupancy is small . in addition , mn - oxide investigations@xcite have shown that an intermediate or large @xmath13 coupling acts similarly as @xmath33 , also suppressing double occupancy .
( ii ) a nearest - neighbors anti - ferromagnetic coupling between the mn spins is not included . again
, this is justified as long as @xmath5 is small .
( iii ) finally , also potential disorder is neglected ( together with the spin , the mn sites should in principle act as charge trapping centers ) .
this approximation is not problematic when the fermi energy is larger than the width of the disorder potential ( i.e. large @xmath8 and @xmath5 ) .
however , in the ( @xmath8 , @xmath5 ) range analyzed in this paper it becomes more questionable to neglect this effect .
its influence will be studied in a future publication . in the present study
the carriers are considered to be electrons and the kinetic term describes a conduction band . however , in ga@xmath3mn@xmath4as the carriers are holes and the valence band has to be considered instead .
although the latter is more complicated , the simplified treatment followed here should yield similar results for both cases.@xcite + the local spins are assumed to be classical which allows the parametrization of each local spin in terms of spherical coordinates : @xmath34 . the exact - diagonalization method for the fermionic sector
is described in this section , following ref . .
the partition function can be written as : @xmath35 where @xmath36 is the partition function of the fermionic sector , @xmath37 , @xmath38 is the number operator , and @xmath39 is the chemical potential . + in the following , an hypercubic lattice of dimension @xmath40 , length @xmath41 , and number of sites @xmath42 will be considered .
since @xmath43 is a hermitian operator , it can be represented in terms of a hermitian matrix which can be diagonalized by an unitary matrix @xmath44 such that @xmath45 the basis in which the matrix @xmath46 is diagonal is given by the eigenvectors @xmath47 , @xmath48 , @xmath49 , where the fermionic operators used in this basis are obtained from the original operators through @xmath50=@xmath51 , with @xmath52 running from 1 to @xmath53 .
defining @xmath54=@xmath55 and denoting by @xmath56 the eigenvalues of @xmath55 , the trace can be written @xmath57 since in the @xmath58 basis , the operator @xmath59 is @xmath60 , and the number operator can be replaced by its eigenvalues .
thus , eq . ( [ eq : partition1 ] ) can be rewritten as @xmath61 which is the formula used in the simulations .
the integral over the angular variables in eq .
( [ eq : partition ] ) can be performed using a classical monte carlo simulation.@xcite the eigenvalues must be obtained for each classical spin configuration using library subroutines .
finding the eigenvalues is the most time consuming part of the numerical simulation .
the integrand is clearly positive .
thus , `` sign problems '' in which the integrand of the multiple integral under consideration can be non - positive , are fortunately not present in our study .
+ although the formalism is in the grand - canonical ensemble , the chemical potential was adjusted to give the desired carrier density , @xmath7 . to do so
, the equation @xmath62 was solved for @xmath39 at every monte carlo step by using the newton - raphson method.@xcite this technique proved very efficient in adjusting with precision the desired electronic density and , as a consequence , our analysis can be considered as in the canonical ensemble as well , with a fixed number of carriers .
+ usually , 2,000 - 5,000 monte carlo iterations were used to let the system thermalize , and then 5,000 -10,000 additional steps were carried out to calculate observables , measuring every five of these steps to reduce autocorrelations .
+ since for fixed parameters there are many possibilities for the random location of mn impurities , results are averaged over several of these disorder configurations .
approximately 10 - 20 disorder configurations were generally used for small lattices ( @xmath63 and 10@xmath6410 ) and 4 - 8 for larger lattices
( @xmath65 and 12@xmath6412 ) .
the inevitable uncertainties arising from the use of a small number of disorder configuration does not seem to affect in any dramatic way our conclusions below .
this is deduced from the analysis of results for individual disorder samples .
the qualitative trends emphasized in the present paper are present in all of these configurations .
quantities that depend on the mn degrees of freedom ( @xmath66 and @xmath67 in the previous formalism ) are calculated simply by averaging over the monte carlo configurations .
note that any observable that does not have the continuous symmetry of the hamiltonian will vanish over very long runs .
thus , it is standard in this context to calculate the absolute value of the magnetization , @xmath68 , as opposed to the magnetization vector .
another useful quantity is the spin - spin correlation , defined by : @xmath69 where @xmath70 is the number of non - zero terms in the sum .
the correlation at a distance @xmath21 is averaged over all lattice points that are separated by that distance , but since the system is diluted , the quantity must be normalized to the number of pairs of spins separated by @xmath21 , to compare the results for different distances .
+ the observables that directly depend on the electronic degrees of freedom can be expressed in terms of the eigenvalues and eigenvectors of the hamiltonian matrix @xmath43 .
the dos , @xmath71 , is simply given by @xmath72 .
however , the majority and minority dos , @xmath73 and @xmath74 , were also calculated in this study .
@xmath73 indicates the component that aligns with the local spin , i.e. , @xmath73 is the fourier transform of @xmath75 , where @xmath76 .
then : @xmath77 } , \label{eq : nofomega}\end{aligned}\ ] ] where for sites @xmath27 without an impurity @xmath78 is assumed .
a similar expression is valid for @xmath74 .
the optical conductivity was calculated as : @xmath79 where the current operator is : @xmath80 with @xmath81 the unit vector along the @xmath5-direction . for @xmath82 ,
@xmath83 can be written as : @xmath84 both @xmath71 and @xmath83 were broadened using a lorentzian function as a substitute to the @xmath85-functions that appear in eqs .
( [ eq : nofomega ] ) and ( [ eq : sigma ] ) .
the width of the lorentzian used was @xmath86 in units of the hopping , @xmath14 .
+ the optical conductivity in @xmath21 dimensions obeys the sum rule : @xmath87 where @xmath40 is the drude weight and @xmath88 is the kinetic energy : @xmath89 although the drude weight gives a measure of the d.c .
conductivity properties of the cluster , the `` mesoscopic '' conductance was also calculated in order to gather additional information .
the details are explained in section [ sec : conductance ] .
in this section , the density - of - states calculated at several couplings is shown and discussed .
a few basic facts about the dos of the model , eq . ( [ eq : ham ] ) , are presented first .
if carrier localization effects are not taken into account and a ferromagnetic state is considered at strong @xmath0 coupling , then two spin - split bands ( `` impurity bands '' ) will appear for this model , each with weight proportional to @xmath5 , at each side of the unperturbed band which would have weight @xmath90 ( fig .
[ fig : doscartoon ] ) . for partially compensated samples , @xmath91
, the chemical potential will be located somewhere in the first spin - split band , and , as a consequence , only this band would be relevant in the model eq .
( [ eq : ham ] ) . in practice ,
monte carlo simulations show that , in the regime of interest where @xmath1 is optimal , the impurity band is not completely separated from the unperturbed band . for this reason ,
both the unperturbed band and the impurity band have to be considered in quantitative calculations . to illustrate this , figure [ fig : dosparamj ] shows @xmath71 vs. @xmath92 for different @xmath0 s in two dimensions . with growing @xmath0 ,
the `` impurity band '' begins to form for @xmath93 , and at @xmath94 it is already fairly separated from the main band . however , note that for this extreme regime of @xmath0 , @xmath1 is far lower than for @xmath95 , as discussed before in ref . .
in fact , the optimal @xmath0 for the @xmath96 parameters used here was found to be in the range 2.0 - 4.0 .
this implies that the model shows the highest @xmath1 when the `` impurity band '' is about to form , but it is not yet separated from the main band .
this introduces an important difference between our approach to dms materials , and the theoretical calculations presented in ref . .
figure [ fig : dosparamj3d](a)-(c ) indicates that a similar behavior is found in three dimensions . in this case , once again , the optimal @xmath1 occurs for intermediate @xmath0 , as seen in fig .
[ fig : dosparamj3d](d ) where the phase diagram obtained from a @xmath65 lattice with @xmath5=@xmath97 and @xmath8=@xmath98 is shown .
+ the physical reason for this behavior is that at very large @xmath0 , the states of the impurity band are highly localized where the mn spins are .
local ferromagnetism can be easily formed , but global ferromagnetism is suppressed by the concomitant weak coupling between magnetized clusters . both , figures [ fig : dosparamj ] and [ fig : dosparamj3d ] represent results obtained at particular values of parameters @xmath5=@xmath97 , @xmath8=@xmath98 , and @xmath99 .
however , results for several other sets @xmath100 were gathered as well ( not shown ) .
the @xmath5 dependence of the dos is simple : increasing @xmath5 produces a proportional increase in the number of states of the impurity band , and a corresponding decrease of the main band weight ( see fig.[fig : doscartoon ] ) . in terms of @xmath8 , increasing the effective carrier concentration by a small amount
was found to simply shift the chemical potential to the right .
concerning the temperature dependence of the dos , at low temperature the states contributing to the impurity band are polarized , i.e. the system is ferromagnetic , as shown by the different weights of the majority and minority bands ( figs .
[ fig : dosparamj ] and [ fig : dosparamj3d ] ) . as the temperature increases
, spin disorder grows due to thermal fluctuations , and when @xmath101 , i.e. in the paramagnetic regime , the bands become symmetric .
it appears that the only change that a small increase in the carrier concentration produces is an increase in the chemical potential . however , experimentally the appearance of a pseudo - gap at the fermi energy has been reported.@xcite although the finite size effects in the present theoretical calculation do not allow for a detailed study of @xmath71 at or very near the fermi energy with high enough precision , in certain cases a pseudo - gap was indeed observed in the present study at the chemical potential , particularly when the system is in the clustered regime . while this result certainly needs confirmation , it is tempting to draw analogies with the more detailed calculations carried out in the context of manganites , where the presence of a pseudo - gap both in theoretical models and in experiments is well established.@xcite pseudo - gaps are also present in under - doped high temperature superconductors
. given the analogies between dms materials and transition - metal oxides unveiled in previous investigations,@xcite it would not be surprising that dms presents a pseudo - gap in the clustered regime as well .
more work is needed to confirm these speculations .
results presented in this section for the dos qualitatively agree with those found in ref .
using the dynamical mean - field technique ( dmft ) in the coupling regime studied there .
however , the dmft approach is a mean - field approximation local in space and , as a consequence , the state emphasized in this paper with randomly distributed clusters can not be studied accurately with such a technique
. experimental evidence of the formation of the impurity band for ga@xmath3mn@xmath4as has been provided by okabayashi _
et al . _ in refs .
- ( see also ref . ) .
it is interesting to remark that half the total width of the impurity band obtained in our simulations is about @xmath102 ( see , e.g. , figs .
[ fig : dosparamj ] and [ fig : dosparamj3d ] ) , and using @xmath14=@xmath98ev@xcite this is equal to 0.6ev , in good agreement with the width 0.5ev estimated from the photoemission experimental measurements mentioned before.@xcite the local dos is studied next .
the purpose of this analysis is to further understand the inhomogeneous state that forms as a consequence of mn - spin dilution and concomitant carrier localization .
the local dos is shown in fig .
[ fig : ldos ] on an 8@xmath648 lattice at @xmath103 , @xmath104 , @xmath105 , and @xmath106 . in this particular case
only , an arrangement of classical spins was introduced ` by hand ' such that the clusters are clearly formed .
in this way , sites can easily be classified in three groups , as discussed below . despite using a particularly chosen spin configuration ,
the resulting dos is similar to those obtained previously ( fig.[fig : dosparamj ] ) using a truly random distribution of spins .
lattice sites are classified as follows : the first group is defined to contain lattice sites that have a classical spin and belong to a spin cluster .
the second class is composed of sites that have an isolated classical spin , i.e. nearby sites do not have other classical spins . finally , lattice sites without classical spins in the same site or its vicinity belong to the third class . @xmath107 ,
@xmath108 , and @xmath109 denote the local dos at sites corresponding to each of the three classes , respectively .
it can be observed that @xmath107 contributes to states inside the impurity band ( note the sharp peak near -1 ) , but also contributes to the main band since the @xmath0 used is intermediate .
@xmath108 contributes a sharp peak near zero , and also has weight in the main band .
this can be explained as follows : the electron is weakly localized at an isolated site yielding a state at @xmath110 or , since @xmath111 for @xmath112 , @xmath113 ( see fig.[fig : ldos]a ) .
this interesting result shows that many of the states near zero in figs.[fig : dosparamj ] and [ fig : dosparamj3d ] may be _ localized _ , and they do not contribute to the conductivity .
finally , in @xmath109 the empty sites contribute weight only in nearly unperturbed states , i.e. , states that belong to the main band . in conclusion , the local carrier density is very inhomogeneous and this fact has an important effect on the form of the site - integrated dos .
our results show that scanning tunneling microscopy ( stm ) experiments would be able to reveal the clustered structure proposed here , if it indeed exists , when applied to dms materials .
in the previous section , it was argued that at the optimal coupling @xmath0 where @xmath1 is maximized the impurity band is not completely separated from the main band .
this conclusion is supported by results obtained from the optical conductivity as well , which is shown in fig . [
fig : swparamj](a ) for a two - dimensional lattice .
this optical conductivity has two main features : ( i ) a zero - frequency or drude peak and ( ii ) a finite - frequency broad peak which is believed to correspond to transitions from the impurity band to the main band , as argued below . from fig .
[ fig : swparamj](a ) , it is observed that inter - band transitions are not much relevant at weak coupling ( @xmath0=@xmath114 ) , but they appear with more weight at intermediate couplings ( @xmath0=@xmath115 ) , where @xmath1 is optimal .
it is worth remarking that at @xmath0 strong enough , e.g. @xmath0=@xmath116 , when the impurity band is well - formed , the finite - frequency peak is weaker in strength than for the optimal @xmath0 due to localization .
in fact , if @xmath0 were so large that the `` impurity band '' is completely separated from the unperturbed band , then carrier localization would be so strong that inter - band transitions would not be possible .
the drude peak is plotted separately in fig .
[ fig : swparamj](b ) as a function of @xmath0 at low temperatures . as expected , when @xmath0 increases and localization sets in , the conductivity of the cluster decreases .
similar results for the optical conductivity and drude weight are found in three dimensions ( see fig .
[ fig : swparamj3d ] ) . vs. @xmath92 for a 10@xmath6410 periodic system with 26 spins ( @xmath117 ) , 10 electrons ( @xmath118 ) , @xmath99 , and for different @xmath0 s as shown .
( b ) drude weight , @xmath40 , vs. @xmath0 for the same lattice and parameters as in ( a ) .
[ fig : swparamj],width=340 ] vs. @xmath92 for a @xmath65 periodic system with 54 spins ( @xmath117 ) , 16 electrons ( @xmath118 ) , @xmath99 , and for different @xmath0 s as shown .
( b ) @xmath40 vs. @xmath0 for the same lattice and parameters as in ( a ) .
[ fig : swparamj3d],width=340 ] the rise in absorption that appears for @xmath119 at intermediate frequencies is due to inter - band transitions , i.e. transitions from the impurity band to the unperturbed band .
the frequency of this peak , @xmath120 , depends upon @xmath0 as well as carrier concentration @xmath8 , but for @xmath119 ( which is an intermediate value ) and @xmath104 ( a realistic value ) , it was observed that @xmath121 .
the model parameter @xmath14 was previously estimated@xcite to be @xmath14@xmath2@xmath122 , yielding @xmath123 , in agreement with experiments ( see below ) .
the temperature dependence of @xmath83 is shown in fig .
[ fig : swt ] , and it is as follows : for @xmath124 , the drude weight , @xmath40 , decreases in intensity as temperature increases and the same behavior is observed for the finite - frequency peak .
when @xmath125 , @xmath40 increases slightly at first and then remains constant as the temperature is increased further .
this slight raise in @xmath40 is subtle and could be linked to the decrease in resistivity observed for @xmath126 for intermediate @xmath0 as explained in the next section , since the system is clustered just above @xmath1 . since both the optical conductivity and drude weight decrease with temperature , to satisfy the sum rule the kinetic energy must also decrease as temperature increases .
this is indeed observed ( not shown ) and is to be expected since localization of the wave function is stronger above @xmath1 . ) .
shown is @xmath83 vs. @xmath92 for a 10@xmath6410 periodic system with 26 spins ( @xmath117 ) , @xmath105 , @xmath127 , and for different temperatures , as indicated .
_ inset _ : drude weight , @xmath40 , vs. temperature @xmath18.[fig : swt],width=283 ] a broad peak at around @xmath128ev is experimentally observed in the optical conductivity of ga@xmath3mn@xmath4as , as shown in fig .
[ fig : katsumoto ] which is reproduced from ref . .
this feature has been explained before in two different ways : ( i ) as produced by transitions from the impurity band to the gaas valence band@xcite or ( ii ) as caused by inter - valence band transitions.@xcite due to the frequency range and temperature behavior observed , our study supports the first possibility , i.e. , that the finite - frequency peak observed for ga@xmath3mn@xmath4as at around @xmath128 ev is due to transitions from the impurity band to the main band .
recent observations@xcite also appear to support this view .
spectra for a metallic sample prepared by low temperature annealing .
the temperatures from down to top are 300 , 250 , 200 , 150 , 120 , 100 , 80 , 60 , 40 , 20 , 10 and 4.2 k. ( from katsumoto _ et al._@xcite)[fig : katsumoto ] , width=264 ] the drude weight increases in intensity with increasing carrier density , as can be seen from fig .
[ fig : swp ] .
in fact , the ratio @xmath8 of carriers to mn concentration , which is a measure of the compensation of the samples , could explain , within the framework of this model , the different behavior observed at low frequencies for gamnas and inmnas . in the former case , no tail of the drude peak is found .
however , for inmnas , a clear drude tailing is observed,@xcite with increasing intensity as the temperature increases . at large enough doping , our calculations based on model eq .
( [ eq : ham ] ) predict a drude - like peak for @xmath83 which could correspond to the regime valid for inmnas , while for low doping this study predicts a very small drude peak , which is the case for gamnas .
unfortunately , the precise carrier concentrations for both materials are not known experimentally with precision .
-dependence of the optical conductivity .
shown is @xmath83 , including the drude weight , vs. @xmath92 for a 10@xmath6410 lattice , @xmath103 , @xmath105 , @xmath99 , and different values of @xmath8 . in the direction of the arrow , @xmath8 takes the values 0.08 , 0.12 , 0.20 , 0.30 , 0.38 , and 0.5 .
_ inset _ : drude weight , @xmath40 , vs. fraction of carriers , @xmath8 .
[ fig : swp],width=283 ]
the conductance , @xmath129 , is calculated using the kubo formula adapted to geometries usually employed in the context of mesoscopic systems.@xcite the actual expression is @xmath130 , \label{eq : conductance}\ ] ] where @xmath131 is the velocity operator in the @xmath5 direction and @xmath132 is obtained from the advanced and retarded green functions using @xmath133 , where @xmath134 is the fermi energy . the cluster is considered to be connected by ideal contacts to two semi - infinite ideal leads , as represented in fig .
[ fig : leads ] .
current is induced by an infinitesimal voltage drop .
this formalism avoids some of the problems associated with finite systems , such as the fact that the conductivity is given by a sum of dirac @xmath85 functions , since the matrix eigenvalues are discrete .
a zero - frequency delta peak with finite weight corresponds to an ideal metal having zero resistance unless an arbitrary width is given to that delta peak .
in addition , in numerical studies sometimes it occurs that the weights of the zero - frequency delta peak are negative due to size effects.@xcite for these reasons , calculations of d.c .
resistivity using finite close systems are rare in the literature .
all these problems are avoided with the formulation described here .
, width=302 ] the entire equilibrated cluster as obtained from the mc simulation is introduced in the geometry of fig .
[ fig : leads ] .
the ideal leads enter the formalism through self - energies at the left / right boundaries , as described in ref . .
in some cases a variant of the method explained in ref . was used in this paper to calculate the conductance ( this modification was suggested to us by j. vergs ) . instead of connecting the cluster to an ideal lead with equal hoppings ,
a replica of the cluster was used at the sides .
this method , although slightly slower , takes into account all of the monte carlo data for the cluster , including the periodic boundary conditions .
in addition , averages over the random mn spin distributions are carried out .
the physical units of the conductance @xmath129 in the numerical simulations are @xmath135 as can be inferred from eq .
( [ eq : conductance ] ) . in figure
[ fig : rho4d3 ] the inverse of the conductance , which is a measure of the resistivity , is plotted for a three - dimensional lattice at weak , intermediate , and strong coupling , at fixed @xmath5=@xmath97 and @xmath8=@xmath98 . for the weak coupling regime ( @xmath0=@xmath114 )
the system is weakly metallic at all temperatures . in the other limit of strong couplings
, @xmath136 decreases with increasing temperature , indicating a clear insulating phase , as a result of the system being in a clustered state at the temperatures explored,@xcite with carriers localized near the mn spins . at the important intermediate couplings emphasized in our effort
, the system behaves like a dirty metal for @xmath137 , while for @xmath138 , @xmath136 slightly decreases with increasing temperature , indicating that a soft metal to insulator transition takes place near @xmath1 . for @xmath139 , where the system is paramagnetic , @xmath136 is almost constant .
note that for strong enough @xmath0 , @xmath140 and therefore no metallic phase is present .
, with temperature in three dimensions .
shown is @xmath141 vs. @xmath18 on @xmath63 lattices , 16 spins ( @xmath103 ) , and 5 carriers ( @xmath127 ) for the @xmath0 s indicated .
an average over 20 disorder configurations has been performed in each case .
units are shown in two scales , @xmath142 on the left and m@xmath143 cm on the right , with @xmath144 and assuming .
the estimated critical temperatures are also shown ( the arrows indicate the current accuracy of the estimations ) . [
fig : rho4d3],width=264 ] similar qualitative behavior is found for the two - dimensional case ( figs .
[ fig : rho10x10 ] and [ fig : rho12x12 ] ) . furthermore , in fig .
[ fig : rho12x12 ] the spin - spin correlations have been plotted to show the location of @xmath1 and @xmath15 and their relation to the resistivity .
, vs. @xmath18 calculated on a 10@xmath6410 lattice with 26 spins ( @xmath117 ) , 8 electrons ( @xmath145 ) , and two values of @xmath0 as indicated .
shown is an average over three disorder configurations .
@xmath136 has units of @xmath146 in two dimensions.[fig : rho10x10],width=226 ] , spin - spin correlations , @xmath147 , at maximum and minimum distance , and inverse of the conductance , @xmath136 , vs. @xmath18 on a 12@xmath6412 lattice with 22 spins ( @xmath148 ) , 6 electrons ( @xmath145 ) , and @xmath0=@xmath114 .
the approximate vanishing of spin - correlations at different temperatures , depending on the distance @xmath21 studied , allowed us to obtain an approximate determination of @xmath1 and @xmath15 , similarly as carried out in previous studies.@xcite [ fig : rho12x12],width=245 ] it is interesting to compare our results with experiments . for ga@xmath3mn@xmath4as ,
data similar to those found in our investigations have been reported.@xcite a typical result for resistivity versus @xmath18 can be found in ref . .
the qualitative behavior of the resistivity in these samples agrees well with the theoretical results presented in fig .
[ fig : rho4d3 ] if intermediate couplings are considered . on the other hand ,
stronger or weaker couplings are not useful to reproduce the data , since the result is either insulating or metallic at all temperatures , respectively ( defining metallic and insulator regimes by the slope of the resistivity vs. temperature curves ) .
furthermore , even the experimental numerical values of the resistivity are in agreement with the results presented here ( even though model eq .
( [ eq : ham ] ) does not include a realistic treatment of the gaas host bands ) .
this can be shown as follows .
the conductivity of a three - dimensional sample is related to the conductance by @xmath149 , where @xmath41 is the side length of the lattice .
hence the resistivity is @xmath141 .
the units of @xmath129 are , as explained before , @xmath146 , and in our case @xmath150 , where @xmath151 is the lattice spacing .
assuming @xmath152 , then the values shown in fig .
[ fig : rho4d3 ] are obtained .
note that in fig .
[ fig : rho4d3 ] the minimum resistivity for @xmath0=@xmath153 is 3.3 m@xmath143 cm , whereas the minimum possible value for that carrier doping and cluster size used is 1.5 m@xmath143 cm , which corresponds to the case @xmath0=@xmath154 . in spite of the label ` metallic ' for these results , the absolute values of the resistivity are high .
similarly , in the metallic phase of the sample shown in the experimental results of ref . , the minimum resistivity is only @xmath155 m@xmath143 cm .
both in theory and experiments , the metallic phase appears to be ` dirty ' , which is likely due to the reduced number of carriers , and localizing effect of the disorder . for @xmath0 intermediate or strong , localization of the wave function
is observed at intermediate temperatures.@xcite this implies that carriers tunnel between impurity sites without visiting the main band , leading to insulating behavior .
below @xmath1 conduction is favored by the ferromagnetic order and , as a consequence , as temperature decreases conductance increases . to further understand why the conductance ( resistance ) has a minimum ( maximum ) around @xmath1 it is helpful to consider the three states : fm , clustered , and paramagnetic ( pm ) as depicted for a special spin configuration in fig .
[ fig : conduction ] on an 8@xmath648 lattice .
this configuration is not truly random but it was chosen ` by hand ' for simplicity so that the spin clusters can easily be recognized . in the case of a random configuration of spins a similar reasoning can be drawn . for the fm state of fig .
[ fig : conduction ] , conduction is possible and indeed the measured conductance is @xmath156 .
for the two possible clustered states represented in figs .
[ fig : conduction](b ) and ( c ) , the conduction is reduced due to the different alignment of spins in different clusters .
this is verified by calculating the conductance which is @xmath157 in those cases . for a paramagnetic spin arrangement with spin pointing in random directions (
not shown ) , the conductance is small but finite , @xmath158 .
it follows that the clustered state has the minimum conductance , and this explains the observed behavior of the resistivity with a maximum near @xmath1 , i.e. , when the system is clustered . 8 lattice and with spin configurations also selected ` by hand ' to illustrate our ideas .
for a ( a ) fm state at @xmath0=@xmath153 and @xmath8=@xmath159 , conduction is possible due to the alignment of magnetic moments and the conductance was found to be @xmath129=@xmath160 .
for the `` clustered '' state regime , two typical configurations ( b ) and ( c ) are shown where the conduction channels are broken and as a consequence @xmath161 .
for the same clustered state but with randomly selected spin orientations ( not shown ) the conductance is small but finite @xmath158 . the radius of the solid circles represent the local charge density .
[ fig : conduction],width=283 ] the inverse of the drude weight , @xmath162 , also shows a peak around @xmath1 as seen in fig . [
fig : drude ] , where the same mn - spin configuration as in fig .
[ fig : rho12x12 ] was used .
this provides further evidence for the metal - insulator transition near @xmath1 described here . for a 12@xmath6412 lattice and parameters as in fig .
[ fig : rho12x12 ] .
note that both the inverse conductance in fig .
[ fig : rho12x12 ] , and @xmath163 provide evidence for a metal - insulator transition at approximately the same temperature @xmath1.[fig : drude],width=207 ] it is also interesting to remark that the optimal value of @xmath0 changes with @xmath5 ( e.g. , for small @xmath5 , @xmath0 optimal is also small ) . since @xmath164 approximately separates the metal from the insulator , working at a fixed @xmath0 ( as in real materials ) and varying @xmath5 , then an insulator at high temperatures is found at small @xmath5 , turning into a metal at larger @xmath5 .
this is in excellent agreement with the experimental results of ref . using a careful annealing procedure ( while results of previous investigations with as grown samples had found an insulator - metal - insulator transition with increasing @xmath5.@xcite ) .
the issue discussed in this paragraph is visually illustrated in fig .
[ fig : jvsx ] , for the benefit of the reader .
the magnetoresistance percentage ratio was calculated using the definition @xmath165 .
figure [ fig : l12x12field ] shows the magnetization , @xmath166 , and the resistivity as a function of the applied field @xmath167 on a 12@xmath6412 lattice , @xmath5=@xmath168 , @xmath8=@xmath159 , @xmath0=@xmath114 , and low temperature .
note that in two dimensions at @xmath5=@xmath97 and @xmath8=@xmath98 , the intermediate coupling corresponds to @xmath0=2.5 , but at @xmath5=@xmath168 , now @xmath0=@xmath114 corresponds to optimal coupling since as @xmath5 decreases the optimal @xmath0 also decreases . for these parameters ,
@xmath1 was estimated to be 0.02@xmath14 and @xmath1690.06@xmath14 .
the value in teslas of the magnetic field was calculated assuming @xmath170=@xmath171 and @xmath172=@xmath173 for the localized spins .
the units of the resistivity are @xmath146 for the two - dimensional lattice .
in our simulations it was observed that @xmath166 increases with increasing magnetic field . at zero magnetic field @xmath166
is 60% of its maximum value , while at 12 t it has reached @xmath280% .
this is in agreement with the experimental results in fig . 3 of ref . where the magnetization is 50% of its maximum value at zero field , but it reaches 70% of that maximum value at 4 t. in our studies it was also observed that @xmath136 decreases with increasing magnetic field , showing at all fields a negative magnetoresistance . near @xmath1
the resistivity decreases by 20%-30% increasing the field from zero to 12 t , while the decrease in resistivity at higher temperatures is much smaller .
the present computational results agree very well with the experimentally measured dependence of the resistivity and magnetization on magnetic fields .
for example , in fig .
2(b ) of ref .
, the decrease in resistivity is 25% near @xmath174 k when increasing the field from zero to 15 t. a much smaller decrease in resistivity is observed at higher temperatures ( @xmath175 k ) .
it is also indicated in that experimental reference that the magnetoresistance is between 30% and 40% at 12 tesla . , and inverse conductance , @xmath136 , vs. magnetic field , @xmath167 , on a 12@xmath6412 lattice with pbc , @xmath5=@xmath168 , @xmath8=@xmath98 , @xmath13=@xmath114 , and @xmath176=@xmath177 ( same parameters as in fig . [
fig : rho12x12 ] ) .
the units of @xmath167 are tesla , assuming @xmath170=@xmath171 and @xmath172=@xmath173 .
the units of @xmath136 are @xmath178 .
the approximate values of the magnetoresistance are 18% at 4 tesla , and 30% at 12 tesla.[fig : l12x12field],width=226 ] it is interesting to remark that substantial magnetoresistance effects have been reported in thin films consisting of nanoscale mn@xmath179ge@xmath180 ferromagnetic clusters embedded in a dilute ferromagnetic semiconductor matrix.@xcite the characteristics of these materials are analogous to our clustered state , that also show a robust mr effect due to magnetization rotation of spontaneously formed clusters .
it was discussed before in this paper and in previous publications that a clustered state is formed above @xmath1 for intermediate and large @xmath0 couplings .
this state is a candidate to describe dms materials since it explains both the resistivity maximum around @xmath1 , as well as the decrease in resistivity with increasing applied magnetic field .
in addition , it provides an optimal @xmath1 .
this clustered state is formed only for intermediate or larger @xmath0 and when the compensation is strong , @xmath181 . since for large @xmath0
the carriers are localized , the problem becomes one of percolation theory , which has already been treated using different approaches and approximations.@xcite here , only a very simple way of visualizing this state will be presented .
first , consider a two - dimensional lattice with 5% mn spins represented as black dots , as shown in fig .
[ fig : percolation]a .
the carrier wave function , @xmath182 , is considered to be localized around a mn spin and it is assumed to be a step function for simplicity , i.e. @xmath182 is non - zero only if @xmath183 where @xmath184 is the localization radius introduced by hand
. sites where the wave function is not zero are the black areas of fig .
[ fig : percolation](b ) . in this case
, sites can be classified into connected regions , and that feature is indicated in fig .
[ fig : percolation](c ) using different shades .
each region is correlated and will correspond to a fm cluster . in this case
all spins belong to some large cluster .
as the concentration @xmath5 grows , the clusters will tend to merge .
as @xmath5 decreases , these clusters will become more and more isolated .
recent experimental work on ( ga , cr)as have revealed unusual magnetic properties which were associated with the random magnetism of the alloy .
the authors of ref . explained their results using a distributed magnetic polaron model , that resembles the clustered - state ideas discussed here and in refs . .
in this paper , dynamical and transport properties of a single - band model for diluted magnetic semiconductors have been presented .
the calculations were carried out on a lattice and using monte carlo techniques .
the optical conductivity , density of states , and resistance vs. temperature agree with experimental data for ga@xmath3mn@xmath4as if the model is in a regime of _ intermediate _ @xmath0 coupling . in this region , the carriers are neither totally localized at the mn sites nor free , as in previous theories .
the state of relevance has some characteristics of a clustered state , in the sense that upon cooling from high temperatures first small regions are locally magnetized at a temperature scale @xmath15 causing a mild insulating behavior while at an even lower temperature @xmath1 the alignment of the individual cluster moments occurs causing metallic behavior .
this is in agreement with previous investigations.@xcite one of the main conclusions of the paper is summarized in fig.[fig : pd ] where the ` transport ' phase diagram is sketched .
the low - temperature ferromagnetic phase is metallic ( although often with poor metallicity ) .
the clustered state between @xmath15 and @xmath1 has insulating properties , and the same occurs in a good portion of the phase diagram above these two characteristic temperatures .
note that here the terms metallic and insulating simply refer to the slope of the resistivity vs. temperature curves .
the values of the resistivities in the two regimes are not very different , similarly as observed experimentally .
this and related efforts lead to a possible picture of dms materials where the inhomogeneities play an important role . in this respect , these materials share characteristics with many other compounds such as manganites and cuprates , where current trends point toward the key importance of nanocluster formation to understand the colossal magnetoresistance and underdoped regions , respectively .
`` clustered '' states appear to form a new paradigm that is useful to understand the properties of many interesting materials .
the authors acknowledge the help of j. a. vergs in the study of the conductance . the subroutines used in this context were kindly provided by him .
conversations with a. fujimori are also gratefully acknowledged .
g. a. and e. d. are supported by the nsf grant dmr-0122523 .
additional funds have been provided by martech ( fsu ) .
part of the computer simulations have been carried out on the csit ibm sp3 system at the fsu school for computational science and information technology ( csit ) .
j. knig , j. schliemann , t. jungwirth , a.h .
macdonald ferromagnetism in ( iii , mn)v semiconductors in electronic structure and magnetism of complex materials , edited by d.j . singh and d.a .
papaconstantopoulos ( springer verlag 2002 ) ; and references therein .
j. sinova , t. jungwirth , s .-
eric yang , j. kucera , and a. h. macdonald , phys .
b * 66 * 041202 ( 2002 ) . s .-
eric yang , j. sinova , t. jungwirth , y.p .
shim , a.h .
macdonald , phys .
b * 67 * , 045205 ( 2003 ) .
j. sinova t. jungwirth , j. kucera , a.h .
macdonald , cond - mat/0301405 ( 2003 ) .
see also t. jungwirth , jrgen knig , jairo sinova , j. kucera , and a. h. macdonald , phys .
b * 66 * , 012402 ( 2002 ) .
john schliemann , jrgen knig , and a. h. macdonald , phys .
b * 64 * , 165201 ( 2001 ) .
malcolm p. kennett , mona berciu , and r. n. bhatt , phys .
b * 66 * , 045207 ( 2002 ) .
mona berciu and r. n. bhatt , phys .
* 87 * , 107203 ( 2001 ) .
see also : c. timm , f. schfer , and f. von oppen , phys .
* 90 * , 029701 ( 2003 ) and mona berciu and r. n. bhatt phys .
lett . * 90 * , 029702 ( 2003 ) .
k. c. ku , s. j. potashnik , r. f. wang , m. j. seong , e. johnston - halperin , r. c. meyers , s. h. chun , a. mascarenhas , a. c. gossard , d. d. awschalom , p. schiffer , n. samarth , appl .
. lett . * 82 * , 2302 ( 2003 ) . in ref .
, a hybrid mc study of the multi - band model for dms was presented . in that reference
the model was treated in k - space , approximating the fermionic trace for its zero - temperature value . on the contrary
, our present work treats the fermionic trace exactly , and calculations are performed in real space .
photoemission results for dms materials have also been presented in h. asklund , l. ilver , j. kanski , j. sadowski , and r. mathieu , phys .
rev . b*66 * , 115319 ( 2002 ) .
these experiments indicate that the fermi energy is slightly above , but very near to , the valence - band maximum .
since in our results the fermi energy is within the ` impurity band ' , this provides further evidence that the ` impurity band ' and the main band are not separated , but that they overlap in gamnas for the samples experimentally studied . | dynamical and transport properties of a simple single - band spin - fermion lattice model for ( iii , mn)v diluted magnetic semiconductors ( dms ) is here discussed using monte carlo simulations .
this effort is a continuation of previous work ( g. alvarez _ et al .
_ , phys .
. lett . * 89 * , 277202 ( 2002 ) ) where the static properties of the model were studied .
the present results support the view that the relevant regime of @xmath0 ( standard notation ) is that of intermediate coupling , where carriers are only partially trapped near mn spins , and locally ordered regions ( clusters ) are present above the curie temperature @xmath1 .
this conclusion is based on the calculation of the resistivity vs. temperature , that shows a soft metal to insulator transition near @xmath1 , as well on the analysis of the density - of - states and optical conductivity .
in addition , in the clustered regime a large magnetoresistance is observed in simulations .
formal analogies between dms and manganites are also discussed . |
the central exclusive production ( cep ) of new particles has received a great deal of attention in recent years ( see @xcite and references therein ) .
the process is defined as @xmath0 and all of the energy lost by the protons during the interaction ( a few percent ) is used in the production of the central system , @xmath1 .
the final state therefore consists of a centrally produced hard subprocess , two very forward protons and no other activity .
the @xmath2 sign denotes the regions devoid of activity , often called rapidity gaps . in double pomeron exchange ( dpe ) , the central system contains remnants from the diffractive exchange in addition to the hard subprocess .
more details on the diffraction physics program of the atlas experiment can be found e.g. in @xcite .
at low luminosity , the diffractive processes can be detected using rapidity gaps .
a possible l1 trigger would be based on a requirement of a rapidity gap on one or both sides from the ip and an activity in the central detector with energy over a certain threshold .
the gap may span the region from the forward calorimeters of the atlas detector @xcite or cms detector @xcite over the luminosity detectors ( lucid @xcite in atlas or totem @xcite in cms ) up to zdc detectors in atlas @xcite or in cms @xcite .
measurements which would be straightforward and hence suitable for analyses of the very early lhc data are ratios of the kind of x@xmath3gaps / x(incl . ) , where x may be w , z , dijet , heavy quark and dilepton , and x(incl . )
means measuring x without requiring rapidity gaps .
measurements of ratios are convenient since many sources of systematic uncertainties are canceled , particularly that of the luminosity at the early phase and among other , they also serve as valuable checks of different components of the formalism used to predict the cep cross section by the kmr group @xcite .
the soft survival probability , @xmath4 , can be studied in electroweak processes , such as w@xmath3gaps or z@xmath3 gaps .
@xmath4 is defined as a probability that additional soft secondaries will not populate the gaps and it explains the factorization breaking observed at hadron colliders when diffractive parton density functions ( dpdf ) obtained in single diffraction ( sd ) at hera were applied in measurements of sd by cdf @xcite .
the generalized gluon distribution , @xmath5 , can be probed in exclusive @xmath6 production proceeding via either a photon or an odderon exchange .
the higher - order qcd effects , especially sudakov - like factors and also a possible role of the enhanced absorptive corrections can be studied in exclusive two- or three - jet events .
when the proton tagging becomes available , the t - dependence of the elastic , sd and dpe cross sections can be obtained and hence effect of individual components of the pile - up background can be evaluated . without proton tagging , the dijet production in dpe and cep can be measured by requiring two central jets and rapidity gaps on both sides of the ip in forward calorimeter , lucid / totem and zdc . in dpe
, the rapidity gaps may be spoiled by particles from the pomeron remnants and although the cross section is about two orders of magnitude larger than the cep cross section at the same dijet mass , the cep cross section will dominate if the forward calorimeter is required to be devoid of activity .
the measurement of the dijet production in cep at 714 tev may be compared with a similar measurement made at tevatron from which models used to describe the data may be constrained . by selecting events with two jets
each in opposite side of forward calorimeter and a rapidity gap in the central detector , atlas and cms can improve an existing measurement of this type by d0 @xcite at centre - of - mass ( c.m.s . )
energy of 1.8 tev .
different colour singlet exchange models can be tested by comparing data with predictions for the gap fractions as functions of rapidity between the jets .
a great attention is recently devoted to the possibility of complementing the standard lhc physics menu by adding forward proton detectors ( fpd ) to the atlas and cms detectors .
they would detect a great part of the energy flow that escapes undetected by the main detectors .
the forward proton tagging will provide an exceptionally clean environment to search for new phenomena at the lhc and to identify their nature . of particular interest in this context
is the cep which gives access to the generalized ( or skewed ) pdfs .
the cep of a sm ( or mssm ) higgs boson is attractive for two reasons : firstly , if the outgoing protons remain intact and scatter through small angles then , to a good approximation , the central system @xmath1 must be produced in a @xmath7 , cp even state , therefore allowing a clean determination of the quantum numbers of any observed resonance .
here @xmath8 is the projection of the total angular momentum along the proton beam axis .
secondly , from precise measurements of proton momentum losses , @xmath9 and @xmath10 , the mass of the central system can be measured much more precisely than from the dijet mass measured in the calorimeters , by the so - called missing mass method , @xmath11 , which is independent of the decay mode .
the simplest decay mode from an experimental perspective is the ww decay mode , in which one ( or both ) of the w bosons decay leptonically . with standard single and double lepton trigger thresholds at atlas ( or cms ) , approximately 6 events are expected for higgs boson mass around 160 gev with luminosity of 30 fb@xmath12 @xcite . in the b@xmath13 decay mode , the quantum number selection rules in cep strongly suppress the qcd b - jet background , nevertheless severe requirements necessary to get rid of the pile - up background make the event yield rather modest @xcite .
full details of the calculation of the background to this channel are described in @xcite . in certain regions of the mssm parameter space the cross section for the cep of higgs bosons is significantly enhanced and hence making the b@xmath13 decay mode attractive @xcite . in fig .
[ mass_spectrum ] an example mass spectrum is shown for mssm higgs boson candidates of mass of 120 gev decaying into b@xmath13 for @xmath14 ( corresponding to the final cross section of about 18 fb ) after 3 years of data taking at luminosity of @xmath15 @xmath16s@xmath12 or 3 years at @xmath17 @xmath16s@xmath12 . at the low luminosity
, the pile - up background can be completely eliminated and the statistical significance is around 3.5@xmath18 . at the highest luminosity ,
fast timing detectors are necessary to reduce the pile - up background - significance of 5@xmath18 is achieved with time resolution of 2 ps ( see section [ timing ] ) .
b@xmath13 signal and its backgrounds for 3 years of data taking with atlas and the 420@xmath3420 detector configuration ( a ) at @xmath15@xmath16s@xmath12 ( 60 fb@xmath12 ) .
the significance of the fit is @xmath19 .
( b ) at @xmath17@xmath16s@xmath12 after removing the pile - up background completely .
the significance is @xmath20 .
figures from @xcite.,title="fig:",height=132 ] b@xmath13 signal and its backgrounds for 3 years of data taking with atlas and the 420@xmath3420 detector configuration ( a ) at @xmath15@xmath16s@xmath12 ( 60 fb@xmath12 ) .
the significance of the fit is @xmath19 .
( b ) at @xmath17@xmath16s@xmath12 after removing the pile - up background completely .
the significance is @xmath20 .
figures from @xcite.,title="fig:",height=132 ] + encouriging results have recently been obtained for the cep higgs bosons in the b@xmath13 and @xmath21 decay channels in the model which extends the sm by a fourth generation of heavy fermions @xcite .
nmssm provides the @xmath22 channel which in cep gives a very good signal - to - background ratio and is interesting from the experimental perspective @xcite . also , an excellent mass measurement of the neutral higgs boson decaying into b@xmath13 in the triplet model is possible in the forward proton mode @xcite .
the tagging of both protons in fpds will enable a measurement of the proton transverse momenta and azimuthal angles which allows us to study the opacity of the incoming protons , and more generally , to test the dynamics of the soft survival probability by studying the correlations between the outgoing protons @xcite .
this can be carried out with the cep of dijets as the cross section is large .
thanks to the @xmath23 selection rule which is applicable to all cep processes , quark jet production is suppressed and the cep can then be recognized as reduced ratio of b - jets to all jets when compared to other production processes .
the sd and dpe processes serve to provide an information about the low - x structure of the proton and the dpdfs .
inclusive jet and heavy quark production are mainly sensitive to the gluon component of the dpdfs , while vector boson production is sensitive to quarks .
the kinematic region covered expands that explored at hera and tevatron , with values of @xmath24 ( the fractional momentum of the struck parton in the diffractive exchange ) as low as @xmath25 and of q@xmath26 up to @xmath27 gev@xmath26 .
more information about the sd dijet and w boson production can be found elsewhere in these proceedings @xcite . as the lhc beams act also as a source of high - energy photons
a rich program of photon - photon and photon - proton physics can be pursued .
photon - induced processes have been extensively studied at lep and hera . however at lhc , these processes can be investigated in an unexplored region of the phase - space .
the final state topology is similar to cep , i.e. a central system , x , separated on each side by large rapidity gap from a very forward proton detected in the fpd .
different average proton transverse momenta make it possible to separate between diffractive and photon - induced events .
the w- and z - pair production ( as a tool to study anomalous triple and quartic gauge couplings ) is discussed elsewhere in these proceedings @xcite .
the susy particle production is described in @xcite .
two - photon exclusive production of muon pairs has a well known qed cross section , including very small hadronic corrections @xcite .
very recently , such event candidates have been observed by the cdf @xcite and their cross sections found in a good agreement with theory .
after applying simple selection criteria such as @xmath28 10 gev , @xmath29 2.5 and requiring one forward proton tag , the cross section is 1.3 pb @xcite .
this corresponds to approximately 50 muon pairs detected in a 12 hour run at a mean luminosity of @xmath30@xmath16s@xmath12 .
the large event rate coupled with a small theoretical uncertainty makes this process a perfect candidate for the absolute lhc luminosity calibration @xcite and also of the fpd system at 420 m @xcite .
the @xmath31 production can also be studied , although the trigger thresholds will be larger and hence the final event rate reduced .
the high luminosity and the high c.m.s .
energies available for photoproduction processes at the lhc allows us to study electroweak interactions and to search beyond the standard model up to the tev scale @xcite .
as shown in @xcite , the cross section for the associated wh production ( @xmath32 after applying selection criteria and considering five different final states is 0.17 fb at m@xmath33 = 115 gev and 0.29 fb at m@xmath33 = 170 gev .
the most promising channel seems to be the @xmath34 at m@xmath33 = 170 gev where the signal to irreducible background ratio is 0.22 fb/0.28 fb , so luminosity of 100 fb@xmath12 might reveal the @xmath35 gauge coupling .
photoproduction of single top quark ( @xmath36 ) is dominated by t - channel amplitudes in association with a w boson which all are proportional to the ckm matrix element @xmath37 .
the ratio of associated wt production cross section to the sum of all top production cross sections is 5% for parton - parton interactions , while it is 50% in photoproduction . in @xcite
two topologies were studied , namely @xmath38 and @xmath39 . the signal cross section after selection cuts of about 44 fb with a signal to irreducible background ratio of 0.6 suggest that this mechanism and hence @xmath37 may be easily measurable even with luminosity of 1 fb@xmath12 . at lhc
the exclusive single top quark photoproduction can only occur via flavour changing neutral current processes which are not present at tree level of sm but appear in many extensions of sm such as two higgs - doublet models or r - parity violating supersymmetry .
the final state of this @xmath40 process is composed of a b - jet and a w boson . in @xcite
the leptonic @xmath41 topology was studied and only photoproduction @xmath42 background considered . with an integrated luminosity of 1
fb@xmath12 the expected limits for anomalous couplings @xmath43 and @xmath44 at 95% cl are greatly improved with respect to existing best estimates . in photoproduction of jets ,
the fraction of the photon , @xmath45 , and proton , @xmath46 , four - momentum carried by a parton involved in binary hard scattering is calculated using the energies and angles of jets in the central detector and of the protons in the fpds .
the direct photon processes are characterized by @xmath47 and resolved photon processes by @xmath48 .
the h1 and zeus collaborations have constrained the region @xmath49 for diffractive photoproduction . at lhc
we expect to reach values of @xmath46 and @xmath45 of an order of magnitude lower than at hera .
furthermore , the diffractive photoproduction of dijet systems at the lhc promises to shed light on the issue of the qcd factorization breaking recently reported in the same process by the h1 experiment @xcite .
exclusive @xmath6 photoproduction , @xmath50 can be studied using fpds @xcite , although only one proton can be tagged due to the low mass of the @xmath6 .
the cross section is expected to be approximately 1.25 pb for the decay channel @xmath51 and is sensitive to the same skewed unintegrated gluon densities of the proton as the cep of higgs boson . measuring
this process thus helps to constrain the @xmath5 as the soft survival factor is expected to be close to 1 .
the @xmath50 process can also occur via odderon exchange and this channel could be the first evidence for the odderon s existence .
the fp420 r&d collaboration @xcite , with members from atlas , cms and lhc studied the possibility of installing high precision tracking and timing detectors at 420 m from the ip .
detection of the protons will be achieved by two 3d silicon detector stations at each end of the fp420 region .
this novel technique provides high radiation - resistive detectors close to the beam with an insensitive area as small as 5 @xmath52 m and with a resolution of about 15 @xmath52 m .
the tracking and timing detectors will be attached to a movable beam pipe .
as the beam pipes in the 420 m region are contained in an interconnecting cryostat and the sensitive detectors are best operated at room temperature , a new connection cryostat has been designed using a modified arc termination module at each end .
the afp project ( atlas forward proton ) @xcite is a combined effort to install fpds at 220 and 420 m from the ip of atlas .
the 420 m region has been described above and the proposed equipment for the 220 m region is very similar ; differences are mainly in no need to change the cryostat and an addition of a detector to be used for l1 trigger .
the detectors will be read out by standard abcnext chip being developed for the silicon detectors in atlas . with the position resolution of 15 @xmath52 m
we expect a mass resolution of the order of 12% for the 420 + 420 and about 3% for the 420 + 220 configurations over a mass range of 120200 gev . for given dipole apertures and collimator settings and a thin window of 200 @xmath52 m ,
the expected @xmath53 range is 0.0020.02 for 420 + 420 and 0.010.15 for the 420 + 220 configuration .
the low-@xmath53 ( and therefore low mass ) acceptance depends critically on the distance of approach of the active area of the sensitive detectors from the beam .
the final distance of approach will depend on the beam conditions , machine - induced backgrounds and collimator positions , and the rf impact of the detector on the beams . at 420
m the nominal operating position is assumed to be between 5 and 7.5 mm , at 220 m it is between 2.0 and 2.5 mm .
for masses above about 120 gev , the 220 m detector adds to the acceptance with increasing importance as the central mass increases .
the differences between atlas and cms acceptances for the 420 + 420 as well as 420 + 220 configurations ( see fig .
[ mh - acc ] ) are due to a different crossing angle which is in the vertical plane for ip1 ( atlas ) and horizontal plane at ip5 ( cms ) .
the most prominent background to many diffractive physics analyses comes from an overlap of two soft sd events from pile - up and one nd event produced at a hard scale .
fast timing detectors with an expected sub-10 ps time resolution corresponding to a vertex resolution of better than 2.1 mm should be able to assign a vertex to the proton detected in the fpd and to reject about 97% of cases that appear to be cep events but where the protons in reality originated from coincidences with pile - up events .
presently two detector options are studied , namely quartz and gas cerenkov which may be read out with a constant fraction discriminator allowing the time resolution to be significantly improved compared to usual electronics .
due to a limited l1 trigger latency , detectors at 420 m are far away from the central detectors to be included in the l1 trigger in normal running conditions , while detectors at 220 m can in principle be included .
the trigger strategy depends on the mass of the diffractively produced object @xcite .
demanding standard l1 triggers such as those for high mass h@xmath54ww / zz or high-@xmath55 dijet trigger would result in an acceptable output rate which may be further reduced by requiring the double proton tag at 220 m. triggering on low mass objects is more difficult but in principle feasible as documented in @xcite where diffractive l1 triggers for a case of h @xmath54 b@xmath13 at @xmath56=120 gev have been proposed .
if the fpd trigger at 220 m is capable of triggering only on hits in the inner 4 mm part and if the l1 calorimeter is capable of defining exclusivity criteria using @xmath57 and @xmath1 , then the final output rate is well below a 2 khz limit at @xmath58@xmath16s@xmath12 and slightly above this limit at @xmath59@xmath16s@xmath12 .
other reductions are under study @xcite .
supported by the project av0-z10100502 of the academy of sciences of the czech republic and project lc527 of the ministry of education of the czech republic . 99 fp420 r&d collab . , j. inst . : 2009-jinst-4-t10001 , arxiv:0806.0302 [ hep - ex ] .
m. taevsk , nucl .
* 179 - 180 * ( 2008 ) 187 ; atl - phys - conf-2008 - 019 .
atlas collab . ,
0803 : s08003 ( 2008 ) .
cms collab . ,
0803 : s08004 ( 2008 ) .
atlas collab .
, cern - lhcc-2004 - 010 ( 2004 ) ; a. zoccoli , these proceedings .
totem collab . , cern - lhcc-2004 - 002 ( 2004 ) ;
s. giani , these proceedings . atlas collab . , cern - lhcc-2007 - 001 ( 2007 ) .
v. a. khoze , a. d. martin and m. g. ryskin , eur .
j. * c 23 * ( 2002 ) 311 . v. a. khoze , a. d. martin and m. g. ryskin , arxiv:0802.0177 [ hep - ph ] .
cdf collab .
* 84 * ( 2000 ) 5043 .
d0 collab .
, phys . lett . *
b440 * ( 1998 ) 189 . b. e. cox _ et al . _ ,
j. * c45 * ( 2006 ) 401 .
m. albrow _
_ , cern - lhcc-2006 - 039 ( 2006 ) .
a. brandt , v. jurnek , a. pal and m. taevsk , atlas note in preparation .
et al . _ , eur.phys.j.*c25 * ( 2002 ) 392 ; a.g .
et al . _ , eur.phys.j.*c56 * ( 2008 ) 467
. s. heinemeyer _
_ , eur.phys.j.*c53 * ( 2008 ) 231 ; arxiv:0801.1974 [ hep - ph ] ; arxiv:0811.4571 [ hep - ph ] .
cox , f.k .
loebinger , a.d .
pilkington , jhep 0710:090 ( 2007 ) .
p. h. frampton , p. q. hung and m. sher , phys.rept . * 330 * ( 2000 ) 263 , arxiv : hep - ph/9903387 .
s. heinemeyer _
et al . _ , these proceedings , arxiv:0909.4665 [ hep - ph ] .
j. forshaw _
et al . _ , jhep 0804:090 ( 2008 ) , arxiv:0712.3510 [ hep - ph ]
. m. chaichian _
_ , arxiv:0901.3746 .
cdf collab . , arxiv:0712.0604 [ hep - ex ] .
v. jurnek , these proceedings .
w. carvalho , these proceedings .
o. kepka , these proceedings .
k. piotrzkowski and n. schul , arxiv:0806.1097 [ hep - ph ] ; k. piotrzkowski , these proceedings .
v. a. khoze , a. d. martin , r. orava and m. g. ryskin , eur .
j. * c 19 * ( 2001 ) 313 .
cdf collab .
* 98 * ( 2007 ) 112001 .
m. albrow , these proceedings .
j. favereau de jeneret _ et al .
_ , cp3 - 09 - 37 ( august 2009 ) , arxiv:0908.2020 [ hep - ph ] . k. piotrzkowski , phys . rev .
* d63 * ( 2001 ) 071502 . | diffractive physics program for atlas and cms is discussed with emphasis on measurements of central exclusive processes . at low luminosities ,
a l1 trigger based on requiring rapidity gaps can be used , while at high luminosities , the use of proton taggers proposed to be placed at 220 m and 420 m from the interaction point is foreseen . |
supersonic outflows from young stellar objects ( ysos ) are often seen at optical wavelengths as herbig - haro ( hh ) objects .
hh objects are produced by shocks from the collision of the outflow with the external medium or with previously ejected outflow components , producing collisionally excited forbidden lines . in the optical ,
they can be readily identified by their h@xmath0 and [ s ii ] line emission , non - stellar appearance , and nearly collinear alignment .
see , e.g. , @xcite for a more detailed discussion . despite their distinctive characteristics
, hh objects can be difficult to find in the dense , complex environment in which ysos are forming . in this paper
we report on the search for hh objects in the regions of and ( b175 ) .
is a reflection nebula illuminated by the herbig be star .
it is embedded in a giant molecular cloud complex known as the cepheus flare " shell @xcite and is on the northern edge of the @xmath1 complex . the measured distance
to hd 200775 via _ hipparcos _ parallax is @xmath2 pc @xcite , which we adopt here ; although @xcite and references therein argue for other distances .
no prior discoveries of hh objects in ngc 7023 have been reported .
b175 , which is also embedded in the cepheus flare , is a cometary - shaped bok globule .
it is illuminated on the southern edge by the b9.5v type star , producing the reflection nebula .
the distance to b175 is estimated to be at a distance of about 400pc ; see @xcite for discussion .
@xcite also report the discovery of embedded in ced 201 .
it is the only known hh object within b175 .
ngc 7023 and b175 were observed with the mosaic camera on the mayall 4-meter telescope at kitt peak national observatory .
mosaic is an optical camera that consists of eight 2048@xmath34096 ccd detectors arranged to form a 8192@xmath38192 array with 35 to 50-pixel - wide gaps between the ccds . with a scale of @xmath4 pixel@xmath5 ,
the field of view is approximately @xmath6 . to fill in the gaps and bad columns ,
all observations are completed in a five - exposure dither pattern with offsets of approximately 100 pixels .
observations of ngc 7023 were obtained on 29 august 2009 with the harris @xmath7 ( mosaic filter k1002 ) , harris @xmath8 ( k1003 ) , nearly - mould " @xmath9 ( k1005 ) and h@xmath0 ( k1009 ) filters .
five exposures each of 600 sec in @xmath7 , 420 sec in @xmath8 , 300 sec in @xmath9 and 600 sec in h@xmath0 were obtained .
excellent seeing of @xmath10 in h@xmath0 was achieved .
after the outflows were first detected in ngc 7023 , follow - up observations were completed in [ s ii ] ( ha16 , h - alpha+16 nm , k1013 ) on 21 june 2012 .
five exposures of 600 sec in [ s ii ] were obtained .
unfortunately the seeing was @xmath11 ; but it was sufficient for confirmation purposes .
observations of b175 were obtained on 23 october 2011 with the @xmath7 ( k1002 ) , @xmath8 ( k1003 ) , @xmath9 ( k1005 ) and h@xmath0 ( k1009 ) filters .
five exposures each of 480 sec in @xmath7 , 300 sec in @xmath8 , 180 sec in @xmath9 and 600 sec in h@xmath0 were obtained .
seeing of 11 in h@xmath0 was achieved . at the time of the august 2009 observations mosaic
was equipped with thinned , science - grade site ccd cameras . in the summer of 2012 mosaic
was upgraded to eight thinned , science - grade e2v ccds .
the observations of b175 and follow - up [ s ii ] observations of ngc 7023 were therefore obtained after the upgrade .
both ccd types have 15 pixels ; thus the pixel scales of the ccds are nearly identical .
and the dqes of the two cameras are very similar at the wavelength range of interest ( at 6500 , an average of 86% for the site chips and 80% for the e2v chips ) .
the bandpasses of the h@xmath0 ( k1009 ) and [ s ii ] ( k1013 ) filters are @xmath12 , with transmission efficiencies of @xmath13 at @xmath14 and @xmath15 respectively .
thus they should be roughly equivalent in sensitivity to h@xmath0 and [ s ii ] emission .
the data were reduced with the iraf package mscred in the standard manner .
ten bias frames and five dome flats in each filter were used .
the world coordinate system ( wcs ) was determined via stars from the usno - b1.0 catalog @xcite with a global solution rms of better than @xmath16 in all cases .
this is assumed as the accuracy for all measured positions .
all of the images were projected to a common wcs to remove geometric distortions and to allow for the images to be aligned .
the images in each filter were then combined to fill in the gaps .
as a means of visualizing the structure , the data for each object were combined to form a color - composite image with the methodology described in @xcite .
an advantage of searching for hh objects in a color - composite image is that the h@xmath0 emission is distinct and readily visible if assigned a unique color .
further , the broadband filters reveal the amount of obscuration from dust and gas .
thus , faint outflows can be found more easily in the complex environments typical of star - forming regions .
the results for each object are discussed below .
a region 365 @xmath3 363 in size , centered on the location @xmath17 was searched for new hh objects .
at least two distinct outflows were discovered in the northwestern lobe " of ngc 7023 .
four new hh objects are labeled in figure [ fig-1 ] .
all of the hh objects are visible in the h@xmath0 and [ s ii ] filters , but are not detected in the broadband filters .
thus we are confident they are sources of line emission only .
the objects were named in order of increasing right ascension . when objects are clearly part of the same outflow ( e.g. , hh 1067 ) they share a single hh number .
when it is uncertain if individual knots are part of the same flow , they are given separate hh numbers .
their coordinates are given in table [ tbl-1 ] .
the given positions correspond to the brightest knot in each object .
figures [ fig-2 ] and [ fig-3 ] show the h@xmath0 and [ s ii ] images for each object .
no yso progenitors are seen in the optical .
however , @xcite completed a survey of the cepheus flare for yso candidates with the _ spitzer _ irac and mips cameras . in the western lobe of ngc 7023 they identified
tight group b " in the region of our newly discovered outflows .
the group consists of four ysos with an associated bubble structure detected at 3.6 and 4.5 .
based upon their infrared spectral indices , @xcite classify all four of the ysos in tight group b as embedded
class i " protostars .
the irac archival data ( aor 25037312 ) of this region are shown in figure [ fig-4 ] . the hh1067 outflow between components
a and b , as well as hh1069 and hh1070 , are visible at 3.6 and 4.5 .
this is consistent with deeply embedded sources , as emission at these wavelengths is enhanced in molecular shocks at relatively high temperatures or densities as compared to 5.8 and 8 emission @xcite .
hh1068 is not in the field of view of the irac observations .
hh1067 appears to be a bipolar outflow with an origin in tight group b , whose position is marked in figure [ fig-1 ] . in both the optical and infrared , hh1067a and hh1067b are clearly connected by a faintly visible and gently curved segmented line that runs through the group and points towards hh1067c .
hh1067c may be a termination shock , as its structure is nearly perpendicular to the structure of hh1067b and no hh objects are seen to the west of hh1067c .
it is not clear which , if any , of the four ysos detected in the group is the progenitor for this outflow .
hh1068 , hh1069 and hh1070 are also roughly collinear with tight group b. the structure of hh1070 strongly suggests a bow shock with an origin in the direction of this group .
it is also not clear which yso is the progenitor of this outflow .
however it is worth noting that hh1069 and hh1070 are equidistant and collinear with yso # 100 as identified in @xcite , suggesting that both hh objects were produced by that object as part of the same ejection event .
if hh1068 is also part of this outflow , the angular distance between yso candidate # 100 and hh1068 is 357 . assuming a distance to ngc 7023 of 430pc , this is a projected distance of 0.75pc , which is a notable length for an embedded outflow . assuming a transverse motion of 100 km s@xmath5 of the outflow implies a dynamical age of about 7500 yr .
if hh1068 , hh1069 and hh1070 are components of a bipolar outflow , then the yso is likely precessing in a manner similar to @xcite .
the hh1068-hh1070 outflow appears to be embedded within or behind the reflection nebulosity of ngc 7023 , as both hh1069 and hh1070 are visible in cavities . hh1068 and the western half of hh1070 are fainter and appear to be partially obscured by dust and gas visible in the broadband filters .
the overall structure observed in ngc 7023 is indicative of two distinct outflows with a common origin .
it is similar in appearance to the and cloud core @xcite .
since they are seen as nearly perpendicular , it is hard to envision a scenario where hh1067 is produced by the same progenitor as hh1068 - 70 , although it could be a binary system . because hh1067 is visible along most of the outflow
, it suggests that it is on the near edge of the ngc 7023 nebula and therefore is likely closer than the hh1068-hh1070 outflow .
a region 406 @xmath3 363 in size , centered on the location @xmath18 was searched for new hh objects .
the field of b175 is particularly interesting because it contains the outflow as well as the supernova remnant , both discovered by @xcite .
no new hh objects were found in b175 .
however , we do reclassify a small filament located between hh 450 and two large filaments of snr g110.3 + 11.3 .
@xcite originally classified this filament as hh 450x and interpret it as evidence of interaction between the hh outflow and the supernova remnant .
unlike hh 450 and snr g110.3 + 11.3 , which are detected only in h@xmath0 , we detect hh 450x in all of the broadband and narrowband filters , indicating it is a continuum source . based upon its colors and morphology ( figure [ fig-5 ] )
we reclassify hh 450x as a background galaxy . the discovery that hh 450x is not a shock front eliminates a key piece of evidence that there is a physical connection between b175 and snr g110.3 + 11.3 and makes their relationship less clear .
we are grateful to b. reipurth and j. bally for their helpful discussions .
we also wish to thank the anonymous referees as well as kitt peak national observatory and its excellent support staff .
the figures in this paper were created with the help of the esa / eso / nasa fits liberator .
bally , j. & reipurth , b. 2001a , 546 , 299 bally , j. & reipurth , b. 2001b , 552 , l159 gomez , m. , kenyon , s.j . , & whitney , b.a .
1997 , 114 , 265 hubble , e.p .
1934 , 79 , 8 kirk , j.m . ,
ward - thompson , d. , di francesco , j. , bourke , t.l . , evans , ii , n.j . , mern , b. , allen , l.e . ,
cieza , l.a . ,
dunham , m.m . , harvey , p. , huard , t. , jrgensen , j.k . ,
miller , j.f .
, noriega - crespo , a. , peterson , d. , ray , t.p . & rebull , l.m .
185 , 198 kun , m. kiss , z. t. & balog , z. 2008 ,
star forming regions in cepheus " , handbook of star forming regions , volume i : the northern sky , asp monograph publications , vol .
4 . edited by bo reipurth , 136 monet , d.g .
2003 , 125 , 984 rector , t.a . ; levay , z.g . , frattare , l.m . ,
english , j. & puuohau - pummill , k. 2007 , 133 , 598 reipurth , b. , yu , k.c .
, moriarty - schieven , g. , bally , j. , aspin , c. & heathcote , s. 2004 , 127 , 1069 takami , m. , karr , j.l . ,
koh , h. , chen , h .- h .
& lee , h .-
2010 , 720 , 155 van den ancker , m.e .
, the , p.s .
, tjin a djie , h.r.e . ,
catala , c. , de winter , d. , blondel , p.f.c . & waters , l.b.f.m .
1997 a&a , 324 , l33 lll hh1067a & 21:00:24.5 & + 68:13:19 + hh1067b & 20:59:59.0 & + 68:13:04 + hh1067c & 20:59:50.4 & + 68:13:21 + hh1068a & 21:00:17.6 & + 68:18:58 + hh1068b & 21:00:14.8 & + 68:19:12 + hh1069 & 21:00:17.9 & + 68:10:13 + hh1070 & 21:00:23.7 & + 68:16:04 + | wide - field optical imaging was obtained of the cluster and reflection nebula and the bok globule .
we report the discovery of four new herbig - haro ( hh ) objects in ngc 7023 , the first hh objects to be found in this region .
they were first detected by their h@xmath0 and [ s ii ] emission but are also visible at 3.6 and 4.5 in archival _ spitzer _
observations of this field .
these hh objects are part of at least two distinct outflows .
both outflows are aligned with embedded
class i " ysos in a tight group on the western edge of the nebula
. one of the outflows may have a projected distance of 0.75pc , which is a notable length for an embedded source .
no new hh objects were discovered in b175 .
however , we reclassify the knot , in b175 , as a background galaxy .
the discovery that hh 450x is not a shock front weakens the argument that hh 450 and snr g110.3 + 11.3 are co - located and interacting . |
in 1988 one of us found a preprint by vaughan jones , `` on a certain value of the kauffman polynomial '' @xcite .
we immediately saw that the metaplectic representation for @xmath2 therein had to be related to our chiral potts model work @xcite and it soon became clear that it was even related to the 5-state fateev
zamolodchikov model @xcite .
after reporting this to vaughan we received an invitation to the workshop on integrable systems in statistical mechanics , quantum field theory , and knot theory , at msri , uc berkeley , january 1989 .
there we both had detailed further discussions on the relationships between knot theory and integrable models of statistical mechanics and some of that got incorporated in the paper `` on knot invariants related to some statistical mechanical models '' @xcite .
one evening during the workshop , vaughan invited us both to his apartment , but helen could not go as our baby was not well . at one point vaughan
came up with five copies of hndel s messiah for five of us present .
he then assigned the four voice parts , `` jacques , do you want to sing tenor or bass?''and anthony wasserman got the piano accompaniment .
thus we performed the entire piece from cover to cover .
it was a deeply spiritual experience with the entire text taken from the king james version of the bible .
we have met vaughan later at several other meetings and have played several games of snooker together . once during a conference honoring baxter in canberra we had a bus trip to tidbinbilla .
vaughan said , helen , you have to sit somewhere else ; i have to sit next to jacques . " and we sang several parts of the messiah together . even during the great wall expedition of the qinhuangdao conference we sang ` comfort ye ' and
` every valley ' in the bus after helen fell asleep and her earplug came loose from her ipad .
associated with the crossing of oriented rapidity lines with rapidities p and q. spin variables live on line pieces , faces and vertices .
faces are colored alternatingly black and white in a checkerboard pattern .
[ fig1 ] the yang
baxter equation @xcite is a generalization of artin s braid equation in knot theory with spectral variables called rapidities @xmath3 , living on oriented lines , see figure [ fig1 ] .
to each crossing of the rapidity lines one assigns boltzmann weights @xmath4 or @xmath5 depending on a black - white checkerboard coloring assigned to the faces .
most generally , one can have spin variables on all faces , vertices and line pieces , while each spin variable can be chosen independently from a finite or infinite set .
one can sum over the spin @xmath6 at the vertex , obtaining the irf - vertex model with new boltzmann weights @xmath7 , @xmath8 .
such a reduction can always be applied and has been useful in certain spin models , see e.g. @xcite . assuming that there are no spins at the intersections of rapidity lines
, one can define several special cases @xcite .
first , if the spins on the line pieces ( @xmath9 ) only take a single value , one can omit them , arriving at the checkerboard interaction - round - a - face ( irf ) model with weights @xmath10 , @xmath11 .
second , assuming all spins on the faces have a single value ( and thus can be omitted ) one receives the checkerboard vertex model with weights @xmath12 , @xmath13 . forgetting about the coloring of the faces one gets the usual vertex model with @xmath14 .
finally , if one only leaves spins on the black ( or white ) faces , one ends up with a spin model . the most general yang
baxter equation is depicted in figure [ fig2 ] and can be expressed more complicatedly in formula as & & _
i_j_k____d w^(i)^_|^ad_cb(p , q ) [ eq1 ] + & & w^(j)^_|^ab_dc(q , r ) w^k)^_|^dc_ba(p , r ) + & & = r(p , q , r ) _ i_j_k____d w^(i)^_|^bc_da(p , q ) + & & w^(j)^_|^cd_ba(q , r ) w^(k)^_|^ab_cd(p , r ) , + & & r(p , q , r ) _
i_j_k____d w^(i)^_|^ad_cb(p , q ) [ eq2 ] + & & w^(j)^_|^ab_dc(q , r ) w^(k)^_|^dc_ba(p , r ) + & & = _ i_j_k____d w^(i)^_|^bc_da(p , q ) & & w^(j)^_|^cd_ba(q , r ) w^(k)^_|^ab_cd(p ,
one can read more about this and the various formal equivalences and specializations in @xcite .
as said before , the sums over @xmath6 , @xmath15 and @xmath16 can be taken introducing new sum weights and one can specialize further to get the various special versions of the yang
baxter equation in the literature . finally , in some examples that we have studied the scalar factors @xmath17 and @xmath18 can be factorized and absorbed into the @xmath19 s and @xmath20 s by properly redefining these . also ( [ eq1 ] ) and ( [ eq2 ] ) are the same equation in the integrable chiral potts model , for example .
let us from now on specialize to the integrable chiral potts model @xcite .
potts means here that there is a translation invariance in the spin variables , meaning that each weight is of the form @xmath21 or @xmath22 , depending only on the difference of two spin variables @xmath23 and @xmath24 modulo an integer @xmath25 and integrability implies the existence of the two rapidities @xmath26 and @xmath0 .
the chiral property means that there is no reflection invariance , i.e. @xmath27 , @xmath28 in general .
the two weights are depicted in figure [ fig3 ] and have the form ( 3.6,0 ) ( 1.2,0 ) node[left ] @xmath0 ; ( 0,0 ) ( -2.4,0 ) node[left ] @xmath0 ; ( 2.4,-1.2 ) ( 2.4,1.2 ) node[left ] @xmath26 ; ( -1.2,-1.2 ) ( -1.2,1.2 ) node[right ] @xmath26 ; at ( -1.2,0.75 ) @xmath29 ; at ( 2.5,0.75 ) @xmath30 ; ( 1.8,0.6 ) circle ( 2pt);at ( 1.75,0.6 ) @xmath24 ; ( 3,-0.6 ) circle ( 2pt);at ( 3,-0.6 ) @xmath23 ; ( -1.8,-0.6 ) circle ( 2pt);at ( -1.85,-0.6 ) @xmath23 ; ( -0.6,0.6 ) circle ( 2pt);at ( -0.6,0.6 ) @xmath24 ; ( -1.75,-0.55 ) ( -0.65,0.55 ) ; ( 2.95,-0.55 ) ( 1.85,0.55 ) ; & & w_pq(n)=w_pq(0)^n_j=1 ( ) , [ bweights ] + & & _
pq(n)=w_pq(0)^n_j=1(_p_q),where @xmath31 , @xmath32 is an @xmath25th root of unity , and the two rapidities @xmath33 and @xmath34 lie on the high - genus curve x_p^n+y_p^n = k(1+x_p^ny_p^n),_p^n== , with some @xmath16 and @xmath35 satisfying @xmath36 . here
we shall choose the normalization @xmath37 .
these weights satisfy the reflection relation w_pp(a - b)=1,w_pq(a - b)w_qp(a - b)=1 , and inversion relation _
b=0^n-1w_pq(a - b)w_qp(b - c)=r_pq_a , c .
these relations correspond to reidemeister moves i and ii of knot theory . indeed , ( 0,-3 ) ( 0,0 ) .. controls + ( up:1 mm ) and + ( left:5 mm ) .. ( 0.5,0.5 ) ; ( 0.5,0.5 ) .. controls + + ( right:1 mm ) and + + ( up:5mm) .. (1,0 ) .. controls + + ( down:5 mm ) and + + ( down:5 mm ) .. ( 0,0) ( 0,3 ) node[above ] @xmath26 ; at ( 2,0.5 ) @xmath38 ; ( 4,-3 ) ( 4,3 ) node[above ] @xmath26 ; ( 0.4,0 ) circle ( 2pt ) ; at ( 0.27,0.1 ) @xmath23 ; ( -0.6,0 ) circle ( 2pt ) ; at ( -0.5,0.1 ) @xmath24 ; ( -0.6,0 ) ( 0.4,0 ) ; at ( 2,0.5 ) @xmath38 ; ( 3.6,0 ) circle ( 2pt ) ; at ( 3.6,0 ) @xmath24 ; ( 4.4,0 ) circle ( 2pt ) ; at ( 4.4,0 ) @xmath39 ; ( 1,1.2 ) .. controls + + ( left:4 mm ) and + + ( up:5 mm ) .. ( 0,0 ) .. controls+(down:5 mm ) and + ( left:4 mm ) .. ( 1,-1.2 ) ; ( 1,1.2 ) circle ( 2pt ) ; at ( 1,1.2 ) @xmath39 ; ( 1,-1.2 ) circle ( 2pt);at ( 1,-1.2 ) @xmath40 ; ( 3.5,-3 ) .. controls ( 2,0 ) .. ( 3.5,3 ) node[right ] @xmath26 ; ( 2,-3 ) .. controls ( 3.5,0) .. (2,3 ) node[left ] at ( 2,3 ) @xmath0 ; ( 1.9,0 ) circle ( 2pt ) ; at ( 1.78,0 ) @xmath24 ; ( 2.75,0 ) circle ( 2pt ) ; at ( 2.8,0 ) @xmath41 ; ( 3.6,0 ) circle ( 2pt ) ; at ( 3.72,0 ) @xmath39 ; ( 2.75,-2.6 ) circle ( 2pt ) ; at ( 2.8,-2.6 ) @xmath23 ; ( 2.75,2.6 ) circle ( 2pt ) ; at ( 2.8,2.6 ) @xmath42 ; ( 1.9,0) .. controls + + ( up:4 mm ) and + + ( left:5 mm ) .. ( 2.75,1.45 ) .. controls + + ( right:4 mm ) and + + ( up:5mm) .. (3.6,0 ) ; ( 1.9,0) .. controls + + ( down:4 mm ) and + + ( left:5 mm ) .. ( 2.75,-1.45 ) .. controls + + ( right:4 mm ) and + + ( down:5mm) .. (3.6,0 ) ; ( 2.75,-2.6 ) ( 2.75,0 ) ; ( 2.75,0 ) ( 2.75,2.6 ) ; at ( 4.5,0.5 ) @xmath38 ; ( 6,-3 ) ( 6,3 ) ; ( 7,-3 ) ( 7,3 ) ; ( 5.5,0 ) circle ( 2pt ) ; at ( 5.5,0 ) @xmath24 ; ( 6.5,0 ) circle ( 2pt ) ; at ( 6.5,0 ) @xmath23 ; ( 7.5,0 ) circle ( 2pt ) ; at ( 7.5,0 ) @xmath39 ; & & w_pp(n)=()^n_j=1^n , , + & & w_pp(n)=(_p_p)^n_j=1^n , . corresponds to reidemeister move i and w_pq(n)=()^n_j=1^n = w^-1_qp(n ) , + , + _
d=0^n-1w_pq(a - d)w_qp(d - a)=r_pq_a , a , with some factor @xmath43 , corresponds to reidemeister move ii .
these are depicted in figure [ fig4 ] .
finally there is the star - triangle relation ( yang baxter relation for the chiral potts model ) , see also figure [ fig5 ] , & & ^n_d=1w_pr(a - d)w_pq(d - c)w_rq(d - b ) [ startriangle ] + & & = r_pqrw_pq(a - b ) w_pr(b - c)w_rq(a - c),corresponding to reidemeister move iii .
( 3,5 ) node[left ] @xmath26 ; ( 5,5 ) (-2,-2 ) node[left ] @xmath0 ; ( 5,-5 )
( -2,2 ) node[left ] @xmath44 ; ( -1.5,0 ) circle ( 3pt ) ; at ( -1.5,0 ) @xmath24 ; ( 1.8,0 ) circle ( 3pt ) ; at ( 1.8,0 ) @xmath41 ; ( 3.6,4.5 ) circle ( 3pt ) ; at ( 3.6,4.5 ) @xmath39 ; ( 3.6,-4.5 ) circle ( 3pt ) ; at ( 3.8,-4.5 ) @xmath23 ; ( 3.6,-4.5 )
( 2.7,-2.25);(2.7,-2.25 )
( 1.8,0 ) ; ( 1.8,0 )
( 2.7,2.25 ) ; ( 2.7,2.25 )
( 3.6,4.5 ) ; ( 1.8,0 ) ( -1.0,0);(-1.0,0 ) ( -1.5,0 ) ; at ( 7,0.5 ) @xmath38 ; ( 12,-5 )
( 12,5 ) node[right ] @xmath26 ; ( 16,2 ) (9,-5 ) node[left ] @xmath0 ; ( 16,-2 )
( 9,5 ) node[left ] @xmath44 ; ( 10,0 ) circle ( 3pt ) ; at ( 10,0 ) @xmath24 ; ( 14,4 ) circle ( 3pt ) ; at ( 14,4 ) @xmath39 ; ( 14,-4 ) circle ( 3pt ) ; at ( 14,-4 ) @xmath23 ; ( 14,-4 )
( 14,2 ) ; ( 14,2 )
( 14,4 ) ; ( 14,-4 )
( 11,-1 ) ; ( 11,-1 )
( 10,0 ) ; ( 10,0 ) (13,3 ) ; ( 13,3 )
( 14,4 ) ; as is well known , repeated application of the star - triangle relation , see figure [ fig6 ] , implies that transfer matrices commute . defining in 2, ... ,9 ( 2*,80pt )
( 2*,160pt ) node[above ] @xmath26 ; ( 19,150pt ) (3.5,150pt) .. controls + + ( left:8 mm ) and + + ( up:8mm) .. (1,80pt ) node[left ] @xmath0 ; ( 19,90pt ) (3.5,90pt) .. controls + + ( left:8 mm ) and + + ( down:8mm) .. (1,160pt ) node[left ] @xmath44 ; at ( 19,150pt ) @xmath45 ; at ( 19,90pt ) @xmath46 ; @xmath38 in 2, ... ,9 ( 2*,-40pt ) ( 2*,40pt ) node[above ] @xmath26 ; ( 20,-40pt) .. controls++(up:8 mm ) and + + ( right:8mm) .. (18.5,30pt) ( 3,30pt ) node[left ] @xmath44 ; ( 20,40pt ) .. controls++(down:8 mm ) and + + ( right:8mm) .. (18.5,-30pt) ( 3,-30pt ) node[left ] @xmath0 ; at ( 2,30pt ) @xmath47 ; at ( 2,-30pt ) @xmath48 ; & & ( t_q)_,=_j=1^lw_pq(_j,_j ) w_pq(_j+1,_j ) , + & & ( _ r)_,=_j=1^lw_pr(_j,_j ) w_pr(_j,_j+1),with periodic boundary conditions @xmath49 and
@xmath50 , we can prove @xcite more precisely , combining the chiral potts weights in two successive diagonal rows , as indicated in figure [ fig6 ] , and summing over all spins in the middle row , we get the two products of the two transfer matrices , after closing the horizontal @xmath26 and @xmath0 rapidity lines . inserting an inversion relation ,
we can then repeatedly apply ( [ startriangle ] ) .
the boltzmann weights @xmath51 and @xmath52 for edges between spins @xmath23 and @xmath24 , @xmath31 , in ( [ bweights ] ) can be rewritten in terms of @xmath53-pochhammer symbols as w_pq(n)=()^n_j=1^n = _ pq^n , + w_pq(n)=(_p_q)^n_j=1^n = |_pq^n , where & & ( x,)_n=_=0^n-1(1-x^),=e^2/n , + & & _ pq=_p y_q/_q y_p,_pq = x_q / y_p,_pq = x_p/ y_q , + & & |_pq=_p_q x_p/ y_q,|_pq = x_q/ x_p,|_pq = y_p/ y_q , it is obvious that the star - triangle relation is related to identities of basic hypergeometric series .
l=0^z^l , [ defbhs]with the @xmath0-pochhammer symbol ( x;q)_l=_j=0^l-1(1-xq^j ) , |q|<1 , generalizing the usual pochhammer symbol ( a)_l = a(a+1)(a+l-1 ) .
the series defined in ( [ defbhs ] ) is then the @xmath0-deformation of the generalized hypergeometric series = _ l=0^z^l .
[ defgf]for the root - of - unity case with @xmath54 , we find _ j=0^n-1(1-x^j)=(1-x^n)(;)_n=0 .
thus in the limit @xmath55 , the summand in the series in ( [ defbhs ] ) is divergent . to have a well - defined summand
, we have to let @xmath56 to make it a terminating series , that is = _
l=0^n-1 z^l .
[ defbhsc ] many well - known theorems in basic hypergeometric series may not hold in the root - of - unity case , and need to be modified , such as the transformation formula * _ theorem _ 10.2.1 * @xcite _ n=0^ x^n= , |x|<1 [ 10.2.1]which is the @xmath0-analog of the binomial theorem _
n=0^ x^n= ( 1-x)^-a|x|<1 . only for the case with @xmath57 , can ( [ 10.2.1 ] ) be extended to the root - of - unity case as _
n=0^ x^n = ( ^-x ; ) _ =
_ n=0^(-x)^n^ n(n-1)-n .
this well - known formula is due to rothe , according to @xcite .
similarly , the @xmath0-analog of euler s formula = ( 1-x)^c - a - b = , does not hold for @xmath55 .
however , for some particular values of @xmath23 , @xmath24 and @xmath39 , we find & & [ 10.10.1c ] + & & = ( ^+-t;)_n--+ .the proof was rather non - trivial , as will be outlined later .
this shows that the summation formulae of the basic hypergeometric series can not be extended to the root - of - unity case , unless further restrictions are imposed .
if we impose the periodicity requirement for the finite sum in ( [ defbhsc ] ) to be z^n=_j=1^p_j^n,^n= , we obtain a cyclic hypergeometric function " with summand periodic mod @xmath25 .
the fourier transform of the chiral potts weight is a cyclic @xmath58 , i.e. w^(f)(k)=_n=0^n-1^nkw(n)= , where w(n)=^n , w(n+n)=w(n ) , ^n=. it is summable as shown in @xcite , namely 21 , x yz= , [ phi21c]where , p(x ) = _ j=1^n-1(1-^jx)^j / n , + ( x)=(1-x^n)^1/n , _ 0=e^i(n-1)(n-2)/12n.however , the periodic restriction make the riemann sheet structure very complicated .
it was also shown in @xcite the following relations hold 21,xyz 21,y / xz / zx = n .
[ wff ] and & & 21,x^my^nz^k [ rec4 ] + & & = 21,xyz ( ) ^ k ( z^k)^-n .
it is found in @xcite that the cyclic hypergeometric @xmath59 satisfies the transformation formula - pochhammer symbol . ]
32,x_1,x_2y_1,y_2z = a 32 , , , x_2 z / y_2z_1 , [ phi32c]where the periodic restriction is , [ periodic]and the constant a = n^-1 21,x_1y_1z_1 21,x_2y_2 , z_1^n=. if @xmath60 , we find @xmath59 on the right - hand side of ( [ phi32c ] ) becomes @xmath61 , so that it is a product of three cyclic @xmath61 , which are summable .
therefore we find @xmath59 is also summable at @xmath60 . now
( [ periodic ] ) becomes ( 1-x_1^n)(1-x_2^n)=(1-y_1^n)(1-y_2^n ) , which gives rise to a very complicated riemann sheet structure .
we next outline the proof of ( [ 10.10.1c ] ) . consider the series in ( [ phi32c ] ) with @xmath62 , @xmath63 , @xmath64 , @xmath65 and @xmath66 . in this case
, it becomes 32,x,^-y ^x , y^= @xmath67 and @xmath68 , and then use ( [ phi32c ] ) twice , we obtain 32,^-y , x^x , y^&= & a32,^z,^+-zx / y z,^z x / y^- + & = & a b 21 ^ -,^-^ ^+-x y , where the constants are & & a = n^-1 21,^-y^x1/z 21,xy^z , + & & b = n^-1 21,^+-zx / y^zx / y1 21,^zz^- .
now we may use ( [ rec4 ] ) , ( [ wff ] ) and ( [ phi21c ] ) to find ab/|b=(^+-x / y;)_n--+ .
clearly , the star - triangle relation ( [ startriangle ] ) gives a summation formula for @xmath69 . to convert the left - hand side of the star - triangle equation into @xmath69 , we must rewrite the pochhammer symbols in the weights in terms of the rapidities .
we find & & _ 1=^c - a-1y_r / y_p,_2=x_p / y_q , _ 3=^c - b x_q / x_r , + & & _ 1=^c - ax_p / x_r,_2=x_q / y_p , _ 3=^c - b+1y_r / y_q , + & & _ 1=y_p/_p_r x_r,_2= _ p y_q/_q y_p , _
3=_q_r x_r / y_q.this gives ^2_1_2_3=_1_2_3,_1_2_3=. it is known that when the well - balanced condition q a_1 a_2a_p+1=b_1 b_2b_p , z = q , is satisfied , there exist many summation formulae for the basic hypergeometric series .
the most well - known summation formula is = , which is the @xmath0-analog of the pfaff - saalschtz formula = . by taking the @xmath70 limit @xcite ,
the star - triangle relation becomes the summation formula for double sided series _
n=-^ = , if both the well - balanced condition and the periodicity condition hold , i.e. & & x_1+x_2+x_3 + 2=y_1+y_2+y_3 , + & & x_1x_2x_3= y_1y_2y_3 , where & & g(x_1,x_2,x_3y_1,y_2,y_3 ) + & & _
j=2 ^ 3(x_j)(1-x_j ) _ i=1 ^ 3(y_i - x_1)(1-y_i+x_1),which is invariant under * permutations of @xmath71 and @xmath72 , * reflections @xmath73 , @xmath74 simultaneously , * translations @xmath75 , @xmath76 for @xmath77 or @xmath78 .
we note that there are also other @xmath79 limits with spins taken from a finite or infinite continous interval @xcite .
at the qinhuangdao meeting we also discussed the superintegrable subcase of the integrable chiral potts model , which has an additional underlying onsager loop group structure , and we discussed how the spectrum is then dominated by an affine quantum group at an @xmath25-th root of unity , where @xmath25 can be odd or even .
even though the theory of this quantum group is much better understood for odd roots of unity , we presented new approaches to establish the higher quantum serre relations also for the even root - of - unity case .
we shall not go in more detail here as a more complete account has been presented at the meeting in honor of baxter s 75th birthday @xcite . | in this paper we discuss the integrable chiral potts model , as it clearly relates to how we got befriended with vaughan jones , whose birthday we celebrated at the qinhuangdao meeting .
remarkably we can also celebrate the birthday of the model , as it has been introduced about 30 years ago as the first solution of the star - triangle equations parametrized in terms of higher genus functions . after introducing the most general checkerboard yang baxter equation , we specialize to the star - triangle equation , also discussing its relation with knot theory .
then we show how the integrable chiral potts model leads to special identities for basic hypergeometric series in the @xmath0 a root - of - unity limit .
many of the well - known summation formulae for basic hypergeometric series do not work in this case .
however , if we require the summand to be periodic , then there are many summable series . for example , the integrability condition , namely , the star - triangle equation , is a summation formula for a well - balanced @xmath1 series .
we finish with a few remarks about the relation with quantum groups at roots of unity . #
1#2 @c@#1 + # 2 = cmr12 at 14pt # 1#2#3#4#5_#1_#2 # 1#2#3_p+1_p # 1#2#3#4#5_#1f_#2 |
current gravitational wave detectors such as ligo , virgo , geo and tama are now starting science runs " and the plans for upgrades to these detectors are already in the works .
this paper attempts to model a likely occurrence in galactic nuclei i.e. the capture of compact stellar - sized objects by the supermassive black holes that exist in the center of most galaxies .
these type of events will generate gravitational waves within the sensitivity band of nasa s space - based lisa project .
this problem has been treated in the past , quite extensively , using a frequency domain approach to linear black hole perturbation theory @xcite . a similar calculation in the time domain was only recently presented for the special case of equatorial , circular particle orbits around a rapidly rotating black hole @xcite . in this work ,
we extend the time domain calculation to treat more generic types of orbits , i.e. elliptic and inclined , and compare energy and angular momentum fluxes lost from gravitational wave emission to current frequency domain calculations .
we ignore the effects of radiation reaction in this work , with the hope of addressing those in detail sometime in the future .
we note that for highly elliptic orbits , both approaches develop problems and only a very detailed study , possibly with the inclusion of radiation reaction effects will lead us to favor one over the other .
this paper is organized as follows . in the next section we review ,
briefly , the basics of our time domain approach .
details can be found in the second section of our past work @xcite .
then in the following section , we present our results for waveforms and radiative fluxes and compare them with frequency domain results . in the last section , we present our conclusions and emphasize that future work will be based on the inclusion of radiation reaction in our calculations .
the general approach to this problem in the time domain , with details regarding numerical implementation , stability , second - order convergence , etc . have been presented in the earlier work on the subject @xcite , so we simply summarize some of the essentials here .
the teukolsky equation in boyer - lindquist coordinates with the matter - source term , appears below @xcite , @xmath0 \frac{\partial^2 \psi}{\partial t^2 } + \frac{4 m a r}{\delta } \frac{\partial^2 \psi}{\partial t \partial \phi } + \left [ \frac{a^2}{\delta } - \frac{1}{\sin^2 \theta } \right ] \frac{\partial^2 \psi}{\partial \phi^2 } - \delta^{-s } \frac{\partial}{\partial r } \left ( \delta^{s+1}\frac{\partial \psi}{\partial r}\right ) \nonumber \\ & & - \frac{1}{\sin \theta } \frac{\partial}{\partial \theta } \ , \left ( \sin \theta \frac{\partial \psi}{\partial \theta } \right ) -2 s \left [ \frac{a ( r - m)}{\delta } + \frac{i \ , \cos \theta}{\sin^2 \theta } \right ] \frac{\partial \psi}{\partial \varphi } - 2 s \left [ \frac{m(a^2 - r^2)}{\delta } - r - i\ , a\ , \cos \theta \right ] \frac{\partial \psi}{\partial t } \nonumber \\ & & + \left [ s^2 \cot^2 \theta - s \right ] \psi = 4 \pi ( r^2+a^2\cos^2\theta ) t\end{aligned}\ ] ] where @xmath1 .
the methodology for numerically evolving the homogeneous part of this equation has been presented in detail at various places in the literature @xcite . in this work
we focus on the matter - source term . in the time domain teukolsky equation
this source term is given by @xmath2 where @xmath3 + ( \delta^ * - \tau^ * + \beta^ * + 3 \alpha + 4 \pi ) \nonumber \\ & & \times [ ( \delta + 2 \gamma + 2 \mu^*)t_{nm^ * } - ( \delta^ * - \tau^ * + 2 \beta^ * + 2 \alpha ) t_{nn}].\end{aligned}\ ] ] all the greek symbols used above are the ones as defined in the original teukolsky equation paper @xcite . by choosing their values in boyer - lindquist coordinates and expanding everything we obtain an explicit but quite complicated form of the source term .
we treat the delta functions that appear in the energy - momentum tensor of the particle in orbit by using an approximation in which we substitute the delta function by a very narrow gaussian function ( whose width is only of a few grid points ) , @xmath4 we handle the deltas in the radial and angular direction through the gaussian approximation , whereas the @xmath5 function we handle analytically .
we use code generation software to convert the source term expressions to fortran code that can be used with our teukolsky equation evolver .
we verify that this substitution is adequate by showing convergence of the obtained waveform and radiative flux values as @xmath6 .
we also take special care that the gaussian integrates to unity on the spatial hypersurface . to achieve this with the three - metric of the background black hole spacetime , we normalize the mass - density of the particle by the factor , @xmath7 where @xmath8 is the 3-metric of flat space and @xmath9 is the 3-metric of a slice of constant boyer - lindquist time of the kerr spacetime .
the initial data we take for the fields is zero . since we are considering the teukolsky equation with a source this would correspond to the particle `` appearing suddenly '' and that generates an artificial burst of radiation .
we handle this by evolving to late times and computing fluxes only when the system has settled down . sometime in the future
, we will examine the implementation of initial data in this context , but at this time we will simply rid the system of this initial burst , by evolving `` long enough '' .
lastly , we also need a high accuracy general geodesic equation integrator , to obtain the position of the particle at every time step .
the main problem with integrating the geodesic equations is the presence of turning points . at the turning points ,
since the right - hand sides of the geodesic equations go to zero ( see below ) , the integrators have trouble resolving the motion correctly and large amounts of error accumulate .
@xmath10 ^ 2- \delta\left[r^2 + ( l_z - a e)^2 + q\right]\;,\label{eq : rdot } \\ % \sigma^2\left({d\theta\over d\tau}\right)^2 & & = q - \cot^2\theta l_z^2 -a^2\cos^2\theta(1 - e^2)\;,\label{eq : thetadot }
\\ % \sigma\left({d\phi\over d\tau}\right ) & & = \csc^2\theta l_z + ae\left({r^2+a^2\over\delta } - 1\right ) - { a^2l_z\over\delta}\;,\label{eq : phidot } \\
% \sigma\left({dt\over d\tau}\right ) & & = e\left[{(r^2+a^2)^2\over\delta } - a^2\sin^2\theta\right ] + al_z\left(1 - { r^2+a^2\over\delta}\right)\;.\label{eq : tdot } \label{eq : geodesiceqns}\end{aligned}\ ] ] fortunately , there is such a public domain integrator , called `` geod '' written by scott hughes @xcite . in geod , this problem is solved by transforming the equations to orbital parameters that do not have turning points .
more information can be found about this approach in the documentation of this software package .
we call this geodesic equation integrator from within our teukolsky code to update the particle s position at every time step .
@xmath11 & @xmath12 & @xmath13 @xmath14 & @xmath15 & @xmath16 @xmath17 & @xmath18 & @xmath19 in this section , we present sample waveforms and compute radiative fluxes from numerical evolutions of the teukolsky equation with matter - source term as discussed in the previous section . as detailed before @xcite
, the numerical implementation of this work is in the form of a set of 2 + 1 dimensional linear pde s that uses the lax - wendroff evolution scheme .
also discussed therein is the convergence and stability of the entire approach , therefore that study shall not be presented here .
we calculate the energy and angular momentum fluxes emitted due to gravity wave emission from the binary for a sample equatorial , elliptic orbit and show close agreement with current results of high accuracy frequency domain calculations @xcite .
the code monitors the value of the teukolsky function at a set radial location in the grid far from the horizon and computes the energy and angular momentum fluxes in gravitational radiation according to the formulae @xcite : @xmath20\right\}\ ; \;\;\;\;\ , { \rm d}\omega=\sin\theta{\rm d}\theta{\rm d}\varphi . \label{jfl}\end{aligned}\ ] ] in tables [ comp1 ] and [ comp2 ] we present representative comparisons with frequency domain calculations for equatorial , elliptic orbits .
the level of agreement between the two approaches is about the same ( within @xmath21 ) over a wide range of eccentricities ( @xmath22 through @xmath23 ) .
it is interesting to note both frequency domain approaches and time domain approaches begin to have computational issues for orbits of high eccentricity ( above @xmath23 ) and also for hyperbolic orbits . in the frequency domain , where the method is based on variable
separating the teukolsky equation the need to compute for a large number of @xmath24 and @xmath25 modes for a given accuracy , makes it computationally very demanding ( see figure 1 ) . in the time domain case ,
since highly elliptic orbits have large orbital periods , we need to evolve for very long times ( several tens of thousands times the mass of the hole ) in order to compute average radiation fluxes ( see figure 3 ) .
therefore it is not clear at this point , which of the two approaches is better suited to address the general problem of extreme mass ratio inspiral .
however , this question will be revisited , once more work is done on the inclusion of radiation reaction and a more detailed comparison study is undertaken .
we also include a representative waveform in figure 2 from the evolution of a particle in an equatorial , elliptic orbit .
it shows the real part of the @xmath26 mode of the teukolsky function , and as mentioned before , the initial burst of radiation is unphysical , since it corresponds to the particle `` appearing out of nowhere '' .
it is only after the evolution has settled down to a time - periodic pattern , do we compute average fluxes . in figure 3
we show the variation of energy fluxes with time , from the evolution of a particle in an equatorial , elliptic orbit and also in an inclined , circular orbit .
its is interesting to note that if we keep the semi - latus rectum fixed , and study the energy - fluxes as they change with eccentricity for elliptic - equatorial orbits , they increase at first , but then begin to drop after reaching a maximum at an eccentricity of about @xmath27 .
it turns out that this variation can be understood easily by studying the same in the context of a binary newtownian system where this turn - over point can be computed to be about @xmath28 @xcite .
lastly , we suggest here an alternate approximate approach to treating delta functions on a spatial grid . this approach is significantly more efficient ( more than two orders of magnitude faster in a numerical implementation ) and yields significantly more accurate results ( in terms of their agreement with frequency domain ) compared with
the above mentioned narrow gaussian approach .
one can think of this as a limiting case of the gaussian particle approach , where the width of the gaussian is made so narrow that it is comparable to the numerical grid separation . in this approach , we implement a `` discrete '' version of the delta function on the spatial grid . more specifically
, we define a discrete - delta function as being unity at the appropriate grid point and zero everywhere else .
we then define the derivatives of the delta function as the result of applying the usual finite - difference derivative stencils to the above mentioned discrete - delta function .
of course , the discrete - delta needs to be normalized suitably , i.e. scaled by a factor such that its integral on the spatial grid is unity .
note that such a delta function satisfies the usual delta function properties , such as @xmath29 , on the discrete grid .
care has to be taken to study convergence of the code using such an approach , however , just the efficiency and accuracy obtained by using this simple idea , makes it worth mentioning and investigating further .
we can now treat different types of particle orbits in the context of extreme mass ratio inspiral , using black hole perturbation theory in the time domain .
we get good agreement in comparisons with current frequency domain results for energy and angular momentum radiative fluxes . at this time , it is not clear which of these two approaches is better suited to the generic case and will provide more accurate results , more efficiently . in future work
, we hope to be able to further widen the scope of these calculations by incorporating the case of a spinning particle in orbit about a kerr hole and we also expect to investigate methods to include radiation reaction
. the first approach will be to correct the orbit trajectory using energy - balance arguments , as for instance in @xcite .
but , later other proposals for the inclusion of radiation reaction forces will be incorporated .
we express our gratefulness to daniel kennefick for working very closely with us and generating results for comparisons using his frequency domain codes .
we also thank jorge pullin , pablo laguna , ramon lopez - aleman , eric poisson , karl martel , lee samuel finn and scott hughes for input and discussions all along , relating to this research .
this work was supported by nsf grant number phy-0140236 .
we also thank the long island university for its continued research support .
some of the numerical simulations were performed at the boston university s scientific computing and visualization group .
we thank them for allocating us supercomputing time on their ibm p690 server . | we extend the treatment of the problem of the gravitational waves produced by a particle of negligible mass orbiting a kerr black hole using black hole perturbation theory @xcite in the time domain , to elliptic and inclined orbits .
we model the particle by smearing the singularities in the source term using narrow gaussian distributions .
we compare results ( energy and angular momentum fluxes ) for such orbits with those computed using the frequency domain formalism . |
models for new physics at the tev scale are typically motivated by the hierarchy problem .
they strive to explain the hierarchy between the electroweak ( ew ) scale and the planck scale , or at least remove the incredible fine tuning required in the standard model ( sm ) for having such a hierarchy .
however , the sm is stubbornly surviving the first tests at the lhc and there are no traces of new physics yet .
in addition , the mass of the recently discovered higgs - like particle suggests that the sm can be a consistent theory all the way to the planck scale .
this gives more weight to speculations that there is no mechanism , no new physics , that stabilizes the hierarchy , or that the ew scale is selected based on anthropic reasoning .
however , even when we ignore the hierarchy problem , the sm is still not very satisfactory .
the three gauge couplings , all couplings of the higgs boson to fermions , the higgs mass , and the higgs quartic coupling are free parameters .
this motivates us to explore extensions of the standard model in which at least some of these parameters could be understood .
we have recently showed that extending the standard model by three complete vector - like families ( sm+3vfs ) with masses of order 1 tev - 100 tev allows for the unification of gauge couplings @xcite .
predictions for gauge couplings at the ew scale are highly insensitive to fundamental parameters : the grand unification scale , the unified gauge coupling , and the masses of vector - like fermions .
their observed values can be mostly understood from infrared fixed point behavior . in this paper
we discuss gauge coupling unification in detail in models with additional 1 to 4 complete vector - like families ( vfs ) , and derive simple rules for masses of vector - like fermions required for exact gauge coupling unification .
we then focus on scenarios with 3 or more vector - like families that lead to insensitive unification of gauge couplings . requiring the smallest splitting between masses of vector - like fermions
we show that the best motivated grand unified theory ( gut ) scale is at @xmath0@xmath1 gev .
we provide examples of the spectrum as a function of the gut scale , which can be as large as the planck scale .
we discuss constraints from proton decay and show that predictions from the best motivated region are close to current limits . however , due to insensitivity of predicted ew scale values of gauge couplings to gut scale parameters , no sharp predictions can be made without knowing the spectrum of vector - like fermions .
the focus on complete families follows from the fact that quantum numbers of quarks and leptons in the sm nicely fill representations of a gut symmetry , @xmath2 and @xmath3 of @xmath4 or @xmath5 of @xmath6 .
this provides a support for the idea of grand unification and the unification of gauge couplings @xcite .
additional complete families represent some of the simplest extensions of the sm that can be embedded into simple guts .
consequently , there are many studies exploring various features of vector - like families ( mostly in supersymmetric models ) , see for example refs . @xcite .
in addition , vector - like fermions , not necessarily coming in complete gut multiplets , are often introduced on purely phenomenological grounds to explain various discrepancies between observations and sm predictions .
examples include discrepancies in precision ew z - pole observables @xcite , and the muon g-2 anomaly @xcite .
however , with arbitrary new particles there are many possibilities for gauge coupling unification .
therefore , we generalize the mass rules and the method to classify scenarios consistent with gauge coupling unification to an arbitrary extension of the standard model . the method to classify scenarios consistent with gauge coupling unification in terms of physical masses of extra particles starts with finding the mass scales that represent average " masses of all particles charged under given gauge symmetry required for gauge coupling unification ( they are defined precisely in the next section and are referred to as crossing scales ) .
these crossing scales are easy to obtain and they immediately give us information about the required spectrum .
first of all , if they do not exist between the ew scale and the gut scale , the gauge coupling unification in a given model is not possible , no matter what the splitting between masses of extra particles is .
second , the splitting between crossing scales represents the minimum necessary splitting in the spectrum required .
third , from the mass formulas that define crossing scales in terms of masses of extra particles one can immediately see the basic features of the spectrum required , and the spectrum can be calculated .
in addition , these formulas also indicate the freedom one has in imposing further relations between masses of extra particles
. this might be useful when searching for models that relate masses of particles at a given scale .
the mass rules given in terms of particle masses can be evolved to an arbitrary scale , _
e.g. _ the gut scale , which would provide the boundary conditions that need to be satisfied . however , the renormalization group ( rg ) evolution of the mass rules depends on additional assumptions one has to make about the origin of the masses and the scale at which these masses are generated .
this paper is organized as follows . in sec .
[ sec : rge ] we discuss rg evolution of gauge couplings in models with extra vfs .
we start with the discussion of ir fixed point predictions for gauge couplings , then add threshold corrections from a universal mass of vector - like fermions , and , finally , we add effects from splitting masses of vector - like fermions . we discuss sensitivity of predicted values of gage couplings to fundamental parameters . finally , we derive simple mass rules that have to be satisfied in order to get exact gauge coupling unification .
we generalize the method to classify all solutions consistent with gauge coupling unification to an arbitrary extension of the sm . in sec .
[ sec : discussion ] we discuss constraints from proton decay , the stability of the ew minimum of the higgs potential , and discuss a possible origin of masses of vector - like fermions .
we give few concluding remarks in sec .
[ sec : conclusions ] .
the one - loop renormalization group equations ( rges ) for three gauge couplings , @xmath7 , are given by : @xmath8 where @xmath9 with @xmath10 representing the energy scale at which gauge couplings are evaluated .
the beta function coefficients , @xmath11 , in the sm with @xmath12 families are given by @xmath13 for @xmath14 we get the usual sm result , @xmath15 .
with extra @xmath16 pairs of complete vfs we have @xmath17 ( a vector - like partner contributes in the same way ) .
for example , in sm+3vfs we find @xmath18 which indicates that all three gauge couplings are asymptotically divergent ( this result obviously holds for 3 or more pairs of vfs ) .
( top solid line ) , @xmath19 ( middle solid line ) , and @xmath20 ( bottom solid line ) in the sm extended by three vector - like families for @xmath21 at @xmath22 gev . dashed lines in the same order show the running of gauge couplings in the sm .
masses of 3 vfs are neglected in the left plot , and fixed to 10 tev ( indicated by @xmath23 ) in the right plot .
the crossing points in the evolution of gauge couplings in the sm+3vfs and the sm indicated in the left plot define the common threshold scales , @xmath24 , for masses of particles charged under given symmetry required for exact gauge coupling unification .
, title="fig : " ] ( top solid line ) , @xmath19 ( middle solid line ) , and @xmath20 ( bottom solid line ) in the sm extended by three vector - like families for @xmath21 at @xmath22 gev .
dashed lines in the same order show the running of gauge couplings in the sm .
masses of 3 vfs are neglected in the left plot , and fixed to 10 tev ( indicated by @xmath23 ) in the right plot .
the crossing points in the evolution of gauge couplings in the sm+3vfs and the sm indicated in the left plot define the common threshold scales , @xmath24 , for masses of particles charged under given symmetry required for exact gauge coupling unification .
, title="fig : " ] the evolution of gauge couplings in the sm and an example of the evolution in the sm+3vfs case are shown in fig .
[ fig : thresholds ] .
the numerical analysis closely follows that of ref .
@xcite . for the sm evolution we use the z - scale central values of @xmath25 , @xmath26 , and @xmath27 , together with the top quark mass @xmath28 gev , which can be found in ref .
the @xmath29 and @xmath30 are related to @xmath31 through @xmath32 where , assuming the @xmath4 normalization of the hypercharge , @xmath33 .
we set the higgs boson mass to @xmath34 gev @xcite .
the example of the rg evolution of gauge couplings in the sm+3vfs starts with unified gauge coupling @xmath21 at @xmath22 gev .
the crossing points in the evolutions of gauge couplings in these two cases , which will be important for the discussion of threshold corrections , are indicated in the left plot by @xmath24 . in all numerical results we use full two loop rges @xcite
, we integrate out all particles with masses above @xmath35 at their mass scale , and include one - loop matching corrections for @xmath36 and @xmath37 @xcite .
we assume that yukawa couplings of vector - like fermions are negligible , and we also neglect yukawa couplings of all fermions in the sm except the top quark . the results of the numerical analysis we present can be understood from approximate analytic formulas .
the one - loop rges can be solved , and we can express gauge couplings at the ew scale in terms of the gut scale , and values of gauge couplings at @xmath38 : @xmath39 assuming gauge coupling unification , @xmath40 , and neglecting threshold corrections both at the ew scale and the gut scale , we can express one gauge coupling in terms of the other two .
for example : @xmath41 where @xmath42 , and @xmath43 .
for the measured values of @xmath29 and @xmath44 the su(5 ) embedding of the sm predicts @xmath45 which is about 40% below the experimental value . adding complete chiral or vector - like families at the ew scale does not change at all the one - loop prediction given in eq .
( [ eq : prediction_1loop ] ) since complete families contribute equally to all three beta function coefficients , see eq .
( [ eq : bi ] ) .
furthermore , the scale of unification ( more precisely the scales where any two couplings meet ) does not change at one - loop , only the value of the unified gauge coupling increases . with increasing the number of extra families , at some point ,
the couplings become non - perturbative before they meet , and eventually reach the landau pole .
further increase of the number of families lowers the energy scale at which the landau pole occurs .
however , the sm extended with a sufficient number of complete vector - like families so that all couplings are asymptotically divergent offers a new possibility .
vector - like families introduce an additional scale to the problem associated with masses of vector - like fermions , @xmath46 , and they contribute to the rg evolution of gauge couplings only above this energy scale .
this allows us to consider models with a large ( but still perturbative ) unified gauge coupling at a high scale , higher than the scale at which the landau pole would occur if the vfs were at the ew scale .
consequently , in the rg evolution to lower energies , gauge couplings run to the ( trivial ) infrared ( ir ) fixed point .
thus , at lower energies , the values of gauge couplings are determined only by the particle content of the theory and how far from the gut scale we measure them . since the exact value of @xmath47 becomes irrelevant , instead of one prediction of the conventional unification , eq .
( [ eq : prediction_1loop ] ) , we have two predictions for ratios of gauge couplings . at the @xmath46 scale ,
the vector - like fermions are integrated out , and below this scale gauge couplings run according to the usual rg equations of the standard model . in a way ,
the two parameters of the conventional unification , @xmath38 and @xmath47 , are replaced by @xmath38 and @xmath46 .
the discrepancies of ir fixed point predictions from observed values can be explained by threshold effects of extra vector - like fermions .
the ir fixed point predictions were discussed in detail in ref .
@xcite . in models with asymptotically divergent couplings ,
these can be easily obtained if the 1-loop rges are good approximations . assuming a large enough unification scale and large ( but still perturbative ) unified gauge coupling , the first term in eq .
( [ eq : sol_1loop ] ) dominates , and the ratios of gauge couplings are given by ratios of beta function coefficients , @xmath48 this can be translated into the prediction for @xmath30 : @xmath49 where @xmath50 .
numerically , we find @xmath51 in the case of sm+3vfs , which is identical to the value obtained assuming 9 chiral families @xcite . similarly ,
in sm+4vfs we find @xmath52 . in the case of sm+3vfs , the one - loop rge for @xmath53 given in eq .
( [ eq : rge ] ) is not a good approximation because of the accidentally small @xmath54 coefficient .
the two - loop contribution to the beta function is well approximated by the term proportional to @xmath55 , @xmath56 where @xmath57 for sm+3vfs @xcite .
thus , the two loop contribution is larger than the one - loop contribution for @xmath58 . ,
proportional to @xmath59 , represents a @xmath60 correction to 1 + 2-loop beta function for @xmath61 @xcite . ]
the rge for @xmath53 can be solved by adding the 1-loop contribution as an expansion in @xmath62 to the solution obtained from the 2-loop contribution only @xcite , @xcite .
alternatively , we can solve the full rge given in eq .
( [ eq : rge2l ] ) and find : @xmath63 neglecting @xmath64 , we obtain the second prediction : @xmath65 numerically , for @xmath66 , it predicts @xmath67 in the case of sm+3vfs .
the beta function coefficients for @xmath53 in the sm+4vfs scenario are @xmath68 and @xmath69 .
the 1-loop term in the rg equation ( [ eq : rge2l ] ) dominates for @xmath70 in this case . the proximity of predictions from the ir fixed point , eqs .
( [ eq : s2w ] ) and ( [ eq : prediction_1 + 2loop ] ) , to observed values is certainly intriguing .
although they are not a perfect match to measured values , the discrepancies can be easily accommodated by taking into account threshold corrections from vector - like fermions that should be integrated out at the @xmath23 scale . the existence of a scale associated with masses of vector - like fermions is strongly suggested by the overlay of the rg evolution of gauge couplings in the sm and those in the sm+3vfs assuming unified gauge coupling at a high scale given in fig . [ fig : thresholds ] .
all three gauge couplings in these two scenarios cross at comparable scales suggesting a common threshold at which particles from vfs are integrated out . indeed , for the example given in fig .
[ fig : thresholds ] , fixing all the masses of 3 vfs to 10 tev , shown in fig .
[ fig : thresholds ] ( right ) , the ew scale values of gauge couplings are predicted within 8% from measured values . in the next subsection
, we will show that the measured values of gauge couplings can be precisely reproduced by splitting the masses of vector - like fermions .
first , however , we would like to discuss general features of this result assuming the common mass of vfs .
the fairly good agreement of predicted values of gauge couplings from 3 vfs at @xmath71 tev with observed values does not rely on the specific choice of the gut scale and the value of the unified gauge coupling .
the ew scale values of gauge couplings are highly insensitive to these parameters , which can be understood from ir fixed point behavior .
the low sensitivity of predicted values of gauge couplings to fundamental parameters is demonstrated in fig .
[ fig : m_g_vs_m_vf_3vf ] ( left ) .
it shows a large region of the gut scale , @xmath38 , and the universal mass of fermions from 3 vector - like families , @xmath23 , from which the values of gauge couplings at the ew scale are simultaneously predicted within @xmath72 from the measured values .
it also shows the best fit that predicts all couplings within @xmath73 .
the gut scale is the best motivated between @xmath74 gev and @xmath75 gev with the best fit close to @xmath1 gev . for completeness ,
a similar plot for @xmath29 , @xmath76 , and @xmath53 is presented in fig .
[ fig : m_g_vs_m_vf_3vf ] ( right ) .
however , as we will see from the discussion of threshold corrections in the next subsection , the plot on the left for @xmath20 , @xmath19 , and @xmath53 is more indicative of the best motivated values of @xmath38 and @xmath23 .
, @xmath20 ( green ) , @xmath19 ( blue ) , and @xmath53 ( red ) , as functions of the gut scale , @xmath38 , and the universal mass of fermions from 3 vector - like families , @xmath23 , for fixed @xmath21 .
solid lines represent the central experimental values of three gauge couplings , the shaded regions represent @xmath77 ranges , and the dashed lines in unshaded areas represent @xmath78 ranges .
the lightly shaded area corresponds to a @xmath79 range of @xmath53 . in the overlapping ( bright red ) region , all three gauge couplings are simultaneously predicted within @xmath72 from the measured values , and the small black area in the red region represents the best fit with all three couplings within @xmath73 from the measured values .
the gray region corresponds to @xmath80 ; @xmath81 becomes non - perturbative very fast with increasing @xmath23 from the value that corresponds to the boundary of this region .
right : the same as in the plot on the left but for @xmath29 ( orange ) , @xmath76 ( purple ) , and @xmath53 ( red ) .
the dotted purple line represents @xmath76 being @xmath82 from the central value.,title="fig:",width=278 ] , @xmath20 ( green ) , @xmath19 ( blue ) , and @xmath53 ( red ) , as functions of the gut scale , @xmath38 , and the universal mass of fermions from 3 vector - like families , @xmath23 , for fixed @xmath21 .
solid lines represent the central experimental values of three gauge couplings , the shaded regions represent @xmath77 ranges , and the dashed lines in unshaded areas represent @xmath78 ranges .
the lightly shaded area corresponds to a @xmath79 range of @xmath53 . in the overlapping ( bright red ) region , all three gauge couplings
are simultaneously predicted within @xmath72 from the measured values , and the small black area in the red region represents the best fit with all three couplings within @xmath73 from the measured values .
the gray region corresponds to @xmath80 ; @xmath81 becomes non - perturbative very fast with increasing @xmath23 from the value that corresponds to the boundary of this region .
right : the same as in the plot on the left but for @xmath29 ( orange ) , @xmath76 ( purple ) , and @xmath53 ( red ) .
the dotted purple line represents @xmath76 being @xmath82 from the central value.,title="fig:",width=278 ] in order to understand the sensitivity of the ew scale values of gauge couplings to fundamental parameters quantitatively , it is instructive to estimate separate contributions to @xmath83 from @xmath47 , @xmath38 , and @xmath23 . note that the values of @xmath84 are approximately 59 , 30 , and 8.4 , respectively . from eqs .
( [ eq : sol_1loop ] ) and ( [ eq : prediction_1 + 2loop ] ) , we see that @xmath85 contributes less than @xmath86 to the ew scale values of gauge couplings .
it is the least important parameter .
plots in fig .
[ fig : m_g_vs_m_vf_3vf ] for any @xmath87 would look almost identical .
increasing @xmath47 moves all the contours slightly to the right .
the largest contribution to ew scale values of couplings originates from the @xmath88 term multiplied by corresponding beta function coefficients .
the second largest contribution to gauge couplings comes from masses of vector - like fermions .
the ir fixed point predictions for the gauge couplings at the ew scale , obtained from eq .
( [ eq : sol_1loop ] ) for @xmath89 with @xmath90 , and from eq .
( [ eq : sol_2loop ] ) for @xmath91 , are modified by threshold corrections @xmath92 : @xmath93 that depend on masses of the extra vector - like fermions .
these threshold effects are well approximated by the leading logarithmic corrections : ) and eq .
( [ eq : sol_2loop ] ) below their mass .
it is an excellent approximation for @xmath89 and sufficient approximation for @xmath91 since , for the ir value of @xmath91 , the 1-loop term in the rg equation dominates . ]
@xmath94 where @xmath95 is the contribution of a given fermion @xmath96 , with mass @xmath97 , to the corresponding beta function coefficient @xcite . for particles originating from vector - like families , these contributions , summarized in table [ tab : b ] , are identical to contributions from fermions in the standard model .
the contribution from the complete family is identical to all three beta function coefficients and equal to 4/3 for a chiral family , and 8/3 for a vector - like pair ( @xmath98 in the so(10 ) language ) .
.quantum numbers and contributions to beta function coefficients of particles from extra vector - like families .
the names are chosen to mimic those of the standard model particles with the same quantum numbers . for each particle
, there is a corresponding vector - like partner , and its contributions to the beta function coefficients are identical .
the @xmath99 coefficients correspond to the su(5 ) normalization of the hypercharge . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] [ tab : b ]
the correction to @xmath53 of about @xmath100 is crucial in order to reproduce the measured value . as can be seen in figs .
[ fig : thresholds ] and [ fig : m_g_vs_m_vf_3vf ] , it is indeed @xmath53 that determines @xmath101 gev and , consequently , @xmath102 gev .
the other two couplings are within 10% from measured values in much larger ranges of @xmath23 and would actually prefer smaller @xmath23 and @xmath38 . out of the three parameters ,
the ew scale values of gauge couplings are the most sensitive to changes in @xmath23 . however , since @xmath23 is only responsible for at most @xmath103 of the ew scale values of couplings , the overall sensitivity is still very small .
most of the ew scale values of couplings originate from the ir fixed point . since no precise cancellations between separate contributions are required , there are large ranges of fundamental parameters from which the predicted values of gauge couplings at the ew scale are close to observed values .
but for the sm extended by 4 vector - like families . , title="fig:",width=278 ] but for the sm extended by 4 vector - like families . , title="fig:",width=278 ]
the standard model extended by 4 vector - like families ( sm+4vfs ) allows for insensitive unification of gauge couplings in a similar way as the sm+3vf .
predicted values of gauge couplings at @xmath35 as functions of the gut scale and the universal mass of fermions from 4 vector - like families for fixed @xmath21 are shown in fig .
[ fig : m_g_vs_m_vf_4vf ] .
there are , however , notable differences from the sm+3vfs case .
first of all , the common mass of vector - like families moves to @xmath104 gev .
this is easily understood from the fact that more matter makes gauge couplings run faster , and thus the vfs must stop contributing to rg evolution at a higher scale ; otherwise , the ew scale values of gauge couplings would be too small .
second of all , the 1-loop ir fixed point value of @xmath105 is 0.234 which is larger than in the sm+3vfs case and actually very close to the measured value .
overall , this however does not make the predictions much better than in the sm+3vfs case , since @xmath53 requires @xmath23 larger than the one needed to reach the measured value of @xmath105 . finally , as a result of larger masses of vfs required in the sm+4vfs scenario ,
the sensitivity of ew scale values of gauge couplings to fundamental parameters increased , which is visible in fig .
[ fig : m_g_vs_m_vf_4vf ] as narrower @xmath72 bands compared to those in fig .
[ fig : m_g_vs_m_vf_3vf ] corresponding to the case of sm+3vfs .
it is easy to extrapolate to a larger number of vfs . increasing the number of vfs requires larger @xmath23 closer and closer to the gut scale .
the sensitivity of predicted values of gauge couplings to fundamental parameters is increasing and approaching the sensitivity in the sm . in the sm extended by 1 or 2 vector - like families the predictive power
is lost , since the unified gauge coupling is small and its specific value is crucial for predictions for gauge couplings at the ew scale in a similar way as in the sm .
the difference form the sm is that the exact unification of gauge couplings is now possible with split masses of vfs .
we will include these solutions as a curiosity in the next subsection .
let us now turn our attention to precise predictions for gauge couplings rather than a @xmath010% agreement .
for this we need to consider threshold effects from splitting masses of vfs .
the necessity to split masses of particles from extra 3vfs is indicated in fig .
[ fig : thresholds ] ( left ) by slightly different scales at which the rg evolutions of gauge couplings in the sm and sm+3vfs cross .
for the example in this figure the crossing scales for @xmath20 , @xmath19 , and @xmath53 are @xmath106 tev , @xmath107 tev , and @xmath108 tev .
these scales determine threshold corrections @xmath109 , see eq .
( [ eq : ti ] ) , required for gauge coupling unification .
any spectrum that leads to required threshold corrections will reproduce the measured values of gauge couplings .
the crossing scales are increasing with increasing @xmath38 and depend very little on @xmath47 for @xmath110 . for different values of @xmath38 , they can be read out of fig .
[ fig : m_g_vs_m_vf_3vf ] ( left ) as corresponding values of @xmath23 for which we obtain the measured value of given gauge coupling .
similarly , in the case of sm+4vfs , the values of @xmath24 can be read out of fig .
[ fig : m_g_vs_m_vf_4vf ] ( left ) . for values of @xmath38
not shown , or for other scenarios , the crossing scales can be easily calculated from rg equations as functions of @xmath38 and @xmath47 . in general , for @xmath16 pairs of vector - like families , once we know values of crossing scales @xmath24 for chosen gut scale , the masses of fermions must satisfy : @xmath111 in order to get exact gauge coupling unification at given gut scale . in the case of universal masses of particles with the same quantum numbers , _
@xmath112 , these mass rules can be written in a simple form : @xmath113 inserting the beta function coefficients from table [ tab : b ] , we find : @xmath114 these formulas hold for any number of complete vector - like families , only the values of @xmath24 depend on the specific scenario . in the case of non - universal masses of particles with the same quantum numbers the above formulas are still valid with the replacement : @xmath115 from eqs .
( [ eq : m3rule ] ) - ( [ eq : split ] ) we can immediately see that splitting fermions with the same quantum numbers does not help to find a solution if a solution does not exist with universal masses .
thus , it is sufficient to assume universal masses , @xmath116 , of particles with the same quantum numbers , and eqs .
( [ eq : m3rule ] ) - ( [ eq : m1rule ] ) classify possible solutions .
in addition , for each solution with universal masses , there are other solutions with split masses , and the only constraint is that their geometric mean is the universal mass needed for a given solution .
there are many solutions available , since we have 5 different masses that have to satisfy three conditions ( [ eq : m3rule ] ) - ( [ eq : m1rule ] ) .
however , it is not guaranteed that for a given gut scale there is a phenomenologically viable solution .
clearly , the crossing scales have to be above the ew scale , and even then the solution might require new fermions below experimental limits or some fermions above the gut scale or the planck scale .
is fixed to 0.3 .
smaller values of @xmath38 ( not shown ) can still be consistent with gauge coupling unification for smaller @xmath47 .
the spectrum shown is just an example , it is not unique . ]
representative examples of the spectrum for various values of @xmath38 in the case of sm+3vfs are given in fig .
[ fig : spectrum3vf ] .
the value of @xmath47 is fixed to 0.3 ; however , the spectrum is not very sensitive to this choice as previously discussed .
the spectrum shown is just an example , motivated by the smallest splitting between masses required , and it is not unique . a specific example with exact numerical values was also given in ref .
the gut scale motivated by the lowest splitting required between masses of vector - like fermions is at @xmath117 gev in agreement with what is suggested in fig .
[ fig : m_g_vs_m_vf_3vf ] ( left ) , and the masses are split between @xmath01 tev and @xmath0100 tev .
there is a lower bound on the possible gut scale at @xmath118 gev . for smaller @xmath38 the crossing scale @xmath119
is too small , see fig .
[ fig : m_g_vs_m_vf_3vf ] ( left ) , and thus a phenomenologically viable solution does not exist . with increasing @xmath38 , the splitting of fermion masses is increasing , which can also be inferred from a larger splitting of crossing scales
. the gut scale can be as high as the planck scale .
however , in that case the masses of vector - like fermions have to be split over 6 orders of magnitude .
note that quark doublets , @xmath10 , are typically predicted at @xmath01 tev .
the preference for @xmath10 being the lightest of vector - like fermions can be understood from @xmath120 . however , there are also solutions with @xmath121 being the lightest .
keep in mind , however , that these masses represent geometric means of masses of particles with the same quantum numbers . therefore , when considering split masses of fermions with the same quantum numbers , any fermion can be the lightest one and as light as current experimental limits . for sm+4vfs ,
examples of the spectrum are given in fig .
[ fig : spectrum4vf ] .
the main features are very similar to the case of sm+3vfs .
the gut scale motivated by the lowest splitting required between masses of vector - like fermions is also at @xmath117 gev , in agreement with what is suggested in fig .
[ fig : m_g_vs_m_vf_4vf ] ( left ) , and about two orders of magnitude splitting of masses of vector - like fermions is required .
the main difference from the sm+3vfs case is that the spectrum shifted to @xmath122 gev .
but in the case of sm+4vfs .
] for completeness , we also include examples of the spectrum needed for exact gauge coupling unification in the case of sm+1vf in fig .
[ fig : spectrum1vf ] and sm+2vfs in fig .
[ fig : spectrum2vf ] . for these cases ,
the ew scale values of gauge couplings are highly sensitive to @xmath47 .
thus @xmath47 in these examples is not fixed but rather optimized for the given @xmath38 . in both cases ,
the exact unification can be achieved even in the region consistent with limits on proton lifetime .
however , the required splitting between masses of vector - like fermions is sizable .
but in the case of sm+1vf . in this case , values of @xmath47 are optimized for given gut scale , and are close to 0.03 for all @xmath38 shown . for smaller or larger values of @xmath38 ,
the unification is not possible for any spectrum.,width=336 ] but in the case of sm+2vf . in this case
values of @xmath47 are optimized for given gut scale , and vary between 0.042 and 0.048 .
smaller values of @xmath38 ( not shown ) can still be consistent with gauge coupling unification . ]
the mass rules we have just derived can be generalized to any extension of the sm .
the existence of crossing scales is a necessary condition for achieving gauge coupling unification in a given model .
this follows from the fact that integrating out extra fields above the ew scale can only increase gauge couplings at the ew scale .
therefore , values of predicted couplings at the ew scale without considering the mass effect of extra matter fields have to be smaller than the measured values .
the crossing scales depend only on @xmath47 and @xmath38 .
thus , requiring that the crossing scales exist leads to limits on possible values of the gut scale and @xmath47 . for chosen @xmath47 and @xmath38 , we can find the crossing scales @xmath24 for all three gauge couplings .
if one loop rges are good approximations , these crossing scales can be easily found by applying eq .
( [ eq : sol_1loop ] ) separately between @xmath35 the @xmath123 scales using the sm beta function coefficients , @xmath124 , starting with the observed values of @xmath125 , and between @xmath126 and @xmath38 scales using beta function coefficients in the given extension , @xmath11 , assuming gauge couplings exactly unify .
we get : @xmath127 the meaning of crossing scales is the same as in extensions of the sm with vfs ; namely , they represent the threshold corrections , @xmath128 , that masses of extra particles must generate in order to reproduce the measured values of gauge couplings starting from the given @xmath47 and @xmath38 .
the rest follows what we did for complete vfs .
once we know values of crossing scales @xmath24 , in order to get exact gauge coupling unification , the masses of extra particles must satisfy : @xmath129 where the product is over all extra fermions ( or scalars ) charged under a given gauge symmetry . for a vector - like pair of fermions , the corresponding mass on the right - hand side appears twice . as in the case of complete vfs , it is sufficient to consider the universal mass of all particles with the same quantum numbers .
the universal mass that enters eq .
( [ eq : mrulegen ] ) represents their geometric mean . for any model with an arbitrary particle content , the crossing scales ( [ eq : crossing ] ) as functions of @xmath47 and @xmath38 together with the mass rules ( [ eq : mrulegen ] )
classify all the solutions consistent with gauge coupling unification in terms of physical masses of extra particles .
let us illustrate the usefulness of crossing scales and the mass rules on one example .
let us ask if there is any spectrum of extra particles in the sm extended by one vector - like family that leads to exact gauge coupling unification for @xmath130 gev .
this choice corresponds to one of the points in fig .
[ fig : spectrum1vf ] , and so we already have the answer we can compare with
. however , this answer is obtained by a fairly complicated numerical procedure that iteratively solves coupled differential equations with masses of extra vector - like fermions varied untill the ew scale values of gauge couplings are precisely reproduced . using our method
, we can get the basic features of the required spectrum fast . for @xmath131 that corresponds to
the given example in fig .
[ fig : spectrum1vf ] , the crossing scales @xmath24 , easily calculated from eq .
( [ eq : crossing ] ) , are @xmath132 gev , @xmath133 gev , and @xmath134 gev . this immediately tells us that there will be more than 8 orders of magnitude splitting between masses required . knowing the crossing scales
, we can easily see the basic features of the spectrum that will work . from eqs .
( [ eq : mrulegen ] ) , which in this case are the same as eqs .
( [ eq : m3rule ] ) - ( [ eq : m1rule ] ) , we see that @xmath135 , which heavily weighs on @xmath119 , should be less than @xmath119 , while everything with large hypercharge ( especially @xmath136 and @xmath137 ) should be above @xmath138 in order to find a solution . for a specific example , one can choose two masses and calculate the rest of the spectrum from eqs .
( [ eq : m3rule ] ) - ( [ eq : m1rule ] ) . given a large splitting between @xmath119 and @xmath138 in this case
, it would be easiest to choose the masses of q and u as a starting point .
once we have one solution , varying the starting masses of q and u and calculating the rest of the masses from eqs .
( [ eq : m3rule ] ) - ( [ eq : m1rule ] ) will give us all possible solutions for the given @xmath47 and @xmath38 .
this procedure can be repeated for any @xmath47 and @xmath38 , or the solutions can be plotted as functions of these variables .
so far , we have only considered constraints on the gut scale and masses of vector - like fermions from gauge coupling unification . in order for this scenario to be easily embedded into simple grand unified theories , based on @xmath4 or @xmath6 , the constraints on proton lifetime and the stability of the ew minimum of the higgs potential should be satisfied .
the most stringent limits on proton lifetime come from super - kamiokande . for the dominant decay mode from dimension-6 operators ,
the limit is @xmath139 yrs @xcite .
assuming naively , that the proton lifetime is @xmath140 , where @xmath141 is the mass of the proton , this limit translates into the lower bound on the gut scale : @xmath142 gev for @xmath21 which we use in our examples .
however , the prediction for the proton lifetime is somewhat model dependent ( see for example refs .
@xcite and references therein ) , and so we do not impose the strict limit in the plots we present . in addition , the plots would look very similar for any large value of @xmath47 , but the limits would differ .
the interested reader can easily impose the limit on any scenario by simple rescaling of the mentioned limit using the formula for the proton lifetime .
it is interesting to note that the best motivated value of the gut scale is in the @xmath0@xmath1 gev range which is basically at the current limit .
it is , however , not possible to make precise predictions without knowing the masses of vector - like fermions .
for example , a scenario with the gut scale larger by a factor of 3 results in @xmath02 orders of magnitude enhancement of the proton lifetime but would only require modest changes in the spectrum of vector - like fermions in order to have exact gauge coupling unification .
this inability to make precise predictions of gut scale parameters is a direct consequence of the insensitivity of the ew scale couplings to gut scale boundary conditions .
the rg evolution of the top yukawa and higgs quartic couplings in the sm+3vfs for @xmath143 gev and @xmath144 was given in ref .
the higgs quartic coupling remains positive all the way to the gut scale , and thus the electroweak minimum of the higgs potential is stable .
this result holds in a large range of @xmath38 and @xmath47 , especially in the best motivated region .
therefore , these scenarios represent some of the simplest possible extensions of the standard model that can be embedded into grand unified theories , with a sufficiently long lived proton , and a stable ew minimum of the higgs potential .
we have not investigated the origin of masses of vector - like fermions needed for gauge coupling unification .
this would require additional assumptions about the mechanism that generates them and the scale at which boundary conditions are set .
the masses of vector - like fermions may be fundamental lagrangian parameters , or they can originate from yukawa couplings to one or several additional scalars ( singlets under sm gauge symmetry , but possibly charged under family symmetries ) that acquire vacuum expectation values at any scale between the gut scale and the @xmath23 scale .
in addition , vector - like fermions can have non - zero yukawa couplings to the sm higgs doublet , which add another layer of complexity by contributing to the physical masses and possibly significantly affecting the rg evolution of other parameters that directly determine their masses .
the study of gauge coupling unification is , to a large extent , unaffected by these assumptions , only the physical masses of particles matter in the leading order .
fundamental lagrangian masses would not affect the running of gauge couplings at all , and the yukawa couplings to extra scalars may only contribute to the rg evolution of gauge couplings at 2-loop level .
however , in any specific scenario , the mass rules ( [ eq : m3rule ] ) - ( [ eq : m1rule ] ) can be evolved to the gut scale ( or other relevant scale ) , and the freedom to choose some of the masses can be used to search for simple boundary conditions that are consistent with gauge coupling unification . finally , it is intriguing to consider a connection with the anthropic solution to the hierarchy problem , or the ew scale @xcite . adding vfs to the sm
makes this possibility more appealing , since in the sm , special values of gauge couplings either at the ew scale or some high scale have to be selected . in scenarios that we discussed , the ew scale values of gauge couplings close to the observed values are not very special , but rather quite a generic outcome from large ranges of fundamental parameters . for example , in the sm+3vfs case , as far as @xmath145 tev , for any @xmath146 , and @xmath38 anywhere between @xmath147 gev and the planck scale , the predicted values of gauge coupling at the ew scale are always within 50% of the measured values
( @xmath89 typically well within 20% ) .
this is indicated by the lightly shaded region in fig .
[ fig : m_g_vs_m_vf_3vf ] , and a similar region is indicated in fig . [
fig : m_g_vs_m_vf_4vf ] for sm+4vfs .
furthermore , if the ew scale and @xmath23 have the same origin , it would also explain the proximity of the qcd scale to the ew scale .
the beta function of @xmath53 changes the sign at @xmath23 , and below this scale , it starts running fast toward @xmath148 .
we have discussed gauge coupling unification in models with additional 1 to 4 complete vector - like families . in scenarios with 3 or more vector - like families
the values of gauge couplings at the electroweak scale are highly insensitive to the grand unification scale , the unified gauge coupling , and the masses of vector - like fermions .
their observed values can be mostly understood from infrared fixed point behavior . starting with a large ( but still perturbative ) unified gauge coupling at a high scale ,
the values of gauge couplings at lower energies are determined only by the particle content of the theory and how far from the gut scale we measure them . since the exact value of @xmath47 becomes irrelevant , instead of one prediction of the conventional unification , we have two predictions for ratios of gauge couplings .
these predictions are modified at the @xmath46 scale , where the vector - like fermions are integrated out , and below this scale , gauge couplings run according to the usual rg equations of the standard model .
assuming first a common mass of vector - like fermions , @xmath23 , we showed predictions for three gauge couplings at the ew scale as functions of @xmath38 and @xmath23 .
we found that the observed values of gauge coupling are reproduced with good precision from a large range of parameters .
especially , @xmath38 can be varied over several orders of magnitude while having all three gauge couplings within 20% from observed values .
the best fit , which predicts all three couplings within 6% from measured values , suggests @xmath149 gev , @xmath101 gev in the case of sm+3vfs , and @xmath150 gev in the case of sm+4vfs . the best motivated gut scale , @xmath0@xmath1 gev , predicts a proton lifetime close to current limits .
however , due to insensitivity of the predicted ew scale values of gauge couplings to gut scale parameters , no sharp predictions can be made without knowing the spectrum of vector - like fermions .
in addition , it was previously shown that the higgs quartic coupling remains positive all the way to the gut scale , and thus the electroweak minimum of the higgs potential is stable .
this result holds in a large range of @xmath38 and @xmath47 , especially in the best motivated region .
therefore , these scenarios represent some of the simplest possible extensions of the standard model that can be embedded into grand unified theories , with a sufficiently long lived proton , and the stable ew minimum of the higgs potential .
the discrepancies of ir fixed point predictions from observed values can be explained by threshold effects of extra vector - like fermions .
we showed examples of the spectrum for the gut scale varied between @xmath147 gev and @xmath151 gev .
we derived simple rules for masses of vector - like fermions required for exact gauge coupling unification .
in addition , we generalized the mass rules and the method of using crossing scales of evolutions of gauge couplings in the sm and the given extension to classify scenarios consistent with gauge coupling unification to an arbitrary extension of the standard model .
the problem of finding all possible mass spectra in a given model consistent with gauge coupling unification is reduced to solving a set of simple algebraic equations that masses of extra particles have to satisfy .
with respect to the sensitivity to fundamental parameters , the model with 3 extra vector - like families stands out . in the best motivated region , it requires vector - like fermions with masses of order 1 tev 100 tev , and thus at least part of the spectrum may be within the reach of the lhc .
notably , quark doublets , @xmath10 , are typically predicted at @xmath01 tev .
however , only geometric means of masses of particles with the same quantum numbers are constrained by gauge coupling unification .
therefore , when considering split masses of fermions with the same quantum numbers , any fermion can be the lightest one and as light as current experimental limits . besides direct production of these particles at the lhc
, it may be also possible to observe their effects in a variety of processes .
however , they typically affect standard model predictions only through mixing with light fermions , which is highly model dependent .
the discussion of gauge coupling unification that we focused on here is negligibly affected by such mixing .
* acknowledgments : * r. d. thanks h.d .
kim and l. hall for useful comments and discussions .
r. d. also thanks cern theory institute , the center for theoretical underground physics and related areas ( cetup * 2012 ) , and the galileo galilei institute for theoretical physics for hospitality and for partial support during various stages of this work .
this work was supported in part by the department of energy under grant number de - fg02 - 91er40661 .
v. agrawal , s. m. barr , j. f. donoghue and d. seckel , phys .
d * 57 * , 5480 ( 1998 ) [ hep - ph/9707380 ] .
v. agrawal , s. m. barr , j. f. donoghue and d. seckel , phys .
lett . *
80 * , 1822 ( 1998 ) [ hep - ph/9801253 ] . | we discuss gauge coupling unification in models with additional 1 to 4 complete vector - like families , and derive simple rules for masses of vector - like fermions required for exact gauge coupling unification .
these mass rules and the classification scheme are generalized to an arbitrary extension of the standard model .
we focus on scenarios with 3 or more vector - like families in which the values of gauge couplings at the electroweak scale are highly insensitive to the grand unification scale , the unified gauge coupling , and the masses of vector - like fermions .
their observed values can be mostly understood from infrared fixed point behavior . with respect to sensitivity to fundamental parameters ,
the model with 3 extra vector - like families stands out .
it requires vector - like fermions with masses of order 1 tev 100 tev , and thus at least part of the spectrum may be within the reach of the lhc .
the constraints on proton lifetime can be easily satisfied in these models since the best motivated grand unification scale is at @xmath0@xmath1 gev .
the higgs quartic coupling remains positive all the way to the grand unification scale , and thus the electroweak minimum of the higgs potential is stable . |
coorbital systems are those in which at least two bodies share the same mean orbit .
coorbital objects that librate around the lagrangian points @xmath8 or @xmath9 are said to be in tadpole orbits , while those that librate around @xmath8 , @xmath10 and @xmath9 are said to be in horseshoe type orbits .
although lagrange have described the motion of bodies around these equilibrium points in 1788 , when he published his _ analytical mechanics _ , only in 1906
a body showing this kind of motion was discovered by max wolf in heidelberg .
he found an asteroid librating around the point @xmath8 of jupiter in the system sun - jupiter .
this asteroid is ( 588 ) achiles and at the moment there are more than 3200 know asteroids coorbiting with jupiter ( http : + www.cfa.harvard.edu/iau/lists/jupitertrojans.html ) .
apart from jupiter there are other planets with coorbital bodies . in 1991
the asteroid 5261 eureka was discovered librating around the @xmath9 of mars ( innanen , 1991 ) .
later three more asteroids were discovered : 1998 vf@xmath11 ( tabachnik & evans , 1999 ) , 1999 uj@xmath12 ( connors et al . , 2005 ) and 2007 ns@xmath13 ( http : + www.cfa.harvard.edu/iau/lists/marstrojans.html ) .
neptune has six coorbital asteroids .
all of them librating around its lagrangian point @xmath8 ( zhou et al . ,
2008 ) . on the other hand , all the coorbital satellite systems known are around the planet saturn ( table 1).the satellite thetys has two coorbital satellites .
telesto around its @xmath8 point and calypso around its @xmath9 point .
they were discovered in 1980 by observations made from earth ( veillet , 1981 ; reitsema , 1981 ) .
the satellite dione also has two coorbital satellites .
helene around its @xmath8 point and polydeuces around its @xmath9 point .
helene was discovered in 1980 by observations made from earth ( lecacheux et al . , 1980 ; reitsema et al . , 1980 ) and
polydeuces was discovered through images from the cassini spacecraft ( murray et al .
, 2005 ; porco et al . ,
2005 ) . in 1981
, images from voyager not only confirmed the existence of the coorbital system janus and epimetheus but also determined their masses and orbital elements . in a suitable
rotating coordinate system , both satellites perform horseshoe orbits and reach a distance of only 15000 km from each other ( yoder et .
al , 1983 ) . [ cols="<,^,^,^,^,^",options="header " , ] dermott & murray ( 1981a , b ) presented a theory on the tadpole and horseshoe type orbits for the case of circular and elliptic restricted three - body problem .
they also described the coorbital motion of janus and epimetheus through a numerical study combined with perturbation theory .
the study was made for a third body of negligible mass .
then , they generalized some of these results for the case when the third body disturbs the other two bodies of the system .
one of the main results found was a relation between the coorbital trajectory and its associated zero velocity curve .
there are several studies on the origin of the trojan asteroids , but not too many on the origin of the coorbital satellites .
the origin of these satellites and trojans asteroids are believed to be associated to one of the following mechanisms ( yoder et al.,1983 ; smith et al.,1981 ) : ( i ) capture by gas drag , ( ii ) congenital formation or ( iii ) rupture of a parent body .
chanut et al .
( 2008 ) studied the mechanism of capture by drag .
they considered several values for the relative mass of the secondary body , from @xmath14 to @xmath15 .
they explored the orbital evolution of planetesimals under orbital decay toward the secondary body orbit .
there are three possible outcomes along such evolution : collision with the secondary body , capture in the coorbital region or crossing the coorbital region toward the central body .
once a planetesimal is captured in a coorbital trajectory , its amplitude of oscillation around one of the triangular lagrangian points shrinks until it is fixed over the lagrangian point .
they also showed that the location of the lagrangian points change according to the density of the gas . with the numerous discoveries of the extrasolar planets
beaug et al .
( 2007 ) analyzed the formation of hypothetical earth - type coorbital planets in extrasolar systems .
the central body was taken as equal to one solar mass and the giant planet with jupiter s mass .
they considered two scenarios of formation , one gas - free scenario and another rich - gas scenario . in both of them
there were the formation of a single coorbital of the terrestrial type , but due to gravitational instability with other bodies , the accretion process was not efficient .
in this paper we study the mechanism of congenital formation of coorbital satellites and in some points this mechanism is similiar to adopted in beaug et al .
according to the modern theories of planetary formation , in the beginning of the solar system formation there was a cloud of gas and dust that settled in a disk .
due to several factors the gas dissipated along the time . in the first stages the particles were of the size of sub - micrometers up to the order of centimeters .
this disk of gas and dust gradually generated larger bodies through physical collisions leading to agglomeration .
successive encounters among them resulted in a large number of meter sized aggregates .
from these objects originated planetesimals of the order of kilometers .
further encounters between planetesimals led to the formation of the planet s progenitors of terrestrial size or even ten times this size ( armitage , 2007 ) .
the formation of planetary satellite systems are thought to have followed a very similar way ( safronov , 1969 ; wetherill 1980 , 1990 ; hayashi et al .
1985 ; greenberg , 1989 ) . in the present work we consider an intermediate stage of saturn s satellite system formation .
in this stage all the major satellites are almost formed ( here called proto - satellites ) , there is almost no gas left , but there is still plenty of small planetesimals in the disk .
our model of study is characterized by a central body ( saturn ) , a secondary body ( proto - satellite ) and a cloud of plantesimals . since it is assumed that the proto - satellite has already a significative mass , which is several orders of magnitude larger than the planetesimals , it defines ( together with the planet ) the structure of the phase space with the five lagrangian points .
then , we consider a cloud of planetesimals in the coorbital region around one of the triangular lagrangian points , under the gravitational influence of saturn and the proto - satellite .
the collision between planetesimals are considered constructive and through this accretion process larger bodies are built in the coorbital region .
this paper has the following structure . in the next section we present the methodology and assumptions adopted in our numerical simulations .
the results of the simulations are presented in section 3 . finally , in the last section we present some final comments and our conclusions .
the dynamical system considered is composed by saturn ( the central body ) , a proto - satellite and a cloud of planetesimals . the mass of the proto - satellite was chosen to be @xmath0 of saturn s mass , which is close to the masses of dione and thetys ( table 2 ) .
it is placed in a circular orbit with the semi - major axis equal to that of thetys .
the mass of the proto - satelite might vary along the simulation due to some collisions with planetesimals , but its small growth do not affect significantly the dynamic of the planetesimals around the lagrangian points ( fleming and hamilton , 2000 ) .
the planetesimal clouds were always initially randomly distributed in a sector around @xmath8 or @xmath9 .
the sector is delimited by an arc of @xmath16 , centered on the lagrangian point , and the extreme orbital radii of the largest horseshoe orbit ( figure 1 ) .
the half width of the largest horseshoe orbit is given by dermott & murray ( 1981a ) as @xmath17 where @xmath18 and @xmath19 are the relative mass and the semi - major axis of the proto - satellite , respectively .
studying the overlap of first order mean motion resonances , wisdom ( 1980 ) found an expression that gives the width of a chaotic region from the planar circular orbit of a secondary body .
therefore , it is important to have in mind that just outside the coorbital region ( interior and exterior ) there is a chaotic region . particles in this region usually tend to have close encounters with the proto - satellite and get large eccentricities , leaving the neighborhood . and @xmath9 were the planetesimals are initially distributed .
each sector is delimited by an arc of @xmath16 , centered on the lagrangian point , and the extreme orbital radii of the largest horseshoe orbit .
the half width of the largest horseshoe orbit , @xmath20 , is given by equation ( 1 ) . ] since we are interested in the formation of the small counterparts of coorbital satellite systems , in general , we considered clouds of planetesimals with the total mass of the order of the smaller coorbital satellites , given in table 1 .
we also considered that all planetesimals have the same initial mass , @xmath21 .
we performed simulations for different clouds of planetesimals .
we adopted clouds of @xmath4 , @xmath5 and @xmath6 planetesimals each one for three different values of @xmath21 , @xmath3 , @xmath2 and @xmath1 masses of saturn .
the simulations were made through numerical integrations using the burlish - stoer integrator from the package mercury ( chambers , 1999 ) .
the length of the integration , @xmath22 ( in orbital periods of the proto - satellite ) , varied according to the value of @xmath21 . for @xmath23 it was used @xmath24 , for @xmath25 it was used @xmath26 and for @xmath27 it was used @xmath28 .
that is necessary due to the gravitational attraction between the planetesimals , which is smaller as smaller are their masses , leading to a longer time for the system to evolve . for each cloud of planetesimals with a specific value of @xmath21 we performed at least two independent simulations .
one placing it around @xmath8 and the other around @xmath9 .
we also tested the effect of saturn s oblateness by running extra simulations with the inclusion of the @xmath29 term in the gravitational potential of the central body .
the planetesimals in all simulations were initially placed in circular orbits .
along the numerical simulations the planetesimals collide among them generating lager planetesimals .
the collisions are always considered constructive .
i.e. inelastic collisions . by the end of the integration time
there are only a few ( at most four ) large planetesimals left .
the final masses of the remaining planetesimals are represented by @xmath30 . in order to have an idea of the time evolution of these systems we present a sample of snapshots for two representative simulations .
figure 2 shows the dynamical evolution of a cloud of @xmath6 planetesimals with @xmath25 , initially distributed around @xmath8 . in each frame
are shown the values of the osculating semi - major axis and eccentricities of the remaining planetesimals at a given moment .
the orbital radius of the proto - satellite is considered as unity .
the color grade indicates the mass of the planetesimal .
since we are interested in the formation of small coorbital satellites , the best candidates should be the remaining planetesimals with low eccentricity and semi - major axis inside the coorbital region .
so we included in the plots the dashed curves that indicate the coorbital region defined by the largest horseshoe width ( @xmath31 ) . in general , the planetesimals that leave the coorbital region present chaotic behavior and collide with the proto - satellite . from the plots of figure 2 we notice that the eccentricities of the planetesimals increase as they start to grow
however , they are limited to low values ( @xmath32 ) . after about 1068 years
there are only three planetesimals left and it takes about a thousand years to reach a single planetesimal . that last one has a mass of @xmath33 in figure 3 we present the dynamical evolution of a cloud of @xmath4 planetesimals with @xmath27 , initially distributed around @xmath9 .
the evolution is very much similar to that presented in figure 2 .
however , at least two major differences should be pointed out .
the eccentricities do not get much larger then 0.001 and the time scales are much longer then in the previous case .
these two features are naturally due to the lower values of the initial mass of the planetesimals , @xmath21 , and the total mass of the cloud of the planetesimals .
larger masses produce stronger gravitational interactions accelerating the evolution of the system and increasing the values of eccentricities .
[ figura 1 ] the evolution of all the other numerical simulations we have performed are similar to these ones .
however , the plots presented in figures 2 and 3 do not guarantee that the remaining planetesimals are in coorbital orbits with the proto - satellite .
therefore , we analyzed the evolution of the angle @xmath34 between the remaining planetesimals and the proto - satellite , called libration angle .
this angle is defined as the difference in mean longitudes , @xmath35 , where @xmath36 and @xmath37 are the mean longitudes of the planetesimal and secondary body respectively .
a representative sample of such evolution is presented in figures 4 to 7 . in each figure we present two plots .
the top plot shows the evolution of @xmath34 and the bottom one shows the evolution the of planetesimal s mass .
all plots cover the whole integration time and they are shown in a logarithmic scale .
the temporal evolution of the libration angle confirmed that in all simulations the remaining planetesimals are in coorbital trajectories .
the plots show a complete variety of evolutions .
planetesimals that started in tadpole orbits around @xmath8 or @xmath9 and ended in horseshoe orbits ( figures 5 and 6 ) .
planetesimals that started in tadpole orbit around @xmath8 and ended in tadpole orbit around @xmath9 ( figure 4 ) or in tadpole orbit around @xmath8 ( figure 7 ) .
the evolutions are not always smooth .
there are cases of tadpole orbits that `` jump '' from @xmath8 to @xmath9 and then again back to @xmath8 ( figure 7 ) . and
others that `` jump '' from @xmath9 to @xmath8 and then to a horseshoe orbit ( figure 6 ) .
the changes in the amplitude of oscillation of the coorbital planetesimals along their evolution are associated to the collisions with other planetesimals .
several of these changes can clearly be seen by comparison between the top and the bottom plots of each figure .
for example , in figure 6 the last visible step in the growth of the planetesimal coincides with the change from tadpole to horseshoe orbit .
planetesimals with @xmath25 , initially distributed around @xmath8 .
after about 60 years the planetesimal reaches its final mass , @xmath38 . and starts to librate around @xmath9 . ] planetesimals with @xmath23 , initially distributed around @xmath9 . by the end of the simulation the planetesimal reaches its final mass , @xmath39 and
stays in a horseshoe orbit . ]
planetesimals with @xmath25 , initially distributed around @xmath8 . by
the end of the simulation the planetesimal reaches its final mass , @xmath40 and stays in a horseshoe orbit . ]
planetesimals with @xmath27 , initially distributed around @xmath8 .
after about 51 thousand years the planetesimal reaches its final mass , @xmath41 and starts to librate in a tadpole orbit around @xmath8 .
] a summary of the whole set of simulations are presented in tables 3 and 4 .
we performed 24 simulations neglecting the oblateness of saturn ( table 3 ) and other 18 simulations taking it into account ( table 4 ) . in the last two columns the kind of coorbital trajectory and the final mass of the survivors planetesimals
are shown .
as we can see , in several simulations there were more then one planetesimal left . in some cases we followed the simulations for longer times .
for example , simulation 13a ( table 3 ) was integrated for more then @xmath42 years and no extra collision occurred along the second half of such integration time . in the case of simulation 14a ( table 3 ) , the integration was extended to one million years and there were still two planetesimals left .
one in horseshoe orbit and other in tadpole around @xmath9 .
from table 3 we have that 2/3 of the survivors are left in horseshoe trajectories , while the remaining 1/3 are equally divided in tadpole trajectories around @xmath8 and @xmath9 .
we also note that , in each simulation , the total mass of the survivors is very close to the corresponding total mass of initial cloud of planetesimals .
the difference between these two values is the mass of planetesimals that collided with the proto - satellite . comparing the simulations presented in table 3 with those presented in table 4
, we can note that the oblateness of saturn did not introduce any significant change in the results as a whole .
@lcccccc@ + simulation & number of & location of & integration & @xmath21 & trajectories of & final masses + & plantesimals & initial distribution & time ( years ) & & the survivors & @xmath30 + 1 & 10000 & @xmath8 & 10005 & @xmath1 & h & @xmath43 + 2 & 10000 & @xmath9 & 20238 & @xmath1 & h & @xmath43 + 3 & 10000 & @xmath8 & 1068 & @xmath2 & @xmath44 & @xmath45 + 4 & 10000 & @xmath9 & 99 & @xmath2 & h & @xmath46 + 5 & 10000 & @xmath8 & 51 & @xmath3 & h & @xmath47 + 6 & 10000 & @xmath9 & 50 & @xmath3 & h & @xmath48 + 7 & 5000 & @xmath8 & 51000 & @xmath1 & h , @xmath49 & @xmath50 , @xmath51 + 8 & 5000 & @xmath9 & 51000 & @xmath1 & h , @xmath44 & @xmath52 , @xmath53 + 9 & 5000 & @xmath8 & 3325 & @xmath2 & h & @xmath54 + 10 & 5000 & @xmath9 & 989 & @xmath2 & h & @xmath55 + 11 & 5000 & @xmath8 & 52 & @xmath3 & h & @xmath56 + 12 & 5000 & @xmath9 & 61 & @xmath3 & @xmath44 & @xmath57 + 13a & 1000 & @xmath8 & 51000 & @xmath1 & h , h , @xmath49 & @xmath58 , @xmath59 , @xmath60 + 13b & 1000 & @xmath8 & 51000 & @xmath1 & @xmath49 , h , h & @xmath61 , @xmath62 , @xmath63 + 14a & 1000 & @xmath9 & 51000 & @xmath1 & h , @xmath44 , h , h & @xmath64 , @xmath65 , @xmath66 , @xmath67 + 14b & 1000 & @xmath9 & 51000 & @xmath1 & h , h , h , @xmath44 & @xmath68 , @xmath69 , @xmath70 , @xmath71 + 15a & 1000 & @xmath8 & 5100 & @xmath2 & h , @xmath49 & @xmath72 , @xmath73 + 15b & 1000 & @xmath8 & 5100 & @xmath2 & @xmath49 , h & @xmath74 , @xmath75 + 16a & 1000 & @xmath9 & 5100 & @xmath2 & @xmath49 , h , @xmath44 & @xmath76 , @xmath77 , @xmath78 + 16b & 1000 & @xmath9 & 5100 & @xmath2 & h , @xmath44 & @xmath79 , @xmath80 + 17a & 1000 & @xmath8 & 510 & @xmath3 & @xmath49 , h & @xmath81 , @xmath82 + 17b & 1000 & @xmath8 & 229 & @xmath3 & h & @xmath83 + 18a & 1000 & @xmath9 & 192 & @xmath3 & h & @xmath84 + 18b & 1000 & @xmath9 & 277 & @xmath3 & h & @xmath7 + @lcccccc@ + simulation & number of & location of & integration & @xmath21 & trajectories of & final masses + & plantesimals & initial distribution & time ( years ) & & the survivors & @xmath30 + 1 & 10000 & @xmath8 & 51000 & @xmath1 & h , h & @xmath85 , @xmath86 + 2 & 10000 & @xmath9 & 14672 & @xmath1 & h & @xmath43 + 3 & 10000 & @xmath8 & 654 & @xmath2 & h & @xmath87 + 4 & 10000 & @xmath9 & 218 & @xmath2 & h & @xmath88 + 5 & 10000 & @xmath8 & 55 & @xmath3 & h & @xmath89 + 6 & 10000 & @xmath9 & 59 & @xmath3 & h & @xmath90 + 7 & 5000 & @xmath8 & 51000 & @xmath1 & h , @xmath49 & @xmath91 , @xmath92 + 8 & 5000 & @xmath9 & 51000 & @xmath1 & h , @xmath49 & @xmath93 , @xmath94 + 9 & 5000 & @xmath8 & 2308 & @xmath2 & h & @xmath95 + 10 & 5000 & @xmath9 & 1663 & @xmath2 & h & @xmath96 + 11 & 5000 & @xmath8 & 95 & @xmath3 & h & @xmath97 + 12 & 5000 & @xmath9 & 49 & @xmath3 & h & @xmath98 + 13 & 1000 & @xmath8 & 51000 & @xmath1 & h , h , h , @xmath49 & @xmath99 , @xmath58 , @xmath100,@xmath101 + 14 & 1000 & @xmath9 & 51000 & @xmath1 & h , h , @xmath44 & @xmath102 , @xmath103 , @xmath104 + 15 & 1000 & @xmath8 & 29768 & @xmath2 & h & @xmath105 + 16 & 1000 & @xmath9 & 51000 & @xmath2 & @xmath44 , h & @xmath106 , @xmath107 + 17 & 1000 & @xmath8 & 218 & @xmath3 & h & @xmath108 + 18 & 1000 & @xmath9 & 173 & @xmath3 & h & @xmath109 + beaug et al .
( 2007 ) found that the final mass of the formed trojans the kind of terrestrial planets did not depend significantly neither on the initial mass , nor on the number of the initial population of planetesimals . in our simulations
we do realize that the number of planetesimal and its initial mass play an important role in the final outcome . in table 3
, we can see that the initial populations with small mass ( @xmath2,@xmath1 ) and small number of planetesimal ( 1000 ) have a larger tendency to yield more than one final body .
this effect is caused by the weak gravitational attraction when the planetesimals are smaller and more distant from each other , consequently few collisions occur .
however , when we have a lot of planetesimals , even with small mass , they are closer , then the gravitational attraction is larger and collisions become more frequent .
although in most cases the final total mass is almost equal to the initial value , in one case there is a significant mass loss ( table 3 , simulation 3 ) .
this occurred due to a collision of a larger planetesimal with the secondary body .
however , for a similar initial planetesimal population distributed around @xmath9 there is no mass loss , this loss mass might be one statistical fluke , but this has to be checked with further studies .
the temporal evolution of the total number of the remaining planetesimals , @xmath110 , in the system gives an idea of the different speeds at each stage of the system evolution . in figure 8
we present a representative sample of the temporal evolution of the total number of the remaining planetesimals of our simulations .
in all the cases studied , the evolution of the system can be well represented by three stages . in the first stage
there is a `` cold '' cloud , the planetesimals are in near circular orbits and do not collide frequently .
then , along the time the gravitational interaction among the planetesimals `` heat '' the cloud , the trajectories get some eccentricity and the rate of collisions increase , producing larger planetesimals .
larger planetesimals excite the cloud even more and also increase their sphere of influence , resulting in larger cross sections and faster growing .
when most of the planetesimals collided , there are a few left , then the collision rate decreases significantly and the evolution of the system slows down .
a comparison of the curves presented in figure 8 shows that the rate of decay of the total number of planetesimals is strongly dependent on the total mass of the initial cloud of planetesimals .
( top ) , @xmath5 ( middle ) and @xmath6 ( bottom ) planetesimals around @xmath8 with initial masses @xmath27 , @xmath25 and @xmath23 , title="fig : " ] ( top ) , @xmath5 ( middle ) and @xmath6 ( bottom ) planetesimals around @xmath8 with initial masses @xmath27 , @xmath25 and @xmath23 , title="fig : " ] ( top ) , @xmath5 ( middle ) and @xmath6 ( bottom ) planetesimals around @xmath8 with initial masses @xmath27 , @xmath25 and @xmath23 , title="fig : " ] in order to infer possible effects due to mean motion resonances with nearby satellites on the mechanism of coorbitals formation , we performed simulations considering mimas and thetys in 4:2 resonance of the type - inclination .
we considered a system composed of an oblate central body ( saturn ) , two satellites ( mimas - thetys ) which are in mean - motion resonance 4:2 and a cloud of planetesimals .
the simulations were made for six cloud of 1000 planetesimal considering in each case @xmath21 equal to @xmath3 , @xmath2 or @xmath1 around @xmath8 and @xmath9 .
it is important to have in mind that the previous simulations were in a planar system , and when we considered mimas and thetys in a 4:2 inclination - type resonance , the problem becomes three - dimensional . in general , these new simulations did not show any significant difference in comparison with the simulations presented in table 4 . in figure 9
we present a comparison of the temporal decaying number of the planetesimals for the simulations of a cloud of @xmath4 planetesimals initially located around @xmath8 and with @xmath23 .
it shows that only in the final stages ( after reducing the number of planetesimals to about 30% ) the results present some notable difference .
the results under resonant perturbation evolves more slowly than the system without such perturbation . probably , this is only due to the fact that the system under resonant perturbation is in a three - dimensional space , which reduces the chances of encounters between the planetesimals in comparison with the planar system case . with initial masses
@xmath23 initially distributed around @xmath8 .
the red and dashed line correspond to a simulation only with thetys and the blue and solid line correspond to a simulation with thetys and mimas in mean motion resonance 4:2 . ]
in the present work we present results of numerical simulations in order to study the viability of the congenital formation approach as a mechanism for the origin of the coorbital satellites of thetys and dione .
we considered that the collisions are always inelastic and make the growth of the planetesimals .
the results show that almost all the initial mass of the clouds are converted in one or a few planetesimals in coorbital motion with the proto - satellite .
the final masses are of the same order of the small coorbital satellites helene , polydeuces , telesto and calypso .
therefore , we found that this is a promising mechanism for this purpose .
the distribution of librational amplitudes of the final bodies that remain in tadpole orbits varies from 4 up to 75 degrees.the real coorbitals have a libration amplitude of which varies from 1.3 up to 25.9 degrees .
however , some mechanisms could reduce the amplitude of libration , for example , the gas drag ( chanut et al . , 2008 ) , mass accretion or radial migration of the secondary body ( fleming and hamilton , 2000 ) . in this paper
we do not consider these mechanisms , but might be included in future works .
the effects of mean motion resonances of nearby satellites shows to be of no significative relevance for the present model of congenital formation of coorbitals .
the results presented here might look like obvious .
however , in beaug et al ( 2007 ) it was found that a growing planetary embryo generates a chaotic region around the lagrangian point , inhibiting additional accretion and setting a limit in the final mass of the coorbital body .
they studied the possibility of congenital formation of earth like planets in coorbital orbits of exoplanetary systems . in their work
they adopted systems like the jupiter - sun system , with relative mass of @xmath15 , and did not manage to form coorbital planets with mass larger than @xmath111 this phenomena does not occur in the present work because the results of the simulation are strongly dependent on the relative mass of the secondary body ( izidoro et al . , 2010 ) .
it is important to comment that in the present congenital model the final coorbitals formed are of similar masses .
however , dione has one coorbital , helene whose mass is about two orders of magnitude higher than the other coorbital , polydeuces . therefore , the model as considered in the present work is not able to explain by itself alone such configuration .
consequently , there is still plenty of work to be done .
the comments and questions of an anonymous referee helped to improve this paper .
this work was supported by the brazilian agencies capes and cnpq , and by the foundation of the state of s~ ao paulo fapesp .
99 armitage p. j. , 2007 , arxiv : astro - ph/0701485 beaug c. , sndor zs . , rdi b. , sli . , 2007 , a&a , 463 , 359 chambers j.,1999 , mnras , 304 , 793 chanut t. , winter o. c. , tsuchida m. , 2008 , a&a , 481 , 519 christou a. a. , namouni f. , morais m. h. m. , 2007 , icarus , 192 , 106 connors , m. , 2005 , planetary and space science , 53 , 617 dermott s.f .
, murray c.m . , 1981a , icarus , 48 , 1 dermott s.f . , murray c.m . , 1981b , icarus , 48 , 12 fleming j. h. , hamilton d. p. , 2000 ,
icarus , 148 , 479 giorgini , j.d . , yeomans , d. k. , chamberlin , a. b. , chodas , p. w. , jacobson , r. a. , keesey , m. , lieske , j .
h. , ostro , s. j. , standish , e. m. , wimberly , r. n. , 1996 , baas , ( 28 ) greenberg r. , in origin and evolution of planetary and satellite atmosphere ( eds . atreya , s. k. , pollack j. b. , matthews m. s. ) , univ .
arizona press , tucson , 137 hayashi c. k. , nakazawa k. , nakagawa y. , 1985 , in protostar and planets ii ( eds . black d. c. , matthews m. s. ) , univ .
arizona press , tucson , 1100 innanem k. a. , 1991 , j. r. astron .
, 85 , 151 izidoro et al .
, in preparation , 2010 jacobson r. a.,bull .
soc , 36 , 1097(2004 ) lecauchex j. , laques p. , vapillon l. , auge a. , despiau r. , 1980 , icarus , 43 , 111 mouro d. m. , winter o. c. , tokoyama t. , cordeiro r. r. , 2006 , mnras , 372 , 1614 murray c. d. , cooper n. j. , evans m. w. , beurle k. , 2005 , icarus , 179 , 222 porco c. c. , thomas p. c. , weiss j. w. , richardson d. c. , 2005 , science , 307 , 1226 reitsema h. j. , 1981 , icarus , 48 , 140 reitsema h. j. , smith b. a. , larson s. m. , 1980 , icarus , 43 , 116 safronov v. s. , 1969 , nauka moscow , english translation nasa tt f-677 smith b. a. , voyager imaging team , 1981 , science , 212 , 504 tabachnik s. , evans n. w. , 1999 , astrophys . j. , 517 , l63 veillet c. , 1981 , a&a , 102 , l5 wetherill g. w. , 1980 , ann .
astrophys . , 18 , 77 wetherill g. w. , 1990 , ann .
earth planet sci .
, 18 , 205 wisdom j. , 1980 , aj , 85(8 ) //
yoder c. f. , 1979 , icarus , 40 , 341 yoder c.f .
, colombo g. , synnott s.p . , yoder k.a .
, 1983 , icarus 53 , 431 yoder c.f .
, salo h. , 1988 , a.a .
, 205 , 309 zhou l. , dvorak r. , sun y. , 2008 , arxiv:0811.2491v1 [ astro - ph ] | saturn is the only known planet to have coorbital satellite systems . in the present work we studied the process of mass accretion as a possible mechanism for coorbital satellites formation .
the system considered is composed of saturn , a proto - satellite and a cloud of planetesimals distributed in the coorbital region around a triangular lagrangian point .
the adopted relative mass for the proto - satellite was @xmath0 of saturn s mass and for each planetesimal of the cloud three cases of relative mass were considered , @xmath1 , @xmath2 and @xmath3 masses of saturn . in the simulations each cloud of planetesimal was composed of @xmath4 , @xmath5 or @xmath6 planetesimals .
the results of the simulations show the formation of coorbital satellites with relative masses of the same order of those found in the saturnian system ( @xmath2 - @xmath7 ) .
most of them present horseshoe type orbits , but a significant part is in tadpole orbit around @xmath8 or @xmath9 .
therefore , the results indicate that this is a plausible mechanism for the formation of coorbital satellites .
[ firstpage ] planets and satellites : formation - planets and satellites : individual : saturn |
the phenomenon of anomalous magnetoresistance at low temperature in dirty " metals and doped semiconductors was explained by the theory of quantum corrections to the conductivity.@xcite the interference correction to the conductivity gives the main contribution to the negative magnetoresistance in 2d structures at low temperature and low magnetic field . a unique analytical expression for the magnetic field dependence of negative
magnetoresistance has been found in ref .
@xmath0 where @xmath1 , @xmath2 , @xmath3 is the mean free path , @xmath4 is a digamma function , @xmath5 and @xmath6 stand for the elastic scattering and phase breaking time , respectively .
the parameter @xmath7 is equal to unity for the non - interacting case .
this expression was obtained in the diffusion approximation for the isotropic scattering by randomly distributed scatterers with a short - range potential .
nevertheless it is universally used in the analysis of experimental data to extract the phase breaking time and its temperature dependence through the fitting of experimental curves .
it should be pointed out that @xmath8 determined in this way is a fitting parameter rather than the phase breaking time because some deviation of experimental curves from eq.([eq3 ] ) takes place in almost without exception .
this deviation may result from some correlations in distribution of scatterers , scattering anisotropy , or long - range potential fluctuations in real 2d systems .
this may be a possible reason of the saturation of @xmath9 with decreasing temperature as it is seen in some experiments.@xcite in the previous paper we have developed a new approach to the analysis of anomalous magnetoresistance which makes it possible to obtain the information about the statistics of closed paths from the magnetic field dependence of magnetoresistance .
this approach provides the basis for a new method of analysis of negative magnetoresistance due to weak localization suppression . in the present paper
we demonstrate the potentials of this method as it is applied for interpretation of concrete experimental results obtained for gaas / ingaas / gaas quantum wells .
the essence of the method is clear from eq .
( 8) of the previous paper .
@xcite one can see that the fourier transform of negative magnetoresistance is given by @xmath10 @xmath11 is the elementary flux quantum , @xmath12 , @xmath13 and @xmath14 are the area distribution function of closed paths and the area dependence of the average length of closed paths respectively , introduced in section ii of ref .. thus , it is clearly seen from eq .
( [ eq6 ] ) that @xmath15 contains the information on the area distribution function of closed paths @xmath13 , and on the function @xmath14 .
if @xmath16 tends to infinity when @xmath17 , the extrapolation of @xmath18 to @xmath19 gives the value of @xmath20 . to determine the area dependence of @xmath21 we assume that for actual areas @xmath21 is a power function of area , @xmath22
the numerical calculations of the function @xmath23 ( see fig . 4 in ref . ) have shown that this assumption is valid in a wide range of @xmath24 , @xmath16 . in the diffusion approximation (
i.e. for @xmath25 ) the value of @xmath26 is about @xmath27 .
it is clear that in this approximation the value of @xmath26 is independent of scattering anisotropy . beyond the diffusion approximation
the value of @xmath26 is lower and depends on @xmath28 ratio . to extract the value of @xmath26 from experimental data
one can measure @xmath29 at two temperatures , i.e. at different @xmath16 , then find the function @xmath30 = s^{\beta } ( f(l_{\varphi } ^{t_1 } ) -f(l_{\varphi } ^{t_2 } ) ) \label{eq7}\ ] ] and finally determine @xmath26 from @xmath31 * curve*. =
we have measured the conductivity in heterostructures n - gaas / in@xmath32ga@xmath33 as / n - gaas of two types .
the heterostructures with 200 in@xmath32ga@xmath33as quantum well , @xmath34-doped by si in the centre , relate to the first type .
the heterostructures with 50 in@xmath32ga@xmath33as well and doped barriers relate to the second type .
the @xmath34-doped by si layers are arranged in them on both sides of the well at the distance 100 . in this paper
we present the experimental results for two structures of different types which are refereed as structure i and structure ii , respectively . the measurements carried out in wide ranges of magnetic fields ( up to 6 t ) and temperatures ( 0.4 - 40 k ) show that in structures investigated only one size - quantized subband is occupied . the main contribution to
the conductivity comes from the electrons in the quantum well of in@xmath32ga@xmath33as . the electron density and mobility for structure i are @xmath35 @xmath36 and @xmath37 @xmath38/(v sec ) , respectively . for structure
ii they are the following @xmath39 @xmath36 and @xmath40 @xmath41/(v sec ) .
the magnetic field dependencies of the conductivity for both structures for low magnetic fields and different temperatures are shown in fig .
the negative magnetoresistance is observed in the whole range of magnetic fields ( up to 6 t ) .
the main contribution to the negative magnetoresistance in the range @xmath42 t for structure i and @xmath43 t for structure ii comes from the weak localization effect , while in the range @xmath44 t for structure i and @xmath45 t for structure ii it results from the correction to the conductivity due to electron - electron interaction .
notice that the positive magnetoresistance due to weak antilocalization effect was observed in analogous structures for low magnetic fields in refs . .
this effect is observed when the spin relaxation time @xmath46 is less than the phase relaxation time .
the positive magnetoresistance is absent in both our structures ( for example , see the inset in fig .
[ fig1]b , where magnetoresistance of structure ii for very low magnetic fields is shown ) .
it should be mentioned that the conductivity of the structures studied in ref .
was order of magnitude larger than that in our case therefore the phase relaxation time was longer , too .
the phase relaxation time in our case lies in the range @xmath47 sec ( see below ) . comparing this value with @xmath48 sec determined in ref .
we have @xmath49 for our structures .
another reason for the absence of positive magnetoresistance in our case is the fact that , in contrast to structures investigated in ref .
our structures are symmetric in the growth direction .
usually the expression ( [ eq3 ] ) is used to analyze the negative magnetoresistance , taking @xmath7 and @xmath9 as fitting parameters .
the solid curves in fig .
[ fig1 ] have been obtained in this way and , at first glance , they are in good agreement with the experimental data in the range of magnetic field @xmath50 . for structure
i this procedure gives @xmath51 , @xmath52 sec for @xmath53 k. however , the more detailed analysis reveals the difference between the theory and the experimental data ( dashed curve in fig .
[ fig1]a ) . as a consequence ,
the parameters @xmath7 and @xmath54 vary in the range of @xmath55 and @xmath56 sec , respectively , when the fitting procedure is undertaken in different intervals of @xmath57 within the range @xmath58 .
thus , the accuracy of determination of @xmath7 and @xmath9 values is @xmath59% .
the ratio @xmath60 for this structure is @xmath61 for the temperature range @xmath62 k. analogous data treatment for structure ii gives @xmath63 , @xmath64 sec for @xmath53 k and the ratio @xmath65 for the temperature range @xmath66 k. in the strict sense the expression ( [ eq3 ] ) is not valid for this structure because the scattering potential is smooth and the scattering is anisotropic .
nevertheless it provides a good agreement with the experimental data ( fig.1 b ) .
it is commonly believed that lower than unity value of @xmath7 results from the electron - electron interaction ( maki - tompson term ) .
however , below it is shown that such value of @xmath7 in structure ii is the result of failure of the diffusion approximation due to poor @xmath28 ratio .
the temperature dependencies of @xmath9 are plotted in fig.[fig2 ] for both structures , and as is seen @xmath67 with @xmath68 .
this means that the inelasticity of electron - electron interaction is the main mechanism of the phase relaxation.@xcite = now we demonstrate new possibilities of analysis of experimental data provided by the method described above .
it is obvious that the method is applicable for low magnetic fields , @xmath69 , where the interference contribution to the negative magnetoresistance is dominant . as is seen from eq.([eq6 ] )
the information about the statistics of closed paths can be obtained from the fourier transform of @xmath70 .
but the value @xmath71 , not @xmath72 , is experimentally measured . it is easily shown that the fourier transform of the experimental curve @xmath73 padded with zeros at @xmath74 is close to that of @xmath75 at @xmath76 . in fig .
[ fig3 ] the fourier transforms of @xmath77 for different temperatures are presented .
the area range where the fourier transform @xmath15 is shown is bounded .
the maximum value of @xmath24 is determines by signal - to - noise ratio in the experimental curves @xmath78 .
the minimum value is determined by the range of magnetic field , @xmath79 , where the weak localization effect is dominant .
we assume that @xmath80 is about @xmath81 t for structure i and @xmath82 t for structures ii .
= = as is seen from eq .
( [ eq7 ] ) , the function @xmath83 must be linear with respect to @xmath84 with the slope @xmath26 .
it is seen from fig.[fig4 ] that the corresponding curves are really close to straight lines for both structures , but the slopes are different : @xmath85 for structure i and @xmath86 for structure ii .
thus , for structure i with @xmath87 the value of @xmath26 is close to that obtained in the improved diffusion approximation @xmath88 , but somewhat larger than the value @xmath89 obtained in the numerical simulation.@xcite in structure ii the ratio @xmath90 is significantly larger and the diffusion approximation fails . besides , the fact that the impurities are arranged in the barriers in this structure leads to smooth scattering potential and anisotropic scattering .
to our knowledge , there are no theoretical results for this case .
however , it is valid to say that the anisotropy of scattering and smooth scattering potential do not change the statistics of closed trajectories with the lengths significantly larger than the mean free length .
the numerical calculations beyond the diffusion approximation for isotropic scattering @xcite show that @xmath91 at @xmath92 .
it is close to @xmath26 value for structure i with @xmath93 .
thus we believe that the main reason for small value of @xmath26 in this structure is failure of diffusion approximation .
= the method put forward in this paper gives a possibility to determine the area distribution function @xmath13 .
the temperature dependencies of @xmath94 for several @xmath24 are plotted in fig .
the value of @xmath95 for a given @xmath24 increases when @xmath17 due to increase of @xmath16 .
thus the extrapolation of curves @xmath96 to @xmath19 ( see eg .
( [ eq6 ] ) ) gives the value of @xmath97 for the corresponding value of @xmath24 .
the results of such a data treatment for both structures are shown in fig .
[ fig5]b . in fig .
[ fig5]b the area distribution functions obtained within the improved diffusion approximation@xcite with parameters corresponding to the structures investigated are presented too .
one can see that at low @xmath24 the experimental area dependencies of @xmath98 are close to the theoretical ones for both structures . for @xmath99 @xmath41 the more rapid decreasing of the experimental curves is observed and for @xmath100 @xmath41 the experimental values of @xmath101 is @xmath102 times lower than the theoretical values .
there are two reasons for such a discordance : ( i ) the number of closed trajectories with large areas in real samples is smaller than the theoretical one due to , for instance , the long - range potential fluctuations ; ( ii ) the saturation of the phase breaking length with decreasing temperature from @xmath103 k has to lead to underestimating the value of @xmath101 for large @xmath24 in the data processing described above .
it should be noted that some evidence for @xmath104 saturation was obtained only for temperatures @xmath105 k. @xcite our measurements were carried out at significantly higher temperature , @xmath106 k. so , we believe that the rapid decreasing of @xmath101 for @xmath99 @xmath41 ( fig . [ fig5 ] b ) results from the shortage of large trajectories , rather than from the saturation of @xmath104 .
thus , one can see that the area distribution functions of closed paths coincide practically for both structures , but the area dependencies of the average length of closed trajectories are distinguished .
this distinction results from different @xmath107 ratio .
just this fact leads to the lower than unity value of prefactor @xmath7 in structure ii rather than the electron - electron interaction .
the new method of the analysis of negative magnetoresistance is used .
this method provides a possibility to obtain an information on the statistics of closed paths .
the experimental studies of negative magnetoresistance show that the area dependence of average length of closed paths depends on @xmath107 ratio : @xmath108 at @xmath109 ; @xmath110 at @xmath111 .
this fact leads to the lower than unity value of prefactor when one fits the experimental results to hikami expression rather than contribution of electron - electron interaction ( maki - tompson term ) .
the experimental area distribution functions of closed paths are close to these obtained in improved diffusion approximation at low area , but distinct at large one .
the shortage of large trajectories by long range potential fluctuation is a possible reason for such distinction .
this work was supported in part by the rfbr through grants 97 - 02 - 16168 , 98 - 02 - 17286 , the russian program _
physics of solid state nanostructures _ through grant 97 - 1091 , and the program _
university of russia _ through grant 420 . | it is shown that a new kind of information can be extracted from the fourier transform of negative magnetoresistance in 2d semiconductor structures .
the procedure proposed provides the information on the area distribution function of closed paths and on the area dependence of the average length of closed paths .
based on this line of attack the method of analysis of the negative magnetoresistance is suggested .
the method has been used to process the experimental data on negative magnetoresistance in 2d structures with different relations between the momentum and phase relaxation times .
it is demonstrated this fact leads to distinction in the area dependence of the average length of closed paths . 2 |
the critical behavior of antiferromagnetic frustrated systems is still a debated question forty years after the first studies of these systems @xcite .
the key difference between frustrated and nonfrustrated systems is that the order parameter is a vector in the nonfrustrated case and a matrix in the other cases . when frustration originates from the geometry of the system as in stacked triangular antiferromagnets ( sta ) , the symmetry of the hamiltonian is @xmath5 for @xmath1-component spins and the order parameter is a rectangular @xmath6 matrix @xcite .
depending on @xmath1 and the dimension @xmath0 of space , the nature of the phase transition changes , being first order for low values of @xmath1 and dimensions close to four and second order otherwise .
one of the key questions is thus the determination of the line @xmath3 separating the first and second order regions .
it turns out that the value of @xmath7 is certainly close to 3 and its precise determination is crucial to know whether the transition is first or second order for the systems realized in nature that are either ising , xy or heisenberg .
numerical simulations of several frustrated antiferromagnets such as xy and heisenberg sta show unambiguously that the transition is first order for these systems @xcite .
however , depending on the theoretical approach considered , the determination of @xmath3 varies much when @xmath8 and , as a result , it is not yet settled whether all @xmath5 symmetric systems undergo first order phase transitions in @xmath9 for @xmath10 . the two - dimensional physics of the xy and heisenberg systems is also debated because the relevance of topological defects is not yet understood , in particular the possibility that they trigger a phase transition at finite temperature @xcite .
the different theoretical approaches tackling with the problem of the calculation of @xmath3 can be roughly divided into two classes : the perturbative and the nonperturbative renormalization group ( nprg ) calculations .
the class of perturbative calculations can be again divided into several different subclasses depending on whether they are performed directly in @xmath9 ( at six loops ) @xcite or in an @xmath11- or pseudo-@xmath11-expansion ( respectively at six and five loops)@xcite . in the latter case ,
the value of @xmath7 is systematically found larger than 3 ( of order 6 ) as it is also the case for the nprg calculations that find @xmath12@xcite . on the contrary , the perturbative calculation performed directly in @xmath9 at six loops
yields a fixed point for @xmath13 and 3 and thus predicts that several @xmath5 symmetric systems should undergo a second order phase transition .
recently , a completely different method based on the conformal bootstrap has been used to study matrix models in @xmath9 and in particular the @xmath5 frustrated systems @xcite .
a critical behavior has been found in the heisenberg case with exponents in good agreement with those of the six - loop fixed dimension approach .
this approach has the advantage of being unbiased by convergence problems since it is not based on series expansions , contrary to rg methods and , when applied to the ferromagnetic @xmath14 models , it leads to an extremely accurate determination of the critical exponents , at least when it is truncated at large orders @xcite .
the situation of the nprg approach , that we re - examine here , is therefore the following .
either the conclusions drawn from its results are correct and then both the fixed dimension perturbative rg approach and the conformal bootstrap are wrong or , conversely , it is wrong ( together with the @xmath11-expansion approaches ) and this implies that the approximations used are too drastic to reproduce the correct physics . in both cases ,
something very unusual is at work because the methodologies that have been used in these studies lead in many cases to correct and accurate results . as for the nprg , which is based on an exact rg equation , the approximations used so far to tackle with frustrated systems consists in performing a derivative expansion @xcite and a field expansion of the gibbs free energy @xcite .
the rationale behind this choice is ( i ) that the critical behavior of thermodynamic quantities such as the specific heat or the susceptibility for instance are dominated by long wavelength fluctuations which justifies expanding the correlation functions in their momenta ( derivative expansion ) and ( ii ) that the impact of the @xmath15-point functions with @xmath15 large on the rg flow of the zero or two - point functions should be small ( field - expansion ) .
it is the aim of this article to eliminate one source of inaccuracy of the nprg approach , the field expansion , which is known to be inaccurate at low dimensions even for simple models such as the ferromagnetic @xmath14 models @xcite .
the price to pay to get rid of this approximation is to work functionally , that is , to follow the rg flow of functions of the fields instead of a limited number of coupling constants . in the case of nonfrustrated systems , this is relatively simple since the @xmath14 symmetry implies that all functions involved in the rg flows depend on the fields only through the unique @xmath14-invariant : @xmath16 . for frustrated systems , there exists two @xmath5 invariants and the resulting flow equations are partial differential equations that are rather involved .
we show in this article how to simplify the numerical problem and point out why the numerical difficulties are so severe at low dimensions that our method does no longer work when approaching @xmath2 .
we provide the results thus obtained for the curve @xmath3 between @xmath17 and @xmath18 .
our results confirm what was previously found within a nprg approximation involving a field expansion of the potential and the @xmath11-approaches and thus contradict both the fixed - dimension perturbative approach and the results obtained with the conformal bootstrap .
as the archetype of frustrated spin systems , we employ the stacked triangular antiferromagnets ( sta ) .
this system is composed of two - dimensional triangular lattices that are piled - up in the third direction . at each lattice site @xmath19 ,
is defined a @xmath1-component vector @xmath20 of modulus 1 .
the hamiltonian of this system is given by @xmath21 the sum @xmath22 runs on all pairs of nearest neighbor spins .
the coupling constants @xmath23 are given by @xmath24 for a pair of sites inside a plane and @xmath25 between planes .
we assume that the interactions inside a plane are antiferromagnetic : @xmath24 is positive .
the long distance effective theory for the sta has been derived by yosefin and domany@xcite .
the order parameter consists of the @xmath26 matrix @xmath27 that satisfies @xmath28 for @xmath29 .
then , the effective hamiltonian in the continuum is given by @xmath30\right).\ ] ] the constraint @xmath31 for @xmath29 can be replaced by a soft potential @xmath32 whose minima are given by @xmath33 and the ginzburg - landau - wilson hamiltonian for sta reads @xmath30+u\left(\mathbf{\boldsymbol{\phi}}_{1},\mathbf{\boldsymbol{\phi}}_{2}\right)\right).\label{eq : effective hamiltonian}\ ] ] instead of @xmath34 , it is convenient to work with the invariants of the o@xmath35o(2 ) group that can be chosen as : @xmath36 with this choice , the ground state configuration corresponds to @xmath37 and @xmath38 . up to the fourth order @xmath39
can be written as @xmath40 where @xmath41 and @xmath42 are positive coupling constants .
a typical ground state in terms of @xmath43 is given by @xmath44 , that is : @xmath45
the nprg method is based on wilson s idea of integrating statistical fluctuations step by step . in this paper
, we employ the effective average action method as an implementation of the nprg in continuum space @xcite .
the first step is to introduce a @xmath46-dependent partition function @xmath47 in the presence of sources : @xmath48= \int\mathcal{d}\boldsymbol{\phi}_{i}\exp\left(-h[\boldsymbol{\phi}_{i}]- \delta h_{k}[\boldsymbol{\phi}_{i}]+\boldsymbol{j}_{i}\cdot\boldsymbol{\phi}_{i}\right),\ ] ] where @xmath49 and @xmath50 .
the idea underlying the effective average action is that in @xmath47 only the fluctuations of large wave - numbers ( the rapid modes ) compared to @xmath46 are integrated over while the others ( the slow modes ) are frozen by the @xmath51 term .
as @xmath46 is decreased , more and more modes are integrated until they are all when @xmath52
. the function @xmath53 , which is the fourier transform of @xmath54 , plays the role of separating rapid and slow modes : it almost vanishes for @xmath55 so that the rapid modes are summed over and is large ( of order @xmath56 ) below @xmath46 so that the fluctuations of the slow modes are frozen
. we define as usual @xmath57=\ln\mathcal{z}_{k}[\boldsymbol{j}_{i}]$ ] .
thus , the order parameter @xmath58 at scale @xmath46 is defined by @xmath59}{\delta\boldsymbol{j}_{i}\left(\mathbf{x}\right)}\ , .
\label{eq : def=00003d00003d0000a5phi}\ ] ] the running effective average action @xmath60 $ ] is defined as the ( modified ) legendre transform of @xmath61 : @xmath62=-w_{k}\left[\boldsymbol{j}_{i}\right]+\mathbf{j}_{i}\cdot\boldsymbol{\varphi}_{i}-\delta h_{k}\left[\mathbf{\boldsymbol{\varphi}}_{i}\right]\ ] ] where @xmath63 is defined such that eq .
( [ eq : def=00003d00003d0000a5phi ] ) holds for fixed @xmath64 . from this definition
one can show that @xmath65 where the cutoff @xmath66 is the inverse of the lattice spacing @xmath67 .
equations ( [ gammakboundaries ] ) imply that @xmath68 interpolates between the hamiltonian of the system when no fluctuation has been summed over , that is , when @xmath69 , and the gibbs free energy @xmath70 when they have all been integrated , that is , when @xmath52 .
we define the variable @xmath71 , called `` rg time '' , by @xmath72 .
the exact flow equation for @xmath68 reads @xcite : @xmath73= \frac{1}{2}\mathrm{tr}\int_{x , y}\partial_{t}r_{k}(x - y)\left(\frac{\delta^{2}\gamma_{k } \left[\mathbf{\boldsymbol{\varphi}}_{i}\right]}{\delta\varphi_{i}^{\alpha } \left(\mathbf{x}\right)\delta\varphi_{i'}^{\alpha'}\left(\mathbf{y}\right)}+r_{k } \left(\mathbf{x - y}\right)\delta_{i , i'}\delta_{\alpha,\alpha'}\right)^{-1 } , \label{floweq}\ ] ] for @xmath74 and @xmath75 .
it is generally not possible to solve exactly the above flow equation ( [ floweq ] ) and approximations are required in practice . in this paper , we employ the approximation of lowest level in the derivative expansion dubbed the local potential approximation ( lpa ) and some of its refinements . within the lpa ,
@xmath68 is approximated by a series expansion in the gradient of the field , truncated at its lowest non trivial order : @xmath62= \int d^{d}\mathbf{x}\left(\frac{1}{2}\left[\left(\partial\boldsymbol{\varphi}_{1}\right)^{2}+ \left(\partial\boldsymbol{\varphi}_{2}\right)^{2}\right]+u_{k}\left(\rho,\tau\right)\right).\ ] ] only a potential term @xmath76 is thus retained in this approximation which is accurate as long as the impact of the renormalization of the derivative terms on the flow of the potential is small .
this is most probably the case when the anomalous dimension is small and @xmath77 .
the next level of approximation consists in including in the approximation a running field renormalization @xmath78 @xmath79 this approximation has been used in @xcite where the function @xmath80 was further expanded in powers of the invariants @xmath81 and @xmath82 .
this is what we improve here to avoid any artifact coming from this field truncation .
this approximation , that we call lpa , yields the one - loop result obtained within the @xmath11-expansion in @xmath83 and also , in the @xmath14 case , the one - loop result of the @xmath84 expansion of the nonlinear sigma model .
although the situation is a little more involved in our case , it is very probable that the lpa is very accurate close to @xmath2 and our numerical results confirm this , see the following .
our approach is therefore at least a clever interpolation between the results obtained either in @xmath17 or @xmath2 .
the @xmath46-dependent effective potential @xmath76 is defined by @xmath85\ ] ] where @xmath86 are constant fields and @xmath87 is the volume of the system .
the running field renormalization @xmath88 is set to one in lpa : @xmath89 , which leads to a vanishing anomalous dimension : @xmath90 . in lpa calculations
, the anomalous dimension @xmath91 is obtained from the flow of @xmath88 since it can be shown that at criticality : @xmath92 the flows of @xmath93 and @xmath78 have been derived in @xcite and we recall them for completeness in appendix a. these flows are rather complicated and their numerical integration suffers from all the inherent difficulties of the nonlinear partial differential equations .
the first difficulty comes from the choice of variables .
it is tempting to work with the invariants @xmath81 and @xmath82 defined above because the symmetry of the problem is encoded in the very definition of the variables and any smooth function of these variables corresponds to a function that has the right symmetry .
however , @xmath81 and @xmath82 satisfy @xmath94 and it is not easy to deal with this constraint numerically because the domain where the variables @xmath81 and @xmath82 live is nontrivial .
thus , we define another set of variables @xmath95 which is numerically more convenient . for any @xmath96 and @xmath97
, it can be proven that there exists @xmath98 and @xmath99 such that the matrix @xmath100 , where @xmath26 matrix @xmath101 is defined as @xmath102 , becomes `` diagonal '' , namely , @xmath103 because of the @xmath104 symmetry of the model , we conclude that @xmath105 .
this fact shows that we can parametrize the order parameter space using @xmath106 and @xmath107 , instead of @xmath96 and @xmath97 .
the @xmath104 invariants @xmath81 and @xmath82 are expressed in terms of @xmath106 and @xmath107 as @xmath108 from the definitions ( [ eq : def_rho_tau ] ) we find that the symmetries of the original problem imply : @xmath109 thus , to solve the flow equations , it is sufficient to consider the region @xmath110 .
this triangular domain is much more convenient from a numerical point of view than the parabolic domain @xmath94 for the invariants @xmath81 and @xmath82 . at criticality
, the @xmath46-dependent effective action is attracted towards the fixed point solution of the nprg equation once it is expressed in terms of the dimensionless renormalized fields @xmath111 and a dimensionless local potential @xmath112 .
we thus define the dimensionless and renormalized quantities : @xmath113 the flow equation for @xmath114 is given by eq .
( [ eq : flowu ] ) in appendix a. the critical exponent @xmath115 of the correlation length is obtained from the relevant eigenvalue of the linearized flow around the fixed point solution and @xmath91 from the flow of @xmath88 .
the other critical exponents can be deduced from these ones by scaling relations .
the scaled @xmath104 invariants @xmath116 and @xmath117 are defined by @xmath118 , @xmath119 , and the potential and couplings by @xmath120 , @xmath121 , @xmath122 .
notice that as said above @xmath78 does not reach a fixed point but @xmath123 , defined by @xmath124 , does : @xmath125 at criticality with @xmath91 the anomalous dimension of the fields .
from a numerical point of view , there are two possibilities for finding fixed points when they exist .
the first is to dynamically integrate the flow . in this case
, the problem is to find the critical surface which is usually done by dichotomy on the temperature .
once it is found , the fixed point is ( approximately ) reached since it is attractive on the critical surface .
the other method is to look directly for the solution of the fixed point equation ( coupled with eq .
( [ eq : eta ] ) ) : @xmath126 .
this is what we do here .
the advantage of this method is three - fold : ( i ) the numerical scheme is much simpler than integrating the flow ; ( ii ) several numerical instabilities occuring during the integration of the flow are avoided ; ( iii ) the critical exponents are easily obtained from the diagonalization of the rg flow around the fixed point .
we show in the following that although this scheme works very well in dimension @xmath9 , numerical difficulties appear in dimensions close to @xmath2 that make almost impossible to study the physics of frustrated systems in this dimension , at least with our numerical scheme .
the basic idea of this scheme is simple .
it consists in solving the fixed point equations for @xmath127 on a grid in @xmath128 space , taking into account the symmetries ( [ constraint ] ) of this space .
we introduce a cut - off field value @xmath129 and consider the domain @xmath130 .
we then discretize @xmath131 on a square lattice with mesh size @xmath132 , where @xmath133 is the number of lattice points on the axis @xmath134 .
the lattice points are given by @xmath135 for integers @xmath19 and @xmath136 that satisfy @xmath137 .
we define @xmath138 to alleviate the notation .
the fixed point equation for the potential is a differential equation .
we transform it into a set of algebraic equations by discretizing the derivatives of @xmath139 .
we give below some details about this procedure because all our numerical problems come from the boundary of the domain @xmath131 , precisely at the points where the discretization involves exceptional cases .
the formulae for the derivatives @xmath140 for @xmath141 are constructed as follows : \(1 ) in the bulk region ( @xmath142 ) : @xmath143 and @xmath144 as well as @xmath145 and @xmath146 are computed with five points .
@xmath147 is computed with the nine points @xmath148 , @xmath149 , @xmath150 @xmath151 and @xmath152 .
the formulae are exact up to @xmath153 .
notice that for points on the two borders of @xmath131 defined either by @xmath154 or @xmath155 , the derivatives of @xmath156 involve points outside @xmath131 . by using ( [ constraint ] )
, we can compute these values of @xmath156 from those that are inside @xmath131 .
this is one of the advantage of the choice of variables @xmath128 compared to the choice @xmath157 : the derivatives on the two borders @xmath154 and @xmath155 can be computed in the same way as in the bulk .
\(2 ) on the boundary of the domain @xmath131 corresponding to the large field region , @xmath158 , we compute the derivatives in the @xmath159 direction @xmath160 and @xmath161 in the same way as in ( 1 ) , that is , as in the bulk .
the formulae for @xmath162 and @xmath163 are constructed with the five quantities @xmath164 for @xmath165 and are exact at order @xmath166 .
the formula for @xmath167 for @xmath168 involves the six values @xmath169 , @xmath170 for @xmath171 and is exact at order @xmath172 .
finally , for @xmath173 we use twelve points in the region @xmath174 and the formula is exact at order @xmath166 . notice that we have increased the precision of the derivatives on the boundary of the domain @xmath131 corresponding to the large field region in order to test the robustness of our results with respect to the choice of discretization and to try to reduce numerical problems when @xmath0 is close to 2 . in all cases
studied we did not find any significant changes . in particular , the scheme is not more stable when the number of points chosen to compute the derivatives is increased .
once the derivatives are discretized , the fixed point equation @xmath175 becomes a set of coupled algebraic equations for @xmath176 .
we look for a solution to these equations by a newtons - like method .
one of the difficulty of this method is the huge number of unknowns and the possibility for newton s method to get lost in the very complicated landscape of extrema of the set of equations to be solved .
the way out of this difficulty is to deform continuously a solution of the problem .
our strategy in this paper is to follow the fixed point potential @xmath177 by changing the dimension @xmath0 and the number of spin components @xmath1 gradually starting from @xmath178 and @xmath179 where the field - expansion method provides a good approximation of the fixed point potential .
we use as an initial condition of newton s method : @xmath180 and @xmath90 .
the parameters @xmath181 , @xmath182 and @xmath183 are determined by performing a field - expansion of the lpa equation on @xmath139 at order four in the fields and solving the fixed point equation for these parameters in @xmath178 and for @xmath179 .
as expected , we find four fixed points : the gaussian and the @xmath184 fixed points as well as a once - unstable fixed point @xmath185 driving the phase transition and @xmath186 that corresponds to a tricritical fixed point .
once an approximation of @xmath185 is found with the truncation of eq .
( [ eq : field expansion ] ) , we use it as the initial condition of newton s method for the full potential equation ( supplemented by @xmath91 ) and we easily find @xmath187 .
then , we move in the @xmath4 plane by little steps using as new initial condition what was found for the previous value of @xmath0 and/or @xmath1 studied . the fixed potential potential deforms smoothly and the newton s method always works properly this way .
the line @xmath3 separates in the @xmath4 plane the region where the phase transition is of second order and the region where it is of first order .
when @xmath1 is lowered at fixed @xmath0 , this line corresponds to the locus of points where @xmath185 disappears by collapsing with @xmath186 .
there are two possibilities to determine @xmath3 .
either we decrease @xmath1 at fixed @xmath0 and look for the value of @xmath1 where @xmath185 is no longer found and then repeat the same procedure by decreasing @xmath0 . or we compute the smallest eigenvalue of the flow around the fixed point @xmath185 corresponding to an irrelevant direction and look for the value of @xmath1 where it vanishes .
this eigenvalue is a measure of the speed of the flow on the rg trajectory joining @xmath185 and @xmath186 and this speed goes to 0 when the fixed points collapse .
this second method is much more accurate and less demanding than the first one and we therefore use it . for each value of @xmath4 studied , we thus compute the eigenvalues of the stability matrix @xmath188 defined as @xmath189 where we consider @xmath190 and @xmath191 as ( super-)indices . since the rg time @xmath192 is negative , a negative ( positive ) eigenvalue of the matrix @xmath193 corresponds to a relevant ( irrelevant ) eigendirection around the fixed point .
we sort the eigenvalues as @xmath194 .
note that the above stability matrix around any fixed point solution has a trivial relevant eigendirection corresponding to the constant shift @xmath195 with the eigenvalue @xmath196 , which can be easily seen from eq .
( [ eq : flowu ] ) .
hereafter , this trivial eigenvalue is omitted when we discuss the stability of a fixed point .
the critical exponent @xmath115 is given by @xmath197 and the smallest positive eigenvalue we are interested in is @xmath198 .
@xmath199 for each dimension @xmath0 and value of @xmath1 we have to make sure that our results are converged .
once the choice of discretization of the derivatives has been made , there are two parameters that can be tuned : the values of @xmath129 and of the mesh size @xmath132 .
the potential @xmath187 shows a minimum at @xmath200 and we have observed that @xmath129 should be at least 1.5 times larger than @xmath201 to get values of @xmath3 converged with an accuracy of less than @xmath202 .
we have also observed that the smaller the dimension , the smaller @xmath203 must be to get converged results .
this last point has two origins .
first , at small @xmath0 the fixed point potential is steep at large fields because it behaves as @xmath204 and a small mesh size is necessary to accurately describe the shape of @xmath187 .
second , if @xmath205 is too small , we find that even far away from @xmath2 , say @xmath9 , several eigenvalues corresponding to relevant eigendirections appear in the spectrum and spoil the degree of stability of the fixed point @xmath206 .
these eigenvalues are clearly spurious because their values change considerably when either @xmath203 is decreased or @xmath207 is increased whereas the complementary set of eigenvalues , the physical ones , remain unchanged up to the sixth digit , see table [ eigenvalues ] .
we observe that as @xmath203 is decreased , these spurious eigenvalues systematically disappear ( or , at least , get a very large real part which makes them highly irrelevant ) .
the conclusion of this study is that for each @xmath0 , a sufficiently large @xmath205 should be chosen so that the set of first most relevant eigenvalues is converged as for their numbers and values .
we find that in @xmath9 , @xmath208 is sufficient to get fully converged results while leading to numerically feasible calculations .
we also find that as @xmath0 approaches 2 , `` large '' values of @xmath129 favor the presence of spurious eigenvalues that can only be eliminated by increasing @xmath205 .
it turns out that around @xmath209 , very large values of @xmath205 , such as @xmath210 , would be necessary to avoid spurious eigenvalues and that decreasing @xmath0 would impose to increase @xmath205 in a prohibitive way .
we have been able to compute @xmath3 down to @xmath18 by computing directly the value of @xmath1 where no fixed point @xmath206 is found with newton s method but we have not been able to go below this dimension .
the crosses correspond to the calculation performed in this article either with the lpa or lpa. the two continuous curves correspond to the five - loop results @xcite obtained within the @xmath11-expansion resummed either by assuming that @xmath211 in resummation 2 or assuming nothing about the value of @xmath211 in resummation 1.,scaledwidth=65.0% ] we have checked by varying all parameters ( @xmath212 and @xmath213 ) that our results are fully converged in @xmath9 from a numerical viewpoint both at the lpa and lpa levels .
they are also converged down to @xmath214 and are less reliable in @xmath18 at the lpa level although we are not able to give a quantitative estimate of the impact of our numerical errors on the value of @xmath3 in this dimension .
we show our determination of @xmath3 in fig .
[ nc ] together with the results obtained from the @xmath215-expansion at five loops . for @xmath9
, our results confirm the previous results obtained either by the nprg @xcite or the @xmath11-expansion approaches @xcite .
the comparison between the lpa and lpa results strongly suggests that neglecting the effect of the derivative terms on the determination of @xmath3 plays a minor role in @xmath9 .
moreover , @xmath7 increases between the lpa and lpa and becomes closer to the results obtained with the @xmath11-expansion , which is expected .
it seems therefore very difficult to imagine that @xmath7 could be smaller than 3 .
let us also emphasize that the only monte carlo simulation that still finds a second order transition for a value of @xmath1 below our value of @xmath7 , that is , for @xmath216 , has been performed for @xmath13 by calabrese et al .
@xcite on a discretization of the ginzburg - landau model eqs .
( [ eq : effective hamiltonian ] ) , ( [ pot ] ) .
they found that depending on the values of @xmath41 and @xmath42 , the transition is of first or second order : at fixed @xmath41 and small @xmath42 , the transition is of second order whereas it is of first order at large @xmath42 . since nonuniversal quantities , such as phase diagrams @xcite
, can be accurately computed from the integration of the nprg flow equations , it is possible to estimate the magnitude of the correlation length @xmath217 at the transition within the lpa by initializing the flow with the data corresponding to the simulations . by varying these data as well as the cut - off function @xmath218 , it is found that @xmath217 is always finite ( since there is no fixed point ) but very large , typically larger than 2000 lattice spacings @xcite . from a numerical point of view
, there is no doubt that such a large correlation length makes impossible to decide in favor of a second or a ( very weak ) first order phase transition since in both cases the physics will look the same at the scale of the lattice size which was at most 120 lattice spacings in the numerical simulations . we conclude that this monte carlo result does not contradict our conclusion that @xmath219 .
this result shows unambiguously that if our result is wrong , the origin of the problem can only be found by including the renormalization of the functions in front of the derivative terms .
however , considering that the anomalous dimension is small for these systems when they undergo a second order phase transition , that is , for @xmath220 , this hypothesis seems very doubtful .
we therefore suggest that it is useless to study the order two of the derivative expansion in these models that , most probably , would bring only minor modifications as compared to the present study .
we also suggest that only the blaizot - mendez - wschebor approach @xcite , where the full momentum dependence of the two - point functions is retained as well as the full field - dependence of the potential @xmath139 could lead to a very accurate determination of @xmath221 . as for the approach to @xmath2
, we find a remarkable agreement between our results and what was found within the @xmath11-expansion .
two resummations of the @xmath11-expansion were performed by the authors of @xcite , either by assuming that @xmath211 or by letting free the value of @xmath222 .
this agreement is not very surprising because we expect the lpa to be accurate around @xmath2 for @xmath223 ( it is one - loop exact in the nonfrustrated case ) .
notice that our results are not precise enough to determine unambiguously the value of @xmath222 although it seems clear that it can not be very different from 2 .
it is therefore very unlikely that @xmath224 and our results show that the @xmath206 fixed point must exist for all dimensions larger than two in the heisenberg case .
since the nprg flow reproduces the low - temperature expansion of the nonlinear sigma model around @xmath2 , we conclude that the critical behavior of frustrated systems in @xmath225 is driven for @xmath226 by the fixed point @xmath206 corresponding to a critical temperature of order @xmath11 in agreement with mermin - wagner theorem . since we find no other once - unstable fixed point
, we conclude that our study rules out the possibility of having a finite temperature fixed point in @xmath2 for @xmath226 contrary to what was found at five loops in a fixed dimension rg calculation @xcite . to conclude
, we have presented a rather simple method to compute the fixed point properties of matrix models describing frustrated systems without having recourse to a field expansion of the free energy @xmath70 ( but keeping a derivative expansion of @xmath70 ) .
this is especially important in low dimensions where the field expansion is known to fail . in dimension
@xmath9 , our results fully confirm what was previously found within less accurate nprg calculations that involved field truncations on top of the derivative expansion @xcite . in dimension @xmath2 ,
more stable numerical schemes are still needed to study the physics of topological excitations in frustrated systems ( that are of different natures than in nonfrustrated systems ) and we believe that the present work is the first step in this direction .
this work was supported in part by a grant - in - aid for young scientists ( b ) ( 15k17737 ) , grants - in - aid for japan society for promotion of science ( jsps ) fellows ( grants nos . 241799 and 263111 ) , the jsps core - to - core program `` non - equilibrium dynamics of soft matter and information '' .
throughout this paper we employ the following @xmath227 , which is useful for analytical treatments@xcite : @xmath228 where @xmath88 is defined as @xmath229 where the field values are set to the minimum of @xmath230 given by eq .
( [ eq : min ] ) . here
the fourier transform @xmath231 is defined as @xmath232 the scaled nonperturbartive renormalization group flow equation for the porential @xmath114 is given by @xmath240 here , to simplify the notation , we have defined another kind of derivatives @xmath241 with respect to @xmath242 and @xmath243 as @xmath244 | we revisit the critical behavior of classical frustrated systems using the nonperturbative renormalization group ( nprg ) equation .
our study is performed within the local potential approximation of this equation to which is added the flow of the field renormalization .
our flow equations are functional to avoid possible artifacts coming from field expansions which consists in keeping only a limited number of coupling constants
. we present a simple numerical method to follow the fixed point solution of our equations by changing gradually the dimension @xmath0 and the number @xmath1 of spin - components . we explain in details the advantage of this method as well as the numerical difficulties we encounter , which become severe close to @xmath2 .
the function @xmath3 separating the regions of first and second order in the @xmath4 plane is computed for @xmath0 between 4 and 2.2 .
our results confirm what was previously found within cruder approximation of the nprg equation and contradict both the fixed dimension perturbative approach and the results obtained within the conformal bootstrap approach . |
we consider some well known families of two - player , zero - sum , turn - based , perfect information games that can be viewed as specical cases of shapley s stochastic games @xcite .
they have appeared under various names in the literature in the last 50 years and variants of them have been rediscovered many times by various research communities . for brevity , in this paper we shall refer to them by the name of the researcher who first ( as far as we know ) singled them out . * * condon games * @xcite ( a.k.a .
simple stochastic games ) .
a condon game is given by a directed graph @xmath0 with a partition of the vertices into @xmath1 ( vertices beloning to player 1 , @xmath2 ( vertices belonging to player 2 ) , @xmath3 ( random vertices ) , and a special terminal vertex * 1*. vertices of @xmath3 have exactly two outgong arcs , the terminal vertex * 1 * has none , while all vertices in @xmath4 have at least one outgoing arc . between moves ,
a pebble is resting at one of the vertices @xmath5 .
if @xmath5 belongs to a player , this player should strategically pick an outgoing arc from @xmath5 and move the pebble along this edge to another vertex .
if @xmath5 is a vertex in @xmath3 , nature picks an outgoing arc from @xmath5 uniformly at random and moves the pebble along this arc .
the objective of the game for player 1 is to reach * 1 * and should play so as to maximize his probability of doing so .
the objective for player 2 is to prevent player 1 from reaching * 1*. * * gillette games * @xcite . a gillette game @xmath6 is given by a finite set of states @xmath7 , partioned into @xmath8 ( states belonging to player 1 ) and @xmath9 ( states belonging to player 2 ) .
to each state @xmath5 is associated a finite set of possible actions . to each such action
is associated a real - valued _ reward _ and a probability distribution on states . at any point in time of play ,
the game is in a particular state @xmath10 .
the player to move chooses an action strategically and the corresponding award is paid by player 2 to player 1 .
then , nature chooses the next state at random according to the probability distribution associated with the action .
the play continues forever and the accumulated reward may therefore be unbounded .
fortunately , there are ways of associating a finite payoff to the players in spite of this and more ways than one ( so @xmath6 is not just one game , but really a family of games ) : for _ discounted _ gillette games , we fix a _ discount factor _ @xmath11 and define the payoff to player 1 to be @xmath12 where @xmath13 is the reward incurred at stage @xmath10 of the game .
we shall denote the resulting game @xmath14 . for _ undiscounted _ gillette game
we define the payoff to player 1 to be the _ limiting average _
payoff @xmath15 we shall denote the resulting game @xmath16 .
undiscounted gillette games have recently been referred to as _ stochastic mean - payoff _ games in the computer science literature @xcite .
a natural restriction of gillette games is to _ deterministic _ transitions ( i.e. , all probability distributions put all probability mass on one state ) .
this class of games has been studied in the computer science literature under the names of cyclic games @xcite and mean - payoff games @xcite .
a _ strategy _ for a game is a ( possibly randomized ) procedure for selecting which arc or action to take , given the history of the play so far .
a _ pure , positional strategy _ is the very special case of this where the choice is deterministic and only depends on the current vertex ( or state ) , i.e. , a pure , positional strategy is simply a map from vertices ( for gillette games , states ) to vertices ( for gillette games , actions ) .
a strategy @xmath17 for player 1 is said to be _ optimal _ if for all vertices ( states ) @xmath10 it holds that , @xmath18 where @xmath8 @xmath19 is the set of strategies for player 1 ( player 2 ) and @xmath20 is the probability that player 1 will end up in * 1 * ( for the case of condon games ) or the expected payoff of player 1 ( for the case of gillette games ) when players play using the strategy profile @xmath21 and the play starts in vertex ( state ) @xmath10 .
similarly , a strategy @xmath22 for player 2 is said to be optimal if @xmath23 for all games described here , a proof of liggett and lippman @xcite ( fixing a bug of a proof of gillette @xcite ) shows that there are optimal , pure , positional strategies and that a pair of such strategies form an exact nash equilibrium of the game .
these facts imply that when testing whether conditions ( [ opt1 ] ) and ( [ opt2 ] ) holds , it is enough to take the inifima and suprema over the finite set of pure , positional strategies of the players . in this paper , we consider _ solving _ games . by solving a game we mean the task of computing a pair of optimal pure , positional strategies , given a description of the game as input .
to be able to finitely represent the games , we assume that the discount factor , rewards and probabilities are rational numbers and given as fractions
. it is well known that condon games can be seen as a special case of undiscounted gillette games ( as described in the proof of lemma 4 below ) , but a priori , solving gillette games could be harder .
a recent paper by chatterjee and henzinger @xcite shows that solving so - called stochastic parity games @xcite reduces to solving undiscounted gillette games .
this motivates the study of the complexity of the latter task .
we show that the extra expressive power ( compared to condon games ) of having rewards during the game in fact does not change the computational complexity of solving the games .
more precisely , our main theorem is : the following tasks are polynomial time equivalent : 1 . solving condon games ( a.k.a .. , simple stochastic games ) 2 . solving undiscounted gillette games ( a.k.a , stochastic mean - payoff games ) with rewards and probabilities represented in binary notation .
3 . solving undiscounted gillette games with rewards and probabilities represented in unary notation .
4 . solving discounted gillette games with discount factor , rewards and probabilities represented in binary notation .
in particular , there is a pseudopolynomial time algorithm for solving undiscounted gillette games if and only if there is a polynomial time algorithm for this task .
the theorem follows from the lemmas 2,3,4 below and the fact that solving games with numbers in the input represented in unary trivially reduces to solving games with numbers in the input represented in binary .
the proof techniques are fairly standard ( although coming from two different communities ) , but we find it worth pointing out that they together imply the theorem above since it is relevant , did not seem to be known , and may even be considered slightly surprising , as deterministic undiscounted gillette games can be solved in pseudopolynomial time @xcite , while solving them in polynomial time remains a challenging open problem .
an even more challenging problem is solving simple stochastic games in polynomial time , so our theorem may be interpreted as a hardness result .
note that a `` missing bullet '' in the theorem is solving discounted gillette games given in unary notation .
it is in fact known that this can be done in polynomial time ( even if only the discount factor is given in unary while rewards and probabilities are given in binary ) , see littman ( * ? ? ?
* theorem 3.4 ) .
[ limit ] let @xmath6 be a gillette game with @xmath24 states and all transition probabilities and rewards being fractions with integral numerators and denominators , all of absolute value at most @xmath25 .
let @xmath26 and let @xmath27 $ ] .
then , any optimal pure stationary strategy ( for either player ) in the discounted gillette game @xmath28 is also an optimal strategy in the undiscounted gillette game @xmath16 .
the fact that _ some _ @xmath29 with the desired property exists is explicit in the proof of theorem 1 of liggett and lippman @xcite . here
, we derive a concrete value for @xmath29 . from the proof of liggett and lippman
, we have that for @xmath17 to be an optimal pure stationary strategy ( for player 1 ) in @xmath16 , it is sufficient to be an optimal pure stationary strategy in @xmath14 for all values of @xmath30 sufficiently close to @xmath31 , i.e. , to satisfy the inequalities @xmath32 for all states @xmath10 and for all values of @xmath30 sufficiently close to @xmath31 , where @xmath33 ( @xmath34 ) is the set of pure , positional , strategies for player 1 ( 2 ) and @xmath35 is the expected payoff when game starts in position @xmath10 and the discount factor is @xmath30 .
similarly , for @xmath22 to be an optimal pure stationary strategy ( for player 1 ) in @xmath36 , it is sufficient to be an optimal pure stationary strategy in @xmath14 for all values of @xmath30 sufficiently close to @xmath31 , i.e. , to satisfy the inequalities @xmath37 so , we can prove the lemma by showing that for all states @xmath10 and _ all _ pure stationary strategies @xmath38 , the sign of @xmath39 is the same for all @xmath40 .
for fixed strategies @xmath41 we have that @xmath42 is the expected total reward in a _ discounted markov process _ and is therefore given by the formula ( see @xcite ) @xmath43 where @xmath44 is the vector of @xmath45 values , one for each state , @xmath46 is the matrix of transition probabilities and @xmath47 is the vector of rewards ( note that for _ fixed _ positional strategies @xmath41 , rewards can be assigned to states in the natural way ) .
let @xmath48 .
then , ( [ closed ] ) is a system of linear equations in the unknowns @xmath44 , where each coefficient is of the form @xmath49 where @xmath50 are rational numbers with numerators with absolute value bounded by @xmath51 and with denominators with absolute value bounded by @xmath25 . by multiplying the equations with all denominators
, we can in fact assume that @xmath50 are integers of absolute value less than @xmath52 .
solving the equations using cramer s rule , we may write an entry of @xmath44 as a quotient between determinants of @xmath53 matrices containing terms of the form @xmath54 .
the determinant of such a matrix is a polynomial in @xmath55 of degree @xmath24 with the coefficient of each term being of absolute value at most @xmath56 .
we denote these two polynomials @xmath57 .
arguing similarly about @xmath58 and deriving corresponding polynomials @xmath59 , we have that @xmath60 is equivalent to @xmath61 , i.e. , @xmath62 . letting @xmath63
, we have that @xmath64 is a polynomial in @xmath55 , with integer coefficients , all of absolute value at most @xmath65 .
since @xmath66 , the sign of @xmath67 is the same for all @xmath68 , i.e. , for all @xmath40 .
this completes the proof . solving undiscounted gillette games ( with binary representation of rewards and probabilities )
polynomially reduces to solving discounted gillette games ( with binary representation of discount factor , rewards , and probabilities ) .
this follows immediately from lemma [ limit ] by observing that the binary representation of the number @xmath26 has length polynomial in the size of the representation of the game .
solving discounted gillette game ( with binary representation of discount factor , rewards , and probabilities ) polynomially reduces to solving condon games .
zwick and paterson @xcite considered solving _ deterministic _ discounted gillette games , i.e. , gillette games where the action deterministically determines the transition taken .
it is natural to try to generalize their reduction so that it also works for general discounted gillette games .
since condon games allows for vertices making random choices , a natural attempt is to simply simulate a stochastic transition by such random vertices .
such a generalization is made even easier by the fact that zwick and paterson proved that solving `` augmented '' condon games where random vertices are allowed to take choices given by arbitrary discrete distributions with rational probability weights ( represented in binary ) is polynomially equivalent to solving `` plain '' condon games .
we find that the reduction outlined above is indeed correct , even though the correctness proof of zwick and paterson has to be modified slightly compared to their proof .
the details follow .
we are given as input a gillette game form @xmath6 and a discount factor @xmath30 and must produce an augmented condon game @xmath69 whose solution yields the solution to the gillette game @xmath14 .
first , we affinely scale and translate all rewards of @xmath6 so that they are in the interval @xmath70 $ ] .
this does not influence the optimal strategies .
vertices of @xmath69 include all states of @xmath6 ( belonging to the same player in @xmath69 as in @xmath6 ) , and , in addition , a random vertex @xmath71 for each possible action @xmath72 of each state @xmath5 of @xmath6 .
we also add a `` trapping '' vertex * 0 * with a single arc to itself .
it does not matter which player it belongs to .
we construct the arcs of @xmath69 by adding , for each ( state , action ) pair @xmath73 the `` gadget '' indicated in figure [ fig ] . reducing discounted gillette games to condon games , height=226 ] to be precise
, if the action has reward @xmath47 and leads to states @xmath74 with probability weights @xmath75 , we include in @xmath69 an arc from @xmath5 to @xmath71 , arcs from @xmath71 to @xmath76 with probability weights @xmath77 , an arc from @xmath71 to * 0 * with probability weight @xmath78 and finally an arc from @xmath71 to the terminal * 1 * with probability weight @xmath79 . there is clearly a 1 - 1 correspondence between pure stationary strategies in @xmath6 and in @xmath69 .
thus , we are done if we show that the optimal strategies coincide . to see this , fix a strategy profile for the two players and consider play starting in any vertex @xmath5 . by construction , if the expected reward of the play in @xmath6 is @xmath80 , the probability that the play in @xmath69 ends up in * 1 * is exactly @xmath81 .
therefore , the two games are strategically equivalent .
solving condon games polynomially reduces to solving undiscounted gillette games with _ unary _ representation of rewards and probabilities .
we are given a condon game @xmath6 ( a `` plain '' one , using the terminology of the previous proof ) and must construct an undiscounted gillette game @xmath69 .
states of @xmath69 will coincide with vertices of @xmath6 , with the states of @xmath69 including the special terminals * 1*. vertices @xmath5 belonging to a player in @xmath6 belongs to the same player in @xmath69 .
for each outgoing arc of @xmath5 , we add an action in @xmath69 with reward 0 , and with a deterministic transition to the endpoint of the arc of @xmath6
. random vertices of @xmath6 can be assigned to either player in @xmath69 , but he will only be given a single `` dummy choice '' : if the random vertex has arcs to @xmath82 and @xmath83 , we add a single action in @xmath69 with reward @xmath84 and transitions into @xmath82 , @xmath83 , both with probability weight @xmath85 . the terminal * 1 * can be assigned to either player in @xmath69 , but again he will be given only a dummy choice : we add a single action with reward 1 from * 1 * and with a transition back into * 1 * with probability weight @xmath31 .
there is clearly a 1 - 1 correspondence between pure stationary strategies in @xmath6 and strategies in @xmath69 .
thus , we are done if we show that the optimal strategies coincide . to see this , fix a strategy profile for the two players and consider play starting in any vertex @xmath5 . by construction , if the probability of the play ending up in * 1 * in @xmath6 is @xmath64 , the expected limiting average reward of the play in @xmath69 is also @xmath64 .
therefore , the two games are strategically equivalent , and we are done .
undiscounted gillette games can be seen as generalizations of condon games and yet they are computationally equivalent .
it is interesting to ask if further generalizations of gillette games are also equivalent to solving condon games .
it seems natural to restrict attention to cases where it is known that optimal , positional strategies exists .
this precludes general stochastic games ( but see @xcite ) .
an interesting class of games generalizing undiscounted gillette games was considered by filar @xcite .
filar s games allow simultaneous moves by the two players . however ,
for any position , the probability distribution on the next position can depend on the action of one player only .
filar shows that his games are guaranteed to have optimal , positional strategies .
the optimal strategies are not necessarily pure , but the probabilities they assign to actions are guaranteed to be rational numbers if rewards and probabilities are rational numbers .
so , we ask : is solving filar games polynomial time equivalent to solving condon games ? krishnendu chatterjee , marcin jurdziski , and thomas a. henzinger . quantitative stochastic parity games . in _
soda 04 : proceedings of the fifteenth annual acm - siam symposium on discrete algorithms _ , pages 121130 , philadelphia , pa , usa , 2004 .
society for industrial and applied mathematics .
d. gillette .
stochastic games with zero stop probabilities . in m.
dresher , a.w .
tucker , and p. wolfe , editors , _ contributions to the theory of games iii _ ,
volume 39 of _ annals of mathematics studies _ ,
pages 179187 .
princeton university press , 1957 . | we consider some well known families of two - player , zero - sum , turn - based , perfect information games that can be viewed as specical cases of shapley s stochastic games .
we show that the following tasks are polynomial time equivalent : * solving simple stochastic games , * solving stochastic mean - payoff games with rewards and probabilities given in unary , and * solving stochastic mean - payoff games with rewards and probabilities given in binary . |
the investigation of the chaotic processes in the deterministic systems is one of the most important trends of the modern physics @xcite . the successes of this current for nonlinear classical systems that are described via trajectories of the system in hilbert space are obvious @xcite .
yet the conception of deterministic chaos in quantum systems , where the idea of trajectory is inapplicable , stays a subject for serious discussions so far .
there is no common understanding of the problem of quantum chaos for today .
usually the quantum chaos is considered as a set of phenomena taking place in quantum systems , the classical analogues of which display chaos @xcite .
generally when investigating the problems of quantum chaos , one has to deal with the hamiltonians of the form : @xmath0 where p , q are a set of classical coordinates and momentum , @xmath1 is classically integrable part of hamiltonian , @xmath2 is a part of hamiltonian which leads to the nonintegrability of the classical equations of motion when added , @xmath3 is a parameter by variation of which the system may be driven in the domain of chaotic dynamics .
the goal of this work is the investigation of a concrete physical system described by the hamiltonian of the form ( 1.1 ) .
the object of our investigation is a diamagnetic hydrogen atom placed in a strong magnetic field .
diamagnetic hydrogen atom , subjected to the influence of the strong magnetic field , has been considered in the quasi - classical approximation in the ref @xcite .
namely a photo - absorbtion spectrum in the transition regime to chaos has been studied .
the main purpose of our work is the investigation of the irreversible evolution of the quantum - mechanical system and the passage from the quantum - mechanical to the quantum - statistical consideration .
we shall try to compare and find out the link between the phenomena taking place under both classical and quantum considerations .
relatively weak magnetic field applied to atom leads just to the energy levels splitting , i.e. to the zeeman effect . in case when the hydrogen atom is placed in the extreme high magnetic field
, the motion of the electron in the plane perpendicular to the field direction is completely defined by this magnetic field and not by the coulomb field of the nucleus , and because of this , the atom is found to be deformed in the transverse direction . at the same time
the field does not affect on the lengthwise motion , and in this direction the atom keeps its size .
one may easily estimate the value of the magnetizing force when the deformation of the electron shell takes place .
one has to compare the value of bohr radius @xmath4 ( where @xmath5 is a plank constant , @xmath6 and @xmath7 are the electron mass and charge respectively ) with a characteristic size of the localization domain of electrons in the magnetic field in the ground state when @xmath8 , @xmath9 , @xmath10 . if @xmath11 , then the magnetic field has the defining effect .
this condition leads to the following estimate of the extremely strong field @xmath12 in which the atom is deformed .
notice , that the magnetic fields required for this , the intensity of which is defined by the condition ( 1.2 ) , according to the modern concepts , may exist on the surface of some astrophysical objects .
the neutron stars ( or pulsars ) appearing as the result of the collapse of the supernova , belong to them .
that is why the investigation of the diamagnetic atoms in the conditions of extremely high magnetic fields is of doubtless interest .
consider the hydrogen atom placed in the magnetic field directed along z - axis @xmath13 .
we shall use larmor theorem . according to this theorem , for studying the diamagnetism theory for single atoms
it is enough to consider the motion of the electron in the coordinate system , rotating with the larmor frequency @xmath14 , where @xmath15 is a gyromagnetic ratio for electrons . during the transition to the rotating coordinate system ,
the change of the kinetic energy of the electron will occur .
this change is conditioned by the addition to the velocity of in the rotating coordinate frame @xmath16 the cross - product of the larmor frequency @xmath17 and the vector @xmath18 of the location of electron .
taking into account , that the direction of vector @xmath17 is opposite to the vector @xmath19 , and gyromagnetic ratio of the electron is @xmath20 ( here @xmath21 is the velocity of light ) , we shall achieve for the velocity of electron in a motionless coordinate frame : @xmath22=\vec{u}+\frac{e}{2m_ec}[\vec{b}\vec{r}].\ ] ] then the kinetic energy of the electron will be of form : @xmath23\vec{u}+\frac{1}{2}m_e\lvert[\vec{\omega}_l\vec{r}]\rvert^2=\\ = \frac{1}{2}m_eu^2+\frac{e}{2m_ec}bl_z+\frac{e^2}{8m_ec^2}b^2(x^2+y^2 ) .
\end{split}\ ] ] one may easily see , that if z - projection of the angular momentum is equal to zero , @xmath24 the task can be reduced to the diamagnetic kepler problem . finally in the system of atomic units the hamiltonian of the considered problem takes the form : @xmath25 where @xmath26 is a constant describing the connection of the system with the external magnetic field .
the form of the hamiltonian ( 2.1 ) is similar to the one of ( 1.1 ) . for @xmath27 the equations corresponding to ( 2.1 )
are integrable . for @xmath28
the system is nonintegrable . finally with the increase of the amplitude of the magnetic field , starting with some definite value of @xmath29
, the dynamics appears to be chaotic .
the use of the numerical methods for the hamiltonian of the form ( 2.1 ) is connected with definite difficulties because of the appearance of singularity in the @xmath30 point . to get over this problem
it is necessary to use semiparabolic coordinates that are connected with the cartesian ones via formulas : @xmath31 and make a transition to a non - homogeneous time via transformation @xmath32 @xcite . then taking into account that @xmath24 , the hamiltonian ( 2.1 ) can be transformed to the following form @xcite : @xmath33 i.e. the hamiltonian is of form @xmath34 where @xmath35 is the integrable part of the hamiltonian ; @xmath36 is the nonintegrable part of the hamiltonian . as one may see from ( 2.2a ) , the transformed hamiltonian by its form coincides with the hamiltonian of interacting oscillators , and in ( 2.2 ) the singularities which are typical for the hamiltonian ( 2.1 ) are absent .
though one should take into account , that during the direct numerical calculations , right after the integration of the hamilton equations corresponding to ( 2.2a ) , one has to perform a reverse transition to a real time .
henceforth the transformed hamiltonian will be the object of our investigations in case of classical consideration . as it is clear from ( 2.2 ) , the value of the magnetic field @xmath37 is the parameter of nonlinearity . we shall study a case of a strong nonlinearity , i.e. the case when the methods of perturbation theory are inapplicable . in such case
the chaos may appear in the system .
we should expect to get a chaos in case of high magnetic fields @xmath38 .
when chaos appears , the mechanical concepts lose their meaning . for describing the processes taking place in the system , the statistical concepts :
kolmogorov entropy @xcite and the fractal dimension of the phase space of the system @xcite become significant .
the study of the statistical features of the dynamical system can be performed only numerically .
the results of the numerical computations , gained by solving the hamilton equations , corresponding to ( 2.2 ) , are given in fig.1 and fig.2 .
, obtained after the numerical integration of the hamilton equations , corresponding to the hamiltonian ( 2.2 ) . the result ( as all the other numerical calculations given in this work ) is achieved by using the software mathworks matlab 6.0 for the values of parameters @xmath39 , @xmath40 . as it is seen from the graph ,
the motion is of a chaotic type , and the destruction of the invariant tori points to this.,width=283 ] plotted under the conditions similar to the one in fig.1 .
the destructed invariant torus , according to kam theorem @xcite , testifies to the existence of the classical chaos.,width=283 ] as seen from fig.1 and fig.2 , the destruction of the invariant tori took place .
the motion is of the chaotic type . for more convincing proof of
the chaotic character of motion one has to perform fourier analysis of the obtained numerical data . to define kolmogorov
entropy one has to perform a fourier transformation of the correlation function @xmath41 where @xmath42 means averaging over the time , @xmath43 is the correlation time , connected with the kolmogorov entropy by the ratio @xmath44 . in practical calculations of ( 2.3 )
it is more convenient to use a realization of a fast fourier transform algorithm @xcite .
as the numerical calculations show , the value of the correlation time is @xmath45 .
the results of numerical calculations are presented in fig.3 . on the frequency @xmath46 ( @xmath47 ) , plotted using eq .
( 2.3 ) and fft method .
the presence of the finite width of the correlation function @xmath48 is the evidence of the appearance of chaos . in other words ,
one may say , that oscillatory process is not characterized by a definite frequency ( period ) .
all the frequencies ( periods ) in a finite frequency interval @xmath49 are involved into oscillatory process.,width=283 ] there is one more interesting property of dynamical systems connected with the kolmogorov entropy that describes chaotic motion and is called poincar recurrence .
the main point of the poincar recurrence is that any system with a finite phase space after some definite time returns to its initial state @xcite .
the study of the time of poincar recurrence is one of the most powerful methods of the analysis of the nonlinear systems .
it turned out to be that in conditions of strong chaos ( i.e. when the phase plane does nt contain the stability islands ) the distribution of the times of poincar recurrences has a lorenz structure @xcite : @xmath50 where @xmath51 is the average recurrence time connected with kolmogorov entropy as @xmath52 .
thus , calculating the value of the kolmogorov entropy one can define the distribution of the recurrence times .
we in details touched upon the question about classical poincar recurrences , because henceforth we will be interested in quantum poincar recurrences .
the idea of quantum poincar recurrence was introduced in the recently published work by g. m. zaslavsky and a. iomin @xcite .
we shall return to the description of this phenomenon during a quantum consideration .
one of the signs of the appearance of chaos is a discrete ( fractal ) dimension of the phase space of the system . for estimation of the fractal dimension we shall use the grassberger - procaccia algorithm @xcite .
the main point of this algorithm consists into following : assume we have obtained from a numerical solution of the equations of motion a set of state vectors @xmath53 that correspond to the sequential steps of integration . in our case
@xmath54 denote a complete set of variables , characterizing the phase space of the system : @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 , @xmath60 , where @xmath61 $ ] is the time interval for the numerical integration . defining some ( small ) @xmath62
, we can use our set for estimation of the following sum : @xmath63 where @xmath64 is the heaviside step function @xmath65 according to grassberger - procaccia algorithm , if one knows @xmath66 , the fractal dimension of the strange attractor may be defined @xmath67 one should calculate @xmath66 for different values of @xmath62 and represent the results in coordinates @xmath68 and @xmath69 .
the expected dependence of @xmath66 has a form of @xmath70 , so that the obtained graph must have a straight - line form with @xmath71 angular coefficient .
results of the numerical calculations are presented in fig.4 and fig.5 . on @xmath62 .
the graph is plotted via the integration of classical hamilton equations corresponding to the hamiltonian ( 2.2a ) for the values of parameters @xmath39 , @xmath40.,width=283 ] on @xmath62 , plotted in a logarithmic scale using the grassberger - procaccia algorithm @xcite .
the graph is plotted via the integration of classical hamilton equations corresponding to the hamiltonian ( 2.2 ) for the values of parameters @xmath39 , @xmath40 .
the curve line corresponds to the results of the numeric data .
the straight - line is plotted via the processing of the numerical data with least squares algorithm . from this graph
one can estimate the fractal dimension of the attractor using the formula @xmath72 , width=283 ]
as it has been already mentioned , usually the quantum chaos is considered as a set of phenomena that appear in quantum systems , the classical analogues of which display chaos .
the difficulty in definition of quantum chaos is connected with a fact , that the concepts characterizing classical chaos ( local instability of phase trajectories ) lose their sense in quantum consideration .
therefore , for describing quantum chaos one has to use methods and concepts which are typical for quantum mechanics .
our goal in case of quantum consideration , is the demonstration of the fact that one of the signs of the quantum chaos is a formation of the mixed state . for endorsement of this assumption we shall use methods of random matrix theory @xcite . due to this fact
, we have to use basis of eigenfunctions of the integrable part of the hamiltonian ( 2.1 ) and we can not expend the wave function in landau states , as was done in @xcite .
this makes situation more complicated but as we are interested in the state @xmath73 , some analytical results still can be obtained .
let us write down the schrdinger equation for the hamiltonian ( 2.1 ) @xmath74 it is easily seen from ( 3.1 ) , that the schrdinger equation in contrast to the classical equations of hamilton is linear . like in the case of classical systems , this conclusion means that eq.(3.1 ) is completely deterministic and does not contain the properties of chaos , the main sign of which is fortuity and abruptness of the behavior of system .
because of this the question appears : then how can one observe nonregular behavior and quantum chaos in the quantum systems ? to answer this question
we shall use some well - known facts from quantum mechanics [ 23,24,25 ] .
let us define via @xmath75 the required eigenfunction of the eq.(3.1 ) and expand it in eigenfunctions of the hamiltonian @xmath76 @xmath77 here @xmath78 is the eigenfunction of the hamiltonian @xmath79 .
according to the universally recognized rules of quantum mechanics , the coefficients of the expansion @xmath80 taken squared by module , define the probability to find the considered system in state described by wave - function @xmath81 and that is why they satisfy the conditions of normalization : @xmath82 substituting the expansion ( 3.2 ) into schrdinger equation ( 3.1 ) , one can obtain a matrix equation for @xmath83 @xmath84 where the elements of matrix @xmath85 are defined by the integrals of the type : @xmath86 it is shown in quantum mechanics [ 23 - 25 ] that the series ( 3.2 ) exactly converges to the function @xmath75 ,
if when taking sum by @xmath87 a totality ( infinite number ) of functions @xmath88 is considered . however the wave function @xmath75 from a qualitative point of view may be with a high degree of accuracy described by the expansion like ( 3.2 ) using a finite number @xmath89 @xmath90 of states @xmath91the further increase of the number of states @xmath92 is similar to the tendency of the plank constant to zero @xmath93 and leads to the improvement of quantitative fit between classical and quantum chaos @xcite . according to the premises
one may say , that the problem of finding eigenfunctions and eigenvalues of hamiltonian @xmath94 may be solved by a numeric diagonalization of the matrix equation ( 3.4 ) .
but here a question is born .
if the state of the system is described by a finite superposition of regular states @xmath81 then the state @xmath75 is regular too , so where should one look for a nonregular behavior of the system ? by the assumption taken for today @xcite the properties of chaos in quantum system are revealed in that sense that matrix ( 3.5 ) is a random matrix , all the elements of which are random values .
based on this , there must exist some correspondence between matrix ( 3.5 ) and the quantities describing classical chaos . using our problem as a pattern we shall start to examine this question .
let us start with quantum mechanical investigation of the hamiltonian ( 2.2 ) . as it has been already mentioned , when the external magnetic field is absent@xmath27 the hamiltonian @xmath95 is exactly integrable .
when the external magnetic field is turned on @xmath96 the hamiltonian becomes nonintegrable , and for some definite value @xmath97 the solutions of the classical equations corresponding to the hamiltonian @xmath98 display chaos .
by analogy with classical chaos , when a set of phase trajectories with a small dispersion of initial conditions is investigated , we shall consider a small increment of the parameter @xmath99using the basis of functions in which hamiltonian @xmath100 is diagonal , for small values of @xmath101 the hamiltonian @xmath102 may be represented as a sum of two matrices : @xmath103 where @xmath104 is a diagonal matrix , @xmath37 is a banded matrix , the elements of which are random numbers . by a numerical diagonalization , one can define the eigenfunctions @xmath105 @xmath106 and eigenvalues @xmath107 @xmath108 corresponding to the @xmath94 and @xmath98 respectively .
we are interested in the interval of energy @xmath109 which is small from a classical point of view , and large from a quantum mechanical one ( with relation to the large number of quantum levels contained in this interval ) .
let us begin to the quantum mechanical analysis of the hamiltonian ( 2.1 ) .
the eigenvectors of the basis , in which hamiltonian is diagonal , can be found easily and are of the form @xcite @xmath110 where @xmath111 is the normalized radial part of the wave function and @xmath112 @xmath113 are the spherical and hyper - geometrical functions respectively @xcite ; @xmath114 @xmath64 and @xmath115 denotes the spherical coordinates . as was mentioned above , using of the symmetric kepler problem s eigenfunctions is connected to the requirements of the random matrix theory .
calculation of non - diagonal matrix elements in this basis is more complicated then in landau basis @xcite .
but the fact that we are interested in the state @xmath116 @xmath117 makes life and it s still possible to obtain some analytical results .
actually the problem is reduced to the calculation of the following matrix elements : @xmath118 these matrix elements can be calculated analytically . as a result for diagonal matrix elements we get @xcite : @xmath119 and non - diagonal matrix elements
can be reduced to the calculation of the following integral : @xmath120 @xmath121 where @xmath122 direct calculation of the integral @xmath123 is not possible .
but we can use recurrence formula , which connects integral @xmath123 and integrals with the lower values of superscript @xmath124 ; @xmath125 .
after straightforward but laborious calculations one can obtain @xmath123 : @xmath126 @xmath127\ ] ] where @xmath128 and @xmath129 is the euler gamma function @xcite .
at the discussion of the nonreversible evolution of the quantum -mechanical system , naturally the question emerges .
how in a quantum system originates non -reversibility ? if the quantum system evolves according to the schrdinger equation , how can a pure quantum -mechanical state become a mixed one ?
the question is that , in contrast to the classical chaos , quantum mechanically irregular motion can not be characterized by extreme sensitivity to tiny changes of the initial data . due to the unitarity of the quantum dynamics ,
the overlap of two wave functions remains time - independent @xmath130 provided time - dependence of @xmath131 and @xmath132 is generated by the same hamiltonian .
however , an alternative characterization of classical chaos , extreme sensitivity to slight changes of the dynamics does carry over into quantum mechanics .
let us assume , that the amplitude of the external magnetic field is modulated by varying field @xmath133 taking into account ( 4.1 ) the schrdinger equation for the wave function of diamagnetic hydrogen atom takes the form @xmath134 where @xmath135 @xmath136 the solution of the time - dependent schrdinger equation formally can be written with the help of a time - ordered exponential @xmath137\}_{+},\ ] ] where the positive time ordering requires @xmath138_{+}= \left\ { \begin{array}{ll } a(t)b(t^ { ' } ) & \mbox{if $ t > t^{'}$ } \\b(t^{'})a(t ) & \mbox{if $ t < t^{'}$ } \end { array}\right.\ ] ] in our case @xmath139 the evolution operator referring to one period @xmath140 the so - called flouqet operator @xmath141 @xcite , is worthy of consideration since it yields a stroboscopic view of the dynamics @xmath142 the floquet operator , being unitary , has unimodular eigenvalues .
suppose we can find eigenvectors @xmath143 of the floquet operator @xmath144 then , with the eigenvalue problem solved , the stroboscopic dynamics can be written out explicitly @xcite @xmath145 as it was mentioned above , our aim is to proof that one of the signs of the emergence of quantum chaos is a formation of the mixed state .
being initially in a pure quantum mechanical state , described by the wave function @xmath146 the system during the evolution makes an irreversible transition to the mixed state .
after the formation of a mixed state the quantum - mechanical description loses its sense and we have to use a quantum - statistical interpretation . to prove the formation of the mixed state , one has to show the zeroing of non - diagonal elements of density matrix , the equality of which to zero is a sign of a mixed state @xcite . taking ( 4.5 ) into account , we obtain for non diagonal matrix elements of the density matrix : @xmath147 where @xmath148 , @xmath149 means averaging over time . for
averaging let us recollect that value @xmath150 is the eigenvalue of the floquet operator .
owing to the non - integrability of the system , eigenvalues can be obtained only by way of numerical diagonalization of the hamiltonian @xmath151 ( 4.2 ) . according to the main hypothesis of the random matrix theory @xcite , elements of this matrix in the chaotic domain are random values .
so , their eigenvalues also can be considered as random values .
this statement is valid only in the chaotic domain @xcite . according to this ,
the phase @xmath152 also may be considered as random .
therefore when @xmath153 , @xmath154 , we get @xmath155 where @xmath156 denotes ensemble average of random matrices , that corresponds to a small dispersion of the value of magnetic field parameter @xmath157 . expression ( 4.8 ) may be found by more vigorous mathematical substantiation .
for this let us recall some details from the probability theory @xcite .
\a ) in general case , under the characteristic function of the random value @xmath158 , mathematical expectation of the following exponent is meant @xmath159 where @xmath160 is a real parameter .
\b ) mathematical expectation itself is defined as a first moment @xmath161 of the random variable @xmath158 @xmath162 where @xmath158 is the discrete random variable which takes possible values @xmath163 with appropriate probabilities @xmath164 , @xmath156 means average . taking ( 4.9 ) ,
( 4.10 ) into account and considering @xmath165 as a random value , we get : @xmath166 where @xmath167 is the mathematical expectation of the characteristic function of the random phase @xmath165 , and @xmath168 let us assume that random phase @xmath165 has a normal dispersion @xcite . then from ( 4.12 ) one can obtain @xmath169 where @xmath170 is the mathematical expectation of the random phase , @xmath171 is the mean square deviation . substituting ( 4.13 ) in ( 4.11 ) and making transition from summation to the integration @xmath172 finally we get @xmath173+erfi[\frac{a+i(t / t_0)\sigma^2}{\sigma}]),\ ] ] where @xmath174 and @xmath175 is the error function , connected to the ordinary error function @xcite by the relation @xmath176 taking into consideration asymptotical form of the function @xmath177 , @xmath178 and @xmath179 , @xmath178 , in case when distribution of the eigenvalues of the quantum - mechanical chaotic system obeys the normal distribution , we have @xmath180 of non - diagonal matrix element ( 4.6 ) on time @xmath181 .
the graph is obtained using ( 4.8 ) and by way of numerical diagonalization of the hamiltonian ( 4.2 ) for different values @xmath182 , @xmath61 $ ] .
the graph is plotted after taking an ensemble average of hamiltonians @xmath183 , corresponding to the dispersion @xmath184 of the value of magnetic field @xmath185 . as it is easily seen from the graph that when @xmath186 , @xmath187 , zeroing of @xmath188 happens . in the numeric calculations we have used definite initial conditions @xmath189 , @xmath190 , obtained via numerical diagonalization of the hamiltonian ( 2.1 ) for @xmath191 .
the dimension of the diagonalized matrix was @xmath192 , @xmath193,width=283 ] .
[ fig:6 ] zeroing of the non - diagonal elements of density matrix is the sign of quantum chaos beginnings and formation of mixed state . from that moment
the quantum dynamics is non - reversible since the information about the wave functions phase is lost . given above reasoning can be checked by numerical experiment , which consists in : the direct solving of non - stationary schrdinger equation ( 4.2 ) , defining eigenvalues @xmath194 and in estimation of the averaged non - diagonal elements of the density matrix according to ( 4.8 ) .
the result of numerical calculations is represented on fig.6 .
according to the results of the previous section ( see fig.6 ) , for times @xmath195 formation of a mixed state happens due to time evolution of quantum - mechanical system .
consequently the concept of a wave function loses its meaning and the poincar recurrences are out of the question . but what happens when @xmath196 ?
we shall clear up this question in the given section .
as it was mentioned above , in work @xcite the concept of quantum poincar recurrence was introduced . in particular , the authors have studied the quantum dynamics of one - dimensional rotator under the influence of delta - like perturbation . the definition of the `` distance '' between the wave - functions was introduced @xmath197 where @xmath198 is the solution of nonstationary schrdinger equation @xmath199 the authors of the work , cited above , look for the solution of the equation ( 5.2 ) in the form of expansion of the function @xmath198 in an arbitrary orthogonal basis @xmath200 @xmath201 where @xmath202 are the expansion coefficients , which determine the time dependence of the function @xmath203 .
it is obvious , that during the evolution in time of the wave - function @xmath198 , the quantity @xmath204 will change in time .
the authors offer to consider the quantum poincar recurrences as such values of time @xmath160 , that satisfy the condition @xmath205 , where @xmath62 is a small constant value ( @xmath206 ) .
our purpose is the use of this method for the study of quantum poincar recurrences in the system of diamagnetic hydrogen atom .
for that , like in the previous section , we shall assume , that the amplitude of the external magnetic field is modulated by a varying field .
but in contrast to the previous section , we shall examine the solution of the equation in interval @xmath196 , when the system is still in a pure quantum - mechanical state and chaotization of phase and formation of a mixed state can be neglected . before starting the analysis of the equation ( 4.2 ) , let us note some typical differences of our problem from the problem studied in @xcite . in problem
studied in @xcite , the dimension of the hilbert space is finite .
that is why the wave - function @xmath203 , according to ( 5.3 ) , is exactly defined by the finite set of the orthogonal functions @xmath210 .
the authors make an assumption , that the expansion coefficients @xmath202 are the random quantities and satisfy the gaussian distribution . in our case
the dimension of the hilbert space is infinite and the approximation of @xmath203 by the finite set of functions of the orthogonal basis @xmath210 is rough .
our purpose is a proof that in spite of it the quantum poincar recurrences may happen in the system . and to compare it with the mean time of classical poincar recurrences @xmath211 ( see formula ( 2.4 ) ) . as it comes from the graph , @xmath212 , where @xmath213 is the angular coefficient of the approximating straight - line .
the achieved estimate for the mean time of quantum recurrences is in a good agreement with the mean value of the classical poincar recurrences @xmath214 ( see fig.3).,width=283 ] for the solution of the nonstationary schrdinger equation ( 4.2 ) we shall consider @xmath160 as a discrete quantity ( consequently the values of the parameter @xmath215 become discrete too ) .
we shall not assume the coefficients @xmath216 to be _ a priori _ random , but we shall determine their values by solving the matrix equation ( 3.5 ) and taking into account the dispersion of the parameter @xmath157 .
thus , using the formulas ( 3.2 ) , ( 3.5 ) and ( 5.1 ) one can define the distribution of the times of poincar recurrences .
the results of numeric calculations are shown in fig.7 and fig.8 . according to fig.8 ,
the distribution of the times of poincar recurrences is of the lorenz form .
this fact can be easily explained , taking into account the existence of a strong classical chaos ( i.e. the absence of the stability islands in the phase space ) .
the lorentz form testifies that if @xmath217 , quantum poinacr recurrences are improbable .
this coincides with the results of the previous section , according to which the system is found in a mixed state in this time interval .
99 f. haake , _ quantum signatures of chaos _ , springer , berlin ( 2001 ) m. c. gutzwiller , _ chaos in classical and quantum mechanics _ , springer , new york ( 1990 ) a. iomin , g. m. zaslavski , _ phys .
e _ , v.63 , p.047203 ( 2001 ) d. cohen , t. kottos , _ phys .
v.69 , p.055201 ( 2004 ) d. cohen , t. kottos , _ phys .
lett . 85 _ , p.4839 ( 2000 ) a. j. lichtenberg , m. a. lieberman , _ regular and stochastic motion _ , springer - verlang , new york , heidelberg berlin ( 1983 ) r. z. sagdeev , g. m. zaslavsky , _ introduction to the nonlinear physics _
( in russian ) , moscow , nauka ( 1988 ) k. t. alligood , t. d. sauer and j. a. york , _ chaos and introduction to dynamical systems _ , springer , new york ( 1996 ) g. berman , g. m. zaslavsky and a. r. kolovsky , _ phys . lett .
a _ , v.87 , p.152 ( 1982 ) v. ya .
demikhovski , d. i. kamenev , _ phys . lett .
a _ , v.228 , p.391 ( 1997 ) d. cohen , t. kottos , _ phys .
e _ , v.63 , p.36203 ( 2001 ) a. ugulava , l. chotorlishvili , k. nickoladze , _ phys .
e _ , v.68 , p.026216 ( 2003 ) a. ugulava , l. chotorlishvili , k. nickoladze , _ phys .
v.70 , p.026219 ( 2004 ) a. ugulava , l. chotorlishvili , k. nickoladze , _ phys .
v.71 , p.056211 ( 2005 ) h. friedrich , d. wintgen , _ phys .
183 _ , ( 1989 ) t. fabi , j. main , t. bartsch and g. wunner , _ j. phys .
b _ , v.38 , p.219 ( 2005 ) p. grassberger , _ phys . lett .
a _ , v.97 , p.227 ( 1983 ) p. grassberger , _
phys.lett._ , v.97 , p.224 ( 1983 ) p. grassberger , i. procaccia , _ physica d _ , v.9 , p.189 ( 1983 ) g. m. jenkins and d. g. watts , _ spectral analysis and its applications _ , holden - day ( 1969 ) r. k. otnes and l. enochson , _ applied time series analysis _ , jon wiley and sons ( 1978 ) a. iomin and g. m. zaslavsky , _ arxiv : nlin.cd/0506031_ v.1 ( 2005 ) l. d. landau and e. m. lifshitz , _ quantum mechanics , nonrelativistic theory _ , pergamon press ( 1977 ) f. a. kaempffer , _ concepts in quantum mechanics _ , academic press ( 1965 ) d. bohm , _ quantum theory _
, new york , prentice - hall ( 1952 ) h. bateman and a. erdelyi , _ higher transcendental functions _ , v.2 , mcgraw - hill book co ( 1955 ) _ handbook of mathematical functions with formulas , graphs , and mathematical tables _ , national bureau of standards , applied mathematical series , v.55 , u.s .
government printing office , washington d.c .
( 1964 ) h. j. shtockmann , _ quantum chaos _
, philipps universitt , marburg ( 2001 ) m. l. mehta , _ random matrices _ , san diego academic press ( 1991 ) l. d. landau and e. m. lifshitz , _ quantum mechanics .
nonrelativistic theory _ , pergamon oxford ( 1977 ) r. p. feynman , _ statistical mechanics _
, w. a. benjamin , inc .
massachusetts ( 1972 ) a. relano , j. m. g. gomer , r. a. molina , j. retamosa and e. faleiro , _ phys .
_ , 89 , 244102 ( 2002 ) w. feller , _ an introduction to probability theory and its applications _ ,
john wiley & sons , inc .
new york , london , sidney v.1 ( 1958 ) , v.2 ( 1966 ) | the statistics of quantum poincar recurrences in hilbert space for diamagnetic hydrogen atom in strong magnetic field has been investigated .
it has been shown that quantities characterizing classical chaos are in a good agreement with the ones that are used to describe quantum chaos .
the equality of classical and quantum poincar recurrences has been shown .
it has been proved that one of the signs of the emergence of quantum chaos is the irreversible transition from a pure quantum mechanical state to the mixed one . |
the kolkata paise restaurant ( kpr ) problem @xcite is a repeated game , played between a large number @xmath0 of agents having no interaction amongst themselves . in kpr problem ,
prospective customers ( agents ) choose from @xmath0 restaurants each evening simultaneously ( in parallel decision mode ) ; @xmath0 is fixed .
each restaurant has the same price for a meal but a different rank ( agreed upon by all customers ) and can serve only one customer any evening .
information regarding the customer distributions for earlier evenings is available to everyone .
each customer s objective is to go to the restaurant with the highest possible rank while avoiding the crowd so as to be able to get dinner there .
if more than one customer arrives at any restaurant on any evening , one of them is randomly chosen ( each of them are anonymously treated ) and is served .
the rest do not get dinner that evening . in kolkata
, there were very cheap and fixed rate
paise restaurants " that were popular among the daily laborers in the city . during lunch hours
, the laborers used to walk ( to save the transport costs ) to one of these restaurants and would miss lunch if they got to a restaurant where there were too many customers . walking down to the next restaurant would mean failing to report back to work on time !
paise is the smallest indian coin and there were indeed some well - known rankings of these restaurants , as some of them would offer tastier items compared to the others .
a more general example of such a problem would be when the society provides hospitals ( and beds ) in every locality but the local patients go to hospitals of better rank ( commonly perceived ) elsewhere , thereby competing with the local patients of those hospitals .
unavailability of treatment in time may be considered as lack of the service for those people and consequently as ( social ) wastage of service by those unattended hospitals . a social planner s ( or dictator s ) solution to the kpr problem is the following : the planner ( or dictator s ) asks everyone to form a que and then assigns each one a restaurant with rank matching the sequence of the person in the que on the first evening .
then each person is told to go to the next ranked restaurant in the following evening ( for the person in the last ranked restaurant this means going to the first ranked restaurant ) .
this shift process than continuous for successive evenings .
call this solution the _ fair social norm_. this is clearly one of the most efficient solution ( with utilization fraction @xmath2 of the services by the restaurants equal to unity ) and the system arrives at this this solution immediately ( from the first evening itself )
. however , in reality this can not be the true solution of the kpr problem , where each agent decides on his own ( in parallel or democratically ) every evening , based on complete information about past events . in this game ,
the customers try to evolve a learning strategy to eventually get dinners at the best possible ranked restaurant , avoiding the crowd .
it is seen , the evolution these strategies take considerable time to converge and even then the eventual utilization fraction @xmath3 is far below unity . the kpr problem
have some basic features similar to the minority game problem @xcite in that diversity is encourage ( compared to herding behavior ) in both , while it differs from ( two - choice ) minority games in terms of the macroscopic size of the choices .
as already shown in ref @xcite , a simple random - choice algorithm , if adapted by all the agents , can lead to a reasonable value of utilization fraction ( @xmath4 ) .
compared to this , several seemingly `` more intelligent '' stochastic algorithms lead to lower utilization of the services .
@xcite studied a few more such `` smarter '' algorithms , having several attractive features ( including analytical estimate possibilities ) , but still failing to improve the overall utilization fraction beyond its random choice value . here
we develop a stochastic strategy , which maintains a naive tendency ( probability decreasing with past crowd size ) to stick to any agent s own past choice ( successful or not ) , leading to a maximum , so far , value of the utilization fraction @xmath2 ( @xmath5 ) in the kpr problem .
we also estimate here analytically the @xmath2 values for several of such strategies .
let the symmetric stochastic strategy chosen by each agent be such that at any time @xmath6 , the probability @xmath7 to arrive at the @xmath8-th ranked restaurant is given by @xmath9,\hspace{.1 in } z=\sum_{k=1}^n\left[k^{\alpha}\exp\left(-\frac{n_k(t-1)}{t}\right)\right],\label{generalstoch}\ ] ] where @xmath10 denotes the number of agents arriving at the @xmath8-th ranked restaurant in period @xmath6 , @xmath11 is a scaling factor and @xmath12 is an exponent . note that under ( [ generalstoch ] ) the probability of selecting a particular restaurant increases with its rank and decreases with its popularity in the immediate past ( given by the number @xmath13 ) .
certain properties of the strategies given by ( [ generalstoch ] ) are the following : 1 . for @xmath14 and @xmath15
, @xmath16 corresponds to the complete random choice case for which we know @xcite that the utilization fraction is around @xmath17 , that is on an average there is 63% utilization of the restaurants ( see appendix a ) .
2 . for @xmath14 and @xmath18 ,
the agents avoid those restaurants visited last evening and choose again randomly from the remaining restaurants @xcite . with appropriate simulation it was shown that the distribution of the fraction @xmath19 of utilization of the restaurants is gaussian around @xmath20 ( see subsection 2.2 ) . for any natural number @xmath21 and @xmath15
, an agent goes to the @xmath8-th ranked restaurant with probability @xmath22 ; which means in the limit @xmath15 in ( [ generalstoch ] ) gives @xmath22 .
let us discuss the results for such a strategy here .
if an agent selects any restaurant with equal probability @xmath23 then probability that a single restaurant is chosen by @xmath24 agents is given by @xmath25 therefore , the probability that a restaurant with rank @xmath8 is not chosen by any of the agents will be given by @xmath26 where @xmath27 hence @xmath28 therefore the average fraction of agents getting dinner in the @xmath8-th ranked restaurant is given by @xmath29 ) versus rank of the restaurants ( @xmath8 ) for different @xmath21 values .
the inset shows the distribution @xmath30 of the fraction @xmath19 agent getting dinner any evening for different @xmath21 values . ] and the numerical estimates of @xmath31 is shown in fig .
( [ fig1 ] ) . naturally for @xmath14 , the problem corresponding to random choice @xmath32 , giving @xmath33 and for @xmath34 , @xmath35 giving @xmath36 as already obtained analytically earlier ( see appendix b ) .
we consider here the case ( see also @xcite)where each agent chooses on any evening ( @xmath6 ) randomly among the restaurants in which nobody had gone in the last evening ( @xmath37 ) .
this correspond to the case where @xmath38 and @xmath39 in eq .
( [ generalstoch ] ) .
our numerical simulation results for the distribution @xmath40 of the fraction @xmath19 of utilized restaurants is again gaussian with a most probable value at @xmath41 .
this can be explained in the following way : as the fraction @xmath2 of restaurants visited by the agents in the last evening is avoided by the agents this evening , the number of available restaurants is @xmath42 for this evening and is chosen randomly by all the @xmath0 agents .
hence , when fitted to eq .
( [ eq : poisson_mm ] in appendix a ) , @xmath43 .
therefore , following eq .
( [ eq : poisson_mm ] ) , we can write the equation for @xmath2 as @xmath44=\bar { f } .\ ] ] the solution of this equation gives @xmath41 .
this result agrees well with the numerical results for this limit ( @xmath14 , @xmath39 ) . in this section
we start with the following stochastic strategy : if an agent goes to restaurant @xmath8 in period ( @xmath37 ) then the agent goes to the same restaurant in the next period with probability @xmath45 and to any other restaurant @xmath46 with probability @xmath47 . in this process , the average utilization fraction is @xmath48 and the distribution @xmath40 is a gaussian around @xmath49 ( see fig . [ fig2 ] )
. an approximate estimate of @xmath2 : let @xmath50 denote the fraction of restaurants where exactly @xmath51 agents @xmath52 appeared on any evening and assume that @xmath53 for @xmath54 .
therefore , @xmath55 , @xmath56 and hence @xmath57 . given
the strategy , @xmath58 fraction of agents will make attempts to leave their respective restaurants in the next evening @xmath59 , while no intrinsic activity will occur on the restaurants where , no body came ( @xmath60 ) or only one came ( @xmath61 ) in the previous evening @xmath62 .
these @xmath58 fraction of agents will now get equally divided ( each in the remaining @xmath63 restaurants ) . of these @xmath58 ,
the fraction going to the vacant restaurants ( @xmath60 in the earlier evening ) is @xmath64 .
hence the new fraction of vacant restaurants is now @xmath65 . in restaurants having
exactly two agents ( @xmath58 percent in the last evening ) , some vacancy will be created due to this process , and this is equal to @xmath66 .
steady state implies that @xmath67 and hence using @xmath57 we get @xmath68 , giving @xmath69 and @xmath70 . of course
, the above calculation is approximate as none of the restaurant is assumed to get more than two costumers on any evening ( @xmath53 for @xmath71 ) .
the advantage in assuming @xmath60 , @xmath61 and @xmath58 only to be non vanishing on any evening is that the activity of redistribution on the next evening starts from this @xmath58 fraction of the restaurants .
this of course affects @xmath60 and @xmath61 for the next evening and for steady state these changes must balance .
the computer simulation results also conform that @xmath72 for @xmath71 and hence the above approximation does not lead to serious error .
in this section we assume that agents have two possible exogenously given values of @xmath21 : @xmath14 or @xmath34 .
we start by taking some random allocation of @xmath21 over the set of @xmath0 agents .
the strategy followed by each agent thereafter is the following : if an agent starts with an @xmath73 and fails to get dinner for the successive @xmath74 evenings then , in the next evening , the agent shifts to @xmath75 .
the steady state distribution of the @xmath21 values in the population of agents do not depend on the initial allocation of @xmath21 values in the population ( see fig .
[ fig3 ] ) . however , as in obvious , for large values of @xmath76 , the stability of the distribution disappears . .
the same for @xmath34 will be given by just complementary function . ]
in the kpr problem if the rational agents interact then a _ fair social norm _ that can evolve is a periodically organized state with periodicity @xmath0 where each agent in turn gets served in all the @xmath0 restaurants and all agents get served every evening . can we find deterministic strategies ( in the absence of a dictator ) such that the society achieves this fair social norm ?
there is one variant of pavlov s win shift lose stay strategy ( see @xcite ) that can be adopted to achieve the fair social norm and another variant that can be adopted to achieve the fair social norm in an asymptotic sense .
of course , these strategies are deterministic in nature .
the fair strategy works as follows : 1 . at time
( evening ) @xmath77 , agents can choose any restaurants either randomly or deterministically .
2 . if at time @xmath6 agent @xmath51 was in a restaurant ranked @xmath8 and was served then , at time @xmath78 , the agent moves to the restaurant ranked @xmath79 if @xmath80 and moves to the restaurant ranked @xmath0 if @xmath81 .
3 . if agent @xmath51 was in a restaurant ranked @xmath8 at time @xmath6 and was not served then , at time @xmath82 , the agent goes to the same restaurant .
it is easy to verify that this strategy gives a convergence to the fair social norm in less than or equal to @xmath0 periods .
moreover , after convergence is achieved , the fair social norm is retained ever after .
the difficulty with this strategy is that a myopic agent will find it hard to justify the action of going to the restaurant ranked last after getting served in the best ranked restaurant .
however , if the agent is not that myopic and observes the past history of strategies played by all the agents and can figure out that this one evening loss is a tacit commitment devise for this kind of symmetric strategies to work then this voluntary loss is not that implausible .
therefore one needs to run experiments before arguing for or against this kind of symmetric deterministic strategies .
more importantly the fair strategy can be modified to take care of this justification problem provided one wants to achieve the fair social norm in an asymptotic sense .
the asymptotically fair strategy works as follows : 1 . at time
( evening ) @xmath77 , agents can choose any restaurants either randomly or deterministically .
if at time @xmath6 agent @xmath51 was in a restaurant ranked @xmath8 and was served then , at time @xmath82 , the agent moves to the restaurant ranked @xmath79 if @xmath80 and goes to the same restaurant if @xmath83 .
if agent @xmath51 was in a restaurant ranked @xmath8 at time @xmath6 and was not served then , at time @xmath78 , the agent goes to the restaurant ranked @xmath0 .
we consider the kpr problem where the decision made by each agent in each time period @xmath6 is independent and is based on the information about the rank @xmath8 of the restaurants and their occupancy given by the numbers @xmath84 .
we consider here in sec .
@xmath85 several stochastic strategies where each agent chooses the @xmath8-th ranked restaurant with probability @xmath7 given by eq .
( [ generalstoch ] ) .
the utilization fraction @xmath86 of the @xmath8-th ranked restaurants on every evening is studied and their average ( over @xmath8 ) distributions @xmath40 are shown in fig .
[ fig1 ] for some special cases . from numerical studies ,
we find their distributions to be gaussian with the most probable utilization fraction @xmath87 , @xmath88 and @xmath20 for the cases with @xmath14 , @xmath89 ; @xmath34 , @xmath89 ; and @xmath14 , @xmath39 respectively .
for the stochastic crowd - avoiding strategy discussed on sec .
@xmath90 , we get the best utilization fraction @xmath91 .
the analytical estimates for @xmath2 in these limits are also given and they agree very well with the numerical observations . finally , we suggest ways to achieve the fair social norm either exactly in the presence of incentive problem or asymptotically in the absence of such incentive problem . implementing or achieving such a norm in a decentralized way is impossible when @xmath92 limit .
the kpr problem has similarity with the minority game problem @xcite as in both the games , herding behavior is punished and diversity s encouraged .
also , both involves learning of the agents from the past successes etc . of course , kpr has some simple exact solution limits , a few of which are discussed here . in none of these cases considered here ,
learning strategies are individualistic ; rather all the agents choose following the probability given by eq .
( [ generalstoch ] ) . in a few different limits of such a learning strategy ,
the average utilization fraction @xmath2 and their distributions are obtained and compared with the analytic estimates , which are reasonably close .
needless to mention , the real challenge is to design algorithms of learning mixed strategies ( e.g. , from the pool discussed here ) by the agents so that the fair social norm emerges eventually even when every one decides on the basis of their own information independently .
as we have seen , some naive strategies give better values of @xmath2 compared to most of the `` smarter '' strategies like strict crowd - avoiding strategies ( sec @xmath93 ) etc .
this observation in fact compares well with earlier observation in minority games ( see e.g. , @xcite ) .
it may be noted that all the stochastic strategies , being parallel in computational mode , have the advantage that they converge to solution at smaller time steps ( @xmath94 or weakly dependent on @xmath0 ) while for deterministic strategies the convergence time is typically of order of @xmath0 , which renders such strategies useless in the truly macroscopic ( @xmath92 ) limits .
however , deterministic strategies are useful when @xmath0 is small and rational agents can design appropriate punishment schemes for the deviators ( see @xcite ) . in brief
, the study of the kpr problem shows that while a dictated solution leads to one of the best possible solution to the problem , with each agent getting his dinner at the best ranked restaurant with a period of @xmath0 evenings , and with best possible value of @xmath3 ( @xmath95 ) starting from the first evening itself .
the parallel decision strategies ( employing evolving algorithms by the agents , and past informations , e.g. , of @xmath96 ) , which are necessarily parallel among the agents and stochastic ( as in democracy ) , are less efficient ( @xmath97 ; the best one discussed here in sec . @xmath90 , giving @xmath98 only ) .
we also note that most of the `` smarter '' strategies lead to much lower efficiency .
is there an upper bound for the value of utilization fraction @xmath2 ( less than unity ; easily achieved in the dictated solution ) for such stochastic strategies employed in parallel ( democratically ) by the agents in kpr ?
if so , what is this upper bound value ? also , what is the learning time required to arrive at such a solution ( compared to zero waiting time to arriving at the most efficient dictated solution ) in kpr ?
these are the questions are to be investigated in future .
the authors would like to thank anindya sundar chakrabarti and satya ranjan chakravarty for useful comments and discussions .
suppose there are @xmath99 agents and @xmath0 restaurants .
an agents can select any restaurant with equal probability .
therefore , the probability that a single restaurant is chosen by @xmath24 agents is given by a poission distribution in the limit @xmath92 : @xmath100 therefore the fraction of restaurants not chosen by any agents is given by @xmath101 and that implies that average fraction of restaurants occupied on any evening is given by @xcite @xmath102 in the kpr problem .
in this case , an agent goes to the @xmath8-th ranked restaurant with probability @xmath103 ; that is , @xmath7 given by ( [ generalstoch ] ) in the limit @xmath34 , @xmath15 . starting with @xmath0 restaurants and @xmath0 agents , we make @xmath104 pairs of restaurants and each pair has restaurants ranked @xmath8 and @xmath105 where @xmath106 .
therefore , an agent chooses any pair of restaurant with uniform probability @xmath107 or @xmath0 agents chooses randomly from @xmath104 pairs of restaurants .
therefore the fraction of pairs selected by the agents ( from eq .
( [ eq : poisson_mm ] ) ) @xmath108 also , the expected number of restaurants occupied in a pair of restaurants with rank @xmath8 and @xmath105 by a pair of agents is @xmath109 therefore , the fraction of restaurants occupied by pairs of agents @xmath110 hence , the actual fraction of restaurants occupied by the agents is @xmath111
99 ghosh , a. , chakrabarti , a. s. , chakrabarti , b. k. , 2010 , _ kolkata paise restaurant problem in some uniform learning strategy limits _ , in econophysics & economis of games , social choices & quantitative techniques , new economic windows , eds . b. basu , b. k. chakrabarti , s. r. chakravarty , k. gangopadhyay , springer , milan , pages 3 - 9 . | we study the dynamics of a few stochastic learning strategies for the `` kolkata paise restaurant '' problem , where @xmath0 agents choose among @xmath0 equally priced but differently ranked restaurants every evening such that each agent tries get to dinner in the best restaurant ( each serving only one customer and the rest arriving there going without dinner that evening ) .
we consider the learning strategies to be similar for all the agents and assume that each follow the same probabilistic or stochastic strategy dependent on the information of the past successes in the game .
we show that some `` naive '' strategies lead to much better utilization of the services than some relatively `` smarter '' strategies .
we also show that the service utilization fraction as high as @xmath1 can result for a stochastic strategy , where each agent sticks to his past choice ( independent of success achieved or not ; with probability decreasing inversely in the past crowd size ) . the numerical results for utilization fraction of the services in some limiting cases are analytically examined . |
i am grateful to prof .
a.i.olemskoi for inspiring discussions and helpful comments .
* dependence of the probability @xmath119 of realization of various trajectories on the initial value of the order parameter @xmath111 for @xmath160 ( curves 1 , 2 correspond to @xmath158 @xmath157 , 0.6 , curves 3,4 correspond to @xmath159 @xmath157 , 0.6 .
the initial value of the conjugate momentum @xmath161 . | the behavior of the most probable values of the order parameter @xmath0 and the amplitude @xmath1 of conjugate force fluctuations is studied for a stochastic system with a colored multiplicative noise with absorbing states .
the phase diagrams introduced as dependencies the noise self - correlation time vs temperature and noise growth velocity are defined .
it is shown that phase half - plane @xmath2 can be split into isolated domains of large , intermediate , and small values of @xmath0 .
system behavior in these domains is studied by the probability represented as path integral . in the region
@xmath3 , the trajectories converge to the point @xmath4 for @xmath5 and to @xmath6 , @xmath7 for @xmath8 . in the former case ,
the probability of realization of trajectories is finite , while in the latter case it is vanishingly small , and an absorbing state can be formed . a wide class of stochastic systems with multiplicative noise whose intensity is a function of the stochastic variable @xmath0 determining the state of the subsystem can exhibit absorbing states .
an absorbing configuration is one in which a system can get trapped , from which it can not escape @xcite .
the most popular example of this kind is the verchlst model proposed for explaining the population kinetics @xcite ; the analysis of this model remains a topical question @xcite .
another example is the percolation model @xcite in which a linear dependence on @xmath0 is observed for noise intensity .
indicated models can be generalized by representing the noise intensity as a stochastic variable power dependence with an arbitrary exponent .
a kinetic features of stochastic system with that kind noise was studied in @xcite in the white noise approximation and with accounting inhomogeneity of the space .
an evolution of averages in that model was discussed in @xcite .
a distinguishing feature of our approach is that we shall study the dependencies of most probable values vs system parameters using supposition of the noise correlation .
this is realized on the basis of a field scheme in which the probability functional has an exponential form with the exponent that can be reduced ( correct to the sign ) to the standard action @xcite .
consequently , a description of phase transition in terms of the most probable values corresponds to the application of the least action principle .
such an approach makes it possible to extend the self - consistent field method to a description of systems with multiplicative noise .
this reveals the following nontrivial peculiarity : the most probable values can tend either to free energy minima or to points corresponding to unstable thermodynamic states depending on the system parameters .
the singular nature of noise indicating the tendency of its intensity to zero for the value @xmath6 of the order parameter ( more probably magnitude of @xmath0 ) is responsible for the emergence of a region on the phase portrait in which all trajectories converge to the axis @xmath6 over a finite time . in other words ,
an absorbing state is formed , in which the system becomes closed @xcite .
we consider following model .
the time dependence of the hydrodynamic mode amplitude @xmath9 is determined by the langevin equation @xmath10 here the dot indicates the differentiation with respect to time @xmath11 , deterministic force @xmath12 is defined according to the thermodynamic potential @xmath13 as @xmath14 , @xmath15 is the colored multiplicative noise with amplitude defined by function @xmath16 . a specific form for the noise @xmath15 we choose in kind of ornstein ulenbeck process , i.e. @xmath17 where @xmath18 is a self correlation time , @xmath19 is the gaussian distributed variable with zero mean and exponentially decaying correlations : @xmath20 @xmath21 is a white noise namely , gaussian stochastic variable with zero mean and @xmath22correlated : @xmath23 . the amplitude @xmath16 of the multiplicative noise in eq.([1 ] ) is chosen in the simple form @xmath24 where @xmath25 is a positive exponent varying from 0 to 1 @xcite .
we shall describe our system by master parameter , self correlation time and velocity of noise growth defined by exponent @xmath25 .
if we take time derivative of eq.([1 ] ) , replace first @xmath26 in terms of @xmath19 and @xmath27 from eq.([2 ] ) and then @xmath19 in terms of @xmath28 and @xmath0 from eq.([1 ] ) , we obtain the following non markovian stochastic differential equation ( sde ) ( see ref.@xcite ) : @xmath29 where @xmath30 \label{5a}\ ] ] ( throughout the paper , the it interpretation for the sde will be meant ) . in the limit @xmath31 a system behavior was described in @xcite . here , following `` unified colored noise approximation '' ref.@xcite we can recover a markovian sde .
it needs to use adiabatic elimination ( neglecting @xmath32 ) and to neglect @xmath33 so that the system s dynamics be governed by a fokker planck equation . in order to make correct transition to an ordinary , linear in @xmath28 ,
sde we have to use it s differential rule @xcite .
in above mentioned suppositions sde ( 5 ) takes form @xmath34 expression in lhs in eq.([6a ] ) defines time derivative for variable @xmath35 .
the relation between @xmath35 , and @xmath0 is set by expression @xmath36 .
an equivalent equation to eq.([6a ] ) can be write down as @xmath37 with @xmath38 here we stress that drift term @xmath39 appears only by accounting the noise self correlation time .
let us now go over to a new field @xmath40 connected with the initial field @xmath9 through the relation @xmath41 the new stochastic field taking into account eq.([6 ] ) satisfies the langevin equation @xmath42 with additive noise and effective force @xmath43 the obtained equation ( [ 8 ] ) allows us to use the standard field scheme @xcite based on analysis of the generating functional .
the latter has the form of the functional laplace transform @xmath44 for the partition function @xmath45 where @xmath46 denotes integration over all paths starting at @xmath47 for @xmath48 and ending at @xmath49 for @xmath50 .
it is defined as @xmath51 where @xmath52 is a field at time @xmath53 , having sliced the interval 0 to @xmath54 in @xmath55 parts of size @xmath56 .
the argument of the @xmath22-function in eq.([11 ] ) can be reduced to the langevin equation ( [ 8 ] ) , and the determinant ensuring a transition from continual integration with respect to @xmath21 to @xmath57 is equal to unity in it calculus . in the approach developed by zinn - justin @xcite , @xmath58fold variation of functional ( [ 10 ] ) over the auxiliary field @xmath59
makes it possible to find the @xmath58-th order correlator for the hydrodynamic mode amplitude @xmath40 and to construct the perturbation theory .
we shall proceed , however , from expression ( [ 11 ] ) for the conjugate functional @xmath60 whose variation leads to the most probable realization of the stochastic field @xmath40 .
going over to an analysis of functional ( [ 11 ] ) , we write the @xmath22-function in integral form @xmath61 averaging over the noise @xmath27 with the help of the gaussian distribution @xmath62 corresponding to ( [ 3 ] ) and taking into account eq.([12 ] ) , we reduce functional eq.([11 ] ) to the standard form @xmath63 here the probability distribution @xmath64 is specified by the action @xmath65 , where the lagrangian is given by @xmath66 we shall use in subsequent analysis the euler equations @xmath67 in which the dissipative function is defined as @xmath68 as a result , the equations for the most probable realizations of the stochastic fields @xmath40 and @xmath69 assume the form @xmath70 a comparison of eq.([18 ] ) with the stochastic equation ( [ 8 ] ) having the same form shows that the fields @xmath40 and @xmath69 are the most probable values of amplitudes of the auxiliary hydrodynamic mode defined by relation ( [ 6 ] ) as well as by fluctuation of the conjugate force .
obviously , the latter can be reduced to the conjugate momentum @xmath71 . in order to return to the initial stochastic field @xmath9
, we shall use the relation ( [ 6 ] ) and definition ( [ 9 ] ) . as a result ,
the lagrangian ( [ 15 ] ) and the dissipative function ( [ 17 ] ) assume the form @xmath72 @xmath73 where we have used the definition of the conjugate momentum @xmath74 leading to the relation @xmath75 in this case , the euler equations ( [ 16 ] ) assume the form @xmath76
@xmath77-(f+h)/g^2+\partial_x g/2g.\label{24}\ ] ] similar equations can be obtained directly from the system ( [ 18 ] ) , ( [ 19 ] ) using the relations ( [ 6 ] ) and ( [ 22 ] ) as well as definition ( [ 9 ] ) . in order to calculate noise self
correlation time influence on the system behavior we shall use the @xmath78-model for potential @xmath79 \label{26}\ ] ] and the definition ( [ 4 ] ) of multiplicative function . since the lagrangian ( [ 20 ] ) does not change its form upon simultaneous reversal of the signs of @xmath0 and @xmath1 , the phase portraits will possess central symmetry relative to the origin @xmath4 .
on the other hand , the axis @xmath6 on which the noise intensity assumes zero value is singular in accordance with ( [ 4 ] ) .
for this reason , we can confine our analysis only to the upper part of the phase plane corresponding to the value of @xmath80 .
first of all let us investigate steady states of the system .
in the stationary case @xmath81 , we have two equations @xmath82 @xmath83+{1\over 2 } ( \partial_x\ln\sigma)\partial_x\ln g_0-\frac{\partial_x^2g_0}{2g_0}\right\}=0.\ ] ] from eq.([29 ] ) it follows that at noise self
correlation time not exceeding the value @xmath84 ( fig.1 ) the form of the phase portrait is characterized by the presence of a single saddle point @xmath85 whose position is defined by the solution of eq.([30 ] ) and corresponds to @xmath86 .
figure 2 shows that bifurcation takes place at @xmath87 at noise self
correlation time @xmath84 ( fig.2a , 2b ) .
this is accompanied by the emergence of an additional saddle point @xmath85 and an attractive node @xmath88 whose positions are determined by the condition @xmath89 and specified by the coordinates @xmath90 . in fig.3 we show critical magnitude of the noise self
correlation time @xmath91 vs temperature @xmath92 ( as maximum of the dependence @xmath93 ) .
as shown in fig.1 , above @xmath91 the system is always ordered for any velocity of the noise growth , defined by exponent @xmath25 . obviously , the saddle point @xmath85 and the node @xmath88 merge at the point corresponding to the correlation time @xmath91 .
the above analysis shows that in systems with @xmath94 the stationary state corresponds to the point @xmath85 whose coordinates are defined as @xmath95 @xmath96 .\label{35b}\ ] ] the meaning of these coordinates becomes clear if we proceed to the additive limit @xmath97 : condition ( [ 35a ] ) indicates that the most probable value @xmath1 of the fluctuation amplitude of conjugate force for homogeneous systems has the sign opposite to that of @xmath98 ; according to eq.([35b ] ) , the `` susceptibility '' @xmath99 in this case assumes an infinitely large value .
thus , node @xmath85 corresponds to the stationary state of a thermodynamic system which is unstable to a transition to the ordered phase .
then the phase portrait has the form shown in fig.4a . at temperatures @xmath100 , the steady state with the coordinates @xmath101
@xmath102 is formed at the point @xmath88 . in the additive limit
, this point corresponds to the state of thermodynamic equilibrium .
figure 4 shows that the corresponding phase portrait can be obtained not only upon an decrease in temperature ( see fig.4a ) .
here we stress that the system can pass from equilibrium ( ordered ) domain trough unstable ( disordered ) to stable if we increase the velocity of the noise growth ( see fig.1 and fig.4b ) .
naturally , a real thermodynamic system in the process of its evolution tends to the equilibrium state eqs.([36a ] , [ 36b ] ) rather than to the unstable state eqs.([35a ] , [ 35b ] ) .
the equilibrium state corresponds to small and large values of the exponent @xmath25 for any values of the parameter @xmath18 .
the domain of medium values of the exponent @xmath25 at small noise correlations corresponds to unstable states .
comparing fig.4a , fig.4b we see that they are distinguished by the location of the attraction node on the axis @xmath89 .
the general form of phase portraits shown in fig.4 is characterized by the presence of two separatrices with branches @xmath103 and @xmath104 .
they divide the phase plane into three isolated regions corresponding to large , intermediate , and small values of the order parameter @xmath0 .
the first region is characterized by an indefinite increase in the values of the quantities @xmath0 and @xmath1 with time @xmath105 .
it will be sown below that this is not realized in actual practice .
the region of intermediate values of @xmath0 in which the system ultimately goes over to a stationary ordered state is most interesting .
it is this region that determines the phase transition kinetics .
the formation of the region corresponding to values @xmath106 is associated with the multiplicative nature of noise . in this region
, the order parameter @xmath9 tends with time to the value @xmath6 .
let us analyze the behavior of the system in each of these regions .
for this purpose , we consider the probability of realization of a phase trajectory corresponding to different initial values @xmath107 . in accordance with ( [ 14 ] ) , the probability can be written in the form @xmath108 where expressions ( [ 20 ] ) and ( [ 23 ] ) are taken into account and integration is carried out along the corresponding trajectory .
the dependence @xmath109 obtained for an exponent @xmath110 is shown in fig.5 ( curves 1 , 2 ) .
apart from the trivial increase in probability ( [ 37 ] ) which approaches the origin for values of @xmath111 corresponding to separatrices , the jumps near which the value of @xmath109 can increase insignificantly are observed . outside the region
bounded by the ( outer ) separatrix , we have @xmath112 since @xmath9 , @xmath113 for @xmath105 in this case .
such a behavior of the probability @xmath109 can be explained by the form of the time dependencies @xmath9 and @xmath114 during relaxation of the initial value .
far away from the region of @xmath115 ( see fig.4a ) , the quantities @xmath1 and @xmath0 rapidly change their values , the change slowing down as we approach this region .
such a behavior can be explained by the fact that the action @xmath116 changes much more slowly near this region than away from it .
this can be visualized by associating the region @xmath117 with the bed of a large river @xcite
. for values of the exponent @xmath118 , the integrand in formula ( [ 37 ] ) diverges , and the probability @xmath119 assumes zero value for @xmath120 .
typically , this divergence is observed only in the region of phase portrait bounded by the separatrix branch @xmath121 in fig.4b . in order to explain the form of the dependence @xmath109
, we analyze the behavior of the quantities @xmath9 and @xmath114 for various values of exponent @xmath25 . for this purpose
, we put @xmath122 in eq.([24 ] ) .
the obtained quadratic equation gives stationary values of conjugate momentum in the limit @xmath123 : @xmath124 thus , for @xmath110 , the system tends with time to the origin @xmath4 , and the attraction node jumps to infinity ( @xmath7 , @xmath125 ) as @xmath25 exceeds the critical value @xmath126 .
the corresponding integrand in distribution ( [ 37 ] ) , i.e. , @xmath127 is characterized by the sign inversion in the exponent upon a transition through the critical value @xmath126 . substituting eq.([38 ] ) for @xmath110 into the equation of motion (
[ 23 ] ) and retaining the leading term in it , we obtain the equation @xmath128 which gives the time dependence of the order parameter : @xmath129 where @xmath130 is the integration constant defining the time during which the point gets to the axis @xmath6 . the substitution of dependence eq.([40 ] ) into eq.([39 ] ) at @xmath110 and of the obtained expression into the integral in eq.([37 ] ) shows that the probability @xmath109 of realization of a trajectory in the region @xmath106 of the phase portrait differs from zero ( see fig.5 , curves 1 , 2 ) .
a completely different situation is observed for an exponent @xmath118 . in this case ,
eq.([23 ] ) can be reduced to @xmath131 , leading to eq.([40 ] ) in which the factor @xmath132 is replaced by @xmath133 . however , for @xmath134 the expression @xmath135 acquires an exponent with the opposite sign in accordance with eq.([39 ] ) so that it assumes an infinitely large value for @xmath136 .
as a result , the probability eq.([37 ] ) becomes vanishingly small ( see fig.5 , curves 3 , 4 ) .
the physical reason behind such a behavior is that the system gets to the axis @xmath6 over a finite time interval @xmath137 , which ensures an infinitely large value of the conjugate momentum @xmath138 .
this can be visualized as the precipitation of condensate ( absorbing state ) of configuration points from the phase portrait domain @xmath106 onto the abscissa axis for @xmath7 .
note that the condition @xmath137 is satisfied only below the separatrix branch @xmath121 in fig.4b , while in the region @xmath139 we have @xmath140 , and the convergence of the integrand in ( [ 37 ] ) is not manifested .
consequently , the equality @xmath112 holds only below the curve @xmath121 ( see fig.5 and fig.4b ) . in accordance with ( [ 40 ] ) , the time dependence of the most probable magnitude @xmath141 is defined by the exponent @xmath142 whose magnitude determines the fractional dimension @xmath143 characterizing the domain @xmath144 of the phase portrait of a system with multiplicative noise @xcite . in the additive limit @xmath145
, we have dimension @xmath146 of the phase plane as expected .
this means that as the time @xmath147 , the phase trajectories of the system fill the entire phase plane .
the increase of the exponent @xmath148 leads to a decrease in the fractional dimension @xmath149 which assumes the critical value @xmath150 for @xmath126 . as the value of @xmath25 increases further , the fractional dimension of the set of points on the plane @xmath151 , which is the law of motion @xmath9 , becomes smaller than unity .
the physical reason behind such a behavior is the above mentioned absorption of configuration points by the axis @xmath6 for @xmath7 .
finally , it must be stressed that the noise correlation time can play role of master parameter as the temperature and produces transition to the ordered state .
moreover we shown that the increasing of the velocity of the noise intensity growth produces appearance of the domain of disordered states .
this domain disappears if noise self - correlation time does not exceed value @xmath91 .
path integral solution shown that absorbing state appearance is characterized by the vanishingly small probability of the system states realization .
asymptotic time dependence of the order parameter @xmath0 explains picture of absorbing states by the fractal dimension smaller than unity . |
a number of investigations aimed to check absolute stability of half - life of different radioactive isotopes and to search for possible time variations of the half - life constants under action of known and possible unknown natural factors has been carried out during last years . in work @xcite it was shown that , after many years of investigations with scintillation and semi - conductor @xmath2-ray detectors , a conclusion was made about changes of rate of radioactive elements @xmath3-decay with 24 hours and 27 days periodicities .
many years measurement s decay rates data for the @xmath4si ( @xmath3-decay ) measured in the brookhaven national laboratory ( usa , 1982 - 1986 years ) and for the @xmath5ra ( @xmath6 and @xmath3-decays ) measured in the physikalisch - technische - bundesanstalt ( germany , 1983 - 1998 years ) were analyzed in work @xcite , @xcite , @xcite .
decay rate variations with a one year period and maximum amplitude of @xmath70.15% in january - february were found in both data sets .
the authors considered an assumed seasonal variations of the detector systems characteristics and/or the direct annular modulation of the count rate being caused by some unknown factor depending on the sun - earth distance as possible causes of such count rate variations .
the weak point of the count rate long time monitor experiments is a possible influence of meteorological , climatic and geophysical factors on the count rate of a source - detector pair
. this shortcoming could be practically totally avoided in measurements based on a direct registration of a nuclear life time between birth and decay .
such a method allows us to answer the question of possible change with time of a just nuclear decay constant . besides that a direct registration of nuclear life time allows us to study of the exponentially decay low . some theoretical models
@xcite , @xcite predicted that decay curve does not exactly follow an exponential low in the short - and long - time regions in particular due to of the so - called quantum zeno effect @xcite , @xcite , @xcite , @xcite .
experimentally zeno effect was proved @xcite in repeatedly measured two - level system undergoing rabi transitions , but not observed in spontaneous decays .
very important condition for measure possible deviations from an exponential low is the all investigated nuclei could have the same age .
a primary task of the investigation presented in this work was to investigate constancy of a half - life @xmath8 @xmath9 of @xmath0po during several years .
@xmath0po decays with 164.3 @xmath10s half - life @xcite by emitting the 7.687 mev @xmath2-particle .
this isotope appears mainly in the exited state ( @xmath11% ) in the @xmath0bi @xmath3-decay .
half - lives of the exited levels does not exceed 0.2 ps @xcite and they discharge instantly with regard to the scale of the @xmath0po half - life .
energies of the most intensive @xmath12-lines are equal to 609.3 kev ( 46.1% per decay ) , 1120 kev ( 15.0% ) and 1765 kev ( 15.9% ) .
so , the @xmath3-particle and @xmath12-quantum emitted at the moment of a birth of @xmath0po nuclear ( start ) and the @xmath2-particle emitted at the decay moment ( stop ) .
measurement of `` start - stop '' time intervals allows one to construct decay curve at an observation time and to determine the half - life from it s shape .
@xmath0bi isotope is an intermediate daughter product in the @xmath13u series .
it gives almost all @xmath12-activity of the series .
this isotope could be produced with a constant velocity if an intermediate isotope @xmath5ra ( @xmath14 y ) of the @xmath13u series will be used as a source .
a time of an equilibrium reached in a partial @xmath5ra-@xmath0bi series equals to @xmath15 days and is determined by the longest half - life isotope , @xmath16rn ( 3.82 days ) .
a @xmath0bi decay rate does nt change after this time if a source is hermetically sealed to prevent an escape of the radon .
two test sources with @xmath5ra activity of @xmath17 bq and @xmath15 bq were prepared in the v.g .
khlopin radium institute ( st.petersburg , russia ) in march , 2008 .
a thin transparent radium spot is deposited in the center of a polished plastic scintillator disc of 18 mm diameter and 0.8 mm thickness .
a disc s side with the spot is covered by a similar disc hermetically glued on periphery ( 2.5 mm width ) .
a plastic scintillator ( ps ) registers all charged particles generated in the decay series .
a ps light yield for an @xmath2-particle absorption is @xmath18 of that for the electron of the same energy . due to this fact @xmath2-spectra mix with @xmath3-spectra if a scintillator is thick enough to absorb a complete energy of an electron . to prevent this effect and separate these spectra
the scintillator discs were made thin enough so that electrons lose only part of their energy .
two setups named tau-1 and tau-2 were made to measure radiations of these sources .
a schematic sectional view of the tau-1 with the electronics is shown in fig.[fig1 ] .
schematic cross - sectional view of tau-1 setup and electronics block diagram.,width=313 ] the detector d1 contains the photomultiplier ( feu-85 ) which views the source disc through a plastic light guide .
a teflon reflector is placed under the source to improve light collection .
the whole assembly is packed in a stainless steel body .
the detector d1 is placed on the butt - end of the scintillation detector d2 with low background nai(tl ) crystal of 80 mm diameter and 160 mm length .
the crystal has a stainless still cover and a quartz entrance window .
the crystal with photomultiplier ( feu-110 ) is packed in a cooper body .
the two detectors are mounted vertically in a low background shield made from pb(10 cm ) + fe(15 cm ) + w(3 cm ) .
the tau-1 is located in the underground laboratory `` kapriz '' of the baksan neutrino observatory of inr ras at 1000 m of water equivalent depth .
a cosmic ray background is decreased by @xmath19 times due to the rock thickness in comparison with the ground surface one .
walls of the laboratory are covered by a low background concrete thus decreasing by @xmath20 times the @xmath12-ray background from a rock natural radioactivity .
total background suppression inside the shield is equal to @xmath21 times .
signals from the photomultipliers are read by charge sensitive preamplifiers ( csa ) and fed to two inputs of a digitizer ( digital oscilloscope - do ) la - n20 - 12pci which is inserted into a personal computer ( pc ) .
pulses are digitized with 6.25 mhz frequency .
the do pulse recording starts by a signal from the d2 which registered @xmath0bi decay s @xmath12-quanta .
a d2 signal opens a record of a sequence with 655.36 @xmath10s total duration where first 81.92 @xmath10s time is a `` prehistory '' and the last 573.44 @xmath10s is a `` history '' .
duration of a `` history '' exceeds the three @xmath0po half - lives .
a ps disc source in the tau-2 installation is fixed at the end of an air light guide having a smooth wall made of vm-2000 light reflecting film .
the light guide is put into 0.8 mm thick stainless still rectangular frame with inner dimensions of @xmath22 mm .
an open butt - end of the frame is welded into a bottom of cylindrical stainless still body of 45 mm diameter and 165 mm length where photomultiplier ( feu-85 ) placed .
so much for the detector d1 of the tau-2 .
two scintillation detectors ( d2a and d2b ) made of large nai(tl ) crystals ( @xmath23 mm ) are used for a registration of the @xmath12-quanta .
each crystal has a stainless still cover and a quartz entrance window .
photomultipliers ( feu-49 ) are used for the light registration .
the d1 are installed into a gap between d2a and d2b .
a scheme of pulse registration in the tau-2 is similar to the one of tau-1 but signals from d1a and d2b are preliminarily summed in additional summer .
the tau-2 is located in the low background room in the deepest underground laboratory nlgz-4900 of the bno inr ras at the 4900 m of water equivalent depth .
a cosmic ray background is decreased by @xmath24 times by the rock thickness in comparison with the ground surface one .
the room walls are made of polyethylene ( 25 cm ) + cd(0.1 cm ) + pb(15 cm ) .
the detectors are surrounded with additional lead shield of 15 cm thickness aimed to absorb a radiation from decays of @xmath0bi in the air being in equilibrium with the radon in the room .
working characteristics of each installation were tested preliminary by using usual multichannel analyzer ai-1024 . _ * tau-1 installation * _ an amplitude spectrum , from the d1 , collected during 600 s is shown in fig.[fig2 ] . d1 detector pulse amplitude spectrum collected at 600 s.,width=321 ]
d2 detector pulse amplitude spectra ( a ) of @xmath5ra source ( 30 min collection time ) and ( b ) of background normalized at 30 min ( 72 hours collection time).,width=321 ] integral count rate is equal to 179 s@xmath25 .
a wide peak above the 500 channel corresponds to @xmath0po @xmath2-particles .
a count rate within @xmath26 channels is equal to 21.2 s@xmath25 .
an amplitude spectrum , from the d2 , collected during 30 min is shown in fig.[fig3]a .
a main part of the spectrum is due to @xmath12-quanta from the source .
a spectrum of the d2 background was measured in 72 hours when the d1 was removed from the shield .
it s spectrum shown in the fig.3b .
an integral count rate in the spectrum in fig.[fig3]a is 4.90 s@xmath25 ( 1.64 s@xmath25 at the energy above 400 kev ) and that one in fig.[fig3]b is 0.20 s@xmath25 . _ * tau-2 installation * _ integral count rate of the d1 is equal to 70 s@xmath25 .
a background count rate of the d2a detector at the energy above 400 kev is equal to 2.3 s@xmath25 .
a total count rate with the source is equal to 5.0 s@xmath25 .
a light collection coefficient of the d1 in the tau-2 was found to be 0.51 of the one in the tau-1 .
the tau-1 and tau-2 installations were placed into underground low background shields to provide stable environmental conditions and a low cosmic ray background .
it allows us to decrease possible variations of electronics characteristics and to exclude random d1-d2 coincidences caused by d2 registrations of background s muons and @xmath12-quanta . a control of the do work in regimes of on - line event separation , data collection and visualization was carried out by specially designed pc program .
this program allows do to record only coinciding d1-d2 events , with two pulses in the d1 `` history '' and one pulse in the d2 `` history '' .
the last one is in prompt coincidence with the first d1 pulse .
an example of the event is shown in fig.[fig4 ] .
an example of coinciding d1-d2 events with two pulses in the d1 `` history '' ( blue ) and one pulse in the d2 `` history '' ( red).,width=283 ] a number of such events make up @xmath27% of the total amount .
their count rates are equal to @xmath28 s@xmath25 for tau-1 and @xmath29 s@xmath25 for tau-2 for the energy threshold of 400 kev in the channels of d2 detector .
spectra of the d2 pulses ( @xmath12-quanta ) and the second d1 pulses ( @xmath2-particles ) for the tau-1 are shown in fig.[fig5](a , b ) .
spectra of the d2 @xmath12-quantum pulses ( a ) and the second d1 @xmath2-particle pulses ( b ) of the tau-1 setup.,width=302 ] double peak of the fig.[fig5]b spectrum can be explained by different light output of the two ps disks .
the lower energy peak was found to move towards the beginning of the coordinates with time .
this effect could be connected with a degradation of ps surface scintillations properties under the active spot due to diffusion of chemical components or decay products into a surface layer of the source deposited disc .
a rate of the direct data record is 5.8 gb@xmath30day@xmath25 .
such large information flux complicates processes of the data collection and off - line processing .
a number of pulses and their delays for each event are analysed on - line to decrease usage of pc memory .
`` wrong '' events are excluded .
only a time of appearing and amplitude of pulses defined for the `` right '' events record at the pc memory . a rate of an information accumulation decreased up to 8.5 mb@xmath30day@xmath25 .
measurements with the tau-1 setup started in april , 2008 and the ones with tau-2 started in july , 2009 .
a continuous data set is divided on a sequence of equal intervals .
a spectrum of delays between the first and second d1 pulses is constructed for each interval .
this spectrum is approximated by exponential function viewed as @xmath31 corresponding to the decay curve .
a value of the half - life @xmath32 is found from this curve .
the series of @xmath32-values is analysed to find out any regularity or variation .
duration of the analyzed data sets of tau-1 and tau-2 were equal to 1038 days and 562 days correspondingly . a half - life value obtained for a decay curve constructed with the whole set of tau-1 data was found to be @xmath33 @xmath10s .
this curve is shown in fig.[fig6 ] .
summary exponent constructed for the whole data set collected at 1038 days by tau-1.,width=283 ] a half - life value obtained for the tau-2 data was found to be @xmath34 @xmath10s .
the difference between values is connected with uncertainties of the digitizer s frequency calibrations .
each interval of the whole data set was chosen to be 7 days for the analysis of the half - life constant behavior for a long time period .
sequences of @xmath32-values for tau-1 and tau-2 are shown in fig.[fig7 ] ( a ) and ( b ) respectively .
time dependence of @xmath0po half - life determined for one week collection interval data set from tau-1 ( a ) and tau-2 ( b ) .
, width=283 ] dependence of @xmath35 for the tau-1 data.,width=283 ] to search for a possible annular variations of the data they were normalized on the averaged values and compared with a periodical function @xmath36 $ ] where @xmath37 day@xmath25 , @xmath38 and @xmath39 are the amplitude and phase correspondingly .
@xmath40 was taken at the beginning .
a @xmath39-value has been changed from 1 to 365 with the step of 1 day .
a correlation coefficient @xmath41 between the @xmath32-sequences and @xmath42 was calculated for the each @xmath39-value .
a dependence of @xmath35 for the tau-1 data is shown in fig.[fig8 ] .
normalized graph of the time dependence of @xmath0po half - life for tau-2 in comparison with function @xmath43 $ ] , @xmath44 - days.,width=264 ] normalized graph of the time dependence of @xmath0po half - life for tau-2 in comparison with function @xmath43 $ ] , @xmath44 - days.,width=264 ] the maximum value @xmath45 has been reached at @xmath46 .
a selection of @xmath38-value corresponding to the @xmath47 minimum was done for this @xmath39-value , and was found to be @xmath48 .
the values of @xmath49 , @xmath48 and @xmath50 days were found for tau-2 by the similar way .
corresponding graphs for the tau-1 and tau-2 data are shown in fig.[fig9 ] and fig.[fig10 ] .
the estimated values of @xmath38 are the same for both setups but the phases differ by 167 days .
a mean square error @xmath51 of a @xmath41 for a number @xmath52 of degrees of freedom could be estimated as @xmath53 .
it was found that @xmath54 for the tau-1 ( @xmath45 and @xmath55 ) and @xmath56 for the tau-2 ( @xmath57 and @xmath58 ) . the decay dependence of @xmath0po of the detector tau-2 .
points is the experimental data . curve _
1 _ is the result of the approximation of the exponential low . curve _
2 _ is the result of the aproximation using eq.1,width=264 ] a low reliability of the obtained results and disagreement between @xmath39-values allow one to determine only the upper limit of the amplitude of @xmath32-constant possible annular variation which was found to be @xmath59 ( 90% c.l . ) .
an interval of data spacing was chosen to be 1 day in order to search for possible short period variations .
no significant frequency peaks in a range @xmath60 day@xmath25 of the fourier spectra constructed by using two such @xmath61 - series has been found .
we have tested the theoretical consideration of possible violation of the radioactive decay exponential law at low values of time of life @xmath62 of @xmath0po nucleus . for this purpose the total data set of the detector tau-2 was used .
the exponent was constructed for the time interval @xmath63 @xmath10sec .
half - live of the @xmath0po nucleus was determined using data for the time interval @xmath64 @xmath10sec .
then the residual part of the data was used for analysis of the deviations between the data and the exponent which was calculated .
the deviation was search for in form @xcite : @xmath65 limiting n=2 we find : a@xmath66 @xmath67 4.01@xmath3010@xmath68 ,
a@xmath69 @xmath67 3@xmath3010@xmath70 , for time interval @xmath71 . the results of the analysis are shown at the fig.11 .
the points at the figure are the experimental data for the time interval @xmath72 @xmath10sec . curve _
1 _ is the part of the exponent calculated for time interval @xmath73 @xmath10sec . curve _ 2 _ is the result of the approximation of the data for time interval @xmath74 @xmath10sec , using eq.([eq1 ] ) .
two underground installations aimed at monitoring the time stability of @xmath0po half - life have been constructed .
baurov , yu.g .
sobolev , yu.v .
ryabov , v.f .
kushniruk , `` experimental investigations of changes in the rate of beta decay of radioactive elements '' .
phys.of atomic nucl .
v.*70 * , no.11 , 1825 ( 2007 ) .
jere h. jenkins at al .
, `` analysis of experiments exhibiting time - varying nuclear decay rates : systematic effects or new physics ? '' .
arxiv:1106.1678 [ nucl - ex ] ; `` evidence for time - varying nuclear decay rates : experimental results and their implications for new physics '' arxiv:1106.1470 [ nucl - ex ] . | a method and results of an experimental test of the time stability of the half - life of alpha - decay @xmath0po nuclei are presented .
two underground installations aimed at monitoring the time stability have been constructed .
time of measurement exceeds 1038 days for one set up and 562 days for the other .
it was found that amplitude of possible annual variation of @xmath0po half - life does not exceed 0.2% of the mean value .
the limit on the deviation of the decay curve from exponent at @xmath1 range has been found . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.